U.S. Environmental Protection Agency Industrial Environmental Research EPA~600/7-77-1 03b
Office of Research and Development Laboratory _
Research Triangle Park. North Carolina 27711 September 1977
NOX ABATEMENT
FOR STATIONARY SOURCES
IN JAPAN
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S.
Environmental Protection Agency, have been grouped into seven series.
These seven broad categories were established to facilitate further
development and application of environmental technology. Elimination
of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from
the effort funded under the 17-agehcy Federal Energy/Environment
Research and Development Program. These studies relate to EPA's
mission to protect the public health and welfare from adverse effects
of pollutants associated with energy systems. The goal of the Program
is to assure the rapid development of domestic energy supplies in an
environmentally—compatible manner by providing the necessary
environmental data and control technology. Investigations include
analyses of the transport of energy-related pollutants and their health
and ecological effects; assessments of, and development of, control
technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal
Agencies, and approved for publication. Approval does not
signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names
or commercial products constitute endorsement or recommen-
dation for use.
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161.
-------
EPA-600/7-77-103b
September 1977
NOX ABATEMENT
FOR STATIONARY SOURCES
IN JAPAN
by
Jurnpei Ando, Heiichiro Tohata, Katsuya Nagata, and B.A. Laseke
PEDCo. Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
Contract No. 68-01-4147
Task No. 3
Program Element No. EHE624
EPA Project Officer: J. David Mobley
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
PREFACE
This report describes recent developments in NO con-
iC
trol technology in Japan through May 1976, with emphasis on
flue gas denitrification. Flue gas denitrification is
considered necessary because the stringent Japanese ambient
standard for NO2 (0.02 ppm in a daily average) cannot be
attained by combustion control at stationary sources and by
the regulation of automobile exhausts. Many commercial flue
gas denitrification plants are in operation and under con-
struction, and many new processes are being developed.
Section 1 of this report introduces the NO problem in
5C
Japan. This section describes NO sources, emission regu-
j£.
lations, ambient standards, and ambient concentrations in
Japan. It also presents the estimated costs of controlling
emissions from stationary sources to achieve the ambient
standard.
Section 2 reviews combustion modification technology,
particularly for sources firing oil, which has been the
major fuel in Japan.
Section 3 describes dry denitrification processes,
mainly selective catalytic reduction using ammonia, and the
performance of the commercial plants.
11
-------
Section 4 reviews wet denitrification processes,
including many new processes for simultaneous removal of S02
and NO .
x
Section 5 discusses the significance of flue gas
denitrification and the advantages and disadvantages of many
combination systems of flue gas desulfurization and denitri-
fication.
iii
-------
CONTENTS
PREFACE ii
FIGURES vi
TABLES x
CONVERSION FACTORS AND ABBREVIATIONS xi
SECTION 1 INTRODUCTION TO NO REGULATIONS IN JAPAN 1
X
Ambient Concentrations and Standards 1
Emission Standards 1
Air Pollution Control Agreement in Chiba 4
Prefecture
Cost of NOX Abatement to Attain Ambient 6
Standard
SECTION 2 NO ABATEMENT BY COMBUSTION CONTROL 11
-" Ml
Classification of Combustion Control Techniques 11
Change of Operating Conditions 12
Modification of Combustion System Design 14
Other Methods 28
Application of Techniques 32
Investigation of Fuel NOX 34
Further Investigations 36
SECTION 3 DRY PROCESSES FOR NO REMOVAL 41
X
General Description 41
MHI SCR Processes 47
Hitachi Shipbuilding SCR Process 64
Hitachi, Ltd., SCR Proce'ss 68
Japan Gasoline Paranox Process 72
Kurabo SCR Process 76
SCR Processes with Santetsu (SARC) Catalyst 82
Other SCR Processes 89
Ammonia Injection without Catalyst 91
Reaction of Activated Carbon with NOX 93
Simultaneous Removal Processes using Activated 97
Carbon
Ebara-Jaeri Electron Beam Process 104
Other Dry Processes 106
iv
-------
SECTION 4 WET PROCESSES FOR NO REMOVAL 109
a
General Description 109
Tokyo Electric - MHI Oxidation Absorption 116
Process
Kawasaki Magnesium Process 117
Absorption Oxidation Processes 122
Fujikasui-Sumitomo Simultaneous Removal Process 124
MHI Oxidation Reduction Process 126
IHI Simultaneous Removal Process 129
Other Oxidation Reduction Processes for 129
Simultaneous Removal
Kureha Process 132
Mitsui Shipbuilding Process 135
Chisso Process 137
Asahi Chemical Process 139
SECTION 5 EVALUATION AND DISCUSSION 143
Significance of Flue Gas Denitrification 143
Combination of Flue Gas Desulfurization and 144
Denitrification
SECTION 6 REFERENCES 151
-------
FIGURES
Number Page
1 Two-stage combustion type burner for low-NO 16
formation
2 Nox emission from the low-NOx burner shown 17
in Figure 1 '
3 Effect of catalyst on NOX emissions from two- 18
stage combustion burner
4 Two-stage combustion-type burner for oil 19
5 Effect of low-NO burner on NO emissions 21
Ji ji
6 Atomizing nozzles in off-stoichiometric com- 22
bustion-type low-NO burner for oil
Jt
7 Effect of low-NO atomizer on NO emissions 23
X JC
8 Two-stage combustion for small boiler 25
9 Flow diagram of apparatus for producing and 29
supplying emulsified oil
10 NO emission with kerosene 30
jt
11 NO emission levels in oil-fired boilers 30
X '
12 NO emission levels in gas-fired boilers 31
X
13 Air flow in reversely-turned firing 33
14 Relation between nitrogen and sulfur contents 35
in heavy oils
15 Relation between nitrogen content in oil and 37
fuel NOV conversion ratio in boiler
ji
16 NO^ concentration versus nitrogen contents of 38
fuel in many boilers
VI
-------
FIGURES (Continued)
Number Page
17 Reduction of nitrogen by hydrodesulfurization 39
of heavy oil
18 Deactivation of catalyst on y-Al-jO- carried 46
by S0x
19 Formation temperature of NH4HS04 48
20 Criteria for catalyst for clean gas 50
21 Durability of catalyst for clean gas 50
22 Effect of oxygen on NO removal 51
X
23 Flowsheet of pilot plant (clean gas) 52
24 Effects of boiler load on NO removal 53
jt
25 Effects of ammonia converter 55
26 Pressure drop in fired-bed reactor with 57
different catalyst diameters
27 Effect of catalyst size on NO removal 57
.X.
28 Systems of pilot plant tests 58
29 Structure of moving-bed reactor 59
30 Results of first test 59
31 Effect of SV on NO removal and NH_ emission 61
X -j
32 Durability of NOjj reduction catalyst for low- 61
sulfur oil-burning gas
33 Particle size distribution of dust 62
34 Change of pressure drop with moving bed 63
35 Deposit of NH4HSO. in air heater 63
36 Flowsheet of Hitachi Shipbuilding process 65
VII
-------
FIGURES (Continued)
Number Page
37 Hitachi SCR pilot plant test (LNG) at a space 70
velocity of 20,000 hr-1
38 Hitachi SCR pilot plant tests after desulfuriza- 71
tion (heavy oil)
39 Flowsheet of Chiba plant, Kawatetsu Chemical 73
40 Structure of parallel-passage reactor 73
41 Flowsheet of Paranox process 74
42 NOX removal ratio (Kashima plant) 75
43 NO concentration (Kashima plant) 75
a
44 Pressure drop in reactor (Kashima plant) 75
45 Flowsheet of Kurabo SCR process 77
46 Structure of Kurabo moving-bed reactor 78
47 Results with SARC II catalyst (420°C) 85
48 Results with SARC III catalyst (420°C) 86
49 Results with SARC IV catalyst (420°C) 87
50 Gas velocity and pressure drop with SARC 88
catalysts
51 Results with SARC IV catalyst 90
52 NO adsorption capacity 94
X
53 Desorption of NOX adsorbed by activated carbon 96
by washing with water at different temperatures
54 Desorption of NOX adsorbed by activated carbon 96
by heating at different temperatures
55 S02 and NOX removal by activated carbon at 100
different temperatures and space velocities
56 Flowsheet of Unitika process 102
Vlll
-------
FIGURES (Continued)
Number Page
57 Apparatus for tests of electron beam process 105
58 Results with different intensities of electron 105
beam
59 NO absorption in EDTA-Fe(II) liquor 114
60 EDTA-Fe(II) concentration and NO absorption at 115
50°C
61 Flowsheet of Tokyo Electric - MHI oxidation 118
absorption process
62 Flowsheet of Kawasaki magnesium process 120
63 Flowsheet of Moretana simultaneous removal 125
process
64 MHI simultaneous removal process 127
65 Flowsheet of IHI simultaneous removal process 130
66 Effects of CuCl2 and NaCl concentrations on NO 131
removal efficiency
67 Effects of N02/NOX ratio and additives on 131
removal efficiency
68 Flowsheet of Kureha simultaneous removal process 133
69 Flowsheet of Mitsui Shipbuilding process 136
70 Flowsheet of CEC process 138
71 Flowsheet of Asahi Chemical reduction process 140
72 Combination of denitrification and desulfuriza- 145
tion
IX
-------
TABLES
Number Page
1 NOX emission standards for stationary sources 2
issued in 1973 and 1975
2 Expected reduction of air pollutants in Chiba 5
Prefecture due to agreements with industries
3 Estimated NOX emissions from boilers and furnaces 6
4 Consumption of fuels by boilers and furnaces 7
5 Costs of fuels 7
6 Assumed costs of flue gas desulfurization and 8
denitrification
7 Estimated cost of NOx abatement for boilers and 9
furnaces
8 Major plants using denitrification by selective 42
catalytic reduction
9 Denitrification plants planned by companies 43
10 Composition of dust from oil burning 45
11 Examples of gas composition 54
12 Denitrification plant cost 80
13 Dentrification cost (Kurabo process) 81
14 NOX adsorption capacity of activated carbon 95
15 Desorption of NOX in reducing gas 98
16 Effect of addition of base metal compounds to 99
carbon on NOx reduction efficiency
17 Major plants for NOX removal from flue gas by 110
wet processes
18 Specifications for the test plant 119
x
-------
CONVERSION FACTORS AND ABBREVIATIONS
CONVERSION FACTORS
The metric system is used in this report. Some of the
factors for conversion between the metric and English systems
are shown below:
1 m (meter) = 3.3 feet
1 m (cubic meter) = 35.3 cubic feet
1 t (metric ton) = 1.1 short tons
1 kg (kilogram) = 2.2 pounds
1 liter =0.26 gallon
1 kl (kiloliter) =6.29 barrels
1 kcal (kilocalorie) =3.97 Btu
The capacity of NO removal systems is expressed in normal
X
cubic meters per hour (Nm /hr). One Nm /hr = 0.59 standard
cubic foot per minute. For monetary conversion, the exchange
rate of 1 dollar = 300 yen is used.
ABBREVIATIONS
ESP electrostatic precipitator
kW kilowatt
kwh kilowatt hour
LNG liquefied natural gas
MW megawatt
SCR selective catalytic reduction
SV space velocity
XI
-------
SECTION 1
INTRODUCTION TO NOX REGULATIONS IN JAPAN
AMBIENT CONCENTRATIONS AND STANDARDS
In 1973 the Japanese government set an ambient standard
for N0_ more stringent than in any other country of the
world: 0.02 ppm in a daily average of hourly values, which
is roughly equivalent to 0.01 ppm in a yearly average.
Conformance with this standard is to be attained within 5
years in most districts and within 8 years in heavily pol-
luted cities such as Tokyo and Osaka. NO- concentrations
^
in large cities range from 0.03 to 0.04 ppm in a yearly
average and from 0.2 to 0.06 ppm in a daily average. In
January 1975 the Environment Agency reported that in only 3
cities among 147 was the N02 level as low as 0.02 ppm or
lower (yearly average).
Total man-made emission of NO in Japan is now about 2
Ai
million tons yearly. About 65 percent of the total emission
is derived from stationary sources; in large cities, however,
60 to 70 percent is derived from mobile sources.
EMISSION STANDARDS
NO emission standards for large stationary sources
Jt
were first issued in August 1973 (Table 1). These standards
-------
Table 1. NOx EMISSION STANDARDS FOR STATIONARY SOURCES ISSUED
IN 1973 and 1975
(ppm)
Source
Boiler (gas)
Boiler (solid)
Boiler (oil)
Metal-heating
furnace
Heating furnace
Cement kiln
Coke oven
Capacity, Nm /hr
More than 100,000
40,000 - 100,000
10,000 - 40,000
More than 100,000
40,000 - 100,000
10,000 - 40,000
More than 100,000
40,000 - 100,000
10,000 - 40,000
More than 100,000
40,000 - 100,000
10,000 - 40,000
More than 100,000
40,000 - 100,000
10,000 - 40,000
More than 100,000
More than 100,000
For new plants
1975
100
130
130
480
480
480
150
150
150
100
150
150
100
100
150
250
200
1973
130
130
n
480
480
n
180
180
n
200
200
200
170
170
170
n
n
For existing plants
1975
130
130
150
750
750
750
230
190
n
220
220
200
210
210
180
n
n
1973
170
n
n
750
n
n
230
n
n
220
220
n
210
210
n
n
n
°2 in
gas , %
5
6
4
11
6
n No regulation.
-------
take into account the status of abatement technology —
tail-gas treatment at nitric acid plants and combustion
control for other plants — and are similar to those in the
United States. The emission standards were later achieved
through 2 years of effort, but the ambient NO2 concentrations
were still much higher than the ambient standard. As a
result, new emission standards were promulgated in December
1975 (Table 1). To meet the new standards a considerable
number of plants are required to use, in addition to com-
bustion control, low-nitrogen oil or gas in place of the
cheap, grade-C heavy oil that is rich in nitrogen. The 1975
standard is applied to about 3000 plants, whereas the 1973
one was applied to about 1000 plants.
The NO emission standard for new automobiles weighing
X
over 1000 kilograms is 1.2 grams per kilometer and is equiva-
lent to that in California for 1975 and 1976 models. The
standard for new, smaller automobiles is 0.84 gram per
kilometer. Those figures will be vastly reduced in the near
future because of recent improvements in the technology to
reduce NO emissions from automobiles.
X
A serious problem in Japan is that the ambient standard
will be far out of reach even when the new emission standards
are attained. Flue gas denitrification is thus needed not
only for NO -rich tail gas from plants producing or using
-------
nitric acid but also for flue gas from numerous plants using
fossil fuels.
AIR POLLUTION CONTROL AGREEMENT IN CHIBA PREFECTURE
Most prefectural governments have made agreements with
local industries for pollution control. An example of
reductions expected in the Chiba prefecture is shown in
Table 2. Chiba is close to Tokyo and has one of the largest
industrial complexes in Japan. The agreement was made in
February 1976 with 40 companies that have 45 large plants in
the complex. S02 emissions from those plants will be
reduced from 13,395 Nm3/hr in 1973 to 4706 Nm3/hr in 1977 in
order to attain the national ambient standard for S02 ,
namely, 0.04 ppm daily average. NO emissions will be
K.
3
reduced from 9668 to 5119 Nm /hr in the same period. In
addition to combustion modification and change of fuel, flue
gas denitrification is required for several large boilers,
several heating furnaces, and a few iron-ore sintering
plants and coke ovens. As a result of these efforts ambient
NO 2 concentrations will decrease to 0.04 ppm daily average,
an intermediate goal. However, to attain the national
ambient standard of 0.02 ppm daily average, the NO emissions
X
from these sources will have to be reduced to about 2700
Nm3/hr.
-------
Table 2. EXPECTED REDUCTION OF AIR POLLUTANTS IN CHIBA PREFECTURE
DUE TO AGREEMENTS WITH INDUSTRIES
ui
Type of industry
Power
Steel
Oil refining
Petrochemical
Chemical
Other
Total
Number
of
plants
6
2
4
14
12
7
45
^x
T
NmJ/hr
1973
6,150
3,010
2,098
1,414
510
214
13,395
1977
1,745
1,159
645
858
207
92
4,706
Reduction,
%
71.6
61.5
69.3
39.3
59.4
56.7
64.9
N0x
T
Nm-'/hr
1973
5,218
1,930
1,043
1,066
248
163
9,668
1977
2,427
1,142
568
690
184
110
5,119
Reduction,
%
53.5
40.8
45.6
35.3
26.0
32.7
47.1
Partlculates
kg/hr
1973
822
670
478
289
223
48
2,531
1977
244
285
220
231
92
32
1,108
Reduction ,
%
70.4
57.0
54.1
18.0
58.8
34.0
56.2
-------
COST OF NO ABATEMENT TO ATTAIN AMBIENT STANDARD (12)
X
In October 1975 the Nitrogen Oxides Investigation Com-
mittee of the Environment Agency published a report describing
the estimated cost to stationary sources for NO abatement
X>
to attain the ambient standard. For convenience of cost
estimation, Japan was divided into four regions:
A. Large cities (Tokyo, Osaka, Nagoya, Chiba, etc.)
B. Industrial districts (Kashima, Toyama, Fuji,
Handa, Yokkaichi, Omuta, Kitakyushu, etc.)
C. Middle-size cities (Kyoto, Sapporo, Sendai,
Okayama, etc.)
D. Other districts.
NO emissions and abatement required for each region
X
are shown in Table 3. Table 4 lists the amounts of fuels
required for boilers and furnaces in 1980 without control
and with control to attain the ambient standard. Costs of
the fuels are given in Table 5.
Table 3. ESTIMATED NOV EMISSIONS FROM BOILERS AND FURNACES
A
(106 Nm3/year)
1973
1980 (without
control)
1980 (with
control)
Abatement ratio,
%
A
181.2
237.6
49.9
79
B
207.8
272.5
98.1
64
C
23.2
30.4
12.2
60
D
179.5
235.3
117.7
50
Total
591.7
775.8
277.9
64
-------
Table 4. CONSUMPTION OF FUELS BY BOILERS AND FURNACES
(kl/year)
Fuel
High-sulfur heavy oil
Medium- sulfur heavy oil
Low-sulfur heavy oil
Kerosene, naphtha
Gas
Coal
Total
Without control
72,400
101,500
24,800
2,500
15,600
11,200
228,000
With control
49,300
53,100
33,000
16,400
65,000
11,200
228,000
Table 5. COSTS OF FUELS
Fuel
Cost, $/kl or
kl equivalent
High-sulfur heavy oil (S >_ 1.5)
Medium-sulfur heavy oil (ID. 5 < S < 1.5)
Low-sulfur heavy oil (0.1 < S < 0.5)
Kerosene (S < 0.1)
Gas (For plants larger than 40,000 Nm /hr)
Gas (For smaller plants)
66.7
83.3
100.0
106.7
116.7
150.0
-------
The costs of flue gas desulfurization and denitrifica-
tion are based on the assumptions shown in Table 6.
Table 6. ASSUMED COSTS OF FLUE GAS
DESULFURIZATION AND DENITRIFICATION
Desulfurization
Denitrif ication
Investment
$/Nm3/hr
23.2
33.3
$/MW
70
100
Annual cost*
mills/Nm3
1.25
2.43
mills/kWh
3.75
8.29
Including depreciation (7 years) and interest (10% per
annum) at 70% operation.
Table 7 summarizes the estimated cost of NO control
J\.
for boilers and furnaces based on the above assumptions.
The investment costs are $5810 million for denitrification
and $1844 million for desulfurization. The annual costs
are $2950 million for denitrification and $680 million for
desulfurization. A change of fuel will require an additional
cost of $2596 million, whereas combustion modification costs
only $464 million in investments and $92 million in annual
costs. The total investment cost reaches $8118 million, and
the total annual cost reaches $6318 million.
In addition, NO abatement is required for other large
stationary sources, such as iron-ore sintering plants, coke
ovens, glass melting furnaces, and cement kilns. The costs
for those sources will be about $1700 million in investments
and $850 in annual costs.
-------
Table 7. ESTIMATED COST OF NO ABATEMENT
X
FOR BOILERS AND FURNACES
(millions of dollars)
(No. of plants)
NOX control
Flue gas denitrif ication
Investment cost
Annual cost*
Combustion modification
Investment cost
Annual cost*
SO- control
Flue gas desulf urization
Investment cost
Annual cost*
SO- - NO control
^ X
Change of fuel
Annual cost *
Total investment cost
Total annual cost*
Region classification
A ;
(1260)
3870,
1903
90
13:
767
283
V
903
4727
3102
B
(123)
1033
567
147
43
947
350
1193
2127
2153
C
(65)
247
117
27
3
27
10
120
301
250
D
(71)
660
363
200
33
103
37
380
963
813
Total
(1519)
5810
2950
464
92
1844
680
2596
8118
6318
Annual cost includes depreciation (7 years), interest
(10% per annum), labor, fuel, etc.
-------
Although recent progress in denitrification technology
has reduced denitrification costs below those in Table 6,
extremely large expenditures will still be needed to attain
the NO2 ambient standard.
The Environment Agency intends to re-estimate the costs
more precisely to allow public evaluation of NO abatement
5C
programs.
10
-------
SECTION 2
NOX ABATEMENT BY COMBUSTION CONTROL
CLASSIFICATION OF COMBUSTION CONTROL TECHNIQUES
Combustion modification techniques used in Japan for
NO control can be classified into the following three
X
categories:
(1) Change of Operating Conditions
Low excess-air combustion
Promotion of mixing of fuel with air in com-
bustion chamber
Reduction of heat load in combustion chamber
Less air preheating
(2) Modification of Combustion System Design
Modifications of burner design
Staged combustion
Flue gas recirculation
Water or steam injection
(3) Other Methods
Change of fuel
Modification of firing
Although changes of operating conditions can be rela-
tively easy at existing installations, they usually reduce
the NO emissions only slightly and often cause operating
X
difficulties. On the other hand, modification of the
combustion system design is a promising control technique.
11
-------
Fewer problems are encountered when design modifications are
incorporated in new plants; whereas, relatively large recon-
struction costs are required when they are applied to
existing installations.
NO emissions from combustion processes are formed by
J\.
two mechanisms: thermal NO (that generated in the com-
X
bustion process) and fuel NO (that attributable to nitrogen
J\.
content of the fuel). Formation of thermal NO can be
X
suppressed effectively by decreasing the oxygen concentra-
tion in the combustion regions, by shortening the residence
time of combustion gases in the high temperature zones, or
by lowering the flame temperature. Reduction of fuel NO
X
emissions requires reducing both the oxygen concentration in
the reaction zone and the nitrogen content of the fuel.
CHANGE OF OPERATING CONDITIONS
Low Excess-Air Combustion
Low excess-air combustion (0.6 - 1.0% O- in flue gas
from oil burning) has been used in large boilers for several
years to reduce the S03 that causes low-temperature cor-
rosion. Low excess-air combustion is also useful in NO
j*L
abatement, particularly with large boilers burning high-
nitrogen fuels, such as grade B (0.08 - 0.35% N) or grade C
(0.25 - 0.35% N) heavy oil. The technique is not widely
used with small boilers, which usually burn low-nitrogen
12
-------
fuels such as kerosene, grade A heavy oil, or gas. The
small boilers normally are unmanned, and their operation is
not very well controlled. Therefore, operation with low
excess air in these boilers tends to cause incomplete
combustion.
.Promotion of Mixing of Fuel with Air in Combustion Chamber
It has been reported that a considerable NO reduction
J\.
is achieved through a slight adjustment of the air register;
for example, a change in the vane angle. Very few existing
installations use this technique. Changing the air system
makes it necessary to readjust the air fuel mixture. An
improper adjustment may increase the quantity of unburned
products such as soots, carbon dioxide, and hydrocarbons, or
it may cause unstable combustion. In new boilers and also
in existing ones, the air register is often modified when a
low-NO burner is installed.
X
Reduction of Heat Release Rate in the Combustion Chamber
NO emissions can also be reduced by decreasing the
boiler load. The use of this method in existing installa-
tions, however, results in a decrease in fuel efficiency and
output power. Therefore, this method is regarded as an
emergency measure, to be applied, for example, when the
ambient oxidant concentration exceeds 0.3 or 0.5 ppm and the
prefectural governor issues a recommendation or order to
decrease fuel consumption.
13
-------
Manufacturers currently are designing boilers with
lower heat release rates than those of older ones. The new
boilers are larger and more easily adaptable to such NO
5C
control techniques as two-stage combustion.
Less Air Preheating
A reduction in air preheating temperature reduces the
flame temperature and consequently the formation of thermal
NO . This method is seldom used because it also reduces
x
boiler efficiency. Moreover, application is limited to
large boilers with air preheaters, for which better control
techniques are available.
MODIFICATION OF COMBUSTION SYSTEM DESIGN
Modification of Burner Design
Low-NO burners developed in Japan may be classified
X
into the following four types, based on the NO suppression
X
principles and burner configuration:
(1) Good-mixing type,
(2) Divided-flame type,
(3) Self-recirculation type,
(4) Staged-combustion type.
Good-mixing and divided-flame type burners are useful
in reducing formation of thermal NO but have little effect
X.
in reducing fuel NO . Self-recirculation and staged-combus-
X
tion types, by contrast, seem to reduce emissions of both
types of NO .
14
-------
Since three of the categories -- good mixing, divided
flame, and self recirculation -- were discussed in earlier
reports (1,2) only the staged-combustion type burners are
described here. There are two types of staged-combustion
burners: the two-stage combustion type and the off-stoichio-
metric combustion type. A two-stage combustion type burner
(Figure 1) has been developed by Tokyo Gas Company for the
firing of town gas. A more advanced burner of this type
burner has a catalyst at the outlet of the preliminary
combustion region for further reduction of NO (3). Figure 2
jv
shows the emission levels with this burner. HCN and NH~ are
likely to play a significant role in NO formation in the
,X
second stage. The catalyst reduces HCN and NH_ together
with NO (Figure 3). These burners are not yet used com-
J\,
mercially because of problems such as flashback.
The configuration of the two-stage combustion burner
fired with oil is illustrated in Figure 4 (4). This type,
with a precombustion chamber, is designed by Kawasaki Heavy
Industries. The chamber is installed in the wind box, and
the second-stage air inlet is placed at the end of the
chamber. Oil is fired in the chamber with an insufficient
amount of air. Vaporization of the oil results from the
high temperatures of circulating gas and of the refractory.
Considerable reduction of both forms of NO is thus attained.
15
-------
GAS OR
GAS PRE-
MIXED WITH
AIR
FIRST STAGE
SECOND STAGE
AIR
Figure 1. Two-stage combustion type burner for
low-NO formation.
16
-------
50
40
Q.
Q.
2 30
I—
Cd
CJ
20
10
0
40 60 80 100
PREMIXING RATIO
(AIR/AIR + FUEL), %
Figure 2. NO emission from the low-NOx burner
shown in Figure 1.
(Fuel, methane; thermal input, 95,000 kcal/hr;
NO concentration corrected to 0% 0-)
17
-------
CL
QL
2 30
o
o
o
5
20
10
-
—
-
1
2
J
4
1
3
2
4
0
1
3
2
.n
4 :
n.
1
—
3
2
11
-
4
n
CATALYST CATALYST CATALYST CATALYST
NONE A B C
1:
2:
3:
4:
NO AFTER SECOND STAGE (CORRECTED TO 0% 02)
HCN
NH3 !• AFTER FIRST STAGE
x
Figure 3. Effect of catalyst on NO emissions from
two-stage combustion burner.
(Fuel, methane; thermal input, 95,000 kcal/hr;
sv, 65 hr~l; premixing ratio, 0.50;
equivalent ratio in first stage, 1.67)
18
-------
WIND BOX
ATOMIZER
FIRST-STAGE AIR SECOND-STAGE AIR
Figure 4. Two-stage combustion-type burner for oil.
19
-------
NO concentrations from this burner are compared with those
X.
from a conventional burner in Figure 5.
An off-stoichiometric combustion type burner for gas-
firing was mentioned in a previous report (1). Figure 6
shows arrangement of the fuel-injection holes of atomizing
nozzles in an off-stoichiometric combustion burner for oil-
firing (5) , developed by Volcano Company. The main feature
of this atomizer is that all of the fuel-injection holes are
not the same size. As combustion air is uniformly admitted
around the atomizer, the fuel-injection holes with larger
diameters produce fuel-rich combustion zones whereas the
smaller holes produce fuel-lean regions. Therefore, off-
stoichiometric combustion is achieved.
Several different arrangements of fuel-injection holes
are available. The optimum arrangement may be determined by
experimentation. Because of the ease of installation and
the low cost, this burner has been used widely. In some
cases, altering only the nozzle tips can successfully
reduce NO emissions. NO levels produced by such a burner
^C X
are shown in Figure 7. This burner tends to increase soot
emissions when excess air is low because some of the flame
zones are deficient in oxygen.
20
-------
300
Q.
n.
CM
o
o
I—
o
o
LU
E£.
C£.
O
200
100
0
CONVENTIONAL
BURNER
LOW-NOV BURNER
/\
0 1000 2000
FUEL FLOW RATE, liter/hr
Figure 5. Effect of low-NO burner on NO emissions,
A X
(Fuel, grade C heavy oil; N = 0.206%)
21
-------
OIL-
STEAM-
Figure 6- Atomizing nozzles in off-stoichiometric
combustion-type low-NO burner for oil.
J x
22
-------
400
a.
DL
o^ 300
<**
Q
liJ
O
LU
CCL
CC.
O
O
200
100
60
80
LOAD, %
100
Figure 7. Effect of low-NO atomizer on NO emissions
X X
(Boiler capacity, 55 t/hr; fuel, grade C heavy oil;
air temperature, 280°C; four atomizers)
23
-------
Staged Combustion
There are two major categories of staged combustion:
two-stage (the numbers of stages seldom exceed two) and off-
stoichiometric.
Two-stage combustion - In two-stage combustion, about 80 to
90 percent of the stoichiometric air needed for combustion
is admitted to the first stage. Oxygen-starved combustion
in this region reduces both thermal and fuel NO . Reduction
A
efficiencies usually are about 30 to 50 percent for thermal
and less than 50 percent for fuel NO .
x
The application of two-stage combustion is classified
into four types determined by the following locations of
second-stage air ports.
1) On the furnace wall above the burners.
2) Top burners for air injection only.
3) Side or rear walls of the furnace.
4) On the circumference of the burners.
Large utility boilers employ types 1) or 2). Type 2),
called quasi-two-stage combustion, is used in units to which
type 1) cannot be applied (6). Types 3) and 4) are used in
medium and small boilers. Type 3) is not popular for water
tube boilers because it requires considerable remodelling.
Type 4), as illustrated in Figure 8, can be readily applied
to small installations with a single burner because it
24
-------
FUEL
FIRST-STAGE
AIR
SECOND-STAGE
AIR
OUTLET OF SECOND-
STAGE AIR
Figure 8. Two-stage combustion for small boiler.
25
-------
requires little remodelling. Two-stage combustion cannot be
applied to installations with furnace dimensions that cannot
accomodate the greater flame length. Also two-stage com-
bustion tends to increase the amount of unburned products,
particularly CO (7).
Off-stoichiometric Combustion - This method has an effect
similar to that of two-stage combustion and is readily
adapted to medium and small boilers with several burners, to
which two-stage combustion is not easily applied.
Location of fuel-rich and fuel-lean burners or air
ports is decided after systematic tests. Often, it is
effective to place the fuel-lean burners or air ports in the
central upper parts of furnace walls above the burners, or
in the regions of highest heat release (6).
Flue Gas Recirculation
In flue gas recirculation, NO reduction is achieved
X
through the decrease of flame temperature. Thus, fuel NO
A
is not reduced, but thermal NO is reduced by 30 to 40 per-
X
cent. Recirculation ratios are limited to about 30 to 40
percent to prevent unstable firing, although this limit is
lower with larger units. Keeping the recirculation ratios
in suitable ranges may improve combustion conditions and
decrease the quantities of unburned products. however, the
decrease of flame temperature alters the distribution of
26
-------
heat transfer and lowers the fuel efficiency of existing
boilers.
A recirculation fan and additional duct work are
required to implement flue gas recirculation. As a result,
installation cost is considerably higher than that for two-
stage combustion and more installation space is required.
Therefore, this method is not used with small boilers or
furnaces. Many operators of large boilers recirculate flue
gases because of the stabilizing effect, which reduces
operational problems.
Water or Steam Injection
Water or steam injection reduces NO emissions by
decreasing the flame temperature. There are three injecting
methods:
a) Injecting into the combustion air.
b) Injecting into the combustion chamber. (This
includes increasing the steam flow rate in the
steam atomizer.)
c) Mixing water with the fuel (emulsification).
With an equal injection rate, b) and c) offer greater
NO reduction than a) because of the greater reduction in
x
flame temperature. The injection ports close to the burners
are effective for b). The upper injection rate limit is
about 5 kg/104 kcal (8).
27
-------
With the use of this technique, combustion character-
istics are improved, making it possible to reduce excess
air; therefore, the decease in thermal efficiency may not be
very great.
There are two kinds of emulsified fuels: water drop-
lets suspended in oil (W/0 type) and oil particles suspended
in water (0/W type). The former is used in oil firing
because the resulting emulsion viscosity is less than that
of an O/W type emulsion.
Figure 9 shows a fuel emulsion system developed by
Kawasaki Heavy Industries (4). Oil to which a slight amount
of emulsifying agent is added is mixed with water in a
mechanical mixer. This emulsified fuel is reserved in a
service tank and pumped to the burners.
Since it does not reduce fuel NO , this method is
x
useful with low-nitrogen oils. Figure 10 shows NO emission
levels obtained by using emulsified fuel.
OTHER METHODS
Change of Fuel
Except in special cases, NO emissions are related to
A.
nitrogen content of the fuel, which decreases in the follow-
ing order: solid fuels (coal and coke), liquid fuels (petro-
leum fuel oils), and gaseous fuels (town gas, LNG and LPG).
28
-------
TO BURNERS
1: WATER PUMP
2: OIL FLOW CONTROL VALVE
3: EMULSIFIER PUMP
4: MECHANICAL MIXER
5: RECIRCULATION PUMP
6: EMULSIFIER RESERVOIR
7: FUEL SERVICE TANK
Figure 9. Flow diagram of apparatus for producing
and supplying emulsified oil.
29
-------
100
80
e\j
o
•* 60
o
ce.
on
o
40
20
O
O
AA J(£ROSENE,
A A A
1357
STEAM FLOW RATE, t/hr
Figure 10. NO emission with kerosene,
400
. 300
CM
O
o
o 200
a:
o
o
x 100
GRADE A HEAVY OIUW
I
I
10 20 30
STEAM FLOW RATE, t/hr
40
Figure 11. NO emission levels in oil-fired boilers
X
30
-------
300
10 20 30
STEAM FLOW RATE, t/hr
40
Figure 12. NO emission levels in gas-fired boilers
jfX
31
-------
NO emissions are lower with low-nitrogen oil, such as
X
kerosene or grade A heavy oil, than with grade C heavy oil
(Figure 11) (9). The emissions from gas burning decrease in
the order of LPG (C3Hg and C4H10), LNG (CH4), and town gas
(synthetic gas with a heating value of about 5000 kcal/Nm ),
as shown in Figure 12 (9).
Modification of Firing
It is well-known that tangential firing gives lower NO
X
emissions than do front and opposed firing because the flame
temperature is lower. This lower temperature is caused by
better heat emission resulting from a larger flame volume.
For small boilers reversely turned firing is used to
reduce NO emissions. This type of firing is illustrated in
X
Figure 13. NO reduction is obtained by recirculation in
J\.
the flue tube.
APPLICATION OF TECHNIQUES
Combination of Techniques
A combination of the above-mentioned techniques can
increase the reduction of NO emissions. However, a com-
J\.
bination of techniques based on the same suppression princi-
ples, such as the combination of two-stage and off-stoichio-
metric combustion, is not efficient.
With large boilers flue-gas recirculation is generally
combined with two-stage combustion. With smaller boilers,
32
-------
SMOKE TUBE
•'FLUE1 TUBE
(..
FUEL—**
AIR
Figure 13. Air flow in reversely-turned firing.
33
-------
control techniques similar to those used with large ones are
practiced but the use of low-NO burners is more popular.
J\.
Many oil-refining furnaces use a combination of a two-stage
combustion and a self-recirculation-type, low-NO burner.
Ji
Field Tests
Since 1974 the Japan Environment Agency and Tokyo
Metropolis Bureau of Environmental Protection have made
field tests of NO control techniques (7,10). NO reduc-
5C • .JC
tions reported by the Japan Environment Agengy have ranged
from 19 to 67 percent with oil firing and from 19 to 73
percent with gas firing.
INVESTIGATION OF FUEL NOV
X
From 1974 to 1975 the Tokyo Metropolis Bureau of
Environmental Protection investigated the relationship
between the nitrogen and sulfur contents of fuel and the
fuel NO conversion ratio (11).
X
Figure 14 shows the relationship between the sulfur and
nitrogen contents of fuels.
Nitrogen contents of fuels fed to 90 commercial boilers
were measured, along with NO emissions from those boilers.
A.
More detailed tests were done with one boiler to determine
the effects of excess air, boiler load, and nitrogen content
on the fuel NO conversion ratio. These tests showed that
X
nitrogen content was the most significant factor influencing
34
-------
0.4
0.3
Ul
o
CJ
z
LU
CD
s
I—1
o
•
ro
0.1
X *
x
O x
"**•
o
0
8 o
AA
A
SULFUR CONTENT, %
• GRADE A
A GRADE C
x GRADE B
o MIXTURE OF A AND B
Figure 14. Relation between nitrogen and sulfur
contents in heavy oils.
35
-------
fuel NO conversion ratio. The relationship is expressed by
X
the following equation (Figure 15) :
a = (1 - 4.58 n + 9.50 n2 - 6.67 n3) x 100%
Where
a = Fuel NO conversion ratio (%)
J\
n = Nitrogen content in fuel (%)
Applying the above equation to the results of tests
with the 90 boilers (Figure 16) yields the following:
[NO ] = 1550 n - + 110 (ppm)
Where
[NO ] = NO concentration corrected to 0% 09 (ppm)
Jei ^C £*
With this equation it is possible to predict approxi-
mate NO emissions for boilers burning a fuel oil with a
known nitrogen content. When an oil containing 0.2 percent
nitrogen is burned, the amount of fuel NO is about equal to
Jv
that of thermal NO .
x
FURTHER INVESTIGATIONS
A low nitrogen content in fuel is significant for NO
reduction. The Environment Agency has investigated nitrogen
removal by hydrodesulfurization of heavy oil, which has been
carried out commercially in Japan (11) . Figure 17 shows
examples of the reductions achieved. About 30 percent of
the nitrogen can be removed, but this is generally insuf-
ficient to meet emission regulations.
36
-------
100
' 80
n
>
I
I 60
i
I 40
o
0 20
• x a = (1 . 4.58n + 9.50n2 - 6.67n3) x 100
0.2 0.4
NITROGEN CONTENT IN OIL, %
0.6
Figure 15. Relation between nitrogen content in oil and fuel
NO conversion ratio in boiler.
x
(Steam flow rate, 1 t/hr; load, 40-100%; excess air,
5-100%; base fuel, kerosene and grade B heavy oil;
additive containing nitrogen, quinoline and pyridine)
37
-------
300
CM
O
§
o
o
LU
I—
UJ
Od
O
o
200
100
INO.
« + 180
3»,1 ' A55°"
O O
o
•*"ol + 40
0.1 0.2
NITROGEN CONTENT IN FUEL,
0.3
Figure 16. NO concentration versus nitrogen
5C
contents of fuel in many boilers.
38
-------
0.3
0.2
cr>
o
0.1
•O
O CHARGE
A PRODUCT
SULFUR CONTENT, %
Figure 17. Reduction of nitrogen by
hydrodesulfurization of heavy oil.
39
-------
Research and development on the following may be
required for further NO control, in Japan:
J\.
1) Control techniques, such as fluidized bed com-
bustion, for coal- or coke-fired processes.
2) Control techniques for high-temperature furnaces,
such as those used for glass-melting.
3) Low-NO burners for LPG firing.
A,
4) Effective catalysts for better nitrogen removal in
hydrodesulfurization of fuel.
40
-------
SECTION 3
DRY PROCESSES FOR N0,r REMOVAL
v
GENERAL DESCRIPTION
Classification of Dry Processes
Dry processes for denitrif ication that have been
developed in Japan may be classified as follows:
1. Selective catalytic reduction (SCR) with ammonia
2 . Ammonia reduction without catalyst
3. Activated carbon process (simultaneous removal of
N0v and S09)
X £
4. Electron beam radiation (simultaneous removal of
NO and SO.,)
X Z*
5 . Adsorption
6. Catalytic decomposition
SCR has been used commercially in many plants (Tables 8
and 9) . Pilot-plant tests have been carried out with other
processes.
Reduction of NO with Ammonia
^™*™ • " •' — •"•• •* "—• HM^™^»™ - i ^^ . .. - i m~— *^ •-. T i_i-.-i_ •_ _•
Usual reactions of ammonia with NO are shown in
X
equations (1) and (2) .
4NH3 + 6NO = 5N2 + 6H20 (1)
8NH3 + 6N02 = 7N2 + 12H20 (2)
41
-------
Table 8. MAJOR PLANTS USING DENITRIFICATION BY
SELECTIVE CATALYTIC REDUCTION (SCR)
Process developer
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Hitachi Shipbuilding
Hitachi Shipbuilding
Hitachi Shipbuilding
Tokyo Electric-Mitsubishi H.I.
Kurabo
Kurabo
Kansai Electric-Hitachi Ltd.
IHI-Mitsui Toatsu
Chubu-MKK
Mitsubishi H.I.
Kobe Steel
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Hitachi Ltd. -Mitsubishi P.C.
Hitachi Ltd.
Ube Industries
Mitsui S.B. -Mitsui P.C.
Mitsui S.B. -Mitsui F.C.
MKK-Santetsu
MKK-Santetsu
MKK-Santetsu
Seitetsu Kagaku
Japan Gasoline
Japan Gasoline
Asahi Glass
Plant owner
Sumitomo Chemical
Higashi Nihon Methanol
Nihon Ammonia
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Kansai Oil
Idemitsu Kosan
Shindaikyowa Pet. Chem.
Tokyo Electric
Kurabo
Kurabo
Kansai Electric
Chubu Electric
Chubu Electric
Mitsubishi H.I.
Kobe Steel
Mitsui Toatsu
Mitsui Toatsu
San Polymer
Japan Novopan
Mitsubishis P.C.
Kawasaki Steel
Chiba Pet. Chem.
Mitsui Pet. Chem.
Ukishima Pet. Chem.
Okayama Paper
Kawasaki Steel
Nippon Yakin
Seitetsu Kagaku
Kashima Oil
Fuji Oil
Asahi Glass
Plant site
Sodegaura
Sodeqaura
Sodegaura
Anegasaki
Anegasaki
Niihama
Sodegaura
Sodegaura
Sakai
Chiba
Yokkaichi
Minamiyokahama
Hirakata
Hirakata
Sakaiminato
Shinnagoya
Yokkaichi
Hiroshima
Kakogawa
Sakai
Sakai
Osaka
Sakai
Yokkaichi
Chiba
Chiba
Chiba
Chiba
Okayama
Chiba
Kawasaki
Kakogawa
Kashima
Sodegaura
Keihin
Capacity,
Nm3/hr
30,000
200,000*
250,000*
100,000*
200,000*
200,000*
250,000
300,000
5,000
350,000
440,000
10,000*
5,000
30,000
4,000
8,000
100
4,000
600
1,000*
3,000
4,000*
8,000*
150,000
350,000
10,000
200,000
240,000
1,500
1,000
15,000
15,000
50,000
70,000
70,000
Source of gas
Oil-fired boiler
Heating furnace
Heating furnace
Gas-fired boiler
Gas-fired boiler
Heating furnace
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
CO-fired boiler
Oil-fired boiler
Gas-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Sintering plant
Gas-fired boiler
Oil-fired boiler
Gas-fired boiler
Gas-fired boiler
Oil-fired boiler
Coke oven
Oil-fired boiler
. Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Coke oven
Oil-fired boiler
Oil-fired boiler
Heating furnace
CO boiler
Glass furnace
Completion
July 1973
May 1974
Mar. 1975
Feb. 1975
Feb. 1975
Mar. 1975
Mar. 1976
Oct. 1976
Nov. 1973
Nov. 1975
Dec. 1975
Jan. 1974
Nov. 1973
Aug. 1975
Jan. 1975
Oct. 1974
Oct. 1974
Dec. 1974
May 1974
Oct. 1973
Oct. 1974
Oct. 1974
June 1974
Dec. 1975
Oct. 1976
Jan. 1975
Sept. 1975
Aug. 1976
Dec. 1974
Mar. 1975
June 1976
June 1975
Nov. 1975
Mar. 1976
Apr. 1976
* Clean gas; those without asterisks are for dirty gas.
-------
Table 9. DENITRIFICATION PLANTS PLANNED
BY COMPANIES
Company
Hokkaido Elec.
Chubu Elec.
Kyushu Elec.
EPDC
EPDC
Tobata Kyodo
Chubu Elec.
Plant
site
Tomakomai
Chita
Kokura
Isogo
Takasago
Tobata
Chita
Capacity,
MW
300
700 x 2
600 x 2
265 x 2
250 x 2
375
375
Fuel
Coal
LNG
LNG
Coal
Coal
LNG
Oil
Process
N.d.*
SCR1"
4.
AR§
T
AR§
t
N.d.
AR§
Scheduled
completion
1978
1977
1978-79
1977-78
1977-78
1980
1977
Not determined
t SCR (Hitachi, Ltd.)
T SCR (Mitsubishi Heavy Industries)
J Ammonia reduction, possibly by the Exxon Process
43
-------
Since the presence of oxygen promotes the reactions, the
actual reactions may be better represented by equations (3)
and ( 4 ) .
+ 4NO + O = 4N + 6HO (3)
4NH + 2N02 + 02 = 3N2 + 6H20 (4)
In addition, the following reactions between ammonia
and oxygen can take place:
+ 30 = 2N -f 6H0 (5)
4NH + 509 = 4NO + 6H O (6)
,3 ^ ^
4NH3 + 402 = 2N20 + 6H2
-------
developed to minimize the influence of dust. For example,
a parallel-passage reactor has been used commercially, and
moving-bed reactors have been tested in pilot plants.
Table 10. COMPOSITION OF DUST FROM OIL BURNING (12)
A
B
C
9.8
5.9
Si°2
6.0
2.4
S°4
42.6
58.4
v2o5
18.0
12.2
Fe2°3'
4.6
3.6
'A12°3
5.3
1.5
NiO
2.3
2.6
CaO
5.7
3.0
MgO
1.8
2.5
Na2°
11.9
14.8
Another problem is catalyst poisoning by SO in exhausts
Jt
from boilers burning coal and oil and by alkaline vapor in
exhausts from glass-melting furnaces, cement kilns, and
other sources. SO mainly affects the catalyst carrier,
X
which is usually porous alumina. Alumina tends to react
with SO , particularly with S0_ , to form aluminum sulfate,
X -j
leading to a decrease in surface area and catalyst activity
(Figure 18). Titania and silica are more resistant to S0_
than alumina.
Base metal oxides commonly used as catalysts also tend
to react with SO to form sulfates. The sulfates, however,
A.
are usually still reactive and less likely to promote the
decomposition of ammonia by reacting with oxygen. There-
fore, some catalysts are made of base metal sulfates.
A serious problem common to ammonia reduction pro-
cesses with or without catalyst is the formation of ammonium
45
-------
160
en
CVJ
< 120
£ 80
GO
a 40
0 0.02 0.04 0.06 0.08 0.10
WT OF SULFUR DEPOSIT/WT OF CATALYST
Figure 18. Deactivation of catalyst on
y-Al O carrier by SO .
^ J A.
46
-------
bisulfate, as shown in Figure 19. For example, when the gas
contains 10 ppm each of SO- and NH_, liquid ammonium bisul-
fate forms at 210°C. The bisulfite usually is formed in a
heat exchanger used for heat recovery after the reduction.
The bisulfite melt is corrosive, and it disturbs the heat
transfer. At lower temperatures, the bisulfate solidifies
and often reacts with ammonia to form ammonium sulfate. A
corrosion-resistant material must be used for the heat
exchanger. During plant operation, it is necessary to
remove the bisulfate or sulfate from the exchanger occasion-
ally by steam blowing or water washing.
MHI SCR PROCESSES (13)
Clean Gas Treatment
Since 1973 Mitsubishi Heavy Industries (MHI) has tested
SCR in a laboratory and at a pilot plant with a capacity of
treating 100,000 Nm /hr of flue gas from an LNG-fired boiler
at Minamiyokohama Station, Tokyo Electric.
Granular base-metal catalysts with an alumina carrier
(2 to 4 mm in diameter) have been used. Figure 20 shows
results of laboratory tests with various catalysts at
temperatures between 150° and 500°C and their effect on the
denitrification ratio and NH., concentration in the outlet
gas. Catalysts made of Cr20_ and Pt were reactive at low
temperatures (200° to 220°C) but showed a tendency to con-
47
-------
1000
100
Q.
ft
CO
10
10 100
S03, ppm
1000
Figure 19. Formation temperature of NH HSO.
(NH-, + SO., + H00
3 32
Gas Gas Gas
NH4HS04)
Liquid
48
-------
vert a portion of NH_ into NO at higher temperatures.
Catalysts made of Cr.,0 with other_pxide_s _y_iBlded good _:„--
•^ J
results.
Figure 21 shows results of a 5500-hour life test of a
selected catalyst. The effect of 0_ concentration on
denitrification is presented in Figure 22. These results
indicate that when more than about 1 percent 0~ is present
in the gas, 90 percent denitrification is achieved at
320° to 450°C with a space velocity of 10,000 hr"1 and an
NH-/NO mole ratio of 1. Unreacted ammonia in the outlet
•J A.
gas is kept at 10 ppm or below. The catalyst has shown no
tendency to decrease in activity during the 5500-hour
continuous test.
At one pilot plant, flue gas at 270° to 370°C could be
obtained from a boiler either before or after the economizer
(Figure 23). The NO concentration fluctuated between 50
X
and 130 ppm with the boiler load. Figure 24 shows the
effect of boiler load on NO removal ratio and NH_ con-
X j
centration of the outlet gas. When the load was reduced,
the temperature fell. The NO removal rate did not decrease,
jt
however, because the space velocity decreased and O_ con-
centration increased.
At the pilot plant, tests were made to reduc'e ammonia
emissions by using an additional converter with an ammonia
decomposition catalyst. The effect of the converter is
49
-------
TOO
80
ox 60
200 300 400
REACTION TEMPERATURE, °C
A: Cr203-Al203
B: Pt-Al203
C:
D: Fe203-Al203
E: Fe203-Cr203
F: V0-Cr0-
20
10
-
H- CL
CO
n ^
0 O
Figure 20. Criteria for catalyst for clean gas
x-
O I
1.2
^o 1.0
z^ 0.9
. 100
90
40
20
o o
z z:
LU
C£
CO
3:
UJ Q.
—1 O.
NO 140-170 ppm
/\
SV 12,000 hr " 340-360°C
J I
I I i i i i
1000 2000 3000
REACTION TIME, hr
4000
5000
Figure 21. Durability of catalyst for clean gas.
50
-------
100
80
_T60
o
140
X
o
z
20
250 300 350 400 450
CATALYST TEMPERATURE, °C
Figure 22. Effect of oxygen on NO removal.
(SV, 10,000 hr"1; NO , 100 ppm; NH /NO , 1.0)
X J X
51
-------
ECONO-
BOILER MIZER
NH
3 REACTOR
k=X
MIXER
"^T
if-*
r
AJR
AIR
HEATER
TO
STACK
BLOWER
NH3 CONVERTER
Figure 23. Flowsheet of pilot plant (clean gas).
52
-------
** 100
UJ
on
90
so
270
4000
8
i
INLET N0¥
n
50-130 ppm
300
i
20
10
0
CO
—1 Q.
330 350 (°C) o
6000
9000 12000 (sv, hr"1)
2 (0, %)
1/4 2/4 3/4 4/4
BOILER LOAD
Figure 24. Effects of boiler load on NO removal.
53
-------
shown in Figure 25. Ammonia in the treated gas was reduced
to 1 or 2 ppm, and the denitrification efficiency was
increased to about 97 percent.
MHI recently obtained an order from Kyushu Electric for
construction of large commercial SCR plants for two LNG
boilers (600 MW each) to be completed in 1978.
Dirty-Gas Treatment
MHI has made tests with dirty flue gas from low-sulfur
oil burning. Examples of the gas composition are shown in
Table 11.
Table 11. EXAMPLES OF GAS COMPOSITION
Full load
25% load
?2'
1.0
3.0
NOX,
ppm
160
80
S02,
ppm
80
60
S03,
ppm
3
2
Dust,
mg/Nm^
20
15
Screening tests were made on several catalysts. The
S03 resistivity of the catalyst carrier was as follows:
Ti02 = Si02 > a A1203 > n A12O > y A12O
Metals used for the catalyst were Cu, V, Cr, Mn, Fe, Co, and
Ni. These were tested alone and in combination. Selected
catalysts have been tested further.
Bench-scale tests with catalysts in a fixed bed have
shown that a spherical catalyst 3 millimeters in diameter
causes serious dust plugging whereas a catalyst 8 millimeters
54
-------
100
90
REACTOR OUTLET
^CONVERTER OUTLET
200 250 300 350
CATALYST TEMPERATURE, °C
20 Q.
t— CL
I-i I
10
CO
400
Figure 25. Effects of ammonia converter.
(Gas flow, 12,000 Nm3/hr; NH /NO , 1.0; NO , 150 ppm)
J 2*. ji
55
-------
in diameter does not (Figure 26). The larger catalyst,
however, gives poor denitrification efficiency (Figure 27).
Therefore, a catalyst 4 to 6 millimeters in diameter has
been used in later tests.
Tests have been carried out at a pilot plant with
capacity for treating 4000 Nm /hr of flue gas from the
burning of low-sulfur oil. The pilot plant includes two
units, one with a fixed-bed reactor and the other with a
moving-bed reactor (Figure 28). Both have an air heater
after the reactor to allow tests of formation of ammonium
bisulfate.
A hot electrostatic precipitator (ESP) was installed
and used for most of the tests. The shape of the reactor is
shown in Figure 29. The catalyst is placed in a W-shaped
container developed by MHI. The container allows a uniform
gas flow at a low pressure drop, and the catalyst can be
replaced easily. The catalyst layer is 100 to 200 milli-
meters thick.
The relationship of gas temperature and space velocity
to denitrification ratio when the fixed bed was used with
the ESP is shown in Figure 30. About 90 percent removal was
obtained at 360°C, with a space velocity of 10,000 hr"1, and
with an NH_/NO mole ratio of 1. The NH_ content of the
-j x 3
reactor effluent was below 15 ppm (Figure 31). Results of a
56
-------
150
3 mm DIAMETER
8 mm DIAMETER
600
800
200 400
TEST PERIOD, hr
Figure 26. Pressure drop in fired-bed reactor
with different catalyst diameters.
(Dust 10-20 mg/Nm3)
100
90
80
70
2468
CATALYST DIAMETER, mm
10
Figure 21. Effect of catalyst size on NO removal
X
(SV, 10,000 hr"1, Temp, 360-370°C, NH../NO , 1.0)
•J J*.
57
-------
BOILER
NX
r
AIR HEATER
FIRST TEST
ESP
AIR HEATER
-L
ESP
\7
L
NH,
BLOWER
HEATER
X
REACTOR
SECOND TEST
LJUNGSTROM
AIR HEATER
CATALYST
CIRCULATION
SYSTEM
Figure 28. Systems of pilot plant tests.
58
-------
CATALYST
GAS
Figure 29. Structure of moving-bed reactor,
100
90
X
S 80
-SV 4000 (hr'1)
SV 8000
'SV 10)000
SV 12,000
300 320 340 360 380 400
TEMPERATURE, °C
Figure 30. Results of first test.
(Fixed bed; NH /NO = 1.0)
•J J\.
59
-------
catalyst life test are shown in Figure 32. The activity of
the catalyst did not decrease in 1500 hours of testing.
The particle size distribution of the dust is shown in
Figure 33. Before passing through the ESP about 5 percent
of the dust was larger than 1 micron and about 20 percent
was larger than 0.2 micron. After the precipitator, 2 to 3
percent of the dust particles were larger than 1 micron and
about 10 percent were larger than 0.2 micron.
Figure 34 shows results of a test using a moving bed
(with intermittent moving) without an ESP. Pressure drop
increased at a rate of 7.3 to 13.4 millimeters per day-
When it reached 160 millimeters H~O, the bed was moved to
replace 10 to 20 percent of the catalyst. The pressure drop
then was reduced to 60 millimeters HO. This test indicates
that by use of a moving bed an ESP may be eliminated. When
the gas contains much more dust, some dust-removal equip-
ment, such as an ESP or multicyclone, may be needed even
with a moving bed.
Figure 35 depicts measurements of deposit of ammonium
bisulfate on the air heater ,.(heat exchanger) . The maximum
deposit occurred at 200°C, in agreement with Figure 19. The
deposits must be removed by some means.
60
-------
100
90
80
20
-I
LU Q.
10
^J fl
o :c
4,0006,0008,00010,000
SV VALUE, hr"1
Figure 31. Effect of SV on NO removal and NH, emission
JC j
(Catalyst diameter, 4-6 mm; inlet NO , 150 ppir;
Temperature, 380°C; NH?/NO = 1*0).
400 800 1200
TEST PERIOD, hr
1600
Figure 32. Durability of NO reduction catalyst
-------
M 1
UJ
oz 5
£p 10
°-2 20
2£ 30
*-i co
o co
j I
o ESP INLET
A ESP OUTLET
a CONVERTOR
OUTLET
J_LJ_U J 1 1 I I
0.1 0.3 1.0
SHE, ym
3.0
Figure 33. Particle size distribution of dust,
62
-------
MOVED
JL
10.9
(MM H20/DAY)
10 20
TEST PERIOD, day
Figure 34. Change of pressure drop with moving bed.
Figure 35. Deposit of NI^HSO. in air heater,
63
-------
HITACHI SHIPBUILDING SCR PROCESS
Yokkaichi Plant, Shindaikyowa Oil
Hitachi Shipbuilding, after extensive pilot plant
tests, constructed the first commercial SCR unit in the
world for treatment of dirty flue gas in conjunction with
flue gas desulfurization. The unit was constructed at the
Yokkaichi plant, Shindaikyowa Oil, with capacity of treating
440,000 Nm /hr of flue gas from an oil-fired boiler (140 MW
equivalent). The flue gas is first treated by a Wellman-
Lord process plant constructed by Mitsubishi Kakoki Kaisha
(MKK) to reduce SO2 from 1500 to 100 ppm and dust from 140
to 40 mg/Nm . The by-product of the flue gas desulfuriza-
tion plant is sulfuric acid, which is consumed at the
Yokkaichi plant.
A flowsheet of the denitrification system is shown in
Figure 36. At about 55°C, gas from the scrubber contains
•3
150 ppm NO , 150 ppm SO2, and 40 mg/Nm dust. It is heated
in Ljungstrom-type heat exchangers to 310° to 320°C and then
in an auxiliary heater to 420°C. The gas is then injected
with ammonia and introduced into a reactor specially designed
to minimize the effect of dust. Space velocity in the
catalys'; is 5,000 to 10,000 hr"1, and the NH /NO ratio,
( , \. -3 x
about 1.2. The NO concentration after the reactor is 30 to
.X
40 ppm (75 to 80 percent removal); concentration after the
64
-------
BOILER
DUST
COLLECTOR
15Q°C
55°-60°C
HEAT EXCHANGER
160°-165°C
_\r
310°-320°C
HEATER
420°-430°C
DESULFURIZER
420°-430°C
NH.
REACTOR
Figure 36. Flowsheet of Hitachi Shipbuilding process
(an example of dirty-gas treatment).
-------
heat exchanger is 40 to 50 ppm because of inlet gas leaking
into the outlet stream.
No ammonia decomposition catalyst is installed in the
system, but ammonia emissions are reported to be very low,
possibly because excess ammonia is decomposed at the rela-
tively high temperature in the reactor.
The plant started operation in January 1976 and has
since been in operation without serious problems. Tempera-
ture of the heat exchanger effluent gas is about 160°C.
Because ammonium bisulfate deposits in the heat exchanger,
soot blowers have been used several times a day to remove
bisulfate. The heat exchanger is coated with enamel to
prevent corrosion. The catalyst life is expected to be
about one year.
The capital cost was $20 million ($143/kW) for flue gas
desulfurization and $7 million ($50/kW) for denitrification.
In addition, the company spent $3 million for miscellaneous
items for both systems.
Other Plants
Hitachi Shipbuilding a] so constructed for Ideniitsu Kosan,
at its Chiba refinery- a unit with a capacity for treating
3
350,000 Nm /hr of flue gas. The flue gas from the refinery,
containing relatively small amounts of SO., and dust, is
subjected to denitrification without desulfurization. The
66
-------
denitrification system, similar to that of Shindaikyowa Oil,
has been in operation since November 1975.
Hitachi Shipbuilding is to construct a plant for
Kawasaki Steel at its Chiba Works to treat 840,000 Nm /hr of
flue gas from an iron-ore sintering plant. The gas contains
not only 300 to 500 ppm SO™ and a considerable amount of
dust, but also a small amount of an alkaline vapor that
contaminates the catalyst. The gas passing through an ESP
will be treated first by lime scrubbing, then by a wet ESP
to clean the gas, and finally by a denitrification system.
The Cement Producers Association has been operating a
pilot plant constructed by Hitachi Shipbuilding with a
capacity of treating 5000 Nm /hr of flue gas from a cement
kiln at the Nanyo Plant, Tokuyama Soda. The gas, with a low
SO_ concentration but high dust content, is passed through
an ESP and then treated by the denitrification unit. Dust
plugging has been a problem, and Hitachi Shipbuilding is to
change the reactor design and catalyst shape.
Hitachi Shipbuilding has been operating several small
test units (150 to 250 Nm /hr) for SCR with flue gas from a
coal-fired boiler at Isogo Station, Electric Power Develop-
ment Company. The gas from an economizer of the boiler at
about 400°C, containing about 15 g/Nm dust, 300 ppm SO_,
and 250 ppm NO , is treated without dust removal. Details
3C
of operation and efficiency have not been reported.
67
-------
Economics
Hitachi Shipbuilding reports that the capital cost
(battery limits) for a denitrification plant similar to that
of Shindaikyowa Oil is about $6.7 million. The 300,000
Nm /hr plant is equivalent to 100 MW, and the cost includes
the heat exchanger. The cost of removing about 85 percent
of the NO from a 150-ppm gas stream, is estimated at $15
n
per kiloliter of oil (3.3 mills/kWh), based on the following
assumptions:
Annualization of fixed cost: 21% of total capital cost
8000 annual operating hours
Power: 3.3£/kWh
Fuel: $83/t
Steam: $7/t
Ammonia: $233/t
Catalyst: $7500/m3
1 year catalyst life
The cost would be higher for the Kawasaki Steel plant,
which will have a wet electrostatic precipitator in addition
to the above system. Kawasaki Steel is required to con-
struct the plant under an agreement with the Chiba prefec-
tural government. |
/
HITACHI, LTD., SCR PROCESS
Hitachi, Limited, has tested SCR at several'pilot
plants with capacities to treat 200 to 4000 Nm /hr of flue
68
-------
gas from oil- and LNG-fired boilers and a coke oven. Various
types of reactors with fixed and moving beds have been
tested. Hitachi applied for a patent on base-metal cata-
lysts that are deposited on alumina on the same day that
Exxon applied for a similar patent.
Some results of the pilot plant tests are shown in
Figures 37 and 38. More than 90 percent of the NO is
reacted with ammonia at 380° to 400°C, and more than 80
percent is reacted at about 300°C, using an NH_/NO mole
~5 5C
ratio of 0.9 to 1.0 and a space velocity of 10,000 to
20,000 hr . The ammonia concentration in the reactor
effluent gas is about 2 ppm at a space velocity of 10,000
hr and about 5 ppm at a space velocity of 20,000 hr
In a joint venture with Mitsubishi Petrochemical,
Hitachi developed a catalyst based on a titanate, which is
more resistant to SO than alumina, and constructed a com-
X
mercial plant (150,000 Nm3/hr) (Table 8).
Hitachi is now constructing a larger commercial plant
with a capacity to treat 500,000 Nm /hr of flue gas from
coke ovens. Two reactors are used in parallel at the Chiba
plant, Kawatetsu Chemical, a subsidiary company of Kawasaki
Steel (Figure 39). Recently Hitachi received an order from
Chubu Electric Power to construct SCR plants for two 700-MW
LNG-fired boilers to be constructed at the Chiba Station by
the end of 1977.
69
-------
MOLE
B&TQ
0.9 ~ 0.86
0.92 - 0.91
0.95
-1.0
0.87
0.88
1.0
SHUTDOWN
OF PLANT
o.
LU
400 r
30060
200
300
OC M C£
OH- =>0
I— Z OO CM
O LU to ar
So: LU
LU ae.
ac. u. Q.
onn
':uu
10°
Lfb
lOOr
LU
LU i— i
OH O ls«
0 U-
Z LU
80
70
oo
i
2000 2100 2200 2300 2400 2500 2600
PERIOD OF OPERATION, hr
2700
2800
2900
3000
Figure 37. Hitachi SCR pilot plant test (LNG) at a space
_-i
velocity of 20,000 hr
-------
TOO
60
X
o
20
REACTION TEMPERATURE 380°C
15,000
20,000
2000 4000
PERIOD OF OPERATION, hr
6000
Figure 38. Hitachi SCR pilot plant tests after
desulfurization (heavy oil).
-------
JAPAN GASOLINE PARANOX PROCESS
Japan Gasoline Company, an engineering and plant
construction firm that earlier constructed a Shell-process
flue gas desulfurization unit at Yokkaichi for Showa Yokkaichi
Sekiyu Co. (SYS), has developed its own SCR catalyst and has
constructed two commercial units (Table 8) using the Shell-
type parallel-passage reactor, which is less vulnerable to
dust than other designs. The structure of the reactor and a
flowsheet of the process are shown in Figures 40 and 41.
The first commercial plant was completed in November
1975 at the Kashima Refinery of Kashima Oil. The plant has
a capacity to treat 50,000 Nm /hr of flue gas from the
refinery (containing 60 to 250 ppm NO and about 50 mg/Nm
X
dust) and to remove more than 95 percent of the NO . Flue
X
gas entering at 200° to 350°C is heated via a heat exchanger
and an auxiliary heater to 390° to 420°C, injected with
ammonia at an NH7/NO ratio of 1.1/1.3, and introduced into
*3 X
the reactor. Figures 42 through 44 show the denitrification
efficiency, NO concentration at the reactor inlet and
X
outlet, and pressure drop ii, the reactor.
Very high removal efficiency is obtained with a rela-
tively high pressure drop and a relatively low space veloc-
ity. The ammonia concentration of the reactor effluent is
less than 10 ppm. There has been no problem of dust plugging
72
-------
FRONT
BURNER
COKE OVEN FURNACES
O—*
o
BOOSTER
FANS
REACTORS
STACK
NH.
Figure 39. Flowsheet of Chiba plant, Kawatetsu Chemical,
TREATED FLUE GAS
\\//
I'
INCOMING FLUE GAS
Figure 40. Structure of parallel-passage reactor.
73
-------
STACK
REACTOR
FLUE
/E
1
1
IN-LINE
FAN HFATFR
J
X
FLUE
HOT AIR AND
COMBUSTION /
AMMONIA
r-€>
( )
CONOENSAT
&
AIR HEATER NH3 COMPRESSOR NH3 TANK
AIR BLOWER
Figure 41. Flowsheet of Paranox process.
74
-------
^100
z 98
o
3 96
o
o
94
Figure 42.
500 1000 1500
PERIOD OF OPERATION, hr
2000
NO removal ratio (Kashima plant).
X
LU C£.
M °
^j C_5
O <
_x £
gio.o
8.0H
6.0
4.0
2.0
REACTOR INLET
-REACTOR OUTLET
500 ' ' 'lOOO Woo ' ' 200)
PERIOD OF OPERATION, hr
Figure 43.
NO concentration (Kashima plant).
A.
o
CM
g
n
Q.
O
o:
o
LU
00
00
UJ
Of.
a.
220
200
180
160
140
120
100
- ' ' • i • • • • i • • • i • • • • L.
!*C — ^_
• ... i .... i .,, ^ i .... L
500 1000 1500 200<
PERIOD OF OPERATION, hr
Figure 44. Pressure drop in reactor (Kashima plant)
75
-------
during about 6 months of continuous operation since start-
up of the plant. The catalyst life may be longer than the
one year that was initially assumed. When less than 90
percent removal is sufficient, a higher space velocity may
be used at a lower pressure drop. A recent estimate by
Japan Gasoline puts the capital cost at $1.33 million for a
150,000 Nm /hr system giving 85 to 90 percent removal.
KURABO SCR PROCESS (14)
Kurabo has been testing an SCR process at a pilot plant
with a capacity of treating 30,000 Nm /hr of flue gas from
a boiler burning a high-sulfur oil. The plant uses moving
beds to treat gas containing a high concentration of S02 and
a considerable amount of dust. A flowsheet of the process
is shown in Figure 45.
The reactor has three elements (Figure 46), each with
capacity of treating 10,000 Nm /hr of gas. In each element
the spherical catalyst, 5 millimeters in diameter, moves
continuously downward. The gas passes horizontally through
the elements. The catalyst bed is thin, and the gas veloc-
ity through the catalyst is relatively low to maintain the
pressure drop in the reactor below 100 millimeters of H_O.
£*
A space velocity of 7000 to 10,000 hr is normally
used. The flue gas contains about 1600 ppm S0_, 280 ppm
NO , and nearly 100 mg/Nm dust. More than 90 percent of
ji
the NO and more than 80 percent of the dust are removed in
76
-------
AIR
PREHEATER
AMMONIA
GENERATOR
TO DCSUL-
.FURJZER
AIR
KNORCA DENITRIFICATION UNIT
Figure 45. Flowsheet of Kurabo SCR process.
7.7
-------
ELEMENT
CASING
Figure 46. Structure of Kurabo moving-bed reactor.
78
-------
the reactor at 350° to 400°C with an NH0/NO mole ratio of
3' x
about 1.0. Following are typical operating data:
NO concentration, ppm
X
Dust content, mg/Nm
Gas volume, Nm /hr
Space velocity, hr~
S02 concentration, ppm
Gas temperature, °C
Pressure drop in reactor,
mm H_0
Ammonia consumption,
liters/min
280 (inlet)
94 (inlet)
24,000
8000
1650
400
65
106.4
18 (outlet)
17 (outlet)
The catalyst discharged from the reactor is screened to
remove dust, heated to 800°C to recover the catalyst, and
then returned to the reactor. One cycle takes about 100
hours, an indication that only a small portion of the
catalyst always undergoes regeneration. The catalyst has a
high strength of more than 10 kilograms per granule, and the
annual crushing loss is less than 5 percent. The catalyst
life is estimated at about 10,000 hours. The gas from the
reactor is passed through a tubular heat exchanger. The
outlet gas temperature is maintained at 200°C to prevent the
deposit of ammonium bisulfate.
For a utility boiler that undergoes considerable load
fluctuation, an auxiliary burner may be needed to control
79
-------
the gas temperature in the reactor. Because of high SO*
and NH_ concentrations in the reactor, the temperature must
be kept above 350°C to prevent amminium bisulfate deposits.
Estimated denitrification costs are shown in Tables
12 and 13. The estimate is based on the assumption that
the gas from the uncontrolled boiler would be cooled from
360°C to 200°C in an air heater and that with denitrifica-
tion the gas would be heated by an auxiliary burner to
400°C and emitted at 220°C from the air heater.
Table 12. DENITRIFICATION PLANT COST
(thousands of dollars)
Item
Reactor, heater, etc.
Duct
Instrumentation
Transportation and
installation
Subtotal
Catalyst
Total
Capacity, Nm /hr
30,000
330
83
93
44
550
40
590
100,000
887
197
203
113
1400
130
1530
500,000
3347
543
540
310
4740
640
5380
The estimate shows that the denitrification cost is
about $11 per kiloliter (2.3 mills/kWh) for a 100,000 Nm3/hr
unit and $8.53 per kiloliter (1.8 mills/kWh) for a 500,000
Nm /hr unit, assuming 8400 hours operation in a year.
80
-------
Table 13. DENITRIFICATION COST (KURABO PROCESS)
C grade heavy oil: 0.78 kl/10,000 Nm , NOX 280 ppm, 8400 hours annual operation;
Average yearly load: 80%; Gas temperature: 360°C before air heater, 400°C in reactor,
220°C after air heater
Depreciation
Interest
Ammonia
Power (3C/kW-hr)
Catalyst ($8300/m3)
Steam (0.3C/kg)
Labor
Maintenance, etc.
Fuel*
Total
Capacity, 1000 Nm /hr
30
$1000/year
75.9
29.5
21.2
24.4
25.0
0.7
6.0
3.3
21.7
207.8
$/kl
4.82
1.88
1.35
1.55
1.59
0.05
0.38
0.21
1.38
13.21
100
$1000/year
196.7
76.5
61.8
71.3
83.3
2.4
6.0
8.3
72.5
578.9
$/kl
3.75
1.46
1.18
1.36
1.59
0.05
0.11
0.16
1.38
11.04
500
$1000/year
692.1
269.2
154.4
294.6
416.7
14.6
9.0
23.3
326.5
2236.4
$/kl
2.64
1.03
0.59
1.12
1.59
0.06
0.03
0.09
1.38
8.53
00
* For auxiliary burner. Half of this energy is recovered.
-------
Emission of the gas at 220°C causes a fairly large loss of
energy. Use of the Ljungstrom-type heat exchanger will
allow more energy recovery, although soot blowing is required.
Kurabo has been developing a microcomputer system to mini-
mize emissions of ammonia from the reactor without decreas-
ing the denitrification efficiency as boiler load fluctuates.
It would then be possible to attain better heat recovery
with a heat exchanger and thus prevent the deposit of the
bisulfate. In treating flue gas from a coal-fired boiler, a
multicyclone may be used to reduce the dust content of the
gas to a degree compatible with the moving-bed reactor.
SCR PROCESSES WITH SANTETSU (SARC) CATALYST
Santetsu Iron Catalyst (SARC)
Santetsu Kogyo, a chemical company, has developed an
effective catalyst made of solid goethite (Fe2O -H_0). The
ferric oxide is recovered from waste liquor containing
ferric iron by neutralization with ammonia. The price of
the catalyst, now about $8000 per ton, may be substantially
reduced when the catalyst is produced in large quantities.
The catalyst is effective above 300°C and is resistant
to SO . It is easily shaped into rings with a diameter of
J\
13 to 35 millimeters and with a smooth surface to minimize
the effect of dust. The catalyst is used by the companies
discussed in the following paragraphs.
82
-------
Mitsubishi Kakoki Kaisha (MKK) has operated two pilot
plants to treat boiler flue gas and is constructing (1) a
pilot plant (1000 Nm /hr) to treat waste gas from iron-ore
sintering, using funds of the Japan Iron and Steel Federation,
and (2) a commercial plant (15,000 Nm /hr) to treat flue gas
t »
from an oil-fired boiler (Table 8).
Seitetsu Kagaku Kogyo, a chemical company, has operated
successfully since July 1975 a pilot plant with capacity of
treating 15,000 Nm per hour of flue gas from an oil-fired
boiler.
Other companies including Ishikawajima-Harima Heavy
Industries (IHI) also plan to build pilot plants using the
SARC catalyst.
Tests by MKK with SARC Catalyst
MKK has operated a pilot plant with capacity of treating
3
1400 Nm per hour of flue gas from low-sulfur oil burning.
They have used the following three sizes of ring-type SARC
catalyst produced by Santetsu Kogyo:
No.
II
III
IV
Diameter , mm
Inner
20
13
8
Outer
35
23
13
Height or
thickness, mm
15
13
13
83
-------
The flue gas at 330° to 450°C contains 170 to 400 ppm
NO , 30 to 50 ppm SO , 7 to 8 percent CO9, 6 to 7 percent
xx ^
02, 8 to 10 percent H20, and about 40 mg/Nm dust. The gas
is introduced into the reactor without dust removal. The
reactor has a diameter of 635 millimeters, and the packed
height ranges from 600 to 1200 millimeters. In test opera-
tion, catalyst II was packed regularly, and III and IV were
packed at random. Gas volume was variable, from 800 to 1400
Nm /hr. Test results are shown in Figures 47, 48, and 49.
With catalyst IV more than 90 percent of NO was re-
X
moved at a space velocity of 5700 hr~ at 420°C and an
NH-./NO ratio of 1.2. The randomly packed III and the
•j X
regularly packed II gave nearly equal removal ratios —
about 90 percent at a space velocity of 3500 hr~ at 420°C
and an NH-/NO mole ratio of 1.2. Although catalyst III is
•3 J^
smaller than catalyst II, the packed density and pressure
drop were lower with III because of the random packing.
The pressure drops with a 1-meter packed height at a
gas velocity of 2 meters per second were 80 millimeters H2O
for III, 135 millimeters H2C for II, and 150 millimeters HO
for IV (Figure 50). The dust content of the gas is about
1000 milligrams per cubic meter normally and much higher
when soot is blown from the boiler three times a day. But
no effect of dust was observed in continuous tests for 1000
84
-------
TOO
90
80
70
SV (hr'1)
0.8 1.0 1.2 1.4
NH./NO. MOLE RATIO
*3 A
1.6
Figure 47. Results with SARC II catalyst (420°C)
85
-------
100
90
I
80
70
SV (hr"1)
0.8
I I I I
1.0 1.2 1.4
NH3/NOX, MOLE RATIO
1.6
Figure 48. Results with SARC III catalyst (420°C)
86
-------
100
90
80
1.0
SV 2800 (hr"1)
4000
j I
1.2 1.4
NH3/NOX, MOLE RATIO
1.6
Figure 49. Results with SARC IV catalyst (420°C)
87
-------
1000
500
•o
£
o
Q.
E
O
CM
:r
ex.
O
QC
Q
oe
=3
to
UJ
G_
300
100
SO-
10
1 235
GAS VELOCITY, m/sec
10
Figure 50. Gas velocity and pressure drop
with SARC catalysts.
88
-------
hours with catalyst II, 1700 hours with III, and 1000 hours
with IV. In the event of dust plugging, the catalysts can
be washed with water to remove dust. This has not yet been
necessary.
The slip ammonia (NH_ concentration of the reactor
effluent gas) when reacted at 420°C was less than 0.2 ppm at
an NH-/NO ratio of 1.2 and about 1 ppm at a ratio of 1.4 to
•j X
1.6; the SARC catalyst is capable of decomposing the exces-
sive ammonia.
The relationship of reaction temperature and space
velocity to denitrification ratio is shown in Figure 51.
More than 80 percent removal was obtained at a temperature
of 330°C and a space velocity of 4000 hr .
OTHER SCR PROCESSES
Sumitomo Chemical has made extensive tests on SCR and
has constructed several commercial plants, including the
world's first plant for combustion gas treatment, completed
in 1974 (Table 8). Details of the process and the chemistry
involved were described earlier (1). Recently commercial
plants for dirty-gas treatment have started operation, but
no details of operation have been disclosed.
Mitsui Toatsu Chemical also developed its own catalysts
and constructed a few units (Table 8). They plan to cooper-
ate with IHI in building larger plants.
89
-------
100
90
£ 80
70
60
420°C
370°C
350°C
330°C
2000 3000 4000
SPACE VELOCITY, hr
5000 6000
-1
7000
Figure 51 . Results with SARC IV catalyst
(NH-/NO mole ratio, 1.2)
•j X
90
-------
Mitsui Shipbuilding is constructing a few units (Table
8) using a catalyst developed by Mitsui Petrochemical.
Plant cost is estimated at about $5 million for a 67-mega-
watt boiler, with guarantees of 90 percent NO removal and
j\.
1 year catalyst life for treatment of gas containing less
than 100 ppm SO .
Asahi Glass Company has also developed its own catalyst
and is constructing a unit to treat flue gas from a glass-
melting furnace (Table 8).
AMMONIA INJECTION WITHOUT CATALYST
Nippon Kokan applied for a patent in 1970 for a deni-
trification process in which ammonia is injected into waste
gases at temperatures above 500°C and some refractory struc-
ture is used instead of a catalyst to promote the mixing of
ammonia and the gas. A few years later Exxon also made
patent application for ammonia injection without a catalyst.
Both patents are pending.
Laboratory tests have shown that when ammonia is
injected at 980° to 1000°C about 80 percent of the NO is
converted to N2 and that the residual ammonia in the treated
gas is below 20 ppm. In a large-scale operation about 50
percent of the NO removal is expected at an NH-/NO mole
H J
ratio of 1.5 to 2.0 because the suitable temperature range
is narrow and very rapid mixing of NH_ and gas is required.
91
-------
Fluctuation of exhaust gas temperature with load in a
utility boiler may present a problem. When the temperature
drops, unreacted ammonia may be emitted. When it rises
excessively, a portion of the ammonia may be converted to
NO.
Exxon has found that addition of hydrogen with ammonia
can reduce the reaction temperature to 730°C (15). The use
of hydrogen at power plants may be difficult, however.
The ammonia injection process would be suited for
industrial boilers that are not subject to large load fluc-
tuations. It also is suitable for treating flue gases, such
as those from coal-fired boilers and cement kilns, which
contain dust that contaminates t^he SCR catalyst. There is a
!
possibility that ammonium bisulfate is deposited in the air
heater, as in the SCR processes.
Electric Power Development Company (EPDC), which owns
several coal-fired boilers, is testing the Exxon process for
possible commercial use in 1978 (Table 9). Many other
companies also have been testing ammonia injection pro-
cesses, because injection without a catalyst may be the most
economical method of flue gas denitrification, even though
removal efficiency may not be high.
92
-------
REACTION OF ACTIVATED CARBON WITH NOV (16)
A.
Classification of Reactions
Activated carbon has the following functions useful in
removing NO :
(1) Adsorption It adsorbs NO_ below 100 °C.
(2) Oxidation It promotes the oxidation of NO to
N02 below 100 °C.
(3) Catalytic Above 100°C, it promotes the
reduction following reactions:
6NO +
4NO +
2NO + 2CO -> N2 + 2CO2
2NO + 2H2 -*• N2 + 2H20
2NO + C -> N2 + C02
NO Adsorption
Below 100 °C activated carbon adsorbs N02 fairly well,
but does not adsorb NO very well. In the presence of 02/
however, NO is oxidized to NO_ by the catalytic reaction of
carbon, and the resulting N02 is adsorbed. The adsorption
efficiency of carbon is shown in Figure 52 in comparison
with that of silica gel. The amounts of NO adsorbed by
X
carbon under different conditions are shown in Table 14. A
larger amount is adsorbed at lower temperatures and humid-
ities.
93
-------
100
a*
A
<
i
LU
an
80
60
40
ACTIVATED CARBON
10
REACTION TIME, hr
20
Figure 52. NO adsorption capacity
(Temp., 18°C; NO, 1800 ppm; NO , 1200 ppm; SV, 1000 hr"1)
94
-------
Table 14. N0x ADSORPTION CAPACITY OF ACTIVATED CARBON
(rag NO /g activated carbon)
X
Relative humidity, %
14
20
35
40
60
75
Adsorption temperature
30°C
120.0
104.0
67.7
50°C
110.2
95.0
75.4
70°C
56.8
36.8
NO^ Desorption
X —i-ij..._.
NO adsorbed by carbon is desorbed either by washing
X
with water or by heating. Water washing produces a dilute
nitric acid. Higher temperatures are more favorable for the
washing, as shown in Figure 53. The wet carbon after washing
has poor NO adsorption ability and should be dried before
X
it is used again for adsorption. For SO_ adsorption,
however, the wet carbon is effective.
Figure 54 shows the desorption rates when carbon is
heated in a nitrogen gas stream at different temperatures.
Above 150°C desorption occurs fairly rapidly, but carbon is
consumed in reaction (8) to form NO. When the gas is
heated above 450°C, a larger amount of carbon is consumed in
reaction (9) to form N_.
95
-------
1.0
x 0.8
« 0.6
g
5 0.4
QC
fe
o
I—I
s
0.2
0
70°C
20 40
WASHING TIME, min
60
Figure 53. Desorption of NO adsorbed by activated carbon
jC
by washing with water at different temperatures
(activated carbon 6 mm in diameter).
1.0
0.8
x
o
z 0.6
is
i—*
« 0.4
o
2 0.2
0 10 20 '30 40
HEATING TIME, min
Figure 54. Desorption of NO adsorbed by activated carbon
Ji
by heating at different temperatures.
96
-------
2NO, + C -> 2ND + CO- (8)
+ 2C -> N2 + 2C02 (9)
Heating carbon in a stream of reducing gas lowers the
carbon consumption substantially and increases the conver-
sion of NO to N9 (Table 15).
JC £•
Catalysts for Ammonia Reduction
Among the reducing gases mentioned previously, ammonia
is most useful in flue gas treatment because other reducing
gases are readily consumed by O_ in the flue gas. The
effect of carbon on the reaction between NO and NH_ is
X J
further increased by adding base metal compounds (Table 16) .
Copper and vanadium have been found most effective.
SIMULTANEOUS REMOVAL PROCESSES USING ACTIVATED CARBON
Simultaneous Removal of SO and NO
~ "• " ...... "' " " ' "~"r T " X X
Activated carbon has been used commercially as an
adsorbent for S00. Although it also adsorbs NO , the
£• X
adsorbing capacity is not sufficient to treat a large
amount of gas. Takeda Chemical has produced activated
carbon adsorbents that contain metallic components or that
are specially structured to promote the reaction of NO with
X
ammonia to form N_. Higher temperature is favorable to the
reaction but decreases the S02 adsorbing capacity (Figure 55)
Optimum temperature for simultaneous removal by this process
is about 250°C.
97
-------
Table 15
DESORPTION OF NOV IN REDUCING GAS
X
Gas
H2
CO
•
NH3
He
(Inert gas)
Temperature,
°C
200
400
600
200
400
600
200
400
600
200
400
600
Conversion ratio of
adsorbed NOX into
N_ on desorption, %
10
68
100
12
70
100
100
100
100
0
5
70
98
-------
Table 16. EFFECT OF ADDITION OF BASE METAL COMPOUNDS TO
CARBON ON NOV REDUCTION EFFICIENCY (NO 2000 ppm,
X
NH3 3000 ppm, SV 3000 hr"1)
Metal
None
Ti
Cr
Mn
Fe
Co
Ni
Cu
V
Mo
W
NO reduction efficiency, %
J\.
110°C
38
55
50
52
63
91
80
150°C
44
65
70
67
67
75
67
99
88
70
65
250°C
78
95
88
90
98
100
100
99
-------
100
90
80
70
60
100
I
150 200 250
TEMPERATURE, °C
300
350
Figure 55. SO0 and NO removal by activated
^ .Ok
carbon at different temperatures and space velocities
100
-------
Carbon for simultaneous removal of S0_ and NO costs
£ X
about $8000 per ton, whereas carbon used commercially for
flue gas desulfurization costs $3000 per ton. Simultaneous
removal from flue gas leaving a 300-MW boiler will require
about 1000 tons of the carbon, which may be too expensive.
The price would be substantially lower in mass production.
Unitika Activated Carbon Process
Unitika Company recently started operating a pilot
plant with a capacity to treat 4500 Nm /hr of flue gas from
a glass-melting furnace, containing about 400 ppm SO- and
500 ppm NO (Figure 56). The plant has a tower with four
X
compartments, all of which have a fixed carbon bed. About
600 ppm NH- is added to the gas at about 230°C, and the
mixed gas is ducted to three compartments. About 98 percent
of the NO and S0_ is removed. The carbon that has adsorbed
X ^
SO2 is heated to 350°C in a reducing hot gas to release
concentrated S02 for sulfuric acid production. Ammonium
sulfate and sulfite, which tend to form on the carbon, are
decomposed to S02 and N_ in the regeneration step.
101
-------
o
to
STACK
NH
3 r
FLUE
GAS
rO
-L>
"-O
-Q*
^S02
TO H2S04 PLANT
FUEL
INERT GAS
PRODUCER
Figure 56. Flowsheet of Unitika process
-------
The principal design and operating parameters are as
follows:
Tower height
Carbon bed thickness
Pressure drop
Adsorption time for one cycle
Regeneration time for one cycle
SO_ in gas from regeneration step
17 m
1 m
100 mm H_O
3 days
12 hours
5 to 10 percent
A space velocity of about 700 hr is used. The carbon
consumption is estimated to be less than 10 percent of a
charge per year. In the 6 months of operation, carbon loss
has been only 1 to 2 percent. Gas from the regeneration
step, containing 5 to 10 percent S02, may be used for
sulfuric acid production.
Other Activated Carbon Processes
Sumitomo Heavy Industries has constructed a prototype
flue gas desulfurization plant (175,000 Nm /hr) using moving
beds of adsorbent activated carbon, which is regenerated by
heating in a reducing gas. With this plant, the company has
studied simultaneous removal and is constructing a test
plant with a capacity to treat 1500 Nm /hr of flue gas using
moving beds.
Hitachi, Limited, has found that activated carbon
treated with ammonium bromide is effective even at 100°C for
NO reduction by ammonia (17). The low-temperature activity
X
103
-------
may result in energy saving, but deposits of ammonium
sulfate and bisulfate on the carbon may present a problem.
EBARA-JAERI ELECTRON BEAM PROCESS (18)
Ebara Manufacturing Company and the Japan Atomic Energy
Research Institute, a government organization, have developed
jointly a unique process for the simultaneous removal of S02
and NO by electron beam radiation. A pilot plant with
X
capacity to treat 1000 Nm /hr of flue gas from an oil-fired
boiler has been operated by Ebara. A larger pilot plant
with capacity to treat 3000 Nm /hr of waste gas from an iron
ore sintering plant is scheduled for construction at Yawata
Works, Nippon Steel Corporation.
A flowsheet of a bench-scale test unit is shown in
Figure 57. Dimensions of the stainless steel reactor are 50
by 500 millimeters. The electron beam accelerator (Cockcroft-
Walton type) is made by Hitachi, Limited. Flue gas produced
by burning of oil and containing 600 to 900 ppm SO- and 80
ppm NO was passed through an ESP, introduced into the
X
reactor, and exposed to the electron beam. A sulfuric acid
mist and a powdery product containing sulfur, nitrogen,
oxygen and hydrogen were produced and treated by another
ESP.
Figure 58 shows the results of tests at 110°C. About
90 percent of the NO was removed by an electron beam of
104
-------
1: FUEL OIL 3:ELECTRON BEAM ACCELERATOR 5:DUST COLLECTOR
2:BURNER 4-.REACTOR 6: ANALYZER (SOg, NOX)
Figure 57. Apparatus for tests of electron beam process,
0
1 2 3
TOTAL BEAM, Mrad
• 4.31xl05 rad/sec O4.31xl05 rad/sec
A 8.61xT05 rad/sec O1.46xl05 rad/sec
Figure 58. Results with different intensities of
electron beam.
105
-------
about 0.8 Mrad (radiation for 2 seconds of the beam with an
5 >
intensity of 4.31 x 10 rad/sec), whereas about 80 percent
of the SO,, was removed at 4 Mrad (radiation of the beam for
10 seconds). Radiation at lower temperatures slightly
increased the removal efficiency.
3
The larger pilot plant (3000 Nm /hr), including equip-
ment for various tests and measurements, will cost more than
$1 million. Power consumption for electron beam accelera-
tion is estimated at 1 megawatt for 100,000 Nm /hr of flue
gas (about 32 MW equivalent). It may be possible to make an
electron beam accelerator as large as 1 MW. Consequently,
treatment of 1,000,000 Nm3/hr of flue gas from a 320-MW
boiler will require ten accelerators.
The main advantage of the process is the simultaneous
removal of NO and S00 by consuming electric power only.
X £
Power consumption is not high compared with that of other
processes, which may also require ammonia, lime, or some
other chemical. Investment cost, however, seems fairly high
because the process requires accelerators and a highly
efficient ESP. In a large-scale operation, the by-product
will have to be treated.
OTHER DRY PROCESSES
Shell Copper Oxide Process
Copper oxide, used as an absorbent of S0_ in the Shell
process, works as a catalyst in the reaction of NO with
X
106
-------
ammonia. The Yokkaichi plant of SYS, treating 120,000
Nm /hr of flue gas from an oil-fired boiler by the Shell
process, has introduced ammonia into a reactor at 400°C
since 1975. Up to about 70 percent of the NO can be re-
j£
moved. Copper sulfate formed by SO- absorption is reacted
with hydrogen to generate concentrated S02, which is sent to
a Glaus furnace to produce sulfur.
The plant uses two parallel-passage reactors alter-
nately for absorption and desorption-regeneration. The
reactor is fairly free of dust plugging. The flue gas at
400°C is treated without using a hot ESP ahead of the
reactor.
Dry Adsorption Process
Small-scale tests have been conducted in which NO was
X
adsorbed by materials other than activated carbon, including
molecular sieve, silica gel, and calcium silicate. Nissan
Chemical Industries, a large chemical company, is going to
use molecular sieves (to be supplied by Union Carbide,
U.S.A.), for the treatment of tail gas from a nitric acid
plant (20,000 Nm /hr). The unit is operated at a pressure
of 5 atmospheres and is suitable for adsorption. The ad-
sorbed NO will be released by heating and will be returned
X
to the acid plant.
107
-------
Generally speaking, adsorption processes are not
suitable for large amounts of gas, particularly those
containing SO- and dust.
Catalytic Decomposition
The decomposition of NO into N0 and 0- is an exother-
X £m Z,
mic reaction that can theoretically proceed even at room
temperature, but the reaction rate is quite small. Tests
have been made on catalysts that may promote the reaction at
relatively low temperature (150° to 200°C). No data have
been published. Some of the catalysts work fairly well at
700° to 800°C, but it seems that no effective low-tempera-
ture catalyst has been found.
108
-------
SECTION 4
WET PROCESSES FOR NO REMOVAL
X
GENERAL DESCRIPTION
Major Processes and Plants
Major plants using wet processes for NO removal from
H
flue gas are listed in Table 17. In addition there are many
small commercial units for treating waste gas from plants
using nitric acid. The major processes may be classified as
follows:
Oxidation-absorption
Absorption-oxidation
Oxidation-reduction (simultaneous removal of NO and
sox)
Reduction (simultaneous removal of NO and SO )
x x
Essentially all of the NO in the combustion gas is in
X
the form of NO, which has poor reactivity and is not readily
absorbed by most absorbents. NO is oxidized to NO, in air,
£*
but the oxidation occurs slowly. In many processes oxidizing
agents are used to promote absorption of NO .
J\
Oxidizing Agents
Ozone (0.,) and chlorine dioxide (C102) are used mainly
for the oxidation of NO in the gaseous phase. They oxidize
NO to NO- within a second but barely oxidize S09 to S0_.
109
-------
Table 17. MAJOR PLANTS FOR NOV REMOVAL
A
FROM FLUE GAS BY WET PROCESSES
Process developer
(Tokyo Electric
Mitsubishi H.I.)
(Tokyo Electric
Mitsubishi H.I.)
Kawasaki H.i.
Nissan Engineering
Nissan Engineering
(Mitsubishi Metal
MKK, Nihon Chem.)
Kobe steel
Kobe Steel
Hodogaya
(Sumitomo Metal
Pujikasui)
(Sumitomo Metal
Fujikasui)
(Sumitomo Metal
Fujikasui)
Osaka Soda
Shirogane
Chiyoda
Mitsubishi H.I.
Ishikawajima n.i.
Kureha Chemical
Chisso Corp.
Mitsui S.B.
AMhi Chemical
Type of
process
(Oxidation
absorption)
(Oxidation
absorption)
(Oxidation
absorption)
(Absorption
oxidation)
(Absorption
oxidation)
(Absorption
oxidation)
(Absorption
oxidation)
(Absorption
oxidation)
(Absorption
oxidation
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
(Oxidation
reduction)
Reduction
Reduction
Reduction
Reduction
Plant owner
Tokyo Electric
Tokyo Electric
EPDC
IJima Metal
Nissan Chemical
Mitsubishi Metal
Kobe Steel
Kobe Steel
Rodagaya
Sumitomo Metal
Toshin Steel
Sumitomo Metal
Osaka Soda
Mitsui Sugar
Chiyoda
Mitsubishi H.i.
Ishikawajima H.I.
Kureha Chem.
Chisso P.c.
Mitsui P.c.
Asahi chemical
Plant
site
Minami-
Yokohama
Minami-
Yokohama
Takehara
Tokyo
Toyama
Omiya
Kakogawa
Kahogawa
Koriyama
Amagaaaki
Fuji
Osaka
Amagaski
Kawasaki
Kawasaki
Hiroshima
Yokahama
Nishikl
Goi
Chiba
llzushima
Capacity,
Nm'/hr
2,000
100,000
5,000
1,800
3,000
4,000
1,000
50,000
4,000
62,000
100,000
39,000
60,000
48,000
1,000
2,000
5,000
5,000
300
150
600
Source
of qaa
Boilerb
Boilerb
Boiler0
Pickling
HNOj plant
Boiler*
Furnace®
Furnace
Furnace*
Boiler*
Furnace
Boiler*
Boiler*
Boiler"
Boiler*
Boiler*
Boiler"
Boiler*
Boiler"
Boiler"
Boiler"
Completion
Dec. 1973
Oct. 1974
Dec. 1975
July 1973
Mar. 1975
Dec. 1974
Dec. 1973
Mar. 1976
Oct. 1975
Dec. 1973
Dec. 1974
Dec. 1974
Mar. 1976
Aug. 1974
1973
Dec. 1974
Sept. 1975
Apr. 1975
1974
1974
1974
By-product
HNO,
HNO.,
(Gypsum Ca(N03)2l
NaNO2
NaNO.,, NO
KNOj
Gypaum, N_
Gypaum, N-
NaNO.,, NaCl
(NaNOj, NaCl, Na2SO<)
(NaNO.,, NaCl, Na2S04)
(NaN03, NaCl, Na2S04)
(NaNO.,, NaCl, Na2S04l
Na-,S04, NaNO3
(Oypsum, Ca(NO.,)2l
Gypaum, NH^
Gypaum, N,
Gypsum, Nj
(NII4)2S04
H2S04, N2
Gypaun, 1*2
• Oil-fired boiler.
b Gas-fired boiler.
c Coal-fired boiler.
d Metal-heating furnace.
Iron-ore alntering furnace.
110
-------
Ozone can oxidize NO to N2°5 wnen added in an excessive
amount.
NO + O3 -»• N02 + 02
2NO + 303 -> N2O5 + 302
Ozone is fairly expensive, and costs $1.20 to 1.40 per
kilogram (about 1.5 mills/kWh). A large-scale ozone genera-
tor with a capacity of 100 kilograms per hour of ozone is
near completion. It can treat about 230,000 Nm /hr of flue
gas (76 MW equivalent) containing 200 ppm NO. The cost of
ozone is expected to decrease to some extent with the large
generator. The cost of chlorine dioxide is 30 to 40 percent
less than that of ozone, but chlorine dioxide has the
disadvantage of introducing hydrochloric and nitric acids,
which complicate the system.
2NO + C102 + H2O -> N02 + HN03 + HC1
moisture
Solutions of potassium and sodium permanganates, sodium and
calcium hypochlorites, and hydrogen peroxide have been used
for the oxidation in the liquid phase, but these chemicals
are also expensive.
Oxidation-Absorption and Absorption-Oxidation Processes
In oxidation-absorption processes the NO is first
oxidized with a gaseous oxidizing agent and then absorbed.
In absorption-oxidation processes the NO is absorbed in a
111
-------
solution containing an oxidizing agent. Usually NO absorp-
X
tion occurs more slowly in the latter case because NO must
be absorbed in the liquor before it can be oxidized. Most
plants using nitric acid for such processes as metal washing
emit a gas fairly rich in NO (1000 to 10,000 ppm). However,
X
the amount of gas is not great (500 to 5000 Nm /hr). In
many of the plants, all or part of the NO is oxidized to N02,
and the gas is absorbed in a sodium hydroxide solution.
Activated carbon is used in some plants as a catalyst for
the oxidation of NO by air. In other plants NO is absorbed
X
in a solution containing an oxidizing agent such as NaCIO
or H-O-. In both cases the resulting liquor, containing
nitrate and nitrite, is sent to a wastewater treatment
system. Such processes cannot be applied on a large scale
because the treatment does not remove the nitrogen compounds
from the wastewater.
Tests have been made in pilot plants to recover nitric
acid for industrial use or to recover potassium or calcium
nitrate for fertilizer. Those processes do not seem prom-
ising because of the high cost and the limited demand for
the by-products.
Oxidation-Reduction and Reduction Processes (Simultaneous
Removal)
Since 1973 many oxidation-reduction and reduction pro^-
cesses have been developed in which NO and SO are absorbed
X X
112
-------
simultaneously. In the oxidation-reduction process NO is
first oxidized and then absorbed together with SO in a
X
slurry or a solution. In the reduction process NO is
absorbed with SO in a liquor containing ferrous ion, which
X
can form an adduct with NO. Usually EDTA (ethylenediamine
tetraacetic acid, a chelating compound whose present cost in
Japan is about $2700/t) is added to promote the reaction of
ferrous ion with NO (Figures 59, 60) (19).
In both cases various reactions, as shown below, occur
in the liquor or slurry and result in the reduction of NO
J\.
by S02 (or sulfite) to NH_ through imidodisulfonic acid
(HN(SO3H)2), sulfamic acid (H2NSO3H), or a salt or either
acid (20).
REDUCTION
MONO
,NH
k(HO)2NS03H
%
»• nur
t
HONHSO
^ t
^ HON(SO
V.
'"2
3H
\
3H>2
X.
3
N,
H2NS03H HYDROLYSIS
^ 1
HN(S03H)2
k
113
-------
oo
'o
s
o
CO
O\
I I
10 30 50 70
LIQUID TEMPERATURE, "C
Figure 59. NO absorption in EDTA-Fe(II)
liquor (0.01 mole/liter).
(Amount of liquor, 1 liter; gas flow, 1.5 liter/min.;
pH of liquor, 5.65; initial NO concentration, 275 ppm)
114
-------
O)
CO
on
O
00
CO
< 1
I
0 0.01 0.02
EDTA-Fe(II), mole/liter
Figure 60. EDTA-Fe(II) concentration and
NO absorption at 50°C.
(Amount of liquor, 1 liter; gas flow, 1.5 liter/min.;
pH of liquor, 5.65; initial NO concentration, 275 ppm)
115
-------
NO can be reduced to N_. The reactions are complex
x 2 c
but may be simply described:
Tanaka (19) has found that a compound Na2SO3*2NO is
formed when NO is absorbed in an NaSO_ solution. The com-
pound is stable at high pH (above 8) but decomposes to form
Na2SO. and N20 at lower pH. It is likely that in addition
to N» or NH3/ N20 also is formed in some of the wet pro-
cesses.
In some of the processes a considerable portion of NO
A.
remains in the resulting liquor as a nitrite and nitrate,
which would cause a problem in wastewater treatment.
The advantage of such wet processes over dry processes
is that they can simultaneously remove S09 and NO without
«<- X
problems of dust and ammonium bisulfate. They have not yet
been commercialized on a large scale. Five relatively small
commercial plants and seven pilot plants are in operation.
TOKYO ELECTRIC - MHI OXIDATION-ABSORPTION PROCESS (21)
MHI has constructed a pilot plant at Minamiyokahama
Station, Tokyo Electric Power. The pilot plant has a
capacity to treat 100,000 Nm /hr of flue gas from an LNG-
fired boiler to remove NO by a wet process that produces
X
nitric acid. A flowsheet of the process is shown in
116
-------
Figure 61. Specifications for the system are shown in
Table 18.
The flue gas containing about 100 ppm NO is cooled
J*L
with water spray, injected with ozone to oxidize NO to N-O-,
and then washed with water in an absorber. More than 90
percent of the NO is removed when more than 1.7 moles of
X
ozone are used per mole of NO. Resulting dilute nitric acid
(about 10 percent concentration) is concentrated to 60
percent for industrial use. Exhaust gas from the absorber
is treated with a calcium sulfite slurry and a catalyst to
remove residual ozone.
The plant has operated without serious problems, but
the process consumes large amounts of ozone and fuel for gas
reheating and nitric acid concentration.
KAWASAKI MAGNESIUM PROCESS
Kawasaki Heavy Industries constructed a pilot plant to
treat 5000 Nm /hr of flue gas from a coal-fired boiler at
Takehara Station, Electric Power Development Company. The
process was originally designed for S02 removal and equi-
molecular removal of NO and N0~ with a magnesium hydroxide
slurry, followed by oxidation and lime addition to recover
gypsum, calcium nitrate, and magnesium hydroxide (1).
A flowsheet of the process is shown in Figure 62. Flue
gas is first treated in a venturi scrubber to remove S0_,
117
-------
00
COOLER
FLUE GAS
fcSlfcL
REMOVAL
OXIDIZER
AIR
DILUTE
HN00
OZONIZER
TO STACK
MIST
ELIMINATOR
FAN
Q
-------
Table 18. SPECIFICATIONS FOR THE TEST PLANT
(TOKYO ELECTRIC - MHI PROCESS)
Material
Structure
Dimensions, m
Cooler
Oxidizer
NO absorber
Ozone absorber
Mist eliminator
Blower
SS 304L
SS (plastic coated)
SS 304
SS 304
Polypropylene
SS 60 (Impeller)
Packed tower
Square tower
Packed tower
Packed tower
Chevron
Centrifugal
4 x 4 x 11
4 x 4 x 15
4 x 4 x 19.5
4 x 4 x 19.5
2.6 x 1.8 x 3
370 MW
-------
N)
O
FLUE GAS
ABSORBER
J2
H,
i NO
JLL
T
MgSO
CLEANED SAS
Mg(OH).
A
AIR
MgSO,
GYPSUM
Ca(N03)2
Ca(NO:
Figure 62. Flowsheet of Kawasaki magnesium process.
-------
using a slurry containing Mg(OH)2, Ca(OH)2, and Ca(N03)2 at
a pH of 6.2 to 6.5 to produce CaS03, MgSO.,, and Mg(N03)2.
The desulfurized gas is then mixed with N02 and treated in
an NO absorber with a slurry containing Mg(OH)0 and a small
X £
amount of Ca(NO.J2 at a pH of 6 .^8^ to 7.8 to produce Mg(NO-)2/
Mg(N02)2, and a small amount of CaCO_. The discharge from
the NO absorber is sent to a decomposer and treated with
J\.
sulfuric acid to release concentrated NO, which is air-
oxidized to N02 and returned to the absorber. Both dis-
charges from the scrubber and the decomposer are air-oxi-
dized to produce a slurry containing gypsum, MgSO., and
Mg(N03)2. The slurry is treated with Ca(NO.J2 to precipi-
tate gypsum, which is then centrifuged. The separated
liquor containing Ca(NO3)_ and Mg(NO,)2 is treated with lime
to precipitate Mg(OH)2. Most of the liquor containing
Ca(N03)2 is separated and recycled or recovered as a by-
product. Mg(OH)2 containing a small amount of Ca(NO,)2 is
recycled to the scrubber and the absorber.
The absorbing system went into operation in July 1975,
and the whole system went into operation in April 1976. The
flue gas contains about 1000 ppm SO, and 200 to 400 ppm NO
^ X
at 130°C. The volume ranges from 1250 to 6250 Nm /hr. More
than 95 percent of the SO2 but less than 50 percent of the
NO has been removed at a liquid-to-gas ratio of 10.
X
121
-------
Tests on ozone-oxidation of NO have been carried out to
increase the denitrification ratio. The addition of 1.3
mole or 1.5 mole of ozone to 1 mole of NO increased the
removal ratios to 80 and 95 percent, respectively.
Equimolecular absorption is not suitable when the NOx
concentration is low. The limitation on use of calcium
nitrate is another drawback of the process.
ABSORPTION-OXIDATION PROCESSES
Nissan Engineering Process (1)
Nissan Engineering, a subsidiary of Nissan Chemical,
developed a manganate process and constructed small com-
mercial units for a user and a producer of nitric acid
(Table 17). NO in waste gases is absorbed by a sodium
.X
manganate (or permanganate) solution to form a sodium ni-
trite solution and manganese dioxide sludge. Sodium nitrite
is either sent to a wastewater treatment system or reacted
with nitric acid to produce sodium nitrate, which is used
for some purpose, and concentrated NO, which is sent to a
nitric acid plant. The process is not suitable for flue gas
containing S02 because the S02 consumes the manganate or
permanganate.
MON Permanganate Process (1)
Mitsubishi Metal, MKK, and Nippon Chemical Industries,
a chemical company producing permanganates, have jointly
developed a permanganate process (Table 17).
122
-------
NO is reacted with a potassium permanganate solution
X
to form a potassium nitrate solution and manganese dioxide
sludge. The manganese dioxide is converted to potassium
permanganate by a conventional process including electroly-
sis. The potassium nitrate can be used as fertilizer. The
process seems expensive. For flue gas treatment, SO-
should be removed prior to the denitrification because the
S02 consumes the permanganate.
Kobe Steel Process (1)
Kobe Steel has operated a pilot plant to treat 1000
Nm /hr of waste gas from an iron-ore sintering plant. SO,
£»
is first removed in a calcium chloride solution containing
lime. The treated gas is then reacted with a calcium chloride
solution containing Ca(OCl)2 as the oxidizing agent. In
this reaction, calcium nitrate is formed and chlorine is
evolved and then caught by a calcium sulfite slurry from the
SO- removal system to produce calcium chloride and gypsum.
Calcium nitrate is decomposed chemically and reduced to
produce N.. A larger pilot plant (50,000 Nm /hr) is near
completion. The process requires highly corrosion-resistant
materials.
Hodogaya Chemical Process
Hodogaya Chemical has developed a process to remove NO
K
by using an NaC10_ solution (Table 17). When the gas con-
tains S02/ sulfuric acid is also formed.
123
-------
FUJIKASUI-SUMITOMO SIMULTANEOUS REMOVAL PROCESS
(MORETANA PROCESS)
Fujikasui Engineering and Sumitomo Metal Industries
jointly developed a sodium scrubbing process for removal of
SO, and NO (Figure 63) and constructed three plants (Table
£• J^.
17). Flue gas containing 1200 to 1300 ppm S02 and 240 to
280 ppm NO is first cooled by a water spray. Gaseous C102
is added to the gas just before the scrubber and oxidizes NO
into N02 within 0.5 second. The gas is then introduced into
a Moretana scrubber with specially designed perforated
plates and is reacted with a sodium hydroxide solution.
More than 98 percent of the S0_ is absorbed to produce
sodium sulfite. About 90 percent of the NO in the gas is
JC
removed. About half of the removed NO is converted into N_
J\, £•
by reaction with sodium sulfite, and the rest is converted
into sodium nitrate.
The liquor from the scrubber contains 18 to 22 percent
sodium sulfate, 0.5 to 1.8 percent sodium sulfite, 0.4 to
0.8 percent sodium chloride, and 0.4 to 0.9 percent sodium
nitrate. The liquor is concentrated to separate most of the
sodium sulfate in a crystal form. The remaining liquor is
sent to a wastewater treatment system.
Capital cost ranges from $60 to $90 per kilowatt.
Operating cost, including depreciation (7 years), is roughly
$30 per kiloliter of oil (7 mills per kWh).
124
-------
QUENCHING TOWER ABSORBER
Na2S04, NaNO,
NaCl
ABSORBENT PRETREATMENT
HOLDER CRYSTALLIZATION
I
Figure 63. Flowsheet of Moretana simultaneous removal process,
125
-------
Fujikasui recently started tests on ozone oxidation to
reduce NO to N- or NH.,. The process is followed by lime
J^ £• O
scrubbing to produce gypsum.
MHI OXIDATION-REDUCTION PROCESS (21)
MHI has modified the wet lime/limestone flue gas desul-
furization process for simultaneous removal of NO and has
Ji
operated a pilot plant (Table 17, Figure 64). Flue gas from
an oil-fired boiler is cooled to 60°C by a water spray.
Then ozone is introduced into the flue gas prior to scrub-
bing. A water-soluble inorganic catalyst is added to a
lime/limestone slurry to promote the reaction of NO-.
About 80 percent of the NO is removed together with more
X
than 90 percent of the S02 when the gas contains more than 3
moles of SO- per mole of NO The slurry discharged from
^ 1\.
the scrubber contains solid gypsum and dissolved nitrogen-
sulfur compounds (reactions 10, 11, and 12). The slurry is
centrifuged to recover gypsum. Most of the liquor is
returned to the absorber after lime or limestone is added.
A small portion of the liquor is treated in a decomposer at
100° to 130°C to decompose more than 95 percent of the
nitrogen and sulfur compounds (reactions 13 through 16). A
small portion of the nitrogen is converted to N2. Most of
it is converted to NH HSO., which is then treated with lime
to recover gypsum and gaseous ammonia, 5 to 10 percent in
126
-------
1 Fan 2 Cooling Tower
4 Mist Catcher 5 Reheater
7 Absorbent Makeup 8 Thickener
10 Decomposer 11 Neutralizer
Figure 64. MHI simultaneous removal process.
3 Scrubber
6 Ozonizer
9 Centrifuge
127
-------
concentration, which may be used for some other purpose
(reaction 17) .
2N02 + Ca(OH)2 + CaS03«l/2H20 + Aq
Ca(N02)2 + CaS04'2H20 (10)
4(CaS03'l/2H20) + Aq
2CaNOH(S03)2 + 3Ca (OH) 2 (11)
CaNOH(SO3)2 + CaSO '1/2H20 + Aq ->
CaNH(S03)2 + CaS04-2H20 (12)
2Ca[NOH(S03)2] + Aq +
Ca(NOH'HS03)2 + Ca(HSO4)2 + Aq (13)
Ca(NOH'HS03)2 + Aq -»•
2/3N2 + CaSO.-2H2) + 2/3NH4«HS04
+ 1/3H2S04 + Aq (14)
2Ca[NH(S03)2] + Aq -»•
S03)2 + Ca(HS04)2 + Aq (15)
+ Ca(HS04)2 + Aq +
HS04 + 2CaS04'2H20 + Aq (16)
Ca(OH)2 + Aq ^
NH3 + CaSO4'H2
-------
IHI SIMULTANEOUS REMOVAL PROCESS (22)
IHI has been testing an oxidation reduction process at
a pilot plant that can treat 5000 Nm /hr of flue gas from an
oil-fired boiler containing about 1000 ppm SO^ and 200 ppm
Figure 65).
The flue gas is cooled, injected with ozone to oxidize
NO (Figure 65).
X
NO to NO2, and treated in a scrubber with a lime/limestone
slurry at a pH of 5 to 6 that contains small amounts of
CuCl9 and NaCl as catalysts for NO absorption (Figures 66
^ X
and 67). A lower pH is favorable to NO absorption. More
H
than 80 percent of the NO and 90 percent of the S02 are
absorbed, resulting in various reactions in the liquor. The'
following reactions are assumed to take place:
2NO2 + 4CaSO3 -> N2 + 4CaSO4
4N02 + 4CaS03 + 2H20 •*• Ca(NO2)2 + Ca(N03)2 + 2Ca(HS03)2
2NO2 + 3CaS03 ->• N2O + 3CaSO4
2N00 + SCaSO- + Ca(HSO_)0 + H0O •*• Ca (NH^SO.,) _ + SCaSO.
2. 6 524 £ 3 £. 4
N2O5 + 2CaS03 + H20 •* Ca(NO3)2 + Ca(HS03)2
About half of the NO is reduced to N~, and half stays
X «
in the liquor as nitrate and other compounds. Tests are in
progress for further reduction of NO to N9.
Jrk, £t
OTHER OXIDATION-REDUCTION PROCESSES FOR SIMULTANEOUS REMOVAL
Chiyoda has made a simple modification of the Thorough-
bred 101 process to remove NO (1). Ozone is added to the
X
gas prior to scrubbing. More than 60 percent of the NO is
129
-------
U)
o
OZONIZER
COOLER
(DUST REMOVAL)
CLEANED GAS
/"^ABSORBER
A A A
OXIDIZER
CaO
CATALYST
CENTRIFUGE
GYPSUM
Figure 65. Flowsheet of IHI simultaneous removal process.
-------
0 0.5 1.0 1.5 2.0 2.5 x 10
,, mole/liter
-2
Figure 66. Effects of CuCl2 and NaCl concentrations
on NO removal efficiency.
X
(CaSO_ 5%, pE 5.5, N09/N0 95%)
•J
-------
removed along with about 90 percent of the SO-. A portion
of the removed NO is converted into nitric acid, which
X
forms calcium nitrate, and the rest is converted into N£ and
N_0. Wastewater treatment is required to remove the nitrate.
Osaka Soda, a chemical company, developed a process
similar to the Fujikasui-Sumitomo process and constructed a
prototype unit (Table 17). Tests on wastewater treatment
are in progress.
Shirogane Company, an engineering company, has built a
system (Table 17) based on a process similar to the Fujikasui-
Sumitomo process except that ozone is substituted for chlo-
rine dioxide. The wastewater containing sodium sulfate and
nitrate is sent to a treatment system along with other
wastewaters.
KUREHA PROCESS
Kureha Chemical has developed a process to remove NO
j),
in combination with the sodium acetate flue gas desulfuriza-
tion process (Figure 68).
S02 is absorbed by a sodium acetate solution to produce
sodium sulfite and acetic acid (reaction 18). NO is absorbed
by a sodium sulfite solution in the presence of acetic acid
and a soluble metallic catalyst to produce sodium imidodi-
sulfonate (reaction 19).
132
-------
TREATED
GAS
i
i
i f—
WATER
ACETIC
ACID
RECOVERY
S02,NOX
REMOVAL
FLUE GAS
OXIDATION
r*---
_J
No
2
<
J
•i-ADL
•*-Ca(OH)2
*-H2S04
CaO
(CaC03)
I
'2>2
I
GYPSUM
Figure 68. Flowsheet of Kureha simultaneous
removal process.
133
-------
S02 + 2CH3COONa + H20 -*• Na2S03 + 2CH3COOH (18)
2NO + 5NaS03 + 4CH3COOH -> 2NH(S03Na)2
+ Na2S04 + 4CH3COONa + H20 (19)
The remaining sodium sulfite is air-oxidized into sulfate.
The sulfate is treated with calcium acetate as in the flue
gas desulfurization process.
Sodium imidodisulfonate is reacted with slaked lime to
precipitate and separate sodium calcium imidodisulfonate
(reaction 20) , which is then hydrolyzed in the presence of
sulfuric acid into sulfamic acid (reaction 21) . The sul-
famic acid is treated with calcium nitrite to release
nitrogen (reaction 22) .
NH(S03Na)2 + Ca(OH)2 + CH-jCOOH ->
NNa(SO3)2Ca + CH3COONa + 2H_O (20)
2NNa(SO3)Ca + H2S°4 + 2H2° "*
2CaSC>4 (21)
Ca(N02)2
2N2 + CaS04 + H2S04 + 2H20 (22)
Kureha has been operating a pilot plant with a capacity
to treat 5000 Nm /hr of flue gas from an oil-fired boiler.
The process seems fairly complicated with many reaction
steps. Recently, the sodium imidodisulfonate has been found
useful as a builder of detergents to replace sodium tri-
polyphosphate, which has been causing eutrophication problems.
134
-------
Tests have been in progress on the effect of the disulfonate
on the environment. Possible commercial use of the disul-
fonate will make the process useful.
MITSUI SHIPBUILDING PROCESS
Mitsui Shipbuilding has developed a simultaneous removal
process that produces concentrated S02, which can be used in
sulfuric acid production (Figure 69).
Flue gas is treated with a ferrous compound solution
containing EDTA which absorbs both SO^ and NO.
A portion of the ferrous ion is converted to ferric ion
by the oxygen in flue gas. The ferric ion in the absorbed
liquor is then reduced to ferrous ion by electrolysis, and
the liquor is sent to a stripper, where it releases con-
centrated S02 and NO by steam distillation. The NO is
reduced to N_; the SO2 is used in sulfuric acid production.
The liquor from the scrubber is returned to the absorber.
In tests with a pilot plant (150 Nm /hr) about 95 percent of
the S00 and 85 percent of the NO were removed at a liquid-
2. X
3
to-gas ratio of 1 liter/Nm .
It is estimated that the plant cost is $80 million for
a 67-megawatt plant. EDTA consumption per year is 300 to
400 tons at a cost of $500,000 to 600,000.
By using H-S in the reduction step, elemental sulfur
may be produced. Tests with a larger plant are required for
further evaluation.
135
-------
CLEANED
GAS-*
ABSORBER
COOLER
FLUE GAS _
S02,NO
REDUCTION
WASTE-
WATER
COOLING
WATER
STRIPPER
Figure 69. Flowsheet of Mitsui Shipbuilding process,
136
-------
CHISSO PROCESS (CEC PROCESS)
Chisso Engineering, a subsidiary of Chisso Corporation,
has developed a process for simultaneous removal of S02 and
NO from flue gas by ammonia scrubbing using a catalyst
X
(chelating compound) to produce ammonium sulfate. A pilot
plant treating 300 Mm /hr of flue gas from an oil-fired
boiler has been operating (Figure 70) .
SO_ and NO from the flue gas is absorbed in an
z, x
ammoniacal solution containing a soluble catalyst to reduce
the absorbed NO to NH_ by ammonium sulfite and bisulfite,
X o
which are formed from SO- and ammonia. Most of the catalyst
is separated from the product solution, containing ammonium
sulfate and sulfite and the intermediate compounds. The
solution is oxidized by air and then heated to convert the
intermediate compounds into ammonium sulfate. The product
solution is concentrated in an evaporator to crystallize
ammonium sulfate, which is separated by a centrifuge. The
mother liquor, which contains a small amount of the catalyst,
is returned to the catalyst separation step. The over-all
reaction may be expressed as follows:
2NO + 5S0 +
At an NO concentration of 300 ppm it is desirable to
X
have more than 1200 ppm SO_ in the flue gas in order to
recover 80 percent of the NO .
X
137
-------
so
OJ
CO
WATER
"GAS"*
i
i
i
i
i
r-
i
\
^_i
NH,
Y "
\
i
1
k
OX IDAHO
A
1
I
AIR
•«
?'
V
i
A
NH,
en 3
S04
CATALYST DECOM- NEUTRAL I- CRYSTALLI
RECOVERY POSITION ZATION ZATION
•to — ^
1 r
(NH4)?SOA
Figure 70. Flowsheet of CEC process.
-------
Reaction of NO with the sulfite liquor is not rapid,
-------
ABSORB
BER
FLUE
GAS
NaOH
FILTER
CRYSTAL-
LIZER
DUST
REACTOR (REDUCTION)
V
Na2S2°6
Na2S206
SEPARATOR
FURNACE FOR
DECOMPOSITION
o
CO
CM
-------
The NO adduct reacts with the sulfite to form sodium sulfate
and nitrogen by the following reaction:
Fe++-EDTA'NO + Na2S03 + Fe-EDTA + 1/2N2 + Na2S04
Most of the resulting liquor is returned to the absorber,
A portion is sent to a crystallizer, where sodium dithionate
Na2S2Og'2H20 iTs crystallized. The dithionate is separated
and heated at 300°C to be decomposed to Na2S04 and SO2, both
of which are sent to a reactor and reacted with calcium
sulfite to precipitate gypsum.
Na0S0Oc'2H06 -> Na.SO. + SOn + 2H0O
2 2 b 2 24 2 2
Na0SO. + SO0 + CaSO. + H_0 -> 2NaHS00 + CaSO.
242 32 3 4
2NaHS03 + Ca(OH)2 -*• Na2S03 + CaS03
The sodium bisulfite solution formed by the reaction is
treated with lime to precipitate calcium sulfite and to
regenerate sodium sulfite. The former is sent to the
reactor, and the latter is recycled to the absorbing system.
Chlorine, derived from the fuel, accumulates in the scrub-
bing liquor and can be eliminated by ion exchange. Asahi
Chemical has had much experience in ion exchange.
Asahi Chemical estimates that the plant cost for a
500,000 Nm /hr unit (160 MW equivalent) is $16 million and
3
that requirements for the treatment of 10,000 Nm of gas
containing 2000 ppm of S0_ are as follows:
141
-------
Ca(OH)2 6.7 kg
FeSO4«7H20 1.0 kg
EDTA 1.0 kg
NaOH 4.2 kg
Oil (Gas reheating) 30 kg
Oil (Thermal decomposition) 5 kg
Steam 60 kg
Cooling water 6 tons
Power 150 kWh
The system is a combination of several feasible unit
processes. Operating data from a larger pilot plant may be
needed for a reliable evaluation.
142
-------
SECTION 5
EVALUATION AND DISCUSSION
SIGNIFICANCE OF FLUE GAS DENITRIFICATION
The present development in flue gas denitrification in
Japan is a result of the stringent ambient standard for
NO2, 0.02 ppm daily average of hourly values, about equiva-
lent to 0.01 ppm yearly average. The ambient N00 concen-
£•
trations in large cities such as Tokyo and Osaka range from
0.02 to 0.05 ppm (yearly average) and are not higher than
concentrations in Los Angeles and Chicago. With the approach
of the time limit for attaining the standard — 1978 in most
regions and 1981 in polluted regions — doubts have been
raised as to the necessity for such a stringent standard
because it will be difficult to attain and will require tre-
mendous expenditures.
In any case, it is necessary to develop further tech-
nologies for both flue gas denitrification and combustion
modification as long as consumption of fossil fuels con-
tinues to increase. Particularly in countries where coal
consumption is increasing, denitrification will be needed
sooner or later because it is difficult to reduce the NO in
143
-------
the flue gas from coal burning to below 400 ppm by combus-
tion modification without increasing the particulates. On
the other hand, NO concentrations in flue gas from burning
X
of gas or oil can be reduced to 50 or 100 ppm by combustion
modification.
Because the denitrification processes are not yet fully
developed and are still too expensive to be used in treating
dirty gases containing much SO and dust, industries in
A.
Japan have been moving toward the use of clean fuels such as
LNG, kerosene, naphtha, and low-sulfur oil (less than 0.3%
sulfur). The clean fuels, however, are expensive and are
limited in supply. Under the stringent NO control in
A.
Japan, flue gas denitrification will be indispensable,
particularly for new plants, even though clean fuels are
barned with combustion modifications.
COMBINATION OF FLUE GAS DESULFURIZATION AND DENITRIFICATION
Clean gas can be easily denitrified with more than 90
percent efficiency by SCR processes using a fixed bed of a
catalyst that has more than 2 years life. Many processes
for treatment of a dirty gas have been developed and are
shown schematically in Figure 72.
System 1 in the figure is an ideal dry process for
simultaneous removal of SO and NO by which a flue gas at
5C X
a normal temperature of 140°C, after passing through an
144
-------
No.
i^WX^ 14°
14° ^/^DDtT 140
*"( DOS
250 /7m\ 250 x-v 160
*\ DOS ' WAH* *
No. 8
©
HEAT
LOSS. %
5.5
3.5
3.5
1.0
3.5
1.0
3.0
BOILER (AH) AIR HEATER (ESP) ELECTROSTATIC PRECIPITAT0R
DON ) DRY DENITRIFICATION
DOS ) DRY DESULFURIZATION
WON ) WET DENITRIFICATION
WDS ) WET DESULFURIZATION
H } HEATER
HE) HEAT EXCHANGER
Figure 12. Combination of denitrification and
desulfurization (Figures show gas temperature, °C).
145
-------
air heater and ESP, may be treated without changing the
temperature. Such a process is not yet fully developed.
The heat requirements shown in Figure 72 are caused by
losses in heat recovery and by required reheating of the
gas. The amount is expressed as a percent of the heat for
boiler operation, taking the loss in System No. 1 as zero.
Heat needed to operate desulfurization and denitrification
units, such as the heat used to concentrate a liquor or to
regenerate the catalyst, is not included in the heat loss
shown in the figure.
Systems 2 through 7 include dry denitrification pro-
cesses — ammonia reduction with or without a catalyst. For
these systems, the air heater outlet temperature of flue gas
is assumed to be 160°C, which is 20°C higher than in System
No. 1. This temperature accounts for 1 percent heat loss
because ammonium bisulfate will deposit in the air heater at
the lower temperature. The temperature may vary with S0_
and NH-. concentrations of the gas, with types of air heater
(or heat exchanger), and with cleaning procedures. In many
cases the gas temperature will have to be kept above 160°C.
System 2 shows a combination of wet desulfurization
(WDS) and dry denitrification (DON), as used in the Hitachi
Shipbuilding process, which has been operated commercially
in Japan since 1975. In addition to a standard air heater,
146
-------
the process requires a heat exchanger and much energy for
gas heating, although the process seems feasible for treat-
ing flue gases from oil burning.
In systems 3 and 4, flue gas from the air heater at
400°C is first subjected to SCR and then to wet desulfuri-
zation. A considerable energy savings is attained compared
with System 2. System 3 uses a moving-bed reactor, as in
the Kurabo process. Treatment of flue gas from oil burning
may not require an ESP because most of the dust is caught on
the catalyst bed, but the catalyst must be treated contin-
uously or intermittently. Treatment of flue gas from coal
burning will require an ESP or other dust removal facility
ahead of the reactor. System 4 uses a parallel-passage
reactor like the Japan Gasoline - Shell reactor. The
catalyst bed is hardly contaminated by the dust, although an
ESP may be needed after the air heater. The advantage of
the parallel-passage reactor may be greater if it can treat
flue gas from coal burning without using a hot ESP ahead of
the reactor.
System 5 shows simultaneous S09 and NO removal at
^ X
400°C, as in the Shell process. A dry flue gas desulfuriza-
tion process is usually much more expensive than a wet
process, although it does not require gas reheating. In
system 5, however, the disadvantage might be compensated for
by the capability of simultaneous S09 and NO removal.
^ X
147
-------
System 6 shows ammonia injection into the boiler
without a catalyst, followed by wet desulfurization. The
process is simple and is less expensive than others, but
even a large-scale system may not easily attain more than 50
percent NO removal.
For Systems 3 through 6, any fluctuating gas tempera-
ture resulting from a change of the boiler load will present
a problem. In Systems 3 through 5 ammonium bisulfate may
deposit on the catalyst and poison it when the temperature
drops below 350°C. In order to maintain the gas tempera-
ture, it may be necessary to install an auxiliary burner or
a device to take some high-temperature gas from the boiler.
The problem is more serious for the ammonia injection pro-
cess without a catalyst because the process has a narrow
range of suitable reaction temperatures around 970°C.
Although the addition of hydrogen with ammonia can reduce
the temperature to 870°C, use of hydrogen at power plants
may not be practicable. The process may be better suited to
boilers that have smaller load fluctuations.
System 7 is a dry simultaneous removal process using
activated carbon, such as the Unitika process. More than 90
percent of the SO9 and the NO is removed with little heat
£ X
loss. Actually, a considerable amount of heat would be
required for the removal of the S0~ from the carbon. The
148
-------
required reaction temperature of 250°C is not favorable
because two air heaters may be needed, as shown in Figure 72.
For the treatment of flue gas from coal burning, an ESP will
be required ahead of the reactor. The activated carbon,
reactive at 100° to 150°C, (developed by Hitachi, Limited)
permits treatment similar to that by System 1. However,
considerable amounts of ammonium bisulfite and sulfate may
deposit on the carbon and necessitate frequent regeneration.
System 8 is a wet simultaneous removal process. The
wet process has advantages over the dry process in achieving
simultaneous removal of more than 80 percent of the NO and
X
90 percent of the SO2 without the problems of dust and
ammonium bisulfate. The process involves some problems, as
described in Section 4, and has not yet been commercialized
on a large scale.
Many other processes are under development, but they
may not be as useful in treating large amounts of dirty flue
gas as those mentioned above.
Consumption of ammonia would present a problem if the
dry processes were to be used extensively. It is estimated
that about 400,000 tons of ammonia will be consumed yearly
to attain the ambient standard by dry processes alone. This
amount is about half the total annual consumption of nitro-
149
-------
gen fertilizers in Japan. If the dry processes should be
used widely throughout the world, a serious shortage of
nitrogen fertilizer could occur, affecting the supply of
food. Since it is expected that the need for denitrifica-
tion will increase in many countries, development should be
concentrated not only the dry processes but also on the wet
processes capable of producing ammonia or nitrogen fertilizers,
Both the MHI and Chisso processes can produce ammonia
from NO by use of S0» as the reducing agent. For power
X ^
plants where no by-product is desired, the ammonia may be
fed to the boiler to convert a portion of the NO in the
flue gas to N? without a catalyst and thus facilitate
simultaneous removal by a wet process because a higher
SO9/NO ratio is more suitable for these processes.
^ X
In any case, combustion modification should be carried
out for NO abatement because it is much more economical
Ji
than flue gas denitrification and will enable the denitri-
fication to be accomplished with less expense when the NO
concentration of the gas is low.
150
-------
SECTION 6
REFERENCES
Descriptions in this report are based primarily on the
authors' visits to the denitrification plants, their dis-
cussions with the users and developers of each process, and
data made available by them. In addition, the following
publications were used as references:
1. Ando, J., H. Tohata, and G. Isaacs. NOX Abatement for
Stationary Sources in Japan, EPA-600/2-76-013h (in
English). January 1976.
2. Ando, J., and H. Tohata. NOX Abatement Technology in
Japan, EPA-R2-73-284 (in English). June 1973.
3. Yamagishi, K., et al. Low NOX Burner Developed by
Tokyo Gas Co., Journal of the Japan Society of Mechanical
Engineers, Vol. 77, No. 663. 1974. p. 225.
4. Idehara, S. NOX Control Techniques Developed by
Kawasaki Heavy Industries Co., Environmental Creation,
Vol. 6, No. 4. 1976. p. 62.
5. Kanamori, S. Low NOX Burner Developed by Volcano Co.,
Heat Management and Pollution Control, Vol. 26, No. 12.
1974. p. 47.
6. Nagaoka, M. NOX Control Techniques Used in Utility
Boilers, Environmental Creation, Vol. 6, No. 4. 1976.
p. 153.
7. Japan Environment Agency, Control Techniques used for
NOV Emissions from Stationary Sources. March 1975.
f±
8. Kobayashi, H., and K. Huruyaho. Techniques for NOX
Reduction in Boilers, Environmental Creation, Vol. 6,
No. 4. 1976. p. 87.
151
-------
9. Mizobuchi, I. Concrete Measures for NOX Reduction.
Symposium on Pollution Control, Japan Management
Association. June 1974.
10. Tokyo Metropolis Bureau of Environmental Protection,
Basic Investigation on NOX Control Techniques. March
1975.
11. Tokyo Metropolis Bureau of Environmental Protection,
Investigation on Fuel NO Conversion Ratio. March
1975. x
12. Report of Nitrogen Oxides Investigation Committee,
Japan Environment Agency. October 1975.
13. Atsukawa, M., et. al. Development of NOX Removal Pro-
cesses with Catalyst for Stationary Combustion Facilities,
Mitsubishi Juko Giho, Vol. 13, No. 2. 1976.
14. Shima, T. Problems of Denitrification Facility for
Boilers, Netsukanri to Kogai, Vol. 27, No. 12.
15. Lyon, R.K., and J.P. Longwell. Selective, Noncatalytic
Reduction of NO by NH , EPRI NO Seminor, San Francisco
(Feb. 1976). x J x
16. Ninomiya, N. Simultaneous Removal of NOX and SO2 by
Activated Carbon, Report of Takeda Chemical. 1975.
17. Seki, M. et al. Ammonium Hallide Activated Carbon
Catalyst to Decompose NOX in Stack Gases at Low Temperatures
around 100°C. American Chemical Society, Chicago.
August 1975.
18. Kawakami, W., and K. Kawamura. Treatment of Oil-fired
Flue Gas by Electron Beam, Denkikyokai Zasshi, 29.
December 1973.
19. Tanaka, T., M. Koizumi, and Y. Ishihara. Wet Process
for Nitrogen Oxides Removal from Flue Gases (Part 3).
Denryoku Chuokenkyujo Hokoku 275017. April 1976.
20. Audrieth, L.F., et al. Sulfamic Acid, Sulfamide and
Related Aquo-ammonosulfuric Acids, Symposium on the
Chemistry of Liquid Ammonia Solutions. American
Chemical Society, Milwaukee. September 1938.
152
-------
21. Atsukawa, M., et al. Development of Wet-type NO
Removal Processes. Mitsubishi Juko Giho, Vol. 1*, No.
2. 1976.
22. Yamada, S., T. Watanabe, and H. Uchiyama. Bench-scale
Tests on Simultaneous Removal of SC>2 and NOX by Wet
Lime and Gypsum Process. Ishikawajima-Harima Engi-
neering Review. January 1976.
153
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-6QQ/7-77-lQ3b
2.
3. RECIPIENT'S ACCESSION1 NO.
PR J
4. TITLE AND SUBTITLE
5. REPORT DATE
NOx Abatement for Stationary Sources in Japan
September 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jumpei Ando, Heiichiro Tohata (Chuo
University), Katsuya Nagata (Waseda University),
8. PERFORMING ORGANIZATION REPORT NO.
Laseke
9. PERFORMING ORGANIZATION NAME AND ADDRESS
PEDCo. Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-01-4147, TaskS
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final: 3/76-8/77
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer for this report is J. David Mobley,
Mail Drop 61, 919/541-2915.
16. ABSTRACT The report describes the status of NOx abatement technology for stationary
sources in Japan as of August 1976. The report emphasizes flue gas treatment pro-
cesses for control of NOx. It also features processes for the simultaneous removal
of NOx and SOx from flue gases. It examines the major Japanese dry and wet pro-
cesses, with respect to their applications, performance, economics, major technical
problems, developmental status, byproducts, and raw materials. It discusses the
application of dry processes, primarily selective catalytic reduction of NOx with
ammonia, to commercial scale gas- and oil-fired sources. It presents a review of
NOx cou-bustion modification technology in Japan, along with background information
on NO2 ambient concentrations, NO2 ambient standards, and NOx emissions standards
in Japan. The fact that NOx abatement technology in Japan is the most advanced in
the world is probably the result of the NO2 ambient standard in Japan's being the most
stringent in the world.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Nitrogen Oxides
Flue Gases
Sulfur Oxides
Catalysis
Ammonia
Natural Gas
Manufactured
Gas
Fuel Oil
Combustion
Air Pollution Control
Stationary Sources
Japan
Simultaneous NOx/SOx
Removal
Combustion Modification
Catalytic Reduction
13B
07B
2 IB
07D
21D
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
163
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
154
------- |