United States Office of Air Quality
Environmental Protection Planning and Standards
Agency Research Triangle Park NC 27711
EPA-450/3-81-007
May 1981
Air
Summary of Technical
Information for Selected
Volatile Organic
Compound Source
Categories
-------
EPA-450/3-81-007
Summary of Technical Information
For Selected Volatile Organic
Compound Source Categories
by
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
May 1981
-------
This report was furnished to the Environmental Protection Agency (EPA) by
Engineering-Science, 125 West Huntington Drive, Arcadia, California 91006
in partial fulfillment of contract No. 68-01-4146. This report has been
reviewed by the Emission Standards and Engineering Division of the Office
of Air Quality Planning and Standards, and approved for publication.
Approval does not signify that the contents necessarily reflect the views
and policies of the EPA and mention of trade names or commercial products
is not intended to constitute endorsement or recommendation for use. Copies
of this report are available through the Library Services Office (MD-35),
U. S. Environmental Protection Agency, Research Triangle Park, N. C. 27711,
or from National Technical Information Services, 5285 Port Royal Road,
Springfield, Virginia 22161.
Publication No. EPA-450/3-81-007
-------
ABSTRACT
This document is a compilation of existing information of varying
detail based on a review of the technical literature, published and
unpublished EPA and state or local agency reports, and state and local air
pollution control regulations concerning various sources of volatile organic
compound emissions. Because of its relatively limited treatment of these
sources, this document should not be accorded the status of a Control
Techniques Guideline (CTG) document.
This document may, however, serve in providing basic information on
the various processes and emission points and, as such, may be a starting
point for a state or local agency in considering VOC control on a given
industry. Agencies are cautioned not to write a RACT regulation based
solely on the technical material presented.
iii
-------
TABLE OF CONTENTS
Abstract
List of Figures
List of Tables
iii
viii
x
SECTION 1
ADHESIVES APPLICATION
Processes and Emissions
Control Technology
Status of Regulatory Activities
References
SECTION 2 ASPHALT AIR BLOWING
Processes and Emissions
Control Technology
Status of Regulatory Activities
References
SECTION 3
BARGE AND TANKER CLEANING
Processes and Emissions
Control Technology
Status of Regulatory Activities
SECTION 4 BARGE AND TANKER LOADING OF VOLATILE
ORGANIC LIQUIDS
Processes and Emissions
VOC Loading Operations
Control Technology
Regulatory Status
References
SECTION 5 BEER MAKING
Processes and Emissions
Status of Regulatory Activities
References
SECTION 6 FABRIC PRINTING
Processes and Emissions
Control Technology
Regulatory Status
References
SECTION 7 FLARES
Processes and Emissions
Control Technology
Status of Regulatory Activities
References
1-1
1-1
1-11
1-12
1-12
2-1
2-1
2-3
2-9
2-10
3-1
3-1
3-3
3-4
4-1
4-1
4-3
4-5
4-17
4-17
5-1
5-1
5-2
5-2
6-1
6-1
6-8
6-12
6-13
7-1
7-1
7-13
7-14
7-18
-------
TABLE OF CONTENTS—Continued
SECTION 8 LUBE OIL MANUFACTURING 8-1
Processes and Emissions 8-1
Control Technology 8-9
Status of Regulatory Activities 8-9
References 8-10
SECTION 9 OIL AND GAS PRODUCTION STORAGE TANKS 9-1
Processes and Emissions 9-1
Emissions 9-7
Control Technology 9-9
Status of Regulatory Activities 9-13
References 9-13
Appendix (Ventura County Air Pollution Control
District, State of California, Rule 71) 9-15
SECTION 10 PETROLEUM COKING PROCESSES 10-1
Processes and Emissions 10-1
Control Technology 10-3
Status of Regulatory Activities 10-3
References 10-3
SECTION 11 PLASTIC PARTS PAINTING 11-1
Processes and Emissions 11-1
Control Technology 11-2
Status of Regulatory Activities 11-3
References 11-4
SECTION 12 RAILROAD TANK CAR LOADING OF VOLATILE ORGANIC
LIQUIDS 12-1
Process and Emissions 12-1
Control Technology 12-13
Status of Regulatory Activities 12-27
References 12-30
SECTION 13 SOLVENT EXTRACTION PROCESSES 13-1
Processes and Emissions 13-1
Control Technology 13-4
Status of Regulatory Activities 13-8
References 13-9
SECTION 14 SURFACE COATING OF LARGE AIRCRAFT 14-1
Processes and Emissions 14-1
Control Technology 14-4
Status of Regulatory Activities 14-5
References 14-8
vi
-------
TABLE OF CONTENTS—Continued
SECTION 15 SURFACE COATING OF LARGE SHIPS
Processes and Emissions
Control Technology
Regulatory Status
References
SECTION 16 SURFACE COATING OF WOOD FURNITURE
Processes and Emissions
Control Technology
Status of Regulatory Activities
References
SECTION 17 WASTE SOLVENT RECOVERY INDUSTRY
Process and Emissions
Control Technology and Associated Costs
Regulatory Status
References
SECTION 18 WINE MAKING
Processes and Emissions
Control Technology
Status of Regulatory Activities
References
SECTION 19 STYRENE-BUTADLENE COPOLYMER LATEX
Processes and Emissions
Control Technology
Regulatory Status
References
15-1
15-1
15-3
15-7
15-8
16-1
16-1
16-10
16-18
16-20
17-1
17-1
17-4
17-6
17-6
18-1
18-1
18-4
18-6
18-6
19-1
19-1
19-7
19-13
19-15
vii
-------
LIST OF FIGURES
1-1 Viscosity-Concentration Relationship of Corn
Dextrin at 80°F 1-5
2-1 Asphalt Air Blowing 2-2
2-2 Schematic of Dilute Fume Incineration 2-5
2-3 Coupled Effects of Time and Temperature
on Rate of Pollutant Oxidation 2-6
2-4 Typical Effect of Operating Temperature on
Effectiveness of Thermal Afterburner 2-7
4-1 Model Marine Terminal 4-8
4-2 Refrigeration Vapor Recovery Module 4-9
4-3 Carbon Adsorption Module 4-11
4-4 Lean-Oil Absorption 4-12
4-5 Thermal Incinerator Module 4-14
5-1 Beer Making Flow Schematic 5-3
6-1 Roller Print Machine 6-4
6-2 Rotary Screen Print Machine 6-4
6-3 Roller Printing Line and Associated Steam Can
Drying Process 6-9
6-4 Rotary Screen Printing Line and Associated
Drying and Curing Oven 6-9
7-1 Diagram of Steam-Assisted Smokeless Elevated-
Flare System 7-3
7-2 Typical Modern Refinery Slowdown System 7-4
7-3 Different Designs of Elevated Flare Heads 7-6
7-4 Waste-Gas Flare System Using ESSO-Type Burner
Regulated with slotted Orifice 7-8
7-5 Diagram of Waste-Gas Flare System Using a
Sinclair Burner Using In-Line Orifice for
Regulation 7-9
7-6 Typical Venturi Ground Flare 7-10
7-7 Flow Diagram of Multi-Jet Flare System 7-11
7-8 Schematic of an Air Assisted Smokeless Flare 7-12
7-9 Cost Effectiveness of VOC Destroyed by an
Elevated-Flare System 7-15
7-10 Ground-Flare System Gross Annual Operating
Cost 7-16
7-11 Cost Effectiveness of VOC Destroyed by a Fuel-
Gas System 7-17
8-1 Schematic Diagram of a Refinery for Producing
Lubricating Oils 8-2
viii
-------
LIST OF FIGURES—Continued
8-2 Lube Oil Hydrotreating 8-5
8-3 Flowchart of Lube Oil Refining by Furfural
Extraction 8-7
8-4 Typical Solvent Refining Process Employing
Furfural and Provided with Solvent-Water
Separation and Recovery 8-7
9-1 Production Operations and Associated Emissions 9-2
9-2 Typical Tank Sequence in Tank Battery 9-5
9-3 Example Operations of Fixed-Roof Tanks 9-6
10-1 Delayed Coking Process 10-2
12-1 General Service Railroad Tank Car 12-5
12-2 Small Railroad Tank Car Loading Terminal 12-7
12-3 Magnetic Liquid Level Measuring Device 12-11
12-4 Vapor Balance in VOL Loading into Liquified
Gas Service Tank Car 12-16
12-5 VOC Vapor Collection by Suction Pipe During
Open Hatch Tank Car Loading of VOLs 12-17
12-6 VOC Vapor Collection from Top Unloading Piping
of General Service Tank Car 12-18
12-7 Dilution of Collected VOCs with Air During
VOL Loading of Railroad Tank Cars 12-21
12-8 Thermal Incinerator 12-23
12-9 Elevated Flare 12-25
12-10 Cost Effectiveness for Control of Acetone
Emissions during Railroad Tank Car Loading 12-29
13-1 Schematic of a Soybean Processing Plant -
VOC Emission Sources 13-2
14-1 Schematic of Disc Type Electrostatic Spraying
Installation 14-5
14-2 Schematic View of a Bell Type Electrostatic
Spraying Installation
17-1 Simplified Flow Pattern Through Typical Solvent
Recovery Plant 17-2
18-1 Rate of Loss of Alcohol Entrained in Carbon
Dioxide during Stirred Laboratory Scale
Fermentation with Temperature Maintained
at 34°C 18-3
18-2 Rate of Loss of Alcohol Entrained in Carbon
Dioxide during Stirred Laboratory Scale
Fermentation with Temperature Maintained
at 21°C 18-3
19-1 Schematic Flow Diagram for Latex Production by
Emulsion Polymerization 19-4
ix
-------
LIST OF TABLES
1-1 Major Uses of Various Adhesives 1-3
1-2 Preliminary Draft of Industry Classifications,
Estimates of Adhesives Consumed, and
Solvent Emissions 1-11
1-3 Summary of Tentative Priority Areas 1-12
4-1 Hydrocarbon Emission Factors for Gasoline
Loading Operations 4-6
6-1 Model Printing Line Parameters 6-9
6-2 Control Costs for Thermal Incineration Systems
With 70 Percent Recuperative Heat Recovery
Capabilities 6-12
6-3 Summary of State VOC Emission Regulations 6-14
8-1 Treatment of Oils 8-4
9-1 Major Tank Battery Processing Equipment 9-4
9-2 Crude Oil Storage Tank Uncontrolled and
Controlled VOC Emissions 9-12
9-3 Crude Oil Storage Tank Control Costs 9-12
12-1 Quantities of Top Ten VOLs Loaded in Railroad
Tank Cars - 1978 12-3
12-2 Total VOL Tank Car Loadings and Emissions - 1978 12-14
12-3 Examples of Costs and Cost Effectiveness for
Thermal Incineration of VOCs from Railroad
Tankcar Loading 12-28
13-1 Production of Soybean Oil in U.S. 13-5
13-2 Composition of Vent Streams from Soybean
Processing Plants 13-5
15-1 Marine Coatings: Types, Applications, and
Solids Contents 15-2
15-2 Summary of Marine Coating Sales and Resultant
Emissions in California, 1976 15-4
15-3 Comparative Costs of Six Coatings Systems
(Base Year 1976) 15-6
16-1 Wood Furniture Industry Structure 16-2
16-2 VOC Emissions from Industrial Coatings 16-3
16-3 Wood Furniture Categories 16-4
16-4 Wood Household Furniture Plants by EPA Region 16-5
16-5 Typical Wood Furniture Finishing Schedule 16-7
16-6 Relative Emissions from a Typical Conventional
Furniture Coating System 16-9
16-7 Approximate Solid Content of Wood Furniture
Finishes 16-11
x
-------
LIST OF TABLES—Continued
17-1 Typical Solvents Recovered 17-3
17-2 Emission Factors for Solvent Reclaiming 17-5
19-1 Domestic Producers of Styrene-Butadiene Latex 19-2
19-2 Emulsion Latex Model Plant 19-5
19-3 Annualized Costs of Implementing RACT • 19-11
19-4 Summary of State VOC Emission Regulations 19-14
xi
-------
SECTION 1
ADHESIVES APPLICATION
PROCESSES AND EMISSIONS
Aside from the introduction of rubber and pyroxylin cements 100
years ago, there was little advance in adhesives technology until the
twentieth century.
Adhesives may be either either organic, inorganic, or hybrids.
The organic materials are classified according to origin as:
(1) Natural: starch, dextrins, asphalt, animal and vegetable
proteins, natural rubber, shellac.
(2) Semi-synthetic: cellulose nitrate and the other cellu-
losics, polyamides derived from dimer acids, castor-oil
based polyurethanes. We may look to this group for many
new products in the future.
(3) Synthetics:
(a) Vinyl-type addition polymers, both resins and elastomers:
polyvinyl acetate, polyvinyl alcohol, acrylics, unsatur-
ated polyesters, butadiene/acrylonitrile, butadiene/
styrene, neoprene, butyl rubber, polyisobutylene.
(b) Polymers formed by condensation and other step-wise
mechanisms: epoxies, polyurethanes, polysulfide rub-
bers, and the reaction products of formaldehyde with
phenol, resorcinol, urea, and melamine.
Alternatively, adhesives may be categorized according to the
solubility and fusibility of the final glue line:
(1) Soluble, including thermoplastic (soluble and fusible):
starch and derivatives, asphalts, some proteins, cellulosics,
vinyls, acrylics.
1-1
-------
(2) Thermosetting (insoluble and infusible): phenol- and
resorcinol-formaldehyde, urea- and melamineformaldehyde,
epoxies, polyurethanes, natural and synthetic rubbers if
vulcanized.
Table 1-1 on the following page lists various adhesives and their
industrial uses.
In the context of this report, the discussion of adhesives is
included because of the potential for VOC emissions from organic sol-
ventborne adhesives. Thus, further discussion will be limited to
this latter class of adhesives, substitutes for them, and to control
technology. For the most part the natural adhesives, particularly
those derived from starch, dextrins , animal and vegetable proteins and
shellac, are not directly competitive with organic solvent-borne
synthetic adhesives and will not be discussed further. Natural
rubber-based adhesives are organic solvent-borne and will be included
in the discussion of synthetic polymer type solvent-borne adhesives.
Rubber type adhesives are still used to bind paper, rubber, plastic
films, leather, wood, ceramic, plastic tile, metals, etc. to each
other and to other materials.
Hot-Melt Adhesives
Hot-melt adhesives are generally sought for economic reasons with
regard to the application process, since it is simply or quickly
applied and lends itself to mechanization. If these factors are
unimportant, a solvent cement or a thermosetting-type adhesive pro-
vides more durability and greater strength.
The sequence of operations in applying hot-melt adhesives is a
rapid one, hence hand application is rare.
There is a wide variety of low-molecular-weight, natural and
synthetic waxes and resins suited to hot-melts when formulated. Among
these are:
Coumarone-indene resins
Resin and its derivatives
1-2
-------
en
£
hH
en
w
33
o
<3
en
rH P
1 0
^J 1 1
r^ p™i
B!
w <
.j >
CO
<3 fa
H O
en
W
en
:=
«
O
r>
^
^
•a
CO rJ
0) • O
1— I r-H CJ
•H 0
4J C 0)
X rH rl
0) i-l
H H
J-i <->
rl 0) •
01 * CX (50
CX CO cfl J", ^
CO 4-> CM 4J O-i
(Xi O O «
3 T3 rH rH
• *a a) o u
o o s x
en iu s ^ ai
•H O-i 3 ^— '
2 o
1 T3
>. O
rH O
a. s
•
00
^!
a,
^
CU •
,£• en
4J T3
co O
CU *-l
i-J Oi
01
J-I
3
4J
•H
(3
t-l
3
fa
en
CO
co >.
rH 4.) CO
O 0) en
<4H CO
!J CO rH
0) en O
^3
•H ^3
fa
00
1 rH C
O CO -H
3 t3 -H 4J
J-t rl Cfl
4J C 4-1 O
en o en -H
C -H 3 w
O 4-> 13 ,O
u e co
M fn
eo
C
J -H CO
*rH 4J »H
W O >-l
Cfl -H CU
M ^4 -W
JO fa CO
<: s
ta
0)
>
iH
CO 0)
0) CX
£. >,
T3 H
<
X XXXXX XXX
XXXX XX XX XXX
X XXX X
XXXXXX XX X XX
X XXXX X
X X XX XXXX
XX XXXXXXX XX
X XX XXXXXXX XX
XXX XXX
XX X XX XX
s-\
0)
CD
S-l
U-l
rH CO CO
•rH o) en >-i co
O3 u30>cncnrH 13
en rH •H4J4JcoairiO C c
CU »00 Cfl cflcflOTSOlC 01 cfli-i
4-1 rH 4-> OCflCC'H-H4J-HenGC S-lOl
CfleOOrHrH CrHCU COrHECOCJ>jOc0434Jp Cfl
rHT3CcOcOT3-H3-H -P^OCflOlricUCJCUOXcU rHcn
>-,!>i •HSj3Oi>1>OJ3i-l43M(UMC8>,,COQ >-,0) C ><
CJrH CeOf— IC001CX3CD43OO013-HO-I-1 4-1 M -H tfl
<< <<3cacjuwOrHas
rH
CU
a
C
01
SJ
01
UH
O)
oi
1-3
-------
Mineral, vegetable, and petroleum waxes
Alkyds
Terpene resins
Heat-stable phenol-formaldehyde resins
All of these waxes and resins have low strength and melt easily to
low viscosity fluids. They are reinforced or toughened by blending
with limited proportions of higher molecular weight polymers, in order
to be converted into useful adhesives. These higher molecular weight
polymers include the following:
Ethyl cellulose
Polyvinyl acetate and its derivatives
Butyl methacrylates
Polyethylene
Polystyrene and styrene copolymers
Polyisobutylene
The ideal hot-melt adhesive is solid at room temperature, capable
of being stored and handled easily without blocking, and is light in
color. When a heat source is applied, it melts sharply and flows
freely. Also, it is stable during prolonged heating and able to
withstand local overheating.
Hot-melts are available in many different forms such as: tapes or
ribbons, films or thin sheets, granules, pellets in various shapes such
as cylinders and cubes, block and cordlike formations.
Hot-Melt Adhesive Application
Many different methods exist for the application of hot-melt
adhesives. Two common methods are doctored rolls and intaglio tech-
niques. Individual pieces are frequently dipped.
Over the past few years, a system which uses an extruder as a
melting and feeding device or heat pump has been widely used. The
latest system, also widely in use, is a concept which incorporates a
hot-melt adhesive in cordlike form and an applicator to melt and apply
it.
1-4
-------
There are two principal types of equipment for the application of
hot-melt adhesives: melt reservoir and "progressive feed". A melt
reservoir involves a quantity of adhesive melted in a pot and delivered
by a metering pump from the pot to a heated nozzle. The adhesive is in
the form of blocks, chips, or granules. The melting pot is subjected
to a charge, then heated to a predetermined temperature under thermo-
static control, and the melt is fed to a nozzle or cementing wheel by
a pump. The components are immediately bonded together with pressure.
Only a minimum quantity of hot-melts must be molten at any one
time with the "progressive feed" application method. This is because
the adhesive is fed continuously in the form of an elongated cord. The
adhesive is supplied as a flexible, grooved cylindrical cord coiled on
reels. The delivery rate of the melt through the nozzle is synchronized
with the rate of feed of the cord. Between the feed and delivery
point, the adhesive passes around a heated melt wheel running in an
eccentric groove. The melt chamber is formed by the small tubular
space between the wheel and the casing. Since the internal capacity
of the unit is so small, only a few grams of material are held above
the melt temperature inside the applicator. Using thermostatically
controlled heating elements located at carefully selected points enables
the adhesive to be maintained at its maximum application temperature
without overheating.
Uses of Hot-Melts2)
There are many diverse applications of hot-melts. As production
speeds increase, economics of hot-melts will become more favorable, and
new uses will certainly be found.
At present, the most common uses are the following:
1. Protective coating of paper, cloth, foil, and plastic film.
2. Lamination of the same materials.
3. Structural bonding of paper, wood, and other materials.
4. Pick-up gums and spot labeling.
5. Bonding of ceramics, cork, and metal.
1-5
-------
6. Production of articles made from paper, foil or film webs
including several forms of containers and packages.
With modern application equipment using "progressive feed", and
especially with the wide range of adhesive formulations available in
cord-like form, a number of varied bond properties have been made
possible. Some of these are:
(1) Effective bonding of flexible package at -40°F.
(2) Oil-resistive bonds retaining strength on papers with 4 to 6
percent mineral oil.
(3) Effective bonding of lightly waxed surfaces.
(4) Strong uniform bonding and sealing of polyethylene films and
coatings.
(5) Strong foil bonding.
(6) Fast, effective bonding of leathers.
Process Emissions and Emission Points from Hot-Melt Adhesives
According to a contact at Bostik Chemical, Mittleton, Massachu-
setts, the hot-melt adhesive application method is the most widely
practiced presently, and one of the reasons for this is the absence of
volatile organic compound emissions. In the future, water-base ad-
hesives will become increasingly popular for the same reason.
The only time that organic emissions are a possible threat during
hot-melt adhesive application is in the event of accidental over-heating
of the material, resulting in thermal decomposition.
Solvent Adhesives
Approximately 75 percent or more of all rubber-based adhesives are
used in the form of solvent cements^-'. Noncrystalline, amorphous-type
thermoplastics can be joined to themselves by applying the appropriate
solvents, solvent-polymer solutions, or monomer compositions. Polymeric
adhesives are generally the most likely choice for the bonding of
dissimilar plastics.
1-6
-------
Solvent Adhesive Application
The following are methods of application commonly used for solvent
cementing of similar plastics.
Brush Application - Cement is brushed onto both adherends, and the
parts are joined while wet. As in all solvent cementing, it is usually
preferable to jig the parts in place so that they are in intimate
contact until the bond is firmly set. Last traces of high-boiling
solvents will often take several days or more to completely evaporate.
Therefore, the joint should not be subjected to very high stresses
before the setting is completed. It is sometimes desirable to mask the
areas immediately adjacent to the joint to be made, since running cement
will ruin the appearance of the plastic or other material being glued.
Spraying - This method is often used when either large or numerous
pieces are to be cemented. Uniform application of cement is a prime
advantage of this method but masking of the plastic is usually required.
Adequate ventilation is necessary for protection of personnel. Although
spraying provides a rapid means of application, more cement is often
used than when brush coating-*-).
Dipping - Surfaces to be bonded are often dipped into the cement
for periods ranging from 1 second to more than 30 minutes, depending on
the plastic and cement used. The surfaces to be butt-joined are immersed
in the cement until softened, then removed and brought into contact.
Felt Pad - This method involves the use of a thick felt pad
partially immersed in a pan of solvent. The solvent "wicks" through
the felt fibers by capillary action so that the surface of the pad is
constantly wet. Plastic surfaces to be joined are brought into contact
with the pad until properly softened. This method is especially suit-
able for quick assembly where extremely strong bonds are not a prime
criterion and masking is undesirable.
1-7
-------
Hypodermic Syringes - These are sometimes used to introduce cement
into small, otherwise inaccessible areas where a bonded joint is
desirable-'-).
Process Emissions and Emission Points for Solvent-Based Adhesives
VOC emissions from solvent-based adhesives result from the evapor-
ation of solvents in the adhesive as applied. These include aliphatic
and aromatic hydrocarbons, alcohols, and ketones for the most part.
As in the case of surface coatings, the solvent content of adhesives
can constitute the major portion of the adhesive-solvent mixture.
Emissions arise mainly at the point of application and in many cases
are removed with local ventilation systems. Some of the higher boiling
solvents can continue to be evaporated with after removal of the bonded
products from the application area. Tables 1-2 and 1-3 present the
major adhesive consumers and resulting volatile organic emissions.
Adhesives and the Shoe Industry and Associated Emissions
The .footwear industry is a major user of adhesives for the per-
manent bonding of leather and man-made materials. The adhesives are
mostly neoprene (aliphatic, aromatic hydrocarbons) or urethane (ketones
such as acetone or MEK).
These glues are applied automatically (extruded or rolled) or they
are brushed on by hand. In what is referred to as the "fitting" room,
latex (a water-base adhesive) is used. This is one of the few water-
based adhesives utilized since, according to a contact at Red Wing
Shoes, the water-based glues dry very slowly and do not penetrate the
leather very effectively.
Until adhesive technology advances substantially toward improved
water-based glues, the shoe industry must depend on solvent-based
adhesives-^). In limited instances, hot-melt adhesives are used to
assemble various parts of shoes"). The volatile organic compounds
generated at the Red Wing Shoe manufacturing facility in Minnesota
are vented to the atmosphere. They are not subjected to an after-
burner since the sources of VOCs are spread out through the operation-^) .
1-8
-------
CVJ
-H
W
rJ
pa
<<
H
CO
w
<3
1— 1
H
CO
w
M
CO
z
O
H
^ti
u
M
JV .
M
CO
CO
cj
S-i
H
CO
0
Q
S
L 1
ft*
O
r i
•S
2
Q
pa
0
CO
z
<'
eT
w
J§
CO
o
CJ
CO
i_j
CO
W
1
fit
O
c
01
3
"c
CA
C
5
i-H
O
CO
X
"C
f-
c
4>
C "O •"••
2 -1) 41 X
•- > X 4)
r* -H «j ,e
-t o 2Q t3
c/3 <
JJ W
C S
41 4)
W >
u —
O- CO
4J
C
•
Cfl iH
CC O
01
X
vO
p^.
ON
**7
t
"U
—
-
~
.0
2
^
CO
g
^
.0
rH
y
i
T3
OJ
W
C8
.a
01
CJ
u
3
CO
.4
51
<->
O
u
u
^*
£
fH
£
^
•a
^^
A
fH
T"
C
C
o
•rt
4-1
•H
Q.
O.
O
u
a
CN
\O sO CN r^
CN rH m r**
ft -^ O >n
m CN ON co
•••T ^H CN n
O in o o
in
m (N co
fH CO •* • > O i- 1
O O 0 O
0 0 1-4 U u
cC
4-t
o * • • •
H < CC < CO
c
o
u crj 1—1
y ^H .ja
3 U C
4-1 XX
CC 3 CO
c -c <
p c
tN
*«D in
->j r^
"
CN -ff
•^ "^
ON in
-3
S m
O w
CN crj
cfl x
u jj
X CQ
CC fl
CJ O
^
t- Ul
0-, CM
O vO
p*i in
<• co
CN m
rH
.n
S CN
-31 00
00 CN
CO -H
4)
OJ
C
CC
T3 O
C CC
CO
O)
o "3
o u
O -H
3 u
£ °
< 09
X
i* CA
CJ OJ
3 -H
•3 I-
O u
^ CO
*" ^
t: c
O M
O
3
1
1
1
1
1
t
I
CO
=
o
u
K
A
U
•o
OJ
a.
•*
xO
0
o
ON
CTN
C^l
CN
OJ
3 en
CO 01
CO -i-l
cu cc
E
£ CO
COfJ
•H
CJ
ir
—
CN
~*
^
0
CN
to
1
z
in
CN
u
a
u
•a
OJ
u
0
•»
CN
CO
CN
fN
>,
fH
e
X
Cfl
<
u
^
c
^
3
U-
a
r^r^-rsi o>oo i— < r^- n ^r *n
-3- m ,-H i i rvt so -^ o O NON
"
r^ m in
O ^ CN r*-|n .>isDin OOCN ^r-.
»o -3" \o l I oo fsj O I-H co ON ?*} >n r*» t CN so
m ,-H
OOOIlO O OOltO
f*^ n in i i tn n c*^ in i t tn
CN -H ON sc p*. m 1 m
II • . . i t .
r--OCNI|r-. -T CNF-^I ?st
d JN CN -3" rH sO —t
00 fN
*J W U3 M
Qy3C/3 ZSE V3-:
C/3 COti^*^1^ W t/3 W CO "~3 S W
3 3322 n nooo-j-ss n
S .S Q £ ZfHrHfHSZ: U-i S
000 0
O OOllO OO OiJiJ^-iOO ) 4jsn
fH i— (in iH ^co f^ccccra-H,— i ccr*>.
C --
C , =
u xxcc^x xuxnXcncjccxx
cc cctOcrjcrj crjcccccoccc^C iw C3 cc
CJ •" U OCJ"H-n CJ O CJCJCJ CJCJ 33'JCJ
-O O O -O-oai-H T3 -^ t3-^-C •C'C -HCa*CT3
i) a- OH ot(.u.aa; a) --H QJ cu i* 41 DU ™ s D i)
U 3 3 WUCU U.^ Iu u 1- 1- UU O. JH U
SH caa-cua a* S O-.sua,a-a.,£., 1-J
X X 4J 4)
X f- 1 4U CO Cu rH
w 4) CJ CJ C ?C CO
CJ J3 rH3X 3CL,— fC-HQ^tn
o xcc-a-H3C4Jx j=uu*J
o -o 41 HXOIOCOIIOCJ COICQOCW — -HO
3 i- Q. r- 1 4-t CC fH -H CJ 3 I— ' > O C -H U3
v- oxH "cu^aj'cjc'rHr-'o x"w«*J-^ci.i"«"? xi-i
CU .^4) £cB>CC'H34f,H MCTJ^C3^i4^<^ 4)4)
O fH fH 32 33 0- U O O C < < Cu 3£ a- HO
CQ CC • *
U U rH CN
a Hs-^uawwafc < = cJCMi. u -
X
u to
U i- C
fH "O 4> C£
-H O 3 « U
4j u 4j ^ ai
X ^ 0 U J=
CU O to u
H ii a« o
CM
rsi
•*
•H
sC
-n
^
fN
rH
in
CN
^o
"^
m
CO
,
^
~i
^
< -i
o cu
01
01
ca
1-9
-------
£
CO
w
c_>
^
O
CO
ca
o
-3
2
"
,
J*"S
^
[T^
rfjj
S
§
CO
CO
c
O s~^
co cfl
CO CU
•H >,
b! "^
rH
4-> CO
C 60
CU
j> §
O *~*
oo
S-i
o
4-1
CJ
cu
CO
C
o
•H
4-1
cfl
O
•H
rH
a.
ex
CO
•H o 0)
4J TH C
•H 4J CO
Cfl Cfl rH
C 0 rH
01 TH CU
CO rH CJ
CX CO
11 O, TH
r< < 2
3
co s-i -o
co cu C
0) 42 CO
S-I 4J
(X O
• •
i-H CN
**O CO
• •
u"l \o
60
C
•H
60
cfl
^
O
CO
ft
01
rH
43
•H
X
CU
rH
<4H
T3
cu
4-1
cu
^
C '*~N
0 42
CJ CU
*•— ' 3
1
CO -i-l
O> 4-1
4J rH
Cfl 3
Cg CO
•H 4-J
e i o
CO 1 3
1—3 T3
CO O
60 4-1 |_i
CO O-i
•H 3
60 -a s-i
CO O o>
v! j_^ 43
GO- 42
CO 3
• •
ro
CN CO CO
• • •
ro
T3 T3
rl C
cO •* cfl
O l-i
42 cu -o
a. cu
T3 CO 4J
c a co
CO 60
rH 3
S-I rH S-<
CU cfl S-i
ft 3 0
aG
M
CO 60
4-J 4-1 C
3 0) -H CO
O rH T3 0)
43 3 ft
CU CO rH o
•H 4J G rH
T3 >
1 C (2
1 -H CO 01
4-1 S-I
CO vH 0) •>
4J S-I C CO
CJ 3 TH J^ &Q
3 CO o C
•O " 4-1 cfl -H
O co C co l-i
i-i 4-1 o cu
O-i G G T3 >
3 C 0
l-i T! - CO O
CU O CO
CX S-I rH CO rH
Cfl ft CO 60 rH
0* -H cfl cfl
S-i M 43 S
"O CU 0)
CU ft 4-1 )j TJ
4-1 cfl aj 0) C
l-i ft B ft TJ
01 CO
> X 60 ft 01
C S-i C rH
O cfl -H « S-i TH
U 4-1 T3 SJ CO H
•H rH CU 0)
S-l C TH 42 3 |_,
Ol Cfl 3 -H 4-1 O
42 Cfl 42 MH O O
4-1 O rH
O fa ft,
• • •
r*** cp o*\
CO
MH C
O 0
•H
C cn
O CO
TH iH
4J B
ft cu
0)
G 42
X G
CU -H
42
4J B
0
42 l-i
4-1 MH
•H
3 -
CO
S-I 4-1
CO G
0) 3
-^. O
rH ft
cfl
60 T)
O
2 O
Wri F^
2- 3
un s-i
• 0)
CN 42
C O
cO
42 -a
t! c
cfl
CU **
4-1 C
cfl O
CU -H
>^ -[„>
60 Cfl
i— 1
CO 3
C co •
O C rH
•H -H o
CO -C
CO CO 0
•H CO CJ
6 <0 rH
01 rH Cfl
60
42 rH
4-J ~ >>
•H T3 42
3 I-l 4J
CO 0)
co O B
cu 42
G ^3
S-l CU C
3 rH CO
0 G
CO vH rH
4J O
•H SJ 42
rH CO O
co ft y
r-H
0) "CO
•a s-i
3 CU rH
rH 0) >,
G C 42
C CU 4-1
•H > 0)
CO "O >•>
CO C rH
CU Cfl 4H
s-i cu
co -a -H
O 42
0) O G
co 3
cu >> cu
42 rH Jj
H ft CO
/-™s
CO
rH
rH
0)
G
c
CU
CU
UH
a)
1-10
-------
At that facility, which employes between 700 to 800 people, the quantity
of VOC losses amounts to about 55 gallons per day (about 50 tons per
year), a snail amount of loss according to the shoe industry-^ since
International Shoe of Jackson, Missouri emits 2,000 tons per year of
VOCs, and various footwear assembly facilities in St. Louis emit between
800 and 1,000 tons per year. These operations currently incorporate
no organic abatement equipment''.
CONTROL TECHNOLOGY
The VOC emission problem from the application of adhesives is
largely restricted to that from organic solvent-based adhesives. The
trend in control technology for solvent adhesives is not to control the
emissions from this type of adhesive, but rather to replace them with a
non-polluting waterborne type adhesive which can perform all the func-
tions of the solvent-borne adhesives^'-^. Another alternative is to
choose hot-melt adhesives which also maintain an excellent non-polluting
reputation").
Hot-melt adhesive application is the most widely used and does not
incorporate control technology since this method does not result in
volatile organic emissions"'. Water-base adhesives will become more
widely used in the future and these too do not contribute organic
emissions^), hence requiring no emission control technology.
According to the South Coast Air Quality Management District (Los
Angeles Basin), they know of no adhesive application operations cur-
rently using any add-on control technology. In addition, a representa-
tive from that agency reported that adhesive application operations do
not pose a serious contribution to the total VOC emissions of the Los
Angeles Basin. However, precipitators are recommended to abate any
visible emissions due to over-heating the adhesive.
There are several advantages to using waterborne adhesives: (1)
the emissions are non-polluting and satisfy the most stringent environ-
mental requirements; (2) as the cost of petroleum rises, the cost of
organic solvents will rise with it; (3) the danger of residual
1-11
-------
odor in a finished product is minimized; (4) fire hazards are eliminated,
reducing insurance rates; and (5) solvent fumes that linger in the
ambient air around processing machinery are absent, therefore making
the workplace safer and more comfortable^).
In addition, there are the minor advantages of easy cleanup and
the absence of attack on solvent-sensitive substrates.
STATUS OF REGULATORY ACTIVITIES
Currently there are no specific or local emission control regula-
tions concerned with the application of adhesives. There are numerous
general volatile organic compound emisson rules. However, the State
of Virginia Air Pollution Rules exempts adhesives from the general VOC
emission regulations.
REFERENCES
1. Irving Skeist, Handbook of Adhesives, Reinhold Publishing, 1962.
2. R. B. Seymour, Hot Organic Coatings, New York, Reinhold Publishing
Corp., 1959.
3. Contact with Red Wing Shoe, Red Wing Minnesota Adhesives Laboratory,
Jan. 1981.
4. Adolph Miller, Dissolving the Myths of Waterborne Adhesives, Paper,
Film and Foil CONVERTER, Dec. 1980.
5. William J. Storck, Adhesives Use to Continue Modest Climb, C & EN,
February 2, 1981.
6. Phone conversation with Leon Pechinsky, Lab Manager of Bostik
Chemical, Dec. 1980.
7. L. Powell Foster, Hot-Melt Adhesives for Bag Production, Paper,
Film and Foil CONVERTER, Dec. 1980.
8. Charles V. Cagle, Adhesive Bonding, McGraw-Hill, 1968.
9. Contact with Superintendent of International Shoe Co., Jackson,
Missouri, March 1981.
10. James O'Leary, Striving for Compliance with High Solids/Solvent
Adhesives, Paper, Film and Foil CONVERTER, Dec. 1980.
11. William L. Johnson, EPA, Durham, NC, October 1980.
1-12
-------
SECTION 2
ASPHALT AIR BLOWING
PROCESSES AND EMISSIONS
Asphalt is produced largely from residues of atmospheric or vacuum
distillation of crude oil. Some is derived from propane de-asphalting.
Over 99 percent of that used is from petroleum refinery sources, with
ever decreasing amounts of natural asphalt from seepage being used-'-'.
Asphalt from these sources can be used directly or modified by air blow-
ing. Air blowing of asphalt stock (flux) converts it to a harder product
by air contact at a temperature of 200°C to 240°C (392°F to 464°F). The
process is largely a dehydrogenation process with the vacated sites
stimulating condensation or polymerization of the asphalt. The reaction
is exothermic and water is a principal by-product. Most of the asphalt
air blowing is carried out in petroleum refineries or asphalt processing
plants, with the remaining amount being done in roofing plants.
Air blowing is conducted in horizontal or vertical vessels, and can
be carried out batch-wise or continuously (Figure 2-1). Vertical stills
are more effective because of longer air-asphalt contact time. The
continuous process has advantages of better operational control, less
pre-heat capacity requirement, and lower blowing losses. The continuous
process usually operates at a constant temperature (260°C = 500°F) and
constant liquid level.
While the largest use of asphalt is for paving (about 75 percent),
relatively little paving asphalt is air blown^). Approximately 20
percent of asphalt production is consumed in manufacture of roofing
products. Some is used as saturant in asphalt roofing felt and some is
used as a coating. All roofing asphalts are air blown. About 5 percent
of asphalt goes to cements, coatings, emulsions, pipe coatings, auto
undercoatings, laminates, water proofings, potting compounds, and nearly
200 other uses. Some of these use air blown asphalt.
Process Emissions and Emission Points
Emissions from asphalt air blowing include both gases and aerosols.
Excluding the nitrogen, residual oxygen and water of formation, most of
2-1
-------
FIGURE 2-1
t—
C/9
" '! *
CO
2-2
ENGINEERING-SCIENCE
-------
the gas phase emissions are hydrocarbons. Relatively little oxygen is
retained in the asphalt, but some oxygenated hydrocarbon products are
formed. The lower molecular weight hydrocarbons, oxygenates, and some
contaminant sulfur and nitrogen compounds are carried out of the blowing
still together with nitrogen, air, and water vapor. It is conventional
practice, even without specific air pollution control regulations, to
pass these off-gases through a knock-out drum or condenser prior to
discharge to the atmosphere vent stack. Some aerosols and condensibles
are removed from the effluent stream prior to discharge to atmosphere.
Relatively little testing has been performed on blowing stills*.
CONTROL TECHNOLOGY
From a practical point of view, only thermal afterburners have been
used to control emissions of volatile organic carbon from asphalt air
blowing. Afterburners are effective in destroying odiferous gases and
will combust non-condensible VOC emissions as well.
Afterburners are classified as either thermal (i.e., direct flame)
or catalytic. The primary advantage of catalytic afterburners is that
they use much less supplemental fuel than an equivalent thermal after-
burner. Catalytic afterburners are not used or recommended for control
of hydrocarbon emissions from asphalt blowing stills because the catalyst
is subject to rapid poisoning and plugging due to constituents of the
fumes from asphalt processes-*'.
Thermal afterburners destroy combustible pollutants through oxida-
tion to C02 and water. Temperatures of 650° to 760°C (1200° to 1400°F),
maintained for 0.1 to 0.3 seconds of fume residence time, are sufficient
to obtain nearly complete oxidation of most combustible pollutants'^'.
Destruction of most hydrocarbons occurs rapidly at 593° to 650°C (1100°
*Data collected by EPA shows uncontrolled gaseous hydrocarbon emis-
sions ranging from about 6,000 to 9,000 ppm. At the plant tested this
is equivalent to an average emission factor of 1.32 Ib. VOC emissions per
ton of saturant asphalt blown and 3.48 Ibs. per ton of coating asphalt.
2-3
-------
to 1200°F), but destruction of some organic compounds, such as methane,
and the oxidation of CO to CC>2 requires longer residence times and higher
temperatures. Temperatures of 760° to 816°C (1400° to 1500°F) may be
required if the methane content of the hydrocarbon is over 1000 ppm '.
Large droplets (50 to 100 micron) require longer residence times at the
above temperatures; however, these large droplets are also easily removed
in simple cyclones and knock-out vessels^).
The steps involved in dilute fume incineration are shown schemati-
cally in Figure 2-2. As shown in the figure, part of the fume stream is
sometimes bypassed around the fuel combustion process to preclude flame
quenching and combustion instability. In the case of exhaust streams
containing emissions from asphalt blowing, it is common to use only
outside air in the combustion of fuel, since burner fouling is a problem.
For other asphalt roofing processes, burner fouling seems to be less of
a problem, and the fume stream is often used as a major source of combus-
tion air. The fume not used for combustion must then be mixed with the
hot combustion products to give a uniform temperature to all fume flowing
through the afterburner. This mixing should be done as rapidly as possible
without causing flame quenching so that sufficient residence time can be
provided at the required temperature. Temperature and residence time
are somewhat interchangeable; a higher temperature allows use of a shorter
residence time and vice versa. This is illustrated in Figure 2-3 which
indicated that for a 0.1 second residence time the efficiency of pollutant
oxidation varies from 90 percent at 666°C (1231°F) to 100 percent at
725°C (1337°F). For a 1.0 second residence time, the efficiency varies
from 90 percent at 623°C (1153°F) to 100 percent at 666°C (1231°F).
The typical effect of operating temperature on the effectiveness of
thermal afterburner destruction of hydrocarbons is shown in Figure 2-4.
The figure shows that the efficiency of hydrocarbon destruction varies
from about 90 percent to almost 100 percent over a temperature range of
about 677°C to 760°C (1250°F to 1400°F). For a given level of pollutant
destruction for different afterburner designs, the major factor that
influences the residence time required at a given operating temperature
(above about 538°C [1000°F]) is the effectiveness with which the fume is
2-4
-------
FIGURE 2-2
0)
4->
3
i-l
•H
T3
(U
cn
3
o
o
c
o
O
c
rH 0)
3 C
14-1 )-l
03 3
M ,O
0) i-l
O —i
2: or ^3 o
o i— i—i
i-> t— < U_
I— < a: u_
:z LU zj
LLJ UJ Q- OO
i— 2: s:
LU ID uj a:
0£ U_ V- O
LU UJ
2! 00
=)
02
s:
o
ca
s:
o
Q_ U.
Q.
00
I
a:
>—4
u.
uj •— «C
U1 1— UJ
^3 00 UJ 21
^> Q =D
UJ O2 >—i U_
O3 2! 00 C£
o I— u_ uj
>- C_) Z3 >—' "Z.
ef. O 0£.
2: >- o ^)
_j UJ CQ
UJ Q_ • Q
2: Q_ a: uj
^3 ra >—i uj
u. oo
-------
FIGURE 2-3
IS
II
§
gs
M-
O
O)
C
O
-------
FIGURE 2-4
I
c
o
I
100
90
SO
60
50
Hydrocarbons
Only
Hydrocarbon •+• CO
650
(1200)
705
(1300)
760
(1400)
815
(1500)
Temperature °C (°F)
Typical effect of operating temperature on effectiveness of thermal
afterburner for destruction of hydrocarbons and carbon monoxide.4)
2-7
ENGINEERING-SCIENCE
-------
mixed with the combustion products. If hydrocarbons are present in the
exhaust gas of any afterburner operating at a nominal combustion chamber
temperature above 760°C (1400°F) (or above 649°C [1200°F] for all but a
few hydrocarbons), it is due to poor mixing and non-uniform treatment of
the fume stream or too short residence time of the fume at temperature.
Typically, afterburners are designed with average residence times which
vary from 0.1 to 0.5 seconds, but the amount of time required to raise
the cold fume up to the desired temperature often exceeds this average
residence time. Also, not all portions of the fume are in the combustion
chamber an equal amount of time; some portions are swept out very quickly
while others are retained for an appreciable time. The variation in
residence time, which is a function of flow patterns in the combustion
chamber, can appreciably affect afterburner performance. In practice,
operating personnel compensate for deficiencies in design by increasing
the operating temperature of the thermal afterburners during the start-up
phase until a temperature is reached which produces the desired pollutant
destruction.
Little maintenance is required on most thermal afterburners. The
main operating problems involve safety controls, erosion or cracking of
refractory linings, heat exchanger fouling, or mechanical failure and
bearing failure in the fans.
The major distinguishing feature of thermal afterburners, as compared
to non-combustion control techniques for hydrocarbons, is the use of
fuel. Because exhaust gases from the afterburner are typically at 649°
to 816°C (1200° to 1500°F), many asphalt roofing plants use heat exchangers
to recover the waste heat. This recovered waste heat may be used for
many of the plant processes.
Thermal afterburners, like all combustion sources, have the potential
for generating secondary pollutants due to oxidation of nitrogen, sulfur,
and metals in the fume or fuel. Thermal afterburners, in comparison
with power plant boilers and industrial furnaces, have low NOX emissions
because of their lower operating temperatures. The low operating tempera-
tures and dilution of combustion products by excess air and fume results
in a NOX effluent concentration of 5 to 15 ppm. Emissions of SC>2 depend
on the sulfur content of the fuel burned and on the sulfur content of the
fune because almost 100 percent of this sulfur will be converted to S02-
2-8
-------
Costs for controlling blowing stills in model plants have been
estimated by the U.S. EPA3'. In all cases the air blowing was conducted
as part of operations at integrated roofing plants. Model plant asphalt
production ranged from 46,200 Mg/yr. (50,931 tons/yr.) to 116,219 Mg/yr.
(128,122 tons/yr.). Still sizes ranged from 75.7 m3 (20,000 gals.) to
94.6 m3 (25,000 gals.). The basis for control was an afterburner having
a design operating temperature of 815°C (1500°F) and a residence time of
0.3 to 0.5 seconds. Depending upon the size of the plant, the cost
effectiveness for particulate phase organics controlled ranged from $98/Mg
($89/ton) to $134/Mg ($121/ton). These figures cannot be used directly
for the vapor phase portion of the organic emissions. Since particulate
phase organic emissions from blowing stills are about six times as high
as vapor phase organic emissions, the cost effectiveness on the vapor
phase portion alone would be about six times as high as the figures
reported above. The proper allocation of control costs, therefore,
depends upon whether one or both of the two classes of organic emissions
require emission controls.
STATUS OF REGULATORY ACTIVITIES
Existing air pollution control rules and regulations at the state
and local level have been directed at particulate emissions, visible
plumes, and odorous emissions from asphalt air blowing. While there are
no regulations directly affecting volatile organic compound emissions from
this source, the Bay Area Air Quality Management District in California
does have a regulation covering asphalt air blowing which requires
incineration of the effluent at not less than 650°C (1202°F) for a period
of not less than 0.3 seconds. This rule is directed towards odors but
would be effective in controlling VOC emissions as well. A copy of the
rule is attached. Performance standards under consideration for the
asphalt roofing manufacturing industry under Section 111 of the Clean
Air Act (42 U.S.C. 7411), as amended would include asphalt blowing stills
at roofing plants and refineries as affected facilities. The numerical
performance standards under consideration are based upon mass emission
rate of particulates and upon opacity. VOC emissions are not directly
affected. However, afterburners are the only demonstrated means of
control for blowing still emissions. Tests conducted on a well designed
2-9
-------
afterburner indicate that VOC emissions, as well as organic particulates
are reduced by the use of afterburners. VOC reduction efficiencies
exceeded 99 percent in all tests conducted for the New Source Performance
Standards study.
REFERENCES
1. Evans, J.V. Asphalt Industry: Kirk-Othmer Encyclopedia of Chemical
Technology, Volume 3, 3rd Ed., Mark, H.F. et at (Ed.). John Wiley and
Sons, a Wiley-Interscience Publication, New York, N.Y., 1978.
2. Telecon. Marker, V., The Asphalt Institute, with Bryan, R., ES Inc.,
November 12, 1980. Usage of Air Blown Asphalt.
3. "Asphalt Roofing Manufacturing Industry - Background Information for
Proposed Standard." EPA-450/3-78. Draft EIS. June 1978.
4. Afterburner Systems Study. U.S. Environmental Protection Agency.
Research Triangle Park, N.C. EPA-R2-72-062. August 1972.
2-10
-------
REGULATION 12
MISCELLANEOUS STANDARDS OF PERFORMANCE
RULE 3
ASPHALT AIR BLOWING
12-3-10O GENERAL
12-3-101 Description: This Rule applies to operations involving the air blowing of asphalt.
12-3-300 STANDARDS
12-3-301 Processing of Gases: A person shall not engage in the air blowing of asphalt unless all
gases, vapors and gas-entrained effluents are incinerated at temperatures of not less than
650°C (1202°F) for a period of not less than 0.3 seconds; or processed in a manner which
is equally or more effective for the purpose of air pollution odor control as determined by
the APCO.
12-3-50O MONITORING AND RECORDS
1 2-3-501 Monitoring: A person incinerating or processing gases, vapors or gas-entrained effluents
pursuant to this Rule shall provide, install, calibrate and maintain in good working order
devices for indicating temperature, pressure or other operating conditions, as specified
by the APCO.
12-3-3
-------
SECTION 3
BARGE AND TANKER CLEANING
PROCESSES AND EMISSIONS
Cargo tanks on tankers or barges carrying crude oil, various
petroleum products or other volatile organic liquids are cleaned
regularly or occasionally for one of the following purposes:
0 Cleaning for removal of sludge
0 Cleaning for de-ballasting
0 Cleaning for shipyard entry
0 Cleaning of product carriers to insure product purity
This cleaning generally is conducted during or after cargo unloading in
the destination port. The cleaning for removal of sludge is essentially
applicable only to crude oil carriers. In the case of sludge removal,
both crude oil wash (COW) and water washing methods are used. Often
both are used with the water wash following the crude oil wash. Washing
for other purposes is usually done with water. When human entry is
required into cargo tanks, such as for repairs to be conducted in a
shipyard, tanks are gas-freed by air blowing following the tank cleaning
operation.
Barges are cleaned much less frequently than tankers. This is
probably because many barges are in dedicated service carrying gasoline,
light fuel oils, and liquid petrochemicals, e.g. benzene.
Tank washing is, of course, associated with a good deal of splashing
action. As a result of this splashing and agitation, the VOC concentra-
tion in a tank can increase after cleaning. Whether and how long this
condition persists is subject to a number of other factors. These
include (1) method of disposal of slops (residual VOL or VOL-water
mixture) — at the loading or unloading ports or at sea, (2) whether
the cleaned tank is purged with inert gas or gas-freed with air after
cleaning, (3) how effective was the cleaning in removing residual
volatile organic liquid, and (4) temperature of cleaning fluid.
3-1
-------
On November 19, 1979 the U.S. Coast Guard published regulations
relating to tanker safety and oil pollution which could have an impact
on tanker cleaning emissions*. Under these regulations new crude
carriers larger than 20,000 DWT would be required to have COW systems
and existing crude carriers larger than 40,000 DWT would be required to
retrofit with either a COW system, clean ballast tanks (CBT), or
segregated ballast tanks (SET). The purpose of these regulations is to
reduce operational oil pollution. COW systems reduce the potential for
oil pollution by reducing the clingage that remains after cargo dis-
charge. In the case of crude oil tankers, the U.S. Coast Guard
regulation assumes that clingage amounts to 0.4 percent of the original
cargo. If tank washing is carried out with water, the oil-water mix-
ture (slops) resulting is often discharged at sea. Clingage after
crude oil washing is reduced to about 1/30,000 (0.0033 percent) of the
original cargo. During washing the run-down puts the semi-solid oily
residue back into liquid suspension. Therefore, crude oil washing is
usually done during cargo discharge.
As mentioned earlier, the crude oil washing results in higher tank
concentrations of VOC after cargo discharge. Therefore, the U.S. Coast
Guard is requiring that tankers with COW systems and no segregated
ballast tanks (used with sea water only) have the capability to conduct
vapors displaced during ballasting back to tanks that are discharging
crude oil. This would be done with internal vapor balance systems.
Obviously, ballasting would have to be conducted simultaneously with
cargo discharge operations. This inter-tank balancing should reduce or
eliminate ballasting emissions at discharge ports, but filling emissions
at loading ports could increase unless onshore vapor recovery systems
were available.
Barges are not ballasted. However, filling emissions of VOC
could increase under some circumstances if tanks are cleaned. In a
*33 CFR, Part 157 - Rules and Regulations for Protection of the Marine
Environmental Relating to Tank Vessels Carrying Oil in Domestic Trade,
Part D - Crude Oil Washing Systems on Tank Vessels, Part 157.132 -
Cargo Tanks: Hydrocarbon Vapor Emissions.
3-2
-------
study sponsored by the API-'-), three compartments in a barge previously
carrying gasoline were washed with hot (180°F) water. Two tanks were
subsequently air blown. Hydrocarbon concentrations in all three tanks
following the washing were higher than the saturation concentration
for the reported cargo temperature and RVP. Subsequent tests made
during gasoline loading showed that the emissions from all three tanks
were higher than those from tanks which had previously contained diesel
fuel and not cleaned prior to loading with gasoline.
There are no satisfactory methods for predicting emissions from
tank cleaning or from operations conducted subsequent to tank cleaning.
A careful analysis of the operations to be conducted should, however,
enable a prediction to be made as to the general direction of the effect
on emissions.
CONTROL TECHNOLOGY
Tanker and barge cleaning emissions essentially result from dis-
placement of VOC vapor mixtures from tanks after cleaning. This
displacement results from ballasting or loading cargo into tanks which
have been cleaned in such a way that the significant VOC concentrations
are still present at time of ballasting or loading. As discussed
earlier, the U.S. Coast Guard published regulations requiring internal
balancing of ballasting emissions on tankers equipped with COW systems.
An alternate scheme of control would be to discharge these emissions
into a shore-based vapor recovery system. Such a system would not be
installed exclusively to handle emissions from cleaned tanks having
high VOC concentrations, but for all loading and ballasting emissions.
Where internal balancing systems are used, the tanks receiving the
displaced vapors will very likely contain higher concentrations of VOC
upon return to the loading port. These high arrival concentrations
could be reduced by inert gas purging or gas-freeing during the return
voyage. Should the loading port be equipped with a vapor recovery
system, VOC emitted during loading would be controlled regardless of
concentration. There are no vapor recovery systems known to be
installed at marine terminals in the United States at the present
time.
3-3
-------
In the case of barge cleaning with water, the use of hot water
could be avoided. Further, the cleaning should be done just after
unloading instead of immediately prior to loading.
STATUS OF REGULATORY ACTIVITIES
There are no federal, state, or local regulations specifically
addressed to VOC emissions resulting from tanker or barge cargo tank
cleaning. The U.S. Coast Guard regulations relating to use of COW
systems (mentioned earlier) do require internal tank vapor balancing
during ballasting. This has the effect of reducing ballasting emis-
sions, but possibly increasing subsequent loading emissions.
REFERENCES
1. "8-31 Marine Emissions Study, Final Report," 8-31 Technical Advisory
Committee, American Petroleum Institute, Washington, D.C., 1978.
3-4
-------
SECTION 4
BARGE AND TANKER LOADING OF VOLATILE ORGANIC LIQUIDS
PROCESSES AND EMISSIONS
The operations covered in this source category are conducted at
marine terminals. These are usually located adjacent to refineries or
chemical plants, but may also be operated as the terminus of a dedicated
or jointly used pipeline. They may be located in deep water ports or
inland waterways. Operations conducted include the loading and unload-
ing of crude oil, gasoline, and other volatile organic liquids such as
benzene. In the United States crude oil loading is rather uncommon;
essentially all the crude oil arrives for unloading in tankers.
The transfer operations carried out at marine terminals have a
certain degree of similarity to those conducted at conventional re-
finery and pipeline truck terminals and bulk plants. The major
differences relate to scale of operations and the necessity to load
ballast water to some vessels which would otherwise depart empty after
cargo unloading. Ballasting is essentially restricted to tankers and
may be done using dedicated ballast tanks or cargo tanks. Most bal-
lasting is done in crude oil tankers. Because tanker and barge
operations are rather different, they will be discussed separately.
In the case of tankers we will concentrate on crude ballasting, and for
barges on gasoline loading.
Crude Ballasting in Tankers
Following the discharge of crude oil cargo, sufficient ballast
water must be loaded on tankers to permit them to depart port safely.
Increasing numbers of tankers have sufficient segregated ballast tanks
so that ballast need not be added in port to cargo tanks which contained
crude oil prior to unloading. However, most of the tankers currently
in service require some ballasting in cargo tanks. A ship's captain
may decide to take on more than the normal amount of ballast if he
decides that sea conditions require additional ballast. In such cases,
it is possible that some ballasting in cargo tanks may take place even
in ships with provision for segregated ballast. The amount of ballast
4-1
-------
received at the dock (exclusive of that introduced to segregated tanks)
is typically about 20 percent of the crude carrying capacity of the
tanker.
When ballast water is added to tanks previously containing crude
oil the vapor laden atmosphere of the tank is displaced in an amount
equal to the volume of ballast water added. The hydrocarbon vapor
concentration is related to the vapor pressure of the crude at time of
arrival, the rate of unloading, the delay after unloading prior to
ballasting, whether the tank was slack (short) loaded at time of arrival,
whether any lightering of cargo took place, and sun exposure of the
cargo tank. Slack loading and lightering results in additional vapor
space prior to dockside unloading and thus a greater opportunity for
vapor concentrations to reach equilibrium with the cargo as compared to
the situation with a full cargo compartment where evaporation must take
place during the period of unloading and standing prior to ballasting.
Delays in the start of ballasting provides time for additional evapora-
tion from the cargo heel in the tank.
During the ballasting operation, current practice is to remove the
ullage caps from the tank being ballasted. These caps are used for
visual inspection of the tank at deck level and are normally closed.
The tern ullage refers to the ullage space, which in tanker terminology
refers to the empty space above the liquid level in the tank. The
vapor laden air in the tank is discharged to atmosphere during ballast-
ing from the ullage openings. Testing of emissions from marine terminal
operations has not been extensive in the sense that sufficient data
has been obtained to develop accurate predictive approaches. Conse-
quently it is not possible to quantify the influence of the parameters
earlier mentioned.
Emission estimates given in this report are largely derived from
testing performed in response to a request from EPA^). Tests were
conducted on 22 different crude oil tankers ranging in size from 42
MDWT to 121 MDWT. Fourteen different crude oils with a volatility
range from 0.7 psi Reid vapor pressure (RVP) to 8.6 psi RVP were in-
volved in the testing. Testing covered locations on the East Coast,
Gulf Coast, and West Coast. A range of crude oil and ambient air
4-2
-------
temperatures were encountered. Emissions ranged from 0.4 to 3.5
pounds of VOC per 1000 gals, of water ballast (to cargo tanks). The
mid-range was slightly above 1.0 Ib./lOOO gals. The emission factors
given for ballasting of tankers in AP-422) are 0.6 Ib./lOOO gals, and
0.8 Ib./lOOO gals, water ballasted for tanks previously containing
crude oil and gasoline, respectively.
VOC LOADING OPERATIONS
Gasoline, single component volatile organic liquids such as ben-
zene , and crude oil are loaded at deepwater and inland ports in the
United States. The amount of crude oil loaded is relatively small and
is concentrated near coastal oil fields on the West and Gulf coasts
not served by pipelines. In principal, the loading operations are
similar to those conducted at truck and rail car loading terminals,
except that the total volumes transferred per unit time are far greater
at marine terminals than at truck or rail terminals. For example, the
loading rate for gasoline into a 10,000 bbl. barge is typically about
2,000 gals./min. as compared to a tank truck loading rate of about 600
gals./tnin.
There are a significant number of parameters affecting the emission
factors for ship and barge loading with VOL. These are broadly
categorized into (1) vessel factors, (2) cargo factors, (3) cargo tank
history, and (4) loading factors. Vessel factors are principally
related to size and shape. Tank ships typically range from 35,000 to
120,000 DWT (cargo capacity 235,000 to 840,000 bbls.) and have rela-
tively deep draft. Barges typically range from about 10,000 to 20,000
bbls. in capacity and are of shallow draft to accommodate passage in
inland waterways. Thus the surface to volume ratio of a barge is
greater than that of a tanker. This in turn permits a relatively
greater potential for evaporation per unit volume of cargo.
The most important cargo factor is vapor pressure. The vapor
pressure on loading conditions is best determined from the RVP and the
bulk temperature of the product. The relationship between RVP and true
vapor pressure is not as straightforward for crude as it is for pure
4-3
-------
compounds or for gasoline, making the estimation of crude oil loading
losses less accurate than that for gasoline.
The arrival condition of empty tankers or barges is related to
previous cargo and cleaning history. Arrival categories commonly used
are as follows:
0 Clean: Prior to loading, compartments were water-washed or gas-
freed; or the prior cargo was non-volatile*.
0 Ballasted: Ballast contained in compartments that previously
held volatile cargo.
0 Uncleaned: No treatment prior to loading and prior cargo was
volatile.
The arrival condition is important in that loading emissions
consist of both the VOC present in the tank prior to loading plus
any additional VOC generated by evaporation of the VOL cargo being
loaded.
Loading factors include fill method, initial turbulence, initial
and bulk loading rates, and amount loaded expressed as a fraction of
tank capacity. While loading factors are generally not taken into
account in estimating marine loading emissions, the use of a lower
initial rate of filling (reduced turbulence) and a rapid bulk filling
rate (reduced time for evaporation from loaded liquid) could reduce
emissions. The degree of tank filling is important because the concen-
tration of emitted vapors increases with time during loading (particu-
larly when loading uncleaned tanks). Near the end of loading when
tanks are completely filled, the concentration may approach the equili-
brium vapor pressure of the previous cargo (applicable in uncleaned
t anks).
Emission factors in AP-42 for gasoline loading into tankers and
barges have been revised fairly recently (1977). These factors are
^Emissions are normally less for clean tanks, other factors being
equal. However, if a tank was cleaned with hot water just prior to
loading, vapor concentrations could be higher than usual.
4-4
-------
shown in Table 4-3.2). Limited testing of gasoline loading into barges
was conducted as part of the API study previously mentioned-*-). In one
case, the barge was previously loaded with diesel fuel. During loading
with motor gasoline of 8.5 psi RVP at a bulk temperature of 92°F, the
average emission rate for three compartments was 2.2 Ibs./lOOO gals.
loaded. In a second test, gasoline of 9.7 psi RVP was loaded into a
barge which previously contained gasoline. Prior to loading, the tanks
were washed with water at 180°F. Two tanks were air blown and one was
not. Emission rates for the two washed and air blown tanks were 2.8
and 3.0 Ibs./lOOO gals. The rate from the washed only tank was 9.7
Ibs./lOOO gals. This condition most likely resulted from production
of excess vapors which were not purged by air blowing after the hot
water wash.
Emissions from crude oil loading are given by AP-42^) as 0.7
Ib./lOOO gals, for tankers and 1.7 Ibs./lOOO gals, for barges. This
assumes crude having a RVP of 5.0 psi. The API study showed a crude
oil loading average emission rate of 1.0 Ib./lOOO gals, when loading
Nigerian crude oil, with an RVP of 6.9 psi and a temperature of 68°F.
Three compartments on one ship were tested.
CONTROL TECHNOLOGY
At the present time there are no VOC control systems on the marine
(vessel) related activities at marine terminals*. The marine activities
include tanker and barge VOL loading and tanker ballasting. Tank
storage facilities on land associated with marine terminals utilize the
same type of floating-roof tanks for volatile organic liquids as would
be used in refinery or pipeline terminals, or tank farms.
Several studies have been conducted on the possible configuration
and feasibility of marine terminal vapor recovery systems3»^,5)^ A
major deterrent to date has been the uncertainty in the performance of
safety devices at the scale-up in size required. Any such systems must
be approved by the U.S. Coast Guard.
*Under new U.S. Coast Guard regulations (33 CFR, Parts 154 and 164, and
46 CFR, Parts 30, 32, and 34, some tankers will be required to have
Crude Oil Washing (COW) systems. Tankers so equipped will also be
required to have inter-tank vapor balancing lines for use during
ballasting. Ballasting will have to take place during cargo unloading.
4-5
-------
CM
oo
2
O
1— 1
H
<
OS
w
OH
O
O
2
M
n
H
CO
l-l
o
4-1
o
cfl
fa
c
o
•H
CO
CO
•H
1
c
o
43
VJ
cfl
O
o
1-1
•o
>,
S3
CO
cu
00
p
cO
pa
cu
00
CO
(-1
cu
$
cu
00
c
cfl
OS
CO
0)
00
s-i
CO
pa
c
cfl
cu
a
0
cu
00
cfl
Vj
cu
>
•5
cu
00
c
3
CO
a
*rH
43
CO
cu
00
CO
p
cu
>
<
cu
oo
C
cfl
03
c
o
iH
4-1
•H
"O
c
O
u
^
c
cfl
H
rH
CU
CO
CO
CU
- rH
s~^ x~>. o W s~*
CO 43 4-1 CO
r^
• •
vO O O
CO CO u~i
• O O X-*
CO O O -U O 4-1 CO
4-1 4-1
O O ^O so
4J 4-1 UO O LO O
• • • •
O O O O O O
CM CTi O\ r->
OrH vOrH <-CM -O-rH
• • •• •• ••
rHO iHO CMO r-lO
vO 00
CO ~ri-
^ ' ^^
• •
CO O O
co CM co "O
4-14-1 -* O -3" O
• • • •
O O O O O O
CU
0 C
•H O
CU CU P CU 4-1 >-llH O PMH
»-l CUW CUCQ CU CUWCJ CUM
h<4HC <4-l C 4-1 4-1 C >4HC
COCO COCfl CO COCO^ COCO
i-i Cl-i CM 0 Cp C Cp
O C04J Cfl4-lTH Cfl4J CO C04J
a P (-1 -a P H (-1
Cfl4-Jp 4-1 l-t 0) 4J >-l 4-11-1
> cu 01 Q o> o ai
rH4J rH4J rH4J(JOrH4J
13 Cfl-H Cfl-H 1 Cfl'H M CO-H
COOrH OOrH OOrHCOOOrH
CO -O T3 U
COCO CUCOCO CUCOCO COCO
T3 OO 4J OO C OO CU OO
01 rHrH CO rHrH CO rHrH OOrHiH
C ^^. *~-^ cfl ~^«. '^^ CU *^^ "*^» cfl ^*«^ *^«
cfl 4300rH 4300 rH 43SO P 4300
CU lH^
cj pa 3 < •
•
•a
cu
• 4-1
CU CO
rH CO
4= rH
CO i— 1
rH CO
•H 43
Cfl
> >->
CO rH
rH
S 1
c c
o
01 C
(-1
CO 4-J
O
en c
0)
3 CU
rH p
Cfl CO
>
Cfl
CU 01
CO 00
ai P
4= cfl
H pa
/-*, x-s
CO 43
4-6
-------
The types of vapor recovery systems which possibly might be
considered include low temperature refrigeration, compression type
refrigeration systems, activated carbon adsorption, lean oil absorption,
and incineration. All but the compression type refrigeration system
will be discussed using the hypothetical facility configuration shown
in Figure 4-1. This facility includes provision for tanker unloading
(and thus ballasting) and barge loading. Vapor controls are shown only
for the transfer and ballasting operations. Open floating-roof tanks
are used for storage on shore. It is expected that these tanks would
use double seals on the floating roofs. If this is the case, the seal
factor currently shown in the standing loss equation for floating-roof
tanks would give conservatively high emission estimates. New factors
are under study by the API but they have not yet been officially
published.
In the discussion of the possible emission control systems, no
direct reference will be made to the safety issue. There is difference
of opinion as to whether or not active systems designed to render the
returned VOC vapor-air mixture non-flammable are necessary. Concepts
involving saturation with gasoline, high ratio air dilution, blanketing
with butane or fuel gas, and blanketing with inert gases (N2 or gases
from an inert gas generator). Later in this section other safety
aspects will be discussed. The impact on cost will also be mentioned
briefly.
Refrigeration
The refrigeration system discussed here utilizes two operating
stages to reduce temperatures of the vapor mixture to below -100°F.
The principal involves condensation of volatile organic substances at
atmospheric pressure. Vapor pressures of hydrocarbons of C% and above
are low enough at these tempertures to give high removal efficiencies.
Figure 4-2 shows a schematic diagram of such a system. The system
requires pre-cooling to remove water prior to passage of the vapor
mixture through the low temperature sections. Water must be removed
from the condensed hydrocarbons prior to the return of the recovered
hydrocarbons to storage. Critical features of the system include
4-7
-------
FIGURE 4-1
I
TUT
ENGINEERING-SCIENCE
-------
FIGURE 4-2
>- cr
/ nCt — ' ' ' •
uj 2^0 o:£
_ CH — 'O^~~ ^UJ "
QO,,, OJ^CJ 0.0
uT-x OUJLU
£go \ \ -.
C*-r ^> .
\—^ . — _ - , - , »-
^* k— I 7, •
<
u
rO
V-
w
J
pi ^..,,_,,. .„, _.._ -« . •.
oQ^ /
Z ^ ± ±\7 ^>t >' — ' . . , , M
g ^^^ JSIO"1,^
g l^O^JJJ
0
M
H
3 1
M
O Q_ T^ ,1-9.
s uj / y^
I • z -^P -^
UJ^ ^ < ^ QJ
^~\ \ — 1 Q_ — ^_
T ^ Hi
1
cc cr / cro _
UJ UJ .— / UJ— C N.
^/) i x — '5?
i iCc^x r>>'-'- —
UJtzO HUJZ
Q• UJ
Z — t 5 7"
UJ — 1 UJ i , irv
i8*i£
83g8l
*" " *1
J ' j
5 — '
(
2_i
5)
1U
JQ-
;co
:Q
w
L
4-9
ENGINEERING-SCIENCE
-------
refrigeration module performance and correct cycling of the defrost
system. The system is relatively simple, occupies little space, but
is energy intensive. Hydrocarbons are recovered.
Carbon Adsorption
Carbon adsorption utilizes the affinity of activated carbon for
hydrocarbon compounds to remove gasoline vapors from air. A typical
carbon adsorption system consists of two or more carbon adsorber beds
and a regeneration system for the carbon beds (see Figure 4-3). Two or
more beds are necessary in order to keep the unit on stream during long
periods of return vapor flow. This is because the adsorption bed is
regenerated on batch basis. The number and size of the beds are
determined by the loading of adsorbable vapors (VOC), the flow rate,
the length of time required for regeneration, and the net or working
capacity of the activated carbon beds. Regeneration is performed by
steam or vacuum in situ.
The system illustrated in Figure 4-3 uses vacuum regeneration at a
vacuum of approximately 25 mm Hg. The stripped VOC vapors (now much
concentrated) are condensed or absorbed into gasoline liquid. Systems
similar to the one described here have been used at gasoline truck
terminals. Efficiencies of VOC control have reached as high as 99
percent.
Lean Oil Adsorption
This system uses chilled diesel oil (40°F) to absorb the VOC
vapors from the air-vapor mixture. This oil which has been stripped of
lower molecular weight compounds effectively absorbs butane and heavier
hydrocarbons when properly designed and sized. Some portion of any
propane present is also absorbed, but methane and ethane if present
pass through. Following absorption of the vapors the so-called rich
oil is regenerated by heating at 200°F at 150 mm Hg absolute pressure.
The stripped or lean oil is cooled and returned to the storage tank.
The VOC vapors flashed off in the stripper are then usually
compressed and cooled to condense the vapors and then blended into
gasoline or other similar product. The lean oil absorption system is
illustrated in Figure 4-4.
4-10
-------
FIGURE 4-3
w
Q
s
o
M
H
o
CO
2
O
PQ
u
i
CC
LU O
UJ I
crcc <
Q uu —' Q:
(XX ^
(T
2
4-11
ENGINEERING-SCIENCE
-------
FIGURE 4-4
o
M
H
o
CO
pa
H
O
I
w
I-J
z
z
UJ
IU
o
(A
4-12
ENGINEERING-SCIENCE
-------
The effectiveness of the lean oil absorption system is dependent
upon the ratio of lean oil flow to vapor flow, the purity and temperature
of the lean oil, the absorption column design, and the concentration of
VOC in the air-vapor mixture. One advantage of the system is the
flexibility in choice of size of the absorption unit. By increasing
the amount of lean oil storage, the actual processing equipment can be
downsized to that required over a relatively long averaging time. Some
reserve capacity must, however, be provided.
Thermal Incineration
Thermal incineration (Figure 4-5) disposes of VOC by burning rather
than by any of the recovery systems previously discussed. It is the
simplest and most direct control system for VOC vapors available.
Space requirements are relatively low. In this system the air-vapor
mixture is injected into the combustion chamber of the incinerator
through a burner manifold. The design combustion chamber temperature
is usually about 1500°F. The residence time in the chamber is on the
order of one second.
Burner and combustion chamber design must accommodate a relatively
large range of flows and VOC concentrations. At the minimum this
involves automatic combustion air dampers controlled by temperature.
Normally, thermal incinerators have VOC removal efficiencies in excess
of 99 percent. Extremely rapid changes in VOC concentration are not
readily handled by the usual combustion controls, but this is not
expected to be a major problem at marine terminals.
Safety Considerations
Two major safety problems are associated with marine terminal
vapor control systems. These are fire (and/or explosion) and vessel
over-filling. Vapor concentrations in air-vapor mixtures displaced by
filling or ballasting operations are often within the flammable limits.
For gasoline vapors this is 1.4 to 8.4 percent in air. Concentrations
are usually lower than at truck terminals where returned vapors during
filling (ballasting does not apply) are often above the upper flammable
limit. Furthermore, the ultimate reservoir volume of potentially
4-13
-------
FIGURE 4-5
o
cO
w
w
u
2
W
5C
H
4-14
ENGINEERING-SCIENCE
-------
flammable materials is much larger in the case of barges and tankers
than is the case with trucks. While incorporation of vapor recovery
systems at marine terminals would eliminate the discharge of vapors at
deck level from open ullage caps, a new problem would be created by the
connection of the onshore processing facility to the vessel. Both
primary (active) and secondary (passive) safety systems can be incor-
porated within the design of a vapor recovery system. The primary
system is designed to render the vapors emitted from the vessel
non-flammable by one of three means:
(1) Dilution with air - This requires an extremely high dilution
rate to reduce the vapor concentration to a safe margin below
the lower flammable limit (1.4 percent) if it is assumed that
the initial concentration is saturated with respect to VOC
(about 50 percent in the case of gasoline).
(2) Dilution with inert gases such as N? or combustion gases free
of CO - In this case the maximum allowable Q£ concentration
is about 12 percent.
(3) Saturation - Saturation of the vapor-air mixture with butane,
natural gas, or gasoline vapors will place the mixture well
above the upper flammable limit.
In general, the dilution, inertion, or saturation processes des-
cribed above would be designed to produce the desired results as
close as possible to the point where the vapor-air mixture leaves the
vessel. The primary safety modules will take up extra space and increase
captial and operating costs. (Note: insurance costs with and without
primary safety modules could alter this factor, assuming that any
applicable safety regulations would permit construction without a
primary safety module.) The secondary (passive) safety systems include
all measures taken to reduce or eliminate sources of ignition, and to
isolate any propagation of flame or explosion beyond barrier limits.
These isolation systems incude both hydraulic and extended surface type
flash arrestors.
4-15
-------
Costs
Although no vapor recovery systems have been installed at marine
terminals, similar control technology to that just described has been
used at truck terminals for a number of years. Several cost estimates
have been made as part of studies on application of controls to marine
terminals. These studies involve differing assumptions such as the size
of the terminal, activities carried out, and the volume of VOL handled
on an annual basis. It is difficult, therefore, to compare the costs
developed. In a study by Robert Brown Associates3) two different model
terminals were considered. The first was a crude oil unloading terminal
having berths for three 120,000 DWT tankers. Ballasting emissions only
were considered. The second terminal included two berths for unloading
35,000 DWT crude oil tankers, one berth for loading gasoline to a
similar tanker, and one berth for loading gasoline into a 50,000 bbl.
barge. Various control scenarios were considered ranging from vapor
recovery on marine operations and use of external floating-roof tanks
onshore to an elaborate vapor balancing and control system interconnec-
ting the marine operations to controlled fixed-roof storage tanks.
Primary (active) safety modules were used in all cases. Capital costs
ranged from $1,390,000 for the lowest cost system on the smaller com-
bined terminal to $64,000,000 for the most expensive system on the
larger tanker terminal. On a cost-effectiveness basis, annualized
costs ranged from $25/ton VOC controlled to $4,000/ton VOC controlled.
Pullman-Kellog evaluated control technology and model systems for
a barge terminal for loading gasoline^). Two berths for 20,000 bbl.
barges were used as the basis for costs. Again, primary (active safety
modules were utilized in all control alternatives. Capital costs ranged
from $460,000 for the lowest cost system utilizing a thermal incinerator
to $902,000 for a low temperature refrigeration system. Cost-effective-
ness on an annualized cost basis ranged from $240/ton VOC controlled
to $480/ton VOC controlled.
MSA Research Corporation prepared an evaluation of a proposed
barge loading test facility^). Controls for unloading one 20,000 bbl.
barge at a time were considered. No primary safety modules were included
4-16
-------
in the analysis. Capital costs ranged from $166,000 for a lean oil
absorber to $292,000 for a low temperature refrigeration system.
Recovered gasoline credits exceeded annualized costs in all cases.
The capital recovery periods ranged from 3.7 to 18.3 years.
None of the costs given above include costs for retrofitting
tankers and barges to accommodate the vapor recovery system hardware,
including overflow protection. While comparisons are difficult, several
factors stand out:
(1) Barge vapor recovery alone is more cost effective than tanker
ballasting control because higher VOC concentrations are
present in barge return vapor flow.
(2) There are some economies of scale although they are difficult
to quantify at this time.
(3) Use of primary (active) safety modules increases costs (not
including insurance) by a considerable amount.
(4) Costs of shore-based tank controls over and above the costs
for double seal external floating-roofs results in major cost
increases.
REGULATORY STATUS
There are no regulations covering emissions from marine terminal
loading or ballasting operations at the federal or state level. Con-
trols have been considered in several cases, but none have yet been
installed. Control regulations will necessarily have to be coordinated
with safety procedures and requirements now under consideration by the
U.S. Coast Guard.
MAJOR REFERENCES
Listed References
1. "8-31 Marine Emissions Study, Final Report," American Petroleum
Institute, Washington, D.C. 20037, December 1978.
2. "Compilation of Air Pollutant Emission Factors, AP-42", including
supplements, U.S. EPA.
4-17
-------
3. "Emission Control Technology for Two Model Marine Terminals handling
Crude Oil and Gasoline," EPA-450/3-78-016, April 1978.
4. "Control Technology Evaluation for Gasoline Loading of Barges", EPA-
600/2-79-069, March 1979.
5. Status Report by MSAR to U.S. EPA on "Demonstration of Vapor Control
Technology for Gasoline Loading of Barges," Contract No. 68-02-
3141, 4 December 1979.
Other
1. "API Bulletin 2514-A: Hydrocarbon Emissions from Marine Vessels
Loading of Gasolines," American Petroleum Institute, Washington,
D.C. 20037, 1976.
2. "Background Information on Hydrocarbon Emissions from Marine Loading
Terminal Operations," Volumes I and II, EPA-450/3-76-038 a, b; U.S.
EPA, RTF, NC 27711.
4-18
-------
SECTION 5
BEER MAKING
PROCESSES AND EMISSIONS
The beer brewing process begins with malting. In this operation the
barley is soaked in water in steeping tanks and transferred to germinating
compartments for generally one week in order to convert the barley starch
into malt sugar. Once the germinated barley is transferred to the malt
drying kiln, the heat terminates the barley's growth and partly carmalizes
the malt. The malt, or dried sprouted barley, is ground finely for use
in the brewing operation. The mixture is heated in the mash kettle to
complete the conversion of starch to malt sugar.
The mash is sent to the brew kettle after being filtered through a
plate and frame filter, resulting in an extract. The spent grain filter-
cake drops from the filter into the hopper and is pumped away.
The spent grain is dewatered in a screw press and dried in a steam
drier. The solids are removed when the spent grain liquor from the screw
press is centrifuged. The solids are marketed as cattle feed once they
are mixed with the spent grain. The centrifuged liquor is returned as
make-up to the rice mash-in operation.
The malt extract is boiled in the brew kettle with the hops which
adds the flavor and the aroma. In the coolship, or clarifier, the
insoluble proteins from the hot malt extract (called wort) coagulate and
settle out. The coagulate, otherwise known as trube, is sent to a
centrifuge to be separated from the wort. The recovered liquid wort is
recycled into the process and the trube solids are added to the spent
grain. The wort flows from the coolship to a cooler where it is cooled
to fermentation temperature. During fermentation the yeast converts
the malt sugar in the wort into alcohol and (X>2 gas. The beer then
proceeds to the secondary fermentation tanks and undergoes the aging
process where, under pressure, the beer builds-up its own natural
carbonation.
5-1
-------
After the beer has aged sufficiently, it is pumped through a series
of filters for clarification. The spent filter pads are removed and
processed for reuse and the beer is sent to a sterile filling room.
See Figure 5-1 for a flow schematic.
There is a very low potential for ethanol emissions from beer
making, but no quantitative data are readily available. Several factors
mitigate against significant emissions. These include:
(1) Fermentation is conducted at relatively low temperatures,
about 4.5°C (40°F) to 14.5°C (58°F).
(2) Alcohol concentration is relatively low, generally less than
6 percent by volume.
(3) Beer is kept under pressure to avoid loss of natural carbona-
tion. It is stored at rather cold temperatures, about 0°C
(32°F).
(4) Carbon dioxide released during fermentation is recovered and
compressed for use in further carbonation of the beer.
Previous EPA reports suggest that there are virtually no volatile
organic emissions from beer making^).
There are no known control techniques directed at VOC emissions
from beer making. All known pollution control methods are directed at
reducing water pollution^'.
STATUS OF REGULATORY ACTIVITIES
Presently there are no known state or local regulations directed
specifically at controlling VOC emissions from beer making.
REFERENCES
1. EPA-450/2-78-022, "Control Techniques for Volatile Organic Emissions
from Stationary Source", May 1978.
2. EPA Technology Transfer, "Pollution Abatement in a Brewing Faci-
lity", prepared by U.S. Environmental Protection Agency, May 1974.
3. Shreve, Norris R. , "Chemical Process Industries", Fourth Edition,
McGraw Hill, 1977.
5-2
-------
FIGURE 5-1
BEER MAKING FLOW SCHEMATIC
Hot ana cold' •$"""' '"tauter
grains (filter Jui)
Brewers malt
Molt adjuncts
Hops
Yeast
Reference No. 3
35-38 Ib. "I
12-14 Ib L
'/j-lSilb.' f
?4- I Ib. -1
Per U.S. barrel
(31 gal.) of beer
5-3
ENGINEERING-SCIENCE
-------
SECTION 6
FABRIC PRINTING
PROCESSES AND EMISSIONS
This chapter discusses the four basic fabric printing operations of
roller printing, rotary screen printing, flat screen continuous prin-
ting, and flat screen unit printing and the associated fabric drying
and curing processes. Emissions generated from these operations and
applicable emission control techniques are discussed. A review of air
pollution regulations applicable to fabric printing is also presented.
Fabric printing is defined as a decorative design applied to a fabric
by intaglio (etched) roller, rotary screen, or flat screen printing
operations.
The fabric printing industry is a subset of the textile finishing
industry, which includes fabric preparation (desizing, bleaching, and
mercerizing), decorative enhancement (dyeing, printing, plisseing, and
felting), and functional enhancement (permanent press, softening, and
soil resistance). There are approximately 200 fabric printing plants
located throughout the United States, many of which perform other tex-
tile finishing operations in addition to fabric printing. A majority
of the existing fabric printing plants are located in the Northeast
and Mid-Atlantic regions of the country. However, the recent growth
in this industry has occurred in the Southeast.
A typical fabric printing plant usually uses one type of print
machine (roller, rotary screen, or flat screen). These printing
operations apply a print paste in a decorative pattern onto the fabric
that is passed through the printing machine. After printing, the fabric
enters a drying process to remove water and organic solvents so that
the fabric retains its color and pattern. Drying involves passing the
printed fabric over steam cans, which are usually exposed to the ambient
surroundings, or through direct-fired industrial drying ovens. Some
fabrics must be further processed in a curing operation. Curing may
occur in separate ovens or in a different zone within the drying oven.
After the drying process, the fabric is washed and dried to remove the
excess water.
6-1
-------
Materials
This section describes the fabrics used in the fabric printing
industry and the constituents of the print pastes applied to the
fabrics.
Fabric
The two basic categories of fabric are knit fabric and woven fabric.
Knit fabric is made by using a single continuous thread of yarn to
create individual loops and to chain each individual loop to neighboring
loops. Woven fabric is defined as "a fabric made by interlacing two or
more systems of yarns at essentially right angles to each other."*•'
Woven fabric can be further divided into broadwoven fabric which is
over 38 centimeters (15 inches) wide and woven terry towels. The yarns
used in fabric construction are spun from either man-made or natural
fibers. Man-made fibers include cellulose acetate, cellulose triace-
tate, nylon, polyester, polyolefin, and rayon. Natural fibers include
cotton, linen, silk, and wool. The composition of these fibers deter-
mines which print pastes the fabric will accept and what texture or
"hand" the fabric will have.
Print Pastes
Print pastes can be classified into the two basic categories of
dispersed dye print pastes or pigment print pastes according to the
coloring agent used in the print paste composition. Dispersed dye
coloring agents impart color onto fabrics by becoming chemically or
physically incorporated onto the individual fabric fibers. Pigment
coloring agents are insoluble particles that are physically bound to
fabric fibers, usually through the use of a polymerizing binder.
The other constituents of print pastes include a clear thickening
agent and an aqueous or organic solvent. Thickening agents are used to
facilitate the transfer of the coloring agent to the fabric during
printing and to adhere the coloring agent to the fabric during drying.
The desired print sharpness, yarn and fabric penetration, color value,
fiber type, fabric construction, dye class, and fixation conditions
dictate the selection of the proper thickening agent.2)
6-2
-------
The remaining portion of the print paste is the solvent, which acts
as the carrier for the coloring agent and the thickener.3) The solvent
may be aqueous, organic, or a mixture. The organic solvent concentra-
tion in print pastes ranges from less than one percent organic solvent
content by weight for dispersed dye print pastes to 0 to 60 percent
organic solvent content by weight for pigment print pastes. Solvents
maintain proper viscosity, aid in dye and binder dispersion, and adjust
color value, sharpness of mark, and brightness of shade. Organic
solvents used in the fabric printing industry are usually mineral
spirits. The mineral spirits used in fabric printing operations have
an average molecular weight of 155, a boiling point of 145°C (293°F),
and a vapor pressure of 8.41 millimeters of mercury (0.16 pound per
square inch) at 25°C (77°F).
Processes
This section describes the four fabric printing operations of roller
printing, rotary screen printing, flat screen continuous printing, and
flat screen unit printing and the associated fabric drying and curing
processes.
Roller Printing
Roller printing is an intaglio process in which an engraved plate
or roller is used to print- the fabric. A print paste is applied to the
engraved roller and the excess paste is removed with a doctor blade.
The fabric is then placed in contact with the roller and the pressure
between the roller and the central cylinder acts to transfer the print
paste from the incised surface of the roller to the fabric. Figure 6-1
is a diagram of a typical roller printing machine.
Fabric, padded by a backing material or back gray, is continuously
fed into the printing machine. The fabric passes between the central
cylinder and a set of printing rollers. A single color print paste is
applied to each printing roller which applies one portion of the pattern
onto the fabric. A maximum of 14 colors can be applied to a fabric by
a roller printing machine, with 8 colors being average. After printing,
the fabric passes into the drying and curing processes. The cylinders
6-3
-------
FIGURES 6-1 and 6-2
Fabri c
FIGURE 6-1 - ROLLER PRINT MACHINE
Gray
Doctor Blade
Engraved
Printing Roll
Doctor Blade
Engraved
Printing Roll
Color Box.
FIGURE 6-2 - ROTARY SCREEN PRINT MACHINE
Roller Press
\ Doctor if-
Blade
Sack Gray
6-4
ENGINEERING-SCIENCE
-------
commonly used for roller printing are copper-coated surfaces. A pattern
can either be engraved or acid-etched on these copper surfaces. The
engraved copper surface may be plated with chrome to create a longer
lasting surface.
Several of the advantages of roller printing over rotary screen
printing include sharper mark, finer line, more accurate register or
fit, finer gradation of tone, smoothness of blotch prints, and better
color bloom (vividness of color).^' Many of these advantages are
attributable to the use of medium-to-high organic solvent content print
pastes. These advantages make roller printing the preferred method for
printing apparel fabrics.
Rotary Screen Printing
A typical rotary screen printing machine is shown in Figure 6-2.
Fabric moves into the printing machine from left to right. The fabric
is pressed between a rotary screen and a back gray as the process
progresses. The fabric is attached to the back gray by a water soluble
glue. Print paste is pumped from drums into a dispenser, which
distributes the flow over the entire width of the application area. A
stationary doctor blade forces the print paste through the screen and
onto the fabric.
A screen is prepared for printing by covering it with a light
sensitive chemical and then attaching a film positive of the design
pattern over the screen. This sandwich is then exposed to light and an
insoluble lacquer forms on the areas of the screen that are exposed to
the light. The screen is then cleaned, leaving a porous screen surface
in the desired design pattern. The screen is then rolled into a tubular
shape. Rotary screens are commonly made of nickel or nickel alloys.
Some machines are able to apply a 16 color pattern, but 12 color patterns
are the maximum for most rotary screen machines.
Flat Screen Continuous Printing
Flat screen continuous printing, which is used to print drapery and
sheeting on a continuous production line, is essentially the same as
6-5
-------
rotary screen printing. The two nain differences between the two
printing processes are the fact that rotary screens are rolled, whereas
flat screens are put in a flat bed frame, and flat screen continuous
printing lines operate at a slower line speed compared to similar rotary
screen printing lines. The same photogrammetric method is used to
prepare screens for flat screen continuous printing machines as is used
for screen preparation for rotary screen printing machines.
Flat Screen Unit Printing
Flat screen unit printing is used in the printing of cut terry
towels on a semicontinuous production line. In this printing process,
cut terry towels are placed on a conveyor belt and moved into position
under a flat screen. The flat screen is then lowered onto the terry
towel and print paste is applied through the use of a squeegee that
forces the print paste through the open screen areas. This process is
repeated for every color in a pattern. The flat screens used for this
printing process are prepared in the same manner as those used in rotary
screen printing and flat screen continuous printing operations.
Dryers and Curers
After the printing process, fabrics are dried and usually cured.
Drying drives off water and organic solvents so that the colors will be
retained (color fixed) on the fabric. Some fabrics are color fixed
during curing, which may be an entirely separate process or merely a
separate segment of the drying process. Resin bonded pigment print
pastes require curing. Dispersed dye print pastes may require curing
but are usually color fixed during drying. Some color fixed dyes
require aging in a high-heat, high-humidity environment. After the
fabric passes through the drying process it is washed to remove unfixed
dyes or pigments. The fabric is then dried again to remove the excess
water.
Drying is accomplished through the use of industrial drying ovens
(convection drying) or steam cans (conduction drying). Industrial
drying ovens are heated by pressurized steam coils (indirect heat
dryers) or by fossil fuel combustion (direct heat dryers). Screen
6-6
-------
printed fabrics are usually dried in industrial drying ovens because
knit fabrics can be dried without tension. Fabrics requiring a soft
hand, such as apparels, are usually dried by conduction on steam cans.
Conduction drying is by direct contact of the fabric with the steam
cans.
Industrial drying ovens have a temperature range of 149 to 177°C
(300 to 350°F). The surface temperature of steam cans ranges from 93
to 135°C (200 to 275°F). Fabric residence time is 2 to 4 minutes in
industrial drying ovens and 1 to 2 minutes on steam cans. Exhaust flow
rates range from 18 to 600 cubic meters per minute (6,000 to 20,000
cubic feet per minute) for industrial drying ovens, and are dependent on
fabric residence time, oven temperature, fabric weight, and organic
solvent content and water content of the print pastes.
Curers are also heated by pressurized steam cans or fossil fuel
burners. Steam can curers usually operate at atmospheric pressure
under saturated steam conditions and temperatures of 163 to 191°C (325
to 375°F). Fossil fuel fired curers operate at 149 to 204°C (300 to
400°F) with exhaust flow rates of 90 to 210 cubic meters per minute
(3,000 to 7,000 cubic feet per minute). Fabric residence times for
curers are 2 to 4 minutes.^) Sheeting is usually dried and cured in
different zones of the same oven. Decorative and apparel fabrics are
usually dried and cured in separate ovens.
Emissions and Emission Sources
The sources of fugitive VOC emissions from fabric printing
operations (roller printing, rotary screen printing, and flat screen
continuous and unit printing) include evaporation from the wastewater
stream, open print paste barrels, printing troughs, the printing rol-
lers and screens, strike through onto the backing material or back
gray, and the printed fabric prior to reaching the drying process.
The steam can drying process and industrial drying ovens used to dry
printed fabric are considered to be the most significant sources of
VOC emissions. The steam can drying lines can be enclosed or exposed
to the ambient surroundings and VOC emissions generated during steam
can drying are exhausted through the use of roof vents and fans. When
6-7
-------
industrial drying ovens are used in the fabric drying process, the VOC
emissions are exhausted through a stack. Most of the VOC emissions
are generated during fabric drying, with only small quantities being
generated during fabric curing. Figures 6-3 and 6-4 indicate sources
of VOC emissions from fabric printing operations.
Operating parameters for two model printing lines are summarized in
Table 6-1. These parameters are based on data obtained from fabric
printing companies."1'' Model printing line one (MPL 1) uses print
pastes with a weighted average organic solvent content of 30 percent by
weight and an associated drying exhaust flow rate of 255 standard cubic
meters (8,500 standard cubic feet) per minute. Model printing line two
(MPL 2) uses print pastes with a weighted average organic solvent
content of 5 percent by weight and an associated drying exhaust flow
rate of 240 standard cubic meters (8,000 standard cubic feet) per
minute.
CONTROL TECHNOLOGY
The methods available for the control of VOC emissions from fabric
printing operations can be classified as process modification or add-on
VOC emission control systems. Process modification consists of reduc-
tion in the amount of organic solvent in the print paste composition
(print paste reformulation). Add-on VOC emission control systems are
primarily thermal incineration.
Print Paste Reformulation
There is little or no cost associated with the reformulation of the
print pastes to reduce their organic solvent content.8) in fact there
is a credit for the reduction in organic solvent usage achieved. As a
result, most fabric printing operations have switched from higher
organic solvent content print pastes to low organic solvent content (10
percent by weight or less) or all aqueous print pastes.9,10,11,12,13)
The major reason for this changeover is the increasing cost of the
organic solvents used in fabric printing. These solvents have increased
in price from about $0.08 per liter ($0.31 per gallon) in 1973 to a
current price of about $0.37 per liter ($1.39 per gallon).
6-8
-------
FIGURE 6-3 and 6-4
FIGURE 6-3 - ROLLER PRINTING LINE AND ASSOCIATED STEAM CAN DRYING PROCESS
STEAM CANS
FUGITIVE VOC EMISSIONS
6RAVURE ROLLER
LIKT DOCTOR
BRUSH ROLLER. , „
DRY BACK GREY
PRINT PASTE
FIGURE 6-4 - ROTARY SCREEN PRINTING LINE AND ASSOCIATED DRYING AND CURING OVEN
i
I
\ FUGITIVE VOC EMISSIONS
| STACK EMISSIONS
FUGITIVE WATER EMISSIONS
VENT TO
ATMOSPHERE
DRYING AND CURING
OVEN
BLEACHED
FABRIC
MIXER
cL
PRINT PASTE
CONTINUOUS BELT
PRINT PASTE -
""PRINT PASTE
TRANSPORTED TO
PRINTING AREA
6-9
ENGINEERING-SCIENCE
-------
TABLE 6-1
MODEL PRINTING LINE PARAMETERS
MPL 1
MPL 2 -
1. Fabric
A.
B.
C.
D.
Type (% woven/% knit) :
Fiber (% natural/% man-made):
Average weight, kilograms /meter:
(pounds /yard) :
Average width, centimeters:
(inches) :
70/30
30/70
0.13
(0.25)
137
(54)
70/30
30/70
0.13
(0.25)
137
(54)
2. Print Pastes
A. Annual print paste consumption,
megagratas (tons):
B. Organic solvent content,
megagrams (tons):
C. Organic solvent (%):
3. Production Data
A. Capacity utilization (%):
B. Operating efficiency (%):
C. Average line speed, meters/minute:
(yards/minute):
D. Annual fabric production,
meters (yards) x 10°:
4. Drying Process
A. Type of drying:
B. Fuel:
C. Temperature:
D. Exhaust flow, m-Vminute:
(ft3/minute):
5. Estimated VOC Emissions from Drying
379 (417)
114 (125)
30
379 (417
19 (21)
5
84
37
41
(45)
84
37
41
(45)
5.4 (5.0)
Enclosed
steam cans
No. 6 fuel oil
116°C (240°F)
255
(8,500)
5.4 (5.0)
Drying oven
Natural gas
163°C (325°F)
240
(8,000)
A.
B.
C.
Annual VOC emissions, megagrams:
(tons) :
VOC emissions per operating
hour, kilograms (pounds):
VOC exhaust stream concentration,
parts per million:
98
(108)
53 (117)
1,240
16
(18)
9 (20)
207
MPL = Model Printing Line.
6-10
-------
There are, however, a few fabric printing plants that use print
pastes with a weighted average organic solvent content of 30 to 45
percent by weight (which can go up to 60 percent in some cases). In
most cases, it appears these plants have established a special market
for their products (the printing of very dark colors with a soft feel
or "hand") and it does not appear that these plants could switch to a
lower organic solvent print paste and still maintain this special market
niche.
Add-on VOC Emission Control Systems
The only technically feasible add-on VOC control technique appears
to be incineration. The high moisture content of the exhaust stream
associated with the drying of printed fabric leads to freezing problems
in the liquid nitrogen condenser stage of inert condensation recovery
systems. ' The relatively high vapor pressure of the organic solvents
used in fabric printing operations and the low VOC concentration in the
drying exhaust stream cause combination cooler/condenser and electro-
static precipitator systems to have a low VOC removal efficiency.15,16)
Activated carbon adsorption systems may be applicable to fabric printing
VOC control but this has not yet been verified experimentally.
Thermal incineration of VOC emissions is a frequently used VOC
emission reduction technique that is applicable to many types of VOC
emissions and industrial processes. In a thermal incinerator, a burner
located at the inlet of a thermal incinerator maintains a temperature
of 649 to 871°C (1200 to 1600°F). Contaminated air enters through the
burner flame zone and is retained in the combustion zone for about 0.50
to 0.75 second. The VOC emissions in the exhaust stream are oxidized
in the combustion chamber and the hot combustion gases are exhausted
through a stack. The main disadvantage of incineration is the large
amount of auxiliary fuel that is required to maintain the high tempera-
tures necessary for oxidation of VOC emissions.
An effective means of reducing auxiliary fuel costs with incinera-
tion is by recovery of useful heat from the incinerator unit. Some
applications of this heat include preheating the dryer inlet air,
preheating the dryer exhaust before it enters the incinerator, and
producing steam for building heat or production processes. In most
6-11
-------
industries, heat is recovered through recuperative shell and tube, air
to air heat exchangers, or by regenerative heat wheels. Recuperative
heat recovery up to 70 percent is achievable through the proper design
of the thermal incinerator and associated heat recovery system.1''
Control cost estimates have been developed for the installation of
a thermal incineration system with 70 percent recuperative heat recovery
capability on each of the two model printing lines presented previously.
The total installed capital cost and total annualized cost for the
thermal incineration systems are presented in Table 6-2. As indicated
in Table 6-2, the retrofit penalty and hooding, ducting, and enclosure
costs associated with MPL 1 are higher than those associated with MPL
2. This is due to the fact that MPL 1 uses steam can drying which is
more difficult to retrofit and enclose compared with the industrial
drying oven used for MPL 2. However, the higher VOC concentration of
the drying exhaust streams for MPL 1 as compared with MPL 2 results
in a lower net fuel requirement for the thermal incineration system,
and, therefore, an incinerator for MPL 1 would have a lower operating
cost than MPL 2. Incinerator equipment costs are an average based on
quotes from three thermal incinerator manufacturers. 18,19,20) Hood-
ing, ducting, and enclosure costs are estimates from a sheet metal
working contractor.21) The fuel requirements were based on energy use
calculations. Operating and maintenance labor, retrofit penalties,
installation costs, capital recovery factors, and taxes, insurance,
and administration factors are from standard references.22)
REGULATORY STATUS
State regulations pertaining to the reduction of VOC emissions
from fabric printing operations are summarized in Table 6-3. State
regulatory agencies have expressed uncertainty about the applicability
of these VOC emission control regulations to fabric printing opera-
tions. 23,24)
6-12
-------
TABLE 6-2
CONTROL COSTS FOR THERMAL INCINERATION SYSTEMS WITH
70 PERCENT RECUPERATIVE HEAT RECOVERY CAPABILITIES
MPL 1
MPL 2
Equipment cost
Installation costs
Retrofit penalty
Hooding, ducting, enclosures
Total installed capital cost
Direct
$145,000
88,000
64,000
45,000
$342,000
$145,000
88,000
46,000
20,000
8299,000
Operating labor
Maintenance labor
Net fuel
Electricity
Subtotal
Indirect
Capital recovery3
Taxes, insurance & administration^3
Subtotal
Total annualized costc
$ 4,000
4,000
16,000
1,000
$25,000
$56,000
14,000
$70,000
$95,000
$ 4,000
4,000
24,000
2,000
$34,000
$49,000
12,000
$61,000
$95,000
aCapital recovery factor of 16.275 percent of total installed capital
cost. Based on 10 year life of incinerator and 10 percent interest
rate.
"Four percent of total installed capital cost.
GDirect subtotal + indirect subtotal.
MPL = Model Printing Line.
Note: All costs are rounded to the nearest $1,000.
6-13
-------
TABLE 6-3
SUMMARY OF STATE VOC EMISSION REGULATIONS3
State
Summary of Regulation
Alabama
Delaware
New York
Georgia
South Carolina
North Carolina
Florida
Maine
New Jersey
Connecticut
Pennsylvania
California
Virginia
Massachusetts
Rhode Island
No regulation unless the source emits over 100 tons per
year of VOC. If greater than 100 tons per year of VOC,
then source has to apply BACT.b
No applicable regulation.
Fabric printing operations may be covered by regulation
for fabric coating operations, depending on percent
of coverage of material. This only applies to sources
that emit greater than 100 tons per year of VOC.
No applicable regulation. If the source emits greater
than 100 tons per year of VOC, or if it is a new
source, then BACT must be met for PSDC requirements.
Regulation for roller printing which sets a VOC
emission limit of 550 pounds per day or 150 pounds
per hour. This may be applicable to fabric roller
printing operations.
General hydrocarbon regulation which limits emissions
of photochemically reactive materials to 40 pounds
per day or an emission reduction of 85 percent.
No applicable regulation.
Any existing source that has the potential to emit
more than 100 tons per year of VOC must reduce
emissions by 90 percent. There are no exemptions,
and compliance is on a source by source basis.
No applicable regulation.
No applicable regulation.
No applicable regulation.
No applicable regulation.
No applicable regulation.
No applicable regulation.
A general regulation on organic solvents. This
regulation would probably apply to the organic
solvents used in fabric printing print pastes.
aThis table addresses State mass emission rate regulations for the con-
trol of VOC emissions. States have "odor" and "nuisance" regulations
that may be applicable to the control of VOC emissions from fabric
printing operations.
bfiest Available Control Technology.
°Prevention of Significant Deterioration.
6-14
-------
REFERENCES
1. Storey, Joyce. The Thames and Hudson Manual of Textile Printing.
London, Thames and Hudson Ltd., 1974. 188 p.
2. Thomas, R. J. Dyeing Mechanism in Textile Printing. Special topics
in Fabric Printing. P.D. No. 41. Clemson, Clemson University
Press.
3. Clark, W. An Introduction to Textile Printing. Textile Book
Service. 1974. 289 p.
4. Letter from Steenland, William H. , ATMI, to Gasperecz, Greg, EPA:
CPB. August 16, 1979. ATMI comments on Phase I document.
5. U.S. Environmental Protection Agency. Phase I: Textile Printing
(Final Report). U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. October 1979.
6. Memo from York, Steven, RTI, to Johnson, William, EPA:CPB. March
28, 1980. Recommendations on Model Plants and Regulatory Alterna-
tives.
7. Memo from Viconovic, George, GCA/Technology Division, to Porter,
Fred L., EPArESED. October 9, 1980. Environmental and Cost
Analyses.
8. Final Trip Report from Viconovic, George, GCA/Technology Division,
to Porter, Fred L. , EPArESED. October 28, 1980. Applikay Textile
Process Corporation.
9. Telecon. Viconovic, George, GCA/Technology Division, with Cuc-
cirelli, Joseph, Dove Processing Company. January 8, 1981.
Organic solvent usage.
10. Telecon. Viconovic, George, GCA/Technology Division, with Murrow,
Jack, A.T.P. Processors. January 6, 1981. Organic solvent usage.
11. Telecon. Viconovic, George, GCA/Technology Division, with Bolton,
Ed, Duro Textile Printers. January 6, 1981. Organic solvent usage.
12. Telecon. Viconovic, George, GCA/Technology Division, with Grabow,
J. , Brewster Finishing Company. January 6, 1981. Organic solvent
usage.
13. Telecon. Tippitt, William, EPA:ESED, with Technical Representative,
Inmont Corporation. January 5, 1981. Industry trends in organic
solvent usage.
14. Airco Solvent Recovery System (Product brochure). March 1980.
Airco Industrial Gases, Murray Hill, New Jersey.
6-15
-------
15. Letter and Attachments from Pannell, K. E. , Exxon Company, to
Patinskas, John, GCA/Technology Division. December 9, 1980.
Technical data on Varsol 1 and Varsol 18.
16. Memo from Ryan, Ron, GCA/Technology Division, to Viconovic, George,
GCA/Technology Division. December 15, 1980. Vaporization and
condensation of fabric printing solvents.
17. Memo from Mascone, David C. , EPArCPB, to Farmer, Jack, EPArCPB.
June 11, 1980. Thermal Incinerator Performance for NSPS.
18. Telecon. Viconovic, George, GCA/Technology Division, with Sales
Representative, Combustion Engineering. November 18, 1980.
Thermal incinerator costs.
19. Telecon. Viconovic, George, GCA/Technology Division, with Sales
Representative, Peabody International Inc. November 18, 1980.
Thermal incinerator costs.
20. Telecon. Viconovic, George, GCA/Technology Division, with Sales
Representative, John Zink Inc. November 19, 1980. Thermal
incinerator costs.
21. Telecon. Viconovic, George, GCA/Technology Division, with Hamlin,
Robert, Hamlin Sheet Metal Company. November 19, 1980. Sheet
metal cost estimates.
22. Neveril, R. B., CARD Inc. Capital and Operating Costs of Selected
Air Pollution Control Systems. U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. December 1978.
23. Telecon. Viconovic, George, GCA/Technology Division, with Hender-
son, Thomas L., Virginia Air Pollution Control Agency. January
26, 1981. State VOC regulations.
24. Telecon. Pilcher, Lester Y., GCA/Technology Division, with Manly,
Paul, New Jersey Department of Environmental Quality. January 27,
1981. State VOC regulations.
6-16
-------
SECTION 7
FLARES
PROCESSES AND EMISSIONS
Flares are used for the disposal of combustible waste gases.
Their use is most commonly associated with petroleum refineries but
they are also used in petrochemical plants, oil fields, steel mills,
sewage disposal plants, and could be necessary in coal gasification
plants. Three broad classifications, of operational conditions are
associated with the use of flares. These are (1) process start-up and
shutdown venting, (2) minor operational upsets not causing unit or
plant shutdown, and (3) emergency release of process materials as a
result of power failure or major equipment failure. An example of the
first use would be controlled venting of a vessel prior to performing
maintenance; the second category could include flaring of gases from
leaking pressure relief valves tied into a manifold system; and the
last category could cover the massive diversion of the full process
flow. From an operational point of view, flares can be distinguished
from thermal incinerators by the requirement to operate under a wide
range of conditions. Flares designed to handle emergency releases
could be required to handle gas flows ranging from several hundreds of
cubic meters over a relatively short period of time. Obviously they
must be designed to handle the maximum flow possible under worst-case
conditions. Further, these gases may range from hydrogen-rich reformer
gases to liquefied petroleum gases.
True flares are classified as either elevated or ground-level.
Simple burning pits have also been utilized but their use is generally
reserved for extremely large gas flows under catastrophic conditions
when the design capacity of the primary flare is exceeded. The elevated
flare is the most commonly used type for emergency purposes. The
taller elevated flares require a support structure which increases
costs, but can be located relatively close to other units and still
meet the distance standards associated with radiant heat transfer from
the open flame at the tip of the flare. Further, in the case of a
flame-out, the flammable gases are not released, even momentarily, at
ground level.
7-1
-------
Modern flares must be designed to be safe, minimize noise, and be
smokeless. The essential parts of a flare are the burner, the stack,
seal, liquid trap, controls, pilot burner, and ignition system. A
simplified sketch of a typical elevated flare is given in Figure 7-1.
A typical refinery blowdown system used in conjunction with the flare
is shown in Figure 7-2.
From an air emissions point of view, the principal design and
operational objective associated with waste gas flares has been the
achievement of smokeless operation. Little attention has been given to
the systematic evaluation or control of VOC emissions. The control
problem associated with smokeless operation is the injection and
thorough mixing of the proper amount of combustion air in order to
eliminate or minimize those flame reactions which lead to the formation
and agglomeration of soot nuclei. This is not easily done in a rela-
tively simple flare which must handle a wide range (up to very large
volumes) of flare gas flow rates. Several possible aproaches to
introducing and mixing sufficient quantities of combustion air are
given below:
0 Direct Air Injection
forced air (mechanical blower)
induced air (venturi burners)
0 Assisted Air Injection
steam jet
other inert gas (e.g. N2)
The most popular smokeless type flare, until recent years, has been the
steam assisted type, although some multi-tip induced air flares have
also been used. Mechanical (forced air) flares were more costly and
difficult to control. As the cost of steam has increased, advanced
forced air flares have gained in popularity. The use of compressed
nitrogen or other inert gas air injection methods have not been cost
competitive.
Steam injection type elevated flares predominate in the emergency
disposal of combustible waste gases by thermal oxidation. The main
action of steam injection is to aspirate air and improve molecular
mixing of fuel and air by mechanical and dynamic means. It has also
7-2
-------
FIGURE 7-1
vP
<
5
H
CO
oo
W
23
nJ
(it
I
§
g
OO
00
w
ij
8
CO
o
w
oo
U)
8
P
cc o
J
61
UJ
£
I \
I
>
1
cC
2
-------
FIGURE 7-2
TYPICAL MODERN REFINERY SLOWDOWN SYSTEM
TO PURE STACK
LIGHT-ENDS CONOENSATE RECOVERY
7-4
ENGINEERING-SCIENCE
-------
been suggested that the steam reacts with fuel to form oxygenated
compounds that burn readily at relatively low temperatures, and that
water gas reactions take place with the same general effect-'-). Tests
conducted recently on an experimental flare tend to show that the effect
of steam addition on chemical reactions is slight compared to its
dynamic contribution towards better combustion^).
There are a variety of combustion tip designs for elevated flares
utilizing steam injection. Variations in method and location of steam
injection and location of air introduction are the most significant
variables. Siegel^) has illustrated a number of flare head designs
(Figure 7-3). In flare heads where air is mixed after the gas exits, the
flame stabilizes away from the flare mouth. Flare heads using this
design operate at lower temperatures in the tip region, and potentially
are less subject to corrosion and clogging of steam jet openings with
soot. Premixing type flare heads have better flame stability and have
less tendency for soot formation.
Important design factors for steam injected elevated flares include
gas flow rate and the ratio of steam to flare gas. The stack gas design
velocity should be as high as possible to provide increased turbulence
and to prevent flashback. The optimum tip velocity range for one
design cited by the Air Pollution Engineering Manual^-) is 90 to 120
mps. This velocity will vary with design and gas composition. In
order to prevent flame blow-out, the maximum velocity is limited to
about 150 taps. The mass ratio of steam to flare gas generally should
be in the range of 0.2 to 0.5D. In the experimental flare tests
reported by Siegel^), the following visual results were observed at
different steam to air mass ratios:
Steam/Gas Ratio (Kg/Kg) Flame Appearance
0.17 slightly sooty peaks
0.38 soot-free
0.52 almost non-luminous
At ratios over 1.0, carbon monoxide concentrations began to increase,
but at ratios as high as about 1.8 there was little increase in unburned
hydrocarbon. It should be noted that the tests conducted by Siegel
7-5
-------
FIGURE 7-3
DIFFERENT DESIGNS OF ELEVATED FLARE HEADS2)
Air-admixing after the exit
of the gas
'
T
Steam
internal steam
nozzle
external steam
nozzle in one
nlane
efore the exi
c~ the g
stcair
Steam
Chs
2. external steam nozzles
in several planes •
air
V
Steam
3. internal steam-
and gas nozzles
(Bunscn system)
4. annular gap
injectors
(wall jot)
5. injector rods
(wall jet)
ENGINEERING-SCIENCE
-------
cannot be considered completely representative of normal flare condi-
tions. The hydrogen concentration of the flare gases was high (mostly
between 50 and 64 percent by volume) and the heat content high. Also,
the partially premixed flane is not typical of flares in this country.
At excessive stean/gas ratios flame cooling could reach the point
where combustion no longer takes place. From the above it is apparent
that steam flow should be controlled as a function of flare gas flow.
This can be done manually or with automatic controls. In the latter
case, various means of sensing gas flow and using this parameter to
control steam flow have been used. Two variations of flow measurement
orifices are shown in Figures 7-4 and 7-5.
Direct air induction using multiple venturi burner sets have been
used on some ground flares. These depend on sequentially placing
additional burners on-line as flow increases. Both pressure activated
valves and water sealed drums set to release at different pressures
have been used to add burners to the circuit as flow increases. Figures
7-6 and 7-7 illustrate two different types of multi-jet burners.
One flare construction firm representative-^) stated that sales of
forced air flares using centrifugal blowers have increased in the past
several years due to the rapidly increasing costs of steam. Capacities
as high as 100,000 Ibs./hr. of flare gas are available. Improved blower
controls have aided in making this type of flare more effective in
smoke suppression. Figure 7-8 illustrates such a flare.
Data on emissions from flares is very sparse. Tests conducted by
Siegel ' showed conversion efficiencies of organic carbon to COj
of 98 to 99 percent. Because this was a relatively small test flare
and hydrogen content unusually high, the results cannot be considered
typical for a full-scale smokeless flare handling relatively high
molecular weight hydrocarbons.
7-7
-------
FIGURE 7-4
WASTE-GAS FLARE SYSTEM USING ESSO-TYPE BURNER
REGULATED WITH SLOTTED ORIFICE
INSTRUMENT Al*
SHALL FLOI
FLOI """ 5"
CONTtOlLEIt
PHESSUM SCNSOU
RATIO
»EHT
IASTE SAJ
1 HESSUHE t»PS
HICK HI
fLOE A»»ESIO«
*T—i n
^SLOTTED III
Lj^J U|
7-8
ENGINEERING-SCIENCE
-------
FIGURE 7-5
DIAGRAM OF WASTE-GAS FLARE SYSTEM
USING A SINCLAIR BURNER USING IN-LINE ORIFICE FOR REGULATION
7-9
ENGINEERING-SCIENCE
-------
FIGURE 7-6
•x
w
§
a
PS
H
2!
PU
><
H
60
C
•H
C
C
3
o
T3 O u"i
C C 0>
CQ *"H
0)
CO !-4 -
M CO <1)
0) 4J
CCS
M O -U
3 t-i i-l
to cn
o a) M
•HOE
a 3
4-1 a)
V-4 O iH
o
-------
FIGURE 7-7
w
I
M
H
Cs
O
4J
0)
7-11
ENGINEERING-SCIENCE
-------
FIGURE 7-8
SCHEMATIC OF AN AIR
ASSISTED SMOKELESS
FLARE (Courtesy of
John Zink Co.)
Flare Gas
7-12 -
-------
CONTROL TECHNOLOGY
Design of flares to meet existing air pollution emission standards
has been directed towards achieving reliable smokeless operation and
not to reduction of VOC emissions. Information presented in the above
section indicates VOC emissions per unit of fuel burned can be higher
from a smokeless flare than those from a properly designed boiler.
Further, it is possible for VOC emissions to increase without con-
current smoke formation in steam assisted flares if the steam/fuel
ratio is too high.
Several alternatives appear to be possible for reducing overall
emissions from flare operation. Not all are feasible or even possible
at all locations. These alternatives are listed below:
1. Process Changes
0 Reduce potential for flaring by improved process control.
0 Reduce quantities of material routinely flared by sub-
stituting vapor recovery systems.
0 Utilize more of the flare gases in fuel gas systems where
combustion takes place in well controlled boilers or fur-
naces.
2. Flare Design Changes
0 Improve combustion by improvement in burner design, inclu-
ding staged use of burners and better combustion air
mixing.
0 Provide closer control of steam/flare gas ratio over wide
range of flare gas rates.
0 Consider other methods of combustion air introduction such
as blowers.
7-13
-------
Additional information is necessary to quantify emissions from
flares and to determine the most effective control measures. The U.S.
EPA has two active contracts to obtain such information^). The first
is a short-term contract to develop test methods. The second is a
longer-range program to determine flare efficiency under various con-
ditions. No data fron these programs are yet available.
In a study for the U.S. EPA conducted by I.T. Enviroscience^)
of the synthetic organic chemicals manufacturing industry (SOCMI),
alternative means of handling flare gases were examined. This report
suggests that energy and cost aspects of flaring make the use of flare
gases as fuel an attractive option where technologically feasible.
Cost benefit curves for elevated flares, ground flares, and a fuel gas
system are reproduced in Figures 7-9, 7-10, and 7-11. Data should be
used for comparison as they were developed using a particular design
basis.
STATUS OF REGULATORY ACTION
There are no source-specific rules pertaining to VOC emissions
from flare gas systems at the federal, state, or local level. Many
states and local agencies, however, regulate smoke from flares under
general prohibitions on opacity of visible emissions. To some extent
these regulations indirectly serve to reduce VOC emissions as compared
to uncontrolled smoke-producing flares.
The California Air Resources Board has conducted a survey of flare
use in California, but a proposed model rule has not yet been developed.
7-14
-------
FIGURE 7-9
in
S
w
H
CO
>•<
CO
w
ai
I
o
w
H
w
hJ
a
2!
<
:*
CQ
Q
W
H
CO
w
Q
CO
CO
Z
w
CJ
w
w
H
CO
o
o
O
O
o
o
o
o
o
o
c
—
>*
a.
o
in
cj
o
n
o.
c
O
O
ji
O
to
•o
LU
CM
CM
O
CM
CO
O
(D
O
1SOQ
CM
O
CM
O
1SO-JI S6UJABS
7-15
-------
FIGURE 7-10
H
CO
O
o
M
H
w
PU
O
CO
CO
§
O
H
CO
>-"
CO
w
Q
§
( 0001$) isoo Bui.iejado lenuuv
O
O
O
O
a
o
CO
a
o
a
c.
ca
O
O
LL
1
T3
O
w
o
7-16
ENGINEERING-SCIENCE
-------
FIGURE 7-11
H
CO
O
I
w
e
<
Q
W
I
H
to
CJ
O
cn
en
w
!25
W
H
U
W
Pti
ft,
H
CO
O
S5U1A2S
/$)
7-17
ENGINEERING-SCIENCE
-------
REFERENCES
1. "Air Pollution Engineering Manual," AP-40, Second Ed., U.S. EPA,
OAQPS, RTP, NC, May 1973.
2. K. D. Siegel, "Degree of Conversion of Flare Gas in Refinery High
Flares. Pollutant Emission from Refinery High Flares as a Function
of Their Operating Conditions," dissertation for the degree of
Ph.D. in Engineering Science at the Chemical Engineering Department
of the University in Karlsruhe, West Germany, February 1980.
3. Personal communication from Dick Bell, the John Zink Company, to
R. J. Bryan, Engineering-Science, Inc., February 26, 1981.
4. Personal communication between Les Evans, U.S. EPA, and R. J. Bryan,
Engineering-Science, Inc., February 3, 1981.
5. "Emission Control Options for the Synthetic Organic Chemicals
Manufacturing Industry: Flares and the Use of Emissions as Fuels,"
prepared by I.T. Enviroscience for the U.S. EPA, OAQPS, ESED, RTP,
NC, Contract No. 68-02-2577, August 1980.
7-18
-------
SECTION 8
LUBE OIL MANUFACTURING
PROCESSES AND EMISSIONS
The heavy distillates and residues provide heavy oils for various
uses such as waxes and lubricating oils. These distillates are also
hydrocracked to lighter distillate fuels and gasoline. In 1972, lubri-
cating oils and greases accounted for 67.8 million barrels of the 4,280
million barrels of crude run to stills-*-). Both solvent extraction
and chemical treatment have been standard practices in the upgrading
of lubricants. The use of additives (0.001 to 25 percent or more)
such as antioxidants, detergents, extreme-pressure agents, antifoam
compounds, viscosity index improvers and antiscuff agents are used to
improve the performance of most lubricants.
Lubricating oils are manufactured from paraffin-based oils and
mixed-base oils. When lubricating oils are produced from the mixed-
base oils, they generally require acid or solvent treatment.
Solvent treating or extraction is one of the most widely used
processes; however, there are numerous methods and a wide variety of
heavy-oil stocks. Figure 8-1 illustrates the general flow of
operations presently conducted in lube oil manufacture. The solvent
treatment rids the oil of a majority of the dark colored materials.
However, the older, more conventional treatments such as clay percola-
tion or contacting usually must also be incorporated.
The Pennsylvanian stocks are finishable by contacting or percola-
tion alone. Both acid and clay treatment are necessary for lube stocks
from superior lube crude oil. Catalytic desulfurization and hydrogena-
tion are also suggested as a method of preparing superior raw lubricating
oil stocks.
Vacuum distillation in the presence of a small amount of caustic
soda, or soda ash, acid treatment and soda ash neutralization of each
of the several lube fractions, and percolation clay treatment result in
the manufacture of low-cold-test lubricating oils. The lightest lube
8-1
-------
FIGURE 8-1
SCHEMATIC DIAGRAM OF A REFINERY FOR PRODUCING LUBRICATING OILS
CRUDE
VACUUM
FRACT10NATION ! FRACTONA- ' £ •
I TION I < <
Gasoline
Oli
Sat
tX (ft
Ul LU
lOEASPHALT-l SOLVENT I SOLVENT, I DECOLORIZING | BLENDING
ING ' REFINING | DEWAXING |
I
£t
I Gas oil
Reduced crude
Raffinate
tor
Medium!
-Ph.
I Heavy | j
CfcbLL
'Extra | —
Iheavy'
' 'II
o «>
Short (
residuum I
Asphalt I Extract
1
' \
\
s
1
1
Solvent
recovery
I
1
' 1
1
tl
Is
& •
ackl wa
Si?
?«2
« 5-0
I
1
'Dew
1 °
Crud
seal
ued
1
|
I
n
O
L_
Paraffi
r~
3
S
X
Clay «nd
e
e
SAE.
motor oils
Paraffin
wax and
petrolatum
Schematic diagram of a refinery for producing lubricating oils.
Reference 2
3-2
ENGINEERING-SCIENCE
-------
fraction, naphtha, is used to dilute the viscous oils before treating.
For the lightest oils, the acid ranges from 10 Ibs./bbl. to 40 Ibs. for
the heaviest (150 to 200 viscosity at 210°F), and the range for the
soda ash neutralization is from 1 to 3 Ibs./bbl.
Batch agitators are usually used for the acid treatment of lube
stocks. Table 8-1 presents the various operations of treating oils (the
characteristics of which vary greatly). Neutralization with soda ash
or caustic is achieved by agitating gently with 3 to 15°Be' caustic
until neutral, spraying with hot water or steaming and settling for 4
to 15 hours; several washings and settlings are generally required.
The oil can then be brightened by heating to 120 to 200°F by steam
coils in an open pan; heating and blowing with air; or by agitating
with dry Sil-0-Cel, etc., and filtering. Neutralization by ammonia is
becoming increasingly popular since the highly stable emulsions which
form during neutralization via caustic or soda do not form with the
ammonia method, nor do they form with the clay contact process.
Treatment of lubricating oils by centrifuges or horizontal mixers
is frequently done. In this process the oil is heated and mixed with
acid for about 10 minutes by a mechanical and/or reaction tank. Water
is then added to prepare the sludge, and discharged through the centri-
fuges. As the sludge accumulates in the hopper, it is continuously
pumped to fuel oil mixing tanks or to an acid recovery system.
Lube Oil Hydrotreating Process
Hydrotreating, though not the most conventional method of petroleum
product production, yields a full scope of finished single-grade or
multigrade lube oils with a wide range yield-viscosity distribution,
in addition to valuable byproducts consisting of gasoline, naphtha,
kerosine, furnace oil, and waxes (see Figure 8-2).
Makeup and recycle hydrogen along with oil feed are charged to the
reactor where viscosity index improvement, desulfurization, denitrogena-
tion, carbon residue reduction, and demetallization are obtained. The
bottoms (total lube product) is dewaxed after the hydrogen-rich recycle
gas stream is flashed from the reactor effluent and the liquid product
8-3
-------
TABLE 8-1
TREATMENT OF OILS
Distilled Residual
Cylinder Cylinder 225 Pale Average
Stock Stock Neutral Oils Oils
Refiner
Pounds water acid, 66°Be',
agitate 1/2 hr.
Hours of settle and draw
Pounds acting acid, 66°Be',
agitate 1/2-1 hr.
Hours of settle and draw
Water wash
Neutralized by
4
6
Clay
Contact
32
4
Clay
Contact
1
2
4
9
II
4
4
III
2-12
2-31
10-18a 20-60
12-20 2-8
Yes
(150 gal. per
1,000 bbl.)
Clay Caustic Caustic
Contact
a) 98 percent acid.
Reference No. 2
8-4
-------
FIGURE 8-2
CJ
s
eo
f—
C3
0= co
u. —i irj
CJ
•«:
_j
CO
O
cc.
O
LJU
CO
CJ
CJ
X
E
—I
1
V
t
CO C3
CO h—
CS CO
CO
1)
U
ol
S-t
D
3-5
ENGINEERING-SCIENCE
-------
is charged to a stripper to remove oil and the lighter products. The
various viscosity lube oils are produced by vacuum distillation of the
dewaxed oil. The distillation can also be done prior to the dewaxing
for other desired fractions.
Process Emissions and Emission Points from Solvent Treating and
Hydrotreating
Extraction involves the removal of a component from a liquid by
means of the selective solvent action of another liquid. This procedure
facilitates the removing of low-viscosity index hydrocarbons, unusable
sludges and colored materials from lubricating oil. Generally, the
extraction is countercurrent. This presents two difficulties: obtain-
ing solution equilibrium and separating the two immiscible phases. As
an example, the extractive refining of lubricating oils with furfural
(C5H402) is illustrated in Figure 8-3. Color bodies, sulfur compounds,
and oxygen-containing molecules are effectively removed from lube-oil
stock by furfural. As shown in Figure 8-4, of a typical solvent-refining
process of lubricating oil, the oil is mixed with the solvent or solvents
in an extractor column. Use of the proper solvent, facilitates the
separation of the mixture into two layers, one of which is rich in
solvent and containing the dissolved impurities (extract), and the
other containing little solvent and most of the desirable oil (raf-
finate). The procedure as presented in Figures 8-3 and 8-4 incorpor-
ates the following unit operations when furfural solvent is used:
0 Continuous countercurrent extraction of the lubricating stock
with furfural at temperatures between 130 and 280°F, depending
on the oil used; suitable heat exchangers are provided.
0 Continuous separation of the raffinate fraction from the extract
fraction.
0 Recovery of solvent (furfural) by vacuum evaporation from
raffinate or refined oil.
0 Stripping by steam distillation of small amounts of remaining
solvent from refined oil, giving wet furfural or water solution
of furfural.
8-6
-------
FIGURES 8-3 & 8-4
FIGURE 8-3
FLOWCHART FOR LUBE-OIL REFINING BY FURFURAL EXTRACTION
Cooler-
Recovered furfural
^ l>.--Wef' SO/venr yerpor
£T~Cono/enser
Vac.oil flash
- tower
Extract
flash tower
Extract (+solvent)-' Heatexch/ pvn:f/ej 0,y.
FIGURE 8-4
TYPICAL SOLVENT-REFINING PROCESS EMPLOYING FURFURAL AND
PROVIDED WITH SOLVENT-WATER SEPARATION AND RECOVERY
Solvent Refining Process ?-f< Solvent-Water -3
Separation
Key'- A-Raffinate and solvent; B= Solvent extract and water; C=Solvent saturated withwate-
D" Water saturated with solvent
Reference 1
8-7
ENGINEERING-SCIENCE
-------
0 Recovery of solvent (furfural) from extract by atmospheric and
by pressure distillation, this wet solvent (furfural) being the
main recovery; fractionation leaves dry solvent behind and
ready for re-use.
0 Stripping with steam of small amounts of solvent left in the
extract, furnishing wet solvent or water solution of solvent.
0 Final stripping of solvent from combined aqueous solutions;
overhead is chilled, and solvent is conducted to fractionator.
° When furfural is used, solvent is recirculated through the
system as many as 15 times each day and with a very small loss,
less than 0.03 percent of solvent recirculated.
The calculated solvent losses of less than 0.03 percent of
recirculated solvent appear low from an industrial standpoint; however,
as is often the case, the larger the operation the greater the solvent
emissions to the atmosphere, which if left unchecked, could become
sizable. The possible loss locations are the following^':
Storage Tanks Pump and Compressor Seals
Wastewater Treatment Pressure Relief Devices
Cooling Towers Drains, Sumps, Hot Wells
Compressor Engines Blind Changing
Stationary Fuel Combustion Sampling
Valves Uncontrolled Slowdown
Flanges and Other Connecting
Devices
Closely related to this area of the petroleum industry is the
re-refining of various oils used in a number of applications. This
process is reputed to be a potential source of considerable volatile
organic emissions, however, a contact at Golden Eagle Oil (Los Angeles,
California) stated that out of the 50 oil refineries in California,
only 3 of them produce lube oil. As for the re-refining of it, it is
very nonvolatile by nature, thus not a likely source of VOCs. The
re-refining process would involve redistilling and filtering used
8-8
-------
lube oil, in which case the emission sources would be much the same as
those listed previously for the production of lube oil.
Both Standard Oil and Amerada Hess Oil companies confirmed that
with hydrotreating to produce lube oil (the newer process), the same
typical VOC emission sources exist as with any other method.
CONTROL TECHNOLOGY
In the event that volatile organic emissions are significant enough
to warrant specific controls on the lube oil production process,
condensers would be a possibility.
According to the EPA report titled "Control Techniques for Volatile
Organic Emissions from Stationary Sources"^', condensers have long been
used as a successful control method (often with additional control
equipment) in abating organic emissions from petroleum refining and
petrochemical manufacturing. Even when used as the primary control
equipment, condensers are usually followed by a secondary air pollution
control system (such as an afterburner) which treats the non-condensible
gases and achieves a high degree of overall efficiency^.
STATUS OF REGULATORY ACTIVITIES
Presently there are no rules or regulations on the state or local
level regarding emissions from the lube oil production process speci-
fically. The existing regulatory work has centered around volatile
organic emissions from the general petroleum industry and is mostly
concerned with storage, loading and unloading and disposal of volatile
organic compounds.
8-9
-------
REFERENCES
1. Shreve, R. Norris, and Brink, Joseph A., Chemical Process In-
dustries. McGraw-Hill, 1977.
2. Nelson, W. L., Chemical Engineering Series, Petroleum Refinery
Engineering. McGraw-Hill, 1969.
3. Hydrocarbon Processing, Page 128, September 1980.
4. EPA-450/2-78-022, May 1978, Control Techniques for Volatile Organic
Emissions from Stationary Sources.
5. Phone Conversation with Harry Chatfield of the South Coast Air
Quality Management District, Dec. 1980.
6. Phone Conversation with Chemical Engineer at Amerada Hess Oil
Company, Purvis, Mississippi, Jan. 1981.
7. Phone Conversation with Engineer at Standard Oil, San Francisco,
California, Jan. 1981.
8. Phone Conversation with Engineer at Golden Eagle Oil, Los Angeles,
California, Jan. 1981.
8-10
-------
SECTION 9
OIL AND GAS PRODUCTION STORAGE TANKS
PROCESSES AND EMISSIONS
This source category covers tanks used in oil and gas production
fields which produce at least some liquid petroleum. These tanks are
utilized in production operations prior to transfer of the field output
from the producer to a transporter or refiner. Further, for purposes
of this section, we are considering onshore production only.
Onshore petroleum production covers the drilling of wells, the
recovery from wells, and the field operations conducted prior to
transport of oil and/or gas from the field to refineries or other
customers. Wells predominantly producing crude oil are called oil
wells, and those primarily producing gas are termed gas wells. However,
most oil wells produce some gas, and almost all gas wells produce some
liquid hydrocarbons through temperature and pressure changes. This
liquid is a condensate and is distinct from crude oil.
Installations in petroleum production fields are conveniently
classified as (1) tank batteries, (2) gas sweetening, and (3) natural
gas liquids plants (Figure 9-1). We will be concerned principally with
the tank batteries. The tank battery receives the gas/water, gas/oil,
and oil/water mixtures and separates them into fractions. The gas
sweetening unit removes hydrogen sulfide (if any) from the natural gas
mixture and further processes it, usually involving a sulfur recovery
plant. The natural gas liquids plant is designed to remove hydrocarbons
heavier than ethane. The condensate so removed is held in pressurized
storage tanks.
In a study for the U.S. Environmental Protection Agency,
estimated the total number of tank batteries in the United States to be
84,541. The greatest number is in Texas (30,828), followed by Califor-
nia (18,114). Again, for purposes of the TRW study, each battery was
assumed to include two tanks. The tanks in a model battery (specified
for cost analysis purposes) were assigned a capacity of 75,000 liters
9-1
-------
FIGURE 9-1
\—
oo
Q
t— i
w o-
§ J
t-^ CZ3
W J
o - 95
W =3
S 5
M Z
O
O
p
w
S
z: ui "•
UJ C_> o
0 I— O P
1— t 1 I 1 Q£ W'
H Ul Ou S
y .so
P oo "
n
0
>-
LU
r—
33
i
I
UJ
5
t—
UJ
w
2
mm
«<
u
f*
^
«•
e
<.
°~
vt :
fl
ui
Ll
j_
tj
U.
^.
3.
41
"5
o f
"
r
j
E CT
1 C
A «
Dt
C
ftl
ai
•3
*— '
3
^
3
J
i
a.
Q
4
|
(-»
]
Compressor
t-^
Lomp lenity
c2cj(
V*
H*
U
tj
o
-o
v a.
— O
_J
c
c
0
*
o.
CO
1
n
tO t>
u-i a
0 cr
c
0) C
o c_
c
to •-
1 — 4J n
•Z. CO 4-
fl}
= | 2 > a
i* CU t
g C
U CU 4-
uj ^>
• 3 Ji
IbO c
^c
^ o 3 *- - ^p-
» 'N O. wi O O t
• •» £ v */i •-;
c
t— t_>
u, /Y"""~~\
J— / It) L. 1 U O
1 UJ 1/1 1 I I W
1 "^ S A
^tj 1 3 z ""
2"~ ' 41 Sl4i >, «- 3
U 1 | 1 z S S 2S ' r1
------>o<^ o ^ ^^ /
«•— z «*^ \
«j — a a* ^- c \
5° . ° || \l
" 2$
3 «
c e
v> e o.
iss
"1 1 5 ssi^.""'^
b -sl
1 f "" 1
U « 1
«
o
1
o
of Messrs. W. |
4 00
1 C
: 4-1
q co
•H
^ co
4 C
3 O
a
^ nl
nmittef
3 0
4 U
i C
0 O
J -U
>1
&$
* 0
, a* —
r'S
'li
^
4t
I.
S
. Griffith (Getty),
H
'U
c
CO
CO
•H
r-1
•H
3
O
^
o"
2
V*
/—
(X
"
and G. Holliday
^
o
PS
£
i-t
iH
Pn
CO /—
S-l iH
O i-H
a
cfl v_
x-
*
S
w
«
C
/
-v
!x -1
T ->
T
I
, -*
1 W
>
9-2
ENGINEERING-SCIENCE
-------
(about 470 bbls.). Engineering-Science, in a study of fixed-roof tanks,
made tests on oil field tanks ranging from 27,000 liters (170 bbls.) to
16,000,000 liters (100,000 bbls.)2).' Many tanks were in the 1,000
bbls. to 2,000 bbls. range, however. Tanks of bolted, riveted, and
welded construction are used. Most are fixed cone-roof in design.
Very few floating-roof tanks are used in oil fields.
A variety of processing techniques are used to carry out the phase
separations conducted in tank batteries. Major items of equipment and
their function are listed in Table 9-1. A typical sequence of tank
battery units is shown in Figure 9-2. Operations and flow patterns in
tanks located in production fields cover a broad range of conditions.
Further, crude oil fed to tanks in tank batteries may at times be
unstabilized (still contains gas) or near flashing temperatures. In
such cases crude oil under slight pressure flashes when introduced into
the tanks. In addition to the functional classification of tanks
such as wash, knock-out, and storage, they can be classified as to mode
of operation. Using this scheme tanks can be classified as batch,
continuous or special service. Batch tanks can be filled or emptied
but not at the same time. Continuous tanks can be filled and emptied
simultaneously. Special service tanks include those where flashing
might occur. Figure 9-3 illustrates various tank operations.
Tanks in continuous service are deserving of additional attention
because the mode of operation complicates the use of equations used to
estimate working loss emissions. This is because the level of the
liquid in the tank does not vary as much as a batch tank having the
same throughput. Two classes are briefly described. They are Lease
Automatic Custody Transfer (LACT) tanks and wash tanks.
The LACT tanks serve a specific function in the oil field. Control
of the crude is automatically transferred from the producer to the
transporter (or pipeline) as it leaves this tank. The liquid level in
these tanks varies between two set points. As liquid enters, the liquid
level slowly rises to the upper set point. The high-volume LACT
withdrawal pump is actuated and the liquid level falls to the lower set
point. The withdrawal pump is turned off, and the procedure starts
again.
9-3
-------
TABLE 9-1
MAJOR TANK BATTERY PROCESSING EQUIPMENT
Equipment
Primary Function
Two-phase separators
Three-phase separators
Gun barrels
Heater treaters
Scrubbers or drips
Free water knockouts
Test separators
Oil, water and condensate
storage tanks
ACT or LACT units
Oil/water and gas separation.
Oil and water and gas separation.
Oil and water separation.
Oil and water emulsion separation.
Remove liquids formed in gas lines.
Water removal.
Isolates wells to determine gas, oil, and
water production rates for individual
wells.
Self explanatory. Condensate is heavier
fractions (C5+) removed from gas phase
following liquid/gas separation.
Automatic or lease automatic custody
transfer units used to test and quantify
crude oil and/or condensate for sales and
lease requirement purposes.
9-4
-------
FIGURE 9-2
CO
LLJ
CO
CO
CO
o
cr
LU
CO
CD
9-5
ENGINEERING-SCIENCE
-------
FIGURE 9-3
OIL
OIL AND
UNSTABILIZED
OIL UNDER ^
SLIGHT PRESSURE
EXAMPLE OPERATIONS OF FIXED-ROOF TANKS
VAPOR
OIL
H
-e
r VAPOR
f i
•- • T . T •
. • ". ' OIL.
f .-
T . .
OIL
••.
Y
UPi
• . . T
APOR ^^
I DOWN
• •'•• ' .'.
OIL' .' :'/.
10
VAPOR
;-.•;• OIL •;-: •: ;
KATER
3_
-OIL
»ATER
ED
-^
1
• ."'
= ^
VAPOR
- .* .
d
• •
•.'.
' ,'.
*
s
t
OIL
1
T "• .
^*tl *
IS-OIL
STANDING STORAGE
SUBJECT TO BREATHING LOSS.
BATCH OPERATIONS
SUBJECT TO CYCLIC FILLING,
EMPTYING, AND BREATHING
LOSSES.
CONTINUOUS OPERATIONS
SUBJECT TO FILLING, EMPTYING
AND BREATHING LOSSES.
CONTINUOUS OPERATIONS (HASH TANKS)
SUBJECT TO BREATHING LOSS.
MOTE: IF TANK FEED CONTAINS UNSTABILIZEO
CRUDE UNDER SLIGHT PRESSURE, THE
TANK HILL BE SUBJECT TO FLASHING
LOSSES ALSO.
SPECIAL OPERATIONS
SUBJECT TO FLASHING, FILUNG, EMPTYING
ANO BREATHING LOSSES.
9-6
ENGINEERING-SCIENCE
-------
Wash tanks are a second particular type of continuous service
tank. They separate oil and water in the incoming crude. Oil normally
exits the tank in an overflow pipe; therefore, the tanks have a constant
liquid level. Frequently, crude comes directly from the various well-
heads to the wash tank. The crude mixture contains oil, water, sand,
and gases such as nitrogen, and carbon dioxide. The mixture arrives
at the wash tank sometimes under pressure (i.e., 30 to 50 psi) and
often at elevated temperatures (i.e., 150 to 200°F). A vertical
cylinder called a "boot" is sometimes attached to the side of wash
tanks. The "boot" relieves any pressure from the crude oil so that
flashing of the trapped gases can occur upstream of the wash tank.
These gases are piped to the vapor space of the wash tank. From the
boot, the oil/water mixture flows into the bottom of the wash tank.
The oil floats to the top and overflows into other wash or LACT tanks
and the water is drained from the bottom of the tank and discarded
into a sump.
EMISSIONS
Emissions from fixed-roof tanks used for crude oil have tradition-
ally been estimated using equations found in API Bulletin 2518-^) or
modification of these equations in AP-42^). These calculations cover
breathing losses and working losses. Breathing losses are those
due to vapor expelled during the diurnal expansion and contraction of
the air-hydrocarbon mixture above the liquid surface in a fixed-roof
tank. The equation for these losses from crude oil given in API 2518
is :
\ " Kc ( 24 ) ( p ) *68 D1.73H.51T.50F c
1000 14.7-P P
Where:
Ly = Breathing loss (bbls./yr.)
Kc = Factor to adjust gasoline breathing loss equation
to breathing loss of crude oil = 0.58
P = True vapor pressure at bulk liquid temperature (psia)
D = Tank diameter (ft.)
H = Average outage including correction for roof volume (ft.)
9-7
-------
T = Average daily ambient temperature change (°F)
Fp = Paint factor
C = Adjustment factor for small diameter tanks
The adjustment factor for small diameter tanks applies to those
with diameters less than 30 ft. A curve giving these adjustment factors
is to be found in API 2518.
Working losses are those associated with the displacement of vapors
by liquid during tank filling. The API equation for crude oil working
losses is:
F - 2.25 PV KT
10,000
Where:
F = Working loss (bbls.)
P = True vapor pressure at bulk liquid temperature (psia)
V = Volume of liquid pumped into tank (bbls.)
KX = Turnover factor (from equation (5) or Figure 11 in API 2518)
The equations given above for breathing and working losses can obviously
be adjusted to other reporting terms, e.g. Mg/yr., Ib./day, Ib./lOOO
gals, throughput (working loss), using liquid densities, etc.
Recent studies have indicated that breathing losses from fixed-
roof tanks are overestimated by a factor of four using the API 2518
equation^}5,6)> Consequently, TRW, in its source category survey report
on onshore production-^-), discounted breathing losses by a factor of
four in estimating per tank losses for the model tank battery. There is
also an indication that working losses can be overestimated, particu-
larly in the case of continuous service2). In the case of boiling
or flashing crudes, the working losses can be underestimated. In
summary, the predictive equations for breathing losses (as adjusted)
9-8
-------
and working losses are probably adequate for overall emission inven-
tory preparation, but are subject to considerable error when used for
individual tanks. Additionally, these predictive equations were not
intended for use for onshore production tanks.
CONTROL TECHNOLOGY
Control technology for petroleum storage tanks at refineries,
pipeline and marine terminals, and bulk plants is well known.
As compared to uncontrolled fixed-roof tanks, emission reductions can
be obtained with external floating-roof tanks, internal floating-roof
tanks, and various types of vapor recovery systems. In general,
regulations do not require controls for liquid petroleum storage where
the true vapor pressure does not exceed 1.5 psig. Stabilized crudes
have vapor pressures ranging from well under to well over this figure.
There are a number of differences between oil and gas production
tanks at those refineries and terminals which impact the applicability
of control technology. Some of these have been mentioned previously
but will be summarized below:
0 Oil and gas field tanks are generally much smaller than those
at refineries and terminals.
0 Tank batteries may be in remote areas without commercial
electric power.
0 Some oil fields do not have access to gas distribution systems
to handle small amounts of gas produced with the oil.
0 Gases evolved from unstabilized crude may contain C02 and H^S.
0 Tanks may be few in number and isolated by distance in production
fields.
0 Tanks in continuous service in production fields may have
relatively small changes in liquid level associated with
throughput.
0 Production tanks may contain unstabilized or flashing crudes.
9-9
-------
Taking into account the special circumstances of production tanks,
two major options exist for VOC emission control: (1) vapor recovery
and/or control, and (2) floating-roofs.
Emissions from fixed-roof tanks can be reduced by collecting and
either recovering or combusting them. In a typical vapor recovery
system, vapors remain in the tank vapor space until a pressure is
reached which actuates the recovery system. The vapors are then
collected by blowers or compressors and moved to the ultimate disposal
system. Where natural gas is recovered, the tank vapor mixture can be
transferred to the gas sweetening system (if any) or the natural gas
liquids plant. An alternate means of disposal is a flare or fume
incinerator (thermal oxidizer system). Such a system should have means
of ignition, controlling combustion air, and preventing flashback to
the tank vapor space. Where an open flare is used, it must be located
in accordance with usual flare safety practices. Vapor recovery systems
can be operated at efficiencies above 95 percent.
Floating-roof tanks reduce VOC emissions as compared to uncon-
trolled fixed-roof tanks by virtue of the fact that the floating-roof
rises and falls with the liquid level during filling and withdrawal of
the stored liquid and by reduction of the vapor space volume. Per-
formance of floating-roofs is influenced by the effectiveness of seal
design in minimizing gaps between the seal and the tank shell which
result in direct exposure of liquid surface to the air.
Tanks with external floating-roofs have no fixed covers. A great
variety of roof and seal designs have been used. All must provide for
roof stability, precipitation drainage, gaging and inspection, roof
support when the tank is empty, and effective perimeter sealing. Double
seals have been shown to be very effective in reducing emissions as
compared to single seals. In general, external floating-roof tanks are
relatively large and field-erected and may be unavailable in the size
range of most existing production storage tanks.
An internal floating-roof tank is essentially a fixed-roof tank
with an internal cover floating on the liquid surface (contact cover)
or suspended several inches above the surface with a perimeter pontoon
9-10
-------
(non-contact cover). Both aluminum and steel pans and aluminum sandwich
(honeycomb) type covers are in use. Tanks can be retrofitted with
internal covers. Internal covers offer several advantages over external
covers for emissions reduction. They are not as susceptible to wind
influenced losses and provide additional protection against solar
radiant heating of the liquid surface. Additionally, they eliminate
the need for precipitation drainage. Regular inspection to check for
roof condition such as tilting or sinking is necessary. The space
between the floating cover and fixed-roof must be ventilated to avoid
build-up of VOC emissions to the lower flammable limit.
Emissions from floating-roof tanks, whether external or internal,
are classified as "standing" and "wetting" (withdrawal). Standing
losses result from gaps between the seal and the tank shell. As a
result, some liquid surface is exposed to the atmosphere. When wind
creates pressure differences above the floating-roof, air flows into
the annular vapor space at locations of high pressure and an air-vapor
mixture flows out at low pressure points. Vapors can also escape from
open hatches or other openings, glands, valves, and fittings.
Wetting is another source of emissions from floating-roof tanks.
Wetting loss is the vaporization of liquid from a wetted tank wall
when a floating-roof is lowered by withdrawal of liquid. TRW has
estimated the annual emissions from various controlled and uncontrolled
tanksl). Table 9-2 shows these losses using a tank of 75,000 liter
capacity in their model tank battery. These figures may be used for
comparison purposes. Costs of these controls from the same study are
given in Table 9-3.
Based upon the figures in Tables 9-2 and 9-3, the vapor recovery
system is the most cost effective (lowest cost/amount controlled). It
should be kept in mind that these are capital, not annualized costs.
No credit for hydrocarbon recovery can be assigned in this case.
9-11
-------
TABLE 9-2
CRUDE OIL STORAGE TANK UNCONTROLLED AND CONTROLLED VOC EMISSIONS
Device Emission (Mg/yr.)
Uncontrolled fixed-roof tank 1.25
Non-contact internal floating-roof 0.8
External floating-roof tank 0.63
(primary seals only)
External floating-roof tank 0.19
(secondary seals)
Contact internal floating-roof tank 0.13
Vapor control system 0.05
TABLE 9-3
CRUDE OIL STORAGE TANK CONTROL COSTS
Control Technique
Vapor control
Non-contact internal floating-roof
Contact internal floating-roof
Cost per tanka
NAC
5,300
6,300
Cost per facility0
$13,000
10,600
12,600
a) Based on 79,500 liter welded storage tank; 1979 dollars.
b) Based on two tanks per facility; 1979 dollars.
c) NA = Not Applicable.
9-12
-------
STATUS OF REGULATORY ACTIVITIES
The U.S. EPA has not regulated oil and gas production storage
tanks as of the present time. The agency is conducting studies to
determine whether air pollution sources within the category — onshore
production of crude oil and gas — should be regulated with standards
of performance promulgated under the authority of the Clean Air Act as
amended August 7, 1977. Emissions of VOC from tank batteries are being
considered under this study. Emissions from petroleum storage vessels
over 40,000 gallons in capacity are regualated under 40 CFR, Part 60,
Subparts K and Ka, but the regulation does not apply to storage vessels
for petroleum or condensate at drilling or production facilities prior
to custody transfer.
In general, state regulations are similar to Subparts K and Ka
in that crude oil storage upstream of custody transfer from production
facilities is exempted from the regulations. The California Air
Resources Board is considering a model rule for onshore production of
oil and gas but technical studies related to such a rule are still
under way.
The Ventura County Air Pollution Control District in California
has adopted applicable regulations under their Rule 71.1, Crude Oil
Production and Separation. This rule, which is attached, requires that
crude oil storage tanks in the tank battery, including wash tanks, be
equipped with a closed-type vapor recovery system or equivalent.
REFERENCES
1. "Source Category Survey Report Onshore Production," prepared by
TRW, Inc., Environmental Engineering Division for U.S. EPA, Economic
Analysis Branch, RTP, NC, March 19, 1980.
2. "Hydrocarbon Emissions from Fixed-Roof Petroleum Tanks," prepared
by Engineering-Science, Inc. for Western Oil and Gas Association,
July 1977.
3. API Bulletin 2518, "Evaporation Loss fron Fixed-Roof Tanks,"
Washington, D.C., 1962.
9-13
-------
4. "Compilation of Air Pollution Emission Factors," U.S. EPA, Report
No. AP-42, RTP, NC, August 1977.
5. "Emission Test Report - Breathing Loss Emissions from Fixed-Roof
Petrochemical Storage Tanks," U.S. EPA, EMB Report No. 78-OCM-5,
RPT, NC, February 1979, p. 173.
6. "Measurement and Determination of Hydrocarbon Emissions in the
Course of Storage and Transfer in Above-Ground Fixed Cover Tanks
With and Without Floating Covers," German Society for Petroleum
Science and Carbon Chemistry (DGMK) and the Federal Ministry of
the Interior (BMI), BMI-DGMK Joint Projects 4590-10 and 4590-11,
translated for EPA by Literature Research Company, Annandale,
Virginia.
9-14
-------
APPENDIX
Ventura County Air Pollution Control District
State of California
Rule 71. Crude Oil and Organic Liquids
(Adopted June 20, 1978; Revised March 17, 1979, July 10, 1979
-------
Rule 71. Crude Oil and Organic Liquids (Adopted S/20/78; Revised
3/17/79, 7/10/79)
A. Applicability
I. Tha provisions of this rule shall apply to the production,
gathering, separation and processing of crude oii and
natural gas and the storage and transfer of crude ail and
organic liquids excluding gasoline.
B. Definitions
1. For the purpose of this rule, a "clasad-type vapor recovery
system" means any organic vapor control system which is
designed not to release or vent any organic gases to the
atmosphere under normal operating conditions.
2. For the purpose of this rule, "vapor loss control efficiency"
means a comparison of controlled emissions to those
emissions which would occur from a fixed or cone roof tank
in the same product service without a vapor control system.
Base line emissions shall be calculated by using the criteria
outlined in American Petroleum Institute Bulletin 2SI8.
3. For the purpose of this rule, "petroleum production permit
unit" means any aggregation of equipment used exclusively
for the production, gathering and separation of crude oil
and natural gas which is included on a single
Permit-to-Operate issued by the Air Pollution Control
Officer.
4. For the purpose of this rule, "custody transfer" means the
transfer of produced oil and/or condensate, after separation
and/or treating in producing operations, from storage tanks
or automatic transfer facilities to pipelines or any other
forms of transportation.
5. For the purpose of this rule, "tank battery" means any
tank, or any aggregation of tanks, used for the purpose of
storing or holding crude oil or for the purpose of
separating crude oil, water and/or natural gas. An
aggregation of tanks will be considered a tank battery only
if the tanks are located so that no one tank is more than
ISO feet from all other tanks.
6. . For the purpose of this rule, "wash tank" means any tank
used for the purpose of the primary separation of crude oil
and "produced water.
7. For the purpose of this rule, "gauging tank" means any
tank -used exclusively for the purpose of measuring the
production rate of crude oil from any petroleum production
well.
9-15
-------
C. Severabilitv
If any portion of this rule shall be found to be
unenforceable, such finding shall have no effect on the
enforcaability of the remaining portions of the rule, which
shall continue to be in full force and effect.
Rule 71.1 Crude Oil Production and Separation
A. Aoolicabilitv
I. The provisions of this section of this rule shall apply to
equipment used in conjunction with the producing,
gathering and separation of crude oil and natural gas from
any petroleum production- permit unit prior to custody
transfer.
B. Storage Tanks
I. No person shall place, hold or store any crude oil in any
tank battery/ unless all crude oil storage tanks in the tank
battery, including wash tanks, are equipped with a
closed-type vapor recovery system, properly installed,
maintained and operated; or any other control technology
considered by the Air Pollution Control Officer to represent
the best available air pollution control technology at the
time of installation. Any tank gauging or sampling device
on a tank vented to a vapor control system shall be
equipped with a gas-tight cover which shall be closed at all
times except during gauging and sampling.
2. The provisions of subsection 3.1 of this Rule shall not
apply to any of the following:
a. Any tank battery, including wash tanks, installed
prior to June 20, 1S73 for the purpose of holding and
staring crude ail, having a true vapor pressure of less
than 1. 5 psi absolute;
b. Any tank battery, including wash tanks, when it has
been demonstrated to the satisfaction of the Air
Pollution Control Officer that the cost of installing air
pollution control technology is economically prohibitive
and substantially exceeds the average cost per unit
mass of air contaminant for all other stationary saurca
ccntrois. In making this determination:
(i) Consideration shall be given to alternatives other
than continued use of a tank battary or construc-
tion of a new tank battary including transfer by
pipeline or other acceptable methods to an exist-
ing tank battery with appropriate air pollution
control equipment; and
9-16
-------
(ii) The cast af controlling a tank battery shall be
averaged with the total cost of air pollution
control for the entire stationary source;
c. Any crude oil production gauging tanks having a
nominal capacity of 500 barrels or less;
d. Any tank battery, including wash tanks, holding or
staring crude oil from any new crude oil production
well, for a period of ninety days following initiaj
production from that well.
3. Any person wishing to come under the provisions of
subsection B.2.a. of this rule shall keep records to
substantiate the applicability of that subsection. Such
records shall include, for any crude oil, the true vapor
pressure under actual storage conditions and the monthly
throughput of the subject tank battery. These records
shall be made available to the Air Pollution Control Officer
on request. The true vapor pressure in psi absolute of
stored liquid may be determined by using the nomographs
contained in American Petroleum Institute Bulletin 2513 or
by any other method approved by the Air Pollution Control
Officer for determination of true vapor pressure.
C. Fugitive Emissions (Revised 11/20/79)
I. Any produced gas shall be routed to/a vapor recovery
system, flared, or controlled in some other manner
considered by the Air Pollution Control Officer to represent
best available control technology. The provisions of this
subsection shall not apply in the following situations:
a. During routine maintenance of any well, or
b. During the first two weeks of production of an
exploratory well, if the composition of the gas being
produced is unknown (i.e., new reservoir) and there
are no existing gss handling systems in the general
vicinity.
2. All equipment used in conjunction with the production,
gathering and separation of crude oil and natural gas from
any petroleum production lease shall be routinely maintained
in a manner representative of good oil industry practices so
as to minimize air contaminant emissions.
0. Effective Dates
I. Any person owning or operating any existing equipment
which requires modification to comply with the provisions of
subsection 3 of this rule shall comply with the following
schedule of increments of progress:
9-17
-------
Tank Sattary
Throughput Increment Increment Increment Final
SSL/Month a b c Ccmolianc;
5001 4 Above S-l-73 12-1-73 . 1-1-79 S-I-7S
400I-50CO 6H-7S 9-1-79 10-1-79 S-l-80
3001-4000 6-1-30 9-1-50 10-1-30 5-1-31
2001-3000 8-1-81 9-1-31 10-1-31 3-1-32
Revises 3/ES/79
9-18
-------
SECTION 10
PETROLEUM COKING PROCESSES
PROCESSES AND EMISSIONS
Low value residual type oils are converted to higher value
products through a thermal cracking process called coking. Thermal
tars and vacuum residuals are cracked at high temperature and atmos-
pheric pressure. The resulting products are petroleum coke, gas oils
and lighter petroleum stocks. Presently, delayed coking is the most
widely used process-'-).
DELAYED COKING PROCESS AND EMISSIONS
The heated charge stock is fed into the bottom of the fractionator,
here the light ends are stripped off. The recycle from the coke drum is
combined with the remaining feed and is quickly heated in the coking
heater to 480-590°C (900-1100°F). The heating rate is controlled by
steam injection. In the coke drum, the vapor-liquid from the heater
is converted to coke under the proper conditions of residence time,
pressure, and temperature. The thermal cracking products are recovered
in the fractionator once the vapors from the top of the drum have re-
turned. Once the onstream coke drum is filled to the proper capacity
with coke, it is taken offstream and is quenched/purged with steam.
When the temperature reaches the desired level, the drum is opened
and the cotce is cut with high pressure water.
PROCESS EMISSIONS AND EMISSION POINTS
Considerable quantities of steam and hydrocarbons may be released
to the atmosphere when the coke drum is opened. The cutting operation
may also emit hydrocarbons when more steam is produced by vaporization
of the cutting water. Also, the cutting itself may release pockets of
trapped steam and hydrocarbons. Some of the hydrocarbons released can
include polynuclear aromatics and other hazardous compounds since the
coker conditions favor their production. See the following Figure
10-1 for flow chart of delayed coking process.
10-1
-------
FIGURE 10-1
00
c/>
LU
O
O
Q±
Q_
CD
O
CJ
LU
CD
UJ
o
10-2
ENGINEERING-SCIENCE
-------
CONTROL TECHNOLOGY
Venting the quenching stream to a vapor recovery or blowdown system
would lessen the hydrocarbon emissions from coking processes. When the
drum cools to 100°C (212°F), the steam purge can be replaced by a water
flood. Additional cooling to nearly ambient temperature will reduce
steam and hydrocarbon vaporization and subsequent escape when the drum is
opened-*-' .
STATUS OF REGULATORY ACTIVITIES
The state and local agency regulations concerning the petroleum
industry do not include specific rules addressing coking. In general,
the regulations are stated in a blanket fashion, and worded as "those
processes involved in the production of gasoline, kerosene, distillate
and residual fuels, or other products through distillation, cracking,
extraction, or reforming of petroleum derivatives, must control the emis-
sions from vacuum producing systems, wastewater separators, and process
unit turnarounds using the methods specified.... ". The rules also specify
the storage, loading and unloading of petroleum products and the required
controls and precautions.
The State of the Illinois air laws regarding organic discharge
to the atmosphere do indicate the following: " any petroleum fluid
coker; or any other waste gas stream from any petroleum or petrochemical
manufacturing process; in excess of 100 ppm equivalent methane (molecular
weight of 16.0).
REFERENCES
1. Control Techniques for Volatile Organic Emissions from Stationary
Sources, EPA-450/2-78-022, May 1978.
2. Nelson, W. L., Petroleum Refinery Engineering. McGraw-Hill, Fourth
Edition, 1969.
3. Hydrocarbon Processing, Page 154, September 1980.
10-3
-------
SECTION 11
PLASTIC PARTS PAINTING
PROCESSES AND EMISSIONS
A wide variety of industries produce plastic products which are
painted. There are other methods of imparting color to a plastic
product, such as use of pigmented molding compounds and use of colored
gel coats in fiberglass reinforced resin products, but these techniques
are not covered here. Both complete assemblies and component parts are
included in those items painted. In the latter case, painting is often
utilized to obtain close color matching with other parts. Some of the
plastic parts categories where painting is used are instrument knobs,
radio grilles, automobile dashboard parts and body filler panels,
sporting goods, cosmetic cases, control panels and nameplates for
appliances, furniture parts, toys, lamp housings, canisters, clock
bezels, and office equipment parts.
No comprehensive emission information for plastic parts painting
by source category was readily available for use in this report, and
no direct survey was attempted. Indirect information was obtained
indicating that VOC emissions from plastic parts painting at some
automotive parts suppliers were in the range of 1000 tons per year.
Both primers and topcoats are applied using solvent-based coatings
containing solvents such as xylene, naphtha, diisobutylketone, and
methylethylketone. Solvent content of coatings ranged from 62 to 79
percent by weight. Some automated electrostatic coating lines were
utilized. An approximate paint transfer efficiency of 55 percent was
estimated for electrostatic painting.
Most plastic parts are relatively small and are painted with spray
techniques. These include compressed air, airless, and electrostatic
methods of application. Coatings used are somewhat more limited as
compared to metal products because the organic substrate limits coating
compatibility and the temperature of curing. Nevertheless, the solvent
content of paints used for plastic parts is not unlike that of coatings
11-1
-------
used on miscellaneous metal products. In the Control Techniques
Guidelines (CTG) document for this latter category, enamels (at about
30 to 40 volume percent solids) and lacquers (at about 10 to 20 percent
solids) were listed as being conmonly used.D
CONTROL TECHNOLOGY
At the present time the most applicable process related control
techniques for reducing VOC emissions from plastic parts painting are:
1) Use of higher solids coatings;
2) Use of radiation-cured coatings; and
3) Use of higher transfer efficiency coating methods such as
electrostatic and airless spraying.
In the case of higher solids content coatings, there is a general
development in this direction in the entire coatings field. To the
extent that such coatings are compatible with plastic substrates, they
could be adapted for use with plastic parts.
Kut in Chapman and Anderson^) discusses the use of radiation-cured
coatings for plastic automobile radiator grilles and instrument panels.
Advantages claimed include rapid curing at low temperatures, minimal
pollution because of the use of "solvent free" coatings, high throughput
rates, and less floor space requirements as compared to thermal
polymerization techniques. The terra "solvent free" coatings refers to
those in which liquid hydrocarbons such as styrene impart fluidity
during painting, but which are polymerized during the radiation cure to
become part of the coatings. Some losses of these compounds do occur
prior to curing. A large capital expenditure is required for radiation
cure equipment.
Techniques to obtain higher transfer efficiencies are potentially
capable of achieving significant reductions in VOC emissions. Both
electrostatic and airless (hydraulic) spraying techniques can result in
higher transfer efficiencies (ratio of coating reaching object to be
11-2
-------
coated to that consumed). In the case of spray application methods,
the excess coating is lost in overspray and bounceback.
The conventional method for applying surface coatings to many
plastic parts is with an air atomizing spray gun. Transfer efficiency
with these systems is generally less than 50 percent.
Electrostatic coating is based on the principle that opposite
electrical charges attract each other. The atomized coating particles
are negatively charged by the application device and the article to be
coated is positively charged or grounded. Transfer efficiencies with
hand-held equipment can typically be improved to 65 percent, and under
optimal conditions, using automated bell and disc systems with conveyor-
ized parts, efficiencies can reach 85 to 95 percent.3) Because
plastic is non-conductive, it must be treated with a sensitizing
material. Plastic products which have been successfully coated using
electrostatic techniques include containers, golf balls, grill work,
automotive grilles and instrument panels, picture frames, women's shoe
heels, and toilet seats.
Airless spraying utilizes hydraulic pressure in lieu of air pres-
sure to atomize the coating. While coating particles are not attracted
to the target such as in electrostatic spraying, there is a reduction
bounceback and in coating material which follows airstream lines instead
of impacting the target.
Add-on controls such as incineration or activated carbon adsorption
are equally as applicable to plastic parts painting as to many other
surface coating categories. A discussion of these techniques, which for
the most part is applicable to plastic parts painting, can be found in
the CTG for Miscellaneous Metal Parts and Products^) and in a general
EPA Guideline for Control Methods for Surface Coating Operations^'.
STATUS OF REGULATORY ACTIVITIES
Many states and local agencies have general rules which limit
emissions of volatile organic compounds from industrial processes and
surface coating operations. These rules are generally expressed in
11-3
-------
terms of mass limits per unit of time. Rule 442 of the South Coast
Air Quality Management District, California; and Rule 2, "Miscel-
laneous Operations" and Rule 4, "General Solvent and Surface Coating
Operations", Bay Area Air Quality Management District, California are
examples of such rules. No rules specifically directed towards
painting of plastic parts were found to exist.
REFERENCES
1. "Control of Volatile Organic Emissions from Existing Stationary
Sources - Volume VI: Surface Coating of Miscellaneous Metal Parts
and Products," EPA-450/2-78-015, U.S. EPA, OAQPS, RTP, NC 27711,
June 1978.
2. B. N. Chapman and J. C. Anderson, Science and Technology of Surface
Coatings, Academic Press, New York and London, 1974.
3. "Electrostatic — A Dark Horse in Finishing Alternatives," Wood and
Wood Products, February 1980, p. 25.
4. "Control of Volatile Organic Emissions from Existing Stationary
Sources - Volume I: Control Methods for Surface Coating Operations,'
EPA-450/2-76-028, U.S. EPA, OAQPS, November 1976.
11-4
-------
SECTION 12
RAILROAD TANK CAR LOADING OF
VOLATILE ORGANIC LIQUIDS
This section reviews the loading of volatile organic liquids (VOL's)
into railroad tank cars. Emissions generated in these loading operations
and suitable techniques for the control of these emissions are discussed
and cost effectiveness calculations for a typical VOL are presented.
Existing air pollution regulations covering railroad tank car loading of
VOL's are reviewed.
PROCESSES AND EMISSIONS
The loading of volatile organic liquids into railroad tank cars -takes
place nationwide at railroad tank car loading terminals serving chemical
plants and petroleum refineries, and at pipeline, marine and independent
(toll) terminals.
There were approximately 200,000 railroad tank cars in service in
1979, about 98 1/2% of which were owned by shippers and private railcar
companies such as General American Transportation Corp. (GATX), North
American Car Corporation, etc. Each of these cars was, on the average,
loaded and shipped about 10 times during 19791).
The 1979 aggregate volume of diverse (organic and inorganic) chemical
and petroleum products shipped by tank car was approximately 45 billion
gallons. Petroleum products account for about 18% of the total, with
volume of tank car shipments tending to decline in recent years. Volume of
tank car shipments of all chemical products represents 82% of the total and
has tended to grow in recent years, increasing about 3.5% from 1978 to
19791).
Only those VOL's falling within a vapor pressure range of 0.5 to 13
psia @ 60°F were considered in assessing the significance of tank car load-
ing of chemical and petroleum products as a source of volatile organic
compound emissions (VOC's). VOL's with vapor pressures below 0.5 psia @
12-1
-------
60°F produce a relatively insignificant quantity of emissions when loaded
into tank cars, even under the worse loading conditions. VOL's with vapor
pressures above 13 psia @ 60°F (which are in most cases liquified gases)
are always loaded under pressure sealed, vapor controlled conditions by
venting the tank car vapor space to the vapor space of the storage vessel
from which the VOL is drawn (termed vapor balance). While this is done
primarily to conserve valuable product, it effectively eliminates VOC
emissions to the atmosphere during the loading operation.
Within the vapor pressure range of 0.5 to 13 psia @ 60°F fall about
4.5 billion gallons of VOL's shipped by railroad tank car, about 10% of
the total annual shipments of chemical and petroleum products during
1978 and 1979.
Table 12-1 summarizes the approximate volumes of the 10 VOL's loaded
and shipped in railroad tank cars in largest quantities in 19782).
The volume of tank car shipments of these ten chemicals accounted in
1978 for 80% of the total volume of all VOL's shipped. Crude oil, gasoline
and associated petroleum products such as naptha and hexane comprise well
over one half of all the VOL's with vapor pressures between 0.5 and 13 psia
shipped by railcar in 1978^). Based on information obtained in a survey of
the railcar loading industry covering 33 terminals,^ the average railcar
loading terminal loads about 15 million gallons of VOL's/year (roughly 800
railcars/year), with range of throughput volume extending from about 1
million up to 50 million gallons/year. The average VOL loaded in these
terminals has a vapor pressure of about 1.7 psia @ 60°F.
Description Of Equipment
Equipment used to load VOL's into railroad tankcars consists of
loading arms and miscellaneous pumps, valves, meters, etc., generally
mounted on a structure termed a loading rack, positioned within the
railroad siding area designated as a railcar loading terminal.
Railroad Tank Cars. Railroad tank cars consist of cylindrical tanks
mounted on appropriate underframes and wheeled trucks (Figure 12-1).
Recently constructed tank cars are of welded or seamless construction,
12-2
-------
TABLE 12-1
QUANTITIES OF TOP TEN VOL's LOADED IN RAILROAD TANK CARS
1978
MM
VOL GALLONS
LOADED
CRUDE OIL 1162
ETHANOL 558
GASOLINE 555
NAPTHA 449
METHANOL 262
ACRYLONITRILE 202
PROPYLENE OXIDE 179
ACETONE 92
CYCLOHEXANE 79
HEXANE 61
12-3
-------
with domed welded heads. Some tanks of older riveted construction are
still in limited use. Materials of construction are primarily mild
steel, stainless, or aluminum. Interiors of tanks may be lined with
materials such as rubber or poly (vinyl chloride) suitable for specific
service with various corrosive chemicals.
Various classes of tanks are insulated to retard heating or cooling
of the tank contents during transit. Tanks may be equipped with interior
steam coils to facilitate unloading by reducing the viscosity of the cargo,
especially during cold weather-*).
Structural requirements for railroad tank cars are generally covered
in the Code of Federal Regulations by Title 49, Part 179, Subpart D. Part
179.202 encompasses regulations specific to particular chemicals. In
addition, 49 CFR 172 presents regulations on classification and identifi-
cation of hazardous materials, and specifies marking and placarding
requirements for tank cars in which they are shipped.
VOL's are generally suitable for shipment in tank cars designated
as general service tank cars. Such tank cars are designed to accomodate
moderate interior pressures not in excess of 100 psig, and are generally
equipped with over pressure relief valves set at 35 to 100 psig. Figure
12-1 shows such a general service tank car, with insulation and interior
steam coils.
Tank cars are classified, tested and certified by U.S. Department
of Transportation in cooperation with the Association of American Rail-
roads (AAR). The general service classification appropriate for the
shipment of VOL's encompasses DOT classes 103, 104 and 111. The full
specification number (e.g., 111A60ALW) designates the class (111), a
test pressure (60 psig), a material of construction (Aluminum), and a
type of construction (welded).
Different sizes of tank cars are available within each classifica-
tion. Common sizes within the general service classification are 8000,
10,000, 20,000 and 23,500 gallons. Also available for specialized
service are various other sizes up to a maximum of 33,500 gallons.
Maximum size is restricted by allowable maximum total weight of tank car
12-4
-------
FIGURE 12-1
cc
«t
o
O
ce
cc
UJ
o
>
or
UJ
CO
UJ
ICO
12-5
ENGINEERING-SCIENCE
-------
and cargo, particularly if the shipment is to qualify for unrestricted
movement among the various railroad lines-*'.
Various loading and unloading arrangements are provided on the types
of tanks within the general service railroad tank car classification. For
loading, all tank cars within the general service classification are
provided with a hatch (manway) of various diameters through which the
loading device (arm) may be inserted. For unloading, all general service
tank cars are equipped with a bottom unloading valve positioned directly
under the tank. This valve may be operated from the top of the tankcar
by a valve extension handle. Many tank cars of modern construction have
been fitted with a top unloading piping arrangement comprising an inlet
for pressurized inert gas and an outlet pipe extending from the top
unloading port approximately to the bottom of the tank. The cargo is
forced by gas pressure from the tank through the outlet pipe, and is
discharged at the top of the tank.
Railroad Tank Car Loading Terminal. A railroad tank car loading
terminal is broadly considered to be the general area within which are
located the railroad siding(s), structures (loading rack), and equipment
necessary to load liquids into tank cars.
Within a tank car loading terminal may be one or more loading racks.
The loading rack is generally a raised platform positioned adjacent to
the railcar siding and includes the loading arm(s), piping, valves,
pumps, meters and other equipment necessary to the loading process. The
working area of the rack is elevated to the approximate level of the
loading hatch of the tank car positioned on the siding. Most loading
racks have a number of tank car loading stations, and, as a result, fre-
quently load a number of tank cars simultaneously. The rack structure
may include a roof covering the tank car stations on the siding. Individ-
ual racks within a terminal are generally designed to accomodate loading
of specific classes of materials. Much of the equipment associated with
the rack is dictated by the properties of the class of materials intended
to be loaded. A terminal may contain one or more of these racks, and
may be as simple as a single rack accommodating a single arm or multiple
arms, or as complex as two or more racks with multiple arms. Figure 12-2
shows a simple terminal comprising a single siding and positions for 2
railcars.
12-6
-------
FIGURE 12-2
ce
UJ
CD
O
«e
O
_l
cc
-------
Loading Am. The loading ana is the extension of the piping which
carries the VOL from intermediate storage to the tank car.
For liquids of relatively low volatility, this extension may consist
of a large diameter flexible hose fitted with an angled pipe which can
be inserted into the open tank car hatch. The angled pipe may vary in
length from a few feet to a length long enough to reach to the bottom of
the railcar tank. In place of the flexible hose, the arm may be construc-
ted completely of pipe and fitted with double swivel joints to permit
positioning of the pipe into the tank car.
A third type of loading aria in general use for the loading of highly
volatile liquids and compressed gases is also adaptable for the loading of
VOL's in general. This loading arm terminates in a union or coupling to
permit connection to existing piping leading through the railcar dome and
extending approximately to the bottom of the tank. In conjunction with
this type of loading arm, a vapor return line is installed to return vapors
vented from the tank car to vapor collection ducting or piping along the
rack. Since loading of VOL's with this type of loading arm takes place
with the hatch closed and sealed, a sensing device such as a magnetically
activated float within the tank is frequently used to monitor liquid
level within the the tank car.
Preparation of Railcars for Loading. Railroad tank cars are oc-
casionally cleaned prior to loading. Most railroad tank cars however,
are owned and. leased by shippers, and are used exclusively and repeatedly
for shipping a single chemical (dedicated service). These cars are only
cleaned when testing or inspection reveals contamination. Tank cars
owned or leased by shippers of petroleum products are considered to be
in non-specific dedicated service. Since the types of products shipped
in these cars are frequently not cross-contaminating to a serious degree,
these cars are used to ship various essentially similar products without
cleaning. These cars are only cleaned when inspection or sampling
reveals contamination, or when the material to be loaded differs sub-
stantially from the material previously shipped.
12-8
-------
Tank cars in non-dedicated service represent a relatively small pro-
portion of the total cars in chemical service. Such cars are routinely
cleaned between loadings unless sampling, testing and inspection can
establish the compatibility of the VOL to be loaded with the residual
liquid from the previous loading.
Tank cars are cleaned at either the shipping and receiving terminals
or at railcar maintenance and repair facilities. The cars are cleaned with
the use of a steam hose, pressure wand, or rotating spray heads placed
through the manhole. The residual chemical and the cleaning solution are
flushed from the tank with water. Although much of the VOL cleaned from
the tank car will eventually be lost to the atmosphere, the quantity of
emissions occurring as the result of cleaning of railroad tank cars is
considered to be negligible.
Open Hatch Loading. Railroad tank cars are almost exclusively loaded
from the top of the tank. The principal methods of loading tank cars are
the splash loading method and the submerged fill pipe method. In the
splash loading method the fill pipe dispensing the liquid is only partially
lowered into the tank car. Significant turbulence and vapor-liquid con-
tacting occurs during splash loading, resulting in high levels of
vapor generation. If turbulence is high enough, entrainment of liquid
droplets in the vapor may take place. In the submerged fill pipe method,
the fill pipe descends almost to the bottom of the tank car. As a result,
the fill pipe opening is below the liquid during the major portion of
loading, resulting in a significant reduction in liquid turbulence and
vapor-liquid contacting.
Both the splash loading and submerged fill loading methods are con-
ducted at rates as low as 50 gallons/minute and as high as 400 gallons/
minute, depending on the viscosity of the VOL being loaded, the available
pressure and the piping sizes in the system. Accordingly, tank car filling
times vary from as long as 8 hours for a 23,500 gallon tank car filled at
the slowest practiced rate to as quickly as 1/2 hour for a 10,000 gallon
tank car filled at the highest practiced rate.
Railroad tank cars are filled to slightly less than their nominal
capacity, as prescribed by U.S. DOT regulations. The remaining unfilled
volume, termed outage, is volume reserved for liquid expansion occurring
as a result of increasing temperature. Maintenance of a vapor space
12-9
-------
within the tank also insures that vapor and not liquid is vented through
the tank car pressure relief valve under transient conditions of excess
temperature and pressure^).
Closed Hatch Loading. Because of the hazardous or highly volatile
nature of some VOL's loaded into railroad tank cars, closed hatch loading
techniques are also practiced. As described earlier, some general service
tank cars have been fitted with a top unloading piping arrangement. This
piping arrangement has been adapted for VOL loading through use of the
unloading pipe for loading and the use of the pressurization port for
collection of VOC emissions.
Since the number of general service tank cars fitted out with this top
unloading piping arrangement is limited, railroad tank cars designed for
liquified gas service are sometimes used for loading of hazardous VOL's or
VOL's with high vapor pressures. These cars are of course specifically
designed for sealed hatch loading.
Measurement of Liquid Level. Since loading must not proceed to full
nominal tank capacity, liquid level measurement during loading is necessary.
In splash and submerged fill loading, visual sighting of the liquid level
and/or use of a graduated dip stick is a frequently used liquid level
measuring technique. A type of level sensing device used in sealed hatch
loading comprises a magnetic float surrounding a closed tube extending
below the required final liquid level. The magnetic float, which rises
with the rising VOL, actuates a corresponding magnet on the end of the
measuring rod positioned within the closed end tube (Figure 12-3).
Sampling of VOL Cargo During Loading. Conventional industrial quality
control practice requires sampling of the VOL loaded into the tank.
Ideally, sampling is done in such a way that the sample is representative
of the composite material in the tank. In splash loading or submerged
fill loading samples are conveniently dipped from the tank through the
open hatch as required. In closed hatch loading samples are taken at
appropriate intervals through a valve in the piping to the loading arm.
Temperature Measurement of Loaded VOL. VOLs are loaded at a wide range
of temperatures. Since density and volume vary with temperature, it is
necessary for the shipper to measure the temperature of the VOL loaded and
12-10
-------
FIGURE 12-3
MAGNETIC LIQUID LEVEL MEASURING DEVICE
COCxXxXO
GRADUATED MEASURING ROD
TOP OF TANK CAR
TANK CAR SHELL
CLOSE-END TUBE (NON-MAGNETIC)
MAGNETIC FLOAT
LIQUID LEVEL
FLOAT STOP
-------
correct the volume of VOL from standard conditions in determining the
outage required by USDOT regulations. Such temperature measurements are
made either through the open hatch during splash or submerged fill loading,
or, in the case of closed hatch loading, by inserting a thermometer through
the thermometer well.
Unloading of VOL's. Since many VOL's are not chemically reactive with
oxygen, air is vented or pressurized into the car during unloading. If
such materials are chemically reactive with oxygen, an inert gas such as
nitrogen is vented or pressurized to the tank while unloading.
Sources of Emissions
Emissions occur during the VOL loading of railroad tank cars as VOC's
in the vapor space of the tank car are expelled from the tank by the rising
level of the VOL being loaded. Residual VOC's from the previous unloading
operation are present in the vapor space of the tank (unless the tank car
was cleaned after unloading). In addition, VOC emissions are also produced
during the loading operation through evaporation of the VOL being loaded.
In the case of splash loading through the open hatch, some entrainment of
VOL droplets takes place.
The following equation") can be used to calculate an emission factor
for a VOL loaded into a railroad tank car:
LL = 12.46 SPM
T
where: LL = Loading loss (lb/10 gal of liquid loaded)
M = Molecular weight of vapors (Ib/lb-mole)
P = True vapor pressure of liquid loaded (psia)
T = Bulk temperature of liquid loaded (°R)
S = Saturation factor (see Table 2-1).
= 0.5 for submerged loading of a clean tank car
=0.6 for submerged loading of a dedicated service tank car
= 1.45 for splash loading of either a clean or dedicated
service tank car
Data on quantities of VOL's shipped during 1978 was obtained from the
U.S. Department of Transportation Waybill Study^). This data was used with
emission factors obtained from the above equation to calculate approximate
12-12
-------
total potential emissions for the continental United States for 1978. A
saturation factor of 0.6 was used in the calculations as being typical of
a reasonable "worst case" situation, although it is recognized that signifi-
cant quantities of VOL's are loaded with some type of existing emission
control.
On this basis, Table 12-2 presents the maximum potential emissions for
the quantity of VOL's with vapor pressures between 0.5 and 13 psia loaded
into railroad tank cars during 1978.
CONTROL TECHNOLOGY
This section discusses suitable methods currently practiced for the
collection of VOC's emitted during the loading of VOL's into railroad tank
cars, and presents effective techniques for the control of these VOC's.
Vapor Collection. Vapor collection is the physical collection of the
VOC vapor expelled from the tank car during the loading of a VOL. Three
vapor collection techniques are currently practiced within the railroad
tank car loading industry:
0 Collection by vapor balance, in which VOC emissions vented from the
railcar tank are returned to the vapor space of the VOL storage
vessel,
0 Suction pipe collection in open hatch loading of VOL, in which the
VOC emissions are collected at the open hatch with a suction pipe,
0 Collection through the vent port, in which VOC emissions are
collected directly from the vent port of the top unloading piping
arrangement incorporated into many railcars.
A fourth technique, use of a hatch-sealing plug, is in common use for
top loading of gasoline tank trucks. This technique is satisfactory in
the gasoline truck loading industry because of the relatively few standard
hatch diameters installed in gasoline tank trucks, and because gasoline
tank trucks to be loaded at a bulk terminal are generally in captive use and
accordingly are fitted with a specific hatch size to mate with the hatch
sealing plug in use at that terminal. Railroad tank cars however, have been
built with hatches of many different diameters. It would be impractical to
have loading arms fitted with hatch sealing plugs of the diameters required
to mate with each of the many railcar hatch diameters in the existing railcar
fleet.
12-13
-------
TABLE 12-2
TOTAL VOL TANK CAR LOADINGS AND EMISSIONS
1978
Gallons of VOL1s loaded Maximum Potential VOC Emissions,
Millions Tons
4,908 5,300
12-14
-------
Collection by Vapor Balance. VOLs with high vapor pressures (i.e.,
above 10 psia @ 60°F) are occasionally loaded into railroad tank cars
designed for transport of liquified gases (Figure 12-4). The loading
equipment of such cars is designed to permit sealed loading with measure-
ment of outage and temperature while venting the vapor space of the tank
car to the vapor space of the liquified gas storage tank (vapor balance).
Vapor balancing permits sealed hatch loading of high vapor pressure
VOLs with no loss (other than leaks, spills etc.) of VOCs to the atmosphere.
This loading technique is also frequently used in loading of flammable,
toxic, or reactive chemicals, the vapors from which must be contained for
reasons of health or safety.
Collection of VOCs During Open Hatch Loading of VOL. VOLs of moderate
vapor pressure (i.e., up to 5 psia @ 60°F) are commonly loaded through the
open hatch of railroad tank cars. Collection of VOCs expelled from the
open tank car hatch is difficult. Occasionally, a flexible suction hose
is positioned within the open hatch (Figure 12-5), and some collection of
VOCs by this suction hose will take place. Collection efficiency of such
a device is very poor, however, and, a significant portion of the expelled
VOCs escape collection and are released to the atmosphere.
Collection of VOCs Through Top Unloading Piping. As noted earlier, many
general service tank cars are fitted with a top unloading piping arrangement
in which gas pressure is used to force the liquid to be unloaded out the
unloading pipe which extends to the bottom of the tank car.
This piping arrangement has been adapted for VOL loading (Figure 12-6).
The unloading pipe is used as a VOL loading pipe, and the gas pressurization
port is used for VOC venting. As the VOL is loaded through the unloading
pipe, the rising liquid level expels VOC vapors from the tank through the
gas pressurization port, thereby collecting the VOC vapors for control.
VOC Vapor Conditioning
VOC vapor conditioning refers to the saturation of the vapor with
additional VOC or the dilution of the vapor with air. The VOC vapor stream
is saturated with additional VOC to increase the total VOC concentration
12-15
-------
FIGURE 12-4
_ o
CD ¥•
O LU
_J O
o ce
> LU
CO
— CO
ex a
o —
a. —i
C9
f
60
LU CO
Sss
S ^ c=
12-16
ENGINEERING-SCIENCE
-------
FIGURE 12-5
CD
—
*-
0
• >•
CD
UJ
a:.
o
o
o
ae
-t
CJ
es
ea
CU
CO
CJ
<=9
o
a
12-17
ENGINEERING-SCIENCE
-------
FIGURE 12-6
CD
Q_
Q_
CD
a ce.
0. H-
O
ce
UJ
oo
i— o:
CJ UJ
cr
C3
a.
o
o
12-18
ENGINEERING-SCIENCE
-------
to a level above the upper explosive limit (UEL), or is diluted with air to
a level below the lower explosive limit (LEL). VOC vapor conditioning
should occur as soon as possible after VOC collection, to avoid transport
of a potentially explosive VOC stream.
In the loading of VOLs into railroad tank cars the composition of the
expelled stream is influenced by the type of service the tank car is in, the
prior unloading method, and whether cleaning is performed between unloading
and loading.
Not all vapor streams collected from tank cars would require condi-
tioning. In those cases where the tank car is in dedicated service and
has not been cleaned following its last unloading, the VOC concentration
in the vapor space within the tank car is frequently, but not always,
above the UEL.
Tank cars which were top unloaded under nitrogen pressure will contain
VOCs diffused in nitrogen. Such a mixture is not explosive at any VOC con-
centration because of the absence of a significant concentration of oxygen.
No vapor stream conditioning would be necessary if such a tank car were
subsequently loaded under sealed hatch conditions.
Railroad tank cars are cleaned between unloading and loading if the
tank has been contaminated in some way, or if the VOL to be loaded differs
chemically from the VOL previously loaded. Tank cars which have been
cleaned will have little or no VOC in the tank. Sealed hatch loading of a
clean car would not require conditioning of expelled vapor if the tank car
interior was purged with nitrogen as a final step in the cleaning process,
or purged with nitrogen prior to loading, as is frequently done.
There are some cases however, where vapor conditioning of the vapors
collected from loading of VOL into railcars would be appropriate to ensure
safe operation.
Saturation of Vapor Stream with VOL. The equipment used for vapor
stream saturation is, in its simplest form, a chamber incorporating nozzles
through which VOL may be injected as a mist into the VOC stream to be
saturated. To achieve maximum efficiency, VOL is usually injected counter-
current to the flow of the VOC stream, and baffles are frequently included
12-19
-------
in the saturator to increase turbulence and promote maximum vapor-liquid
contact. The saturator should be located as near as possible to the VOC
collection point to minimize the distance over which a potentially explosive
VOC stream is transported before saturation.
The VOL saturant may differ chemically from the VOC in the vapor stream
to be saturated. The VOL saturant is selected on the basis of cost, volatil-
ity, availability, and compatibility with the vapor stream to be saturated.
Dilution of Vapor Stream with Air. In diluting the VOC stream col-
lected from the railcar during the loading operation, the VOC stream is
discharged into ductwork leading away from the collection point(s) (Figure
12-7). Air is drawn into the ductwork upstream from the collection
point(s) by a blower located downstream from the collection points. The
ductwork and blower are sized to dilute the maximum possible amounts of
collected VOCs to 25 percent LEL.
This ductwork collection system provides immediate dilution of the VOC
stream at its point of collection, and safe transport of the VOC stream
away from the terminal. Considering that VOC vapor streams from loading of
railroad tank cars are frequently of VOC concentrations near or above the
DEL, considerable dilution with air is required to reduce these concentra-
tions to 25 percent LEL. Consequently, large volumes of dilute VOCs are
produced, and cost of control or destruction is increased significantly.
Vapor Control And Destruction Equipment
Although a number of emission control techniques can be used to reduce
VOC emissions from tank car loading, incineration and flaring are the most
common.
Prior to flaring, however, the VOC's which have been saturated (or which
were above the UEL concentration as collected) are frequently stored temp-
orarily in bladder storage tanks, or are vented to the vapor space of the
intermediate VOL storage tank from which the loaded VOL was drawn. Be-
cause of the large volumes involved, VOC streams which are diluted during
the collection process are not held in temporary storage before destruction
or control.
12-20
-------
FIGURE 12-7
cc.
5
II
30A
O
d
o
o
30A
O
O
o
o
12-21
ENGINEERING-SCIENCE
-------
The purpose of the bladder storage tank is to accumulate and store
the irregular flow of VOC emissions from the intermittent tank car loading
operations, thus evening out the surges and lulls in the flow of VOC's to the
control device. The storage tank consists of a tank which contains an
expandable bladder. Following saturation, the VOC's flow to the storage
tank under pressure, inflating the expandable bladder. When the bladder
has expanded to pre-set volume, the VOC stream flows to the control device.
VOC's vented to the vapor space (vapor balanced) of the intermediate
storage tank from which the loaded VOL is drawn are considered to be in
temporary storage. These vapors are emitted when the storage tank is
refilled with VOL. Since this emission point is remote from the railcar
loading terminal, however, control or destruction of these emissions is
not considered here.
Thermal incineration and flaring are the usual methods for controlling
VOC emissions from loading of VOL's into railroad tank cars.
Catalytic oxidizers are not as widely used as thermal incinerators
because the catalysts can be poisoned by sulfur-and halogen-containing
compounds.
Thermal incineration is a widely used method for control of VOC emis-
sions from railroad tank car loading of VOL's because it is applicable to
a variety of chemicals and VOC vapor stream conditions. Incineration is
commonly the method of control for streams with heating values below the
LEL.
A thermal incinerator is usually a refractory-lined chamber containing
a burner at one end and generally operated at about 760°C (1400°F) with a
residence time of from 0.3 to 1 second^) (Figure 12-8).
Combustion chamber temperature is an important parameter in the design
of a thermal incinerator, since oxidation rates are highly temperature-
dependent. Incineration of a low heating value VOC vapor stream necessi-
tates the burning of an auxiliary fuel to achieve the desired chamber
temperature. Destruction of VOC's with about 90 percent efficiency occurs
rapidly at temperatures over 760°C (1400°F).
12-22
-------
FIGURE 12-8
.
I x
CC
o
o
as
o
i—
V)
CO
eo
CO
00
12-23
ENGINEERING-SCIENCE
-------
Mixing is crucial in achieving good thermal incinerator performance. A
properly designed incinerator will rapidly combine the VOC, combustion air,
and hot combustion products from the burner to ensure that the VOC will be
in contact with sufficient oxygen at a temperature high enough to start
the oxidation reaction. Improper mixing can enable pockets of VOC to pass
through the incinerator intact and can lead to poor temperature distribu-
tions where not all the VOC vapor stream reaches or remains at the
combustion temperature.
Residence time is the time available for the oxidation reaction to
occur within the combustion chamber. Residence times from as low as 0.3
seconds to several seconds have been used in thermal incinerator design^).
Based on a study of thermal oxidizer efficiency, cost, and fuel use,
98 percent VOC reduction is the highest reasonable control level consis-
tently achievable by new incinerators, considering current technology-^).
This degree of VOC reduction is based on thermal incinerator operation at
870°C (1600°F) and 0.75 second residence time.
Flaring is generally used in the petroleum and chemical industry as a
means of controlling rapid and unexpected releases of combustible gases.
Flares can handle a wide variation in gas flow rates and gas heat capacities.
The concentration of VOC's in vapors to be flared is frequently greater than
the upper explosive limit.
Flares are generally less expensive to install and operate than thermal
incinerators. Flares are also generally noisier than incinerators and can
produce vibrations which may be considered a nuisance.
There are three types of flares; elevated, ground, and pit. Elevated
flares are most common (Figure 12-9). As in all combustion processes, time,
turbulence, and temperature control the effectiveness of combustion. Flared
gases must be kept at or above their auto-ignition temperature to combust
completely. The turbulence is sometimes supplied by steam injection at the
ports in the stack. Time is controlled by the rate of release of the flared
vapor.
12-24
-------
FIGURE 12-9
ELEVATED FLARE
STEAM JET
GNITER
t
FLARED GAS
PILOT
GAS-AIR MIXTURE
12-25
ENGINEERING-SCIENCE
-------
Flared VOC vapor streams are ignited by a pilot light just as they reach
the top of the stack. Before thorough mixing can occur, part of the VOC
vapor may burn. This may cause an oxygen deficiency which can result in
carbon formation. The as-yet-unburned VOC vapors may also "crack" to form
smaller reactive molecules of olefins and paraffins, which may allow some
molecules to polymerize into long chain hydrocarbons. More carbon may be
created from the combustion of these long chain hydrocarbons. Steam
injection can reduce the amount of carbon formation by the mixing action
resulting from its injection. Reduction in the amount of carbon produced
will help to eliminate flare smoking problems. For VOC vapor streams with
heating values below 100 Btu/ft-^> 11/ auxiliary fuel may be required
for effective combustion.
The design of a flare depends on the VOC vapor stream to be flared. The
actual VOC composition of the vapor stream effects the time, turbulence, and
temperature requirements for complete combustion. The temperatures of
combustion can be controlled by adding supplemental fuel and additional
pilot lights. The turbulence can be controlled by flare stack design and
steam injection. The time of burning can be controlled by flare stack
design and VOC vapor stream flow rates. Flow rates to the flare are
controlled through design and pre-set operation of the temporary storage
bladder tank which provides even flow rate of the VOC vapor stream to the
flare.
To ensure ignition and provide good flame stability the normal opera-
ting range of a flare should be 1-5% of its maximum capacity. The pilot
light must be able to withstand high wind and heavy rain. The steam
injection requirements to prevent smoking depends upon the hydrogen to
carbon ratio of the VOC. A hydrogen to carbon ratio of 0.33 requires no
steam injection, but a hydrogen to carbon ratio of 0.25 may produce smoke
and therefore requires steam injection. The height of the flare is
dictated by heat, fire, and safety considerations.
The available information on flare efficiency is very limited because
of the difficulty of measuring VOC emissions from flares. Literature data
indicates that flare efficiency may vary from as low as 60% VOC destruction
to as high as 99.9% VOC destruction.
12-26
-------
Costs
To develop information on approximate cost of VOC emission control by
incineration for railroad tank car loading terminals, cost effectiveness
calculations have been performed on acetone, a typical VOL frequently loaded
into railroad tank cars. The vapor pressure of acetone approaches the
weighted average vapor pressure of the diverse VOL's included in the
responses to a a partial industry survey^' covering the railroad tank
car loading industry. Calculations were based on annual loading volumes
spanning the range of throughputs found in typical railroad tank car
loading terminals. In this example the acetone emissions were collected by
dilution. The results of these calculations are tabulated in Table 12-3,
and the cost effectiveness results are presented graphically in Figure
12-10, which shows that cost of control decreases as volume of VOL loaded
in the terminal increases. The cost effectiveness modeled in this example
reflects the control of characteristic amounts of VOC's emitted in typical
railroad tank car loading operations, and should not be construed as
typical cost effectiveness values for incineration in other applications.
STATUS OF REGULATORY ACTIVITIES
Regulations covering control of emissions from tank car loading of
VOL's for those States which presently do not meet the national ambient air
quality standards for ozone, and which have been granted extentions beyond
Dec 1982, have been reviewed. Several of those states currently require
the recovery or destruction of vapors emitted as a result of railcar loading
of VOL's. Six States, however do not require this operation to be con-
trolled, while nine others limit the scope of the requirement to the
loading of gasoline. Still others require only that loading be accom-
plished by means of bottom filling or by submerged fill pipe. Those that
require the recovery or destruction of VOC's have basically used as a
guide the control techniques guideline for the loading of gasoline into
tank trucks.
12-27
-------
TABLE 12-3
EXAMPLES OF COSTS AND COST EFFECTIVENESS FOR THERMAL
INCINERATION OF VOC'S FROM RAILROAD TANKCAR LOADING
Example
VOL loaded
MM gallons loaded/year
Equipment cost
Installation costs
Piping and ducting
Taxes, freight, and instrumentation
Total installed capital cost
Direct
Operating labor
Maintenance labor
Electricity
Natural gas
Subtotal
Indirect
II
$5,400
Capital recovery3 $16,000
Taxes, insurance & administration^3 4,000
Subtotal $20,000
Total annualized costc $25,400
Tons controlled VOC emissions/year 4.5
Cost effectiveness $5600/ton
$21,000
5,000
III
Acetone
4
60,000
37,000
4,000
11,000
Acetone
18
$68,000
41,000
8,000
12,000
Ace tone
42
$80,000
49,000
19,000
14,000
$112,000 $129,000 $162,000
$1,500
2,000
100
2,000
$1,500
2,000
200
4,000
$ 1,500
2,000
500
11,000
$7,400 $15,000
$26,000
6,500
$26,000 $32,500
$33,400 $47,500
20 47
1700/ton $10007 ton
aCapital recovery factor of 16.275 percent of total installed
capital cost. Based on 10 year life of incinerator and 10 percent
interest rate.
percent of total installed capital cost.
cDirect subtotal + indirect subtotal.
Note: All costs are rounded to the nearest $100, and are in 1980
dollars. No heat recovery from the incinerator is included
in these example calculations.
12-28
-------
FIGURE 12.10
COST EFFECTIVENESS FOR CONTROL OF ACETONE EMISSIONS
DURING RAILROAD TANK CAR LOADING
6000
Q
UJ
O
o:
o
o
o
UJ
=c
O
t—
oo
o
oo
O
o
5000
4000
3000
2000
1000
Note: VOC emissions
diluted during collection.
O
20
30
40
50
THROUGHPUT, GALLONS ACETONE/YEAR x 10C
12-29
-------
REFERENCES
1. Yearbook of Railroad Facts. Association of American Railroads, Econo-
mics and Finance Department, Washington, DC, June 1980.
2. Carload Waybill Statistics, U.S. Department of Transportation, Washing-
ton, DC.
3. Association of American Railroads, Washington, DC, private communica-
tion, Patrick Student to Fred Porter, December 23, 1980.
4 Survey Responses from Partial Industry Survey, private communications to
Don Goodwin, December 1980 through January 1981.
5. Tank Car Manual, General American Transportation Corp., Chicago, 111.
1979
6. Compilation of Air Pollutant Emission Factors, Third Edition and Sup-
plements, AP-42, U.S. Environmental Protection Agency, Research
Triangle Park, NC, August 1977.
7. Evaporation Loss from Tank Cars, Tank Trucks, and Marine Vessels API
Bulletin 2514, American Petroleum Institute, Washington, DC, November
1959.
8. Reed, R. J. North American Combustion Hanbdook. Cleveland, North
American Manufacturing Co., 1979. p. 269.
9. Chemical Rubber Company (CRC). Handbook of Chemistry and Physics,
49th Ed.
10. Memo and addendum from Mascone, D., EPA, to Farmer, J., EPA. June 11,
1980.
11. Control Techniques for Volatile Organic Emissions From Stationary
Sources. Office of Air and Waste Management. U.S. Environmental
Protection Agency. Research Triangle Park, NC. Publication No.
EPA-450/2-78-022.
12-30
-------
SECTION 13
SOLVENT EXTRACTION PROCESSES
PROCESSES AND EMISSIONS
The extraction of fats and oils is accomplished through rendering,
pressing and solvent extraction. The process to be used is determined by
the characteristics of each oil-bearing material. Tallow and grease from
fatty animal tissues is recovered by rendering as are marine oils from
fish and whale blubber. Mechanical expression (by continuous pressing or
centrifugal expression) is used to obtain the major portion of the oil
from various oil seeds such as copra, sunflower seeds, castor bean, and
peanuts, as well as from fruit pulp, generally olives. This process is
also used to recover oil from rendering residues. However, for maximum
recovery of oil from oil seeds, extraction is done especially with oil
seeds having a low oil content. The final recovery of the oil is accom-
plished through mechanical expression. Soybeans are almost always ex-
clusively solvent extracted directly. A process called "prepress solvent
extraction" is used for cottonseed, flaxseed, and for high oil content
seeds, including copra, peanuts, corn germ, castor beans, sunflower
seeds, olives and wheat germ.
Soybean Oil Extraction
The emphasis in this section will be on soybean oil extraction since,
according to the Research Triangle Institute (RTI), 80 percent of all
the vegetable oil produced in the United States is extracted from
soybeans-'-). Finelt of Fluor Engineers and Constructors reports that
soybean oil accounts for 60 percent of the total food use of fats and
oils2).
Soybean oil is obtained through solvent extraction via hexane.
This is a continuous processing operation (see Figure 13-1). This figure
illustrates a typical plant operated to produce crude soybean vegetable
oil and a soybean residue meal suitable for ultimate use as animal feed.
Mechanical pretreatment of soybeans containing about 18 percent oil is
required to facilitate the separation of the oils from the solids. In
mechanical pretreatment, foreign materials are removed, the beans are
13-1
-------
FIGURE 13-1
gd-—4
CO
o>
o •
a
(/>
c
CO
o
00
(T3
OC
CO
0=
a
-
o
oo
13-2
ENGINEERING-SCIENCE
-------
dehulled, reduced in size and crushed. Prepared soybeans are then fed
directly to an extractor where they are then introduced to a series of
washings with warm hexane solvent to extract the vegetable oil. The
solvent must then be removed from the oil in the solvent recovery section
by means of multiple stages of evaporation and stripping. The crude
soybean oil is stored and the recovered hexane solvent is passed to a
solvent work tank. The recovered hexane is recycled and used again at
the extractor.
As the meal (soybean residue) leaves the extractor, it is subjected
to a stream stripping operation in the meal desolventizer. This removes
most of the hexane retained in the meal. The water-hexane vapors, in
addition to the liquid solvent and oil, are passed to the solvent recovery
section for oil separation and solvent recovery, while desolventized
meal is dried and cooled to produce a soybean meal product.
Process Emissions and Emission Points
There are three vent streams containing hexane which are discharged to
the atmosphere during soybean processing. These three emission points
are: (1) the main vent stream issuing from the solvent recovery section,
(2) the dryer vent stream issuing from the meal dryer, and (3) the cooler
vent stream exiting from the meal cooler (see Figure 13-1).
The incentive to reduce hexane emissions to the atmosphere at soybean
processing operations is high since hexane is expensive, thus making
recovery operations very desirable. The profit is dependent to a large
extent upon keeping the hexane losses as close to zero as possible in
extraction operations. In most cases, according to RTI, 99.9 percent
of the hexane is recoverable. The remaining 0.1 percent which is lost to
the atmosphere, at first glance may not seem like a significant amount
to the industry; however, in terms of total hexane emissions as an air
contaminant, this 0.1 percent can be highly significant. This is so
because the typical operational ratio of hexane weight to dehulled oil
seed weight is about one to one. Obviously, the larger the operation,
the greater the hexane emissions.
13-3
-------
CONTROL TECHNOLOGY
As reported by Finelt in the November 1979 APCA Journal, the losses
of total hexane to be expected from soybean processing plants may vary
from 0.5 gals./ton of soybean processed up to 5.0 gals./ton, with a
mid-range of 0.7 to 2.7 and an average of about 1.0 to 1.4 gals./ton.
As would be expected, the newer plants tend to have lower emissions due
to the incorporation of additional hexane collection equipment such as a
mineral oil adsorber/stripper unit to the main vent stream. Assuming an
average value of 1.2 gals./ton and 18 percent oil in soybeans, Table 13-1
shows that in 1977, approximately 29.5 MM gals, of hexane were discharged
to the atmosphere as a result of soybean process operations^).
As mentioned previously, the consideration of emission points at such
an operation can be summed up as follows: the main vent stream and a
combined dryer/cooler vent stream. These later two can be combined due to
the similarities of their origins and the low potential hexane content
of each stream. Table 13-2 presents the expected flow quantities of the
two vent streams from any given soybean plant. It is anticipated the
air pollution abatement facilities that would be installed would reduce
the hexane level of the emitted stream to 100 ppm or less^'.
The two primary air pollution abatement systems which could be ap-
plied to soybean processing operations are thermal oxidation and mineral
oil adsorption (currently the most widely used).
Thermal oxidation systems possibly applicable for the reduction of
hexane emissions from solvent extraction include direct thermal incinera-
tion and catalytic oxidation. However, vent streams from the dryer and
cooler contain particulate matter which could foul the catalyst in a
catalytic oxidation system^). Thus, direct thermal incineration appears
to be the most feasible of two oxidation approaches. Because uncon-
trolled VOC emissions from the dryer and cooler are rather low in
concentration, the cost-effectiveness of thermal oxidation would be
rather poor unless emissions from the dryer/cooler vents were combined
with those from the main vent condenser. One other alternative oxidation
approach would be to vent VOC emissions from the dryer/cooler vents to
the plant's main boiler.
13-4
-------
TABLE 13-1
PRODUCTION OF SOYBEAN OIL IN U.S.
Year
MM Ib.
Year
MM Ib,
1960
1965
1966
1967
1968
1969
1970
4392
5236
5811
6150
6150
6805
8086
1971
1972
1973
1974
1975
1976
1977
8082
8084
7540
8705
7862
9640
8836
TABLE 13-2
COMPOSITION OF VENT STREAMS FROM
SOYBEAN PROCESSING PLANTS
Main Vent
Combined
Dryer/Cooler Vents
Solvent
Air Volume, ACFM
Temperature (°F)
Quantity of Solvent, Ib./day
Hexane
1070
80°
209-20900
Hexane
55860
210°
397-39750
Reference 2
13-5
-------
Carbon adsorption has been considered as a VOC control technique and
a few systems have been installed on solvent extraction plant vent streams.
However, high cost and the potential risk of fire and explosion have made
these systems unattractive to industry.
According to a draft Background Information Document (BID) prepared
on Vegetable Oil Processing for the U.S. Environmental Protection Agency,
the most widely used VOC control system on solvent extraction main vents
is the Mineral Oil Scrubber^). These scrubbers can control main condenser
vent VOC emissions at an efficiency of about 95 percent.
The type of oil used for hexane absorption is white mineral oil of
technical grade with a molecular weight of between 290 and 330, and a
viscosity at 100°F of 50/60 second Saybolt Universal. This recovery
process is located at the main vent and is reported to be approximately
95 percent efficient-^-).
Volatile organic compound losses occur throughout the process in
varying amounts. In addition to the emissions of the main vent, the dryer
vent also can experience considerable VOC losses. It is possible that
the hexane concentration at the dryer vent can be in the flammable range.
At present there is no applicable recovery technology for this particular
vent, according to Chessin of RII-^.
Fugitive loss of VOCs (which are not proportional to plant size^))
vary greatly from operation to operation. Generally, they range from
0.13 gals./ton of soybeans processed to 0.33 gals./ton.
Fugitive emissions result from motor driven shafts with packing
glands and stuffing boxes of solids processing units, sight glasses,
cleanouts, and access doors also leak vapor if they are not maintained.
Control Technology Costs
Solvents such as the commonly used hexane are expensive, and
efficient solvent recovery is favorable for both air pollution abatement
and for the reuse of the solvent.
13-6
-------
Using the soybean industry as an example of a solvent recovery
operation, volatile organic compound losses to the atmosphere can amount
to about 0.5 percent of the original quantity used-^). In 1979, the mid
price for the solvent was $0.70/gal.2). In a reasonably well-operated
plant, operating at 1,000 tons/day production, the solvent losses can
amount to 3.15 gals./ton of soybeans processed-^). This amounts to a
dollar loss of $2,205/day (this is in about the mid-range for VOC losses,
as discussed earlier in this section)
Available control cost data were investigated for three control
alternatives; these were: an activated carbon adsorber on the main vent
or combined with the dryer/cooler vent and treated as a whole; a thermal
incineration unit on the combined dryer/cooler vent, or combined with
the main vent and incinerated together^); finally, the mineral oil
scrubber, which is considered by the EPA to be an integral part of the
entire process and not necessarily an add-on abatement technique.
The cost (in 1979) for a carbon adsorption system is approximately
$108,200, with operating costs of $67,860 to $102,420 per year, depending
on the amount of hexane to be handled (209 Ibs./day or 1,045 Ibs./day,
respectively^'). With this type of abatement equipment, the losses of
hexane can be reduced to less than 0.5 gals, per ton of soybeans
process^).
According to a representative of the Petroleum Branch of the Environ-
mental Protection Agency, the cost of a thermal incineration unit on the
main vent is $69,500 and $101,200 on the dryer vent. However, hexane
revenues are not realized until the hexane content exceeds 18,000 Ibs./
day.
As for the cost of a mineral oil scrubber, this particular type of
equipment is so integrated into the total operation, it would be
difficult to single out the price of the unit, its installation or its
operational costs, as reported by a representative of the Petroleum
Branch of the EPA where considerable cost analysis has been done on
such equipment.
13-7
-------
Coffee Roasting and Decaffeination
Thirty percent of the coffee beans processed in the United States
are used for instant coffee and only five percent of the beans are used
for decaffeinated coffee. The majority of the beans are simply cleaned,
blended, roasted, and packaged for sale^'.
Volatile organic compounds are emitted during the roasting and
decaffeination processes. A direct-fired afterburner operating in a
range of 650 to 750°C (1200-1400°F) will almost completely eliminate the
emissions from the continuous batch roasters. Prior to roasting, decaf-
feination is accomplished with trichloroethylene. As of 1974, 1,500
tons per year of volatile organics were emitted to the atmosphere. For
the solvent loss (trichloroethylene) which is the major emission, there
is no control technique used'*).
STATUS OF REGULATORY ACTIVITIES
Currently, the Fresno County Air Pollution Control District (Cali-
fornia) has a rule (No. 409) on organic solvents. However, if the
solvent does not come in contact with a flame, the particular process
is exempt from the rule. The South Coast Air Quality Management
District (SCAQMD [California]) also has a rule on organic solvents and
the associated emissions (No. 442).
Similar to the Fresno County rule, the SCAQMD exempts those
industries utilizing organic solvents under certain conditions such as
having the emissions reduced by 85 percent or the following:
(1) Organic materials that come into contact with flame or are
baked, heat cured or heat polymerized, are limited to 1.4
kilograms (3.1 pounds) per hour not to exceed 6.5 kilograms
(14.3 pounds) per day.
(2) Organic materials emitted into the atmosphere from the use of
photochemically reactive solvents are limited to 3.6 kilograms
(7.9 pounds) per hour, not to exceed 18 kilograms (39.6
pounds) per day, except as provided in subsection (a)(l).
13-8
-------
All organic materials emitted for a drying period of 12
hours following their application shall be included in this
limit.
(3) Through and including November 30, 1980, organic materials
emitted into the atmosphere from the use of non-photo-
cheinically reactive solvents are limited to 180 kilograms
(396 pounds) per hour not to exceed 1350 kilograms (2970
pounds) per day, except as provided in subsection (a)(l).
All organic materials emitted for a drying period of 12
hours following their application shall be included in this
limit.
(4) On and after December 1, 1980, organic materials emitted into
the atmosphere from the use of non-photochemically reactive
solvents are limited to 36.8 kilograms (81 pounds) per hour
not to exceed 272 kilograms (600 pounds) per day, effective
December 1, 1980.
REFERENCES
1. Telephone discussion with Carl Parker and Bert Chessin of RTI, Dec.
1980.
2. Finelt, Stanley, Air Pollution Abatement Facilities at Soybean
Processing Plants. Journal of Air Pollution Control, Nov. 1979.
3. Draft NSPS, February 1980, supplied by EPA, February 1981.
4. Control Techniques for Volatile Organic Emissions from Stationary
Sources. U.S. EPA-450/2-78-022, May 1978.
5. Chemical Engineering Journal, Aug. 6, 1973.
6. Cheretaisinof f, Paul N. , and Young, Richard A., Pollution Engineering
Practice Handbook. Ann Arbor Science Publishers, Inc., 1976.
7. Discussion with representative of the Petroleum Branch of the
Environmental Protection Agency, Durham, NC, March 1981.
13-9
-------
SECTION 14
SURFACE COATING OF LARGE AIRCRAFT
PROCESSES AND EMISSIONS
Aircraft coating operations are centered about major aircraft
manufacturing and repair centers in the United States. In the South
Coast Air Quality Management District (SCAQMD [Southern California])
the coating of aerospace components (a term used by the SCAQMD inclu-
ding aircraft coating) alone contributes 4.4 tons per day of volatile
organic emissions.
Painting operations result in the generation of several types of
contaminants: various quantities of nonvolatiles (solid or semi-
solid nonorganic materials) and vapors of organic solvents. These
volatile organic compounds (VOCs) are not captured and retained by
air filters or waterfall curtains typically used for particulate con-
troll^,3). in the case of aircraft painting and coating, afterburners
or incinerators essentially do not play a role in pollution abatement
efforts^*5). However, activated carbon adsorption has recently been
applied as a method of VOC abatement in aircraft coating.
The aviation industry, from an operational point of. view, is
generally classified into three categories: commercial, general ser-
vice, and military. For the most part, commercial aircraft are large
multi-engine fixed-wing airplanes. Planes used in general service are
generally much smaller, with many in the single engine, two or four
passenger category. There are also increasing numbers of 6 to 20
passenger aircraft used for business purposes. Military aircraft cover
all sizes and types, from light single engine planes to the largest
multi-engine cargo planes. A variety of helicopters are also in the
military category. Coatings used for military aircraft must meet
military specifications. Most military aircraft are coated with flat
camouflage paints as compared to the gloss decorative paint used on
civilian aircraft.
14-1
-------
Fron a point of view of coatings and coating application methods,
the most important distinguishing features seem to be (1) size of
aircraft, and (2) whether the aircraft is for civilian or military use.
Painting operations of both small and large aircraft manufacturers
are described here.
Three manufacturers of light aircraft (Beechcraft, Cessna, and
Gruman) reported that very few, if any, of the exterior components of
the aircraft come to them prepainted or even preprinted. Prior to
assembly, light aircraft components are generally coated with two
primers and a topcoat. The first coat is called a filling primer
which smooths out the surface of the component; the second coat is an
anticorrosion primer which is applied to protect the surface of each
exterior component; finally, the topcoat (either clear or colored) is
applied for the final protection and decoration of the craft. Most
light aircraft, however, have colored (pigmented) topcoats.
Both the primers and the topcoat use MEK, M1BK, or toluene as
solvents. The primers are of the epoxy-polyamide type and the topcoat
paint is generally a polyurethane.
Depending on the normal practice of the aircraft manufacturer, the
first primer coat is applied either electrostatically in an automated
fashion within a waterfall booth, or is conventionally sprayed with
compressed air. The second primer and the topcoat are both applied
using a hand-held compressed air spray gun. This process is also done
in a waterfall booth. The air from the painting area is filtered, and
the water from the waterfall booth is subjected to paint reclaiming
(paint is disposed of) and then is recycled. These two control methods
are well suited to the capture of particulates, however, the volatile
organic compounds are not captured by these methods.
Large aircraft are generally painted with only one primer coat
and one topcoat. Both of these coatings are applied by the conventional
compressed air spray in a controlled environment paint hangar. In
essence, this means that hangar inlet air is filtered.
Large aircraft are usually painted as an entire unit. there is
one epoxy-type primer coat and one polyurethane topcoat. Some airlines
14-2
-------
prefer their planes without color, and as such are coated with a pro-
tective acrylic prior to the riveting; this coating is then removed
after the assembly of the craft. The sheet stock is polished aluminum
coated with "Alodine'"" which allows the metallic finish of the craft
to show through.
The solvents used are generally MEK and toluene, however, there
are many blends which may contain any combination of ketones, alcohols,
and glycol ethers^).
In the case of at least one manufacturer (Lockheed) the wings are
painted prior to attachment to the fuselage. In addition, they use
two primer coats, one called a wash primer which is approximately 85
percent solvent. This does present a high potential for volatile
organic losses; however, according to a finishes expert at Lockheed,
the amount applied per unit area is low. The second primer is an
intermediate coat, a polyurethane which is high in solids, as is the
following topcoat which is a polyesterurethane. The process used for
applying the primers and the topcoat involves a gantry on wheels which
runs along a track. Several painters positioned on each side of the
gantry at different levels use compressed air spray guns for even
paint application. The wings are painted in the open area of the
hangar. Fans are used to direct the over-spray toward the fiberglass
filters. According to a contact at Lockheed in Burbank, California,
it is a general practice for most airlines to completely strip off the
old primer and paint and reprime and repaint the aircraft every five
years.
As with all surface coating operations, the emissions of volatile
organic compounds are dependent upon VOC content of the coating, the
amount of coating applied, and the efficiency of the application method
in transferring the coating to the object to be coated. No detailed
emission factors are available for this category, but solvent-borne
coatings similar to those used in coating other metal products are
utilized. Coatings for main body and wing assemblies must be air dried.
No waterborne coatings were found to be used at the present time.
14-3
-------
CONTROL TECHNOLOGY
The surface coating of aircraft involves the coating of large
irregularly shaped objects at relatively low production rates. This
very nearly precludes the use of add-on controls for VOC losses be-
cause of the very large exhaust flow rates and flow VOC concentrations
in the exhaust. Both capital and operating costs would be extremely
high. Consequently, control techniques principally involve the use of
high solids or waterborne coatings and the conversion to higher transfer
efficiency coating application systems. There are some special problems
relating to aircraft coatings which have an impact on the ability to
change coating formulations. These include the importance of corrosion
resistance, the need to tolerate high altitude exposure conditions
(extreme temperature ranges and high levels of ultraviolet radiation),
and the need to meet military specifications on a portion of the aircraft
manufactured.
Tinker Air Force Base (Oklahoma City, Oklahoma) has the most modern
aircraft painting hangar presently in operation. The hangar, which can
accomodate the painting of two aircraft the size of a 747 simultaneously,
was designed with two carbon adsorption air pollution abatement units
at either end of the building.
This hangar, which went operational in 1979, was designed with
the control equipment to facilitate compliance with volatile organic
compound regulations. Each carbon adsorber contains three large acti-
vated carbon canisters each, with eighteen smaller canisters. The
two units combined move 136,000 CFM of air through their filters.
Approximately 90 percent of the air in the hangar is recirculated,
requiring only about 10 percent make-up air.
According to representatives of Tinker Air Force Base'), the
carbon adsorption units are effective, but would be even more efficient
if the steam used to regenerate the carbon was at a higher temperature
than the 250°F steam now used. Presumably this would aid in stripping
higher boiling point solvents. No test data were reported.
With these two particular units, the carbon is regenerated every 8
or 9 hours; however, it could maintain its efficiency for as long as 16
hours without regeneration.
14-4
-------
The only major problem encountered so far has been with the small
canisters becoming clogged with paint from overspray. In response to
this, Tinker Air Force Base plans to install bag filters to catch the
overspray that the water-wash booths do not capture.
Coating specification data obtained during this study shows vola-
tile organic compound content in coatings ranging from about 30 to 90
percent by volume range. Reductions in VOC content of aerospace
coatings used in the Los Angeles Basin area (South Coast Air Quality
Management District) are scheduled to begin in August 1982 with a
further reduction of VOC content in primers in 1985 (see Rule 1124 in
Status of Regulatory Activities below). The estimated reduction will
be partially accomplished by using high solids coating in 1982 and
waterborne coating for primer in 1985. The estimated cost of the 1124
SCAQMD rule is approximately $16,000 per ton of organic emissions
reduced.
The application of coatings with improved transfer efficiency could
be considered equivalent to VOC control using high solids or water-
borne coatings. Application methods utilizing electrostatic coating,
hydraulic spray equipment, and special low air volume spray guns
(sometimes described as mistless sprayers) are possible alternative
approaches. Reduction in paint use of 50 percent or more are thought
to be achievable with some of these methods.
It was suggested by the Board of the SCAQMD that aircraft coating
could be eliminated altogether, thereby dramatically reducing organic
emissions. However, representatives of the aircraft manufacturing
industry expressed the concern that hand-polishing the bare metal
surface (in order to keep it smooth for flight) was too costly due to
the required number of person-hours and the price of labor.
STATUS OF REGULATORY ACTIVITIES
Several states maintain regulations on volatile organic compound
emissions from the coating industry, however, these rulings do not
apply specifically to aircraft coating.
14-5
-------
The SCAQMD (Los Angeles Air Basin) in July 1979 proposed its
Rule 1124, Aerospace Assembly and Component Coatings Operations. Its
requirements are stated as follows:
"(b) Requirements
(1) After August 1, 1982, a person shall not:
(A) Apply to aerospace components any primer or topcoat
which contains volatile organic compounds in excess
of:
(i) 650 grams per liter of primer less water, as
applied.
(ii) 600 grams per liter of topcoat less water as
applied.
(B) Apply to aerospace components any temporary protec-
tive coating that contains more than 250 grams of
volatile organic compound per liter of material
less water, as applied.
(C) Use volatile organic compounds of composite vapor
pressure of 77.6 mm Hg (1.5 psia) or greater at a
temperature of 21.1°C (70°F) for surface preparation
or cleanup, excluding paint removal.
(D) Use other than closed containers for disposal of
cloth or paper impregnated with solvent containing
volatile organic compounds which are used for sur-
face preparation cleanup and paint removal.
(E) Use volatile organic compounds for the cleanup of
spray equipment used in aerospace component coating
operations unless 85 percent of the volatile organic
compounds are collected and properly disposed such
that they are not emitted to the atmosphere.
(F) Use stripper which contains more than 400 grams per-
liter of volatile organic compounds, or has a com-
posite vapor pressure of volatile organic compounds
more than 10 mm Hg (0.19 psia) at 21.1°C (70°F).
(2) After January 1, 1985, a person shall not apply to
aerospace components any primer in excess of 350 grains
per liter, less water, as applied.
(3) Until and including August 1, 1982, aerospace industry
and aerospace components subject to the provisions of
this rule shall comply with Rule 442 and any other
applicable requirements of the Rules and Regulations of
the South Coast Air Quality Management District.
14-6
-------
(c) Equivalency
Notwithstanding the provisions of subparagraph (b) a person
nay comply with subparagraph (b) by reducing emissions from
such coating operations provided that:
(1) The emission reducitons are at least equal to those
which would be obtained by the use of coatings specified
in subparagraph (b); and
(2) The emission reducition methods are applied to the
coating operations subject to the provisions of this
rule and are approved by the Executive Officer; and
(3) The owner or operator submits applications for new
permits to construct or operate both basic and control
equipment for such reductions.
(d) Methods of Analysis
The volatile organic content of coatings subject to the
provisions of this rule shall be determined by the procedures
detailed in Rule 107.
(e) Exemptions
(1) Until December 31, 1982, coating of aerospace assemblies
and components procured by the Federal Government shall
be exempted from the provisions of subparagraph (b)(l)(A)
of this rule.
(2) The provisions of subparagraphs (b)(l)(A), and (b)(l)(B),
and (b)(2) of this rule shall not apply to the following
materials:
(A) Coatings for masking in chemical etching operations.
(B) Adhesive bonding paper.
(C) Flight test coatings.
(D) Space vehicles coatings.
(E) Fuel tank coatings.
(3) The provisions of subparagraph (b) of this rule shall
not apply to the following materials:
(A) 1,1,1-Trichloroethane
(B) Methylene Chloride
(C) Trichlorotrifluoroethane
14-7
-------
(4) The provisions of this rule shall not apply to a facility
which emits a total of less than 9072 grams (20 pounds)
of volatile compounds in any one day."
Additional existing air pollution control rules and regulations on
the state and local level have not been specifically directed at this
area of industry, but rather at VOC emissions in general.
REFERENCES
1. Charles R. Martens, Technology of Paints, Varnishes and Lacquers,
Reinhold Book Corporation, 1968.
2. Brian N. Chapman and J. C. Anderson, Science and Technology of
Surface Coating, Academic Press, London and New York, 1974.
3. Contact with the South Coast Air Quality Management District,
December 1980.
4. Contact with Cessna Aircraft, Beechcraft, and Gruman Aircraft,
January 1981.
5. Contact with Boeing Aircraft, Northrop Aircraft, and Lockheed
Aircraft, January 1981.
6. Contact with Bostik, Inc. of Torrance, California, paint and
solvent suppliers to the aircraft industry. January 1981.
7. Personal communication with representative of Tinker Air Force
Base, March 1981.
14-8
-------
SECTION 15
SURFACE COATING OF LARGE SHIPS
PROCESSES AND EMISSIONS
There are about eight different types of marine coatings with
application methods uniquely suited to each. These paint types include:
epoxyamine, epoxypolyamide, polyurethane, vinyl, chlorinated rubber,
alkyd, inorganic zinc, and acrylic. Table 15-1 presents each of these,
and their applications and solids contents.
The aforementioned coatings are air dried and can be applied by
brush, roller, or spray. The application of marine coatings and the
associated solvents used for thinning and cleaning result in the emis-
sion of volatile organic compounds. On the California coast over ten
tons per day of these compounds are emitted-^-).
Since marine coatings are applied to large surface areas and are
dried in the open air, it is not practical to duct the emissions through
abatement devices such as afterburners or carbon adsorbers.
Because of the stringent requirements of marine coatings, applica-
tion characteristics vary with the type of coating. The properties
which must be evaluated are abrasion, impact, flexibility, hardness,
corrosion resistance and cathodic protection, and adhesion.
Some of these coatings require a certain degree of temperature
control to cure properly, while others require heaters within the
paint delivery line. With some paints, humidity is the crucial factor.
Specialized application equipment is required for those coatings which
are solvent-free, while other paints are potentially toxic to the
persons applying them.
According to the California Air Resources Board (CARB), a survey
conducted by them indicated that approximately 26 percent of the marine
coatings sold in California were based on epoxy resins. Half of these
coatings were based on amine-cured resins, and the remaining half were
polyamide cured epoxies. Data from the CARB survey yielded that all the
15-1
-------
rH
i
m
rH
W
_!
S3
<;
H
CO
H
M
H
a
0
cj
CO
Q
1— 1
J
O
CO
Q
§j
A
CO
2
0
rH
H
<
CJ
l_j
rJ
O-l
su
<;
M
CO
w
Oi
>i
H
CO
o
a
HH
s
<<
o
o
w
a
M
CSi
T3
O)
H
rH
&
a.
<
Cfl
03
0)
00
c
03
OS
rH
03
4-1
O
<
0)
6
3
rH
0
>-,
^3
cn
T3
•rH
rH
O
CO
4->
C
0)
o
S-l
s
t 1
0)
4-J
C
rH
S-l
at
>r-
f->
Oi
01
a.
>,
4J
O
•r
S-i
01
c
0)
o
in m m o O CM m
i in m co > 03 01
•H -H f 4J 0
S O 4-1 CO -H
< 0-t OI C CO
M iH 03 -H
>,>,3rH S-IT3 OOrH
XX?^^O>^S-i>^
OOrH CrH-ii OS-I
Q,O.O-HXiH C O
MWO-i>CJ
-------
amine-cured epoxies sold in the state exceeded the required 55 percent
solids recommended by the GARB and that 88 percent of the polyamide cured
epoxies either meet or exceed the 50 percent solids level. According
to paint manufacturers, the 55 percent solids level set for epoxy
coatings (by GARB) can be achieved by the mid-eighties.
Chlorinated rubber coatings have gained wide acceptance in the
marine industry and according to the same aforementioned survey, 82,000
gallons of chlorinated rubber coatings were sold in 1976. Approximately
90 percent of the coatings meet the proposed 35 percent solids level
of the GARB rule.
Polyurethane coatings, with their high gloss and durability,
are expected to meet the constraints of the GARB.
Acrylic coatings offer excellent protection against chalking and
discoloration because of their resistance to ultraviolet radiation.
The GARB did not expand on solids information in this coating category.
Of the vinyl coatings, approximately 45 percent comply with GARB
rule recommendations. The vinyl coatings are used primarily by the
Navy and since the Navy paint specifications require only a 20 percent
solids level, considerable effort on the part of the Navy would have to
be made if a rule such as the CARB's were to be put into effect.
At least three paint manufacturers market waterborne inorganic
zinc primers which meet solids levels of the proposed GARB ruling^'.
Table 15-2 presents a summary of marine coating sales and the
resultant emissions in California. The proposed rule of the GARB is
discussed in detail in the "Regulatory Status" section of this report.
CONTROL TECHNOLOGY
The use of low solvent coatings or of high performance coatings
can achieve emission reductions from ship painting. The number of
low solvent coatings which would be suitable for ships are rather
15-3
-------
OS
a
fa
CO
2;
O
M
CO
CO
M
s
w
cn -
C
O
•H •
cn cn
C
O
cn
C
o
CN
CM
I
CQ
H
S3
CO
W
Q
CO
M
o
cn
d)
cn
co
CO
CO
w
CO
o
o
u
w
z
M
c*
i-J
CO
0)
cO
TJ
O
m
CN
C
o
CO
0)
c.
;>>
H
l-i
0)
C
0)
CJ
•a
0)
c3
C
O
•H
N
a
•H
C!
cO
00
CO
I—I
5^
C
0)
c
CO
0)
5-1
3
00
•H
rH
3
O
C
a)
i—i
rH O
cfl CO
4-1
O 5-)
4J CL)
•° r£
3 4J
CO O
•a
a)
cn
CO
CO
-a
a
-------
limited. However, according to the CARB, the very low solvent coatings
which were originally formulated for such applications as tank linings
can be used on other areas of ships with a little modification.
Emission reductions can also be attained through the use of high
performance coatings, even though they are not exceptionally low in
solvent content. Due to the excellent physical durability, chemical
resistance, and extended coating life of these high performance paints,
less frequent recoating of the ship is required, thus resulting in
reduced VOC emissions.
By incorporating the two control strategies, the use of low solvent
(high solids) coatings, and the use of high performance coatings, the
total VOC emission reduction in California would be to over 3.5 tons
per day by 1985.
Conventional coatings are relatively low in solids content and
generally fail in about one year. Examples of these types of con-
ventional coatings include alkyds, modified alkyds, phenolic, and
resin-base coatings. The high performance coatings can be high or low
in solids content, but possess coating lives at least twice as long as
those of conventional coatings. The high performance coatings include
vinyls, chlorinated rubbers, modified acrylics, polyurethanes, and
inorganic zincs. Table 15-3 presents both conventional and high per-
formance coatings with cost and durability comparisons. By examining
this table, it can be seen that though high performance coatings are
more expensive than their conventional counterparts, the life of a
high performance coating is sufficiently longer to yield greater cost
efficiency.
Availability of High Performance Coatings
In preparation for the writing of this rule, the CARB surveyed
numerous coating manufacturers and found that they all manufactured
some coatings which would comply with the proposed model rule. This
implys that no great economic hardship would be encountered by paint
manufacturers as a result of air quality rulings.
15-5
-------
1
in
rH
W
J
ta
<3^
H
/-— s
^O
r-»
ON
rH
Cd
^
ta
CO
<£
pa
^
CO
w
H
CO
^~*
CO
CO
CJ
z
rH
H
O
CJ
rH
CO
fa
o
C/3
H
CO
o
CJ
rT"|
^
M
H
sg
CM
S
O
CJ
VJ
4-1 ^.
cncM
O 4J
CJ UH
CU
<4H /— N
•H cfl
rJ rJ
M)s_/
C *
•H >>
4J CJ
CO C
O cfl
O
• cu
oo P.
> X
< w
4-1
o
CJCM
4-1
rH 4H
CO ~-~
4-1
O
H
01
00
Cfl 4-1
rH Cfl
rH O
s
,_!
•H
•U 0
OCM
Uii
4-J
<4H
en
>, rH
|£S jg
en
4J
cfl
0
U
6
cu
4-1
C/3
M-H
0
CU
a.
>^
H
rH
CO
0
rH
rH
CO
O
•
en CO
en • P^
^ ^
rH rH
<*
^H i— 1 rH
000
in o o
rH en en
CU
cfl
•H
•a
CO 0 rJ
O )H CU
o cu 6
Cu 4J -H
O C S-4
HMO,
CU CU 0)
-Q ,£5 -Q
-O o ^
333
SH S-l rJ
•a -a TJ
cu cu cu
cO cfl cfl
C fi C
•H -H t-l
rJ !-l rJ
000
rH rH rH
r^ pC -*-
a cj cj
•
SQ
rH
0
-*
en
CM
rH
•
rH O
en '•o
0 O
in in
rH rH
O O
0 0
CM ^t
CU
4-1
Cfl
•H
"O
4-1 0)
CO 0
O SJ
O CU
a 4J
o c
H M
CU CU
•0 -0
•H -H
0 0
cfl co
^"i P^
0 0
a a
^ ^
X X
o o
a a
W M
c
CJ
CM
en
0
^o
rH
o
o
CM
J_j
HI
0
•H
J_j
a.
CU
•H
a
cfl
>,
J
O
a.
^
^
o
a.
w
CO
-^-
0
S -C >H
C 00 O
•H vH C
•
fa
•
rH
C -H
o
CU
4-i a
CO C
O cu
0 r4
&. CU
O UH
H CU
•K Cd
15-6
-------
In the following section, the status of the regulatory activities
will be discussed concerning primarily the CARB's "Proposed Model Rule
for the Control of Volatile Organic Compounds from Marine Coatings
Operations".
REGULATORY STATUS
The GARB in June 1978 proposed a rule entitled "Control of Volatile
Organic Compounds from Marine Coatings Operations", which read as
follows:
"1. After January 1, 1982, no person shall offer for sale or sell
for use in this state or apply any marine coating which contains
more than 295 grams of volatile organic compounds per liter of
coating, excluding water.
2. Until January 1, 1985, the provisions of this rule shall not
apply to the sale or application of:
a. chlorinated rubber coatings containing less than 540 grams
of volatile organic material per liter of coating, exclu-
ding water.
b. Vinyl coatings containing less than 540 grams of volatile
organic material per liter of coating, excluding water.
c. Polyurethane coatings containing less than 420 grams of
volatile organic material per liter of coating, excluding
water.
d. Epoxy coatings containing less than 380 grams of volatile
organic material per liter of coating, excluding water.
3. The provisions of this rule shall not apply to:
a. Antifouling coatings.
b. Wash primers.
c. Marine coatings sold in containers of one gallon or less."
This rule did not advance beyond the proposal stage since, accor-
ding to a representative of GARB, the California shipyard owners made
a strong case that business would be lost to out of state shipyards.
The high performance, high solids coatings are more expensive than the
15-7
-------
conventional paints and apparently ship owners do not see enough advan-
tage in using the more expensive paints even though they possess greater
durability and have a longer potential life (as shown in Table 15-3)2).
A similar comment was obtained from the South Coast Air Quality
District (SCAQMD, Los Angeles Air Basin)-^). A staff member explained
that when the GARB dropped all activity concerning the marine coating
rule, the SCAQMD attempted to promulgate a similar regulation; however,
they also ran up against some similar obstacles. Further, the U.S.
Navy was not prepared to accept the paint limitations in the proposed
rule.
There is a unique quality to developing an emissions regulation
for an industry such as ship coating, simply because the shipyards are
in certain set locations, and ships, no matter where registered,
are free to go from port to port and be painted at the most competitive
prices. Ultimately, these shipyards in areas with air quality regula-
tions requiring the use of high performance, high solids paints could
be out of business by competitors whose yards are in regions with no
regulations. The GARB and the SCAQMD feel the regulation should be
applied at the federal level. However, since some shipyards are in
attainment areas, and shipowners in some cases have the alternative of
using foreign yards, the effect of Control Techniques Guidelines
requirements would also be blunted.
REFERENCES
1. California Air Resources Board, Consideration of a Proposed Model
Rule for the Control of Volatile Organic Compounds from Marine
Coating Operations.
2. Phone discussion with staff member of the California Air Resources
Board, February 1981.
3. Phone discussion with staff member of the South Coast Air Quality
Management District, February 1981.
15-S
-------
SECTION 16
SURFACE COATING OF WOOD FURNITURE
PROCESSES AND EMISSIONS
General discussion
The wood furniture industry which includes the products listed in
Table 16-1 is the second largest source of VOC emissions among surface
coating industries, due to the use of very high volumes of low solids
coatings. Total nationwide volatile organic compound (VOC) emissions
for 1973 from coatings used in this industry were 149,000 metric tons.
In addition, approximately 83,000 metric tons of miscellaneous organic
solvents were used for cleanup and thinning. Table 16-2 compares the
emissions from wood furniture coatings which comprise over 10 percent
of all industrial coating emissions, to other industry coating cate-
gories-^ .
Table 16-1 gives the structure of the wood furniture industry.
There are six SIC codes (Standard Industrial Classification, U.S.
Department of Commerce) which include those companies engaged in the
manufacture of wood furniture. Table 16-3 presents the relative total
number of establishments in each class. Table 16-4 gives the number of
wood household furniture plants by EPA region^).
All wood furniture products are coated in a roughly similar way
although furniture will usually receive a much more elaborate series of
finishes during the coating process than kitchen cabinets.
Wood furniture, especially household, is generally categorized as
being of three levels of quality: namely, high end, medium end, and
low end. On a per unit basis, the relative percentages of each of
these quality levels manufactured are 8, 28, and 64 percent respec-
tively^). The quality of the furniture is determined in part by the
number of coating operations performed on the piece. A low-end piece
might typically undergo from three to six finishing operations, while
a high-end piece of furniture could require 15 or more operations to
obtain the desired finish. This is not universally true because some
of the softer woods used in low-end furniture require additional
finishing steps to achieve a satisfactory finish.
16-1
-------
W
hJ
Pfl
<:
H
I
|
CO
H
co
p
H
M
P
fa
n
o
§
Number of Establishments^- Number of Etiployees^
Small Medium Large Total Small Medium Large Total
CO
4J
O
3
-o
O
u
M
a
CJ
•r-t
a.
>,
H
>,
14
u
M
3
•a
c
i— i
*a
p
u
tn
35
o"
I
en
^o
^o
n
n
— i
in
CN
.-4
m
r-t
00
fH
°'
ON
rN
.— I
r^.
*-T
in
<—*
C
•H
1 OJ -O
•H B
3 e
,0 >s O
LI O
4) O J-i
J3 4J .C
O eg a
en u co
OJ QJ "H U
C C J-i ^
.0 -a c o
fl 3 ca
u u >
c
OJ
1^ iH
£}
o c3
3
-c
fn
-
0
n
-a-
-*
-3-
i— t
4-t
0>
3
*
V
CO
a
o
o
J
*
•o
•o
F— t
o
jn
01
(0
3
S
"8
o
3
-H
m
CN
0)
U-l
o
m
CO
a
.c
u
X
^
u
41
u
3
w
C
3
CC
•H
^j
A
0}
01
T3
CO
U
CO
H
o
£>
cs
>H
<•
^3-
2
.•8
0) U
U OJ
3 -u
4J U
•H r-t
c o
L) _Z
3 C-
<•
CN
r>.
00
O
m
CN
-T
^o
iH
C^
CO
m
ts
r^.
CO
r^.
-H
r*.
OJ
H
"S
O
r*»
fH
m
EN
•o
CO
CO
OJ
c
.0
cC
0
tJ
CO
M
n
w
cO
ej
I
O
|
PH
CO
o
•o
CO
CC
o
OJ
Ll
QJ
4-J
OJ 00
c
,£ CO
CQ m
C a
o
00
c -a
•H C
3 CO
CO CO
M <]J
to c
D -H
W Xl
0! CO
CJ 0
(
J
« ai
"O *H
C j=
(0 O
a
CO tt
00 — >
u
4J
01
c
•H
4^
«
CJ
e
o
X
•H
QJ
OJ
CO
01
c
JD
O
CO
4J
D
C
.C
CO
CN
^O
rH
FH
•
-Sl U
CO
•fl 4)
Li
- 3
U 4-J
L| TH
^
m
•^
m"
m
t— t
CN
CN
-*
r*.
r^-
cn
CN
-H
in
50
0
CN
CO
Li
CO
3
u
(O
pH
a
OJ
JS
c
OJ
ffi
M
d
-H
TJ
•H
3
ca
o
JD
3
CM
!— (
m
m
CN
CJ
3
U
CO
Li 0)
OJ CO
JZ 0)
CJ CO
CD OJ
Li Li
OJ 3
•C *J
U -H
S £
fH 3
j3 14^
•a
0)
4-1 OJ
r3 1-
-( 3
0) 4-1
3C -H
C
•o n
C 3
CO
o
, 3
a o
iH C/!
16-2
-------
TABLE 16-2
VOC EMISSIONS FROM INDUSTRIAL COATINGS
Industry
Estimated Emissions
(1000 metric tons/yr.)
Wood Furniture
Large Appliances
Magnet Wire
Automobiles
Cans
Metal Coils
Paper
Fabric
Metal Furniture
Flat Wood Products
Other Metal Products
Others
TOTAL
230
42
10
140
130
30
390
100
90
84
230
310
1,786
16-3
-------
TABLE 16-3
WOOD FURNITURE CATEGORIES
Percentage of
SIC Code Total Facilities*
2434 - Wood Kitchen Cabinets 31
2511 - Wood Household Furniture, 34
Except Upholstered
2512 - Wood Household Furniture, 23
Upholstered
2517 - Wood Television, Radios, 1
Phonograph and Sewing
Machine Cabinets
2521 - Wood Office Furniture 4
2531 - Public Building and 7
Related Furniture
TOTAL 100
*Based on 5,477 establishments for the six SIC codes.
16-4
-------
TABLE 16-4
WOOD HOUSEHOLD FURNITURE PLANTS BY EPA REGION
(no data available for States not listed)
Region I
New Hampshire
Massachusetts
Connecticut
Region II
New York
New Jersey
Region III
Pennsylvania
Maryland
Virginia
Region IV
North Carolina
South Carolina
Tennessee
Kentucky
Mississippi
Alabama
Georgia
Florida
Region V
Ohio
Indiana
Michigan
Illinois
Wisconsin
Region VI
Texas
Louisiana
Arkansas
Oklahoma
Region VII
Iowa
Missouri
Region IX
California
Arizona
Region X
Washington
Oregon
Total Plants
19
90
21
258
69
107
16
55
153
20
81
28
28
59
92
165
51
75
62
81
26
81
13
38
21
10
33
369
22
34
31
Plants with 20 Employees
or More
13
29
5
59
18
43
4
42
103
9
26
8
10
18
30
28
18
43
31
31
16
23
2
18
4
5
14
111
5
10
9
16-5
-------
Processes
The wood furniture finish is applied in series of steps. There is
a great variety in the number and complexity of coating steps that may
be applied, but the basic ones and their purpose are as follows:
Finish Purpose
Body Stain Gives color uniformity. Develops wood
grain and character.
Wash Coat Seals wood surface; prevents subsequent
unwanted staining from filler coat. Stif-
fens the wood fiber for subsequent sanding.
Filler Fills large pores of wood.
Sealer Seals the wood for application of
subsequent coats.
Glaze, shading Small amounts of color coating that are
stains, padding, added to highlight and give character
spatter to the wood. These are often hand wiped.
Topcoat Usually nitrocellulose lacquer. Provides
clear, durable final finish.
Table 16-5 gives a typical wood furniture schedule. Notice the
number of persons required to perform each step. Wood furniture coating
is a very labor-intensive industry and much of the labor force is
unskilled. This dependency on large amounts of relatively unskilled
labor is one reason the furniture industry may have difficulty adopting
new technology. Furniture finishing is still something of an art and
the techniques, equipment, and procedures may vary considerably from
plant to plant.
In larger furniture factories, furniture pieces are loaded onto
a conveyor line. Workers stationed along the line perform specific
finishing operations as the furniture piece is carried by on the con-
veyor. The various coatings are usually applied by air spray although
dip coating is sometimes used. There is usually a separate spray booth
for each of the coating operations and typically one to three spray gun
16-6
-------
TABLE 16-5
TYPICAL WOOD FURNITURE FINISHING SCHEDULE
Operation
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Operation
Name
Load
Spray uniform stain
Dry
Spray NCR stain
Dry
Spray washcoat
Dry
Sand lightly
Spray filler
Flash-off filler
Wipe filler
Dry
Spray sealer
Dry
Sand
Spray sealer
Dry
Sand
Spray glaze
Wipe and brush
Dry
Distress
Spray lacquer
Dry
Spray lacquer
Dry
Unload
Return to load
TOTAL
Operation Time
Allowed in Minutes
5
1.5
15
1.5
20
1.5
20
1.5
1.5
2
4
45
1.5
30
3
1.5
30
3
1.5
5
60
2
1.5
45
1.5
75
5
15
399
No. of Persons
Per Operation
1
2
2
2
4
2
8
2
7
2
7
2
13
4
2
2
1
l „
63
Source: Technical paper, Society of Manufacturing Engineers, MS 75-251
16-7
-------
operators are in each spray booth. After the furniture piece is sprayed,
the conveyor carries it either to an oven or to the next spray booth.
If the piece travels from spray booth to spray booth, it air dries.
Sometimes there is an oven after a spray booth to give quicker drying.
The wood cannot be heated above approximately 130°F, however, because
at higher temperatures the natural moisture in the wood may be driven
out and damage the coating. (The actual oven temperature may be somewhat
higher than 130°F since the furniture piece does not remain in the oven
very long.) The oven is used mainly to flash-off and dry solvent since
the oven temperature is not hot enough to bake or cure the coating.
Oven drying may be used after any of the coating operations. The
finishing step on which ovens are actually used varies widely from
plant to plant. Sometimes ovens are not turned on in the summer when
the temperature in the finishing room is high.
The coating operation may include various glazes, shading stains,
wiping stains, and padding stains which are usually added between the
sealer and topcoat, but may be added elsewhere as there is a great deal
of variation in furniture schedules. These glazes and shading stains
are usually sprayed on and then hand wiped to give a special appearance
to certain parts of the furniture such as edges and corners. The more
expensive pieces of furniture usually have more of this hand-type rub-
bing work. Hand wiping may also be used after other coating operations
such as application of the body stain.
Types of Emissions
As stated in Table 16-2, annual VOC emissions from wood furniture
coatings are estimated at 230,000 metric tons per year. Individual
furniture factories vary greatly in size, but a moderately large factory
can emit around 1300 kilograms of solvent per day or over 300 metric
tons per year. These emissions are almost totally from coatings used
to finish the furniture. The amount of solvent emissions from a piece
of furniture depends on the amount of each type of coating used as
well as its solvent content, although this will vary, Table 16-6 gives
the approximate relative emissions from different coatings in a coating
system for a specific line of furniture. Conventional coatings often
16-8
-------
TABLE 16-6
RELATIVE EMISSIONS FROM A TYPICAL CONVENTIONAL
FURNITURE COATING SYSTEM
Finish
Stain
Washcoat
Wipe Stain
Sealer
Glaze
Fly Speck
Cowtail
Shade
1st Topcoat
2nd Topcoat
Ibs. VOC/1000 ft2
26
38
24
36
25
0
0
10
29
29
216
Percent
of Total
12
18
11
17
11
0
0
5
13
13
100%
16-9
-------
consist of nitrocellulose resins and organic solvents. The approximate
solids content of furniture coatings are given in Table 16-7.
The stain emissions listed in Table 16-6 are from an organic
solvent-borne stain. There are various types of stains; one of the
most common is organic-borne NCR (nongrain raising) stain. It is
composed of a dye in an organic solvent. it dries rapidly and does
not raise the grain very much. Waterborne stains also exist and have
been used for many years.
More solvent comes from the topcoat than from any other step in
the coating system. Usually two or three coats of topcoat are applied.
More than 85 percent of topcoats now used are nitrocellulose lacquers.
These have been used for over 30 years and are well accepted by the
industry. Nitrocellulose lacquer has many desirable properties of
which one of the most important is the excellent appearance of the
finish.
About 15 percent of topcoat materials are other synthetic organic
resins such as urea formaldehyde and catalyzed urethane finishes.
They are very durable and are used on cheaper furniture and on
institutional furniture where heavy use is expected.
The organic solvent content of coatings is the main interest to
an air pollution control agency. The typical solvent types employed
in wood furniture coatings are:
Acetone
Acetates
Alcohols
Aromatic hydrocarbons
Esthers
Glycol ethers
Ke tones
Mineral spirits
CONTROL TECHNOLOGY
General Discussion
Volatile organic carbon emissions from wood furniture finishing
operations can be reduced by changes in materials or processes, and/or
by the application of add-on emission control devices. Material changes
16-10
-------
TABLE 16-7
APPROXIMATE SOLID CONTENT OF WOOD FURNITURE FINISHES
Finish Percent Solids by Volume
Body Stain 1
Washcoat 8
Filler 40
Sealer 14
Glaze 24
Topcoat 14
16-11
-------
involve reduction in the quantity of VOC components in the coating.
One approach is the use of waterborne coatings in place of organic
solvent-borne coatings. Another alternative involves increasing the
solids/solvent ratio in the coating as applied.
Operational changes are directed towards reducing the quantity
of wasted coating. This is material which does not actually get ap-
plied to the product but, for example, in the case of a sprayed
coating, is exhausted from the booth to the atmosphere. The ratio of
coating adhering to the substrate to the coating used is defined as
"transfer efficiency".
Permanent or add-on controls for VOC emissions in the furniture
industry have not traditionally been used. Incineration would be
the most likely approach and could be considered as technologically
feasible.
Waterborne Coatings
Waterborne coatings include all those coatings in which all or
most of the volatile (or non-solids) portion of the formulation consists
of water instead of organic solvents. Such coating materials have been
manufactured for a number of years and are used on metal products, par-
ticularly electrodeposited prime coats, baked pigmented top coats,
trade sales for painting of wood, and especially for architectural
surfaces such as walls and floors. In the case of wood furniture
manufacturing, coatings have been available for a short period of
time. In 1974, the Reliance Chemical Company displayed waterborne
coated furniture at the Louisville Furniture Supply Fair^'. The
stain, sealer, glaze, and clear topcoat were entirely waterborne
formulations. Other paint companies have also displayed waterborne
coated furniture at Louisville.
Additionally, the U.S. EPA sponsored a display of waterborne
coated wood furniture at the Southern Furniture Manufacturing Associa-
tion show in High Point, North Carolina in 1979. This furniture was
observed by both manufacturers and visitors. There was no definitive
consensus on the visual appearance of the furniture. There were both
positive and negative reactions, as might be expected with such a
subjective rating factor as appearance.
16-12
-------
European companies also produce waterborne coatings for wood
furniture. In 1976 a Swedish company introduced a waterborne system
consisting of stains and lacquers for both light and dark wood sur-
faces 5). This company reports that drying times are less than for
many conventional lacquers.
There are a limited number of manufacturers who are using complete
waterborne coating systems for furniture, including case work and office
furniture manufacturers. Some use waterborne topcoats, while waterborne
body stains have been used in the industry for several decades and are
fairly well established in the furniture industry.
The topcoat is generally considered to be one of the more difficult
coats to formulate as a waterborne coating. In part this is due to the
problem of achieving the desirable gloss, clarity, and surface sheen
with waterborne clear topcoats as compared to the industry standard of
solvent based nitrocellulose lacquers. Also, the furniture industry
produces a wide product mix ranging from simple box-like cases through
open work chairs, to complex and ornate large cabinets and tables. The
lack of a standard product and the subjectivity in judging acceptable
appearance probably are major factors in the slow acceptance of water-
borne coatings"'.
Several recent articles in the trade literature discuss waterborne
coatings for wood furniture'»°>°). These articles recognize that there
are problems, but take an optimistic view of the waterborne coatings
for wood furniture.
Below are some general comments on the characteristics of water-
borne coatings for several properties important in wood furniture
coatings.
1. The appearance of waterborne coatings can differ in some
respects from solvent-borne coatings according to industry
critics. The most important differences relate to gloss and
clarity. Waterborne finishes can have less luster and color
depth. The average person may not see these differences as
acutely as would a furniture specialist.
16-13
-------
2. The repairability of coatings is reportedly not as good with
waterbornes since repair coats do not blend as well with
previous waterborne coats. Roughly 10 to 20 percent of wood
furniture must undergo refinishing at the manufacturing site.
Waterborne coatings can be repaired at the dealer by using
conventional solvent-borne touch-up coatings. Roughly 10 per-
cent of furniture which reaches the dealer has to undergo
repair due to abrasion of finish during shipping.
3. Ease of application is somewhat controversial. It has been
reported that "certain waterborne emulsions can be formulated
to air spray and dry in a manner virtually identical to con-
ventional lacquers"'). Other coating suppliers have stated
that waterborne coatings do not dry as quickly as solvent
coatings. At any rate, dry tme does not appear to be a major
problem. Waterborne stains are currently being applied on
conventional coating lines.
Another problem is the difficulty with which the paint is
applied. Some furniture manufacturers state that in experi-
mental runs, sprayers have more difficulty in applying the
paint evenly and have more rejects. This problem can probably
be corrected as the spray gun operators get more experience
with waterborne systems.
Although waterborne coatings are more sensitive to variations
of humidity and temperature during application than conven-
tional coatings, they are known to be applied to other
substrates under a variety of these conditions. High humidity
is also a problem with conventional solvent coatings due to
solvent "blushing" where the solvent picks up moisture and
causes a clear coating to turn white.
4. Compatibility with previous layers of coating, although re-
portedly not as good as with waterborne coatings, is
acceptable.
5. Mar resistance or "print resistance" (resistance to blemishes
caused by pressure on the coated surface) is somewhat of a
16-14
-------
problem when fast line speeds and minimum oven heat are used.
Some furniture makers store pieces of furniture stacked on top
of each other in the warehouse and are afraid this will be a
problem if waterborne coatings are used.
6. Water and alcohol resistance are acceptable. Although there
is limited field data, waterborne furniture appears to be as
tough as solvent coated furniture in standing up to household
wear.
7. Waterborne topcoats have more difficult polishing or "rubbing"
characteristics than nitrocellulose topcoats. New rubbing
techniques will have to be developed.
In summary, according to furniture makers, the main problems with
waterborne coatings (especially topcoats) are "clarity", "rubbing"
characteristics, and print resistance. Also, training of workers to
apply the coatings may be a short-term problem.
There are several operational aspects relating to the use of
waterborne coatings that should be recognized. The most important are
given below.
It is generally agreed that flash-off areas may have to be extended
to allow proper flash-off of the water. Similarly, curing ovens may
have to be modified to drive the water out of the finish, potentially
increasing energy costs. Additionally, there could be a requirement
for humidity control.
Corrosion of the bulk storage system can occur with waterborne
finishes but can be avoided by replacing lines, pump materials, and
spray gun fittings with stainless steel units and by lining the inside
of storage vessels. This will require additional capital expenditures.
Freezing of the waterborne finishes is also a major concern since most
furniture manufacturers store their conventional solvent-borne finishes
outdoors because of fire hazards. Inside storage of waterborne finishes
may be required. (Fire hazards are reduced with waterborne finishes.)
16-15
-------
Emissions reduction potential by the use of waterborne coatings
can best be expressed as a range. Reductions obtained in the trial
program sponsored by EPA ranged from 26 to 94 percent. Replacing
wash coat and sealer with waterborne substitutes resulted in a 26
percent emission reduction, while a 30 percent reduction was obtained
by converting color coats to waterborne and retaining conventional
clear coats. Totally converted finishes yielded up to 94 percent
emission reduction. Reductions expressed as a percentage are, of
course, sensitive to the original amount of VOC emissions from solvent-
borne systems. In the case of the EPA trials emission levels from the
standard systems used ranged from about 85 kg per 100 m^ of surface
covered to 160 kg per 100 m^.
Electrostatic Coating Application
The conventional method for applying surface coatings to wood
furniture is with an air atomizing spray gun. In most cases these guns
are hand-held. The transfer efficiency with these systems is usually
less than 50 percent. It can be greater with large solid objects and
can be considerably less with smaller items such as open work chairs,
chair legs, spindles, and small panels. This poor transfer efficiency
is the result of both overspray and bounceback.
Electrostatic coating is based upon the principle that opposite
electrical charges attract each other. Since the atomized coating
particles are negatively charged by the application device, the ar-
ticle to be coated must be positively charged or grounded. Transfer
efficiencies with electrostatic equipment typically can be improved
to 65 percent, and under optimal conditions, using automated bell and
disk systems together with conveyorized parts, can reach 85 to 95
percent-^).
Several types of atomizing devices are used with electrostatic
application systems. These include rotating disk and bell units where
atomization is produced by centrifugal forces, air atomizing hand guns,
and airless (hydraulic) spray guns. Disk or bell atomizers are most
often used in automated systems where the atomizer is mounted on a
reciprocater which moves up and down. The parts are held on a conveyor
16-16
-------
which transports them past the atomizer in a loop pattern. As the
parts pass along the loop they are rotated or indexed to maximize
exposure to the spray pattern.
As mentioned earlier, the objects to be electrostatically coated
must be positively charged or grounded. Because wood (particularly
when dry) is not conductive, it must be treated with a sensitizing
material. These conductive chemicals are applied in a very thin layer
by spraying, dipping, or flow coating. The coating is transparent and
does not interfere with or detract from subsequent coating steps.
Following application of the sensitizer, the object to be coated must
be passed through a humidification booth or similar chamber where it is
exposed to an atmosphere of about 50 percent relative humidity for
approximately one minute-'--'-'. Following this treatment objects to be
coated can easily be grounded if they are transported through the
coating process on a metal conveyor. In the situation where objects
are pallet-mounted (generally larger case goods) individually devised
grounding techniques must be utilized.
A fairly modest improvement in transfer efficiency can result in
significant reduction in VOC emissions. As an example, an improvement
from 40 to 60 percent in transfer efficiency is equivalent to nearly 47
percent reduction in VOC emissions.
Add-On controls (Incineration)
Incinerators have been rather widely used control VOC emissions at
the point of discharge to atmosphere. In most such cases, however,
they have been used in conjunction with unmanned operations such as
gasoline storage and transfer, asphalt air blowing, can and coil
coating, automotive surface coating, and resin cooking where VOC con-
centrations are fairly high. In the case of fume incinerators serving
baking or drying ovens, heat recovery has been successfully applied.
Considering the cost of fuel, other opportunities for heat recovery
are constantly being explored. In principle, fume incineration should
be technologically feasible for destruction of VOC emissions from
furniture coating as the actual design of hooding, ducting, and
incinerators does not require new design principles.
16-17
-------
At the present tine there are no known fume incinerators used in
conjunction with wood furniture coating in the United States. Cost,
absence of regulations, and safety considerations have a bearing in
this situation. A substantial portion of the VOC emissions from wood
furniture coating arises from manual spraying operations where ventila-
tion rates must meet OSHA standards. This generally results in rather
dilute VOC concentrations in the spray booth exhaust and therefore a
relatively larger volume of exhaust air per unit of VOC emissions than
in many systems where fume incinerators are used. Also, there can be
relatively large nuiaers of potential emissions points per plant because
of the multiple operations performed. While these sources could be
manifolded together, there has been concern in the industry about the
fire risk from overspray deposits of highly flammable nitrocellulose
resins in long exhaust ducts.
It is possible that incinerators could be selectively applied
as part of an overall control strategy using some of the following
criteria:
1. Apply incinerators only to fairly high concentration streams
such as drying ovens, flash-off areas, or automated booths.
2. Avoid nitrocellulose buildup problems by either (a) restrict-
ing application to non-spray booth uses, or (b) removing
particulate nitrocellulose with filters or water wash prior
to entry into ductwork.
3. Use fire control systems such as water deluge, inert gas
blanketing, etc.
Where incinerators are used, control efficiencies of at least 90
percent can be attained on VOC emissions reaching the incinerator.
STATUS OF REGULATORY ACTIVITIES
Existing rules and regulations at the state and local level are
all generic rather than being directed specifically at surface coating
of wood furniture. An example is Rule 442 of the South Coast Air
Quality Management District, California, covering the usage of solvents.
16-18
-------
This rule generally restricts solvent emissions by requiring an 85
percent reduction in emissions or by limiting daily and/or hourly emis-
sion rates. The general limitation for emission of non-photochemically
reactive solvents is 272 kilograms (600 pounds) per day, effective
December 1, 1980. Organic materials which come into contact with flame
or are baked, heat cured, or polymerized are limited to 6.5 kilograms
(14.3 pounds) per day. Emissions of photochemically reactive solvents
are limited to 14 kilograms (3.1 pounds) per hour, not to exceed 18
kilograms (39.6 pounds) per day. Information obtained from a District
spokesman indicates that none of the wood furniture coating operations
within their jurisdiction have elected to use add-on controls to achieve
compliance with the rule^-2). Instead, they operate within the hourly
and daily limits on emissions by limiting solvent use and by solvent
substitution (for photochemically reactive solvents originally used).
At least one plant manufacturing chairs has made a partial conversion
to waterborne coatings. Under the agenda of the California Suggested
Control Measures Committee, the South Coast Air Quality Management
District is taking the lead to develop a Model Rule for the Surface
Coating of Wood Furniture. This rule may be considered during the
year 1981. At the present time emphasis is on developing a rule to
require improved transfer efficiency such as with airless or electro-
static spray equipmnt. However, a decision on the exact content or
numerical standards has not yet been reached^).
A major consideration will be the form of an industry specific
rule. Most rules applying to other industries have been based upon an
equivalent limit on organic solvents per quantity of coating (less
water) as applied. This is a possibility for the wood furniture coating
industry, but there is great variability in the number of coatings
applied and the solvent content of the these coatings depending upon
the type of product and the manufacturer. A more uniform limit could
be based upon the amount of VOC emissions emitted per unit area coated,
such as kilograms VOC per 100 m^ of surface coated. This latter approach
would probably involve a more involved compliance checking procedure.
16-19
-------
REFERENCES
1. Control Techniques for Volatile Organic Emissions from Stationary
Sources, U.S. Environmental Protection Agency, May 1978. EPA-450/2-
78-002.
2. 1972 Census of Manufacturers, Wood Household Furniture, U.S.
Department of Commerce, February 1974.
3. Control Technique Guidelines for the Control of Volatile Organic
Emissions from Wood Furniture Coating, Draft, U.S. Environmental
Protection Agency, OAQPS, ESED, RTF, North Carolina, April 1979.
4. Connelly, Herbert H., "What's New in Furniture Finishing?," Furni-
ture Design and Manufacturing, April 1976, page 14.
5. "Water or Solvent in Wood Stains and Lacquers," Pigment and Resin
Technology, March 1978, page 20.
6. Personal conversation between R. J. Bryan, Engineering-Science,
Inc., and Merrill Evans, Lilly Industrial Coatings, Montebello,
California, Dec. 4, 1980.
7. "Waterborne Furniture Coatings," Finishing Highlights, Jan./Feb.
1978, page 22.
8. Ocko, B., "Fast Air Dry Acqueous Systems Answer End-Use Concerns,"
Modern Pain and Coatings, March 1977.
9. "An Exception to the Rule," Wood and Wood Products, Feb. 1980.
10. "Electrostatic — A Dark Horse in Finishing Alternatives," Wood and
Wood Products, Feb. 1980, page 25.
11. Information from Randsburg Electrostatic Equipment, Division of
Randsburg Corp., Indianapolis, Indiana.
12. Personal communication between R. J. Bryan, Engineering-Science,
Inc., and A. Wilson, South Coast Air Quality Management District,
Dec. 1, 1980.
13. Personal communication between R. J. Bryan, Engineering-Science,
Inc., and A. Rawulka, South Coast Air Quality Management District,
March 16, 1981.
16-20
-------
SECTION 17
WASTE SOLVENT RECOVERY INDUSTRY
Organic solvents used by industries for extractions, for cleaning,
and as chemical mediums or intermediates are available as halogenated,
aliphatic, and aromatic hydrocarbons; as alcohols, esters, glycol
ethers, ketones, and nitroparaffins; and as miscellaneous compounds
such as tetrahydrofuran. These classifications include methyl ethyl
ketone, benzene, perchloroethylene, and isopropanol.
A solvent not consumed during industrial use usually becomes
contaminated and unacceptable for further use. A reclaimed solvent can
be reused for its original purpose or for different industrial needs.
Reclaiming waste solvents has gained importance because the costs of
these petroleum derivatives are dependent on the cost of crude oil.
Furthermore, the rising costs of virgin solvents and waste solvent
disposal have provided incentives for industries to recover solvents
for reuse.*•'
PROCESSES AND EMISSIONS
The waste solvent recovery industry is composed primarily of small
plants using straightforward chemical-processing technology. Typical-
ly, a single plant has an analytical laboratory, a distillation device,
and storage tanks. The usual flow through the plant follows the
diagram in Figure 17-1: the solvent-bearing material entering the plant
is analyzed in the lab, pumped into cone-bottom separation tanks where
heavy solids and water settle out, and put through a flash distillation
device which is typically a thin-film evaporator. For most plants,
this is the end of the recovery process.
The composition of materials handled varies widely in vapor
pressure and type of solvent. A list of solvents most frequently
recovered appears in Table 17-1. This clean solvent-blend product
usually goes to storage for later sale or for return to the contracting
company, and the sludge is disposed of by sending it either to an
incinerator or to a landfill.
17-1
-------
FIGURE 17-1
SIMPLIFIED FLOW PATTERN THROUGH TYPICAL SOLVENT RECOVERY PLANT
17-2
-------
TABLE 17-1
TYPICAL SOLVENTS RECOVERED
Solvent
Vapor Pressure PSIA @ 68°F
Halogenated Hydrocarbons
Carbon tetrachloride
Fluorocarbons
Methylene chloride
Perchloroethylene
Trichloroethylene
1,1,1-Trichloroethane
Hydrocarbons
Hexane
Benzene
Toluene
Xylene
Cyclohexane
Ethers
Mineral spirits
Napthas
Ketones
Acetone
Methyl ethyl ketone
Methyl isobutyl ketone
Cyclohexanone
Alcohols
Methanol
Ethanol
Isopropanol
Butyl alcohol
Ainyl alcohol
Esters
Amyl, butyl, ethyl acetates
1.5
6.8
0.2
0.9
1.6
1.9
4.4
0.3
0.1
1.5
7.0
1.3
2.9
1.2
0.2
9.4
1.5
0.9
0.5
0.2
17-3
-------
Sone large plants are capable of separating the clean solvent
blend into two or more components, usually by using a fractionating
column with a center-feed and with overhead and bottom-product takeoffs.
Plants with high-technology recyclers have large multistage columns
capable of reclaiming solvents with a high degree of purity, but the
throughput of these plants is usually limited to a specific industry.
Storage tanks are the largest source of emissions. Typically,
plants have a relatively large number of small tanks, ranging from
1000 to 20,000 gallons; many have a few tanks outside this range.
Most tanks are fixed roof, straight sided, and erected vertically
above ground; however, some horizontal tanks are in use primarily as
receivers.
Emissions from storage tanks are from working losses and breathing
losses. Working loss emissions come from filling an empty tank with
material that pushes the organic vapors within the tank out through
vents or open hatches to the ambient air. If a residue remains in
a tank after previous unloading, concentrations of VOC's are often near
saturation. Breathing loss emissions come from an inactive partially
filled tank being exposed to temperature differences between day and
night; as the tank heats up, inactivity allows the material to saturate
the air with vapor, so the saturated air expands and vents to the
ambient air.2)
Other points of emissions are fugitive sources such as pump seals,
valves, flanges, and spills. Spills occur when connections are un-
coupled and opened because residual material remains at the coupling.
Other spills are due to line ruptures, tank overflowing, and plugged-
line repairs.
The amount of emissions from waste solvent recovery plants is
small. Emissions for a large plant with an annual throughput of 2.5
million gallons are less than 60 tons per year. Emission factors for
solvent-recycling facilities are in Table 17-2.
CONTROL TECHNOLOGY AND ASSOCIATED COSTS
Few control options are available to recyclers. For control of
emissions from storage tanks, an internal floating roof with primary
17-4
-------
CN
I
w
pa
H
Z
M
2
h- 1
rl
0
S
H
Z
u
0
cn
o
cn
&
O
H
QJ
• CN
rH CO rH <5 •<
. • 0 Z Z
0 O 0
o
o
X~V
CN
-^
•
rH
1
O CN -*
CN r^. CN <; <;
• • O Z Z
o o o
o
•
o
\^/
cn cn cn cn
a o o o
•H -H -H -H
d d c c
cfl cfl CO cfl
60 00 OO OO
J-i j_i j^ j_i
o o o o
cu cu cu cu
i— 1 rH rH i— I
•H -H -H -H
4-14-1 4-14-1
Cfl CO CO Cfl
rH rH rH rH
O O O O
*> ^* ^> P*
cn
01
a
s_i
CU 3
60 00 O
cfl C cn
rH -H Cn
rH -O ^ C
•H CO Cfl CU
Cu O 0) O-
cn >«4 >J o
UH
o
cu
o
•H
>4H
O
o
rH
1
rH
CO
4-1
C
cu
e
CU
i-H
G.
0,
3
CO
00
d
•tH
-O
3
rH
o
C
•H
*i
d
o
•iH
4-1
•H
*rj
CU
•a
(-4
CO
M
cn
j-i
0
4-1
O
cfl
fa
d
O
•H
CO
CO
•H
£
P-
4-1
C
CO
4-1
3
i — i
rH
O
P-l
S-I
•r-l
4n
O
C
0) O
[ ,— J
r^ *n
r*i 4-1
CO cfl
rH rH
•H -H
Cfl D,
> 6
^ cS
4-1
0
Z
cu
1 0
S-4
< 3
Z 0
-jc cn
4-1
O
z
M
^
S-i
cfl
P-i
cu
rH
c°
CO
•H
H
f.
O
^_j
CO
CU
CO
0)
M
cn
•a
cfl
•o
d
CO
4-1
CO
-a
d
CO
60
d
•iH
d
d
cfl
rH
P-l
^
iH
rH
CO
3
o-
j_i
•H
<£
U-l
O
0)
o
•H
MH •
I ^ j ^*
O r~>
« rH
4-1
d *->
01 CO
e 3
cu to
00 3
CO <
d
S fN
CU 1
4-1 O-<
cn
-------
seals is available, but it would require fairly expensive retrofits
of all existing vertical-side tanks. Capital costs would start at
$24,000 per tank, and increase with size. At $1.2 million per inegagram
($1.1 million per ton) of emissions-*), this option would not be cost
effective.
Emissions from fugitive sources are best controlled by monitoring
and maintenance. This option would require two workers to go through
the plant periodically with a portable hydrocarbon detection device to
determine if there are leaks, and if so, where and how much. At a
designated cut-off level, the offending component would either be
repaired or replaced. Although not as expensive as the tank retrofit,
there would be large expenditures required for two detectors, one for
use and one spare; two workers for proper monitoring; an inventory of
spare parts and replacement equipment; and an allotment of labor time
for indicated repairs. The overall cost effectiveness would be about
$2000 per megagram ($1818 per ton) of emissions.^)
REGULATORY STATUS
Currently, no air pollution regulations exist that specifically
pertain to the waste solvent recovery industry. There are many general
VOC emission rules that may apply to specific processes.
REFERENCES
1. Source Assessment: Reclaiming of Waste Solvents, State-of-the-Art,
Environmental Protection Technology Series, Industrial Environmen-
tal Research Laboratory, Cincinnati, Ohio, EPA-600/2-78-004f,
April 1978.
2. Compilation of Air Pollutant Emission Factors, 3rd edition, Inclu-
ding supplements 1-10, Office of Air and Waste Management, Office
of Air Quality Planning and Standards, Research Triangle Park,
North Carolina, AP-42, August 1977.
3. Personal correspondence to Lawrence Lloyd (ES) from Rebecca Sommer
(GCA).
4. Synthetic Organic Chemical Manufacturing Industry, draft chapters
2, 3, and 5 from Fred Porter of U.S. Environmental Protection
Agency, Environmental Standards and Engineering Division, August 28
and 29, and October 28, 1980.
17-6
-------
SECTION 18
WINE MAKING
PROCESSES AND EMISSIONS
Wine is made by the controlled fermentation of the juice of certain
fruits, chiefly grapes. During the fermentation the natural sugar
content of the fruit is converted into ethanol by the reaction:
C6H12°2 2C2H5OH + 2C02
(sugar) (ethanol)
This process takes place in the presence of specially cultured yeast.
The final alcohol content of natural wines ranges from about 7 to
14 percent, depending upon the original sugar content of the grapes. A
typical sugar content is about 20 percent. Theoretically, the yield of
ethanol should be 51.1 percent by weight of the initial sugar content.
The actual yield is found to be around 47 percent. The remaining sugar
is lost as alcohol or byproducts of complex chemical mechanisms, or it
remains in the wine as the result of incomplete fermentation.
While the basic elements of wine making appear to be relatively
simple and straightforward, there are significant numbers of variables
which are closely controlled by modern wineries, depending upon the
type of wine being produced. A number of these variables influence the
quantity of alcohol which could possibly be lost to the atmosphere.
Among the most important are:
(1) Initial Sugar Content of Fermenting Material - As seen from
the chemical equation for alcohol fermentation from sugar,
the theoretical alcohol content of the wine varies directly
with the amount of sugar present. Furthermore, the amount of
gaseous carbon dioxide formed also increases with sugar
content. Increases in alcohol content and rate of carbon
dioxide evolution both increase the potential for alcohol
loss to atmosphere.
18-1
-------
(2) Nature of Fermentation Material - Red wine is fermented from
crushed black or red grapes where of all the solid material,
only a portion of the stems are removed. This mixture of
skins, pulp, other solids, and juice is termed "must". White
wine is fermented from juice extracted from crushed or
macerated grapes. Limited testing by wine manufacturers and
research organizations have shown slightly higher alcohol
losses from must as compared to juice fermentation-"-). the
reason for this is not known.
(3) Temperature of Fermentation - Wine is fermented at tempera-
tures ranging from about 11°C to 30°C. White wines are
fermented at the lower temperature ranges, while red wines
are fermented at higher temperatures. Heat is generated
during the fermentation which must be removed through use of
cooling coils. Both laboratory and wine cellar fermentation
tests have shown higher alcohol losses at higher fermenta-
tion temperatures. Figures 18-1 and 18-2 demonstrate this
relationship for controlled laboratory fermentations.
(4) Agitation - During fermentation the fermenting mixture is
internally mixed or air blown under certain circumstances.
Limited testing shows instantaneous increases in alcohol los
during and immediately following such agitation-"-).
When the fermentation is complete, the wine goes through a finish-
ing process for clarification. Common clarification procedures are
filtration, fining refrigeration, pasteurization, and aging. The wine
is then bottled, corked or capped, labeled, and cased. Little is
known about the loss of alcohol during the post-fermentation stages of
wine production.
Emission estimates of typical alcohol losses during fermentation
have been made as a result of limited laboratory scale and field testing.
These losses have been expressed in terms of the fraction of the total
theoretical alcohol available and as an emission factor in terms of
18-2
-------
FIGURES 18-1 & 18-2
FIGURE 18-1
RATE Of LOSS
—»— REDUCING SUCJR
20 JO «0 50 6070"
HOURS
Rate of loss of alcohol entrained
in carbon dioxide during stirred laboratory
scale fermentation with temperature main-
tained at 34°C.
FIGURE 18-2
0 lo !0 30 «0 50 60 70 80 90 100 »0 UO 130
KOUBS
Rate of loss of alcohol entrained
in carbon dioxide during stirred laboratory
scale fermentation with temperature main-
tained at 2 I °C.
18-3
ENGINEERING-SCIENCE
-------
weight of alcohol lost per unit volume of fermenting material. Zimmer-
man, et all), reports alcohol losses ranging from about 0.1 to 1.5
percent of total alcohol available. The smaller losses are associated
with low fermentation temperatures (5 to 10°C) and a lower range of
initial sugar content (10 to 20 percent by weight of fermenting
material).
Emission factors reported in AP-42^) are given in terms of grams
of alcohol lost per kiloliter of fermentation material and lb./l,000
gals. Typical losses for white wine are given as 125 g/kl (1.06
lb./l,000 gals.) at 11.1°C (52°F) to 574 g/kl (4.79 lb./l,000 gals.)
at 26.7°C (80°F) and 20 percent initial sugar content. For red wine
fermented from must, a supplemental factor of 288 g/kl (2.4 lbs./l,000
gals.) is added to the higher temperature figure. In addition, an
empirical equation is given to calculate emission factors at other
conditions. This equation is:
EF = [0.136T - 5.91] + [(B - 20.4) (T - 15.21) (0.00685)] + [C]
where
EF = emission factor, pounds of ethanol lost per 1,000 gals.
of wine made
T = fermentation temperature, °F
B = initial sugar content, °Brix (percent by weight)
C = correction term, 0 (zero) for white wine or 2.4 lbs./l,000
gals, for red wine
A loss of 8.9 lbs./l,000 gals, is equivalent to a loss of 1 percent of
total alcohol available. Thus the loss levels given by Zimmerman in
percent terms are reasonably consistent with the emission factors given
in AP-42.
CONTROL TECHNOLOGY
There are no regularly used commercial control devices applied in
the wine industry to reduce alcohol losses as these losses are generally
18-4
-------
under 1 percent of the total alcohol produced, and presumably there has
not been a sufficient economic incentive to do so. There has been some
experimentation with control devices including scrubbers. No efficiency
data were reported but alcohol recovery by the control apparatus in two
different studies-^»^' was found to be about 1 percent of the total
alcohol produced. Since this is equivalent to the higher loss levels
reported, it would appear that the control devices were reasonably
efficient.
Conceivably, refrigeration, adsorption, or absorption (scrubber)
type control devices could be used to capture alcohol emissions. These
devices, to be fully effective, would have to provide a means of re-
covering and utilizing the alcohol collected. Further, the alcohol
loss shows a rather sharp peak rate at about the middle of the fermenta-
tion period and is associated with significant carbon dioxide evolution.
This would present a problem in design capacity of any proposed control
system. If designed for peak rate of emissions, the capacity might be
utilized for only a relatively small portion of the total fermentation
period. Because there are no commercially used controls, control costs
cannot be directly estimated. However, the scale of any such equipment
can be roughly estimated. Because most of the alcohol losses are
associated with the active evolution of carbon dioxide during fermenta-
tion, any control equipment would have to be sized to handle the
volumetric flow rate C02» The fermentation reaction shows that there
is about 49 Ibs. of C02 evolved per 100 Ibs. of sugar used. This is
o
equivalent to roughly 7,500 ft of CC^ evolved per 1,000 gals, of
fermenting material (20 percent initial sugar content). If the fer-
mentation takes 250 hours, the C02 is evolved at an average rate
of 30 ft^ per hour. As seen by the laboratory experiments, the peak
rate may be several times the average rate. For a peak rate three
tmes the average rate and a 10,000 gal. fermentation tank, the peak
3
C02 evolution rate would be about 15 ft per minute. The ethanol
content would be about 1 percent of this flow. Compared to many
industrial pollution sources, this is a very low volumetric rate of
gas discharge.
18-5
-------
STATUS OF REGULATORY ACTIVITIES
Currently there are no regulations covering the discharge of
ethanol to the atmosphere from wine making. A joint study currently
is being conducted of possible ethanol losses from wineries by the
Fresno County, California, Air Pollution Control District and the
California Air Resources Board (GARB). Additional source testing and
evaluation of control technologies are covered in the study plan.
Depending upon the results of the study, it is possible that the GARB
will propose a model rule.
REFERENCES
1. "Alcohol Losses from Entrainment in Carbon Dioxide Evolved During
Fermentation," H. W. Zimmerman, E. A. Rossi, Jr., and E. Wick,
American Journal of Enology, JL5:63-68, 1964.
2. "Compilation of Air Pollutant Emission Factors, AP-42," Supplement
10, 6.5.2-1, U.S. Environmental Protection Agency, February 1980.
3. "The Recovery of Alcohol Carried Away by the Carbonic Gas During
the Alcoholic Fermentation and Refrigeration of the Vintage," E.
Negre and M. Marichal, Prog. Agr. et Vitic., 133:250-257, 281-291,
307-313, 358-361, Montpellier, France, 1950.
4. "Alcohol Losses During Fermentation of Grape Juice in Closed
Fermenters," H. Warkentin and M. S. Nury, Am. J. Ecol. Vitic.,
14:68-74, 1963.
18-6
-------
SECTION 19
STYRENE-BUTADIENE COPOLYMER LATEX
PROCESSES AND EMISSIONS
This chapter discusses the processes used in the manufacture of
styrene-butadiene latex and the sources of volatile organic compound (VOC)
emissions from these processes. The applicable emission control techniques
are discussed, and a review of the air pollution regulations applicable to
styrene-butadiene latex manufacturing is presented. Styrene-butadiene latex
is defined as any latex polymer in which styrene and butadiene together
comprise more than 85 percent by weight of the reacted monomer content, and
butadiene alone comprises more than 15 percent by weight of the reacted
monomer content. This definition would not cover polymers such as ABS
(acrylonitrile-butadiene-styrene), in which all three monomers are present as
a significant fraction of the total, but would cover latexes of styrene and
butadiene which are modified with small amounts of vinylpyridine, acrylic
acids, or other monomers.
Copolymers of styrene and butadiene can be made by a solution polymerization
process or by an emulsion polymerization process. The solution polymerization
products are sold as solid thermoplastic elastomers. The emulsion polymeri-
zation products can be sold in either a solid form, known as crumb, or in a
liquid form, known as latex. The crumb product is used extensively in the
manufacture of rubber tires, and has an average styrene content of 23.5 percent
by weight. The latex products have a wide variety of uses, depending on the
styrene content of a particular grade. The rubbery types of approximately
23 percent styrene are used for dipped goods, carpet underlay, adhesives, and
moldings. Small quantities of vinylpyridine monomer may be added if the latex
is to be used for dipping tire cords. The high styrene latexes, ranging from
46 to 85 percent styrene, are used for paper coatings, paints, carpet backsizing,
and adhesives. Small quantities of a carboxylic monomer, such as acrylic
acid, are sometimes added to provide a "self-curing" material for use in
carpet backsizing. Table 19-1 lists the U.S. producers of styrene-butadiene
latex, the plant locations, and the capacities.
19-1
-------
TABLE 19-1. DOMESTIC PRODUCERS OF STYRENE-BUTADIENE LATEX
2,3
Company
Plant
location
Annual
capacity
(103 Mg)
Dow Chemical Company
Bayport, TX
Dal ton, GA
Gales Ferry, CT
Midland, MI
Pittsburg, CA
20
Firestone Plastics Company
General Tire and Rubber Company
BF Goodrich Company
Goodyear Tire and Rubber Company
W.R. Grace and Company
Reichhold Chemicals, Inc.
Rubber Research Elastomers
Polysar Resins, Inc.
Union Oil Company of California
U.S. Steel Corporation
Pottstown, PA
Mogadore, OH
Louisville, KY
Calhoun, GA
Owensboro, KY
South Acton, MA
Cheswold, DE
Kensington, GA
Minneapolis, MN
Chattanooga, TN
Chattanooga, TN
Beaver Valley, PA
Charlotte, NC
La Mirada, CA
Scotts Bluff, LA
4
47
NAa
NAa
7
7
20
25
NAa
30
30
20
8
9
NAa
aNA = not available.
19-2
-------
Process
As shown in the general flow diagram depicted in Figure 19-1, fresh
styrene and butadiene are piped separately to the manufacturing plant from the
storage area. The butadiene stream is passed through a caustic soda scrubber
to remove any inhibitors that have been added to prevent premature polymerization
during shipment and storage. Soap solution, activator, catalyst, and modifier
are added to the monomer mixture prior to entering the polymerization reactor
train. The soap solution is generally a mixture of a rosin acid soap and a
fatty acid soap used to maintain the monomers in an aqueous emulsion state.
The activator is usually a hydroperoxide or a peroxysulfate which initiates
the polymerization reaction by supplying free radicals. The catalyst assists
in generating the free radicals more rapidly and at lower temperatures than is
possible with thermal decomposition of the activator alone. The modifier is
an additive used to adjust the chain length and molecular weight distribution
of the product during polymerization.
Most emulsion latex polymerization is done in a batch process, rather
than continuously. The batch reaction is normally carried out at 50 C
(122 F) , and is taken essentially to completion (97 to 99 percent conversion).
As a result, the recovery of unreacted monomers is not economical and the
process is directed towards maximum conversion on a once-through basis. Some
latex manufacturers (approximately 15 percent ) use the "cold" polymerization
process, which operates at 4 C (40 F). The degree of conversion of monomers
to copolymer for the cold process is less than for the hot process, so recovery
and recycle of unreacted monomers is used in cold process latex plants. Also,
some crumb plants produce small amounts of latex as an end product using the
cold polymerization process. The discussion that follows is for the more
predominant hot process.
Completion of the polymerization process requires from 6 to 24 hours.
The latex is sent to a blowdown tank where, under an absolute pressure of
6.5 kPa (28 in. of mercury vacuum) and steam agitation, any unreacted butadiene
and some unreacted styrene are removed from the latex. The overhead stream
from the blowdown tank is sent to a water-cooled condenser where any condensibles
are removed from the vapor stream and sent to a wastewater treatment facility.
Noncondensibles from the condenser are discharged to the atmosphere.
19-3
-------
fO
X t-J
CU T-
+-> S-
ro 01
o o
M- O.
n3 O
S- •!-
CD in
re i—
•I- 3
-O E
CD
i— .a
o o
re o
E 3
-------
After discharge from the blowdown tank the latex is stored in process
tanks where additives are introduced based upon product specifications.
Starting from this point in the manufacturing process to the final product
storage for shipment, latex is processed on a continuous basis.
Subsequently, the latex is screened using shaker screens to remove any
large, agglomerated solids present in the latex. If the unreacted styrene
content of the latex has not been reduced sufficiently to meet product
specifications in the blowdown step, the latex is introduced to a series of
steam-stripping steps to reduce the unreacted styrene content. Any steam and
styrene vapor from these stripping steps is taken overhead and sent to a
water-cooled condenser. Any noncondensibles leaving the condenser are verted
to the atmosphere. The liquid stream from the condenser is discharged to the
wastewater treatment system.
The stripped latex is then passed through a series of screen filters to
remove unwanted large solids and is stored in blending tanks where antioxidants
are mixed with the latex. Finally, latex is pumped from the blending tanks to
be packaged into drums or bulk loaded into railcars or tank trucks.
Emissions and Emission Sources
The operating parameters of a model latex plant are presented in Table 19-2.
Based on source sampling data and industry supplied information, the VOC
emissions from the model plant sources were developed. These data represent
the expected annual emissions from the model plant in the absence of any
add-on air pollution control equipment. Process control devices which are
routinely applied for reasons of material recovery or plant safety are assumed
present in the model plant. The use of these control devices is reflected in
the emissions listed in Table 19-2.
TABLE 19-2. EMULSION LATEX MODEL PLANT3
VOC sources
Monomer removal - butadiene (1)
Monomer removal - styrene (1)
Blend tanks (15)
Emissions
(Mg/yr)
224
4
3 total
(0.2 /tank)
Vent flow
(scfm)
175
30
275 total
(18/tank)
Concentration
(ppm)
37,000
2,000
150
a
Capacity of model plant is 30,000 Mg/yr and production rate is 27,000 Mg/yr.
19-5
-------
Table 19-2 also reflects the composition of the copolymer product and the
degree of conversion of monomers to polymer. The model latex plant is based
on the use of a hot polymerization process with a 98.5 percent conversion of
monomers to polymer, and a weight ratio of 46 percent styrene and 54 percent
butadiene. Emissions will vary greatly in this type of plant depending upon
the percent conversion, because the unreacted butadiene is vented without any
reduction due to recovery operations. Emissions will also vary for different
ratios of styrene and butadiene. Latex operations with less butadiene and
more styrene will have lower emission rates because butadiene is more easily
volatilized than styrene. The following factor can be used to estimate total
emissions from plants producing latexes with different butadiene contents and
percent conversions:
Emission factor = , 100
percent conversion
( ~ : 1) (933 X. + 67) g VOC
percent conversion D f-
kg net copolymer
where )L = weight fraction of butadiene in net copolymer.
"Net copolymer" means the reacted monomer content of the latex. The percent
conversion to be used in this expression should be the percent of butadiene
only fed to the reactor which reacts to form part of the polymer mass.
Table 19-2 lists the three major VOC emission sources for the emulsion
latex model plant. Emissions from monomer storage tanks and reactor vents are
not quantified. The reactors are normally sealed using rupture discs, with
venting to a flare occurring only under extreme upset conditions within the
reactor.
Monomer removal produces two vent streams in the model latex plant.
Vented emissions from the blowdown process step consist mainly of unreacted
butadiene, though some unreacted styrene is also emitted during this step.
Because this portion of the manufacturing process is a batch operation, the
venting of emissions is intermittent. Emissions from this operation account
for approximately 97 percent of the VOC emissions from the emulsion latex
process.
After the blowdown step, the latex is stored in process tanks. Product
additives may be added in these tanks. The tanks also serve as flow-regulating
holding tanks, because the remaining process operations are run on a continuous
basis. The latex then flows through shaker screens to remove any large agglomerated
19-6
-------
solids. These screens are open to the atmosphere, but emissions are estimated
to be very small. Concentrations of 42 ppm butadiene and 44 ppm styrene have
been measured above the screens, but the air flow rate could not be determined.
If the residual styrene content of the latex is greater than product
specifications, the latex is processed in steam-stripping columns. The
overhead styrene and water vapor stream is condensed in water-cooled condensers,
and the noncondensibles are vented to the atmosphere. Table 19-2 shows that
this noncondensible stream carries approximately 1.7 percent of the latex
plant's VOC emissions with it. The stripped latex is then held in blend
tanks, where antioxidants may be added. Evaporative emissions from the vents
on the blend tanks are estimated to contribute another 1.3 percent of the plant's
VOC emissions. This small amount of VOC would be emitted from 15 tanks in the
model plant. The natural ventilation rates and low mass of VOC produce a
stream of only 150 ppm organics.
CONTROL TECHNOLOGY
This section describes the emission control techniques available for
control of VOC emissions in the styrene-butadiene latex manufacturing industry.
The air pollution control techniques discussed are condensation and thermal
and catalytic incineration. The costs of controlling the model latex plant with
a thermal incineration system are also presented.
Condensers
Vapor condensation involves the removal of sufficient heat to liquify one
or more of the vapors contained in the exhaust stream. The most common type
of condenser for this application is a surface condenser consisting of a shell
and tube heat exchanger. A coolant liquid is passed through the tubes of the
exchanger while the gas stream passes on the shell side.
The primary disadvantage of condensation is its limited efficiency. The
emissions reduction rate of a condenser is a function of the inlet VOC vapor
concentration, because the outlet concentration is fixed by the temperature,
pressure, and specific organic compound present. For example, an air stream
at 4 C (40 F) and standard atmospheric pressure must contain over 2,800 ppm
styrene before any styrene condensation will occur. However, condensers make
19-7
-------
excellent control devices for saturated streams (e.g., reduction of a stream
containing 90 percent styrene to 6,000 ppm equals a removal efficiency of
99.3 percent). Therefore, condensers are only applicable to streams with high
VOC content.
Currently, condensers are in use as process devices in most domestic
latex plants. These units are an integral part of the manufacturing process
and are primarily used for styrene vapor recovery to maintain desirable
process economics, rather than as air pollution control devices. These
condensers still provide a major reduction in VOC emissions, because the
recovered styrene would otherwise be released to the environment.
Incineration
Incineration is the oxidation (combustion) of organic vapors present in
a process vent stream. The factors governing the completeness of this reaction
are time, temperature, turbulence, and the type and concentration of vapor
present. Combustion air must mix thoroughly with the vapor at sufficient
temperatures, usually 760 to 871°C (1400 to 1600°F), and for a sufficient
period of time (0.5 to 0.75 seconds) to complete the combustion reaction.
Combustion of vapors resulting from latex production results in the formation
of carbon dioxide, water, carbon monoxide, and particulate.
In most applications it is necessary to provide auxiliary fuel to increase
the heat content of the gas stream such that an adequate combustion temperature
is reached. When sufficient VOC is present in the gas stream the combustion
reaction will be self-sustaining, requiring no auxiliary fuel.
Plant insurance and safety regulations require sufficient margins of
safety to ensure that the VOC vapor concentration of the gas stream remains
above or below the explosive range of that vapor. This can be accomplished by
keeping the stream above the explosive range by ensuring that the quantity of
oxygen introduced into the gas stream (either through process requirements or
inadvertent air leakage) does not create an explosive mixture or by keeping
the stream below the explosive range by diluting it with ambient air. Most
insurance companies require that the gas stream remain below 25 percent of the
lower explosive limit in the absence of continuous VOC monitors. Concentrations
as high as 50 percent of the lower explosive limit are allowed if a continuous
monitor and alarm system are included in the design.
Two types of incinerators are discussed in this section, thermal (direct
flame) and catalytic.
19-8
-------
Thermal or Direct Flame Incineration
The basic thermal incinerator consists of an insulated combustion chamber
fed by the burner(s) for the vapor stream and auxiliary fuel (if necessary)
and a combustion air source. The vapor-laden stream is transported by a
process blower through a preheat burner to raise the stream to a temperature
suitable for combustion in the combustion chamber. Based on available data,
thermal incinerators can reduce VOC emissions by 98 percent, if the stream
contains more than 1,000 ppm organics. For more dilute streams the reduction
efficiency is governed by the fact that approximately 20 ppm is the minimum
outlet concentration that can be reached.
The main advantage of thermal incineration is that it can be applied to
all streams. This can enhance the cost effectiveness of this option by
having one large incinerator control all VOC streams. Another advantage of
thermal incineration is that generated heat can be recovered using either
recuperative heat exchangers or waste heat boilers. Recuperative heat exchangers
preheat the incoming vapor stream to reduce the heat addition requirements for
combustion. This method can reduce incinerator energy requirements by 30 to
12
70 percent. Similarly, high temperature exhaust gas from the incinerator
can be used to generate process steam using a waste heat boiler in cases where
the generated steam can be used elsewhere in the latex manufacturing process.
One manufacturer has developed a thermal incineration system for the
13
oxidation of dilute VOC emission streams. The oxidation occurs by passing
the exhaust stream through regenerative combustion beds. Thermal recovery
efficiencies of 85 to 90 percent can be achieved by this system. The system
utilizes a vertical cylindrical combustion chamber surrounded by a series of
packed stoneware beds. The VOC exhaust stream is preheated in a hot stoneware
bed and passes through the combustion chamber, which is kept at a temperature
of 760 C (1400 F). The hot combustion gases pass through other stoneware
beds, transferring the heat of combustion to these beds. Inlet and exhaust
valves on each bed control the gas flow as the beds are depleted or saturated
with heat. This heat recovery system can substantially decrease the amount of
auxiliary fuel required by the thermal incinerator.
19-9
-------
Potential disadvantages associated with the use of thermal incineration
systems at latex plants include variations in the stream flow rate which can
lead to either automatic shutdown from surges or repositioning of the flame
14
zone in the exhaust gas entrance to the vapor control unit. The flow rate
problem can be minimized by proper equipment design, such as the installation
of fast response temperature feedback controls for combustion air and auxiliary
fuel flows to maintain constant combustion characteristics in the thermal
incineration system.
Two styrene-butadiene latex manufacturers have attempted to use thermal
incineration as a VOC emission control technique. One unit encountered
problems in handling the surges of emissions from the batch process of
butadiene removal. This unit is no longer operating. The second unit has
been reducing emissions from the monomer removal vents for ten years. Test
results have shown that this unit reduces emissions by 99 percent based on
organic carbon. When properly designed, thermal incineration is considered
a suitable technique for VOC emissions control in the latex industry.
Control cost estimates have been developed for the installation of
thermal incineration systems with 50 percent and 70 percent recuperative
heat recovery capabilities controlling the butadiene removal vent of the
model plant detailed above. The total installed capital costs and the
annualized costs for these systems are presented in Table 19-3. These
incineration systems are sized for 12,000 scfm, in order to handle the
surges of emissions from the butadiene removal vent and the dilution air
that will be needed to keep the waste gas stream below 25 percent of the
lower explosive limit. The average flow rate of 1300 scfm at 5000 ppm
butadiene was used to determine electricity and natural gas requirements.
Incineration temperature is 1600 F and retention time is 0.5 seconds.
Table 19-3 shows that no natural gas is required in the system that recovers
70 percent of the flue gas heat. No natural gas would be required for the
model parameters used here for heat recoveries of 55 percent or more. Any
additional heat recovery would require diluting the waste gases even more, or
utilizing the heat for steam generation or space heating. No recovery
credits have been included for this heat utilization because it would be
very site specific and would also require additional piping, controls, and
heat exchangers.
19-10
-------
TABLE 19-3. ANNUALIZED COSTS OF IMPLEMENTING RACT
(June 1980 dollars)
Cost item
50% heat
recovery
70% heat
recovery
Installed cost ($)
Purchased equipment
Total installed9
Annualized cost ($/yr)
Annualized capital charges
Capital recovery
Taxes, insurance, and
administration0
Subtotal
Direct costs
Operating labor
Maintenance labor
Natural gasf
Electricity 9
Subtotal
Recovery credits
Net annualized cost
154,500
336,300
54,700
13,500
68,200
6,100
6,000
5,000
1,000
18,100
86,300
180,600
393,000
64,000
15,700
79,700
6,100
6,000
0
1,000
13,100
92,800
Reference 15.
Capital recovery factor of 0.16275 is based on 10 year life and 10 percent
interest rate.
cTotal of 4 percent of total installed cost per year; 1 percent for taxes;
1 percent for insurance; 2 percent for administration.
Based on $11.10 per hour.
eBased on $10.90 per hour.
fBased on $2.40 per 1000 SCF.
9Based on $0.0490 per Kw-hr.
19-11
-------
Catalytic Incineration
Catalytic incineration is similar to thermal incineration in both
principle and operation. The difference is that combustion is completed in
a catalyst bed which reduces requirements for both temperature and residence
time. The temperatures required for combustion range from 204 to 538 C (400
to 1000 F), with shorter residence times than thermal incineration. The
lower temperature and shorter residence time requirements can lower operating
costs by as much as 30 percent compared to thermal incineration in some
16
applications.
The catalyst employed may consist of any of several compounds, deposited
in thin layers on an inert substrate. The substrate is formed into specific
geometric configurations designed to maximize surface area and structural
integrity while simultaneously minimizing pressure drops and ensuring a
uniform, evenly distributed air flow. The performance of the catalyst is
dependent on the contact time and temperature, as well as the organic composition
and concentration of the vapor stream. Deactivation of the catalyst
occurs periodically from sintering and accumulation of poisons. Although
the accumulation of poisons is to some extent reversible, sintering will
eventually deteriorate the catalyst, requiring its replacement. Catalyst
12
lifetimes, in continuous use, usually range from about 1.5 to 2 years, but
lifetimes of up to 7 years have been reported.
Catalytic incinerators have the advantages of smaller size and lower
heat requirements compared to thermal incinerators with low heat recoveries.
These factors decrease both the capital costs (smaller size and lower cost
materials of construction due to lower temperatures) and operational costs for
auxiliary fuels. Similar to thermal incinerators, both recuperative heat
exchangers and waste heat boilers may be employed to further reduce operating
costs.
The principle disadvantage of catalytic incinerators is the high cost of
the noble metals used for catalysts, which require periodic replacement. In
addition to poisoning and sintering, as described earlier, the deposition of
polymeric material on the catalyst could occur, decreasing the catalyst
efficiency or partially plugging the catalyst bed.
19-12
-------
Although no manufacturer of latex has attempted to use catalytic
incineration as an emissions control technique, this method is applicable to
1 fi
any dilute hydrocarbon stream without heavy metal poisons. Catalytic
12
incinerators should provide 90 percent removal efficiency of VOC vapors.
REGULATORY STATUS
State regulations which would apply to the control of VOC emissions from
styrene-butadiene latex manufacturing are summarized in Table 19-4. These
regulations are not specific for this industry, but rather are the general
regulations which apply to any VOC sources greater than a certain size,
usually 100 tons or more of VOC per year.
19-13
-------
TABLE 19-4. SUMMARY OF STATE VOC EMISSION REGULATIONS0
State
Summary of regulation
California
Connecticut
Delaware
Georgia
Kentucky
Louisiana
Massachusetts
Michigan
Minnesota
North Carolina
Ohio
Pennsylvania
Tennessee
Texas
Pumps and compressors must have mechanical seals
or equivalent. Pressure relief valves must be
vented to a vapor recovery or disposal system,
protected by a rupture disc, or maintained by an
inspection system.
Pumps and compressors must have mechanical seals
or equivalent. Waste gases must be burned by
smokeless flare or equivalent.
No applicable regulation.
No applicable regulation.
No applicable regulation.
Pumps, compressors, and valves must be equipped
with mechanical seals or equivalent. Waste gases
must be burned at 1300°F for 0.3 seconds or more
in a direct flame afterburner or equivalent.
No applicable regulation.
No applicable regulation.
No applicable regulation.
No applicable regulation.
Liquid organics must be reduced by 85 percent before
discharge to the atmosphere. Waste gases must be
burned by smokeless flare or equivalent.
Gaseous emissions must be burned by smokeless flare.
No applicable regulation.
Vent gases must be burned at 1300°F in a smokeless
flare or direct flame incinerator.
In addition to the regulations in this table, States may have "odor" and
"nuisance" regulations that may be applicable to the control of VOC emissions
from styrene-butadiene latex operations.
19-14
-------
REFERENCES
1. International Institute of Synthetic Rubber Producers, Inc. The Synthetic
Rubber Manual. 8th edition. Houston, Texas. 1980. 74p.
2. U.S. Environmental Protection Agency. Industrial Process Profiles for
Environmental Use: Chapter 9. The Synthetic Rubber Industry.
EPA-600/2-77-023i. Cincinnati, Ohio. February 1977. p. 76-81.
3. Memo from Ryan, R.B., GCA/Technology Division, to SBC project file,
EPA. March 2, 1981. Summary of 114 responses and telephone conversations.
4. Telecon survey from EPA to representatives of companies using the
emulsion latex polymerization process. October 1979. Requesting
information of polymerization reaction temperature.
5. Letter and attachments from Laundrie, R.W., The General Tire & Rubber
Company, to Goodwin, D.R., EPA. June 6, 1978. Response to Section 114
request.
6. Letter and attachments from Walker, T.C., Firestone Plastics Company,
to Goodwin, D.R., EPA. May 26, 1978. Response to Section 114 request.
7. Letter from Kulka, A.R., Reichhold Chemicals, Inc., to Goodwin, D.R.,
EPA. November 14, 1980. Response to Section 114 request.
8. Letter from Arnold, S.L., Dow Chemical U.S.A., to Goodwin, D.R., EPA.
December 31, 1980. Response to Section 114 request.
9. Letter from Stark, F.J., Rubber Research Elastomerics, Inc., to
Goodwin, D.R., EPA. January 8, 1981. Response to Section 114 request.
10. Draft Emissions Test Report, General Tire Plant. U.S. Environmental
Protection Agency. Research Triangle Park, N.C. EMB Report No. 79-RBM-4.
11. Memo from Mascone, D.C., EPA, to Farmer, J.R., EPA. June 11, 1980.
Thermal Incinerator Performance for NSPS.
12. Oxy-Catalyst. Engineered Emission Control Systems. Technical Bulletin.
13. REECO Re-Therm (Product brochure). September 1980. Regenerative
Environmental Equipment Company, Morris Plains, New Jersey.
19-15
-------
14. Williamson, J.E., R.T. MacKnight, and R.L. Chass. Multiple-Chamber
Incinerator Designs for Los Angeles County. EPA, Institute for Air
Pollution Training. Research Triangle Park, North Carolina. October 1960.
15.
Memo from Ryan, R.B., GCA/Technology Division, to SBC project file, EPA.
June 9, 1981. Costing of increased heat recovery incinerators.
16. Romero, P.L., and A. Warsh. Combustion Evaluation — Sources and Control
Devices. EPA, Office of Manpower and Development. Research Triangle
Park, North Carolina.
17. Surface Coating of Metal Furniture. Background Information for Proposed
Standards (Draft Document). EPA, Emission Standards and Engineering
Division. Research Triangle Park, North Carolina. October 1979.
18. Danielson, J.A. (Editor). Air Pollution Engineering Manual, Second
Edition. EPA Office of Air and Water Programs. Research Triangle Park,
North Carolina. AP-40. May 1973.
19-16
-------
TECHNICAL REPORT DATA
{I'Icutr read Instructions on the rt'tersc be/ore competing)
1. REPORT NO. 2
EPA-400/3-81-007
4. TITLE AND SUBTITLE
Summary of Technical Information for Selected Volat
Organic Compound Source Categories
7. AUTHOR(S)
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Engineering Science
125 West Huntington Drive
Arcadia, California 91006
12. SPONSORING AGENCY NAME AND ADDRESS
DAA, Office of Air Quality Planning and Standards
Office of Air, Noise, and Radiation
US Environmental Protection Agency
3 RfcCIPIENT'S ACCESSION NO.
5 REPORT DATE
ilp M™ 1QR1
6. PERFORMING^ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
1O. PROGRAM ELEMENT NO.
11 CONTRACT/GRANT NO
68-01-4146
13 TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 200/04
15. SUPPLEMENTARY NOTES
16 ABSTRACT
This document is a compilation of existing information of varying detail based
on a review of the technical literature, published and unpublished EPA, State
and local agency reports, and State or local air pollution control regulations
concerning eighteen stationary sources of volatile organic compound emissions.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS b.lDENTIFI
Air Pollution Air Po"
Pollution Control
18. DISTRIBUTION ST AT EMENT 19 SECURI
Unc
Unlimited 20 SECURI
Uncl
ERS/OPEN ENDED TERMS C. COSATI l-'icld/Group
lution Control 13B
TY CLASS (This Report} 21. NO. OF PAGES
:lassified 230
TY CLASS (This page) 22 PRICE
assified
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION 'S OBSOLETE
------- |