STATIONARY SOURCE ENFORCEMENT SERIES
TECHNICflL SUPPORT DOCUfllENT
FOR THE RECOmmENDED
LOWEST ACHIEVABLE EfTllSSION RATE
FOR SO2 EfniSSIONS FROm
-OVEN GAS COmBUSTION
\
A Hi :' '
U.S. ENVIRONMENTAL AGENCY
OFFICE OF ENFORCEMENT
DIVISION OF STATIONARY SOURCE ENFORCEMENT
WASHINGTON, DC 20460
JANUARY 1977
-------
TECHNICAL SUPPORT DOCUMENT FOR THE RECOMMENDED
LOWEST ACHIEVABLE EMISSION RATE FOR SO2 EMISSIONS
FROM COKE-OVEN GAS COMBUSTION
Prepared by the U.S. Environmental Protection Agency,
Division of Stationary Source Enforcement
January 1977
-------
This report was developed by the Technical Support Branch of the
Division of Stationary Source Enforcement between November 1976 and
January 1977. Bernard Bloom was the principal investigator.
Material in this report and its appendices is subject to the
Environmental Protection Agency business confidentiality regulations
(40 CFR Part 2, Subpart B 41 Fed. Reg. 36906, et. seq.). This document
has been circulatd to EPA offices and not to vendors, coking companies,
state or local agencies or other members of the public. Freedom of
Information Act or other requests for this report from outside EPA
should be forwarded to the Director, Division of.Stationary Source
Enforcement, 401 M Street, S.W., Washington, D.C. 20460. If distri-
bution to a member of the public is desired by an EPA office, the
Director of Division of Stationary Source Enforcement should be
notified before such disclosure is made.
-------
TABLE OF CONTENTS
PAGE
LIST OF FIGURES IV
LIST OF TABLES V
Section I. INTRODUCTION ' 1
Section II. CONTROL TECHNOLOGY 8
A. Liquid Absorption Processes 8
1. Vacuum Carbonate 8
2. Sulfiban 15
3. Carl Still 19
4. Diairox 24
B. Sulfur Recovery Processes 24
1. Glaus 24
(a) SCOT ' 27
(b) Beavon 28
(c) IFF 28
(d) Vfellman-Lord 28
.2. Sulfuric Acid 31
C. Liquid Oxidation Processes 32
1. Stretford 33
2. Takahax 37
Section III. ORGANIC SULFUR IN COKE-OVEN GAS 45
Section IV. PERFORMANCE LEVELS 48
Section V. MEASUREMENT METHOD 59
-------
TABLE OF CONTENTS
-(Continued)
PAGE
Section VI. IMPACT OF CONTROL TECHNOLOGY 65
1. Cost Inpact 65
2. Energy Impact 76
REFERENCES 79
APPENDICES
A. List of Vendors 82
B. J & L Steel Organic Sulfide Data 84
C. Process Appendix (separate folder)
-------
LIST OF FIGURES
Figures
1. Coke-Oven Gas Distribution in an Example Steel Plant
2. Vacuum Carbonate Process Flowsheet
3. Two-Stage Vacuum Carbonate Process Flowsheet
4. (^S) vs. MEA Circulation Rate for Sulfiban at BSC, Bethlehem, PA.
5. Carl Still Process Flowsheet
6. Clairton Works Sulfur Recovery Flowsheet
7. Stretford Process Flowsheet
8. Takahax Process Flowsheet
9. Liquid Absorption Options for COG Desulfurization
10. Compliance Measurement locations
IV
-------
LIST OF TABLES
Tables
1. Composition of Coke-Oven Gas
2. Technologies for Coke-Oven Gas Desulfurization
3. Glaus Plant Tail Gas Treatment Technologies
4. Vacuum Carbonate Plants in U.S.
5. Sulfiban Plants in U.S.
6. Sulfiban Performance Data
7. Firma Carl Still Plants in U.S.
8. Construction Records of Large Scale Units in Japan
9. Relative Takahax Costs
10. Takahax Performance Data
\
11. Takahax Waste Air Stream Flows
12. Organic Sulfide Concentrations in COG
13. Liquid Absorption Process Comparisons
14. Sulfur Recovery Performance Comparisons
15. Overall Best Obtainable Performance by Liquid Absorption Systems
16. Liquid Oxidation Process Comparisons
17. Ranking of Coke-Oven Gas Desulfurization Technology Performance Levels
18. COG Sampling Field Worksheet
19. Costs for COG Desulfurization Systems at Republic Steel - Cleveland
Works, $103
20. Costs for 60 x 10 scfd COG Desulfurization Systems at YS & T,
$103, May 1976
21. Battery Limits Costs for Vacuum Carbonate Sulfiban, and Carl Still
COG Desulfurization Systems
V
-------
LIST OF TABLES
(Continued)
Tables
22. Recent Vendor Bid
23. Comparison of COG Desulfurization Costs
24. Comparison of Alternative Emissions Control System Costs for
a 10 LTD Sulfur Plant
25. Energy Demands for Coke-Oven Gas Desulfurization Technologies
26. Relative Energy Demand of COG Desulfurization Technolgies
VI
-------
Section I. INTRODUCTION
This document presents a description of technology for the desulfurization
of coke oven gas. It is intended to provide guidance to EPA technical staff
and policy makers in the implementation of significant deterioration and new
source review policies of EPA. The information is also intended to be useful
in the development, revision, or the enforcement of existing State Implementation
Plan (SIP) standards. This document has been prepared in conjunction with an
Office of Enforcement memorandum, "Guidance for Establishing the Lowest Achiev-
able Emission Rate for S02 from the Combustion of Coke-Oven Gas, January 5, 1977."
EPA has not yet developed a New Source Performance Standard (NSPS) for this
source category.
Coke-oven gas (COG) is produced during the coking of metallurgical coal
in by-product coke ovens. As a result of the coking process a fraction of
the sulfur contained in this coal (25-30%) is transferred to COG in the form
of hydrogen sulfide (H-S), carbon disulfide (CS2), and carbonyl sulfur (COS).
Upon subsequent combustion these gases release S02 to the ambient air. It is
the purpose of COG desulfurization systems to reduce the amount of SO2 emitted
to the ambient air by the removal of sulfur compounds from COG prior to com-
bustion.
COG principally consist of hydrogen and light hydrocarbons, as illustrated
in Table 1. Depending upon its exact composition, which in turn depends upon
coal analysis and coke oven operation, COG has a heating value of 500-560
Btu/scf and an average molecular weight of about 10. Very iirportantly it typi-
cally has a sulfur content of 250-600 gr H2S/100 dscf and 5-25 gr/dscf of total
-------
~ '-'.''
COKE
Component
Hydrogen ...
Methane
Nitrogen
Carbon Monoxide
Ethylene
Carbon Dioxide
Ethane
Oxygen
Hydrogen Sulfide
Benzene
Propylene
Propane
Acetylene
Naphthalene
Carbonyl Sulfide
Carbon Disulf ide
Hydrogen Cyanide
Argon
TABLE 1
OVEN GAS ANALYSIS
Range, %
55.83-59.69
24.28-26.94
4.52- 8.94
3.78- 5.24
2.01- 2.31
1.58- 2.02
0.68- 0.82
0.38- 0.87
0.38- 0.59
0.01- 0.19
0.12- 0.17
0.06-0.12
^0.04- 0.10 ''';-
0.00- 0.02
\ ' .'.'*'..
.0.006
0.0009 ,
0.008- 0.12 ,
- traces
.'-'- . - ' *'
Mean, %
57.69
25.40
6.67
4.25
2.16
1.72
0.76
0.59
0.43
0.11
: 0.14
0.08 .
O.08
0.01
0.006
/ 0.0009
0.10
traces
2
-------
organic sulfur (the sum of RSH, COS, & CS-). One useful rule of thumb is that
the H2S concentration in COG, expressed in gr/100 dscf, is approximately 365 x %S
in coal . The volume of COG produced in the coking process is in the range
(2)
10,000 - 13,000 scf per ton of coal charged , again ranging with initial
coal volatility and coking practice. Therefore, a coke battery producing
500,000 tons of coke per year (e.g., a four meter tall, 77 oven battery) will
produce about 8.6 billion scf of COG annually. S02 emissions are in direct
proportion to COG consumption. When burned by itself, uncontrolled, coke-
oven gas will produce 5.7 - 11.5 tons of SO2 per 1000 tons of coke pushed
at 250-500 gr H2S/100 dscf, respectively. A 100 tpy SO2 emission rate
from COG combustion is equivalent to coking 8700 tons per year of coal
producing COG of 500 gr H2S 100 dscf or 87,000 tons per year at 50 gr H2S/
100 dscf. New coke production in excess of 100,000 tons/year will produce
more than 100 tpy SO2, even within the allowable level of all SIP's in the
United States. Hence, even under the most stringent SIP, new coke battery
construction is subject to New Source Review IAER criteria in SO2 non-attain-
ment area.
It should be understood that coke-oven gas is almost never consumed at
one point in a coke plant. Where coke batteries are heated (underfired) with
COG, typically, only 40-45% of total COG production is needed for this purpose.
Distribution to in-plant boilers, steel heating furnaces, open hearth furnaces,
COG flairs, and to out-of-plant consumers is the rule. One average distribu-
tion for a major steel producer is shown in Figure 2. Therefore, desulfuriza-
tion is effected at the point of generation by removal of H-S and organic
sulfur. The existence of a distribution system for coke-oven gas poses a
-------
BY-PRODUCT
BOILERS
12%
_
BLAST
FURNACE
OPEN
HEARTH
SLAB HEATING
FURNACE
10%
ELIZA
BOILERS
8%
SOUTH SIDE
BOILERS
3%
ANNEALING
GALVANIZING
PI
BATTERY
8%
J&L
COKE-OVEN GAS
P2
BATTERY
8%
P3N
BATTERY
8%
_
"C" SOAKING
PITS
10%
"D" SOAKING
PITS
6%
10" BAR
MILL
3%
14" BAR
MILL
2%
MISCELLANE-
OUS
3%
FIGURE! 1973-1974 COG DISTRIBUTION OF J&L
PITTSBURGH WORKS
-------
complex issue for the EPA regional engineer. At each use point, SIP emissions
standards for SO- may exist. An industrial boiler may be limited to 1.2 Ib
SCL/10 Btu; a soaking pit to 1000 ppm S02- On an average basis then, it is
conceptually possible for SO2 emissions from COG combustion within an in-
tegated steel mill to comply with an existing SIP despite the non-existence
(4)
of a direct COG regulation. Dunlap & Massey note that 500 gr H-S/lOO dscf
is approximately equivalent to a 1.3% S coal. Upon combustion in a boiler
this fuel will produce about 2 Ib SO-/10 Btu; when burned in a reheat furnace
the waste gas stream will contain approximately 1000 ppm SCL. However, since
all new coke-oven gas will be consumed, the Offset Policy applicability deter-
mination shall be made for the total gas volume itself. Each end use stream,
of which an integrated mill may have dozens, is not to be separately considered
against the 100 tpy criterion.
-------
STATE IMPLEaVENTATICN PLAN REQUIREMENTS FOR COKE-OVEN GAS DESULFURIZATION
The SIP's for California, Kentucky, Pennsylvania, Ohio, New York, and
West Virginia require COG desulfurization. Coke plants are also located in
states which do not have SIP's for coke-oven gas. These are Illinois, Indiana,
Alabama, Tennessee, Minnesota, Wisconsin, Colorado, Michigan, and Utah. The
most stringent SIP standards are shown below:
COG
TAIL GAS
TOTAL
California
Pennsylvania
Kentucky
Lorain, Ohio
50 gr/100 scf(a) 500 ppm S02
10
(a)
2000 ppm SO-
50 gr/100 scf
(b)
-V20-35 gr/100 scf
35 (b)
(c)
(a) H2S
(b) Total Sulfur
(c) No regulation per se. This is the equivalent total.
-------
The Kentucky coke-oven gas SIP is apparently the most stringent in the
U.S. However, it applies only to "Priority I" areas, which are defined by
the SIP as having (S02) ambient levels in' excess of 0.04 ppm (annual) and
0.17 ppm (24-hr). The 2000 ppm process emission limit, when applied to acid
or Glaus plant tail gas implies a sulfur recovery efficiency of 95-98%, de-
pending upon process details. The Kentucky SIP is equivalent to 20-35 gr
ELS/100 dscf of COG produced. To comply with the Kentucky SIP all of the
liquid oxidation and absorption methods of Section II are available. High
efficiency sulfur recovery is necessary to comply with this regulation if
liquid absorption is chosen.
The Pennsylvania SIP regulates total sulfur, as H2S. Hence, for an
organic sulfide level of 15-20 gr/100 scf, this regulation is as stringent
in overall allowable SCL as the Kentucky SIP. But, the Pennsylvania rule
is more flexible in that only the total equivalent SO2 emission rate is
fixed. The rule applies Commonwealth-wide. The Ohio SIP for U.S. Steel,
Lorain requires certain process streams to the less than 35 gr "H2S"/100 dscf
overall, including tail gas emissions. This regulation is about as stringent
as the Kentucky SIP.
The California SIP is the most stringent with respect to tail gas emis-
sions, requiring 99.5% sulfur yield. In fact, Kaiser Steel chose a liquid
oxidation-no tail gas process in order to comply with this rule. The rule
allows 50 gr/100 scf of H2S in the COG and places no organic sulfide limit.
-------
Section II. CONTROL TECHNOLOGY
Each of the technologies for coke-oven gas desulfurization involves two
separate steps: (1) the removal or stripping of H2S and related sulfur com-
pounds from the coke-oven gas and (2) the recovery of the stripped compounds
as elemental sulfur, sulfuric acid, or ammonium sulfate. Available systems
fall into two broad categories: liquid absorption followed by Glaus or acid
plant sulfur recovery or liquid absorption plus liquid phase oxidation of re-
duced sulfur gases. Altogether, there are at least six basic technologies
commercially available in the U.S. for removing reduced sulfur from coke-oven
gas as shown in Table 2. In addition, there exist a number of technologies
for recovering this sulfur. Improving the sulfur recovery from both Claus
plant tail gas (via tail gas treatment) and sulfuric acid plants must be
considered as part of these desulfurization technologies. These processes
are shown in Table 3.
A. LIQUID ABSORPTION TECHNOLOGIES
1. Vacuum Carbonate Process
This process uses a solution of sodium carbonate to wash countercurrently
an upward rising flow of COG in an absorption tower. The absorber removes H2S,
HCN, and C02 but not COS or CS2, from the coke-oven gas. The rich solution is
then steam stripped in a second tower, called the actifier, which releases the
acid gases overhead and regenerates the lean absorbing solution. In the Vacuum
Carbonate process the steam, stripping is accomplished under partial vacuum, in
order to lessen the steam demand. The basic process flowsheet is shown in
Figure 2.
8
-------
Table 2. COKE OVEN GAS DESULFURIZATICN TECHNOLOGIES
STEP 1 - SULFUR REMDVAL
STEP 2 - SULFUR RECOVERY
PRINCIPAL U.S. VENDORS
Liquid Absorption
Vacuum Carbonate
Sulfiban
Diamox
Carl Still
-Glaus Process
Sulfur Recovery
or
Sulfuric Acid Production
Glaus, Acid or Stretford
Hoppers Co.
Applied Tech. Corp. (BS&B)
Mitsubishi Chemical Industries
Dravo Corp.
Liquid Oxidation
Stretford
Takahax A,B
Takahax C,D
Elemental Sulfur
Elemental Sulfur
Ammonium Sulfate
Wilputte Corp.
Chemico and/or Nippon Steel
-------
Table 3. GLAUS PLANT TAIL GAS TREATMENT TECHNOLOGIES
SYSTEM CHEMISTRY
VENDOR
SCOT
S and S02 hydrogenation,
amine absorption to con-
centrate H2S, feed to
Glaus inlet
Shell
IFF - 1 Catalytic conversion of
H2S and SO2 to elemental
sulfur
Institute Francis Petrol
BEAVON
S, SO2 hydrogenation and
COS, CS2 hydrolysis to H2S,
Stretford sulfur recovery
R. M. Parsons
WELLMAN-LORD Sulfite/bisulfite absorption
and concentration of S02
Davy Powergas
10
-------
STEAM JET EJECTORS
CLEAN COG
30grH2S/100SCF
NO. 1
150 psig stm
317.000 Ib/DAY
NO. 2
150 psig stm
143,300 Ib./DAY
I
>
^
V
y.
r (-
r\ C°N
f
4
VAPOR
c
INTER ^
CONDENSER
ACID GAS TO
TREATMENT
55-70%H2S
5-15% HCN
15-18% CO?
5% H2 ETC.
AFTER
CONDENSER
H2S ABSORBER ACTIFIER
SOUR COG
500 gc H2S/100 SCF
60 gr HCN/100SCF
(5.7 gal/ton CCAL)
(3.02 gal/ton
COAL)
17,200 gal/DAY
ABSORPTION SOLUTION
SLOWDOWN (QUANTITY
VARIABLE)
TOTAL VACUUM JET 38'200 Q^I/DAY
CONDENSER SLOWDOWN
55.400 gal/DAY (CONTAINING
HCH, H2S.CO2)
FIGURE 2. PROCESS FLOWSHEET: VACUUM CARBONATE PROCESS
BASED ON 60 MILLION SCF/DAY COG AND 93 PERCENT
SULFUR COLLECTION EFFICIENCY (REFERENCE 18)
r\y
-------
GAS
OUTLET
H GAS
E3 ACID GASES
P FOUL SOLUTION
Q ACTIFIED SOLUTION
3 CONDENSATE5
Q FLUSHING LIQUOR
ACID GASES T
HCN REMOVAL St
ABSORBER SOLUTION CIRCULATING
ACTIFIER
INTER
CONDENSER
AFTER
CONDENSER
TANK SOLUTION FLASH SOL.
FRESH COOLER HEATER
PUMPING SOLUTION
IMS MIXING VAPOR
. TANK FLUSHING CONDENS
SOLUTION To'S.UPeh
-.». .. . IU rLAon /-s : i_» a^,
g°OLER ' SOL. HEAT Copyright©
; 1 EXCHANGER
__ : ! .
CONDENSATE TANKS
IB
VAPOR KNOCK ,
OUT DRUM
Koppers Company, Inc. 1976
A
-.
_
.-_...
..,
.....
_,,.:.u-^:» -,..,
~ 1
wcm u> fuKu
l{~\ t s\ A
ID DO P
FIGURE 3.
-------
Recently, the Kbppers Co. has proposed a two stage vacuum carbonate process
(Figure 3) which is intended to produce still lower sweet gas levels of H_S.
Hoppers maintains that by a double stage H-S absorber (see Process Appendix)
bench scale levels of <_ 10 gr H2S/100 dscf have been achieved . Kbppers is
offering this technology to stay competitive in the COG desulfurization field.
The single stage vacuum carbonate process is capable of reducing COG H-S
levels to 30-35 gr/100 dscf^ ' , independent of inlet concentration. Hence,
foul gas concentrations of 500 gr/100 dscf will be desulfurized by 93%. How-
ever, COG with H-S concentrations of 250 gr/100 dscf will be desulfurized only
86% to reach the 35 gr/100 scf level. On the other end of the efficiency spec-
trum, U.S. Steel's Clairton Works produces one KLS stream containing 2000-4000
gr/100 scf and it is desulfurized ^ 97% by a Vacuum Carbonate plant . Due to
these unusually high H-S inlet levels, the driving force for H-S absorption at
this Clairton Works Vacuum Carbonate plant is correspondingly high.
A number of vacuum carbonate plants, summarized in Table 4, have been
constructed over the years, and there is no doubt as to the H-S removal effi-
ciency or basic reliability of the technology. A major concern for the Vacuum
Carbonate process has been HCN - caused corrosion in Claus sulfur recovery plants
(8)
used in conjunction with the Vacuum Carbonate process . Serious corrosion and
catalyst fouling at Burns Harbor and Wierton caused major downtimes at each
facility. It is generally recognized now that HCN must be removed from the
acid gas stream leaving the still prior to admittance to a Claus plant. In
acid plants, however, conversion of H2S to sulfuric acid may not require HCN
13
-------
Table 4. VACUUM CARBONATE PLANTS
PLANT
Bethlehem Steel
. Burns Harbor
. Lackawanna
. Sparrows Point
. Bethlehem
. Johnstown
COKE OVEN GAS CAPACITY (MMSCFD)
120
50
60
X (down)
X (down)
SULFUR RECOVERY
Glaus Plant, HCN Destruct
Acid Plant
Acid Plant (down now)*
None for V.C. Plant*
None for V.C. Plant*
National Steel
. Wierton
70
Claus Plant, HCN water
wash
U.S. Steel-Clairton
. Keystone V.C.
. No. 1 V.C.
90
60
Two Claus Plants, HCN
water washing
Inland Steel
50
Claus, water wash for HCN
There are or will be Sulfiban-Claus technology at these plants.
14
-------
removal because HCN combustion in acid plant converters is higher than in
e
Glaus sulfur recovery plants. (The first stage in the Glaus plant inten-
tionally only partially combusts incoming H_S to SCL so that combustion is
really occurring in an oxygen-lean environment. Consequently, HCN will not
destruct in the Glaus burner.) Section IIB discusses HCN removal techniques
in more detail.
For the single stage Vacuum Carbonate process the best coke oven-gas
desulfurization produces a clean gas of 30 gr H-S per 100 dscf of COG pro-
duced. No organic sulfur is removed. See Section III for a discussion of
organic sulfur in COG. Double absorption Vacuum Carbonate has been shown
capable of achieving as low as 10 gr H-S/100 dscf.
2. The Sulfiban Process
This is a technology developed by Black, Sivalls & Bryson, Inc. and
Bethlehem Steel Corporation. Sulfiban is sold by Applied Technology Corporation
(ATC), a subsidiary of B.S.& B.
As with Vacuum Carbonate, Sulfiban is a liquid absorption/steam stripping-
solution regeneration process. It produces at the outlet of the still column
an acid gas rich stream containing H-S, reduced organic sulfur gases, CO2/
and HCN which must be treated in the same manner as the Vacuum Carbonate acid
gas stream. Sulfiban employs an amine absorber solution (^ 15% monoethanol
amine, "MEA") for sulfur removal. Its still column operates at atmospheric
pressure.
Due to the formation of certain salts, Sulfiban employs a reclaimer for
15
-------
distillation of a side stream of the MEA which is returned to the absorber for
(9)
salts recovery. As described by Williams and Homberg , the key variables in
the achievement of sweet gas H2S levels are still column and reboiler steam
rates, absorber solution temperature, and the liquid circulation rate. Figure 4
taken from this reference illustrates the strong dependence on liquid circulation
rate of outlet H2S concentration. It is true of all the liquid absorption pro-
cesses that the sweet gas H2S level is a variable; high lean solution acid gas
levels and low liquid circulation rates detract from the best achievable levels.
The control of these variables allows an operator to lower operating costs.
Therefore, continuous monitoring of sweet gas levels will be needed to insure
continual emissions performance.
Sulfiban plants have been constructed at three locations in Pennsylvania
(see Table 5 for details) with commitments at three other locations. Each
of these facilities was brought on stream during 1975-1976, and then brought
down because of mechanical start-up problems. These experiences are described
in an attachment (see the Process Appendix from the vendor) and are summarized
here'10'.
Shenango Inc. - Neville Island, Allegheny County
Purchase order in spring 1973 - Start-up in May 1975 with a spray tower
absorber - Conversion to a packed tower absorber by December 1975 to de-
crease outlet from 30 gr/100 scf to <10 gr/100 scf - Plant taken down in
winter 1975-6 for winterizing, installation of recycling cooling water
system aimed at increasing still column efficiency (see letter), putting
in epoxy lining in still (HCN corrosion protection), and correcting for
improper still column tray construction - Scheduled to be on-stream
in February 1977.
16
-------
OUTLET H2S LEVEL
(gr/l6bscf)\
45
35:
25
15
5
200
250
300
350
400
MEA CIRCULATION RATE
(gpm)
INLET: 325-375 gr
105
56 MM scf D
75 - 80° F
FIG. 4 SULFJBAN PERFORMANCE BSC, BETH.i
-------
Table 5. SULFIBAN PLANTS
PLANT
COKE OVEN GAS CAPACITY
SULFUR RECOVERY
Bethlehem Steel
(Bethlehem, PA.)
two, each 60 MMSCFD
3-Stage Glaus, HCN Destruct
Jones & Laughlin
Steel
(Pittsburgh Works)
90 MMSCFD
Single Contact Sulfuric Acid
Shenango Inc.
(Pittsburgh, PA.)
32 MMSCFD
3-Stage Glaus, HCN Destruct
There are presently coitmitments for Sulfiban plants at Bethlehem Steel;
Johnstown (two) and Lackawanna.
18
-------
Bethlehem Steel - Bethlehem, PA.
Started up in August 1975 and operated through March 1976 - Severe
corrosion occurred in still column below a stainless-carbon steel
weld (the top 16' uses stainless) - Column was entirely replaced with
stainless, lined and placed into service in May 1976 - Operated until
September 1976 when a mechanical problem in the reboiler (see Process
Attachment for explanation of function) occurred - System now in service.
J & L Steel - Pittsburgh Works
System began start-up in October 1975 was partially destroyed by an
explosion during welding as the still column - It ran for eight days
thereto - Relined, new packing, and scheduled for start-up in March
1977 - Acid plant has not run as of December 31, 1976.
Sulfiban absorption efficiencies have been measured at all three plants
as well as at a pilot plant run at Bethlehem's Lackawanna Works. These data
are summarized in Figure 4 and Table 6. The important distinction between
Sulf iban and other COG desulfurization systems is that the MEA solution removes
COS and CS2 as well H2S.
The Sulfiban process has been demonstrated to produce 5 gr H«S per 100 dscf
of sweet COG and 2 gr of organic sulfur (as H2S) per 100 scf of COG^35'10^.
3. Carl Still Process
This technology is based upon commercial absorption of H-S from coke-oven
gas, steam stripping of the acid gas, and sulfur recovery by Claus, sulfuric
acid or Stretford processes. It was developed by Firma Carl Still of
Recklinghausen, West Germany and is marketed in the U.S. by the Dravo
Corporation. Two Still/bravo systems are or have been built in the U.S.
19
-------
Table 6. SULFIBAN PERFORMANCE DATA
H2S* ORGANIC SULFUR*
PLANT INLET OUTLET INLET OUTLET
Bethlehem 325-375 5 (high MEA) 8 1.2
(Bethlehem, PA.) 40 (low MEA)
Jones & Laughlin 369-416 0.3 - 3 N.D. N.D.
(Pittsburgh, PA.)
Shenango 536-606 8-13 N.D. <100ppm
(Pittsburgh, PA.)
*Units: gr H2S/100 dscf
20
-------
Table 7. STILL/DRAVO PLANTS IN THE UNITED STATES
PLANT COG CAPACITY SULFUR RECOVERY
Armco Steel 60 MMSCFD Single Contact Sulfuric Acid
(Middletown, Ohio)
Wheeling-Pittsburgh 90 MMSCFD . Single Contact Sulfuric Acid
(Follansbee, West Virginia)
(1977 Start-up)
Armco Steel
(Ashland, Kentucky) Proposed to EPA
21
-------
(see Table 7) and apparently one other is being proposed.
The process flow diagram is shown in Figure 5. Since the process is
clearly described in the Still brochure (see Process Attachment) only these
additional comments are provided here. NH~ is the absorbant:
(1) The process is selective for H-S; it does not remove organic sulfur.
(2) The principal attraction of this process is that it simultaneous-
ly treats H2S in COG and NH3 waste waters. No additional reagents
are needed in this process, for example. Hence new plants particu-
larly may be designed for the Still process.
(3) The basic Still process has many variants. The most important ones
are that: (a) The Stretford process can be used to recover sulfur as
can a Glaus plant. An acid plant is a third option, (b) the process
allows for recovery of anhydrous ammonia. While the latter is not
of direct interest to the final COG H2S level/ it is useful for the
EPA engineer to know that this material is a useful by-product.
(4) Since the Still process description mentions the "USS Phosam" pro-
cess (ammonium phosphate scrubbing of COG with steam stripping to
recover NH_ and to concentrate NH3 liquor for H2S stripping) for
recovering anhydrous NH~, a brief description of this process is
provided in the Process Appendix. The Armco-Middletown plant uses
Phosam.
Dravo quotes a guaranteed level of 25 gr H2S/100 scf of COG ,
although Firma Carl Still asserts 10 gr/100 scf is possible ^ '.
22
-------
N)
COG.
CS2
( COS
y IMH3 i
( NapthJ
' HCN
allNH3 |
10-50gr/100scf H2S I
organic sulfur '
H2S
Abs.
H2S, HCN, acjd gases
CO2
Sweet COG
% HCIM
NH3
Abs.
N
f
Deacifier
^
NH3
CO2
NH3
Still
,H
flush,
liquqt
Steam'
liquor of
J
NH3
H2S
CO2 !
FIGURE 5. STILL-DRAVO PROCESS
-------
The distinction is one of economics not technology . The Armco and
Wheeling-Pittsburgh plants are designed to meet an overall 50 gr H-S/lOO scf
standard.
(4) Diamox Process
Diamox is a Japanese developed process (Mitsubishi Chemical Industires,
Ltd.). It too is based upon NH3 absorption of H2S and liquid regeneration by
steam stripping in a still column.
The MCI description in the Process Appendix contains the basic process
(12)
flowsheet as well as a list of facilities at which the process is operating
MCI published data show H-S levels in the sweetened COG of less than
10 gr/100 scf^ . MCI quotes a 97% removal efficiency for a 277 gr/100 scf
coke-oven gas . Since, as noted above, absorber efficiency is dependent
upon inlet (H2S) and various process variables, and since we do not have
the specifics for the MCI tests (yet) we can only state that Diamox is capable
of lOgr H2S/100 scf.
Diamox will require Claus or acid plant recovery and the comments with
respect to these technologies for Sulfiban, Still and Vacuum Carbonate are
relevant for Diamox.
B. SULFUR RECOVERS TECHNOLOGIES
1. Claus Process for Sulfur Recovery for COG Desulfurization
Claus plants operate on the H2S rich acid gas produced by the Vacuum
Carbonate, Still, Sulfiban, or Diamox absorbers to recovery molten elemental
24
-------
sulfur. The process is well known and is described in many articles
A recent EPA publication in support of the proposed refinery Glaus NSPS
is referenced for a description of Glaus technology . The principle
difference between Glaus plant operation on refinery or natural gas H-S
streams and coke-oven gas acid gas is the presence in the latter of HCN.
As stated previously, HCN has created severe Glaus plant corrosion pro-
blems at Wierton and Burns Harbor. At present, for instance, the Wierton
Vacuum Carbonate plant is down due to this effect. Therefore, as previously
noted, prior to the acid gas entering the Glaus plant.
Two techniques for HCN removal are in use in the U.S. Cold water
washing has been employed by Koppers in the 1940's and is utilized by
(14)
U.S. Steel today at the Clairton Works . Wierton is to install water
washing . HCN is removed from the gas stream in a tower which takes
advantage of the different aqueous solubilities of H2S and HCN. HCN is
then stripped from water solution, along with a small amount of H2S
carry-over. The stripped gases may then be incinerated or recombined with
the main clean COG stream. If this latter step is proposed, the sweet
underfire COG may have a bit more H2S than that sampled at the absorber
outlet, due to this blending.
A recent development of Bethlehem Steel is the catalytic "HCN destruct
reactor." Williams and Homberg and Singleton and Homberg^ ' describe this
well in the attached articles and further description is not needed here.
25
-------
Successful operation has been obtained at the pilot unit at Lackawanna and
the full scale units of Shenango and Bethlehem. Scheduled annual downtimes
of this catalyst unit are said by the vendor to be in the range 10 days - 2
weeks, which is a serious matter. Parallel cyanide destruct units can avoid
this long a downtime by providing an alterate path for the acid gases during
catalyst replacement and maintenance periods.
The Dravo Corp. proposes to deal with the Claus-HCN problem by cata-
(18)
lytically decomposing the gas in the Glaus furnace. According to Hallv ,
Dravo's proposal is to completely convert HCN to N2 + H2 + CO. This broad
concept has not yet successfully been implemented at Vacuum Carbonate - Glaus
installations.
Claus plant sulfur yields determine the sulfur content of the Glaus
gas emission. Typically, two thirds of sulfur entering the Claus plant
burner is recovered as elemental sulfur in the burner stage, with increas-
ingly lesser yields obtained in the subsequent catalytic recovery reactors.
As shown in Williams and Homberg, even with four catalytic stages a portion
of the incoming sulfur is emitted to the atmosphere as a tail gas. However,
practical Claus plant yields are closer to 96%^" This was the stated design
target of Shenango and Bethlehem for average long term operation. Yields
as high as 99 % are quoted for very carefully controlled Claus plant design
and operation, but such has yet to be achieved in stable operation and
not for COGderived H_S feeds. Hence, for practical purposes, 4% of input
total reduced sulfur (TRS) should be taken to be the lowest non-treated tail
gas performance attainable for standard Claus plants. That is, at most 96%
of input sulfur should be considered recoverable as elemental sulfur.
26
-------
In terms of emissions, this is equivalent to 0.083 Ib. SO2 from the
Claus plant incinerator* per pound of sulfur recovered. In terms of con-
centration, S02 is found in the range 5,000-20,000 ppm. The exact concen-
tration is a function of the concentration of H_S in the Claus plant feed
and amount of fuel used by the incinerator. For a plant recovering 95% of
input sulfur before the incinerator , for example, a requirement for 500 ppm
SO- in the Claus tail gas** requires an overall yield of 95% +-.Q QQQ (5%) = 99.7%,
Claus tail gas emissions can be treated to reduce substantially the
amount of SCL that is emitted to the atmosphere. Description of these
tail gas treatment processes and their performance levels are contained in
a standard support document recently published by EPA's QAQPS in support of
the Claus sulfur recovery plant proposed standard (F. R. October 4, 1976) for
petroleum refinery applications. The reader is referred to this SSEIS for
an exposition of details of system chemistry and application.
Briefly, three of these processes are described in the following:
(a) SCOT (Shell Claus Off-Gas Treatment)
The Claus tail gas composition is SO-, H-S, and some S vapor. SCOT
first hydrogenates this stream to H_S with H- from sweetened COG. H2S is then
concentrated in an amine absorber/stripper system. (See the Process Appendix
for more details). The concentrated H2S is then fed to the main COG
absorber or to the Claus plant inlet. SCOT went on-stream at U.S. Steel,
At close to optimal yield the tail gas to the incinerator consists of a 2:1
ratio of H2S:SO2- A small amount of S vapor also is contained in the incin-
erator feed. After combustion of course, SO2 is the dominant species.
** (21)
The California SIP requires this at present
27
-------
Clairton in 1975. Figure 6 shows the Clairton flowsheet relevant to SCOT;
note that Glaus yields increase from 95% to 99.9%. EPA test data for SOOT
(see SSEIS) confirm this capability.
(b) Beavon Process
The Beavon process also converts Glaus tail gas to H2S. SO- is hydro-
genated, as per SCOT. COS and CS- are catalytically hydrolyzed (CS- + 2H_0 -»
2H2S + C02; COS + H20 -» H20 + C02). H2S is then recovered in a Stretford
plant. Beavon is a development of the Ralph M. Parsons Co., Los Angeles,
California. Performance levels are discussed in the Claus plant SSEIS.
(c) Institute Francis Petrol (IFP)
Beavon and SCOT are commercially available reduction processes. Other
tail gas treatment (TGT) processes, by IFP, are available. TGT-1500 produces
a 1500 ppm total sulfur gas stream and is commercially available. TGT-150
(150 ppm) has yet to be comnercially proven. The reader is referred to the
SSEIS - - and to the Process Appendix to this document for process descriptions.
Note that reduction to 1500 ppm will provide an overall 99.2% recovery in the
example cited above.
(d) Wellman-Lord
This is an oxidation process, well known in the FGD field. Wellman-Lord
produces SO2 as an of f-gas by means of a sulf ite/bisulfite absorber-stripper
system. The output from the W-L is an SO2 stream which can be either combined
28
-------
to
HCN
WASHER
OVENS
COG
PHOSAM AND
CRYOGENIC
PLANT
KEYSTONE
VACUUM
CARBONATE PANT
H2S
KEYSTONE
CLAUS PLANT
CLEAN
COG FOR
UNDERFIRING
COG
NO. 1 VACUUM
CARBONATE
PLANT
H2S. HCN
NO. 1 GLAUS
PLANT
H2S
S02.H2S
SCOT PLANT
NO. 1
HCN
WASHER
TO
ATMOSPHERE
SO2
INCINERATOR
FIGURE 6. U.S. STEEL, CLAIRTON WORKS FLOW SHEET
-------
with the sweetened COG* or recovered as H2S04 in an acid plant.
Since an acid plant probably would not exist at a coke plant using Glaus
recovery, recycling of the S02 to the Glaus burner is another option. In the
example of the footnote on this page, 2690 Ib S02/day would be available for
a Wellman-Lord tail gas system. The main Glaus burner combusts one third of
incoming H-S to SO- to initiate the Glaus reaction. In this example,
(26,890/3) X 2 = 17926 Ib SOVday are so produced. The Wellman-Lord unit
would supply 15% of this need.
Performance levels for Glaus tail gas treatment increase Glaus yields f
from 95%-96% to 99.5 %. This performance level is documented in the above
referenced SSEIS. The October 4, 1976 F.R. proposes this performance level
directly for new refinery-based Glaus plants. This standard for reduction-based
tail gas treatment is 0.025% (250 ppm) SO2 on a dry no-O2 basis. This level,
implies a 99.9% yield for a straight Glaus/96% efficiency plant producing
tail gas ? with a concentration of 10,000 ppm. as explained above.
In summary, for tail gas Glaus treatment,. EPA has found that technology
is available to produce tail gases of less than 0.5% of Glaus sulfur input.
For example, assume that a 50mm scfd, 410 ?^n 2 .COG is desulfurized to
' A f\f\ Ov^i. /" M M
10 gr H2S in the clean gas, producing (0 x 3°§]'Q x ) or 26,890 Ib
1Q
S/day Glaus recovery. At 95% yield, 25,545 Ib/d will be recovered, leaving
1345 Ib/day of sulfur as a tail gas. This is equivalent to (1345 x ^ x 7000) .
or 10 x 106 gr/day of H2S equivalent. Added to the sweet COG, this will in-
crease its equivalent H2S concnetra£ion from 10 to 10+ (-- x ,-^) = 30 gr equiv-
alent H2S/100 scf of COG produced.
30
-------
2. Sulfuric Acid Recovery
The major alternative for sulfur recovery is the production of sulfuric
acid. Tables 4, 5, and 7 indicate this was the process used by Bethlehem Steel,
J & L, and Armco for Vacuum Carbonate, Sulfiban and Still stripping.
The reader is referred to other descriptions of the basic contact acid
(22)
process and its capabilities. The major point is that sulfuric acid pro-
duction involves a tail gas stream. Single absorption, single contact acid
(22)
plants easily produce a 97% recovery of inlet S . Only single absorption
has been purchased for COG applications and in fact the new J & L plant is
designed for 97% efficiency. One vendor noted that double contact plants
were not being marketed to the steel industry because of competitive forces
between liquid absorption and liquid oxidation technologies.
However, the double contact, double absorption process is commercially
available for the production of sulfuric acid and offers significantly greater
yields and hence lower tail gas emissions of SO2. EPA's NSPS, 4 Ib S0-/ton
acid, is equivalent to 99.7% sulfur recovery. A process description of double
absorption is contained in the Process Appendix. Approximately 35 such plants
exist in the U.S.(23)
The key technical issue is whether this is attainable on H2S from COG.
Since no applications have occurred, the direct demonstration has yet to be
made. However, the sole difference for an inlet H-S stream as compared to S
or SO- is that water is formed in combustion of the H2S to SO2- This water
can be removed in a preliminary drying tower prior to contacting the SO-.
Alternatively, a "wetlf acid plant can be designed to accept this water as "
31
-------
is done for the production of acid from spent acid feeds which contain hydro-
carbons. The technical availability of the double absorption process for coke
(24)
plant feeds is not an issue for Monsanto, Allied Chemical and other designers
As compared to single absorption plants of the same size (the median American
coke plant will produce ^40 tpd acid) capital cost differentials are about
15%-20% or $2.5MMvs. $2.QMM).
Emission rates from double absorption sulfuric acid recovery are at least
99.7% yield or 4 Ib SCL/ton of acid produced as compared to 97% for single
absorption. Therefore, double absorption offers as effective a way of reducing
tail gas emissions as does Glaus plant tail gas treatment technology. The
choice of method can be allowed to be one of plant economics and not limited
by the inherent emissions effectiveness of Claus + TGT vs. double absorption
acid recovery. It is also quite clear that the lowest technically achievable
overall SO- emission rates will require one or the other process combination
for the liquid absorption processes.*
C. LIQUID OXIDATION PROCESSES
These processes differ from the liquid absorption processes described
above in that once H-S is absorbed, it is oxidized to sulfur or ammonium sulfate
in the liquid phase. The separate Claus or sulfuric acid steps are avoided and
therefore no tail gas problem need be faced. On the other hand, a difficult
It is noted that Firma Coal Still suggests that recovery via Stretford
is an option. Stretford produces no tail gas.
32
-------
liquid effluent problem is created in the form of thiosulfate and thiocynate
salts, which are not present in the liquid absorption processes to the same
degree.
Two processes have been investigated by DSSE, the Stretford Process by
the North West Gas Board of the U.K. and the Takahax Process of Nippon Steel
Corp. Two other processes, Fumax, a Japanese process, and Giammarco Vetrocoke,
a German process involving an arsenic solution, have not been studied due to
time pressures. Basic process descriptions of the Stretford and Takahax
processes are provided in the Process Appendix.
1. Stretford Process
The Stretford process produces elemental sulfur from HJ5 in COG. It
does not remove organic sulfur. At present, there is one application in
North America, located at the coke plant of Dominion Foundary & Steel Company
(Dofasco) of Hamilton, Ontario (42mm scfd). Dofasco is building additional
coking capacity (a new No. 6 Battery, 6m wet coal, by Didier) and has ordered
a second Stretford plant to handle the extra COG.
The Stretford process absorbs H2S in a packed tower in a solution of
sodium carbonate, sodium ammonium vanadate, and ADA (anthraquinone disulfuric
r,
acid). The process flow diagram is shown in Figure 7. In the absorber, HS
+5 -f4
is oxidized to.S and vanadium is reduced from V to V . The oxidizer
+4 +5
system allows V to be reoxidized to V by the reduction of ADA. In turn,
ADA is reoxidized by air pumped into the oxidizer tank. Elemental sulfur is
removed and the Stretford liquor is recirculated to the absorber.
33
-------
PRODUCT
GAS
FEED
GAS
H2S ABSORBER
SULPHUR FOAM TO PURIFICATION
PUMPING TANK
FIGURE 7. STRETFORD PROCESS FLOW DIAGRAM
-------
HCN in foul COG causes the formation of the thiosulfates and thiocyanates.
Therefore, removal of HCN ahead of the absorber or of its products in the absor-
\
ber or removal of its products in Stretford liquor or both is necessary. Re-
moval has been achieved ahead of the Stretford abosrber in an absorbing tower,
in which a solution of suspended sulfur and water reacts with the NH3 in COG
to form ammonium thiocyanate and thiosulfate. This is known as the polysulfide
treatment process (due to the intermediate formation of ammonium polysulfide) .
(25)
The polysulfide process can remove up to 98% of the HCN from foul COG
In turn, this liquor, containing SCN and S2O3 needs treatment before disposal.
These same compounds also build up in the Stretford liquor, if allowed, and
either spent Stretford liquor or a steady blowdown thereof require treatment
for these salts.
Peabody-Holmes of the U.K. has developed and operate a Stretford process
waste liquor process (Fixed Salts Recovery) at the Orgreave plant of British
Steel Corp (26 MMSCFD of COG) . The chemistry and other details of this process
are described in the Process Attachment. The essential step is high temperature
oxidation in a reducing atmosphere, generated by substoichiometric combustion of
COG. SCN and S203 are converted to H2S + CO2 + N2 and are recirculated to the
front end of the Stretford tower. A critical question for the Stretford process
is whether the Peabody-Holmes waste liquor process is adequately demonstrated.
26}
To this end Peabody offers the following ' . After laboratory and pilot
scale development, the Orgreave full scale unit was brought on-stream for oper-
ation and testing in August 1975. The plant has operated continuously with two
exceptions for the past six months. In August 1976 an "incident" involving a
COG explosion in a burner, not related to the basic process, brought the plant
35
-------
down. It had run for five weeks prior to then and ran from November 20
to Christmas 1976 when it was again brought down to provide a holiday
for BSC workers. It will be operational in mid-January once more. From
this, one sees that the process is considered a development by P-H and
BSC operators and that this project is proving successful from a system
chemistry perspective. In an attachment, which is_ seen by_ EPA staff to
be quite confidential, Peabody-Holmes believes they have demonstrated
that their combustion process works. In the August 5, 1976 in-house
memo, Peabody-Holmss noted that "steady-state" operation was still
needed at that time to fully prove the system's chemistry. Two re-
cent judgements by the steel industry reflect the diversity of its
views. Dofasco has ordered two Holmes' units; one for its existing
Stretford plant and one for the new Stretford plant mentioned above. The
first Fixed Salt Recovery (FSR) system is to be on-stream September 1977.
(27)
Kaiser Steel has just decided ' to purchase Takahax based in part upon
the longer operating times Takahax has experienced. Wilputte Corp., the
Peabody-Holmes licensee in the U.S. vigorously maintains the Holmes process
is commercially proven
Wbodhall-Ducklam, a British engineering firm, has also piloted a Stretford
waste liquor facility based on the same basic chemistry as the Holmes process.
Wbodhall-Ducklain has not yet built a full scale unit although one is under
(28)
construction at the Redcar plant of British Steel Corp. The reader is also
directed to the Nittetsu process (see Takahax for an additional waste liquor
treatment alternative).
The Stretford process is very efficient at removing H_S from COG. Ludberg
36
-------
quotes concentrations below 1 gr/100 dscf. Wilputte asserts H2S to 10 ppm
/25 26}
(0.6 gr/100 dscf) is well achievablev ' . Massey and Dunlop state removal
efficiencies "in excess of 99%" are possible . Organic sulfur is not re-
moved, however. Therefore, the lowest sulfur level emission rate for desul-
furization by the Stretford process is 1 gr H2S/100 dscf of COG produced plus
organic sulfur in the foul gas.
2. Takahax Process
Takahax was developed and is sold by Nippon Steel Corp. In the U.S.,
Nippon and Chemico Air Pollution Control Company (of Envirotech) have a working
relationship for the marketing of Takahax. Many details of the process chem-
istry are described in the Appendix and are not redescribed here.
The basic flow diagram for Takahax is shown in Figure 8. H^S is absorbed
from COG in a column, using either a Na or NH., based solution. HCN is not re-
moved ahead of the absorber but is allowed to build up as thiocyanate in the
absorber. A blowdown is removed for treatment in one of three ways.
The Nittetsu Chemical (NICE) process, like the Holmes process, is based on
substoichemitric combustion of S-O., and SCN to H2S, 002/ N2, and Na2C03 (in
liquid solution, returned to the absorber) H2S produced from, the NICE process
can be either recycled to the absorber or sent to a sulfuric acid plant. Al-
ternatively, complete combustion of the ammonia waste liquor ((NHJ S203 and
NH.SCN to S02 + C02, and thence SO2 to sulfuric acid) is possible. Wet
oxidation at very high pressure and temperature to ammonium sulfate, for
the NH-j stripping of H-S version of Takahax, is a third waste liquor treatment
bility(29).
Since Takahax is available for each of these waste liquor options, "Takahax"
possibility(29)
37
-------
00
SWEET COG
FOUL COG
w<
1
1
1
1
1
1
A
DC
111
03
.DC' *
0
(A
CQ
1
. . " - . \
WASTE LIQUOR
' ^
H2S RECYCLE
"WASTE AIR" STREAM
w AMMONIUM (NH4)2SO^4
* SULFATE PLANT *
TAIL GAS
t
S, SULFURIC _J
J ' ACID PLANT
ACID
^ TAIL GAS
1 1
. SUBMERGED H2? SULFURIC 1
r COMBUSTION T ACID PLANT
1
FIGURE 8. TAKAHAX FLOW SHEET
-------
is a generic term. Nippon Steel offers four process options.
"A" - NEL absorption + wet oxidation of liquor to recover H2S in COG as
ammonium sulfate.
This version produces two "waste' air" streams containing a small
amount of sulfur both of which need scrubbing before release to
the atmosphere.
"B" - NKL absorption + combustion of waste liquor, producing elemental
sulfur and an SO2 stream from the waste liquor combustion process.
This SO2 requires recovery as H2S04 with attendant considerations
mentioned in Section IIC, above. Again, a waste air stream from
the absorber section carries some sulfur to the atmosphere.
"C","D" - Na2CO2 absorption + NICE waste liquor treatment. ELS is recovered
as elemental sulfur. The NICE process is used on ELS off-gas which
Nippon Steel proposes as feed, along with molten sulfur, to an acid
plant. ELS could also be recycled to the absorber with sulfur the
process output as well. This latter version is termed the "D" type.
Takahax is a conmerically demonstrated process. Nine Japanese facilities
exist as shown in Table 8. Kaiser Steel has decided to construct Takahax-A to
comply with an EPA-Kaiser consent decree and has so notified EPA Region IX.
Relative costs are shown in Table 9.
Table 10 provides Takahax performance data for four Nippon Steel
plants. Note that the Nagoya and Yawata plants achieved 4 gr ELS/100 scf
whereas three other facilities were stated ' to have been tested at
39
-------
Table 8. COKE-OVEN GAS DESULFORIZATION UNITS IN JAPAN (JANUARY 1977 NIPPON STEEL)
NAME
PLACE
GAS VOLUME DESULFUKEZATION
SULFUR TREATMENT
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Steel's
Affiliate
NKK
NKK
Kawasaki
Kawasaki
Kawasaki
Sumitomo
Sumitomo
Sumitomo
Mitsubishi Che.
Mitsubishi Che.
Mitsubishi Che.
Mitsubishi Min.
Amagasaki
Amagasaki
Amagasaki
Tokyo Gas
Hirohata*
Muroran*
Oita*
Nagoya
Nagoya
Tobata
Kimitsu*
Fukuyama*
Ogishima*
Chiba*
Mizushima
Mizushima*
Wakayama
Wakayama
Kashima
Sakaide*
Sakaide
Kurosaki
Hibikinada
Ohatna
Ogimachi
Kakogawa
Tsurumi*
99 MMSCFD
130
206
162
31
116*
227
116
90
80
179
90
70
107
72
215
188
85
107
53
47
132
45
NH-. - Takahax
NH^ - Takahax
NH^ - Takahax
Na - Takahax
Na - Takahax
Na - Takahax
NH-, - Fumax .
3
NH3 - Takahax
NH3 - Takahax
NH3 - Fumax
Diamox
Diamox
Na - Takahax
NH3 - Fumax
NH3 - Fumax
Diamox
Diamox
Diamox
Bischoff, V.C.
Rodax, S tret ford
Rodax, Stretford
Rodax, Stretford
NH3 - Takahax
Hirohax
Hirohax
Hirohax
Combustion, Sulfuric Acid
Combustion, Sulfuric Acid
Combustion, Sulfuric Acid
Combustion, Sulfuric Acid
Combustion, Gypsum
Hirohax
Combustion, Sulfuric Acid
Glaus, IFF
Glaus, IFP
Combustion, Sulfuric Acid
Combustion, Gypsum
Glaus
Glaus, Activated Sludge
Glaus, Activated Sludge
Combustion, Sulfuric Acid
Activated Sludge
Combustion, Sulfuric Acid
Hirohax
Indicates facilities constructed in 1975 and 1976.
-------
Table 9. RELATIVE TAKAHAX COSTS
TYPE RELATIVE COST
A (Hirohax) 100% - $12 million (1975)
B 110%
C 120%
D 130% (producing H-SOJ
D 90% (recycling H2S to absorber)
Notes
1. Above data based on the Hirohata plant: 99M4SCFD
200 gr H2S/100 scf inlet
10 gr H^S/lOO scf outlet
2. Includes cost of duel absorption sulfuric acid plant
41
-------
Table 10. TAKAHAX PERFORMANCE DATA
PLANT
Hirohata
Oita
MDroram
Nagoya
Yawata
H0S^a' INLET
200
260
240
300
340
H0S(a) OOUTLET(b)
I10
<10
I10
1 4
1 4
START-UP
DATE
4/75
6/76
3/76
10/73
3/74
DAYS SINCE
START-UP
630
204
306
1126
1036
DAYS IN
OPERATION
630
204
306
1111
1036
(a) By J.I.S. methodology, gr H?S/100 scf.
(b) No organic sulfide removal in the Takahax process.
42
-------
<_ 10 gr/100 scf. Nippon Steel states this to be a reflection of local pre-
fecture regulations and not system capability.
The NH.,-Takahax process also produces a waste air stream which contains
small amounts of H2S gas. The Takahax liquor is pumped to a liquid/gas
separation bubble tower into which air is pumped. Waste liquor and regen-
erated solution (reoxidized) are separated and a waste air stream containing
H-S gas is created. This stream is scrubbed in a counter-current packed
tower to remove its NH_ content with "mother liquor" (3-4% H2S04). H_S in
the waste stream is of course not absorbed by this acid liquor and is
emitted to the atmosphere. Exactly the same process chemistry gives rise
to the Hirohax waste stream.
Pertinent data for these streams are shown in Table 11. Note that the
quantity of H2S to the atmosphere is very small, < 0.01 gr/100 scf of COG.
As H2S is noticeable even at small concentrations in the ambient air, short
term diffusion model calculations were run to calculate these streams' impact.
For the 99 MMSCFD sized plant from which Table 11 derives, worst case 1-hr
concentrations were ^ 0.0015 ppm or about 40 times below the (H2S) human order
threshold.
43
-------
Table 11. TAKAHAX WASTE AIR STREAM FLOWS
ft3 H-S Equivalent gr H-S
Outlet (H2S) Volums Per Day per 100 dscf of OOG
Takahax Regenerator 2.3 ppn 2.7 - 3.2 6.2 - 7.4 .0042 - .0050
MMSCFD
Hirohax Waste Liquor 2.0 ppm .'.2.0 - 2.9 4.0 - 5.8 .0027 - .0039
Process
Stack Parameters
Tenperature - 120°F
Height - 130 ft
Diameter - 14 in
Velocity - <30 fps
44
-------
Section III. ORGANIC SULFUR IN COKE-OVEN GAS
Only one of the six desulfurization technologies removes COS and CS2 from
COG. Yet both, coirpounds produce SO2 upon combustion by:
CS2 + 3 02 + 2S02 + CO2
and it is the emission of S02 that the technologies are designed to abate.
Organic sulfur has not usually been analyzed with care by coke plant
operaters since H2S predominates the total sulfur level of foul COG. However,
after H2S stripping, uncontrolled organic sulfur may amount to half or more
of the total sulfur load in the sweetened COG. For example, typical foul
gas may contain 450 gr/100 scf H2S and 15 gr/100 scf CS2 + COS. If H2S is re-
duced to 10 gr/100 scf then of the 25 gr/100 dscf of total reduced sulfur in the
sweet gas, 60% is COS or CS2. If the stripped acid gas is processed in a 97%
acid plant, an ELS equivalent emission of 12 gr/100 scf will occur. In this
case, organic sulfur (O.S.) will still account for 40% of SO- emissions.
^
The best estimates we have for foul gas organic sulfur levels are in the
range 5-25gr H2S equivalent/lOOscf . (By ELS equivalent or "ELS", it is meant
that COS or CS2 volume concentration in ppm is converted to the gravimetric
conentration, gr "ELSVlOO scf, by multiplying by .063). No good empirical
or theoretical relationships exist between H2S and COS or CS2 levels nor
does there appear to be any relationship between sulfur (%) in coal and COS/CS2
levels in COG. This seems to be the case, despite the ELS/%S relationship
quoted in Section I. One obvious variable affecting (COS) formation is 0~
in COG induced by stage charging. This is an operating variable not subject
to predictions based on coal analysis.
45
-------
Empirical data for organic sulfur levels in actual. COG streams exist to
some extent. The following data were obtained from various sources.
Table 12. ORGANIC SULF3DE CCNCENTRATICNS IN COG
PLANT
FOUL COG ORGANIC SULFUR CONCENTRATION
1. Jones & Laughlin
-Pittsburgh Works
2. Shenango Inc.
3. Kaiser Steel
4. Bethlehem Steel
(Bethlehem, PA.)
5. Republic Steel
a. 24 gr "H2S"/100 scf, total of CS2 + COs
(Sulfiban design basis)
6. Crucible Steel
7. Bethlehem Steel
a. Sparrows Point
b. Lackawanna
c. Rosedale
d. Franklin
b. 5 gr "H-pS'VlOO scf - COS
7 gr "HpVlOO scf - CS2 See Table 6
25 gr/100 scf (Sulfiban design basis)
19 gr/100 scf (recent gas analysis used
for design)
3-8 gr/100 scf total O.S. (Williams and
Homberg)
Have stated 10-20 gr/100 scf for design
of facilities at Warren, Cleveland, and
Youngstown. This is a RSC generic
estimate for total O.S.
12 gr/100 dscf for COS + CS0
14
14
10
10
46
-------
From this list 5-25 gr/100 scf seems to be a reasonable assumption for foul
gas O.S. However, as coal sulfur levels and coking practices vary from place
to place and with time at any one location direct measurement is encouraged.
Attached is an article by chemists of Jones & Laughlin Steel Corp. who
were faced with a foul COG sampling obligation in 1973 . J & L had
designed a Sulfiban plant based upon a 400 gr H2S/100 dscf presumption.
Because their coal analyses indicated probable higher foul gas H2S sulfur
levels, the Allegheny County Health Department required J & L to develop
and operate a TRS* sampling system. Data developed therefrom (see Appendix
for an example) can be obtained via Section 114 from J & L Steel Corp.
* ,
Total reduced sulfur,
47
-------
Section IV. PERFORMANCE LEVELS FOR COG DESULFURIZATION TECHNOLOGY
In this section, are comparisons of the demonstrated technologies
described in Section II in terras of their ultimate overall so2 emission rates.
Tables 13-15 indicate the H-S equivalent levels attained by liquid
absorption processes and their necessary sulfur recovery adjuncts. Clearly,
the highest desulfurization currently achievable is by the high effeciency
Sulfiban process operated in conjunction with a high yield Claus or sulfuric
acid plant. Also note:
1. Table 15 indicates the best demonstrated performance levels for
other combinations of liquid absorption processes as well. This
table was calculated using a total organic sulfur concentration of
15 gr HLS equivalent per 100 dscf.
2. Sweet coke-oven gas H2S levels of 10 gr/100 dscf of COG are
obtainable by Diamox, Still, Sulfiban, and perhaps by double
stage vacuum carbonate. This only refers to the sweet gas, per se.
3. Only Sulfiban absorbs COS and CS2. From Section III, it is clear
that COS and CS2 can each exist in foul COG at levels from 10-25 gr
H2S equivalent/100 dscf. Sulfiban removes these down to a residual
2 gr "H2S"/100 dscf level.
4. Standard Claus and single absorption sulfuric acid technology pro-
duces a 96%-97% yield.
5. Claus + tail gas treatment or double absorption sulfuric acid
produces a tail gas containing no more than 0.5% of the recovery
plant inlet sulfur level. (See Figure 9 for system options).
48
-------
6. Overall system emissions can be reduced to [ (Foul COG Inlet
"H2S" -5)(0.005 + 5] gr "H2S"/100 dscf of COG produced. At
500 gr "H2S"/100 dscf, this works out to an overall 9 gr "H2S"/
100 dscf of COG.' This is the overall lowest level demonstrated
for liquid absorption.
For the liquid oxidation process, Table 16 summarizes the analogous
performance levels.
The impact of organic sulfur on the overall performance liquid oxidation
process is dramatic. Basically, these technologies reduce SO~ emissions to
twice the foul gas organic sulfur gravimetric concentrations, when operated
at their maximum H2S absorption efficiencies. This is so, because virtually
all H2S is removable.
The overall capabilities of the coke oven gas desulfurization technol-
ogies discussed in this document are ranked in Table 17. Two rankings are
provided, one for H2S and one for total reduced sulfur. This table provides
a direct comparison of all the technology combinations studied in this
document's preparation. Since their purpose is SO2 emissions prevention,
the right hand column of this table, in the units Ib SO2/10 ft GOG, is
most useful. In this table, the phrase "high S recovery" refers to 99.5%
sulfur recovery as shown in Table 14.
49
-------
GAS
COKE
OVENS
FOUL;
GAS
ABSORBER
t/JjZ!
i.rf 1 uj !
o s:
H-
.' :' . i
CLAUS
PROCESS
V
ACID GAS
H2S,HCN
C02
SULFUR RECOVERY
TOATMOSPHERt
SINGLE
ABSORPTION
DOUBLE;
ABSORPTION
I
SULFURIC ACID RECOVERY
FIGURE 9. LIQUID ABSORPTION OPTIONS !
-------
Table 13. SUMMARY OF LIQUID ABSORPTION TECHNOLOGY PERFORMANCE LEVELS
Best Attainable Outlet Sulfur Concentrations
PROCESS
Sulfiban
Vacuum Carbonate
Diamox
Carl Still
H2S
5 gr/100 scf
30**
10
10
ORGANIC SULFUR
TOTAL REDUCED SULFUR
(as H?S)
foul gas level*
(PGL)
FGL
30 + PGL
10 + FDL
10 + FGL
Foul gas levels (see Section III) are usually in the 5-25 gr/100 dscf range.
**
Bench scale demonstration of 10 gr/100 scf has been made with a two stage
vacuum carbonate process.
51
-------
Table 14. SUMMARY OF SULFUR RECOVERS TECHNOLOGY PERFORMANCE LEVELS
PROCESS
SULFUR YIELD (% OF INLET S RECOVERED) TAIL GAS EMISSION RATE
Claus Process
.Three Stage
Catalytic Re-
covery
.With Tail Gas
Treatment
Sulfuric Acid
Plant
.Single Contact
.Double Absorption-
Double Contact
96%
99.5%
97%
99.7%
0.082 Ib S02 per Ib sulfur
recovered.
287 gr H2S equiv. per
Ib S recovered.
0.010 Ib S02/lb S 35 gr
H_S equivalent/lb S
0.063 Ib S02/lb S 215 gr
ELS equiv./Ib S
0.0063 Ib S0-/lb S
22 gr H2S equiv./Ib S
52
-------
Table 15. Best Obtainable Overall Desulfurization Performance
by Liquid Absorption Technology
Sulfur Recovery Process
Three Stage
Removal Process Claus
Sulfiban
HeS / TRS**
300* 17* 19*
500* 25 27
Single Double
Contact Claus & Contact S tret-
Add TGT Acid Ford
H2S/TRS H2S/TRS H2S/TRS H2S/TRS
7 9 14 16 68 X
8 10 20 22 79 X
Vacuum Carbonate***
Diamox
Carl Still
300* 41 56
500* 49 64
300* 22 37
500* 30 45
300* 22 37
500* 30 45
31 46 38 53 31 46 X
32 47 44 59 31 46 X
12 27 19 34 11 26 X
13 28 25 40 12 27 X
12 27 19 34 11 26 10 25
13 28 25 40 12 27 10 25
*Total foul gas reduced sulfur concentrations, gr "^S" per 100 dscf of COG
produced.
**
Total reduced sulfur.
***
Single stage.
53
-------
Table 16. BEST OBTAINABLE OVERALL DESULFURIZATION BY LIQUID OXIDATION
CLEAN COG CONCENTRATION OVERALL
PROCESS H2S TRS* H2S TRS**
Stretford 1 16 1 16
Takahax - A 4 19 4 19
Takahax - B 4 19 6** 21
Takahax - C 4 19 6** 21
Takahax - D 4 19 6** 21
Assumes 15 gr H-S equivalent/100 scf for organic sulfur in foul coke-oven gas.
**
Assumes need for acid plant at 99.5% sulfur yield.
54
-------
Table 17. Technology for Desulfurization of Coke-Oven Gas
H2S Ranking^)
TRS RANKING(a)
Technology
Stretford
Takahax - A
Tahahax - B,C,D
Sulfiban -
high S recovery
Still- Stretford
Still - high S
recovery
Diamox - high
S recovery
Sulfiban
Diamox
Still
Vacuum
Carbonate -
single stage
Clean
COG
1
4
4'
5
10
10
10
5
10
10
30
Tail Gas
0
0
2
2
0
2
2
15
15
15
14
Total H2S
Equivalent
1
4
6
7
10
12
12
20
25
25
44
Technology
Sulfi ban-high
sulfur recovery
Sulfiban
Stretford
Takahax A <
Takahax B,C,D
Still -
Stretford
Diamox -
high S recovery
Still - high
sulfur recovery
Di amox
Still
Vacuum
Carbonate -
single stage
Clean
COG
7
7
6-26
9-29
9-29
15-35
15-35
15-35
15-35
15-35
35-45
Tail
Gas
2 v
15
0
0
2
0
2
2
15
15
14
Total H2S
Equivalent
9
22
6-26
9-29
11-31
15-35
17-37
17-37
30-50
30-50
49-63
S02(b)
Emis-
sion
Rate
24
59
16- 70
24- 77
29- 83
41- 95
46-100
46-100
81-135
81-135
135-170
(a) All units are gr H2S or equivalent
(b) Unit is Ib S02 emitted per 106 ft3
per 100 dscf of coke oven gas produced.
cf coke-oven gas produced.
-------
COG DESULFURIZA1TON PERFORMANCE OCCLUSIONS
1. The overall lowest achievable emission rate is achievable with the
Sulf iban liquid absorption process operated at nigh MEA circulation
and high MEA regeneration rates, followed by either double absorption
sulfuric acid recovery or tail gas treated Glaus plant sulfur re-
covery. This level is 9 gr H2S per 100 dscf of coke-oven gas pro-
duced with all tail gas included as equivalent H-S. This level
corresponds to an SO- emission rate of 25 Ib S02 per 10 ft of
coke-oven gas produced.
2. The Stretford-Holmes or the Takahax processes may achieve this
same or even lower rate of emission. Their lowest equivalent
SO2 emission rates will vary from 16-83 Ib SO2 per 10 ft
COG, depending upon foul gas organic sulfur levels. Therefore,
these two technologies may, at certain plants, be equivalent to
the lowest achievable emission rate and may even surpass the rate
stated in (1) above. Foul gas organic sulfur levels must be less
than 8 gr H?S equivalent per 100 dscf for this to be possible.
3. When the foul gas organic sulfur concentration is between 9-13gr
H2S equiv. per lOOdscf, the two liquid oxidation processes should
be characterized as the second lowest emissive technologies. This
emission rate will be in the range 24-59 Ib SO2 per 10 ft of COG.
4. Above 13gr "H2S"/100dscf of organic sulfur, the second lowest
achievable SO2 rate is provided by the Sulfiban process operated
at high MEA circulation and regeneration rates, with sulfur recovery
56
-------
by a conventional Glaus plant or single contact sulfuric acid plant.
This rate is an overall 22 gr "H2S"/100 dscf or 59 Ib SO2 per 106 ft3
COG produced.
5. Both the Firma Goal Still and Diamox processes can achieve 59 Ib S02
per 10 ft of COG provided organic sulfur levels are below 12 gr
"HgS"/100 dscf. Both processes require high efficiency sulfur
recovery systems (Glaus + TGT, sulfuric acid).
6. Diamox, Still, and Vacuum Carbonate operated with conventional sulfur
recovery will not be able to achieve 59 Ib S02/10 ft GOG because of
the combined impact of organic sulfur and the high tail gas emission
rate.
7. Organic sulfide levels for the specific COG under consideration
should be known.
8. The recommended levels for various EPA regulatory policies are:
(a) Lowsst Achievable Emission Pate: 10 gr/100 dscf of COG
produced of total sulfur compounds, expressed as H-S,
including all tail gas sulfur emitted from sulfur re-
covery equipment.
(b) Best Available Control Technology (considering cost) for
Prevention of Significant Deterioration use: 35 gr/100 dscf
of COG produced of total sulfur compounds, expressed as H2S,
including all tail gas sulfur emitted from sulfur recovery
equipment.
(c) Reasonably Available Control Technology: 50 gr/100 dscf
of COG produced of total sulfur compounds, expressed as
57
-------
H2S, including all tail gas sulfur emitted from sulfur re-
covery equipment.
Section V. MEASUREMENT FOR COMPLIANCE
Since reductions in SO- emissions from COG combustion require removal
of H-S (and possibly organic sulfides) before combustion and since COG
combustion takes place at dozens of separate points/ the compliance measure-
ment is for reduced sulfur.
As described in Section IV, it may be possible to achieve compliance
with the LAER by means of either liquid absorption or oxidation equipment.
Therefore it may be necessary to measure for:
(1) Cswg, concentration of sulfur compounds in the sweetened coke-oven
gas.
(2) Ctg, concentration of sulfur compounds in Glaus, Takahax, or sulfuric
acid tail gas streams.
(3) Vswg, volume flow rate of sweet COG.
(4) Vtg, volume flow rate of tail gas.
(5) Vfg, volume flow rate of foul COG.
The lowest achievable emission rate standard can be written as:
m *
Cswg Vswg + 6.63 x 10 Ctg Vtg <_ 10 gr "H2S" (1)
Vfg 100 dscf of COG
In this relationship the appropriate units are:
[Cswg] = gr "H2S"/100 dscf
[Vswg] = 100 dscf/hr
[Ctg ] = ppmv
[Vtg ] = dscf/hr (of tail gas)
[Vfg ] = 100 dscf/hr
59
-------
Figure 10 is a schematic which indicates the various sampling locations
which potentially are needed to determine compliance. The difference between
the cleaned and foul OOG volumetric flows is the amount of acid gas removed
in the absorber. Since the sum of H2S, HCN, organic sulfides, and CO- is
typically about 1% of the foul gas flow, it is fair to assume Vswg«sVfg in
computing complinace.* If Vswg is measured in lieu of Vfg this will bias
the computed result upwards by about 1%, in the converse case measurement
of Vfg will cause an underestimation of the true overall concnetration by
1% (tail gas flux/total flux 1/2%).
Measurement of clean coke-oven gas flow by built-in orifice or venturi
flow meters can and should be expected to be part of an operator's process
control equipment. Data should be reported to a 24-hour chart located in
the sulfur plant control room. Such instrumentation has a tendency to drift
from calibration so that before using such equipment in a compliance test, it
should be known to be in calibration. Even so, accuracy only to about + 5%
of the true flow should be expected such flow meters tend to change dimension,
particularly on the foul gas side, due to tarry COG constituents. Hence even
long term averaging will not necessarily insure better accuracy.
Clean coke-oven gas measurements of reduced sulfur compounds concentrations
must be sensitive to H2S, COS, and CS-. There are two ways of accomplishing
this. Direct measurement of each compound by gas chromotography (GC) separation
followed by flame photometric or thermal conductivity detection has been
Exception: The Clairton Coke Works cryogenic plant or any other synthetic
NH3 producer using H2 in COG.
60
-------
FIG 10 COMPLIANCE MONITORING LOCATIONS
CTl
COKE OVENS
V^rV^/jLvLJ V^ V 1-iL^ikJ
/
^»
.
COAL
CHEMICALS
PLANT
^
(J
,
X^
i^..
"\
^
,-''~
0
g
P
0
a
. . a
V
3
-.
-~,
if
1
J
j
3
V
-4 -
/?
/ ^
/
\x
\
r\
i^
^
*^^
^~
>-
"CLEAN
CLAUS
SULFURIC ACID
OR
r.TACT'C1 T ~mfY\D
WAolii JaiyUUK
. PLANT
\
COG
6
V:
x_
D"
u
1
/
PS
8
s
f \
%
f
I"*
S
/
\
1 .
I
\
1
i
\ '
-
r*
2)
\
.
FOUL GAS FLOW
SWEET GAS FLOW
AND
CONCENTRATION
TAII. GAS FLOW
1 AND
CONCENTRATION
-------
successfully applied. The reader is referred to Manka . The set-up
used by J & L included a permanent connection to the sweet gas line.
The alternative is to acquire periodic gas samples which would then
be returned to a laboratory for GC separation and analysis. This technique
carries the real risk that water condensation in the evacuated sample bottles
will occur, carrying H2S into solution and hence removal from the sample gas
phase. See the Process Appendix for a discussion of this problem . In
the case of remote sampling and analysis, reheating the sample bottles before
hypodermic needle extraction of the gas sample will be necessary. Reheating
may not completely solve such problems, however. H-S and HOST may react with
NH., to form soluble salts which would thereby falsely remove H-S from the gas
phase. For these reasons on-site GC analysis is preferred.
The alternative to direct sampling of TRS is to combust the sweet gas
sample to CO- and SO?* Measurement of SO2 concentrations then can be made
by EPA Methods 6 or 8. This technique assumes complete conversion of all
reduced sulfur compounds to SO- and the lack of SO- formation. EPA Methods
6 and 8 may therefore provide falsely low readings.
Wet chemistry methods exist for H-S, COS, and CS- concentrations in
coke-oven gas. These are the traditional methods employed by the industry
in routine sampling. The most popular is the Tutweiler titration method
which can be used for either H-S or TRS. While this method is useful at
foul gas concentrations, the method has been reported to be less valid
at (H2S) below 10 gr/100 dscf. Patience in carrying out the titration
to the true end point has been indicated as one cause of poor detectability.
62
-------
(32)
with the method . Another range finding device used by the gas process-
ing industry is generally referred to as the "sniffer tube." The only valid
use of these handheld devices is for quick order-of-nagnitude determinations.
Tail gas sampling is necessitated in the case of the liquid absorption
processes for both concentration and flowrate, per Bq. (1). For Glaus
sulfur recovery plants, with or without tail gas treatment, EPA has pro-
posed tfethod 15 (F.R., October 6, 1976) for (H2S), (COS), & (CS2) deter-
minations. Sulfuric acid plant compliance testing in the formal NSPS is
by Method 6. These methods are recommended for COG tail gas streams.
For Glaus plants, however, which pass the tail gas through an incinerator,
a problem particular to the COG occurs. Since it is necessary to determine
the entire second term in the numerator of Bq. (1), Vtg must be measured.
This was not necessary in the case of the refinery Glaus plant NSPS because
of the absence of parallel sulfur bearing streams. The incinerator produces
a very hot (^1500° F) stream which makes flow rate determinations, particularly
continuous determinations, difficult. One company's solution is to sample
for Vtg Ctg at the inlet to the incinerator with a venturi flow meter. A dis-
advantage to be noted, however, is that the inlet to conventional Glaus plants
contain sulfur vapors in small amounts, which may condense in the Ctg or Cfg
sampling lines.
Averaging time for compliance testing is another monitoring issue. The
refinery sulfur recovery NSPS proposal requires a four hour testing period.
Each hour, four grab samples are to be acquired from a side stream which runs
to an on-site chromatograph. One test comprises 16 grab samples. The arith-
metic average of three tests is to be used to determine compliance. For
63
-------
Table 18. COKE-OVEN GAS SAMPLING FIELD SHEET
Plant
Battery No. (s)
Date
"4
Level = Cswg Vswg + 6.63 x 10 Ctg Vtg
Vswg
HOUR
1
Clock Time
2
Clock Time
3
Clock Time
4
Clock Time
Average
SAMPLE NO.
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Cswg
Vswg
Ctg
Vtg
LEVEL*
Result
64
-------
coke-oven gas flows are variable from minute to minute due to the inherent
batch operation of coke plants. Since it is necessary to determine both
Cswg Vswg and Ctg Vtg to determine compliance, i± is_ imperative that Cswg,
Vswg, Ctg, and Vtg be determined simultaneously.
The suggested monitoring scheme is outlined in Table 18. Compliance
is determined by comparing the appropriate standard to the arithmetic average
of the 16 numbers in the right hand column.
Section VI. ENERGY AND COST IMPACT OF COKE-OVEN GAS DESULFURIZATION
A. Cost Impact
A general discussion of economics of coke-oven gas desulfurization
is very difficult to develop because of the large variability of specific
plant factors, the highly competitive and rapidly developing state of
technology, and the variability and uncertainty in by-product prices.
Each U.S. steel company which has installed or is committed to install
desulfurization technology (U.S. Steel, Bethlehem, Armco, J & L, Shenango,
Kaiser, Republic, Inland, Youngstown Sheet and Tube, and others) has of
course performed site specific cost studies. These studies are limited
in their generality* and are not available to EPA. Published cost studies
are limited either by scope or vintage.
*
In one case a vendor was asked to provide in a bid 12 separate paramatiza-
tions of the costs of the same basic process at a given site,-for a given
COG.
65
-------
Recent cost studies considered in this guidance are:
STUDY SCOPE
Massey & Dunlap Hypothetical parametric case study
. ,q_4 of Stretford and V.C. and Sulfiban
-sprang iy with conventional Glaus recovery.
Massey & Dunlap Hypothetical parametric case study
. 1Q__(34) of Stretford plus waste liquor and
-spring iy/b and y>c^ Suifiban/ 3^ still ^^
both conventional Glaus and single
contact sulfuric acid recovery.
GCA - Spring 1976 Vendor provided costs for plants of
ft ,,, (18) Republic Steel, Youngstown Sheet and
runaea; stretford still & Sulfiban
with Glaus and ' sulfuric acid recovery
conventional.
Done in support of Ohio SO- SIP.
EPA - July 1976 Glaus plant tail gas treatment study
for the refinery NSPS.
From this table it is apparent that the two Japanese processes, Diamox
and Takahax, are not reported. As well due to differences in design assump-
tions these studies are not readily comparable. Nor is there one published
overall cost comparison particularly for the lowest achievable emission rate
technology paths:
(1) Sulfiban + Glaus + Tail Gas Treatment
(2) Sulfiban + Double Contact/Double Absorption Sulfuric Acid
or in the case of <_ 8 gr/100 dscf organic sulfur in foul gas
(3) Stretford + Holmes Waste Liquor
(4) Takahax + Hirohax, Elemental Sulfur, or Sulfuric Acid Recovery
66
-------
Furthermore the representativeness (note, not the accuracy) of
the Massey/Dunlop case studies is uncertain because: (1) they postulate
an inlet (H2S) of 500 grains, which is higher than the bulk of American
coke plants (according to the experience of the various EPA Regional
Enforcement Divisions and DSSE), and (2) competition in the U.S. is
much keener in 1977 because of the positive developments at the Holmes1
Orgreave project for Stretford, the three American Sulfiban plants and
the entrance of Takahax into the American market than when the Massey/
Dunlop studies of 1971-1975 were developed.
One other difficulty in assessing the cost of achieving the LAER
level is the role of organic sulfides. If a given coke-oven gas contains
less than 8 gr H2S equivaliet/100 dscf of COS and CS2/ then it is techno-
logially possible for both Stretford and Takahax to achieve the LAER level
This is not true of all coke-oven gases, however (see Section III and IV).
Prediction of organic sulfide concentrations is difficult. For new coke
batteries, therefore, costing the technology to achieve LAER is difficult.
Extrapolation from existing batteries and coal blend data or from special
field tests of new coals in existing ovens is one possibility for lessening
such uncertainty. However, the fact remains that the cost of achieving the
LAER level will be dependant upon the composition
For all of these reasons the validity and representativeness of the
existing cost figures for coke-oven gas desulfurization are questionable.
Table 19, 20;.and 21 extracted from references (18) and (34) provide some
baseline cost data. The reader is asked to study the specific references
67
-------
Table 19. COSTS FOR COG DESULFURIZATION SYSTEMS AT REPUBLIC STEEL - CLEVELAND WORKS, $10'
''
Opt til requirement '
Battery Haiti plant, Installed (Bl»)
Site preparation and otlUtU«,C 201 of BLP
Filed capital Investment (FCI)
Working capital, 20% of groat operating coat
Total capital Investment
Annual operating coat
labor* .
Administrative and general overhead, 601 of labor
Dtllttlea*
Hatarlali* .
Local taxet and Insurance, 2.7X of FCI
Groat operating coat
Sulfur by-product credit,' $40/ton
Het operating eott
Annuallted coit .
Anmiallted capital coat*
d h
Federal Income tax '
Uet operating cost
Average annual coit
Average annual control coat, $/lh S removed'
RolBet-Slretford
Flint Ho. 1
7=450.0
1.490.0
8,940.0
295.0
9,235.0
210.0
126.0
'304-0
S96.0
241.4
1 ,477.4
233.4
1,244.0
940.7
165.0
1,244.0
2,349.7
0.180
Flant Ho. 2
5,400.0
1.080.0
6.480.0
181, B
6,661.8
173.1
103.9
186.2
270.9
,-m.tO.
909.1
- 95.4
813.7
Total
12,850.0 .
(12,100.0)°
2.S70.0
15,420.0
477.3
15,897.3 .
(15,147.3)*
383.1
229.9
490.1
866.9
416.3
2,386.5
328.8
2,057.7
678.5 1 1.619.2
' (1,S42.8)»
118.5
283.4 .
(270. 5)b
Sulflban
Flent Ho. 1
9,000.0
1.800.0
10.800.0
. 359.0
11.159.0
237.
142.
815.
307.
291.
'1.795.2
- 233.4
1,561.8
1,136.6
199.4
813.7 ; 2,057.7 1,561.8
1,610.7
3,960.3 : 2,897.8
(3,871.0) ;
0.302 0.215
(0.210)
0.222
Plant Ho. 2
4,900.0
980,0
5.880.0
195.9
6,075.9
164.1
98.5
396.4
161.8
158.8
979.6
^
Total
13,900.0
2.780.0
16,680.0
554.5
17.234,9
Dreva-Stlll
Flanf Ho. 1
8,400.0
1.680.0
10.080.0
276.6
10.356.6
402.0
241.2
1.212.2
469.0
450.4
2.774.8
^2JLJL
227.1
136.3
596.0
151.2
272.2
1.382.8
»3-»
884.2 i 2.446.0; 1.149.4
1
1
618.8 1.755.4
108.6
884.2
1.611.6
0.302
1,054.8
308.0 i 184.1
Flent Ho. 2
5.600.0
1.120.0
Totll
14.000.0
2,800.0
6.720.0. j '16.800.0
170.9
6.890.9
176.7
106.0
289.7
100.8
181.4
854.6
- 95,4
759.2
701.9
122.2
447.5
17.247.)
403.
242.3
885.8
252.0
453.6
2.237.4
233.4
2,004.0
1.756.7
306.3
2,446.0] 1,149.4 ; 759.2 2,004.0
4,509.4
0.245
2.388.3 ! 1.583.3 \ 4.067.0
i
0.183 0.296 '. 0.221 '
00
vendor eitlaatei, Decenber 1975. Each plant It Independent, Including tulfut recovery.
Vllpltle Corporation .offered a reduced price If contract vat awarded to provide both facilities tUultaneoutly.
e$lte preparation, utility connection* to battery limit plant, and COG connection to battery Halt plant bated on CCA revlev of similar coatt reported In
the literature.
Bated on Ctlllty Financing Method as codified by the Faxdiandle Eattern Flpelloe Company and detcrlbed In Reference 42.
*See Tablet 16 and i:.
Conservative ettlsate bated on price ef $4&/long too received by SOUIO for by-product tulfer. '
"Capital ccttt are cortlied bated on a discounted cash fleet of 8 percent over 20 yeera.
Average federal lacoce tax 1.731 percent of sun of totel capital resptrecent and working capital
-------
Table 20. COSTS FOR 60 x 10 scfd COG DESULFURIZATION SYSTEMS
AT YS&T S103, MAY 1976
Capital requirement
Battery limits plant Installed (BLP>*
Site preparation and utilities, 201 of BLP
Fixed capital Investment (TCI)
Working capital,6 201. of gross operating cost .
Total capital investment
Annual operating cost
Labor*1
Administrative and gene.al overhead;6 601 of labor
Utilities'1 '
Materials'*
Local taxes and insurance,6 2.7Z of PCI
Cross operating cost
Sulfur byproduct," credit $40/toa
Bet operating cost
Annual tzed cost
Annual ized capital cost
Federal income tax6*8
Net operating cost
Average annual cost
Average annual control cost, $/lb S removed
Holmes-Stratford
$7.500.0
1,500.0
9.000.0
219.2 ,
9.219.2
210.9
126.6
304.0
211.4
243.0
1,095.9
-233.6
862.3
939.0
163.4
862.3
1.964.7
0.150
Sulflban
10,000.0
2,000.0
12,000.0
326.7
12.326.7
253.9
153.6
. S95.4
304. £
324.0
1,633.7
-219.0
1,414.7
1,255.5
219.0
1,414.7
2,889.2
0.236
Dravo-Stlll
9,500.0
1.800.0
11,300.0
. 300.9
11,600.9
245.4
147.3
637.4
169.5
305.1
1,504.7
-219.0
1.285.7
1,181.6
206.0
1,285.7
2,673.3
0.218
"Manufacturer estimates, May 1976.
Site preparation, cose to bring utilities to the battery limit plant, and connect raw COG ducts
to battery Holt plant based on CCA review of similar costs reported in the literature.'
eBas«d on Utility Financing Method as modified by the Panhandle Eastern Pipeline Company and
described in Reference 18.
dSoo Tables 5 and 6. .''.
eConservatlbe estimate baaed on price of $44/ton currently received by SOHIO for byproduct sulfur.
Capital costs are amortized at a discount cash flow rate of 8 percent over 20 years. This method
yields an average annual capital cost, Including depreciation of 10.18 percent of total capital
Investment.
"Average federal Incoae tax 1.731 percent of sum of total capital Investment and working capital.
69
-------
«*. ut-^'l
1 « J4--T
Basis
VACUUM CARBONATE
20HH5CFTr 60MH5CFD
Item
DesulfuHiatlon Plant
Cooling H20, gpro'')
Power, KUH/day
Chemicals, l/day'0'
Steam, f/hr
ActlHer and/or
Ejectors
Condensate Treatment
Total
Claus Sulfur Plant
Steam Credits, f/hr
High Pressure
Low Pressure
Total
Net Desulf. + Sulfur
Plant Stm. demand, f/hr
_ j Sulfurlc Add Plant
0 Cooling H20. gpm(a>c)
Power. KUH/day
Steam Credit (600 pslg).
f/hr
Net System Manpower
Requirements
(1) Operator, man/1
shift
(11) Chemist, hr/day
90tn .
1011
1579
177
5421
1120
6541
558
282
.840
5701
230
(1241)
.1560
(3139)
2220
(4321)
1
2
98*n
1314
2526
177
7179(b
1120
8299
607
J07
914
7385
250
(1564)
1699
(4225)
2417
(5882)
1
2
Win
3032
4737
530
I
' 16,263
3380
19.643
1674
,846
2520
17.123
690
(3722)
4440
(9177)
6670
(12.973)
1
2
98tn
3942
7579
530
it.
21.5371"
, 3380 .
24.917
1824
922
2746
22.171
751
(4693)
4835
(12,414)
7263
(17,654)
1
2
: Sbo'gYalns lUS/lOO SCF at linlet
fk
SULFIBAN
20HH5CFD
gntjt
S30
1300
300
\
' 5840
.
5840
558
282
840
5000
230
(760)
1560
(2860)
2220
(2780)
1
2
98In
1060
1300
300
10,914
.
10.914
607
307
914
10,000
*
250
(1310)
1699
(2999)
2417
(7583)
1
2
PLANT
F1RHA CARL STILL STRETFORD U/EFFLUENT TREATMENT
60HHSCFD
SPla
1590
4148
900
17,520
*
17.520
1674
846
2520
15,000
690
(2280)
4440
(8588)
6670
(8330)
1
2
98Xn
3180
4148
900
32,746
.
32,746
1824
_922
2746
30.000
751
(3931)
4835
(8983)
7263
(22,737)
1
2
2JWMSCFD
93tn
1400
4380(d
-
5500
5500
558
282
840
4660
230
(1630)
1560
(5940)
2220
(3280)
1
2
60HH5CFD 20HHSCFO
u 93«iL 99tn
4200 xO
'll,220(dj 6000
N.A.
16,000
- . -
16,000 2200
1674
846
2520
13.480
690
(4890)
4440
(15,660)
6670
(9330)
11
2 2
60HM5CFQ
99ln
x!2
18,000
N.A.
-
-
6600
-
-
-
,
-
-
1
2
Footnotes; .
(a) Vacuum Carbonate ind Sulflban plants use once-through Hver cooling water. Flrma Carl Still plants employ eooUng tower water. Note that costs
per thousand gallons are different.
(b) Approximately 16 percent of stated steam rate Is required to supply Incremental heat to the actlfler.
(c) Figures In parentheses represent net requirements for a combination of desulfuHiatlon tnd sulfurlc add plants.
(d) %?&&: iK^Tl^^ """H"'t' «» - -1thou* "'Hgeratlon unit ,r, 6000 .nd 2880 KUH/day
(«) Na2C03 for the Vacuum Carbonate plant, aonoethanolamlne for the Sulflban plant. Information not available (MA) for Stretford technology.
-------
for a statement of assumptions and methods. To these data DSSE has added
new supplementary vendor cost estimates.
In order to facilitate comparisons, Table 23 was prepared. The
statistics of Tables 19-21 are adjusted in this comparison to place
all technologies on a before tax and by-product credit basis. Due to
the approximate year difference between the GCA and Dunlap/Massey estima-
tions, the reader may wish to raise Massey/Dunlap costs by about 10% - 15%.
Amortized capitol plus operating costs are: 5.6 - 9.2<£/MCF for
Stretford, 6.3 - 10.8<£ MCF for Sulfiban, and 7.1 - 11.3$ MCF for Vacuum
Carbonate. The Sulfiban capital cost estimate range ($7.07 - 10 million)
appears to be confirmed by an independent estimate of Sulfiban's costs,
shown in Table 22.
Tables 19-23 do not include the necessary cost of tail gas treatment
for achievement of an overall 10 gr "H-S"/100 dscf performance level by
the liquid absorption processes. Nor do these tables show the differential
costs of operating a liquid absorption tower/still column at the lowest
clean COG H-S concentrations. (Higher steam useage occurs at the lowest
ELS levels.) Table 25 indicates the differential steam demand for Sulfiban
between 40 gr/100 scf and 5 gr/100 scf is about 7.9 MMBtu per ft3 COG.
Assuming $2.00/MMBtu* this works out to $506 per day or 1.58
-------
Table 22. RECENT (JANUARY 1977) VENDOR BID
1. COG TO BE TREATED
100,000,000 scfd
250 gr/100 scf inlet H2S
19 gr/100 scf inlet O.S.
2.2 inlet CO2
50 gr/100 dscf overall outlet
2. DESIGN BASIS
Sulfiban Plus Single Contact Sulfuric Acid
t
In round figures:
.$13 million total battery limits capital cost
. $7 million - sulfiban
. $6 million - acid plant
3= (a) This capital cost is equivalent to 13<£/scf (Dunlap and Massey(1975))
- 7.8$/scf; GCAC (1976)
(b) The Sulfiban/acid plant ratio is 1.16 (Dunlap - 1.01)
72
-------
Table 23. COMPARISON OF COG DESULFURIZATICN COSTS
' I 1 . '
COST
Capital ($M4)
Annuaiized Capital
($/day)
Annuaiized Operating
($/Day)
Net Amortized
Cost, Before
Taxes
.$/Day
(Before by-product
credit) .
Estimate Date
500 gr -» 10 gr:COG
95% - CP; 97.5% Acid Plant
V.C. - DUNIAP
V.C. - Claus
2om
2.89
747
1220
L967
9.84
60MM
5.14
1329
2985
4304
7.14
V.C. - Acid
20MM
3.95
1021
1236
2257
11.29
60MM
7.54
950
2964
4914
8.19
500 gr -* 10 gr: COG
98$ CP; 97.5% A.P.
SULFIBAN - DUNIAP
Acid
20M4
3.75
970
1407
3775
6.29
60MM
7.07
1828
3491
5319
8.87
Claus
20MM
2.69
698
1468
2116
10.83
60W
4.67
1208
3743
4951
8.25
SULFIBAN -^ GCA
: Claus
60MM - Operating
(70MM - Peak
10.0
2750
3586
6336
10.56
500 gr * 10 gr
STRETFORD
Dunlap
60MM
.: 4.49
1161
2178
3339
5.57
GCA
60MM
7.5
2063
2337
4400
7.33
Wilputte
60MM
9.2
2493
2986
5479
9.16
. 1st Q.
First Quarter, 1975 Cost Estimates 5/76 1975 5/76 2/77
-------
Table 24. COMPARISON OF ALTERNATIVE EMISSION CONTROL SYSTEM COSTS FOR A 10 LTD SULFUR PLANT
(Cost Adjusted to June, 1975)
Total Costs
Differential Over Preceding Case
Control System
Base Case
Alternative I
Alternative II
Oxidation
Reduction
Investment
($)
v$ 902,000
1,028 ,000 (a)
l,320,00o!a}
l,765,000(a)
Annual
Operating
Cost
($/.vr)
$133,600
198,60o(a)
352,200(a)
442,000(a)
Emission Rate
Total Sulfur
As SOg
(Lbs/hr)
93
19
2
2
Investment
($)
_
126,000
292,000
737,000
Annual
Operating
Cost
($/.vr)
-
65,200
153,400
243,200
Emission Rate
Total Sulfur Unit
As S02 Cost
(Lbs/hr) ' ($/ton)
18 .
74 21o(f>]
UOTAQ* '
£ l""/u1
17 3406^
Notes:
(a) Includes costs of base case Claus sulfur recovery plant.
(b) Incremental costs per incremental ton of S recovered.
-------
Table 25. ENERGY DEMANDS FOR COKE-OVEN GAS TECHNOLOGIES
SULFIBAN - GLAUS PROCESS
Source
X)G flow
(MMSCFD)
tet Process
Steam
(lb/day)
Power ,
(kwh/day)
Power
(Btu per
ft3 COG)
Steam
Energy.
Btu/ftJ)
Total
Energy
Demand
Btu/ft3)
Ib steam/
1000 ft3
Basis Inlet
(gr/
10? Outlet
set
Massey &
Dunlap
:- 60
20,000
4,148
0.72
13.2
13.9
12.0
500
10
GCA,
(Y.S.& T.)
60
672,000
10,920
1.91
12.3
14.2
11.2
464
25
GCA
(RSC-C#1)
65
700,000
13,200
1.31
11.8
13.1
10.8
461
25
GCA
(RSC-C#2)
33
308,000
8,880
2.83
11.8
14.6
9.3
381
25
ATC
32
211,300
3,600
1.18
2.3
8.5
6.6
475
40
ATC
32
447,000
3,000
0.98
15.4
16.4
14.0
475
5
STRETPORD
Massey &
Dunlap
60
158 ,,400
18,000
3.15
2.7
5.9
2.6
"500
5
GCA
Y.S.& T.)
60
68,900
29,760
5.21
1.3
6.5
1.2
464
10
Wilputte
60 -
34,300
21,900
3.84
0.63
4.47
0.6
500
10
-------
The differential capital and operating costs for high sulfur recovery
for the liquid absorption systems with respect to conventional Glaus sulfur
recovery is shown in Table 24, which has been extracted from reference (13),
Page 8-12. The 10 long tons/day example plant cited in Table 24 is the size
of a Glaus plant needed to recover sulfur from a 30 MMSCFD COG flow at 500 gr
H2S/100 scf. This is exactly in the range of the discussion relevant to
Tables 19-22. Table 24 indicates that the increased capital investment
(20 yr, 8%, straight line) for the reduction tail gas systems (e.g., SCOT)
is ^ 2.8* per MCF. The amortized increase for tail gas treatment of Glaus
off-gas is ^ 3.6C/TV1CF or 33% over the baseline Sulfiban-Claus system of
Table 23. This last estimate, however, overstates the differential cost of
tail gas technology. Once in place, TGT equipment allows for a less effi-
cient Glaus system since its tail gas is then being cleaned. In fact, this
is the process selection made for the Clariton coke works at which the two
Glaus plants are only capable of ^ 92% yield. The SCOT TGT system improves
this to 99.9% yield.
B. Energy Impact
The energy demands of COG desulfurizative technologies capable of
meeting the lowest achievable S02 emission rate are shown in Table 25.
Both electrical energy and process steam demands are considered. Elec-
trical energy is rated at 10,500 input Btu per kwh. Steam is assumed
to require 1100 Btu/lb.
76
-------
Table 26. RELATIVE ENERGY DEMAND OF COG DESULFURIZATICN TECHNOLOGY
(COG FLOW - 20 MMSCFD)
COG ENERGY DEMAND
Stretford Sulfiban(a) Sulfiban(b)
COG Heat Content (5-10 gr) H2S** (40 gr) TRS* (10 gr) TRS
16,500 MMBtu 180 MMBtu/day 255 504
Day 1.1% 1.6% 2.9%
(a) Conventional Glaus.
(b) With tail gas treatment.
*
Percentage of COG heat content used as process steam or electricty.
**
Refers to gr/100 dscf of overall sweet COG plus tail qas emission. Add
5-25 gr organic sulfide (as H2S) per 100 dscf to compare to Sulfiban.
77
-------
The range of total energy demands is 5-16 Btu/ft COG processed. Liquid
oxidation requires the least overall power because process steam demands are
lower than for the liquid absorption processes. The energy cost per ft COG
rises sharply as the absorption processes are operated at higher H2S removal
efficiencies.
For example, to desulfurize to the lowest achieved level Sulfiban, 5 gr/
100 scf, an additional 8 Ib of steam per ft of COG, a doubling from the
40 gr/100 scf level, is required.
The energy demand for COG desulfurization is expressed as a fraction
of the energy content of COG produced in Table 26. COG is assessed at
550 Btu/ft . The energy demand, relative to the heat content of COG for
conventional moderate sulfur recovery efficiency H-S removal, and the
highest efficiency case are shown in this table. An additional equiva-
lent 1.3% of available COG heat value is needed to achieve the LAER
value over the base case of an overall 40 gr "H2S"/100 dscf, if Sulfiban
is chosen. Note that the Stretford process is able to produce a high
desulfurization efficiency (at least for H0S) with lower energy re-
&
quirements than for the Sulfiban process.
78
-------
REFERENCES
1. Ludberg, J.E. "Removal of Hydrogen Sulfide from Coke-Oven Gas by the
Stretford Process," a paper delivered before the 1974 meeting of the
Air Pollution Control Association, Denver, 1974.
2. Wilson and Wells, Coal, Coke and Coal Chemicals, first edition,
McGraw-Hill, 1950, p. 240.
3. Pennsylvania Department of Environmental Resources Rules and
Regulations. Title 25, Chapter 123.23.
4. Dunlap, R. W. and M. J. Massey, JAPCA 25(10) 1019-1027 (1975).
5. Stated during a discussion with Frank Vedja of the Koppers Co.,
in Pittsburgh, PA. on December 1976.
6. Massey, M. J., "Comments on the Technology and Economies of Coke-
Oven Gas Desulfurization," a 1971 working paper of the Allegheny
County Air Pollution Control Advisory Committee, 301 39th Street,
Pittsburgh,. PA.
7. This information derives from the Installation Permit application
of U.S. Steel Corporation before the Allegheny County Health
Department, 1972.
8. Homberg, O.A. and Singleton, A. H., JAPCA 25(4) 375-378 (1975).
9. Williams, J. A. and Homberg, 0. A., "Coke-Oven Gas Desulfurization
and Sulfur Recovery Utilizing the Sulfiban Process," a paper pre-
sented to the 34th Iron making conference of the A.I.M.E., St. Louis,
MO., March 1976.
10. Letter from Len Schuster, Applied Technology Corp. to Bernard Bloom,
U.S Environmental Protection Agency, December 23, 1976.
11. Acquired from discussions with J. Gordon Price, Dravo Corporation in
October and December 1976.
12. "Diamox Process for Coke-Oven Gas Clean-Up," Mitsubishi Kakoki
Kaisha, Ltd., July 8, 1975.
13. "Standard Support and Environmental Impact Statement Volume : Proposed
Standards of Performance for Petroleum Refinery Sulfur Recovery Plants",
U.S, EPA (450/2-76-016a), September 1976.
79
-------
14. Massey and Dunlap, Op. Cit.
15. "Process Desciption of Koppers Two Stage Hot Vacuum Carbonate
System," Koppers Company, Inc., 1976.
16. Homberg, Op. Cit.
17. Acquired from a discussion with Marcus Peters, Applied Technology
Corporation on December 27 and 27 1976, and January 16, 1977.
18. "Evaluation of the Technological Feasibility, and Cost of Selected
Control Alteratives Necessary to Meet the Proposed Ohio S02 Regulations
for Industrial Boilers and Processes", Volume III, GCA Corp., June 1976,
(A contractor report to EPA).
19. Op. Cit., reference 4.
20. Berlie, E. M., et. al., "The Role of the Claus Sulfur Recovery Process
in Minimizing Air Pollution", a paper presented at the 1974 meeting
of the Air Pollution Control Association, Denver, 1974.
21. California Air Pollution Control Regulations (South Coast APCD).
22. "Control of Air Pollution from Sulfuric Acid Plants", a draft written
by the U.S. Environmental Protection Agency for internal distribution,
August 1971.
23. Acquired from discussions with Donald Pogue of the Monsanto Corp.,
December 1976.
24. Letter from Donald Pogue to Bernard Bloom, January 10, 1977.
25. Information provided by Frank Smith of Peabody-Holmes in a telephone
call to Bernard Bloom on January 4, 1977.
26. Information acquired from Walter Carbone, Wilputte Corp., in a
telephone call with Bernard Bloom, January 4, 1977.
27. Telephone call between Bernard Bloom and Harold Tauscher, Hellinger
Engineering Corp., (and assigned to Kaiser Steel Corp.), January 5, 1977.
28. Telephone call between John Hemingway, Vfocdhall-Duckham (Pittsburgh, PA.),
and Bernard Bloom on January 5, 1977.
29. "Technical Bulletin for Coke-Oven Gas Desulfurization Equipment," Nippon
Steel Corporation, Plant and Machinery Division, No. PMD 23, APril 1976.
30. Manka, D. P., "Coke-Oven Gas Analysis," Instrumentation Technology,
February 1975, p. 45.
80
-------
31. Letter from Janes R. Zwikl (Shenango) to Ronald J. Chlebaski; Allegheny
County Health Department, May 31, 1974.
32. Letters of Earl F. Young, Jr. (J & L Steel) to the Allegheny County
Health Department, August 31, 1973 and May 20, 1974.
33. Ibid. Young.
34. Massey, M. and Dunlap, R. W., "Assessment of the Technologies for the
Desulfurization of Coke-Cven Gas," presented to the 34th Ironmaking
Conference of the AIME, Toronto, April 1975.
35. Data obtained from a meeting with representatives of Nippon Steel
Corporation, January 27, 1977, New York, NY.
36. Batterton, G. and Singleton, A., "Coke-Oven Gas Desulfurization by
the Sulfiban Process", presented to the 34th Ironmaking Conference
of the AIME, Toronto, April 1975.
81
-------
Appendix A -r .-VENDORS' OF COG TECHNOLOGY
PROCESS
Vacuum Carbonate
VENDORS
Frank Vedja
Kbppers Co.
Coke Plant Project Department
Engineerings & Contraction Division
Chamber of Commerce Building
Pittsburgh, PA. 15219
(412) 391-3300
Carl Still
J. Gordon Price
Dravo Corp.
1 Oliver Plaza
Pittsburgh, PA.
(412) 566-3264
15272
Sulfiban
Mark Peters
Applied Technology Corp.
4242 Southwest Freeway
Houston, TX. 77072
(713) 626-8000
Diamox
Mitsubishi Chemical Ind., Ltd.
277 Park Avenue
New York, NY.
(212) 922-3771
Takahax
Mr. Yamasaki
Nippon Steel Corp.
345 Park Avenue
New York, New York
(212) 486-7150
Stretford
Walter Carbone
12 Floral Park Avenue
Murray Hill, NJ.
(201) 464-5900
82
-------
Appendix A - VENDORS OF COG TECHNOLOGY ( cont'd)
PROCESS VENDORS
Sulfuric Acid Don Pogue
Enviro Chem Division
Lindberg Road
Monsanto Chemical
St. Louis, Missouri
83
-------
APPI5,
: r,
J,(
00
//*.'.tfS-V^^ /^/,/ xU/w ;/Ty.-- vy^^// :^;V-r^^v!.^A^
t'ii?- -:-:--'MJ₯ij£*''-: ^i^^^ ~^:
/*,
- **
bat-
. 77..
7f
.77
/Jf
3'r
3,*
Vo
.Pi-f
77, o
------- |