EPA-450/3-73-006-g
July 1975
PHTHALIC ANHYDRIDE
AP-42 Section 5-12
Reference Number
1
ENGINEERING
AND COST STUDY
OF AIR POLLUTION CONTROL
FOR THE
PETROCHEMICAL INDUSTRY
VOLUME 7: PHTHALIC
ANHYDRIDE MANUFACTURE
FROM ORTHO-XYLENE
U.S. ENVIRONMENTAL PROTECTION AGENCY
Offirr of Air and Waste Maiiu<£rin<'nl
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
EPA-450/3-73-006-g
ENGINEERING
AND COST STUDY
OF AIR POLLUTION CONTROL
FOR THE
PETROCHEMICAL INDUSTRY
VOLUME 7: PHTHALIC
ANHYDRIDE MANUFACTURE
FROM ORTHO-XYLENE
by
W. A. Schwartz, F. B. Higgins, Jr., J. A. Lee,
R. B. Morris, R. Newirth, and J. W. Pervier
Houdry Division
Air Products and Chemicals, Inc.
P.O. Box 427
Marcus Hook, Pennsylvania 19061
Contract No. 68-02-0255
EPA Project Officer: Leslie B . Evans
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
July 1975
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors
and grantees, and nonprofit organizations - as supplies permit - from
the Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711; or, for a
fee, from the National Technical Information Service, 5285 Port Royal
Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency
by Houdry Division/Air Products and Chemicals, Inc. , Marcus Hook,
Pennsylvania 19061, in fulfillment of Contract No. 68-02-0255. The
contents of this report are reproduced herein as received from Houdry
Division/Air Products and Chemicals, Inc. The opinions, findings,
and conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency. Mention of company
or product names is not to be considered as an endorsement by the
Environmental Protection Agency.
Publication No. EPA-450/3-73-006-g
11
-------
PETROCHEMICAL AIR POLLUTION STUDY
INTRODUCTION TO SERIES
This document is one of a series prepared for the Environmental Protection
Agency (EPA) to assist it in determining those petrochemical processes for
which standards should be promulgated. A total of nine petrochemicals produced
by 12 distinctly different processes has been selected for this type of
in-depth study. These processes are considered to be ones which might warrant
standards as a result of their impact on air quality. Ten volumes, entitled
Engineering and Cost Study of Air Pollution Control for the Petrochemical
Industry (EPA-450/3-73-006a through j) have been prepared.
A combination of expert knowledge and an industry survey was used to
select these processes. The industry survey has been published separately
in a series of four volumes entitled Survey Reports on Atmospheric Emissions
from the Petrochemical Industry (EPA-450/3-73-005a, b, c and d).
The ten volumes of this series report on carbon black, acrylonitrile,
ethylene dichloride, phthalic anhydride (two processes in a single volume),
formaldehyde (two processes in two volumes), ethylene oxide (two processes
in a single volume) high density polyethylene, polyvinyl chloride and vinyl
chloride monomer.
-------
ACKNOWLEDGEMENTS
The study reported in this volume, by its nature, relied on the fullest
cooperation of the companies engaged in the production of phthalic anhydride.
Had their inputs been withheld, or valueless, the study would not have been
possible or at least not as extensive as here reported. Hence, Air Products
wishes to acknowledge this cooperation by listing the contributing companies.
Allied Chemical Corporation
BASF-Wyandotte Corporation
Exxon Chemical Company
Koppers Company
Monsanto Company
Stepan Chemical Company
Union Carbide Chemical Company
United States Steel Corporation
Additionally, Air Products wishes to acknowledge the cooperation of the
member companies of the U. S. Petrochemical Industry and the Manufacturing
Chemists Association for their participation in the public review of an
early draft of this document. More specifically, the individuals who served
on the EPA's Industry Advisory Committee are to be commended for their
advice and guidance at these public meetings.
-------
TABLE OF CONTENTS
Section Page Number
Summary i
I. Introduction PA-1
II. Process Description and Typical Material Balance 2
III. Manufacturing Plants and Emissions 8
IV. Emission Control Devices and Systems 18
V. National Emission Inventory 26
VI. Ground Level Air Quality Determination 27
VII. Cost Effectiveness of Controls 28
VIII. Source Testing 31
IX. Industry Growth Projection 32
X. Plant Inspection Procedures 34
XI. Financial Impact 36
XII. Cost to Industry 41
XIII. Emission Control Devices 44
XIV. Research and Development Needs 47
XV. Research and Development Programs 48
XVI. Sampling, Monitoring and Analytical Methods for
Pollutants in Air Emissions 53
XVII. Emergency Action Plan for Air Pollution Episodes 56
References 63
Appendix I 1-1
Appendix II II-l
Appendix III III-l
Appendix IV - Phthalic Anhydride Production from
Naphthalene PAN-1
-------
LIST OF ILLUSTRATIONS
Figure No.
Figure PA-1
Figure PA-2
Table No.
Table PA-1
1A
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Simplified Flow Diagram for*Phthalic Production
from Ortho-Xylene
Phthalic Anhydride Production -
Capacity Projection
LIST OF TABLES
Page Number
PA-4
PA-33
Page Number
Typical Material Balance PA-5
" " " PA-6
Phthalic Anhydride Reactor System Heat Balance PA-7
Summary of U.S. Phthalic Anhydride Plants PA-9
National Emission Inventory for Phthalic
Anhydride Production from Ortho-Xylene
(4 pages) PA-10
Typical Process Vent Gas Composition PA-14
Scrubber and Waste Incinerator Control System PA-19
Thermal Incinerator for Main Process Vent Stream PA-21
Thermal Incinerator plus Waste Heat Boiler System PA-23
Thermal Incinerator for Waste Product Stream PA-25
Cost Effectiveness for Alternate Emission
Control Devices PA-29
Phthalic Anhydride Manufacturing Cost for a
Typical Existing 130 MMLbs./Year Facility PA-37
Phthalic Anhydride Manufacturing Cost for a
Typical Existing Facility with Air Pollution
Control Equipment PA-38
Phthalic Anhydride Manufacturing Cost for a Typical
Most Feasible New 130 MM Lbs./Yr. Facility PA-39
Pro-Forma Balance Sheet PA-40
Estimated 1985 Air Emissions for Alternate
Control Systems (2 pages) PA-42
Detailed Costs for R & D Project A PA-49
Detailed Costs for R & D Project B PA-51
Summary of Sampling and Analytical Methods
Reported for Pollutants PA-55
Financial Impact of Air Pollution Episodes PA-61
-------
SUMMARY
The phthalic anhydride industry has been studied to determine the extent
of air pollution resulting from the operations of the various plants and
processes of the industry. The purpose of the work was to provide the
Environmental Protection Agency with a portion of the basic data required
in order to reach a decision on the need to promulgate air emission standards
for the industry.
It was concluded, because of the advantages of both raw material cost
and process yield, that the newer ortho-xylene feed process will account for
all or nearly all of the future growth in the industry. However, the older
naphthalene feed process will continue to be operated by those producers who
already have an investment in the process and a ready supply of naphthalene.
Some plant conversions might occur but it is not likely because the ortho-
xylene process employs a fixed bed tubular reaction system and the naphthalene
process employs a coil-cooled fluid bed reaction system. For these reasons,
the body of this report is devoted to a study of the air emissions from the
ortho-xylene process while Appendix IV addresses itself to a brief review of
the naphthalene process emissions.
In general terms, the air emissions from both processes fall into the
categories of organic acids and anhydrides, hydrocarbons and carbon monoxide.
Many of the heavier organic materials are solid at ambient/±emperatures and,
therefore, have generally been classified as "particulates". In addition,
operators of ortho-xylene plants have reported sulfur oxide emissions (resulting
from the addition of sulfur to the feed as a catalyst activator) and nitrogen
oxide emissionss(resulting from waste incinerators). As practiced today, an
average emission factor for the ortho-xylene process is about 0.103 Ibs./lb.
of phthalic anhydride produced. This is equivalent to nearly 52 million
Ibs./year. Of this total, about 85 percent is carbon monoxide, about 10
percent particulates (as defined above) and about five percent oxides of
sulfur. Less than one percent is emitted as nitrogen oxides and other
hydrocarbons. In addition, the naphthalene based processes emit about 33
million Ibs./year (an emission factor of about .055 Ibs./lb. of phthalic
anhydride produced) of which about 95 percent is carbon monoxide and most of
the balance "particulates". Assuming that all of the naphthalene based plants
are still operating in 1985 and that all new capacity is in the form of ortho-
xylene based plants with average pollution control as practiced today, the
tatal emission will increase to nearly 220 million Ibs./year of which 186
million Ibs. will be due to the ortho-xylene plants. Of the total, nearly 190
million Ibs. will be carbon monoxide. A further point which has not been
quantified but which was cited at the public review of an early draft of this
report is the trend for future ortho-xylene feeds to be of lower purity.
Any increase in meta- and/or para-xylene in the feed will lead to increases
in emissions beyond those cited above - mainly in the form of toluic acid and
carbon oxides.
Most of the ortho-xylene based plants that were covered by the survey
have some form of pollution control device on the process. These are either a
main process vent gas incinerator (some with heat recovery) or a main process
vent gas scrubber followed by a scrubber water incinerator. Some incineration
of fractionation wastes is also practiced. The reported efficiencies of these
devices were used in estimating the emission factors given above. Thus it can
be seen that there is room for improvement in the design and operation of
pollution control devices for phthalic anhydride plants. The major problems
-------
ii
SUMMARY (continued)
are (1) scrubbing with water does not remove carbon monoxide and (2) incineration
of organic particulates is very difficult.
It was concluded that either a scrubber or an incinerator on the main
process vent will do an approximately equally efficient job if control of
organic particulates (such as phthalic anhydride, maleic anhydride and benzoic
acid) is the primary goal of emission reduction programs. However, if carbon
monoxide is also to be controlled, then an incinerator is the most feasible
demonstrated control device. In either case, incineration of fractionation
wastes is the most feasible demonstrated control technique. Vith a scrubber
this can best be accomplished in the water incinerator but with a main
vent incinerator, a separate fractionator incinerator is recommenced to minimize
control problems. Thus, if a water scrubber is used on all existing uncontrolled
o-xylene plants and on all future plants, 1985 emissions of particulates would
be reduced by about 10 million Ibs. per year out of the estimated totals of 20
million Ibs. per year for the industry and 18 million Ibs. per year for the
o-xylene segment of the industry. No significant reduction in carbon monoxide
emissions would be achieved. On the other hand, the retro-fitting of scrubbers
on all uncontrolled existing o-xylene plants and the installation of dual
incinerators (one on the main process vent and one on the fractionation waste)
on all future plants will achieve about the same 10 million pound per year
reduction in particulate emissions along with a 100 million pound per year
reduction in carbon monoxide emissions out of the estimated carbon monoxide
totals of 190 million Ibs. for the industry and 157 million Ibs. for the
o-xylene segment of the industry. Most of the residual 57 million Ibs./year of
1985 carbon monoxide emissions from o-xylene plants is attributable to existing
plants with scrubbing systems. In summary, the estimated 1985 emission factors
would be reduced to about 0.098 Ibs./lb. if main vent scrubbers followed by
combination water and fractionation waste incinerators are used while the
reduction would be to about 0.042 Ibs./lb. with the dual incinerator system.
It should be noted that no analysis of most feasible emission control
schemes has been made on naphthalene based plants because use of the process
is not expected to expand, and the purpose of the study is to aid the EPA in
decision making relative to new stationary sources of air pollution. However,
in the course of surveying the naphthalene based plants it was determined that
the use of either a scrubbing system or an incineration system is widespread
in this segment of the industry, thus accounting for the relatively low (0.055
Ibs./lb.) emission factor and the predominance (95 percent) of carbon monoxide
in the make-up of this factor.
The costs involved in installing the various pollution control devices on
new 130 MM Ibs./year ortho-xylene based plants have been estimated to be
similar, regardless of the system employed (1973 dollars). Thus, the use of a
scrubber on the main process vent followed by incineration of the scrubber
water and fractionation wastes in a combination unit has been estimated to
cost about $1,450,000 to install and about $420,000 per year to operate.
Incineration systems will have a net operating cost of about $450,000 per year
regardless of whether or not waste heat is recovered in the form of steam.
This is because the extra capital is almost exactly offset by the value of the
steam. The dual incinerator system with waste heat recovery would require
about $1,400,000 in capital as compared with a $1,000,000 investment if the
waste heat is not recovered. It has been estimated that nine new plants will
be built by 1985. Therefore, the capital required is of the order of $13 million
for the industry. This is about 13 percent of the total investment required
-------
iii
SUMMARY (continued)
but is not considered a hardship because most existing plants are using
devices similar to those covered in the study. Thus, the difficult decision
is not whether or not controls should be installed, but whether or not incineration
should be the control technique so as to minimize CO emissions. If the steam
can be used, it is a relatively easy decision, but if not, an additional equivalent
of 3.9 billion SCF of natural gas per year will be required to incinerate off-
gases and achieve the estimated reductions in carbon monoxide emission. The
overall impact of this type of decision on the environment must be carefully
weighed.
In the course of the study, it was concluded that major areas for research
in the industry would be (1) the use of oxygen enriched feeds and off-gas
recycle, (2) improved catalyst and (3) fluid bed reactors. All of these efforts
could result in improved yields (i.e., less carbon monoxide waste) and/or
reduced volumes of total vent gas which could make incineration a more viable
control alternative. All of this research can best be carried out by today's
producers of phthalic anhydride.
-------
PA-1
I. Introduction
Prior to 1945, all phthalic anhydride (PAN) vas produced by oxidation of
naphthalene. During the last fev years, o-xylene has become a major raw
material for making this chemical. At present, about 557, of the 0»9 billion
pound per/year U. S. phthalic anhydride production is obtained by oxidizing
o-xyleneA Since o-xylene represents a cheaper raw material and on a weight
basis yields somewhat more product (about 1.0 Ib. PAN/lb. xylene versus 0.97
Ib. PAN/lb. naphthalene), it is anticipated that all future plants will be
designed to use xylene feed.
Total U. S. annual phthalic anhydride production is estimated to increase
to 2.2 billion pounds by 1985. Currently over 90 percent of U. S. phthalic
anhydride is used for plastics, paints and synthetic resins.1 Approximately
507» of the PAN produced is used to make phthalate plasticizers.
A variety of processes are licensed for PAN production. Presently in the
U. S. all phthalic anhydride is produced by vapor phase oxidation of xylene
and naphthalene. The primary naphthalene processes use fluidized bed converters
whereas all xylene based plants incorporate tubular fixed bed reactors. Except
for the reactors and catalyst handling and recovery facilities required for the
fluid units, these vapor phase processes are similar.
The major atmospheric emission generated in PAN manufacture is the non-
condensibles contained in the reactor effluent. In addition to excess air
provided for the reaction, the reject stream contains carbon oxides plus a
small amount of phthalic. anhydride, maleic anhydride and organic acids. In
most plants this stream is sent to a scrubber of incinerator before being
vented to the atmosphere.
The following report deals with the production of phthalic anhydride from
o-xylene,, Appendix IV presents data for PAN derived from naphthalene.
-------
PA-2
II. Process Description and Typical Material Balance
The basic chemical equation for this air oxidation process is as follows:
3 + 3 02
0-xylene Oxygen
Principal side reactions include;
3 H20
Phthalic
Anhydride (PAN) Water
0-xylene Oxygen
02
8 C02
Carbon
Dioxide
0
H,A
C
£,'
H C
Water
4 C0
4 H20
Maleic Carbon
0-xylene Oxygen Anhydride (HAN") Dioxide Water
[_ Filtered air is compressed to 10-14 PSIG and passed through a preheater.
Liquid ortho-xylene (normally 95 - 96 vt. ?„ purity) is vaporized and mixed
vith the preheated air before entering fixed bed tubular reactors. These
reactors contain vanadium pentoxide catalyst and are operated at 650 - 725° F.
A small amount of sulfur dioxide is added to the reactor feed in order to
maintain catalyst activity. A molten salt bath is circulated around the
reactor tubes in order to remove heat produced in the exothermic reaction.
This heat is transferred to a steam generation system."!
\The reactor effluent contains the above specified reaction products plus
nitrogen, excess oxygen and small amounts of CO, benzoic acid and S02. This
effluent is used to generate lov pressure steam in a vaste heat boiler and then
passes to a separation system where phthalic anhydride is cooled and condensed
as solid crystals in a series of parallel switch condensers. The condenser
effluent gases are normally water scrubbed and/or sent to an incinerator before
being released to the atmosphere.
The individual switch condensers are alternately cooled and heated by
separate heat transfer oil streams in an automatically controlled time cycle.
Crude phthalic anhydride is recovered by melting it from the condenser tutie
fins during the hot oil circulation period.")
fit
The raw liquid phthalic anhydride is heated and sent to a pretreatment
sect km which consists of one or two heated vessels. In this equipment dis-
solved phthalic acid is dehydrated to the anhydride, and the associated water,
maleic anhydride and benzoic acid are partially evaporated. The pretreated
-------
PA-3
liquid stream is then sent to a vacuum distillation section vhere pure phthalic
anhydride (99.8 vt. ?„ purity) is obtained as a distillate vhich can be stored
either in the molten state or processed further by solidifying to flakes in a
flaker and bagged for shipment.
Figure PA-1 shovs a flow diagram for a typical plant and indicates the
various vent streams from this unit.
Commercial grade o-xylene (about 95 - 96 vt. % purity) is used as
feedstock. Feed impurities, which primarily consist of M + P-xylene, are
converted to carbon oxides in the reactors. By modifying the reactor
catalyst it would be possible to supplement or replace the o-xylene vith
naphthalene feed. No other alternate feed materials are available.
/" Table PA-1 presents a typical material balance for an average size
plant producing 130 million pounds per year of phthalic anhydride from ortho-
xylene. Table PA-1A presents the same balance with quantities expressed as
J tons per ton of PAN. Vent gas composition and hydrocarbon losses shown in
/ these balances were derived by averaging survey data from various U. S.
I commercial plants. Xylene feed quantity was calculated by carbon balance.
I This figure agrees within 1 or 2% with both published 3 and plant survey
/ data on feedstock requirements. The material balance shows 0.98 pounds of
PAN product per pound of xylene feed.(a) This corresponds to 73.4% of the
V maximum theoretical yield.
Table PA-2 presents an estimated heat balance around the process
reactor system.
(a) Current technology has improved yield to about 1.0 Ibs./lb. of xylene feed.
-------
MULTlPLt
TO
MOT MJD COOL
CIB.OULA.TIIJ4, OIL
•5>Te.frA.M«> 00.
W4.T6-Z. t
11-6-2 F-ErtD
»/*.Ttt
HEAT EXCHANGER
V&UT
tfc-SIDUfr
LI^MT
f-OB THEr MA\TfcCIA.L feA.LAtUC.tr
x-"
' 1-
2
ill
3
1-
Oi
tJ
•
u
Cl
^
^N
^
^
/^ -.TErA-U
— v^v ^
>
1
3
3
J
O
O
J
a
n
i-
^:^^
^
^
^^
*
Q
(J
J
5
iL
ill
kl
'">•_
1
_ *Tfr«,k/
1
S-' *
T®
VfckJT
PRODUCT
fc.M.
7V
'I ') K
8-UVie.OIJMErlJrAL
COUTg.A.CT
A,. PEM>.
F-lbLltt PA.-I
•>IMPLIF-lfrD
F-OK : PHTMA.LIC. A-UHVOtlOt- PKOO-
Qg.THO-XVUfc.ufe-
-------
TABU; PA-I
Component
Sulfur Oxides
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Phthallc Anhydride
Maleic Anhydride
Benioic Acid
0-xylene
M & P-xylene
Misc. Organlcs
Residue
Water
Total Lbs./Hr.
TYPICAL MATERIAL BALANCE
FOR PRODUCIHG 130 MM LBS . /YR . PRTHALIC ANHYDRIDE
FROM ORTHO XYLENB
12345 6 7 8 9
Ortho Xylene Air to Catalyst Process Pretreatment Pretreatment Distillation Distillation Phthallc Anhydride
Feed Reactor Activator Vent Gas Light Ends Residue Light Ends Bottoms Product
75 <°> 75
2,411
8,326
407,861 407,760 101
123,907 101,940
368 (b) 70 85 16 16,000
694 32 (d> 544 (d> 32
45 (b> 69
15,617
736
19
56
5,780 15,333 (a)
16,353 537,548 75 536,952 203 (a) 56 698 35 16,032
(a) Excludes 4,800 Ibs./hr. of condensate fro* steam ejector.
(b) Normally vapor but can be present as partlculate at low temperature.
(c) Quantity changes with catalyet age. Value shown corresponds to relatively fresh catalyst. Can be 150-200 PPH for aged catalyst.
(d) About 90 percent of maleic anhydride formed is in process vent. Therefore, amount shown in distillation ends is high.
-------
TABLE PA-1A
1 2
Ortho Xylene Air to
Component Feed Reactor
Sulfur Oxides
Carbon Monoxide
Carbon Dioxide
Nitrogen 25.4913
Oxygen 7.7442
Phthalic Anhydride
Maleic Anhydride
Benzole Acid
0-xylene 0.9761
M 6. P-xylene 0.0460
Misc. Organics
Residue
Water 0.3613
TYPICAL MATERIAL BALANCE
FOR PRODUCING 130 MM LBS./YR. PHTHALIC ANHYDRIDE
FROM ORTHO-XYLENE
3456 78 9
Catalyst Process Pretreatment Pretreatment Distillation Distillation Phthalic Anhydride
Activator Vent Gas Light Ends Residue Light Ends Bottoms Product
0.0047 (c) 0.0047 (c)
0.1507
0.5204
25.4850 0.0063
6.3713
•0.0230 (b) 0.0044 0.0053 0.0010 1.0000
.0.0434 (b) 0.0020 (d) 0.0340 0.0020
. 0.0028 0>) 0.0043
o»
0.0012
0.0035
0.9583 (a)
Total Tons/Ton PAN 1.0221
33.5968
0.0047
33.5596
0.0127
0.0035
0.0436
0.0022
1.0020
(a) Excludes 0.3000 T/T of condensate from steam ejector.
(b) Normally vapor but can be present as particulate at low temperature.
(c) Quantity changes vith catalyst age. Value shovn corresponds to relatively fresh catalyst. Can be 0.0095-0.0125 T/T for aged catalyst.
(d) About 90 percent of maleic anhydride formed is in process vent. Therefore, amount shown in distillation ends is high.
-------
PA-7
TABLE PA-2
PHTHALIC ANHYDRIDE REACTOR SYSTEM
HEAT BALANCE*
HEAT OUT BTU/LB. OF PAN
Steam Generation
Reactor Internal Cooling
Effluent Heat Recovery Boiler (From Reactor)
Reactor Heat Losses
Effluent Condenser
Incremental Effluent Heat Content**
Total
HEAT IN
Exothermic Heat of Reaction
Feed Vaporization and Preheat
Total
*Basis
1) Table PA-1A material balance.
2) Feed preheated to 300° F.
3) Reactor outlet temperature 680° F.
**Difference in heat content of effluent @ 150° F and feed xylene @ 80° F
(liquid) plus air @ 100° F.
-------
PA-8
III. Manufacturing Plants and Emissions
Table PA-3 presents a list of U. S. plants producing phthalic anhydride.
This table also shovs published 4,5 capacity figures and type of feedstock
employed in these units. Approximately 40% of the PAN produced from o-xylene
is manufactured in Illinois and the remaining xylene derived production is
in New Jersey, Texas, Louisiana and California. Several of the plants are
located close to major metropolitan areas. One large unit is in Chicago,
Illinois and another is five miles vest of Nev York City. The other
installations are within five miles of towns and cities vith population
ranging between 1000 and 165,000.
Table PA-4 shows individual plant capacity figures and emission data
for most of the major U. S. plants producing PAN from o-xylene. Emissions
from these plants are as follows-
A. Continuous Air Emissions
1. Main Process Vent Gas
This primary air emission stream consists of the gross reactor
effluent after cooling and recovery of crude phthalic anhydride.
Normally the stream is sent to a water scrubber and/or an incin-
erator for removal of hydrocarbons and other pollutants before
atmospheric venting. The process vent gas emissions presented in
Table PA-4 are downstream of pollution control devices and
represent average values, with the actual composition depending
on catalyst activity and reactor operating conditions. Table
PA-5 shows a typical breakdown of components in the vent stream
before entering emission control facilities. Under normal
operating conditions the quantity of hydrocarbon emissions in this
stream are primarily influenced by the amount of air charged to
the reactors. Table PA-5 indicates that the air rate employed in
the various plants and the average overall stream composition does
not vary over a wide range.
Plants incorporate multiple parallel reactors and some of the
larger units have complete parallel trains of equipment in the main
processing areas. Therefore, at reduced plant production it should
be possible to shut down part of this equipment and reduce vent gas
emissions accordingly.
None of the surveyed plants report any odor complaints concerning
the main process vent gas or other vent streams. Plant 53-5 is the
only facility that reported process vent gas odors (PAN & MAN)
passing beyond battery limits and this is the only surveyed unit
that does not have a pollution control device on this vent stream.
A published article ^ indicates nuisance odors and lachrymatory
emissions can exist under certain weather conditions. This reference
shows the following odor threshold and eye irritation concentrations
for known organics in the off-gases from PAN manufacture;
Odor Threshold Eye Irritation
. Component PPM* PPM*
Maleic Anhydride Q.3 0.23
Formaldehyde <1 - 1.0 2-3
Phthalic Anhydride
*By volume.
-------
TABLE PA-3
SUMMARY OF U. S PHTHALIC ANHYDRIDE PLANTS f4l (5)
Conpany
Allied Chen. Corp.
Indust. Chens. Div.
Plastics Div.
BASF Wyandotte Corp.
BASF A.G., subsid.
V. R. Grace & Co.
Hatco Group
Hatco Chen. Div.
Hoppers Co,, Inc.
Organic Materials D
Monsanto Co.
Industrial Chems. Div.
Reichhold Cheng., Inc.
The Shervln-Villlans Co.
Sherwin Williams Chems., Div.
Standard Oil Co. of California
Chevron Chen. Co., subsld.
Indust. ChenB. Div.
Location
El Segundo. Calif.
Frankford (Philadelphia). Pa
Ironton , Ohio
Kearny, N.J.
Fords, N J.
Brtdgaville, Pa,
Chicago, III.
Bridgeport
Texas City. Texas
Morris, 111
Chicago, 111.
Richmond, Calif.
Published Capacity
MM Lbs./Year
40
100
33
130
75
95
130
80
130
100
20
50
Rav Material
0-Xylene
Coal-tar naphthalene
Coal-tar naphthalene
0-Xylene
Coal-tar and petro-naphthalene
Desulfurlced naphthalene
0-Xylen«
Petro-naphthalene
0-Xylene
0-Xylene
Petro-naphthalene
0-Xylene
Exxon Chen. Co., U.S.A.
Stepan Chen. Co.
Indust. Chens. Div.
Union Carbide Corp.
Chens, and Plastics Div.
United States Steel Corp.
USE Chens., Div.
Baton Rouge, La.
Mlllsdale, 111.
Institute, W. Virginia
Neville Island (Pittsburgh, Pa.)
90
50
75
125
0-Xylene
0-Xylene
Petro-naphthalene
Desulfurlzed naphthalene
Total - 1,323
-------
TABLE PA-4
(e)
Plant Code Number
Date on-stream
Capacity - Tons of Phthalie Anhydride/Yr.
Average Production - Tons PAN/Yr.
Range In Production - 7. of Max.
By- Products - Ton«/Yr.
Maleic Anhydride
Benzole Acid
Emissions to Atmosphere
Stream
Flow, <*> Lbs./Hr.
Flov Characteristic
if Intermittent - Hrs. of Flov/Yr.
Composition - T/T PAN
Particulate
Nitrogen Oxides
Sulfur Oxides (»)
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Phthalie Acid <">
Phthalie Anhydride (PAN)
Maleic Acid (e)
Maleic Anhydride (MAN) («>
Benzole Acid (e)
Xylene
Misc. Organics
Water
Simple Tap Location
Date or Frequency of Sampling
Type of Analysis
Odor
Vent Stacks
Number
Height
Diameter
Exit Gas Temp. , °F
SCFM per Stack
Type Emission Control
Date Installed
Total Emissions - Ton/Ton PAN
Hydrocarbons
Participates & Aerosols (m)
CO
Scrubber
Mfnt
530,000
Continuous
0.0011 (d>
0.0050
(j)
X
X
X
) 0.0013
X
75' above grade
Once in 1971
Organic Acids - Flame
SOX - Modified Shell
None
1
110'
8 '6"
113
118,500
Scrubber
1971
NATIONAL EMISSION? INVENTORY
FOR PHTHALIC ANHYDRIDE PRODUCTION
FROM ORTHO-XYLENE Sheet 1 of 4
53-1
1971
64,000
No seasonal variation
3,420 (b>
256 00
Incinerator Xylene Product
Vent Storage Storage
Continuous Continuous Continuous
0.0009 (<0
0.0007 (D
0.0028
0.1200
X
0.0302
) 0.0002
)0.0005
0.00004 (f)
X
Difficult access
Once None None
lonization Orsat Calc'd.
Development
None None Not beyond battery limits (PAN)
1 17
108' 10'
5 '11"
1,700 Ambient 302
2,550 20
Incinerator None Sublimation Traps
1971 1971
0.0040
0.0057 <*>
0.0028
53-2
1970
65.000
60.000
No seasonal variation
Scrubber Incinerator
Vent Vent
540,000 27,900
Continuous Continuous
0.0068
0.1490 0.0020
0.5728 X
25.4382 1.4202
6.2938 0.0482
0.0005
0.0019
0*0011
1.3950 X
Accessible None
None None
Calc'd. Calc'd.
None None
2 1
100' 100'
6'9" 5'
100 1,600
60,600 5,900
2 Scrubbers Incinerator
1970 1970
0.0035
0.0068
0.1510
-------
TABLE PA-4
Plant Cod* Number
Date on-strean
Capacity - Tons of Phthalic Anhydrlde/Yr.
Average Production - Ton* PAN/Yr.
. King* in Production - T of Max.
By-Productc - Toni/Yr.
Malelc Anhydride
Bencoic Acid
biiitont to Atmosphere
Streen
Flo*, (•> Lbi./Hr.
Flow Characteristic
If Intermittent - Rri. of Flov/Tr.
CaBpoaltloa - T/T PAN
Particular*
Nitrogen Oxide*
Sulfur Oxides
Phthallc Anhydride (PAH) (•)
Milalc Acid (•)
mlelc Anhydride (MM) <•>
Benxolc Acid (e)
Xylene
Misc. Organici
Water
Sample Tap Location
Date or Frequency of Sampling
Type of Analysis
Odor
Vent Staeki
Height
Diameter
Exit Ca» Temp., °F
SCFK per Stack
Type Balailon Control
Date Installed
Total Emissions - Ton/Ton PAN
Hydrocarbons
particulates & Aerosols (•>
NOK
s<$
CO
NATIONAL EMISSIONS INVENTORY
FOR PHTHALIC
raon
53-3
1971
43.500
ANHYDRIDE PRODUCTION
ORTHO-XYLEUE Sheet 2 of 4
No seasonal variation
Incinerator
vut
490,000
Continuous
0.0042
400 («>>
Entrgency
VMf
363.000
Intendttent
2 srinutes/yr.
0.0042
0.1700
0.3399
25.9327
8.4585
0.0340
0.0306
0.0023
1.0198
Available
Hone (S)
Mn-Pol*rograph
CO, C02&HCby GC
battery Halts
(PAN & MAN)
150
81,000
0.0110
0.0044 (h)
Distillation
VMt
50
Continuous
X
X
TR
Ron*
Hate
Calc'd.
None
1
100'
6"
135
11
Ejector Hot we 11
53-4
1971
97,500
65.000
No seasonal variation
Start -tip Process Vent
Vent IflfiiUraftf Pffllsir*
57,000 745,000
Intermittent Continuous
0.0002
0.0013
0.001
0.0001
X 1.3251
X 34.4001
X 7.9023
0.0032
X 3.0182
None Available
None 4 tlsjes Oct. 71 -Feb.
None Part. - Fed. Keg. Met
NQ_ - EPA Method 7 -
None Only during f laaja-out
(4 tines per year)
4 2
50' 200'
2' 7'6"
392 482
12.800 84,500
2 Incinerators + W.H. B
1971
0.0096
0.0014
0.0050
0.0013
Distillation
§ater Ef
18,000
Continuous
0.0001
0.0001
0.0012
0.0626
0.6263
0.0939
0.0001
0.3444
Available
72 Twice a
hod 5, NO, - EHk Method 7
Orsat Orsat
: None
1
100'
2'9H
1200- 1600 (o)
4650
Inciaerator
1971
-------
TABLE PA-4
NATIONAL EMISSIOHS INVENTORY
FOR PHTHALIC ANHYDRIDE PHODtlCTION
FROM
Sheet 3 of 4
Plant Cod* Number
Date en-stream
Capacity - Ton* of Phthalic Anhydride/Yr.
Average Production - Ton* PAN/Yr.
Range In Production - % of Mix.
By-Product* - Ton*/Yr.
Malelc Anhydride
Bentolc Acid
Emliilon* to Atmosphere
Stream
Flo*, (•' Lb*./Hr.
Flov Characteristic
If Intermittent - Hr*. of Flov/Yr.
Composition - T/T PAN
Participate
Nitrogen Oxide*
Sulfur Oxide* (ti)
Carbon Monoxide
Carbon Dioxide
nitrogen
Oxygen
Phthalic Acid
Phthalic Anhydride (PAN) (e)
Maleic Acid (e)
mleic Anhydride (MAN) («)
Bentolc Acid («)
Xylene
Misc. Organic*
Water
Sample Tap Location
Date or Frequency of Sampling
Type of Analyil*
Odor
Vent Stack*
Number
Height
Diameter
Exit Gas Temp., °F
SCFM per Stack
Type Emission Control
Date Installed
Total Emissions - Ton/Ton PAH
Hydrocarbons
Particulates & Aerosols'1")
NOX
so*
CO
Process
Vent
Continuous
0.1551
0.0017
0.0016
0.0638
Accessible
Not Normally Sampled
Flame lonization
Yes but no complaints (PAN 6. MAN)
1
85'
3'
110
40,000
53-5
23.000
No seasonal variation
Product Light Ends Light Ends Flaker Bagging
Storage from Pretreatment From Distillation Exhaust Exhaust
Unknovn Unknovn
Continuous Continuous Continuous Continuous Continuous
X
X X
X
0.0016 0.0001 X
Accessible Hazardous Accessible Impossible Accessible
None None None None None
Calc'd. None None Celc'd. None
Not beyond battery limits (PAN) None None None None
1 1111
15- 75' 85' 60' 20'
2" 2" 2" 14" 18" x 18"
300 275 200 Ambient Ambient
14 1 165
0.0688
0.1551
-------
PA-13
Table PA-4 Footnotes Sheet 4 of 4
(a) Based on average production rate if available, otherwise based on design
capacity.
(b) Not recovered, burned in on-site incinerator.
(c) Estimated based on approximate material balance.
(d) Primarily MgC(>3 & CaC(>3 from water supply.
(e) Normally vapor but can be present as particulate at low temperature.
(f) Emission rate 0.00014 T/T PAN during tank filling operations (47 times
per year).
(g) Stream sampled ten times but never during emergency.
(h) Excludes S02 produced from supplemental fuel. Maximum additional S02
from the source is 0.0028 T/T. In 1973 S02 from this source will drop
to 0.0002 T/T.
(j) Stream not analyzed for CO content.
(k) Excludes CO contained in scrubber vent.
(1) Includes 0.0007 T/T of sulfur dioxide resulting from sulfur contained
in fuel.
(m) Includes PAN, MAN and organic acids.
(n) Function of catalyst age and incinerator fuel sulfur content.
(o) Temperature set at -1200° F for light end feed and 1600° F for intermittent
feed of distillation reject liquid.
It should be noted that flow rates and compositions shown for intermittent
streams represent emissions during flowing condition and not yearly averaged
values. The total plant emissions figures shown are averaged emissions for
extended periods of operation.
-------
PA-14
TABLE PA-5
TYPICAL PROCESS VENT GAS COMPOSITION
00
FOR
130 MM LB./YR. PHTHALIC ANHYDRIDE PRODUCTION
FROM ORTHO XYLENE
Component
Sulfur Dioxide
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Phthalic Anhydride
Maleic Anhydride
Benzole Acid
Misc. Hydrocarbons
Water
Normal Range in Average
Composition, Mol
Average Flow Rate
MPH LB./HR.
0.006
0.4
0.6
76
16.5
- .012
- 0.5
- 1.8
- 79
- 16.9
1.2
(.050 - 0.065
4.0 - 5.5
(b)
86.1
189.2
14,563.0
3,185.6
2.5
7.1
0.4
849.9
18,885.0
75
2,411
8,326
407,760
101,940
368
694
45
15,333
(c)
536,952
(a) Up-stream of any pollution control device.
(b) Complete breakdown of organic material not available. Typical
hydrocarbon composition shovn is based on data provided by plants
53-3 and 53-4. Published data (6) indicates some aldehydes CIO -
100 PPM by volume) will also be present.
(c) Represent fresh catalyst value, aged catalyst value could be
150-200 PPH.
-------
PA-15
2. Pretreatment and Product Fractionation Vents
The heat pretreatment and distillation of crude PAN is performed
under vacuum. These operations evolve dissolved non-condensibles
and some light ends which are emitted in the vacuum jet ejector
exhaust stream. In the surveyed plants, this reject stream is
handled as follows:
53-1 - A steam jet ejector with after condenser is used to
generate vacuum. Effluent water is sent to main process
vent scrubbers, which results in most of the contained
hydrocarbons being burned in the downstream waste water
thermal incinerator.
53-2 - Effluent believed to be sent to incinerator.
53-3 - Steam vacuum jets are employed with exhaust gases completely
condensed in a hotwell. Resulting waste water is sent to
sewer.
53-4 - Exhaust steam from ejectors is sent to waste product
thermal incinerator.
53-5 - Exhaust steam vented to atmosphere.
3. Waste Product Incineration
In all but one surveyed plant (53-5) , impurities removed in
product distillation facilities are either sent to the main process
vent incinerator (53-1 and 53-2) or a separate incinerator for waste
products (53-3 and 53-4).
4. Feedstock Storage Vent
Xylene feed is stored at ambient temperature in fixed roof
storage tanks with atmospheric vents. Because of low vapor pressure
(0.25-0.35 PSI) hydrocarbon emissions from this source are small
(0.0001 T/T of PAN).
5. Product Storage Vent
Crude and pure phthalic anhydride is stored at 300° F and
atmospheric pressure in order to be kept molten. The tanks are
normally blanketed with dry nitrogen (53-1, 53-3 and 53-4) to
prevent the entry of oxygen (fire) or water vapor (hydrolysis to
phthalic acid). Consequently, there is a continuous gaseous
effluent. Disposal methods employed for product storage losses
are as follows:
53-1 - Vent stream is permitted to cool for knock-out of solid
PAN prior to atmospheric venting.
53-2 - Vents from all three product storage tanks are collected
by an ejector and burned in main process vent gas
incinerator.
53-3 - The three product storage tanks vent to atmosphere.
-------
PA-16
53.4 . Crude phthalic anhydride storage tanks (two) are vented
to waste product incinerator and the four PAN product
storage tanks vent to the atmosphere.
53-5 - The two tanks involved have atmospheric vents.
6. Flaker and Bagging Exhaust
In most plants PAN product is stored and transported as a
liquid. Occasionally (plant 53-5) the product is shipped as a
flaked solid. Air emissions of PAN from the flaker and bagging
operations are small (about 0.0001 T/T).
7. Heat Transfer Fluid
A small amount of hydrocarbon emissions (0.00002 T/T of PAN)
result from vents on surge tanks in the switch condenser heat
transfer oil circuit.
B. Intermittent Air Emissions
1. Process Vent Gas
At least one plant (53-3) provides for emergency venting of
the process vent gas incinerator feed. If emergency occurs, entire
feed stream is vented until air compressor can be shut down (about
two minutes). Thereafter, flow essentially stops. However, some
hydrocarbon emissions continue for several hours until all switch
condensers can be melted and dumped. Venting occurs about once a
year.
2. Start-Up Vent
Plant 53-4 provides for direct atmospheric venting of reactor
effluent during plant start-up. These vents are used approximately
once per year during heat up of the reactors. Emission consists of
hot air and natural gas combustion products.
3. Product Shipping Losses
Most product phthalic anhydride is shipped by tank truck. Batch
emissions of PAN result from uncontrolled vapor losses during tank
truck loading. These losses are estimated to be 1/4 to 3/8 Ibs./min.
with average yearly loss equal of 0.0001 T/T of PAN product. A fume
scurbber is used by 53-4 to reduce emissions.
C. Liquid Wastes
Most liquid waste streams are sent to on-site incinerators as
previously indicated. The following waste water streams are sent
off-site:
53-1 - Cooling tower blowdown (0.47 T/T of PAN).
53-3 - Distillation vacuum jet condensate equal to approximately
1.2 tons/ton of PAN. In original plant design, the
stream was to be sent to the process vent gas incinerator.
However, incinerator operating problems have prevented the
burning of this material and it is now sent to sewer.
-------
PA-17
53-4 - Steam boiler blowdown (0.52 T/T of PAN).
53-5 - A total of 4.8 T/T of PAN.
D. Solid Wastes
These materials consist primarily of residue and heavy ends
removed during product pretreatment and distillation. Specific
solid vastes sent off-site for disposal by the various surveyed
plants are as follows:
53-3 - Light and heavy ends (0.02 T/T of PAN) rejected in
product distillation are sent off-site.
53-5 - Approximately 0.004 tons of pretreater residue and 0.002
tons of distillation bottoms per ton of PAN product.
E. Fugitive Emissions
All plants surveyed report that fugitive emissions such as
those resulting from minor leaks and periodic equipment decontamination
for maintenance are small. One plant (53-1) estimates these losses
to be less than 0.001 T/T of PAN. Whereas plant 53-4 estimates total
loss is less than 0.0001 T/T of PAN. The other producers indicate
the loss is negligible.
-------
PA-18
IV. Emission Control Devices and Systems
A. Main Process Vent Gas Stream
1. Devices Currently Employed
All but one of the surveyed plants provide an emission control
device on this stream. These devices include:
(a) Scrubber plus Thermal Incineration
Plants 53-1 and 53-2 incorporate a vater scrubber on the
process vent gas and incinerate reject vaste vater from the
scrubbing system.
In plant 53-1 a proprietary co-current scrubber vith a
5000 GPM circulation rate is employed for cleaning about 120,000
SCFM of off-gases. Make-up vater (unknovn rate) is sprayed into
the scrubber system up-stream of one of tvo demisters on the
scrubber o<:f-gases. This arrangement results in minimizing the
amount of entrained vater and at the same time minimizing the
quantity of polluted vater that is included in the total
entrainment. The off-gases from the demisters go to the stack
at about 110 - 120° F. Effluent from steam jet ejectors in
the product pretreatment and distillation areas is also fed to
the scrubber system.
An oil fired thermal incinerator is used to incinerate about
6,000 Ibs./hr. of solution bled from the scrubber recirculation stream
and the light and heavy ends from the product purification columns.
The incinerator operates at about 1700° F and reportedly achieves
efficient combustion. Because of the closed nature of the
system vith respect to vater, all hardness of the make-up vater
leaves the plant as air emissions. This particular incinerator
is also used to incinerate vaste products from a plasticizer
plant vhich is not part of this study.
In plant 53-2 a similar total quantity of process vent gas is
processed in two parallel vater scrubbers. Each scrubber consists
of a venturi contactor (600 GPM vater rate) followed by a packed
column counter-current scrubber and mist eliminator with a 2000
GPM water circulation rate.
A single gas fired thermal incinerator is used to burn waste
water from both scrubbers and other waste streams from PAN
distillation and storage facilities.
Unfortunately neither plant provided information regarding
the scrubber feed composition. Therefore, it is difficult to
make an accurate determination of control efficiency. Based
upon Table PA-1. typical material balance and vent gas emissions
shown by plants 53-1 and 53-2, it is possible to make a material
balance for this type of system. From the balance , which is
shown in Table PA-6, it appears that the overall system is very
efficient in reducing hydrocarbon emissions (96% removal) but_d.oe«
not curtail other emissions such as CO and SC^. ~Pu 61i shedda ta
on vent gas scrubbing indicates aldehyde components are difficult
to absorb.6 Testing of • water scrubber which removed 99% of
organic acids showed poor removal of aldehydes"~*"with concentrations
-------
TABLE PA-6
SCRUBBER AND WASTE INCINERATOR
FOR
130 m LB./YH. FHTHALIC ANHYDRIDE PLAOT
OVERALL MATERIAL BAEAHCE. LB./HJT
Component
partlculetes
Nitrogen Oxide*
Sulfur Oftldea
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Hithane
Ethane
Phthallc Anhydride "
Mslelc Anhydride 0>)
Organic Acids
-------
PA-20
in the effluent varying betveen 8 and 26 PPM as formaldehyde.
Operating problems associated with this type of pollution
control system include:
(1) Maleic acid is formed in the scrubber. This material
is very corrosive on concrete.
(2) Excessive emissions can occur from the scrubber in the
event there is a recirculating water pump failure.
(3) Failure of svitch valves in the reactor effluent
condenser area can overload scrubber resulting in
increased PAN emissions.
(b) Thermal Incineration
Plant 53-3 sends the process vent gas to a thermal incinerator
which was originally also to burn waste water and hydrocarbons
rejected in the PAN purification section. The unit was designed
to operate at 1400° F with a 97% combustion of organics. However,
operating problems have prevented the processing of the auxiliary
streams. In order to get on-stream reliability, it has also been
necessary to cut back operating temperature to 1200° F. Under
these conditions it is believed that only about 907o organic
combustion is obtained. Published data for a small thermal
incinerator processing phthalic anhydride unit vent gas (naphthalene
feedstock) at 1200° F_obtained 95% combustion and 97% odor
reduction efficiencyT^~"
As originally designed.plant 53-3 incorporated an extensive
amount of incinerator feed-effluent heat exchange in order to
reduce supplemental fuel requirements (effluent is presently
vented at 500° F). With excessive preheat it is possible that
the feed gas could approach the auto-ignition temperature of
phthalic anhydride.
Table PA-7 presents a material balance for a process vent gas
thermal incinerator with feed-effluent heat exchange. Heat
recovery has been limited to 55% in order to prevent pre-ignition
of feed components. In order to have essentially complete
combustion of pollutants,the incinerator data are based on a
14000 F combustion zone temperature. In addition to the concern
regarding feed pre-ignition, the following potential problems
exist with vent gas incineration:
(1) Vent gas is available at low pressure.
(2) Investment for required blowers, burning equipment and
control systems is high.
(3) Dislodged slugs of condensed PAN could lead to excessive
temperature or explosion.
(4) Because of negligible heating value of the large vent
stream (2-3 BTU/Ft.3), substantial amounts of fuel are
required in order to achieve complete combustion.
-------
PA-21
TABLE PA-7
THERMAL INCINERATOR
FOR
130 MM LB./YR. PHTHALIC ANHYDRIDE PLANT
MAIN PROCESS VENT STREAM
OVERALL MATERIAL BALANCE - LB./HR.
Component
Participates
Nitrogen Oxides
Sulfur Oxides
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Methane
Ethane
Phthalic Anhydride (a)
Maleic Anhydride (a)
Benzoic Acid (a)
Water
Total Lbs./Hr.
SCFM
Process
Vent Gas
75
2,411
8,326
407,760
101,940
368
694
45
15,333
536,952
119,300
Natural
Gas
240
2,343
517
3,100
4
10
75
121
21,986
407,997
88,049
18
35
2
21.755
540,052
120,200
(a) Particulate at low temperature.
(b) Value depends on catalyst age and
sulfur content of fuel. t Stack Gas
(c) CO emission level depends on
Furnace operating temperature.
700° F
Process Vent Gas
150° F
Natural Gas
1400° F L.
19000 F.
-------
PA-22
(c) Thermal Incineration with Steam Generation
In plant 53-4, two parallel thermal incinerators with waste
heat boilers are used to burn process off-gas. These units were
designed to run at 1400° F. During a test run on one furnace at
design conditions, 97 percent destruction of organics was
measured.
Feed gas to the incinerator is not preheated. This permits
maximum steam generation but substantially increases the fuel
requirement. Since most of the required heat is supplied by
supplemental fuel it is relatively easy to sustain combustion.
The only flame-outs that plant 53-4 has experienced have been
caused by instrument malfunction or power failure.
Except for pre-ignition and particulate carryover, potential
problems listed in Section (b) also apply for the incinerator
with steam generation.
Table PA-8 shows a material balance for this type of unit.
Commercial application of this type of control device has been
published in the literature.10
2. Other Control Devices
(a) Catalytic Incineration
A catalytic incinerator could reduce emissions to similar
levels obtained with a thermal unit. The catalytic facility
would operate at lower temperature (800-1000° F) and, therefore,
would consume less supplemental fuel. Catalytic units are used
in some phthalic anhydride plants (naphthalene feed). Even
though fuel savings may make a catalytic unit attractive, the
application of this type of incinerator is not recommended for
the following reasons:
(1) Only moderate catalyst life with possible danger of
catalyst fouling and poisoning.
(2) Limited oxidation activity. Experience with catalytic
units has not always equalled performance of the direct
flame incinerators. Reported catalytic combustion
efficiencies have been as low as 40-60 percent.**
(b) Flare Systems
(1) Substantial amount of supplemental fuel is required
( 200 MM BTU/Hr.).
(2) Efficiency for removing contaminants is less than for
other combustion devices.
(3) Improper firing of the burner could result in operating
temperatures which favor NOx formation.
(c) Boiler House
Process vent stream could possibly be used as a supplemental
air and fuel source in the plant steam boiler house. This would
be feasible if the phthalic anhydride vent is equivalent to a
-------
PA-23
TABLE PA-8
THERMAL INCINERATOR PLUS WASTE HEAT BOILER
FOR
130 MM LB./YR. PHTHALIC ANHYDRIDE PLANT
MAIN PROCESS VENT STREAM
OVERALL MATERIAL BALANCE - LB./HR.
Component
Particulates
Nitrogen Oxides
Sulfur Oxides
Carbon Monoxide
Carbon Dioxide
Nitrogen
Oxygen
Methane
Ethane
Phthalic Anhydride
-------
PA-24
small portion of the total boiler house requirement.
B. Waste Product Streams
Several of the surveyed plants (53-1 and 53-2) apparently send
all of their waste products to the process vent gas control device.
Plant 53-3 and 53-5 presently send these products off-site for
disposal. Plant 53-4 is the only surveyed facility which has a
separate control device for these wastes. In plant 53-4 ejector
exhaust and reject hydrocarbons from product fractionation are sent
to a thermal incinerator and burned at 1200° to 1600° F. A small
amount of supplemental fuel is required to maintain combustion zone
temperature. Survey data from this unit show that more than 99 percent
of the combustibles are burned. In addition to small amounts of
organic acids, CO and nitrogen oxides, the incinerator stack gas
contains some unidentified particulates (0.0001 T/T of PAN).
Table PA-9 presents a material balance for incinerating the
waste streams shown in Table PA-1.
C. Product Storage Vent
Most plants surveyed either directly vent the storage tanks
to the atmosphere (53-3 and 53-5) or send portions of this material
to incinerators previously described (53-2 and 53-4).
In plant 53-1, the vent stream passes through an uninsulated
sublimation box with a 180° return flow baffle. Flow path cross
sectional area within the box is larger than in outlet piping.
Vaporized phthalic anhydride is condensed and collected in the
sublimation box and non-condensibles are vented to the atmosphere
(110° F). Phthalic anhydride is periodically removed manually
and recycled to the process. Plant 53-1 contains eight PAN storage
tanks and each tank has one of these devices. Total amount of
phthalic anhydride removed from all boxes is about 0.0002 T/T.
D. Best Pollution Control System
The most feasible method of reducing air emissions from either
existing or new plants would be to provide either a scrubber on the
main process vent plus an incinerator for burning scrubber reject
water and waste products removed in product fractionation section or
a dual incineration system with separate incinerators for vent gas
and fractionation wastes. The vent gas unit should have heat recovery
if the steam can be utilized. These represent proven systems for
reducing hydrocarbon and particulate emissions. If it is only
necessary to minimize CO emissions, low temperature catalytic
incinerators could be considered. The small vent from product
storage facilities should either go to one of the incinerators or
be cooled for knock out of PAN.
E. Industry Research Efforts
Current industry effort in air pollution control centers
around development of a catalyst that does not require sulfur
dioxide addition for surface activation. Methods of isolation
and recovery of maleic anhydride from the vent gas stream are
also being studied.
-------
PA-25
TABLE PA-9
THERMAL INCINERATOR
FOR
130 MM LB./YR. PHTHALIC ANHYDRIDE PLANT
WASTE PRODUCT STREAM
OVERALL MATERIAL BALANCE - LB./HR.
Component
Pretreatment
Light Ends
Distillation
Reject
Natural
Gas
Combustion
Air
Particulates
Nitrogen Oxides
Carbon Monoxide
Carbon Dioxide
Nitrogen 101
Oxygen
Methane
Ethane
Phthalic Anhydride (*) 70
Maleic Anhydride («) 32
Organic Acids (•)
Heavy Ends
Water 4.800
Total Lbs./Hr. 5,003
SCFM
16
151
33
10,257
3,103
Stack
Gas _
2
2
40
2,262
10,373
1,275
101
544
69
75
789
200
5
5,538
19,497
4,900
(a) Particulate at low temperature.
(b) 1200° F for pretreatment light
ends and 1600° F for intermittent ptack Gas
feed of distillation reject ,1400° F
liquid. ;
(c) S02 content depends on sulfur
content of fuel.
Pretreatment
Light Ends
400° F
Distillation
Reject (2700 F)
(b)
1400° F
•Combustion
Air (80° F)
.... -Natural Gas
-------
PA-26
V. National Emission Inventory
Based upon the emission factors shown in Table PA-4, total approximate
emissions from U. S. manufacture of PAN from o-xylene are as follows:
Average Emissions (a) Total Emissions '
Component T/T of PAN MM Lbs./Year
Hydrocarbons 0.0001
Particulates (e) 0.0103
NOX 0.0006
SOX (f) 0.0051
CO 0.0872
0.1033
It is estimated that about 10% of the PAN produced from xylene is in
units without emission control facilities. If control devices were added to
these units,total national emissions would be slightly reduced, see Table PA-15
in Section XII.
It should be noted that PAN production is rather constant throughout the
year. As a result there is no seasonal variation in emissions.
(a) Weighted average based on individual surveyed plant emission factors and
PAN production.
(b) Based on 250,000 tons/year PAN production from o-xylene.
(c) Represents xylene lost from feed storage.
(d) Includes 0.0001 T/T of PAN emitted from product storage tanks Cassumes 507.
of product storage vents directly to atmosphere).
(e) Includes PAN, MAN and organic acids.
(f) SOx level based on relatively fresh catalyst.
-------
PA-27
VI. Ground Level Air Quality Determination
Table PA-4 presents a summary of air emissions data for the various
phthalic anhydride (from o-xylene) plants surveyed. This table includes
emissions from the main process and product fractionation vent streams.
This summary table also includes operating conditions and physical dimensions
of the various vent stacks. The EPA vill use this information to calculate
ground level emissions concentration for future reporting.
-------
PA-28
VII. Cost Effectiveness of Controls
Table PA-10 presents a cost analysis for alternative methods of reducing
air emissions from the various vents. Economic data presented are for a
new plant producing 130 MM Ibs./year of PAN and are based on the following
(in 1973 dollars):
A. Investment
Investment costs were derived from cost data provided in plant
surveys plus published data 8 and vendor quotes for similar equipment.
Installation costs were primarily obtained from the plant survey data.
B. Operating Expense
1. Depreciation - 10 year straight line.
2. Interest - 670 on total capital.
3. Maintenance - costs were primarily based on survey data.
Corrosive properties of vent streams tend to increase required
maintenance.
4. Labor - plant survey data.
5. Utilities - unit costs are based con current (1973) typical
values for the Gulf Coast area. After existing contracts
expire, it is possible that fuel gas cost will rise considerably
above figure used in this study.
All of the vent gas emission control devices will increase system pressure
drop (10-14 in. H20). Blower investment and operating costs for providing
this &P are included in the cost analyses.
Because of the large volume of vent gas to be processed, it has been
assumed that parallel equipment would be employed. This provides for more
operating flexibility and follows the practice normally applied in existing
plants.
Table PA-10 shows, for the main process vent gas, that there is not much
economic difference between the various emission control methods studied.
However, it has been assumed that with a water scrubber^incinerator system,
all fractionator waste products can be charged to the waste water incinerator.
In installations without a water scrubber, it has been assumed that a separate
liquid waste product incinerator will be required so as to minimize control
problems. Therefore, the combined operating cost of the water scrubber-
incinerator system is the lowest, while the direct incineration system requires
the lowest capital and the incineration system with a. waste heat boiler results
in the best heat utilization (providing that the steam can be utilized). The
scrubber system has the highest carbon monoxide emissions but is approximately
equivalent in the rate of emissions of other materials.
Because of the low heating value of the process vent gas (2-3 BTU/SCF),
variations in stream composition have limited effect on incinerator operation
and economics.
-------
TABLE PA- 10
COST EFFECTIVENESS FOR ALTERNATE
EMISSION CONTROL DEVICES
(BASED ON 130 MM LBS./YR.
Water Scrubber +
2
50
536,962
119,300
0.0756 (•)
0.0047
0,1507
544,005
122,100
0.0036
0.0047
0. 1507
86
96 (organic*)
275.000
825,000
1,100,000
110,000
66,000
55,000 (5%)
6,500
25,000
1,100
26,100
263,600
420,100
Main Process
Incinerator
1
100
10,460 (b)
0.1221
39,235
9,700
0.0009
0.0002
0.0025
98
99
120,000
230.000
350,000
35,000
21,000
35.000 (107.)
5,000
5,000
55,500
60,500
156,500
PHTHALIC ANHYDRIDE PRODUCTION)
Vent Gas
Direct Incineration (8)
2
50
536,962
119,300
0.0692
0.0047
0.1507
540,052
120,200
0.0036
0.0006
0.0047
0.0076
95
92
575,000
285,000
860.000
86,000
51,600
34,400 (47.)
5,000
19,800
198,300
218,100
395,100
395,100
Incineration 4- Vaste Heat Boiler
2 2
50 50
536 962
119,300
0,0692
0.0047
0.1507
545.767
122.350
0.0036
0.0012
0.0047
0.0076
95
92
625.000
625 . 000
1,250,000
125,000
75,000
50,000 (47.)
20,000
562,300
34,200
596,500
866,500
(465,000)
401,500
V'arte Product?
Direct
5.79?
0.0557
19,497
4.900
0.0004
0.0001
0.0025
97
99
85,000
65,000
150.000
,15.000
9.000
7,500
3.000
1.000
12.800
13.800
48,300
48,300
Incineration
1
100
(57.)
Stream
Type of Emission Control Device
Number of Units
Capacity of each Unit - %
Feed Gas
Total Flow - Lbs./Hr.
SCFM
Composition - Ton/Ton Pan
Hydrocarbons
Particulatas (incl. PAN, MAN & Org. Acids)
NO,
Carbon Monoxide
Gaseous Effluent
Total Flow - Lbf./Hr.
SCFM
Composition - Ton/Ton PAN
Hydrocarbons
Particulates (Incl. PAN, MAN & Org. Acids)
NOX
S0x
Carbon Monoxide
Emissions Control Efficiency ' '
CCR
SERR
SE
Investment - $
Purchased Cost
Installation
Total Capital (e)
Operating Cost - $/Yr.
Depreciation (10 years)
Interest on Capital (67.)
Maintenance
Labor - $4.85/Hr.
Utilities and Chemicals
Power - Ic/KVH
Fuel - 40c/MM BTU (c)
Process Water - 10c/M Gal.
toller Feed Water - 30C/M Gal.
Total Utilities and Chemicals
Total Operating Cost
Steam Production - 59c/M Lbs. (450 PSIG, 750° F)
Net Annual Cost - $/Yr.
(a) Includes 0.0064 T/T of organic material contained in separate liquid reject stream from product fractlonation system ejector (Table PA-1. Stream 5).
(b) Liquid rejected from scrubber system plus light and heavy ends removed in product fractionatlon (Table PA-1, Streams 6, 7 and 8).
(c) It is possible that future fuel cost will be considerably higher than figure used In this comparison.
(d) Emission control efficiencies are defined by the equations given below. For further details, see Appendix III.
CCR • pounds of 02 that react with pollutants to feed device
pounds oT
_
that tneorceteally could react vlth these pollutants
x 100
SERR - veighted pollutants in - weighted pollutants out
weighted pollltantf In
x 100
SE - specific pollutant In - specific pollutant out
specific pollutant in
100
(e) Developed from 1970-1971 coat figure* provided by PAN manufacturer with 10-15 percent added for escalation to 1973 costs.
(f) Shown at fuel plus BFW coat since thia steam only replaces operating cost of stand-by boilers.
(g) With feed preheat.
-------
PA-30
It is estimated that cost for installing the various pollution control
equipment in existing plants vould be about the same or only slightly
higher than for new plant installations shown in Table PA-10. The actual
cost difference would depend on space availability and location in relation
to associated process equipment.
-------
PA-31
VIII. Source Testing
It is recommended that source sampling should be performed on plant
53-4 off-gas incinerator feed and effluent streams. The testing of this
unit has one short-coming in that it is overdesigned, in order to handle a
possible future 50 percent increase in plant capacity. The plant has four
parallel reactor trains which feed two parallel incinerators. There is
cross-over piping up-stream of the incinerators. For short periods of time,
the plant can shut down one reactor train and one incinerator and thus
simulate 100 percent operation.
In order to determine emissions from plants that incorporate waste
gas scrubbing followed by incineration, it is recommended that additional
source sampling should be performed at plant 53-1. Scrubber feed plus stack
gases from the scrubber and incinerator should be analyzed. Two of the
surveyed plants have this type of emission control system. However, plant
53-1 is the only one which has analyzed the vent streams.
-------
PA-32
IX. Industry Grovth Projection
The U. S. annual phthalic anhydride production is estimated to increase
to 2.2 billion pounds by 1985, see Figure PA-2. This represents a 1.3
billion pound per year increase over the 1972 production level (7% increase
per year). It is projected that most if not all of this increase vill be
by catalytic oxidation of o-xylene. Plant investment is 15-20% lover for a
plant designed to use naphthalene feed
-------
PA-33
o
u
z <
5*
S1
U
0.
1000
e
a
U I- -
355
S < >
j n
•••••••••MB**••••••*••••••
••••••••••••••••••' * *»m*-
m m
s
d
100
(71
1960
1964
1968
YEAR
1972
1976
1980
1984
-------
PA-34
X. Plant Inspection Procedures ,,
Plant inspections vill be conducted by the appropriate authorities, either
on a routine basis or in response to a complaint. Usually the inspecting
agent vill only be able to make visual observations. In some instances stack
monitoring equipment may be available or it might be possible to sample the
stack through an accessible sample point. The odor of phthalic anhydride and
maleic anhydride vill sometimes be experienced in the vicinity of PAN plants
that do not incorporate emission control facilities or have control devices
that are not operating properly.
If the inspector has reason to suspect that emissions are excessive, some
factors that he should consider and/or discuss vith plant officials are
itemized belov:
A. During start-ups, some plants directly vent reactor effluent to the
atmosphere until the reactors reach operating temperature. Hot air
is used for this warm-up. Precautions should be taken to prevent
addition of hydrocarbon feed to the reactors until after the direct
venting has been terminated.
B. Proper operation of the switch condensers is essential to limit PAN
losses. Switching frequency and effluent temperature are the main
condenser operating variables which can be adjusted for controlling
these losses. Switching valves associated with the condensers are
potential sources of leakage and resulting air emission of PAN.
C. When scrubbers are provided on the main process vent stream, adequate
make-up and circulating water rates are necessary to insure efficient
removal of hydrocarbons, vapors and particulates. In addition gas
feed rate plus inlet and outlet temperatures influence emissions.
Most plants will keep a record of some or all of these operating
variables and their design limits.
D. Vhen the process vent gas is burned, proper operating of the combustion
device is essential if emissions are to be minimized. Two types of
problems may be expected to be encountered, (1) flame-outs and (2)
excessive smoking. In units employing extensive feed-effluent heat
exchange, it may also be possible to have pre-ignition of the feed.
Plants are likely to periodically record some or all of the following
operating variables. Data will also be available on design limits.
(1) Combustion zone temperature.
(2) Composition and flow rate of feed to the device.
(3) Quantity and heating value of supplemental fuel.
(4) Composition of stack gases.
E. Incinerators employed on the waste water and reject hydrocarbon streams
are sometimes required to handle wide variations in flow rate. Operating
problems are normally limited to flame-outs and smoke production. In
addition to the operating variables listed for the process vent gas
incinerator, the quantity of air used for burning is important in
controlling waste product incinerator performance. Too little excess
air will cause smoking and too much can result in flame-outs. The
quantity of excess air might be indicated by measurements on one or
-------
PA-35
all of the following:
(1) Device draft - inches of water.
(2) Temperature of stack gases.
(3) Air flow rate.
In order to limit emissions (PAN and MAN), temperature of the atmospheric
vent gas streams from product fractionation facilities should not
exceed design values.
-------
PA-3 6
XI. Financial Impact
Table PA-11 presents economics for phthalic anhydride manufacture in a
typical 130 MM Ibs./year plant that incorporates virtually no air pollution
control equipment. Based on present cost/price levels the production of PAN
from o-xylene in this type facility appears to be profitable (ROI - 10.07o).
Table PA-12 provides economics for producing PAN in an existing unit
which has been modified to reduce emissions. Modifications consist of
adding a process vent gas scrubber plus an incinerator for burning scrubber
reject liquid and waste products removed in product fractionation. Data
are also presented for an alternate system using dual thermal incinerators
on the process vent and waste product streams.
Table PA-13 provides similar data for a new most feasible unit which
also incorporates thermal incinerators. However, in this case a waste
heat boiler is used on the process vent incinerator effluent for heat
recovery in place of feed-effluent heat exchange. In addition, the cost
of emission control facilities on product storage tanks have been included
(approximately $20,000). This control consists of either piping the vents to
one of the incinerators or providing sublimation boxes for removal of PAN.
The economic data indicate about a five percent increase in PAN production
cost for both of the modified and new most feasible units. Assuming a PAN
selling price of 8.8c/lb., this corresponds to a 13 percent reduction in
profit and reduces return on investment to about 6.8 percent.
Table PA-14 presents pro forma balances for the above cases. It was
assumed in developing these asset and liability positions that phthalic
anhydride selling price would be held constant and any increase in production
costs would be taken out of profit margin in order to maintain sales volume.
Capital requirement for the most feasible new plant is about 1.5 million
dollars higher than for an existing type plant without emission control.
In addition to financial impact, it is essential to evaluate the overall
environmental impact of the most feasible method of emission control. A
factor for consideration in this evaluation is the effect on fuel compsumption
All of the proposed emission control systems involve incineration. Since
the heating value of the main vent stream is very low, this incineration
requires supplemental fuel in all cases. If vent streams are burned and waste
steam is generated in the plants producing the 1.3 billion pounds per year
increased PAN production between 1972 and 1985, the incineration will consume
supplemental fuel energy equivalent to approximately 14.3 billion standard
cubic feet per year of natural gas. Assuming a thermal efficiency of 87
percent for steam generation, fuel credit of net steam produced would be
equal to 10.6 billion standard cubic feet per year of natural gas. This
results in a net increase in fuel consumption of 3.7 billion SCF per year.
If steam can not be utilized, incineration with feed-effluent heat exchange
or a combination scrubber incinerator could be employed. Assuming all new
plants incorporate incinerators with feed-effluent heat exchange, the pollution
control would consume fuel equivalent to 5.3 billion SCF per year of natural
gas. If all new plants employed the scrubber-incinerator type of control,
fuel usage would be reduced to 1.4 billion SCF per year.
-------
PA-3 7
TABLE PA-11
PHTHALIC ANHYDRIDE MANUFACTURING COST
FOR A TYPICAL
EXISTING 130 MM LB./YR. FACILITY (a)
DIRECT MANUFACTURING COST
Rav Materials
95.5% o-xylene @ 4%c/lb.
S02 @ 5^c/lb.
Labor (3 men/shift @ $4.85/hr.)
Maintenance (5% of investment)
Utilities (includes catalyst)
INDIRECT MANUFACTURING COST
Plant Overhead (110% of labor)
FIXED MANUFACTURING COST
Depreciation (10 years)
Insurance & Property Taxes (2.37* of inv.)
MANUFACTURING COST
GENERAL EXPENSES
Administration (3% of manufacturing cost)
Sales '1% of manufacturing cost)
Research (2% of manufacturing cost)
Finance (6% of investment)
Total Cost
Selling price
Profit before taxes
Profit after 52% tax
Cash flov
ROI (NPAT x 100/investment)
C/LB.
4.34
0.03
0.10
0.23
0.60
5.30
0.11
0.77
0.18
0.95
6.36
0.19
0.06
0.13
0.46
0.84
7.20
8.80
1.60
0.77
$/YR.
10.0%
9,360,000
11,440,000
2,080,000
998,500
1,998,500
(a) Primarily based on published data. 11,12,13
-------
PA-3 8
TABLE PA- 12
PHTHALIC ANHYDRIDE MANUFACTURING COST
FOR A TYPICAL
EXISTING 130 MM LB./YR. FACILITY
WITH AIR POLLUTION CONTROL EQUIPMENT
Type of Emission Control Scrubber & Inciner ator Dual ^ncineratgrs
DIRECT MANUFACTURING COST C/LB.
-------
PA-39
TABLE PA- 13
PHTHALIC ANHYDRIDE MANUFACTURING COST
FOR A TYPICAL MOST FEASIBLE
NEW 130 MM LB./YR. FACILITY
DIRECT MANUFACTURING COST
Raw Materials
95.57. o-xylene @ 4^c/lb. 4.34
S02 @ 5^c/lb. 0.03
Labor 0.12
Maintenance 0.27
Utilities (includes catalyst) 0.71
5.47
INDIRECT MANUFACTURING COST
Plant Overhead (110% of labor) 0.13
FIXED MANUFACTURING COST
Depreciation (10 years) 0.88
Insurance & Property Taxes (2.3% of Inv.) 0.20
1.08
MANUFACTURING, COST 6 . 68
GENERAL JffPENSES
Administartion 0.19
Sales 0.06
Research 0.13
Finance (67. of Investment) £i.53_
0.91
Total Cost 7.59 9,867,000
Selling Price 8.80 11,440,000
Profit Before Taxes 1,573,000
Profit After 527,, Tax 755,000
Cash Flow 1,897,000
ROI (NPAT x 100/Investment) 6.670
ROI Sensitivity
Increased Capital Charges (b) 4.7
Increased Operating Cost (c) 4.0
(a) Economics for pollution control equipment obtained from Table PA-10.
(b) Capital charges for pollution control equipment double values shown in
Table PA-10.
(c) Capital charges and new operating cost for pollution control equipment
double values shown in Table PA-10.
-------
TABLE PA-14
Type of Unit
Type of Emission Control
Current Assets
Cash (A)
Accounts Receivable (B)
Inventories (C)
Fixed Assets
Plant
Building
Land
Total Assets
Current Liabilities (D)
Equity & Long Term Debt
Total Capital
130 MM
Existing
689,000
953,350
1,080,000
10,000,000
100,000
50,000
12,872,350
627,250
12 ,.245 , 100
12,872,350
PRO FORMA BALANCE SHEET
LB./YR. PHTHALIC ANHYDRIDE
Scrubber & inc.
720,400
953,350
1,134,000
11,450,000
100,000
50,000
14,407,750
644,600
13,763,150
14,407,750
FACILITY
Modified
Existing
Dual Incinerators
723,650
953,350
1,135,500
11,010,000
100,000
50,000
13,972,500
652,150
13,320,350
13,972,500
Most Feasible
New Plant
Dual Incinerators
723,650
953,350
1,138,500
11,420,000
100,000
50,000
14,385,500
647,850
13,737,650
14,385,500
(A) Based on one month's total manufacturing cost.
(B) Based on one month's sales.
(C) Based on 15 MM Ibs. of product valued at total cost.
(D) Based on one month's total cost less fixed manufacturing and finance costs.
-------
PA-41
XII. Cost to Industry
As indicated in Section IV, most of the plants surveyed incorporate
pollution control devices. Cost of this equipment represents up to 15%
of the total plant investment. This expenditure plus associated operating
costs equals about 5% of the total PAN production cost (0.35c/lb •) •
Total capital cost for adding a combination scrubber-incinerator system
or dual incinerators 'to the one or tvo small existing units vithout emission
control would be about 1.0 to 1.5 million dollars. In the "most feasible
nev plant" presented in Table PA-13, air emission control equipment represents
13% of the total plant investment. Assuming all nev phthalic anhydride
plants built between 1972 and 1985 incorporate this type of control equipment,
the total incremental capital cost for these plants vill be about 13 million
dollars.
Since most of the existing plants include emission controls similar
to those vhich are proposed, universal use of these controls should not
reduce growth in demand for PAN.
The projected effect of the above expenditures on future air emissions
is shown in Table PA-15.
It should be noted that emission levels shown in this summary exclude
sulfur oxides resulting from sulfur compounds in incinerator fuel. This is
because existing plants use natural gas which contains little or no sulfur.
In the future, it is very likely that low sulfur fuels will be unavailable
or in short supply giving an added incentive to use a pollution control
system with low fuel requirements. In this regard, even though the combination
scrubber incinerator system does not remove CO from the process vent stream,
overall weighted emissions for this system could be lower than for other types
of control if low sulfur fuel is not available.
-------
TAPLE PA-15
ESTIMATED 1985 AIR EMISSIONS
FOR
ALTERNATE CONTROL SYSTEMS
Sheet 1 of 2
Type of Pollution Control
PAN Production,
Average
Emissions
T/T
0.0001
0.0049
0.0006
0.0051
0.0872
0.0979
Scrubber & Incinerator
900 , 000
Total
Emi ss ions
MM Lbs. /Yr.
0.2
8.8
1.1
9.2
157.0
176.3
Vei phted
Emissions
20
530
40
180
160
930
(A) Estimated production frotr o-xylene.
(B) It Is assumed that 107= of total production Is In plants vithout pollution control facilities.
(C) Modification consists of adding devices shovn to units vithout pollution control equipment.
(D) Significant Emission Index, which is based on the following weighting factors: Hydrocarbons « 80. particulate? = 60 NOX
details, see Appendix II.
(E) Includes PAN, MAN and organic acids.
(F) Based on fresh catalyst and low sulfur fuel.
40, SO = 20 and CO = 1. For further
-------
TABLE >A-15 (CONTINUED^
ESTIMATED 1985 AIK EMISSIONS
FOR
ALTERNATE CONTROL SYSTEMS
Sn-'et 2 of 2
Type of Pollution Control
PAN Production, ^ Tons/Yr.
Hydrocarbons
Particulates (E)
NOX
' SOX (F)
CO
Most Feasible Modifications (G>
Existing Plants
250,000
Average Total
Emissions Emissions
T/T MM Lbs./Yr.
0.0001
0.0049
0.0006
0.0051
0.0872
0.0979
0.1
2.5
0.3
2.5
43 6
49.0
New Plants
650,000
Average
Emissions
T/T
0.0001
0.0040
0.0013
0.0047
0.0101
0.0202
....
Total
Emi ssion
MM Lbs./Yr.
0.1
5.2
1.7
6.1
13 I
26.2
— -
Total
Emi FFi onp.
MM I.bs. /Yr
0.2
7 7
20
8.6
56 7
75.2
Total
900.000
Vei fthted
Emi FF i on?
20
460
80
170
60
790 "»
(A) Estimated production from o-xylene.
(B) It is assumed that 107,, of total production Is in plants without pollution control facilities:.
(C) Modification consists of adding devices shown to units without pollution control equipment
(D) Significant Emission Index which is based on the following weighting factors: Hydrocarbons = 80, Particulatec = 60, NOX = 40. SOX = 20, and CO = 1 For further
details see Appendix II.
(E) Includes PAN, MAN and organic acids.
(F) Based on fresh catalyst and low sulfur fuel.
(G) Scrubber and incinerator added to existing units without pollution control. New plants based on dual incinerator with waste heat hollers and product storage tank
vent controls.
-------
PA-44
XIII. Emission Control Deficiencies
Technical deficiencies which hinder reducing the level of emissions
include the following:
A. Process Chemistry and Kinetics
In this vapor phase o-xylene oxidation process, a carrier-supported
vanadium pentoxide catalyst is used to produce PAN. Experimental data
indicate that formation of PAN, by-product maleic anhydride and carbon
oxides are zero order reactions with respect to o-xylene.14 phthalic
anhydride is also produced in a two step reaction involving o-tolualdehyde
as an intermediate product. These particular reactions are first order
with respect to o-xylene and o-tolualdehyde, respectively. All of the
reactions appear to be independent and show a square-root dependence on
oxygen pressure. It is believed that the reactions occur on the
catalyst surface, involving transfer of catalyst oxygen to the
chemisorbed xylene followed by desorption of the products and catalyst
oxidation by gaseous oxygen. The catalyst oxidation step is rate-
determining.
The amount of phthalic anhydride produced is influenced by the
o-xylene feed purity, oxygen concentration, reactor residence time and
other reactor operating conditions.
1. Reactor Feed
(a) Ortho-xylene
Feedstock normally contains 95 - 96 vt. °L o-xylene. Meta
and para-xylene are the primary impurities. These compounds and
any other hydrocarbon impurities are converted to carbon oxides.
Therefore, in order to minimize air pollution and maximize
productivity it is desirable to use high purity feed. Xylene
concentration in total reactor feed is set at 1.0 vol. 7« or
less in order to stay below the lower explosion limit.15
(b) Oxygen Concentration
Air is primary source of oxygen used in phthalic anhydride
production. All surveyed plants producing PAN from xylene use
air exclusively, vhereas some naphthalene based units, incorporate
air plus supplemental pure oxygen. (See Appendix I)
In addition to influencing the rate of xylene conversion,
oxygen concentration controls product distribution. At low
air-xylene weight ratios (1.5 - 5.0), tolualdehyde would be
primary product. Tolualdehyde concentration in reactor product
is insignificant at the higher air to xylene ratios normally
employed (30 - 34).
fc) Sulfur Dioxide
A small amount of sulfur dioxide is used to increase
catalyst activity. The addition rate varies over the catalyst
cycle (.006 - 0.0125 T/T of PAN)..
-------
PA-45
2. Reactor Operating Conditions
Reactor operating conditions influence xylene conversion
rate and the amount of non-selective products. Since unconverted
xylene would be lost in the process vent gas stream, operating
conditions are adjusted to obtain complete xylene conversion.
In the U.S., where raw materials in the past have been relatively
inexpensive, the reactors are operated at higher temperatures
than are used in Europe.16 This results in somewhat lower
yields but at the same time reduces reactor catalyst volume
and plant investment.
3. Catalyst
With the vanadium pentoxide catalyst presently employed,
about one pound of PAN is produced per pound of xylene feed
consumed. Maximum theoretical yield is 1.4 pounds per pound
of o-xylene. Therefore, selectivity is about 75 mol 7<>.
Non-selective material contributes to CO and hydrocarbon
(particulate) emissions in the main process vent gas. In
addition small amounts of waste by-products are produced.
Removal and disposal of this material can also result in
air emission problems.
B. Process Equipment and Operations
1. Reactors
The chemical reaction for PAN production by partial
oxidation of xylene is highly exothermic. It would appear that
fluidized bed reactors could be used for this reaction. However,
successful commercial application of these reactors has been
limited to PAN plants using naphthalene feed. If fluidized
bed units were used in place of the conventional tubular fixed
bed converters, it would be possible to increase the concentration
of xylene in the reactor feed and thereby reduce the volume of
vent gas. With fluidized reactors it is possible to operate
within the flammability limits without problems because of the
inerting effect of catalyst dust particles and the ability of
the fluid bed to dissipate reaction hot spots.13,17
2. Switch Condensers
The cyclic operation of these condensers requires the
utilization of many valves. Improper maintenance or failure of
these valves can result in leakage of PAN and heat transfer
fluid into the atmosphere or the pollution control equipment.
C. Control Equipment and Operations
1. Scrubbers
Failure of the water circulating pumps can result in
atmospheric venting of PAN and other hydrocarbons normally
removed from the main process vent gas.
-------
PA-46
2. Incinerators
As previously noted there have been operating problems
associated with the one surveyed commercial vent gas incinerator
that incorporates feed-effluent heat exchange. At times these
problems necessitate direct atmospheric venting of the incinerator
feed gas.
Thermal incineration with waste heat steam generation is a
dependable and probably most feasible method of controlling
main process vent gas emissions. However, all types of
incinerators require a substantial amount of supplemental fuel.
If liquid fuel is employed, stack gas emissions could increase
as a result of SOx formation and ash carry-over. In addition
all incinerators produce some NOx. Scrubber reject water
incinerators also release the water hardness as particulate air
emissions.
-------
PA-47
XIV. Research and Development Needs
If the technology deficiencies discussed under Section XIII are to be
overcome, additional R & D is desirable in the folloving areas:
A. Existing Plants
1. Improved Catalyst
It would be desirable to have a more selective catalyst in order
to reduce air emissions and produce less unwanted by-products. With
the present commercial operation, xylene to PAN selectivity is
approximately 75%.
It would also be desirable to develop a catalyst that does not
require the addition of sulfur dioxide as a catalyst activator.
Catalyst development work in these areas can best be handled
by the process licensors.
B. Nev Plants
1. Oxygen Feed plus Vent Gas Recycle
Only a small portion of the oxygen contained in the reactor feed
is consumed. Therefore, it should be possible to reduce net emissions
by recycling a portion of the reactor vent gas and by using oxygen
enriched air for make-up.
2. By-Product Recovery
The literature indicates that at least one phthalie anhydride
process licensor offers technology in the recovery of by-product
maleic anhydride.13 Development work and engineering studies would
be necessary to see if by-product recovery is practical. From an
air pollution standpoint, this investigation is not critical since
the PAN process vent streams contain other components which would
not be recovered and would still require clean-up.
3. Reactor Modification
If fluidized reactors vere used in place of fixed bed units, it
would be possible to reduce the volume of vent gas. However, most
previous attempts at processing o-xylene in fluidized converters have
resulted in unsatisfactory yields and excessive by-product
production.^
It should be noted that much of the above suggested R & D probably has
been or is being done by industry.
-------
PA-48
XV. Research and Development Programs
The following proposed programs are for projects within the general
R & D areas listed in Section XIV. These programs are limited to those
projects which would have a good change of success for obtaining methods of
reducing emissions from future phthalic anhydride manufacture. In preparing
these programs, it has been assumed that the researcher has prior experience
or knowledge in PA production.
Project A
1. Title - Oxygen Feed plus Vent Gas Recycle for reduced emissions.
2. Object - Scope the feasibility of reducing emissions by air and
vent gas recycle.
3. Project Cost (See Table PA-16 for Cost Breakdown)
Capital Expenditures $100,000
Operating Costs
Total Manpower 97,200
Services 6,100
Materials 4,000
Contingency 25.000
Total $232,300
4. Scope - On a laboratory scale modify the conventional fixed bed process
by replacing part of the air feed with pure oxygen and by recycling
process vent gas for dilution. These process changes will be studied
in a small pilot plant reactor coupled to an on-line gas chromatograph.
Successful completion of the project could lead to a pilot plant
demonstration utilizing the same pilot plant equipment.
5. Program
(a) Design, Construction and Checkout
This part of the project is concerned with the design, fabrication
and start-up of a laboratory scale unit. The unit will include
effluent condensation and vapor recycle system in order to simulate
equilibrium closed loop operation. Effluent gases from the reactor
and product condenser will be analyzed by an on-line gas chromatograph.
(b) Process Development
The effect of process operating conditions, oxygen partial
pressure and recycle gas rate on conversion, PAN selectivity and
composition of effluent gases will be studied. The standard
vanadium pentoxide catalyst will be used.
(c) Process Engineering
Data from the process development work will be used to design a
model for the xylene oxidation process. This model will define
optimum process parameters for maximum PAN production at low level
of emissions.
-------
PA-49
TABLE PA-16
DETAILED COSTS
FOR
R & D PROJECT A
PILOT UNIT DESIGN CONSTRUCTION & CHECKOUT
Design Manpower: Professional - 6 veeks 5,600
Technician - 12 veeks 6,200
Major Equipment, Installed 100,000
Contingency 15.000
126,800
PROCESS DEVELOPMENT
Operation
Manpover: Professional - 14 veeks 13,000
Technician - 2 men/shift, 3 shifts/day for 14 vks. 60,200
Services: Analytical - 150 hours 2,200
Computational 2,400
Materials 4,000
Contingency 8,000
89,800
ENGINEERING
Process Design and Economic Evaluation
Process Engineer - 20 weeks 12,200
Services: Computational 1,500
Contingency 2,000
15,700
-------
PA-50
6. Timetable
The overall time required for this project including pilot plant
construction, unit operations and engineering evaluation is
estimated to be 13 months (excludes equipment delivery time).
Project B
1. Title - Application of Fluidized Reactors in Phthalic Anhydride
Production from 0-Xylene
2. Object - Scope the use of fluidized reactors to determine if the
volume of vent gas to the emission control facilities can be reduced
while maintaining high feedstock utilization.
3. Project Cost (See Table PA-17 for Cost Breakdown)
Capital Expenditures $ 70,000
Operating Costs
Total Manpower 89,000
Services 21,500
Materials 4,000
Contingency 2 5,OOP
Total $210,500
4. Scope - This project will seek to reduce the volume of vent gas by
incorporating fluidized bed reactors. By using oxygen enriched air
for reactor feed, it may be possible to further reduce the quantity
of vent gas and also lower emissions and reduce production of by-
products. A small pilot plant will be constructed. This unit will
be connected to ah on-line gas chromatograph. Data from the pilot
unit will be used to develop a process model.
5. Program -
(a) Design, Construction and Checkout
The first phase of the program will be the design, fabrication
and start-up of a small pilot unit which consists of a fluid bed
reactor and feed delivery and preheat facilities. Effluent gases
from the reactor will be analyzed by gas chromatograph.
(b) Process Development
The effect of process operating conditions, oxygen partial
pressure and catalyst on conversion, PAN selectivity and composition
of effluent gases will be studied. The standard vanadium pentoxide
type of catalyst will be used initially. Catalyst formulation and
catalyst activators will be modified in an attempt to improve
selectivity. Adjustments of physical properties of the catalyst
(e.g., pore volume distribution, surface pH, total surface area)
will also be studied.
-------
PA-51
TABLE PA-17
DETAILED COSTS
FOR
R & D PROJECT B
Pilot Unit Design Construction & Checkout
Design Manpover: Professional - 6 veeks 5,600
Technician - 12 veeks 6,200
Major Eauipment 70,000
Contingency 15,000
96,800
Process Development
Unit Operation
Professional - 45 weeks 41,800
Technician - 45 weeks 23,200
Services
Analytical - 4weeks 2,400
Catalyst Preparation - 25 veeks 13,000
Physical & Catalyst Testing - 10 veeks 4,600
Materials 4,000
Contingency 9.OOP
98,000
Engineering
Process Design and Economic Evaluation
Process Engineer - 20 weeks 12,200
Services: Computational 1,500
Contingency 2,000
15,700
-------
PA-5 2
Cc) Process Engineering
Data from the process development vork on the most promising
catalyst vill be used to prepare a model for the modified process.
This model vill define optimum process parameters for maximum PAN
production at low level of emissions.
6. Timetable
The overall time reauired for this project including pilot plant
construction, catalyst formulation, unit operations and engineering
evaluation is estimated to be 18 months.
-------
PA-53
XVI. Sampling, Monitoring and Analytical Methods for Vent Streams
A. Methods in Use
Of the five phthalic anhydride plants submitting replies to
questionnaires, four had performed some type of analyses of stack
emissions. Plant number 53-1 sampled for particulates from the
product condenser scrubber vent by isokinetic sampling into
impingers. The analytical method is unknown, but it was stated
that the results represented inorganic salts. Organic acids were
collected in an unknown manner for analysis by flame ionization
gas chromatography, while sulfur dioxide was measured by the modified
Shell Development technique.18 plow was measured by Pitot traverse
and moisture content by wet bulb/dry bulb thermometers.
The same plant measured particulate emissions from the scrubber
water incinerator using a probe and impingers. Velocity and moisture
content were measured as stated above. The particulate emissions
were described as MgC03 and CaC03 originating from the scrubber
feed water. Carbon monoxide and carbon dioxide were measured by
Orsat apparatus and organics by flame ionization.
Limited sampling has been conducted by plants 53-3 and 53-5.
One plant (53-3) determined total organic carbon (TOG) as methane,
CO, C02 and benzoic acid by flame ionization. The samples were
collected from the condenser and purification exhaust incinerator
stack using rubber bladders. Flow was calculated by a material balance.
At the second plant, (53-5), the main process vent gas stream was
sampled for particulates, carbon monoxide and hydrocarbons using a
train consisting of an Alundum thimble followed by impingers. The
methods were not further described except that the sample gas leaving
the impingers was analyzed for hydrocarbons by hydrogen flame gas
chromatography.
Plant number 53-4 determined emissions from both the process vent
and waste product incinerators. Particulates from both units were
determined by EPA Method 5 19, except that non-isokinetic sampling
was used on one stack. EPA Method 7 19 was used for oxides of nitrogen
on both stacks. Organics were collected in impingers containing
deionized water and analyzed for TOC as methane. Orsat analyses were
performed for CO, C02, 02 and N2- The flow in one stack was measured
by Pitot traverse, but in the other a material balance was used.
B. Discussion
The methods in use seem to be quite diverse. With the exception of
particulates and hydrocarbons, however, accepted techniques are
available. Phthalic anhydride is a solid at 266° F and boils at
543° F. The stack temperatures from the various plant units ranged
from a low of 113° F to a high of 530° F. The product material being
sampled thus ranged from a solid to a very high vapor pressure liquid.
Since other volatile by-products may be present, the organic materials
being sampled probably consist of a mixture of solids, liquids and
vapor. Thus, the distinction between particulates and hydrocarbons
would be highly dependent on the form and temperatures maintained in
a sampling train.
-------
PA-54
The EPA Method 5 train, consisting of heated probe and filter, would
probably pass a large portion of phthalic anhydride as vapor. If an
analysis of impinger contents were added, the split between filter
and impinger would still be uncertain and dependent on the state of
the pollutants in the stack, the probe and hot box temperature, and
the duration of sampling. It appears, therefore, that a different
approach may be required for meaningful analyses. A procedure based
on direct impingement and specific analyses of the impinger contents
may provide a meaningful basis for emissions regulations.
C. Future Methods Development
It is recommended that a specific sampling and analytical method
be developed for characterization of emissions from phthalic anhydride
plants. This method should be keyed to any regulations concerning
emissions limitations.
-------
Plant
Scrubber Vent
53-1
53-2
Organic Acids
Participates
SO,
TABLE PA-18
SUMMARY OF
SAMPLING AND ANALYTICAL METHODS
Method
Make
Model
Column Dimensions
Column Packing/Absorbent
Plane lonlzatlon
Isoklnettc Sampling Into Implngers
Modlf. Shell Development
None
Main Process Vent
53-5
Incinerator
53-3
53-4 (Both Incinerators)
Incinerator Vent
53-1
53-2
53-3
53-4 (Both Incinerators)
Hydrocarbons
Participates
CO
PAN & MAN
Benzole Acid
CO
COz
All
Organlcs
Participates
CO & COj
Total Hydrocarbons
Organlcs
Partlculates
NOX
CO, CO2, 02 & N2
plane lonlzatlon
Alundum Filter and Implngers
Not Specified
Polarograph
Gas Chromatograph
Gas Chromatograph
Gas Chromatograph
Gas Chromatograph
Flame lonlzatlon
Implngers
Orsat
None
Flame lonlzatlon
Water Scrub & Analyzed for
Total Organic Carbon
EPA Method 5
EPA Method 7
Orsat
Perkins Elmer
Perkins Elmer
1440
1540
1540
20' x 1/8"
6' x 1/8"
6' x 1/8"
157 ffat on chromosorb W
treated vlth DMCS
13x mole sieve
08 silica gel
-------
PA-5 6
XVII. Emergency Action Plan for Air Pollution Episodes
A. Types of Episodes
The alleviation of Air Pollution Episodes as suggested by the U.S.
Environmental Protection Agency is based on a pre-planned episode
emissions reduction scheme. The criteria that set this scheme into
motion are:
1. Alert Status - The alert level is that concentration of
pollutants at which short-term health effects can be expected
to occur.
2. Warning Status - The warning level indicates that air quality
is continuing to deteriorate and that additional abatement
actions are necessary.
3. Emergency Status - The emergency level is that level at which
a substantial endangerment to human health can be expected.
These criteria are absolute in the sense that they represent
a level of pollution that must not be allowed to occur.
B. Sources of Emissions
As outlined in the foregoing in-depth study of phthalic anhydride
manufacture, there are as many as five continuous and three intermittent
vent streams to the atmosphere.
1. Continuous Streams
(a) Main Process Vent Gas - This stream constitutes the
greatest potential for air pollution. It consists of
the gross reactor effluent after cooling and recovery of
crude phthalic anhydride. The stream is normally either
directed to a water scrubber or an incinerator before
exhausting to the atmosphere.
(b) Pretreatment and Product Fractionation Vent - These
operations are performed under vacuum and evolve dis-
solved non-condensible and light ends which are contained
in the exhaust of a vacuum ejector. In some plants, the
ejector effluent is condensed in an after cooler or hot
well. The resulting waste water is directed to either
the main process vent gas incinerator or vent gas scrubber
and incinerator system. In the cases where ejector
effluent is not condensed, it is directed to the main
process vent incinerator, to a separate waste product
thermal incinerator or to the atmosphere.
(c) Xylene Feed Storage - Xylene is stored at ambient
temperature in fixed roof storage tanks with atmospheric
vents. Because of low vapor pressure, emissions are
negligible except during filling periods.
-------
PA-5 7
(d) Product Storage - Crude and refined phthalic anhydride
are maintained at 300° F to 320° F at atmospheric
pressure in order to hold them in a molten state. The
tanks are usually blanketed with dry nitrogen. Consequently,
there is a small continuous gaseous emission. Usually
in those plants that vent these streams, the gas is first
sent through sublimation boxes or devices wherein the
phthalic anhydride is solidified to crystals and collected
for disposal or recovery. In other instances, the vent
is collected by an ejector and sent to an incinerator.
(e) Flaker and Bagging Exhaust - Usually the phthalic anhydride
is stored and transported as a liquid. There are some
instances where the product is shipped as a flaked product.
This presents another source of emission in the form of
phthalic anhydride dust around the bagging operation.
Usually it is recovered by a ventilating system and
ducted to a cyclone for recovery of product. The cleaned
exhaust presents no problem.
2. Intermittent Air Emissions
(a) Process Vent Gas - Some Facilities provide for emergency
venting of the main process vent stream usually through
the rupture of a "bursting disc". The gross venting in
this event will last only a minute or two or until the
air compressor can be shutdown. Some emissions, however,
will continue for several hours while the switch condensers
are melted and discharged. An emergency of this type occurs
very infrequently.
(b) Start-Up Vent - Some plants may employ a direct atmospheric
vent for exhausting reactor effluent during plant start-up.
The emission during this period consists of hot air and
natural gas combustion products.
(c) Product Shipping Losses - Intermittent emissions result
from uncontrolled phthalic anhydride vapor during loading
of liquid product into tank trucks.
3. Fugitive Emissions
As in any processing plant there are emissions that result from
leaks and safing or purging of equipment in preparation for .
maintenance. This type of emission should be small and infrequent
in nature.
C. Abatement Techniques
As the various levels of the pre-planned episode reduction scheme
are declared (Alert, Warning and Emergency) a progressive reduction in
the amount of air pollutants emitted must be made. This could
ultimately lead to total curtailment of pollutant emissions if the
emergency level become imminent.
-------
PA-5 8
Although these instructions for the "Air Pollution Episode Avoidance
Plan" are designed for phthalic anhydride manufacturing plants the
overall Emergency Action Plan (EAP) will cover all aspects of
environmental air pollution. Consequently, the implementation of the
pre-planned episode reduction scheme, as it applies to phthalic anhydride
manufacture, will be in consideration of reductions made in all sources
of air pollutants as well as to the specific offending constituents in
the atmosphere. Therefore, the extent of required cut back in emissions
from phthalic anhydride plants will depend on the relative amounts of
air pollutants contributed by phthalic anhydride production to the
overall emissions which resulted in the pollution episode. These
factors will be used by the Governing Environmental Protection
Authority in determining the cutback to be made in all air pollution
sources during the various episodes.
Phthalic anhydride manufacturing facilities consist of plants with
multiple parallel reactors and, in some larger installations, complete
parallel trains of equipment in the main processing areas. A multiple
reactor system provides for increased flexibility to affect a partial
reduction in air pollutant emissions during an air pollution alert.
This is possible since individual reactors can be removed from
service resulting in proportionate reduction of absorber vent gas
emissions. A single reactor system in a multi reactor plant can be
taken out of service in about one hour. A start-up of a single reactor
system can be accomplished in less than one hour. Another option
available for a partial reduction in air pollutant emissions is to
reduce the capacity of all reactors (turndown). Reductions of up to
40 percent are possible for short periods of time. It should be noted
that the oxidation of 0-xylene is an exothermic reaction with the
exotherm consumed within the process to generate steam. Turndown
of all reactors to a level below 60 percent will normally result in
a steam deficient condition. Consequently, auxiliary steam generating
facilities will have to be placed into service to prevent potential
safety problems and equipment damage due to solidification of high
melting point materials in the process. Such a condition could result
in a net increase in total emissions.
Reduction in operating rate results in reductions in emissions of
organic acids from the scrubber tail gas (53-1) with an accompanying
reduction in the incinerator effluent. Limited data indicate that
emissions decrease at a rate that is more than a linear proportion at
lower operating levels. Under normal operation conditions a turndown
can be accomplished within one hour. It should be noted that operating
at reduced capacity increases the residence time of xylene in the
reactors which results in lower yields and side reactions.
The curtailment of operation on one or more reactors or a turndown
in capacity must be considered with repsect to emission control
equipment. For example, in plants that employ thermal incinerators, the
curtailment of operation on one or more reactors of a given unit would
result in a decrease in the amount of combustibles flowing to the
incinerating device. However, the composition of the incinerator feed
would remain approximately the same with the total flow rate reduced.
This could result in an increase in furnace residence time with a
lower temperature requirement thus possibly favoring a reduction in
NOx emission. The latter condition would also be applicable on a
turndown in plant capacity.
-------
PA-5 9
Plants employing water scrubbers on the main process vent gas should
continue to run this equipment at design water circulation rates during
an alert. With a reduction in the total flow of the main process vent
gas, the scrubbing efficiency should be improved over that obtained at
normal phthalic anhydride production levels.
During episodes, it might be possible to reduce SC-2 addition to the
reactor feed. In addition, it may be possible to switch to a low sulfur
content fuel.
1. Declaration of Alert Condition - When an alert condition is
declared, the episode emission reduction plan is immediately
set into motion. Under this plan, in addition to notifying
the manufacturer of the alert condition, it may be deemed
necessary by the Environmental Protection Authorities to reduce
emissions from phthalic anhydride manufacture by a small
amount in order to deter further increases in pollution level
which could result in warning or emergency episodes. This may
be accomplished by employing one of the foregoing options.
The specific option to be used for the reduction is at the
discretion of the manufacturer. The time required to affect
the reduction will be approximately as stated in the preceding
discussion. This will reduce the principal source of emission,
represented by the main process vent stream, by a similar amount
to the reduction made in PAN production. The other sources of
emission, represented by the pretreatment and product fractionation
vents, xylene feed and product storage and flaker and bagging
exhaust will be reduced to some lesser degree by viture of the
reduction made in the producing equipment. Usually the alert
condition can be expected to continue for 12 hours or more.
2. Declaration of Warning Condition - When the air pollution
warning episode is announced, a substantial reduction of air
contaminants is desirable even to the point of assuming
reasonable economic hardship in the cutback of production and
allied operations. This could involve a 50-60 percent decrease
in phthalic anhydride production.
3. Emergency Condition - When it appears that an air pollution
emergency episode is imminent, all air contaminants, except
those resulting from storage facilities, may have to be
eliminated immediately by ceasing production and allied
operations to the extent possible without causing injury to
persons or damage to equipment.
The cessation of operation whether wholly or in part should not
result in increased emissions. This is also true for start-up
operations.
D. Economic Considerations
The economic impact on phthalic anhydride manufacturers of curtailing
operations during any of the air pollution episodes is based on the
duration and number of episodes in a given period. It is indicated that
the usual duration of air pollution episodes is one to seven days with
meteorology episode potentials as high as 80 per year.20 The frequency
of air pollution episodes in any given area is indicated as being one
-------
PA-60
to four per year. These data do not differentiate between the
episode levels. Normally, since the alert level does not require
a cutback in production, it will not influence plant economics.
Therefore, in discussing economic considerations resulting from the
air pollution abatement plan, it is only necessary to estimate the
frequency and number of warning and emergency episodes. For the
economic study, it has been assumed that three warning and no
emergency episodes occur per year. Each warning episode is assumed
to require a 50 percent reduction in air contaminants for a period of
5^ days.
The financial impact resulting from this loss in production is shown
in Table PA-19. This table contains comparative manufacturing costs for
an existing 130 MM Ibs./year facility without extensive pollution
control (Table PA-11) and for a most feasible new facility of the
same capacity (Table PA-13).
Economics are shown for each of these plants with and without the
financial impact accredited to the air pollution episodes. It should
be noted that whereas the proposed cutback in phthalic anhydride
production for emission control appears small (2.5 percent on a yearly
basis), it reduced net profit by six to seven percent.
E. Summary of Estimated Emissions
In the foregoing, a reduction in air pollutant emissions was
suggested for the various air pollution levels that may be encountered.
This was primarily predicated on existing plants with limited or no
pollution control equipment. However, most existing plants do provide
efficient control devices which substantially reduce emissions.
Therefore, special consideration should be made in the EAP for Air
Pollution Episode Avoidance for new and existing plants that are
equipped with the "latest state of the art" emission control equipment.
The following presents estimated air emissions for a typical present-day
system without control devices and the most feasible new plant that
incorporates thermal incineration.
Present Plant Without Most Feasible
Pollution Control New Plant
Average Emissions, Average Emissions,
Pollutant T/T T/T
Hydrocarbons 0.0001 0.0001
Particulates (a) 0.1192 0.0040
NOx - 0.0013
SOx 0.0047 0.0047
CO 0.1507 0.0101
0.2747 0.0202
(a) Includes PAN, MAN and Organic Acids
As noted in the above, total emissions for the most feasible new
plant have been reduced to about 7^ percent of that estimated for the
uncontrolled plant. However, some NOx emission is produced by
incineration.
-------
TABLE PA-19
FINANCIAL IMPACT OF AIR POLLUTION EPISODES
ON MANUFACTURING
COSTS
FOR 130 MM LBS./YR. PHTHALIC ANHYDRIDE FACILITY
Type of Operation
Direct Manufacturing Cost,
Raw Materials
95.5% 0-Xylene @ 4£c
S02 @ 5^p/lb.
Labor
Maintenance
Utilities
TYPICAL
EXISTING PLANT
No Cutback In Assuming
Production 8.5 Days Lost
(Table PA-11) Production
$/Yr.
:/lb. 5,642,000
39,000
130,000
299,000
780,000
6,890,000
Indirect Manufacturing Cost, $/Yr.
Plant Overhead
Fixed Manufacturing Cost,
Depreciation, Insurance
Total Manufacturing Costs.
General Expenses, $/Yr.
Administration, Sales,
TOTAL COSTS. $/YR.
Selling Price
Profit Before Taxes
Profit After Taxes
Cash Flow
ROI
143,000
$/Yr.
and Property Taxes 1,235,000
$/Yr. 8,268,000
Research and Finance 1,092,000
9,360,000
11,440,000
2,080,000
998,500
1,998,500
10.0%
5,501,000
38,000
130,000
299,000
760,000
6,728,000
143,000
1,235,000
8,106,000
1,092,000
9,198,000
11,154,000
1,956,000
939,000
1,939,000
9.4%
MOST FEASIBLE
No Cutback In
Production
(Table PA- 13)
5,642,000
39,000
156,000
351,000
923,000
7,111,000
169,000
1,404,000
8,684,000
1,183,000
9,867,000
11,440,000
1,573,000
755,000
1,897,000
6.6%
NEW PLANT
Assuming
8.5 Days Lost
Production
5,501,000
38,000
156,000
351,000
900,000
6,946,000
169,000
1,404,000
8,519,000
1,183,000
9,702,000
11,154,000
1,452,000
697,000
1,839,000
6.1%
-------
PA-62
The particular type and concentration of pollutants in the
atmosphere at the time of the episode would dictate the degree to
vhich a reduction would be made on the most feasible new plant. If
NOx or SOx is the offending material, then a reduction in plant
production may be required as outlined under "Declaration of Alert
Condition". In this case, NOx would be reduced as the cutback is
made in production.
If the offending pollutants are in the form of hydrocarbons,
particulates or CO, the degree of cutback on the most feasible new
plant could be proportionally less severe than on an uncontrolled
facility.
-------
APPENDIX I
BASIS OF THE STUDY
I. Industry Survey
The study which led to this document was undertaken to obtain information
about selected production processes that are practiced in the Petrochemical
Industry. The objective of the study was to provide data for the EPA to use
in the fulfillment of their obligations under the Clean Air Amendments of 1970.
The information obtained during the study includes industry descriptions,
air emission control problems, sources of air emissions, statistics on quantities
and types of emissions and descriptions of emission control devices currently
in use. The principal source for these data was an Industry Questionnaire
but it was supplemented by plant visits, literature searches, in-house back-
ground knowledge and direct support from the Manufacturing Chemists Association.
More than 200 petrochemicals are currently produced in the United States,
and many of these by two or more different processes. It was obvious that
the most immediate need was to study the largest tonnage, fastest growth
processes that produce the most pollution. Consequently, the following 32
chemicals (as produced by a total of 41 different processes) were selected
for study:
Acetaldehyde (two processes)
Acetic Acid (three processes)
Acetic Anhydride
Acrylonitrile
Adipic Acid
Adiponitrile (two processes)
Carbon Black
Carbon Disulfide
Cyclohexanone
Ethylene
Ethylene Dichloride (two processes)
Ethylene Oxide (two processes)
Formaldehyde (two processes)
Glycerol
Hydrogen Cyanide
Maleic Anhydride
Nylon 6
Nylon 6,6
"Oxo" Alcohols and Aldehydes
Phenol
Phthalic Anhydride (two processes)
Polyethylene (high density)
Polyethylene (low density)
Polypropylene
Polystyrene
Polyvinyl Chloride
Styrene
Styrene - Butadiene Rubber
Terephthalic Acid (1)
Toluene Di-isocyanate (2)
Vinyl Acetate (two processes)
Vinyl Chloride
(1) Includes dimethyl terephthalate.
(2) Includes methylenediphenyl and polymethylene polyphenyl isocyanates.
The Industry Questionnaire, which was used as the main source of information,
was the result of cooperative efforts between the EPA, Air Products and the
EPA's Industry Advisory Committee. After receiving approval from the Office of
Management and Budget, the questionnaire was sent to selected producers of
most of the chemicals listed above. The data obtained from the returned
questionnaires formed the basis for what have been named "Survey Reports".
These have been separately published in four volumes, numbered EPA-450/3-73-005a,
b, c, and d and entitled "Survey Reports on Atmospheric Emissions from the
Petrochemical Industry - Volumes I, II, III, and IV.
-------
1-2
The purpose of the survey reports was to screen the various petrochemical
processes into the "more" and "less - significantly polluting processes".
Obviously, significance of pollution is a term which is difficult if not
impossible to define because value judgements ar« involved. Recognizing this
difficulty, a quantitative method for Significant Emission Index (SEI) was
developed. This procedure is discussed and illustrated in Appendix II of
this report. Each survey report includes the calculation of an SEI for the
petrochemical that is the subject of the report. These SEI's have been
incorporated into the Emission Summary Table that constitutes part of this
Appendix (Table I). This table can be used as an aid when establishing
priorities in the work required to set standards for emission controls on
new stationary sources of air pollution in accordance with the terms of the
Clean Air Amendments of 1970.
The completed survey reports constitute a preliminary data bank on each
of the processes studied. In addition to the SEI calculation, each report
includes a general introductory discussion of the process, a process description
(including chemical reactions), a simplified process flow diagram, as well as
heat and material balances. More pertinent to the air pollution study, each
report lists and discusses the sources of air emissions (including odors and
fugitive emissions) and the types of air pollution control equipment employed.
In tabular form, each reports summarizes the emission data (amount, composition,
temperature, and frequency); the sampling and analytical techniques; stack
numbers and dimensions; and emission control device data (types, sizes, capital
and operating costs, and efficiencies).
Calculation of efficiency on a pollution control device is not necessarily
a simple and straight-forward procedure. Consequently, two rating techniques
were developed for each type of device, as follows:
1. For flares, incinerators, and boilers a Completeness of Combustion Rating
(CCR) and Significance of Emission Reduction Rating (SERR) were used.
2. For scrubbers and dust removal equipment, a Specific Pollutant
Efficiency (SE) and a SERR were used.
The bases for these ratings and example calculations are included in
Appendix III of this report.
II. In-Depth Studies
The original performance concept was to select a number of petrochemical
processes as "significant polluters", on the basis of data contained in
completed questionnaires. These processes were then to be studied "in-depth".
However, the overall time schedule was such that the EPA requested an initial
selection of three processes on the basis that they would probably turn out
to be "significant polluters". The processes selected in this manner were:
1. The Furnace Process for producing Carbon Black.
2. The Sohio Process for producing Acrylonitrile.
3. The Oxychlorination Process for producing 1,2 Dichloroethane
(Ethylene Dichloride) from Ethylene.
-------
Acetaldehyde via Ethylene
via Ethanol
Acetic Acid via MetHanoi
via Butane
via Acetaldehyde
Acetic Anhydride via Acetic Acid
Acrylonitrlle (9)
Adipic Acid
Adiponitrile via Butadiene
via Adipic Acid
Carbon Black
Carbon Dlsulfide
Cyclohexanone
Dimethyl Terephthalate (+TPA)
Ethylene
Ethylene Dichlorlde via Oxychlorlnatlon
via Direct Chlorlnation
Ethylene Oxide
Formaldehyde via Silver Catalyst
via Iron Oxide Catalyat
Glycerol via Eplchlorohydrin
Hydrogen Cyanide Direct Process
Isocyanates
Nalelc Anhydride
Nylon 6
Nylon 6,6
Oxo Process
Phenol (
Phthaiic Anhydride via 0-Xylene
via Naphthalene
High Density Polyethylene
Low Density Polyethylene
Polypropylene
Polystyrene
Polyvinyl Chloride
Styrene
Styrene-Butadiene Rubber
Vinyl Acetate via Acetylene
via Ethylene
Vinyl Chloride
Totals
TABLE I
EMISSIONS SUMMARY
ESTIMATED ^ CURRENT AIR EMISSIONS. MM LBS./YEAR
Page 1 of 3
Hydrocarbons (3)
1.1
0
0
40
6.1
3.1
183
0
11.2
0
156
0.15
70
91
15
95.1
29
85.8
23.8
25.7
16
0.5
1.3
34
0
0
5.25
24.3
0.1
0
79
75
37.5
20
62
4.3
9.4
5.3
0
17.6
Participates (*)
0
0
0
0
0
0
0
0.2
4.7
0.5
8.1
0.3
0
1.4
0.2
0.4
0
0
0
0
0
0
0.8
0
1.5
5.5
0.01
0
5.1
1.9
2.3
1.4
0.1
0.4
12
0.07
1.6
0
0
0.6
Oxides of Nitrogen
0
0
0.01
0.04
0
0
5.5
29.6
50.5
0.04
6.9
0.1
0
0.1
0.2
0
0
0.3
0
0
0
0.41
0
0
0
0
0.07
0
0.3
0
0
0
0
0
0
0.14
0
0
TR
0
Sulfur Oxides
0
0
0
0
0
0
0
0
0
0
21.6
4.5
0
1.0
2.0
0
0
0.1
0
0
0
0
0.02
0
0
0
0
0
2.6
0
0
0
0
1.2
0
0
0.9
0
0
0
Carbon Monoxide
0
27
0
14
1.3
5.5
196
0.14
0
0
3,870
0
77.5
53
0.2
21.8
0
0
107.2
24.9
0
0
86
260
0
0
19.5
0
43.6
45
0
0
0
0
0
0
0
0
0
e
Total
1.1
27
0.01
54
7.4
8.6
385
30
66.4
0.54
4,060
5.1
148
146.5
17.6
117.3
29
86.2
131
50.6
16
0.91
88
294
1.5
5.5
24.8
24.3
51.7
47
81.3
76.4
37.6
21.6
74
4.5
12
5.3
TR
18.2
Total Weighted (5
86
27
1
3,215
490
253
15,000
1,190
3.200
30
17,544
120
5,700
7,460
1,240
7,650
2,300
6.880
1,955
2,070
1,280
56
231
2,950
90
330
440
1,940
422
160
6,400
6,100
2,950
1,650
5,700
355
870
425
TR
1.460
1,227.6
49.1
94.2
33.9
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(9)
In most instances numbers are based on less than 1007; survey. All based on engineering judgement of best current control.
Assumes future plants will employ best current control techniques.
Excludes methane, includes HjS and all volatile organics.
Includes non-volatile organics and inorganics.
Weighting factors used are: hydrocarbons - 80, partlculates - 60, NOX - 40, SOX - 20, and CO - 1.
Referred to elsewhere in this study as "Significant Emission Index" or "SEI".
Totals are not equal across and down due to rounding. '
Emissions based on what is now an obsolete catalyst. See Report No. EPA-450/3-73-006 b for up-to-date information.
4,852.6 6,225.9 <7>
Probably has up to 10% lov bias.
110,220 (7)
-------
Hydrocarbon! * '
1.2
0
0
0
12.2
0.73
284
0
10.5
0
64
0.04
77.2
73.8
14.8
110
34.2
32.8
14.8
17.6
8.9
0
1.2
31
0
0
3.86
21.3
0.3
0
210
262
152
20
53
3.1
1.85
4.5
0
26.3
1,547.2
Particulates
0
0
0
0
0
0
0
0.14
4.4
0.5
3.3
0.07
0
1.1
0.2
0.5
0
0
0
0
0
0
0.7
0
3.2
5.3
0.01
0
13.2
0
6.2
5
0.5
0.34
10
0.05
0.31
0
0
0.9
55,9
«"•« *
EMSSIOH SUMMARY
ESTIMATED ADDITIONAL (2>
(*) Oxides of Nitrogen
0
0
0.04
0
0
0
8.5
19.3
47.5
0.04
2.8
0.03
0
0.07
0.2
0
0
0.15
0
0
0
0
0
0
0
0
0.05
0
0.8
0
0
0
0
0
0
0.1
0
0
TO
0
79,5
AIR EMISSIONS IN
Sulfur Oxides
0
0
0
0
0
0
0
0
0
0
8.9
1.1
0
0.84
61.5
0
0
0.05
0
0
0
0
0.02
0
0
0
0
0
6.8
0
0
0
0
1.13
0
0
0.18
0
0
0
80.5
1980, MM LBS./YEAR
Carbon Monoxide
0
0
0
0
2.5
1.42
304
0.09
0
0
1,590
0
85.1
42.9
0.2
25
0
0
66.7
17.0
0
0
85
241
0
0
14.3
0
113
0
0
0
0
0
0
0
0
0
0
0
2,588
Page 2 of 3
Total
1.2
0
0.04
0
14.7
2.15
596
19.5
62.4
0.54
1,670
1.24
162
118.7
77
136
34.2
33
81.5
34.6
8.9
0
87
272
3.2
5.3
18.2
21.3
134
0
216
267
152.5
21.47
63
3.25
2.34
4.5
TR
27.2
4,351.9
Total Weighted (5,6)
96
0
2
0
980
60
23,000
779
3,010
30
7,200
30
6,260
6,040
2,430
8,800
2,740
2,650
1,250
1,445
700
0
225
2,720
194
318
325
1,704
1,100
0
17,200
21,300
12,190
1,640
4,840
225
170
360
TR
2.170
134,213 (?)
Acetaldehyde via Ethylene
via Ethanol
Acetic Acid via Methanol
via Butane
via Acetaldehyde
Acetic Anhydride via Acetic Acid
Acrylonitrlle (9)
Adipic Acid
Adiponitrile via Butadiene
via Adipic Acid
Carbon Black
Carbon Diaulfide
Cyclohexanone
Dimethyl Terephthalate (+TPA)
Ethylene
Ethylene Dichloride via Oxychlorination
via Direct Chlorination
Ethylene Oxide
Formaldehyde via Silver Catalyst
via Iron Oxide Catalyst
Glycerol via Eplchlorohydrin
Hydrogen Cyanide Direct Process
Isocyanates
Maleic Anhydride
Nylon 6
Nylon 6,6
Oxo Process
Phenol
Phthalic Anhydride via 0-Xylene
via Naphthalene
High Density Polyethylene
Low Density Polyethylene
Polypropylene
Polystyrene
Polyvinyl Chloride
Styrene
Styrene-Butadiene Rubber
Vinyl Acetate via Acetylene
via Ethylene
Vinyl Chloride
Totals
(1) In most instances numbers are based on leas than 1007. survey. All based on engineering judgement of best current control.
(2) Assumes future plants vill employ beat current control techniques.
(3) Excludes methane, includes fyS and all volatile organics.
(4) Includes non-volatile organic! and inorganics.
(5) Weighting factors used are: hydrocarbon* - 80, particulates - 60, NOX - 40, SOX - 40, and CO - 1.
(6) Referred to elsewhere in this study as "Significant Emission Index" or "SEI".
(7) Totals are not equal across and dovn duv to rounding.
(9) See sheet 1 of 3.
Probably has up to 107, lou bias.
-------
TABLE I
EMISSIONS
Page 3 of 3
Emissions <2), MM Lba._/Y.Sar
Acetaldehyde via Ethylene
via Ethanol
Acetic Acid via Hethanol
via Butane
via Acetaldehyde
Acetic Anhydride via Acetic Acid
Acrylonitrile (9)
Adlplc Acid
Adiponltrlle via Butadiene
via Adlpic Acid
Carbon Black
Carbon Dlsulflde
Cyc1ohexanone
Dimethyl Terephthalate (+TPA)
Ethylene
Ethylene Bichloride via Oxychlorlnatlon
via Direct Chlorination
Ethylene Oxide
Formaldehyde via Silver Catalyst
via Iron Oxide Catalyat
Glycerol via Epichlorohydrin
Hydrogen Cyanide Direct Process
Isocyanates
Maleic Anhydride
Nylon 6
Nylon 6,6
Oxo Process
Phenol
Phthallc Anhydride via O-Xylene
via Naphthalene
High Density Polyethylene
Low Density Polyethylene
Polypropylene
Polystyrene
Polyvinyl Chloride
Styrene
Stymie-Butadiene Rubber
Vinyl Acetate via Acetylene
via Ethylene
Vinyl Chloride
Total by 1980
2.3
27
0.05
54
22
10.8
980
50
128.8
1.1
5,730
6.3
310
265
94
253
63
120
212.5
85
25
0.5 (10)
175
566
4.7
10.8
43
46
186
47
297
343
190
43
137
7.4
14
9.8
TR
45
Total Weighted (5) by 1980
182
27
3
3,215
1,470
313
38,000
1,970
6,210
60
24,740
150
11,960
13,500
3,670
16,450
5,040
9,530
3,205
3,515
2,000
28
456
5,670
284
650
765
3,640
1,522
160
23,600
27,400
15,140
3,290
10,540
610
1,040
785
TR
3,630
Totals
10,605 (7)
(10)
Estimated Number of New Plants
(1973 - 1980)
6
0
4
0
3
3
5
7
4
3
13
2
10
8
21
8
10
15
40
12
1
0
10
6
10
10
6
11
6
0
31
41
32
23
25
9
4
1
4
10
Total Estimated Capacity
MM Lbs./Year
Current By 1980
1 ,160
966
400
1,020
875
1,705
1,165
1,430
435
280
3,000
871
1,800
2,865
22,295
4,450
5,593
4,191
5,914
1,729
245
412
1,088
359
486
1,523
1,727
2,363
720
603
2,315
5,269
1,160
3,500
4,375
5,953
4,464
206
1,280
5,400
2
1
2
2
3
2
5
1
3
5
40
8
11
6
9
3
2
1
3
3
4
1
8
21
5
6
8
10
5
2
13
,460
966
,800
500
,015
,100
,700 (8)
,200
845
550
,000 (8)
,100
,600
,900
,000
,250 (8)
,540
,800 (8)
,000
,520 (8)
380
202
,120
720
,500
,000
,000
,200
,800 (8)
528
,500
,100
,800
,700
,000
,000
,230
356
,200
,000
244,420
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
In most instances numbers are baaed on less than 100% survey. All baaed on engineering judgement of best current control. Probably has up to 107. low bias.
Assumes future plants will employ best current control techniques.
Excludes methane, includes H^S and all volatile organics.
Includes non-volatile organlca and inorganics.
Weighting factors used are: hydrocarbons - 80, particulates - 60, NOX - 40, SO^ - 20, and CO - 1.
Referred to elsewhere in this study as "Significant Emission Index" or "SEI".
Totals are not equal across and down due to rounding.
By 1985.
See sheet 1 of 3
Due to anticipated future shut down of marginal plants.
-------
1-6
In order to obtain data on these processes, the operators and/or
licensors of each were approached directly by Air Products' personnel.
This, of course, was a slow and tedious method of data collection because
mass mailing techniques could not be used, nor could the request for data
be" identified as an "Official EPA Requirement". Yet, by the time that OMB
approval was given for use of the Industry Questionnaire, a substantial
volume of data pertaining to each process had already been received. The
value of this procedure is indicated by the fact that first drafts of these
three reports had already been submitted to the EPA, and reviewed by the
Industry Advisory Committee, prior to the completion of many of the survey
reports.
In addition, because of timing requirements, the EPA decided that three
additional chemicals be "nominated" for in-depth study. These were phthalic
anhydride, formaldehyde and ethylene oxide. Consequently, four additional
in-depth studies were undertaken, as follows:
1. Air Oxidation of Ortho-Xylene to produce Phthalic Anhydride.
2. Air Oxidation of Methanol in a Methanol Rich Process to produce
Formaldehyde over a Silver Catalyst. (Also, the subject of a
survey report.)
3. Air Oxidation of Methanol in a Methanol-Lean Process to
produce Formaldehyde over an Iron Oxide Catalyst.
4. Direct Oxidation of Ethylene to produce Ethylene Oxide.
The primary data source for these was the Industry Questionnaire,
although SEI rankings had not been completed by the time the choices were
made.
The Survey Reports, having now been completed are available, for use in
the selection of additional processes for in-depth study.
-------
INTRODUCTION TO APPENDIX II AND
The following discussions describe techniques that were developed for
the single purpose of providing a portion of the guidance required in the
selection of processes for in-depth study. It is believed that the underlying
concepts of these techniques are sound. However, use of them without sub-
stantial further refinement is discouraged because the data base for their
specifics is not sufficiently accurate for wide application. The subjects
covered in the Appendix II discussion are:
1. Prediction of numbers of new plants.
2. Prediction of emissions from the new plants on a weighted
(significance) basis.
The subject covered in the Appendix m discussion is:
Calculation of pollution control device efficiency on a variety of
bases, including a weighted (significance) basis.
It should be noted that the weighting factors used are arbitrary.
Hence, if any reader of this report wishes to determine the effect of
different weighing factors, the calculation technique permits changes in
these, at the reader's discretion.
-------
APPENDIX II
Number of New Plants*
Attached Table 1 illustrates the format for this calculation.
Briefly, the procedure is as follows:
1. For each petrochemical that is to be evaluated, estimate what
amount of today's production capacity is likely to be on-stream
in 1980. This will be done by subtracting plants having marginal
economics due either to their size or to the employment of an
out-of-date process.
2. Estimate the 1980 demand for the chemical and assume a 1980
installed capacity that will be required in order to satisfy
this demand.
3. Estimate the portion of the excess of the 1980 required capacity
over today's remaining capacity that will be made up by
installation of each process that is being evaluated.
4. Estimate an economic plant or unit size on the basis of today's
technology.
5. Divide the total required new capacity for each process by the
economic plant size to obtain the number of new units.
In order to illustrate the procedure, data have been incorporated
into Table I, for the three processes for producing carbon black, namely
the furnace process, the relatively non-polluting thermal process, and .
the non-growth channel process.
*The format is based on 1980, but any future year may be selected.
-------
Table 1. Number of New Plants by 1980
Current
Chemical
Carbon Black
Process
Furnace
Channel
Thermal
Current
Capacity
4,000
100
200
Marginal
Capacity
0
0
0
Capacity
on-stream Demand
in 1980
4,000
100
200
1980
4,500
100
400
Capacity
1980
5,000
100
500
Capacity
to be
Added
1,000
0
300
Economic
Plant
Size
90
30
150
Number of
New
Units
11 - 12
0
2
M
M
1
S3
Notes: 1. Capacity units all in MM Ibs./year.
2. 1980 demand based on studies prepared for EPA by Processes Research, Inc. and MSA Research Corporation.
-------
II-3
Increased Emissions (Weighted) by 1980
Attached Table 2 illustrates the format for this calculation.
However, more important than format is a proposal for a weighting basis.
There is a wide divergence of opinion on which pollutants are more noxious
and even when agreement can be reached on an order of noxiousness, dis-
agreements remain as to relative magnitudes for tolerance factors. In
general pollutants from the petrochemical industry can be broken down into
categories of hydrogen sulfide, hydrocarbons, particulates, carbon monoxide,
and oxides of sulfur and nitrogen. Of course, two of these can be further
broken down; hydrocarbons into paraffins, olefins, chlorinated hydrocarbons,
nitrogen or sulfur bearing hydrocarbons, etc. and particulates into ash,
catalyst, finely divided end products, etc. It was felt that no useful
end is served by creating a large number of sub-groupings because it would
merely compound the problem of assigning a weighting factor. Therefore,
it was proposed to classify all pollutants into one of five of the six
categories with hydrogen sulfide included with hydrocarbons.
There appears to be general agreement among the experts that carbon
monoxide is the least noxious of the five and that NOX is somewhat more
noxious than SOX« However, there are widely divergent opinions concerning
hydrocarbons and particulates - probably due to the fact that these are
both widely divergent categories. In recent years, at least two authors
have attempted to assign tolerance factors to these five categories.
Babcock (1), based his on the proposed 1969 California standards for
one hour ambient air conditions with his own standard used for hydrocarbons.
On the other hand, Walther (2), based his ranking on both primary
and secondary standards for a 24-hour period. Both authors found it
necessary to extrapolate some of the basic standards to the chosen time
period. Their rankings, on an effect factor basis with carbon monoxide
arbitrarily used as a reference are as follows:
Babcock
Walther
Hydrocarbons
Particulates
NO*
SOX
CO
2.1
107
77.9
28.1
1
Primary
125
21.5
22.4
15.3
1
Secondary
125
37.3
22.4
21.5
1
Recognizing that it is completely unscientific and potentially subject
to substantial criticism it was proposed to take arithmetic averages of the
above values and round them to the nearest multiple of ten to establish a
rating basis as follows:
Hydrocarbons
Particulates
NOX
SOX
CO
Average
84.0
55.3
40.9
21.6
1
Rounded
80
60
40
20
1
-------
Table 2. Weighted Emission Rates
Chemical_
Process
Increased Capacity_
Pollutant
Hydrocarbons
Particulates
NOX
SOX
CO
Increased Emissions Weighting
Emissions, Lbs./Lb. Lbs./Year Factors
80
60
40
20
1
Weighted Emissions
Lbs./Year
Total
-------
II-5
Increased Emissions (Weighted) by 1980 (continued)
This ranking can be defended qualitatively, if not quantitatively for
the following reasons:
1. The level of noxiousness follows the same sequence as is obtained
using national air quality standards.
2. Approximately two orders of magnitude exist between top and bottom
rankings .
3. Hydrocarbons should probably have a lower value than in the
Walther analysis because such relatively non-noxious compounds
as ethane and propane are included.
4. Hydrocarbons should probably have a higher value than in the
Babcock analysis because such noxious (or posionous) substances
as aromatics, chlorinated hydrocarbons, phenol, formaldehyde, and
cyanides are included.
5. Particulates should probably have a higher value than in the
Walther analysis because national air standards are based mostly
on fly ash while emissions from the petrochemical industry are
more noxious being such things as carbon black, phthalic anhydride,
PVC dust, active catalysts, etc.
6. NOx should probably have a higher value than in the Walther
analysis because its role in oxidant synthesis has been neglected.
This is demonstrated in Babcock *s analysis.
Briefly, the procedure, using the recommended factors and Table 2, is
as follows :
1. Determine the emission rate for each major pollutant category in
terms of pounds of pollutant per pound of final product. (This
determination was made, on the basis of data reported on returned
questionnaires, in the Survey Reports,.
2. Multiply these emission rates by the estimate of increased production
capacity to be installed by 1980 (as calculated while determining
the number of new plants), to determine the estimated pounds of
new emissions of each pollutant.
3. Multiply the pounds of new emissions of each pollutant by its
weighting factor to determine a weighted pounds of new emissions
for each pollutant.
4. Total the weighted pounds of new emissions for all pollutants to
obtain an estimate of the significance of emission from the process
being evaluated. It was proposed that this total be named
"Significant Emission Index" and abbreviated "SEI".
It should be pointed out that the concepts outlined above are not
completely original and considerable credit should be given to Mr. L. B. Evans
of .he EPA for setting up the formats of these evaluating procedures.
-------
II-6
Increased Emissions (Weighted) by 1980 (continued)
(1) Babcock, L. F., "A Combined Pollution Index for Measurement of Total
Air Pollution," JAPCA, October, 1970; Vol. 20, No. 10; pp 653-659
(2) Walther, E. G., "A Rating of the Major Air Pollutants and Their Sources
by Effect", JAPCA, May, 1972; Vol. 22, No. 5; pp 352-355
-------
Appendix III
Efficiency of Pollution Control Devices
Incinerators and Flares
The burning process is unique among the various techniques for
reducing air pollution in that it does not remove the noxious substance
but changes it to a different and hopefully less noxious form. It can be,
and usually is, a very efficient process when applied to hydrocarbons,
because when burned completely the only products of combustion are carbon
dioxide and water. However, if the combustion is incomplete a wide range
of additional products such as cracked hydrocarbons, soot and carbon
monoxide might be formed. The problem is further complicated if the
hydrocarbon that is being burned is halogenated, contains sulfur or is
mixed with hydrogen sulfide, because hydrogen chloride and/or sulfur oxides
then become products of combustion. In addition, if nitrogen is present,
either as air or nitrogenated hydrocarbons, oxides of nitrogen might be
formed, depending upon flame temperature and residence time.
Consequently, the definition of efficiency of a burner, as a pollution
control device, is difficult. The usual definition of percentage removal of
the noxious substance in the feed to the device is inappropriate, because
with this definition, a "smoky" flare would achieve the same nearly 100
percent rating, as a "smokeless" one because most of the feed hydrocarbon
will have either cracked or burned in the flame. On the other hand, any
system that rates efficiency by considering only the total quantity of
pollutant in both the feed to and the effluent from the device would be
meaningless. For example, the complete combustion of one pound of hydrogen
sulfide results in the production of nearly two pounds of sulfur dioxide, or
the incomplete combustion of one pound of ethane could result in the
production of nearly two pounds of carbon monoxide.
For these reasons, it was proposed that two separate efficiency rating
be applied to incineration devices. The first of these is a "Completeness
of Combustion Rating" and the other is a "Significance of Emission Reduction
Rating", as follows:
1. Completeness of Combustion Rating (CCR)
This rating is based on oxygen rather than on pollutants and is
the pounds of oxygen that react with the pollutants in the feed to
the device, divided by the theoretical maximum number of pounds that
would react: Thus a smokeless flare would receive a 100 percent
rating while a smoky one would be rated somewhat less, depending upon
how incomplete the combustion.
In utilizing this rating, it is clear that carbon dioxide and water
are the products of complete combustion of hydrocarbons. However, some
question could occur as to the theoretical completion of combustion
when burning materials other than hydrocarbons. It was recommended
that the formation of HX be considered complete combustion of halogenated
hydrocarbons since the oxidation most typically does not change the
valence of the halogen. On the other hand, since some incinerators will
be catalytic in nature it was recommended that sulfur trioxide be
considered as complete oxidation of sulfur bearing compounds.
-------
III-2
Efficiency of Pollution Control Devices
1. Completeness of Combustion Rating (CCR) (continued)
Nitrogen is more complex, because of the equilibria that exist
between oxygen, nitrogen, nitric oxide, nitrogen dioxide and the
various nitrogen radicals such as nitrile. In fact, many scientists
continue to dispute the role of fuel nitrogen versus ambient nitrogen
in the production of NOX. In order to make the CCR a meaningful
rating for the incineration of nitrogenous wastes it was recommended
that complete combustion be defined as the production of N2, thus
assuming that all NOX formed comes from the air rather than the fuel,
and that no oxygen is consumed by the nitrogen in the waste material.
Hence, the CCR becomes a measure of how completely the hydrocarbon
content is burned, while any NOX produced (regardless of its source)
will be rated by the SERR as described below.
2. Significance of Emission Reduction Rating (SERR)
This rating is based primarily on the weighting factors that
were proposed above. All air pollutants in the feed to the device
and all in the effluents from the device are multiplied by the
appropriate factor. The total weighted pollutants in and out are
then used in the conventional manner of calculating efficiency
of pollutant removal, that is pollutants in minus pollutants out,
divided by pollutants in, gives the efficiency of removal on a
significance of emission basis.
Several examples will serve to illustrate these rating factors.
as follows:
Example 1 - One hundred pounds of ethylene per unit time is burned
in a flare, in accordance with the following reaction:
3C2H4 4- 7 02 ....... fr C -f 2 CO + 3 C02 + 6 H20
Thus, 14.2 Ibs. of particulate carbon and 66.5 Ibs. of carbon
monoxide are emitted, and 265 Ibs. of oxygen are consumed.
Theoretical complete combustion would consume 342 Ibs. of oxygen
in accordance with the following reaction:
+ 3 02 > 2 C02 + 2
Thus, this device would have a CCR of 265/342 or 77.5%
Assuming that one pound of nitric oxide is formed in the reaction
as a result of the air used for combustion (this is about equivalent to
100 ppm), a SERR can also be calculated. It should be noted that the
formation of this NO is not considered in calculating a CCR because it
came from nitrogen in the air rather than nitrogen in the pollutant
being incinerated. The calculation follows:
-------
III-3
Efficiency of Pollution Control Devices
2. Significance of Emission
Pollutant
Hydrocarbons
Particulates
NOX
SOx
CO
Total
SERR = 8000 -
Weighting
Factor
80
60
40
20
1
958.5
Reduction
Rating (SERR) (continued)
Pounds in Pounds out
Actual
100
0
0
0
0
Weighted Actual Weighted
8000 0
14.2 852
1 40
0
66.5 66.5
8000 958.5
8000 x 1UU °°'°
Example 2 - The same as Example 1, except the hydrocarbons are
burned to completion. Then,
CCR = 342
342
x 100 = 100%
and
SERR
8QOO - 40
8000
99.5%
Example 3 - One hundred pounds per unit time of methyl chloride is
incinerated, in accordance with the following reaction.
2 CH3C1 + 3 02
2 C02 +2 H20 + 2 HC1
This is complete combustion, by definition, therefore, the CCR is
100%. However, (assuming no oxides of nitrogen are formed), the SERR
is less than 100% because 72.5 Ibs. of HC1 are formed. Hence,
considering HC1 as an aerosol or particulate;
SERR = 100 x 80 - 72.5 x 60
100 x 80
x 100 = 45.5%
The conclusion from this final example, of course, is that it is
an excellent combustion device but a very poor pollution control device,
unless it is followed by an efficient scrubber for HCl removal.
Example 4 - The stacks of two hydrogen cyanide incinerators, each
burning 100 pounds per unit time of HCN are sampled. Neither has any
carbon monoxide or particulate in the effluent. However, the first is
producing one pound of NOX and the second is producing ten pounds of
NOX in the same unit time. The assumed reactions are:
-------
III-4
Efficiency of Pollution Control Devices
2. Significance of Emission Reduction Rating (SERR) (continued)
4 HCN + 5 02 ' fr 2 H20 + 4 C02 + 2 N2
N2 (atmospheric) + X02 • '> 2 NOX
Thus, CCRi = 100% and CCR2 = 100% both by definition.
However, SERRj^ = 100 x 80 - 1 x 40
100 x 80
and SERRo = 100 x 80 - 10 x 40 _
100 x 80 x 1UO ~
Obviously, if either of these were "smoky" then both the CCR and
the SERR would be lower, as in Example 1.
Other Pollution Control Devices
Most pollution control devices, such as bag filters, electrostatic
precipitators and scrubbers are designed to physically remove one or more
noxious substances from the stream being vented. Typically, the efficiency
of these devices is rated relative only to the substance which they are
designed to remove and for this reason could be misleading. For example:
1. The electrostatic precipitator on a power house stack might be
99% efficient relative to particulates, but will remove little
or none of the SOX and NOX which are usually present.
2. A bag filter on a carbon black plant will remove 99 + % of the
particulate but will remove none of the CO and only relatively
small amounts of the compounds of sulfur that are present.
3. A water scrubber on a vinyl chloride monomer plant will remove
all of the hydrogen chloride but only relatively small amounts
of the chlorinated hydrocarbons present.
4. An organic liquid scrubber on an ethylene dichloride plant will
remove nearly all of the EDC but will introduce another pollutant
into the air due to its own vapor pressure.
For these reasons, it was suggested again that two efficiency ratings be
applied. However, in this case, the first is merely a specific efficiency as
is typically reported, i.e., "specific to the pollutant (or pollutants) for
which it was designed", thus:
SE = specific pollutant in - specific pollutant out
specific pollutant in
The second rating proposed is an SERR, defined exactly as in the case
of incinerators.
Two examples will illustrate these ratings.
-------
IJI-5
Efficiency of Pollution Control Devices
Other Pollution Control Devices (continued)
Example 1 - Assume that a catalytic cracker regenerator effluent
contains 100 pounds of catalyst dust, 200 Ibs, of
carbon monoxide and 10 pounds of sulfur oxides per unit
time. It is passed through a cyclone separator where
95 pounds of catalyst are removed. Therefore,
SE = 100 - 5
X 100 = 95%
and SERR = (100 x 60 + 10 x 20 + 200 x 1) - (5 x 60 + 10 x 20 + 200 x 1) x 100
(100 x 60 + 10 x 20 + 200 x 1)
= 6400 - 700 x 100 = 89%
6400
Example 2 - Assume that an organic liquid scrubber is used to wash a
stream containing 50 pounds of S02 per unit time. All
but one pound of the S02 is removed but two pounds of
the hydrocarbon evaporate into the vented stream. Then
and SERR = (50 x 20) - (1 x 20 + 2 x 80)
(50 x 20) - x 10°
x 100 = 82%
-------
APPENDIX IV
PHTHALIC ANHYDRIDE PRODUCTION FROM NAPHTHALENE
-------
PAN-1
I. Introduction
Initially all phthalic anhydride vas derived form naphthalene. As
early as 1896 BASF patented a process vhereby naphthalene vas oxidized to
phthalic anhydride in a solution of sulfuric acid. Today all U. S.
naphthalene based phthalic anhydride plants employ vapor phase processes,
vith the majority utilizing the Sherwin-Williams/Badger fluid bed technology.
In terms of current production capacity, the naphthalene process is about
on par with the newer o-xylene process (see Table PA-3). However, all future
growth is expected to be based on o-xylene methodology.
The primary source of emissions from either process is the switch
condenser vent. In the naphthalene process, the pollutants associated with
this stream are phthalic anhydride, maleic anhydride, naphthoquinone and carbon
monoxide. The fluidized bed reactor used in most naphthalene based plants is
(indirectly) another source of emissions, since some of the various items of
process equipment required for catalyst storage, transportation, etc, are
responsible for the dispersion of catalyst fines. Additionally, other solid
and liquid-form wastes are produced.
The current U. S. phthalic anhydride ex naphthalene production capacity
(active) is 528 MM Ibs./year. 1985 production capacity is estimated to be
essentially the same.
-------
PAN-2
II. Process Description
Naphthalene may be oxidized to phthalic anhydride in the presence of
an appropriate catalyst. This reaction is shovn as follovs:
2 H20 + 2 C02
Naphthalene Phthalic Anhydride
Mol. Wt. 128.2 148.1
Standard commercial practice is to conduct the reaction in the vapor
phase utilizing a vanadium pentoxide catalyst.*
Naphthalene and air are introduced into the fluidized bed reactor near
the bottom of the catalyst bed. The naphthalene vaporizes immediately and
in the presence of the catalyst and air is oxidized to phthalic anhydride.
The reactor bed temperature is controlled at 650 to 725° F. The exothermic
heat of reaction is removed by cooling tubes, located within the catalyst bed.
The heat is used to produce high pressure steam.
The effluent from the reactor consists of phthalic anhydride vapors,
entrained catalyst and various by-products and non-reactant gases. The
catalyst is removed from the effluent by a series of filters and returned to
the reactor. The phthalic anhydride is removed by condensation.
Purification of the crude phthalic anhydride generally involves tvo
steps. First the crude product is given a "heat treatment" Cheld at elevated
temperature) to boil off vater and to allow contained impurities to form
condensation products. Final purification is by standard distillation. The
product phthalic anhydride may be marketed in the molten form or solidified
and sold as flakes.
*See Figure PAN-1, Table PAN-I and Table PAN-II for a simplified process flow
diagram, a typical process material balance, and a gross reactor heat balance
respectively.
-------
PAN-3
-------
TABLE PAH-1
Coaaonent
Naphthalene
Phthallc Anhydride
Wlelc Anhydride
Naphthoeulnone
Jttec. Organic!
Oxygen
Nitrogen
Carbon Dioilde
Carbon Monoxide
Hater
Total Ton*/Ton of PAN
Miphthala
faad
1.030*
.0052
1.0361
TYPICAL MttIB
ntoM nnm
2 3
ALiAlAUCE
n PBMU.IC AHHTDRIDE
n^nn
456
Air to rroee»i Pretreataant Dlftlllation Dittlllatlon
Reactor Vant Oil Light Enll Light Ball BottoM
.020* 0075 .0007 .0066
.0070
.0007
.0049
.0066
.0052
3.3075 l.W2f .tOSO
10.1*25 10. IMS .0030
.9667
.0504
.3150 .(
MM
13.5000 13.4926 .0175 .0056 .0184
Phthallc Anhydride
Product
1 0000
0020
i.
-------
PAN-5
TABLE PAN-II
GROSS REACTOR HEAT BALANCE*
FOR PRODUCTION OF
PHTHALIC ANHYDRIDE
EX
NAPHTHALENE
HEAT IN BTU/LB. OF PAN
Exothermic Heat of Reaction** 7300
Heat Naphthalene 330
Heat Air 2100
9730
HEAT OUT
Reactor Heat Loss 30
Steam Generation 7000
Reactor Effluent Heat Content*** 2700
9730
*Basis
(1) Table PAN-1 Material Balance.
(2) Napthalene feed at 200° F, air at 80° F.
(3) Reactor outlet 700° F.
**Normal range is from 7,000 to 9,000 BTU/lb. of PAN, depending primarily on
selectivity.
***Difference in heat content of effluent at 700° F and feed naphthalene
at 200o F (molten) plus air at 80° F.
-------
PAN-6
III. Plant Emissions (For details see Table PAN-III)
A. Continuous Air Emissions
1. Switch-Condenser Vent
Emissions from this vent stream represent well over 95 percent
of total reported emissions at every plant surveyed. This is true
in spite of the fact that each plant utilizes reasonably efficient
control devices to minimize pollution from this source. Plants
53-6 and 53-8 both rely on thermal incinerators, which operate
at 80 and 85 percent (SERR) efficiency, respectively. Plant 53-7
water washes the switch-condenser vent stream with a combination
Venturi/spray tower scrubber. On a Ib./lb. basis, hydrocarbon (or
particulate) emissions from plant 53-7 are only about 1/5 of those
reported by plants 53-6 and 53-8.
2. Heat Treatment Section Vent
Plants 53-6 and 53-8 both show emissions from this section of
the product purification 'train1. It is presumed that all plants
that utilize 'heat treatment1 techniques will have associated
emissions. Plant 53-6 incinerates these vapors in the switch-
condenser incinerator. Plant 53-8 washes this vent stream in a
separate scrubbing unit. Emissions after scrubbing amount to
.00005 Ibs. of PAN (emitted)/Ib. of product.
B. Intermittent Air Emissions
1. Emergency Vent
Plant 53-6 is the only plant reporting emergency vent streams
(two). Both are switch-condenser incinerator by-passes, utilized
during emergency shut-downs. The respondent reports that they are
used only a few minutes per year. Consequently, emissions from this
source are negligible.
2. Catalyst Storage Hopper Vent
Plant 53-7 is the only plant reporting emissions from this
source, although one would presume that all (fluid bed) operators
would have similar vents. Emissions occur when catalyst is removed
from the reactor and stored in a hopper. The efficiency of the
associated cyclone and the frequency and duration of the transfer
operation maintain emissions below .00001 Ibs./lb. of product.
3. Spent Catalyst Removal
Again, only plant 53-8 reports these emissions. They are similar
to those mentioned above except they relate to spent catalyst;
emissions are less than .00001 Ibs./lb. of product.
-------
TABLE PAM-III
NATIONAL EMISSIONS INVENTORY
PHTHALIC ANHYDRIDE PRODUCTION
«
NAPHTHALENE
Sheet 1 of 2
Flint - EM Code No.
Capacity - Toot of Phthalle Anhydride/Yr.
Range In Production - X of Mix.
Emissions to Atmosphere
Stream
Flov - Lba./Hr.
Flov Characteristic - Contlnuout or Intermittent
If Intermittent - Hrs./Yr. Flow
Composition, Tom/Ton of Phthallc Anhydride
Nitrogen
Oxygen
Carbon Monoxide
Carbon Dioxide
Water
Phthalle Anhydride
Malelc Anhydride
Naphthoqulnone
Argon
Catalyst Flnei (VjOj)
Vent Stack*
Height - Ft.
Dlamster - Inchat
Exit Gat Temp. - F°
SCFM/Staek
Emission Control Devices
Incinerator
Scrubber
Cyclone
Analycli
Data or Frequency of Sampling
Sample Tap Location
Type of Analyali
Odor Problem
Summary of Air Pollutant!
Hydrocarbons
Aerosols & Partlculatas
*>*
*°x
CO
53-6
62,500
0
Reactor
Section
Emergency Vent
197,167
Intermittent
Q.I
.00012
.00002
4.00001
4.00001
4.00001
4.00001
4.00001
4.00001
(Flue Gas Stack)
None
No
Calc'd.
No
Purification
Section
Emergency Vent
2,413
Intermittent
0.1
4.00001
4\. 00001
41.00001
4.. 00001
(Flue Gas Stack)
None
No
Calc'd.
No
Incinerator
Flue
Gas
SM.5I3
Continuous
8.68213
1.37607
.04247
.91267
2.34947
.00300
.00033
.00007
Yes
1
100
60
450 - 500
48,000
PA- 101
Yes
Several/year
At Stack
GLC & Titr.
No
.00340
.04247
53-7
45.000
0
Scrubber
Vent
Gas
9). 0*7
Continuous
8.02087*
.07130
.00048
.00004
.00009
Tes
1
80
36
100
20.000
PA- 102
No
Calc'd.
No
.00061
.07130
^Represents total flow of nitrogen, oxygen, carbon dioxide and water.
-------
Plant - EM Code No.
Capacity - Tona of Rithallc Anhydride,'Yr.
Range in Production - t of Max.
Emissions to Atmosphere
Stream
Flov - Lb§./Hr.
Flov Charactarlatic - Continuous or Intermittent
if Intermittent - Hrf./Yr. Flov
Composition - Tona/Ton of Phthalic Anhydride
Nitrogen
Oxygen
Carbon Monoxide
Carbon Dioxide
Water
Fhthalic Anhydride
Ma).eic Anhydride
Naphthoquinone
Argon
Catalyst Fines (V205)
Vent Stacks
Number
Height - Ft.
Diameter - Inches
Exit Gas Temp. - F°
SCFM/Stack
Emission Control Devices
Incinerator
Scrubber
Cyclone
Analysis
Date or Frequency of Sampling
Sample Tap Location
Type of Analysis
Odor Problem
Summary of Air Pollutants
Hydrocarbons
Aerosols & Partlculates
CO
BLE PAH-III (CONTINUED)
PCSSIOBS mmiuau
PHTHALIC ANHYDRIDE PROPnCTIOH
EX
NAPHTHALENE
Sheet 2 of 2
53-8
45,000
0
Incinerator
Flue
Gas
126.800
Continuous
8.81818
1.1(335
.04653
.66311
. 79108
)
) .00257
. 15124
Heat
Treater
Vent
U
Continuous*
.00459
.00275
.00005
Yes Yes
1 1
60 | 70
7
550
28,000
PA-103
Yes
Several/year
Stack
M.S. , I.R. , FL. lonir.
No
3
250
20
PA- 104
Stack
Estimate
Ho
Sprat
Catalyst
fant
4,450
Intermittent
50
). 00333
)
•^.otooi
Yes
1
10
30
250
1,250
PA- 105
No
Estimate
No
.00262
.04653
Catalyet
Storage Hopper
Vent __
60S
Intermittent
50 - 100
) . 0005
^.OO001
Yes
1
60
36
350
7,500
M-106
No
Estimate
No
*Flov is continuous but flov rate and comoositlon vary cyclically.
-------
PAN-9
C. Continuous Liquid Wastes
The respondents reported the following:
Stream Flow Treatment
Waste Water 170 GPH Discharged to treated water
system settling basin.
53-7 Water ex Scrubber 6000 GPH "To Treatment"
53-8 Waste Water 4200 GPH To plant waste water
treatment unit.
No other waste liquid streams were reported.
D. Solid Wastes
The respondents reported the following:
Material Amount Disposal Method
Light Ends 58,000 Ibs./mo. Trucked away
Heavy Ends 400,000 Ibs./mo. By contractor
53-7 None Reported
53-8 Waste Solids 45,000 Ibs./day Plant landfill
E. Odors
In general, the production of phthalic anhydride from naphthalene
does not appear to present an odor problem.
None of the respondents reported an odor complaint in the past
year. Of the three plants surveyed only one (plant 53-8) reported
that emission odors were ever detectable off the plant property.
The odorous material was identified as phthalic anhydride and
phthalic anhydride partial decomposition products.
F. Fugitive Emissions
None of the respondents offered an estimate of fugitive emissions.
All indicate that molten product is stored in tanks that vent directly
to the air, i.e., no vapor conservation devices are employed.
G. Other Emissions
All respondents burn fuel gas. Operator 53-8 reports his gas
(ethane) contains no sulfur. Operator 53-7 does not report the sulfur
content of the gas he burns. Operator 53-6 states that the natural
gas he utilizes has a sulfur specification of .001 percent max. This
results in the emission of .000001 Ibs. of sulfur/Ib. of product.
However, these comments relate to 1972 plant operations. Future use
of other fuels could result in different SOx emissions.
-------
PAN-10
IV. Emission Control
The various emission control devices that are employed by operators of
naphthalene process phthalic anhydride plants are summarized in the 'Catalog
of Emission Control Devices'. Table PAN-IV. Device efficiencies are reported
variously as SE, SERR and CCR. A definition of these terms may be found in
Appendix III of this report.
Two types of devices are used on the main process vent strean; water
scrubbers and incinerators. Both plants 53-6 and 53-8 utilize thermal
incinerators. The efficiency of the combustion device is lower than one might
expect. This results, to some extent, from the very low conversion of CO to
CC>2 - at both plants. The flue gases do, however, contain significant amounts
of hydrocarbons, perhaps attesting to the difficulty of burning organic
particulates. Plant 53-7 employs a rather complex water scrubbing system.
Although reported data do not permit calculation of that scrubber's efficiency,
it apparently is quite effective in removing hydrocarbons; emissions of that
type are only about 1/5 of those reported by the plants utilizing incineration.
On the debit side, the scrubber probably removes none of the CO from the vent
stream.
Respondents 53-6 aid 53-8 also report the use of cyclones on the bulk
transfer system, to control catalyst dust emissions. These devices perform
with an efficiency of 90-98 percent. .
There is the possibility that some minor reduction in air emissions could
be achieved through the use of purer raw material. One operator (53-7) reports
using 90 MM Ibs./year of 99.5 percent naphthalene as feed, while another (53-6)
reports using 111 MM Ibs./year of 97 percent naphthalene as feed, obviously
the second of these units is required to dispose of 2.8 x 106 Ibs./year of
additional organic material. Most of this additional material is burned to
C02 and water. Unfortunately, the dissimilarity of pollution control
equipment employed by the two plants precludes an estimate of that anount by
comparison of total emissions from those plants.
-------
TABLE PAN-IV
CATALOG OF ferfgfi ION CONTROL DEVICES
PHTHALIC ANHYDRIDE PRODUCTION
INCINERATION DEVICES
EM Cod* No. for plant using
Device I. D. No.
Type of Compound Incinerated
Type of Device - Flere
Incinerator
Other
Material Incinerated - SCFM (Ib./hr.)
Auxiliary Fuel Req'd. (excl. pilot)
Type
Rate - BTU/hr.
Device or Stack Height - Ft.
Installed Cost - Mat'l. & Labor - $
Installed Cost based on - "year" - dollars
Installed Cost - c/lb. of PAN/Yr.
Operating Cost - Annual - $ (1972)
Value of Beat/Steam Recovered - $/Yr.
Net Operating Cost - Annual
Net Operating Cost - C/lb. of PAN
Efficiency - I - CCR
Efficiency - Z - SERR
NAPHTHALENE
53-6
PA-101
Hydrocarbons
Natural Gas
24,000 (SCFH)
100
280,000
1968
.2240
292.000
0
292,000
.2336
62
80
Sheet 1 of 2
53-8
PA-103
Hydrocarbons
(293)
Fuel Gas
12 MM
60
250.000
1969
2777
55.000
0
55,000
.0611
35
85
CYCLONES
EPA Code No. for plant using
Device I. D. No.
Controls Balsslon of
T-T Height - Ft.
Dlaaater - Ft.
No. of Stages
Installed Cost - Mat'l. & Labor - $
Installed Cost based on - "year" - dollars
Installed Cost - c/lb. of PAN/Yr.
Operating Cost - Annual - $ (1972)
Value of Recovered Product - $/Yr.
Net Operating Cost - Annual - $
Net Operating Cost - c/lb. of PAN
' Efficiency - % - SE
Efficiency - 7. - SERR
53-8
PA-105
Catalyst
1
2000
1962
.0022
3200
0
3200
.0036
98
98
53-8
PA-106
Catalyst
T
1961
0
0
0
90
90
-------
IOS
PHBUXIC AHHTPMBK ItOPOCTIOM
ABSORBER/SCRUBBER
EM Cod* Ho. for plant utlng
Device I. D. Mo.
Control Union of
ScrubblM/ABaorblng liquid
Ty»« - Spray
peeked Coluan
Coluan w/tray*
amber of tray*
tray type
Other
Scrubbing/Absorbing Liquid Rate - cm
Daiign Tea*. (Operating Tea*.) P<>
Gaa Rete, SCFM (Ib./hr.)
T-T Might - Ft.
Maa*ter - Ft.
Vaihed Gaea* to Stack
Steek Height - Ft.
Stack Meatier - Inch**
Installed Co*t - Wt'l. & Labor - $
Installed Coat baaed on - "year" - •oiler*
Installed Coat - c/lb. of HLV/Tr.
Operating Coat - Annual - $ - 1972
value of leee»*iad Product - $/Tr.
Net Operating Coat - Annual - $
Nat Operating Coat - c/lb. of IAN
Efficiency - I - SE
Efficiency - I - SERR
53-8
PA-104
Hydrocarbon*
Water
X
(200 - 250)
20
Ye*
70
3
*,000
1966
.0100
13,000
0
13,000
.0144
98.8
W.8
Sheet 2 of 2
53-7
PA-102
Hydrocarbon*
Plu* venturl, aaparator, mitt elladnator.
125 Total
(100)
20,000
Ye*
80
36
139,000
1953 to 1968
.1144
64,500
0
64,500
.0717
etc.
-------
PAN-13
V. Significance of Pollution
Within the context of this report, 'significance of pollution1 is
related solely to emissions associated with production facilities constructed
in the period 1973 to 1980. Since no growth in capacity is forecast for
the naphthalene based process, then the subject of 'significance of pollution'
is not relavent. Thus, this abreviated report has been appended to the
in-depth study of the 'growth' process for the production of phthalic
anhydride - the o-xylene based process.
The method of calculating significance of pollution is described in
Appendix II of this report and its application to phthalic anhydride production
is illustrated in Table PAN-V.
-------
PA-63
References
1. "Exhaust Gases from Combustion & InduFtrial 'Processes", Engineering Science
Inc., EPA Contract No. EHSD 71-36, October 2, 1971.
2. "Phthalic Anhydride", Hydrocarbon Processing, page 188, November, 1971.
3. Kirk-Othmer; "Encylcopedia of Chemical Technology", 2nd Edition, VoL 15
(1968).
4. "1971 Directory of Chemical Producers - USA", Chemical Information
Services, Stanford Research Institute.
5. "Phthalic Anhydride Chemical Profile". Chemical Marketing Reporter,
August 16, 1971.
6. Fawcett, R. L., "Air Pollution Potential of Phthalic Anhydride Manufacture",
Journal of the Air Pollution Association, Vol. 20, 461*465 (July, 1970).
7. Danielson, J. A., "Air Pollution Engineering Manual, Air Pollution Control
District County of Los Angeles", U. S. Department of Health, Education
and Velfare, Cincinnati, Ohio, 1967, pages 177 & 178.
8. Rolke, R. W., et al, "Afterburner Systems Study", by Shell Development
Company for Environmental Protection Agency ''Contract EHS-D-71-3).
9. "Chemical Economics Handbook", Stanford Research Institute, February, 1970.
10. "Hazardous Waste Air Harnessed to Produce Process Steam", Chemical
Processing, page 12, August, 1971.
11. "Phthalic Anhydride by Vapor-Phase Oxidation", The Oil and Gas Journal,
page 92, March 12, 1973.
12. Ockerbloom, N. E., "Xylenes and Higher Aromatics Part 3: Phthalic Anhydride",
Hydrocarbon Processing, page 162, September, 1971.
13. Spitz, P. H., "Phthalic Anhydride Revisited", Hydrocarbon Processing.
page 162, November, 1968.
14. Emmett, P. H., "Catalysis", Vol. VII, Reinhold Publishing Corporation,
Nev York, N. Y., 1960, pages 212 - 217.
15. Schwab, R. F. and Doyle, W. H., "Hazards in Phthalic Anhydride Plants",
Chemical Engineering Progress, page 49, September, 1970.
16. Hahn, A., "The Petrochemical Industry: Markets and Economics", McGrav Hill,
Inc., Nev York, N. Y., 1970.
17. Graham, J. J., "The Fluidized Bed Phthalic Anhydride Process", Chemical
Engineering Progress, page 54, September, 1970.
-------
PA-'64
References (Continued)
18. "Atmospheric Emissions for Sulfuric Acid Manufacturing Processes",
Public Health Service Publication No. 999-AP-13, 1965.
19. "Standards of Performance for Nev Stationary Sources", Federal Register,
Vol. 36, No. 247, 24876-24895, December 23, 1971.
20. "Guide for Air Pollution Episode Avoidance", Environmental Protection
Agency, Office of Air Programs, Publication No. AP-73, June, 1971.
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing}
1. REPORT NO.
EPA-450/3-73-006-g
3. RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
Engineering and Cost Study of A1r Pollution Control
for the Petrochemical Industry, Volume 7: Phthalic
Anhydride Manufacture from Qrr.hrn
1. AUtHOR(S)
5. REPORT DATE
July 1975
6. PERFORMING ORGANIZATION CODE
7. AUtHOR(S)
W. A. Schwartz, F. B. Higgins, Jr., J. A. Lee,
R. B. Morris, R. Newirth, J. W. Pervier
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG 1NIZATION NAME AND ADDRESS
Houdry Division/Air Products and Chemicals, Inc,
P. 0/Box 427
Marcus Hook, Pennsylvania 19061
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-0255
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
EPA, Office of Air Quality Planning & Standards
Industrial Studies Branch
Research Triangle Park, N.C. 27711
Final Report
14. SPONSORING AGENCY CODE
16. SUPPLEMENTARY NOTES
16. ABSTRACT
This document is one of a series prepared for the Environmental Protection
Agency (EPA) to assist it in determining those petrochemical processes for which
standards should be promulgated. A total of nine petrochemicals produced by
twelve distinctly different processes has been selected for this type of in-depth
study. Ten volumes, entitled Engineering and Cost Study of A1r Pollution Control
for the Petrochemical Industry (EPA-450/3-73-006a through .1) have been prepared.
A combination of expert knowledge and an industry survey was used to select
these processes. The industry survey has been published spearately in a series of
four volumes entitled Survey Reports on Atmospheric Emissions from the Petrochemical
Industry (EPA-450/3-73-OoSa, b, c, and d). ~~
This volume covers the manufacture of phthalic anhydride from ortho-xylene.
Included is a process and industry description, an engineering description of
available emission control systems, the cost of these systems, and the financial
impact of emission control on the industry. Also presented are suggested air
episode procedures and plant inspection procedures.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Air Pollution
Hydrocarbons
Phthalic Anhydride
Petrochemical Industry
7A
7B
7C
116
13B
13H
18. DISTFIBUTION STATEMENT
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
108
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE . , ., .
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
6. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g.. date of issue, date of
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Leave blank.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented at conference of,
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or lea) factual summary of the most significant information contained in the report. If the report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COS ATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19. &20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (9-73) (Revtrie)
-------
ENVIRONMENTAL PROTECTION AGENCY
Technical Publications Branch
Office of Administration
Research Triangle Park, North Carolina 27711
POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
EPA 335
OFFICIAL BUSINESS
AN EQUAL OPPORTUNITY EMPLOYER
Return this sheet if you do NOT wish to leceive this material
01 if change of address is needed Q]. (Indicate change, including
ZIP code.)
PUBLICATION NO. EPA-450/3-73-006-g
------- |