xvEPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 2771 1
EPA-600/7-79-205
August 1979
NOX Abatement
for Stationary
Sources in Japan
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine 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
8. "Special" Reports
9. Miscellaneous Reports
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-agency 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 sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systeftns; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA 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 recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-205
August 1979
NOX Abatement for Stationary
Sources in Japan
by
Jumpei Ando
Chuo University
Tokyo, Japan
Contract No. 68-02-2161
Program Element No. INE624
EPA Project Officer: J. David Mobley
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ACKNOWLEDGEMENTS
The author wishes to acknowledge the assistance of Gary Jones and
Milton Owen of Radian Corporation, Austin, Texas, in editing this report.
In addition, Radian Corporation's typing and graphics support is gratefully
acknowledged. (Radian Corporation's work was provided by EPA under contract
number 68-02-2608, Task 62.)
ii
-------
INTRODUCTION
This report is third in a series of reports on Japanese NOX control
technology authored by Dr. Jumpei Ando. Previous reports in this series are
listed below:
EPA 600/7-77-103b 1977
EPA 600/2-76-013b 1976
Recent efforts for air pollution control in Japan have been concentrated
on NOx abatement since ambient SOz concentrations have decreased to a level
low enough to meet the standard. Although the stringent ambient NO2 standard
promulgated in 1973—a daily average of 0.02 ppm—was amended in July 1978,
the new standard is still much more stringent than that in the United States;
it cannot be attained in large cities and industrial regions by combustion
modification for numerous NOX sources and by strict NOX control of automobile
exhausts, namely 0.25 g/km. Extensive flue gas treatment for stationary
sources is thus required.
Flue gas treatment (FGT) for NOX control has not been applied in the
United States at the present time. However, increasing' concern over NOX
emissions and the rapid development of FGT technology are causing U.S.
regulatory agencies to consider application of this technology to both
utility and industrial boilers. The California Air Resources Board (GARB)
has required existing utility boilers in the South Coast Air Quality Manage-
ment District to achieve 90 percent NOX control by 1990. Currently, FGT is
the only control technology that can achieve this level of control. Also,
FGT is being considered by the U.S. Environmental Protection Agency (EPA)
iii
-------
in their development of a New Source Performance Standard (NSPS) for
industrial boilers.
Application of FGT technology in Japan has preceded any similar appli-
cation in the U.S. Among the flue gas treatment processes, selective
catalytic reduction of NOX that uses ammonia and catalyst is the most
advanced and has been applied to over 60 commercial plants. Other processes,
such as selective noncatalytic reduction and simultaneous SOX and NOX
removal, have also been developed. Many more flue gas treatment plants are
required in order to meet the Japanese ambient standard and also to treat
new large NOX sources which are controlled by stringent local regulations.
The expanding use of coal will further increase the necessity for flue gas
treatment.
The present report will describe the status of energy supply and NOX
concentrations and regulations, as well as the chemistry, technology and
problems of NOX abatement for stationary sources in Japan.
Descriptions in this report are based primarily on Jumpei Ando's visits
to plants that use NOX abatement technologies, his discussions with the users
and developers of each process and data made available by them. An exception
is Chapter 2, which is based primarily on Katsuya Nagata's investigations.
iv
-------
SUMMARY
COMBUSTION MODIFICATION, SELECTIVE CATALYTIC AND NONCATALYTIC REDUCTION, AND
THEIR COMBINATIONS' FOR BOILERS
Among technologies for NOX abatement, combustion modification (CM) and
selective catalytic reduction (SCR) that uses catalyst and ammonia at 300-
400°C have progressed most remarkably. Selective noncatalytic reduction
(SNR) that uses ammonia at 800-1000°C has also been studied extensively and
used commercially.
Table 1 shows practical NOX reduction for flue gas from boilers burning
gas, oil and coal with a relatively low nitrogen content by means of CM, SCR,
SNR, and combinations thereof.
NOX can be reduced about 50 percent by using CM with low oxygen combus-
tion, stage combustion, off-stoichiometric combustion, and a low-NOx burner.
About 60-75 percent abatement can be achieved by adding flue gas recircula-
tion. SCR and SNR should be used when further reduction is necessary.
For SCR, catalysts based on TiOa have been predominantly used because
of the high reactivity and resistance to SOX. Different shapes of catalysts
and types of reactors as shown in Table 2 have been developed for various
gases. A dust rich gas may be treated by parallel flow type catalysts and
reactors without dust plugging problem.
With SCR, over 90 percent of the NOx can be removed when one or more
moles of NHa are used for each mole of NOX. In this case the concentration
of unreacted NHa tends to exceed 10 ppm, causing problems such as air
-------
TABLE 1. PRACTICAL NOX REDUCTION (NOX and NH3, ppm)
Fuel N0x By SCR By SNR By SNR +
NOX NH3 NOX NH3 NOX NH3
Gas
Oil
200a
100b
50C
300a
150b
100°
30
20
10
40
25
20
10
5
5
10
5
5
110
60
30
160
85
60
25
20
10
30
25
20
80
40
20
120
55
40
10
10
5
10
10
10
Coal 500a 50 10 250 50 180 10
250b 35 10 130 30 90 10
200° 30 10 110 25 80 10
without combustion modification.
With combustion modification excluding flue gas recirculation.
with combustion modification including flue gas recirculation.
Placing a small amount of parallel flow type catalyst in a duct at 300-
400°C.
vi
-------
TABLE 2. SCR CATALYSTS AND REACTORS SUITABLE TO VARIOUS GASES
Gas
Type
Clean
Semi-
dirty
Dirtyc
(dusty)
Very
dirtyd
Fuel Dust
(Gas source) (g/Nm3)
Natural gas
„ Below
Kerosene -
Naphtha
Heavy oil
Coal with 0.03-0.3
hot ESPb
Coal 15-25
Glass furnace, cement kiln
sintering machine
(containing alkaline vapor)
Catalyst
Shape
Granule, Ring
Granule, Ring
Honeycomb ,
Plate
Granule
(in elements)
Tube, Plate,
Honeycomb
Granule
Reactor
Type
Fixed bed
Moving bed
a
Fixed bed
Parallel
passage
Fixed bed
Moving bed
Parallel flow type
""Electrostatic precipitator
•»
"Treatment may not be easy
Treatment is difficult
vii
-------
preheater plugging by ammonium bisulfate and accumulation of NH3 in the
scrubber liquor of an FGD plant installed downstream of SCR. The NHa
leakage can be reduced by increasing the amount of catalyst or by adding
an ammonia decomposition catalyst, but this practice raises not only the
investment cost but also the running cost because of the increased gas
pressure drop through the reactor. Therefore, it is preferable to use a
smaller amount of NHs, about 0.85-0.95 moles, to obtain 80-85 percent
removal with a NHa leakage of 10 ppm or less.
Use of CM together with SCR may be appropriate in most cases because
catalyst, NHs, and power consumption can be reduced. Flue gas recirculation,
however, is fairly costly and may not be used with SCR.
SNR is less costly than SCR but gives a low NOX removal efficiency
with a large amount of leak NH , for example, 20-30 ppm for 45 percent NOx
removal. Moreover, SNR is difficult to apply to many existing boilers due
to the limited space for NH3 injection nozzles in a suitable temperature
range.
Application of SCR to existing boilers may also be difficult because
many of them may have no space for installation of an SCR reactor or for
connection of the reactor to the existing duct as shown in Figure 1, A.
Even when the space is available, some boilers may not have the excess fan
capacity to compensate for the pressure drop of 50-70 mmHaO required for
8P-9Q percent NOx removal by System A of Figure 1.
Parallel flow type catalyst may be placed in an existing duct as shown
in B of Figure 1. The major problem with System B is the high gas velocity.
The velocity in a duct is usually about 14 m/sec without a catalyst but
increases when a catalyst is installed, resulting in a large pressure drop.
Moreover, the catalyst may erode with flue gas from coal with a full dust
load. Therefore, the amount of the catalyst may be restricted. By using
a small amount of the catalyst, 30-50 percent removal with about 10 ppm of
viii
-------
BOILER
Eco.
Eco.
APH
APH
Eco.
_NH3
C NH3
Eco,
Eco.
NH3
APH
Eco.
NH3
Figure 1.
Systems for SCR catalyst installation.
(Shaded area indicates catalyst location)
IX
-------
leakage may be obtained with a 10-30 H20 mm pressure drop. If the
catalyst is used in combination with SNR as shown in System C of Figure 1>
50-70 percent removal may be attained with 10 ppm of NH3 leakage.
If the catalyst can be placed as shown in System D of Figure 1, the
gas velocity may not be exceedingly high and 60-70 percent removal may be
obtained with about 10 ppm of NHs leakage and a pressure drop of 30-50 mmHaO.
If there is not enough room for mixing NHa with gas, NH$ may be injected
upstream of the economizer. NHa may be injected in a hot zone (800-1000°C)
for further NOX abatement.
If a portion of the duct between the economizer and air preheater can
be enlarged as shown in System E of Figure 1, any type of catalyst may be
used to obtain 80-90 percent removal with less than 10 ppm of NHs leakage
and a pressure drop of 40-60 rnmHaO. The depth of the catalyst bed (the
length in the direction of gas flow) may be about 2 m with a honeycomb
catalyst and longer with parallel-passage and parallel-plate catalysts.
A larger diameter of the reactor may be needed when a granular catalyst
is used because the catalyst bed requires a smaller gas velocity.
System E is less costly than System A and suitable for a new boiler.
When an additional fan is needed for an existing boiler, System A may be
suitable because it has space for the fan.
There have been arguments about the necessity for CM when SCR is
installed. Flue gas recirculation is costly. It requires extensive duct
work and a fan and therefore may not be used when SCR is installed. Other
measures for CM which are less costly may be used in combination with SCR
or SNR because CM reduces the catalyst and ammonia requirement to some
extent. Decisions on the use of the CM methods should be based on cost
comparison and the allowable NOX and NHs emissions.
-------
HEAT LOSS (*)
ESP
NO. 1 0
AIR
PREHEATER
BOILER
JWQ't L i
150'C
ESP
NO. 2 1.0
BOILER
400 "C
AIR
PREHEATER
150'C
150
"t
FGO
55 'C
HEATER
"Z — -5"
75 'C
STACK
400'C
AIR
PREHEATER
NO. 3 5.0
BOILER
ESP
150'C 150'C
160'C
400'C
HEATER
HEAT
EXCHANGER
300T
55'C
STACK
REACTOR
400'C
AIR
PREHEATER
ESP
NO. 4 2.5
BOILER
150'C
150'C
150'C
HEAT
EXCHANGER
100'C
HEATER
'REACTOR
150'C
WET ESP
70-1361-1
Figure 2. Combined and simultaneous removal systems
(figures show temperatures, °C).
xi
-------
HEAT LOSS (*)
NO. 5 1.0
STACK
400'C
NO. 6 1.0
BOILER
400 *C
HOT ESP
REACTOR
AIR
PREHEATER
400'C,
150'C
HEATER
75 *C _
STACK
ESP
NO. 7 1.5
400'^
SNR
BOILER
AIR
PREHEATER
160*C
160'C
FGD
7 — r
HEATER
_, 55
•c
75'C
STACK
NO. 3 0
BOILER
400'Cj
AIR
PREHEATER
230'C
CATALYST
230*C
AIR
J PREHEATER
HL^
ISO* C
230
• 1
STACK
70-1362-1
Figure 2. Continued.
xii
-------
HEAT LOSS (X)
rtO. 9 0
COPPER OXIDE
CATALYST
4QO»C "— ™ 400
BOILER
i» 1 1
AIR
,c PREHEATER
\L 1I
ESP
150"£-
1502 |
I
;
STACK
NO. 10 0 BOILER
AIR
ESP
150»C ——- 150'C
100JC
COOLER
ELECTRON
BEAM
100'C
100'C,
STACK
ESP
NO. 11 1.0
BOILER
400-C I
AIR
'REHEATER
b=±J
150'C_
ESP
WET SIMUL.
REMOVAL
150'C
55 "C
r^^n 75-c 1
•••^^^^••^
STACK
70-1363-1
Figure 2. Continued,
xiii
-------
A combination of CM, SNR and SCR may be useful when Systems A and E
shown in Figure 1 cannot be used and the combination of CM and the Systems
B or D is not sufficient.
NOX and SOX REMOVAL
Figure 2 shows schematically the temperature and heat loss of various
systems, FGD, combination of FGD and SCR, simultaneous removal processes,
etc.
No. 1 of Figure 2 shows gas temperatures at regular boiler operation,
which is considered standard (no heat loss). No. 2 illustrates the appli-
cation of FGD with gas reheating from 55 to 75°C, which accounts for a heat
loss of about 1 percent (equivalent to about 1 percent of boiler input).
No. 3 shows the application of FGD to clean the gas before SCR. The
combination causes much heat loss (about 5 percent), and yet is not free
from the ammonium bisulfate problem because FGD, although it removes over
90 percent of the SQz, does not remove SOs sufficiently. Moreover, mist
from FGD aggravates the problems of deposits in the heat exchanger and on
the SCR catalyst.
No. 4 is a modification of No. 3 with a wet electrostatic precipitator
after FGD (ahead of the heat exchanger) to remove SOa and mists and a low
temperature catalyst to reduce heat loss. The low temperature catalyst,
however, may not be resistant to SOX and with its use SOX must be reduced
to 1 ppm (Section 4.9). Usually, sodium hydroxide scrubbing is needed in
addition to a conventional FGD system. Although the gas is cleaned well by
such treatments, the system may be fairly costly.
No. 5 is considered a most practical system for NOX and SOX control.
The gas at 350-400°C from a boiler economizer is treated by SCR and then by
FGD. S0x-resistant catalysts and reactors virtually free from particulate
xiv
-------
plugging have been developed and are now commercially available. Parallel
flow type reactors may be able to treat even a flue gas from a coal-fired
boiler with full particulate load. The ammonium bisulfate problem with the
air preheater may not be serious for flue gas from low-sulfur fuels which
contain little SOx and also for flue gas from coal whose fly ash has a
sweeping effect, provided that NHs at the SCR reactor outlet is kept at a
low concentration, preferably below 5 ppm. The bisulfate may cause some
problem with a flue gas from a high-sulfur oil.
The bisulfate deposits on the SCR catalyst treating S0x-rich flue gas
when the gas temperature drops below about 320°C due to the reduction of
boiler load. The deposit may not be a problem if the low-temperature period
is limited to a few hours a day, because the bisulfate is removed when the
gas temperature is raised above 350°C. For some of the boilers, however, a
device to maintain the temperature above about 320°C, such as a by-pass
system of a hot gas or an auxiliary burner, may be needed.
Another problem for the No. 5 system is the accumulation of ammonia
in the scrubber (or prescrubber) liquor of the FGD system. This may not
be a problem for FGD with ammonia scrubbing. For other FGD processes,
it may be necessary to remove the ammonia from the scrubber liquor, as has
been done at the Owase plant, Chubu Electric.1 The Owase plant has no SCR
system but injects a small amount of NHs between an air preheater and
electrostatic precipitator to prevent precipitator corrosion by SOs. Since
a portion of the NHs goes into a lime-gypsum FGD system and is caught by a
prescrubber, lime addition has been used to strip ammonia from the pre-
scrubber liquor.
The No. 6 system involves the use of a hot electrostatic precipitator.
It may be suitable for a flue gas from a low-sulfur coal whose fly ash is
not caught efficiently by a cold electrostatic precipitator. Either the
xv
-------
parallel flow type or the moving bed type SCR reactor can be used because
the ash content of the gas may be reduced to about 200 mg/Nm by the pre-
cipitator. The ammonium bisulfate deposit problem in the air preheater
may be more serious with system No. 6 than with system No. 5 because the
gas of No. 6 contains less fly ash for the sweeping effect for removing
ammonium bisulfite. The problems of the bisulfate deposit on the catalyst
and ammonia accumulation in the scrubber liquor are the same as with system
No. 5.
No. 7 is a combination of SNR (selective noncatalytic reduction) and
FGD. Since the NHa concentration at the boiler outlet is much higher than
at the SCR reactor outlet, the problems of ammonium bisulfate deposit in
the air preheater and ammonia accumulation in the FGD scrubber are more
serious than with systems 5 and 6. NHa must be removed from scrubbing
liquor to eliminate NHa in wastewater. Consequent heat loss might be
slightly higher than with No. 5 and No. 6. It may be preferable to use
a small amount of SCR catalyst to reduce the NHs leakage and to increase
the NOX removal efficiency, particularly when the process is used with an
FGD system other than ammonia scrubbing.
Systems Nos. 8-10 show dry simultaneous removal processes which are
expected to involve no heat loss, although a considerable amount of energy
is required for plant operation. Both the carbon and the CuO processes
have been considered costly for FGD, but the capability of simultaneous
removal may compensate for the disadvantage. The electron beam process has
been tested in a small pilot plant; larger-scale tests are needed to
evaluate its commercial feasibility.
System No. 11 shows a wet simultaneous removal process with the same
level of heat loss as FGD. The process problems are described in Chapter
7; further improvement is required for commercial application.
xvi
-------
SUITABLE PROCESSES FOR VARIOUS SOX AND NOX CONCENTRATIONS
The relationships of SOx and NOX concentrations in gas to suitable
treatment processes are illustrated in Figure 3. The combination of SCR
followed by FGD can be used for all gas compositions, but there may be an
ammonium bisulfate problem for a gas rich in SOX. SNR should be used for a
gas with a relatively low NOX concentration for which about 45 percent NOX
removal is sufficient and with a low SOx concentration which does not
require FGD, because the NHs leakage may be a considerable problem to FGD.
Dry simultaneous removal processes suit gases with a relatively small
SOx/NOx ratio for the following reasons:
1) For the carbon process, carbon consumption increases with SOx
concentration.
2) For the metal oxide processes such as the Shell process, a
high SOX concentration increases hydrogen and steam consumption
and also tends to lower NOX removal efficiency (see Section
6.5.2).
3) For the electron beam process, the by-product is a lower
quality fertilizer when the gas is rich in SOX.
Use of the carbon and electron beam processes may be restricted to
gases containing less than about 400.ppm SOX. For the Shell process, about
70 percent NOX removal may be a practical limit unless the amount of catalyst
(acceptor) is increased considerably.
On the other hand, wet simultaneous removal processes suit a gas rich
in S02 and relatively low in NOX because an S0a/N0x mole ratio of 4-5 is
usually required to achieve over 80 percent NOX removal. Moreover, equip-
ment for conversion of FGD units to accommodate NOX removal is smaller and
xvii
-------
less costly with a low NOX concentration. Application of the oxidation-
absorption-reduction process may be limited to gases containing less than
about 150 ppm of NO because of the cost of the oxidizing agent. Absorption-
reduction processes may be used for gases with relatively high NOX concen-
trations. Although 90 percent NOx removal can be achieved by the wet
processes, 70-80 percent removal may be more practical. Ninety percent
•
removal requires a large L/G ratio or an excessive amount of oxidizing
agent, which tends to form a hard-to-treat nitrate wastewater.
COSTS OF NOX AND SOX REMOVAL
Figures 4 and 5 show investment costs in battery limits of NOX and SOX
removal plants with capacities of up to 1,000,000 Nm3/hr of flue gas from a
boiler, equivalent to 330 MW with oil and 280 MW with coal. Figure 6 shows
annualized operating costs for 8000 hours of yearly operation including 7
years of depreciation. It is assumed that total investment costs are 50
percent more than the investment costs in battery limits. The costs are
based on the investigations by Environment Agency2 with modifications by
Ando based on his study and represent low costs which may be attained. The
costs per kW or kWhr for flue gas from coal may be about 20 percent more
than those for flue gas from oil because of the larger gas volume per kW.
The costs for SCR by the direct process, that is, treating directly
the flue gas from a boiler economizer, are 3000-4000 yen/kW ($15-20/kW)
investment costs and 0.3-0.4 yen/kWhr ($0.00015-0.00020/kWhr) operating
cost with a 200-300 MW size plant for flue gas from an oil-fired boiler,
and are about one-fourth those for FGD.
The costs for SCR with gas heating by a heat exchanger and a heater
are about double those for direct SCR and are about half those for FGD.
xviii
-------
8
<= 6
0)
4-
O
V)
C
o 5
•r- a
CO
3 4
o
-M
C
Ol
TOO
i
Capacity (MW.with oil)
200 300
200
400
I
100 200
Capacity (MW, with coal)
300
FGD+SCR
SCR+FGD
FGO
SCR
(gas heating)
SCR(Retrofit)
SCR(New)
SNR(Retrofit)
SNR(New)
400
600
800
1000 1200
Capacity (1,000 NmJ/hr)
Figure 4. Investment cost in battery limits of NOX and SOX removal plants.
xix
-------
-J400
bak
Q.
A
Q
£
'i
J. 200
0*
sz
n
u
C
•
SCR
SNR
)
SCR + FGD or
Dry Simultaneous Removal
Metal Oxides (CuO, etc.)
^
I1 ,'''
c §1 ^"^
o c .. "^
•e o ^
U ^il x'
Uj 1 ^.
r
L _l — — — _ — —
FGD
I i i
500 1000 1500
x
^s
s'
,' SCR + FGD or
xx Wet Simultaneous
Absorption Reduction
Oxidation Reduction
1
2000 2
Removal
1
500
SOX (mainly S02, ppm)
Figure 3. Suitable processes for various gas compositions.
XX
-------
10
8
SL
JC
I
VI
o
0)
Capacity (MM, with oil)
100 200 300
100 200 300
Capacity (MW, with coal)
-30
30-
o
-C
-20
o
o
o
I/)
o
-10
10-
SCR (Gas heating)
SCR (Direct, Retrofit)
SCR(Direct, new)
SNR(Retrofit)
SNR(New)
0)
J!
(U
200 400 600 800 1000
Capacity (1,000 Nms/hr)
Figure 5. Investment cost In battery limits per unit capacity
of NOX and SOX removal plants.
xxl
-------
1.0-
•12
0.8-
-10
0.6--8
•o
-------
The costs for SNR are half to one-third those for direct SCR, but NOX
removal efficiency is about half that of SCR (40-45 percent versus 80-90
percent). The problems with SNR due to the larger amount of ammonia leakage
might add some actual operation cost.
The combination of SCR followed by FGD, as shown in No. 5 of Figure 2,
costs a little more than the sum of the costs for direct SCR and FGD because
removal of the ammonia from the FGD scrubber liquor is required. The costs,
however, are still considerably lower than those for the combination of FGD
followed by SCR, which requires gas reheating as shown in No. 3 of Figure 2.
The costs for the dry and wet simultaneous removal processes are as yet
uncertain but seem to be close to those for the combination of FGD followed
by SCR.
CONCLUSIONS
There has been considerable progress with combustion modification.
It currently can reduce NOx in flue gases from utility boilers to 50 ppm
for gas burning, 100 ppm for oil burning, and 200 ppm for coal burning.
There have also been considerable advances with selective catalytic
reduction (SCR) to the extent that S0x-resistant catalysts and reactors
free from particulate plugging have been commercialized. By using the
direct process to treat directly the flue gas from a boiler economizer, the
SCR cost has been reduced to about one-fourth the FGD cost. Operation of
SCR is easy and can be controlled from a boiler control room by one of the
boiler operators. The ammonium bisulfate problem has been nearly solved
for flue gases from coal and low-sulfur oil, although further studies are
needed particularly for flue gas from a high-sulfur oil.
xxiii
-------
As a result of these advances, many large SCR plants have been
constructed mainly for boilers which burn gas and low-sulfur oil. A
demonstration plant for an existing coal-fired boiler (about 250 MW) will
be constructed in the near future. Although SCR can achieve over 90 percent
NOX removal without difficulty, it may be appropriate to reduce the amount
of ammonia and obtain 80-85 percent removal with ammonia leakage kept at
less than about 5 ppm. The pressure drop of the gas in the reactor may be
kept at 40-60 mmH20.
Selective noncatalytic reduction (SNR) may be useful for an existing
boiler for which 40-50 percent NOx removal is sufficient and which has no
room for installation of an SCR reactor or no excessive fan capacity to
compensate for the pressure drop caused by the reactor. Since there is a
considerable amount of ammonia leakage with SNR, it may be desirable to
place a small amount of SCR catalyst (parallel flow type) in a duct at
350-400°C to reduce the ammonia leakage and to increase the NOX removal
efficiency. With this combination, an NOX removal efficiency of 60-70
percent with 10 ppm of leak NH3 may be attained with a pressure drop of
15-30 mmH20.
The commercial applicability of the simultaneous removal processes,
dry and wet, to remove over 80 percent of the NOX and SOz has not been
confirmed yet because the processes are still undergoing pilot plant tests.
There should be no heat loss with the dry simultaneous removal processes,
as shown in Nos. 8-10 of Figure 2. These processes, however, require
considerable energy for plant operation and production of reducing agents.
Since all of these dry NOX removal and simultaneous removal processes
require ammonia, it is possible that the world supply of ammonia will run
short when these processes are applied extensively in many countries. On
the other hand, wet simultaneous removal processes can convert NOX to NHs.
Although there are problems with the wet simultaneous processes as already
mentioned, further studies may solve the problems.
xxiv
-------
REFERENCES
1. Ando, J., SOa Abatement for Stationary Sources in Japan. EPA 600-7-78/
210, November 1978.
2. Report on NOX Abatement Technology. Air Preservation Bureau, Environ-
ment Agency, April 1978.
xxv
-------
CONTENTS
Acknowledgements ii
Introduction ill
Summary v
Figures xxxiv
Tables xlvii
Conversion Factors and Abbreviations li
1. NOX CONCENTRATION, REGULATION AND ABATEMENT IN JAPAN 1
1.1 .Introduction . 1
1.2 NOX Emissions, N02 Ambient Concentrations, and Abatement.. 4
1.2.1 NOX Emissions 4
1.2.2 Ambient NOz Concentrations and Abatement Standards. 7
1.3 Reinforcement of NOX Emission Control for Passenger Cars.. 10
1.4 Countermeasures for Stationary Sources 14
1.4.1 Methods of NOX Abatement 14
1.4.2 Emission Standards 16
1.5 Cost Estimates for NOX Abatement for Stationary Sources
and Limit of Abatement 18
1.5.1 Estimate by MITI 18
1.5.2 Estimates by Environment Agency 25
1.6 Revision of Ambient NOa Standard 29
1.6.1 Consultation for Criteria and Guidelines for NOa
That Has Adverse Effects on Human Health 29
1.6.2 New NO2 Standard 31
1.7 NOX Abatement by Industries 33
1.7.1 Introduction 33
1.7.2 NOX Abatement for Utility Boilers 33
xxvi
-------
CONTENTS (Continued)
1.7.3 Chubu Electric's Chita Power Station 36
1.7.4 Efforts by the Steel Industry 37
1.7.5 Nitrogen Removal from Coke 39
1.7.6 Flue Gas Denitrification by the Steel Industry 42
1.7.7 NOX Abatement by Other Industries 44
1.8 References 45
2. NOX ABATEMENT BY COMBUSTION MODIFICATION (CM)
2.1 Classification of Combustion Modification Techniques 46
2.2 Change of Operating Conditions 47
2.2.1 Low-Excess-Air Combustion (LEA) 47
2.2.2 Improvement of Fuel/Air Mixing Conditions in
Combustion Chamber 48
2.2.3 Reduction of Heat Release Rate in the Combustion
Chamber (Reduction of Output Power) 49
2.2.4 Reduction of Air Preheating 49
2.3 Modification of Burner Design 49
2.3.1 Introduction 49
2.3.2 Atomiz ing Type Burners 50
2.3.3 Divided-Flame Type Burners 52
2.3.4 Self-Recirculation Type Burners 52
2.3.5 Staged-Combustion Type Burners 57
2.3.6 Combination Type Burners 64
2.4 Modification of Combustion System Design 68
2.4.1 Reduction of Heat Release Rate in the Combustion
Chamber (Increase of Furnace Volume) 68
2.4.2 Staged Combustion 68
2.4.3 Flue Gas Recirculation (FGR) 70
2.4.4 Water or Steam Injection 70
2.5 Other Methods 72
2.5.1 Change of Fuel 72
2.5.2 Modification of Firing 72
2.5.3 Combination Techniques 73
xxvii
-------
CONTENTS (Continued)
2.6 Commercial Application of Combustion Modification (CM).,.. 74
2.6.1 Boilers 74
2.6.2 Oil Heating Furnaces 79
2.6.3 Metal Heating Furnaces 83
2.7 References 87
3. GENERAL DESCRIPTION OF SELECTIVE CATALYTIC REDUCTION (SCR) 88
3.1 Introduction 88
3.1.1 Basic Reactions 88
3.1.2 Major SCR Plants 90
3.1.3 Operating Parameters • 91
3.2 Composition and Behavior of Catalysts 97
3.2.1 Introduction 97
3.2.2 Behavior of A1203 Carrier 98
3.2.3 Ti02-Based Catalysts 99
3.2.4 Vanadium Catalysts 103
3.2.5 Ammonia Decomposition Catalysts 106
3.2.6 Problems and Further Studies Required 107
3.3 Catalyst Shape and Reactor 108
3.3.1 Introduction 108
3.3.2 Moving Bed and Parallel Flow Reactors 115
3.3.3 Comparison of Parallel Flow Type Catalysts and
Reactors 120
3.4 Commercial Honeycomb Catalysts 124
3.4.1 Introduction 124
3.4.2 Honeycomb Catalysts Produced by Catalyst and
Chemicals Ind. Co 124
3.4.3 Honeycomb Catalysts Produced by NGK Insulators Ltd. 125
3.4.4 Honeycombs Made by Other Producers 131
3.5 Problems of Ammonium Bisulfate Formation 132
3.5.1 Introduction 132
3.5.2 Formation of SOs 132
3.5.3 Formation of Ammonium Bisulfate and Sulfate 133
xxviii
-------
CONTENTS (Continued)
3.5.4 Reaction of Ammonium Bisulfate 139
3.5.5 Deposit in Heat Exchanger and Air Preheater 140
3.5.6 Corrosion Tests 147
3.5.7 Countermeasures for the Bisulf ate Problem 148
3.6 Cost of SCR (Investigation by Environment Agency) 153
3.6.1 Assumptions for Cost Estimation 153
3.6.2 Equipment Cost and Annual Cost 155
3.6.3 Factors that Affect Cost 159
3.7 References 168
4. SCR PROCESSES AND PLANTS 169
4.1 Sumitomo Chemical SCR Processes 169
4.1.1 Introduction 169
4.1.2 Commercial Plant for Clean Gas Treatment 169
4.1.3 Dirty Gas Treatment 172
4.1.4 New Catalysts 176
4.1.5 Economics 176
4.1.6 Evaluation 180
4.2 Hitachi Zosen (Shipbuilding) Process 180
4.2.1 Introduction 180
4.2.2 Catalysts 181
4.2.3 Commercial Plants 187
4.2.4 Economics 191
4.2.5 Evaluation 200
4.3 Hitachi Ltd. Process 201
4.3.1 Introduction 201
4.3.2 Catalysts and Reactors 203
4.3.3 Pilot Plant Tests 211
4.3.4 Kainan Plant, Kansai Electric 214
4.3.5 Chita Plant, Chubu Electric 216
4.3.6 Economics.... 216
4.3.7 Evaluation 218
xx ix
-------
CONTENTS (Continued)
4.4 JGC Process (Paranox Process) 219
4.4.1 Introduction 219
4.4.2 Parallel Passage Reactor 220
4.4.3 Catalyst 222
4.4.4 Commercial Plants 222
4.4.5 Pilot Plant Tests with Sintering Machine Flue Gas.. 229
4.4.6 Prototype Plant for Coke Oven Flue Gas 234
4.4.7 Economics •• 234
4.4.8 Evaluation 234
4.5 Mitsui Engineering and Shipbuilding Process 235
4.5.1 Introduction 235
4.5.2 Chiba Plant, Mitsui Petrochemical 236
4.5.3 Chiba Plant, Ukishima Petrochemical 236
4.5.4 Tests on Tubular Catalysts 241
4.5.5 Economics 244
4.5.6 Evaluation 248
4.6 Mitsubishi Heavy Industries Process (MHI Process) 249
4.6.1 Introduction 249
4.6.2 Chiba (Sodegaura) Plant, Fuji Oil 249
4.6.3 Parallel Plate Catalyst 252
4.6.4 Cooperative Tests with Tokyo Electric Using Plate
Catalyst 254
4.6.5 Large Plants for Utility Boilers 257
4.6.6 Evaluation 257
4. 7 IHI PROCESS 258
4.7.1 Introduction 258
4.7.2 Pilot Plant Tests at Power Stations 258
4.7.3 Large-Scale Plants 265
4.7.4 Economics 268
4.7.5 Evaluation 268
4.8 Kurabo Process (Knorka Process) 269
xxx
-------
CONTENTS (Continued)
4.8.1 Introduction 269
4.8.2 Process Description 270
4.8.3 Economics 273
4.8.4 Recent Studies 276
4.8.5 Pilot Plant for Coke Oven Flue Gas (Kurabo-Niigata
Process) 276
4.8.6 Evaluation..., 279
4.9 Kureha Process 279
4.9.1 Introduction 279
4.9.2 Process Description 280
4.9.3 Economics 282
4.9.4 Evaluation 282
4.10 Kobe Steel Process 283
4.10.1 Introduction 283
4.10.2 Pilot Plant Test with Moving Bed 284
4.10.3 Pilot Plant Test with Honeycomb Catalyst 284
4.10.4 Evaluation Y. 288
4.11 Other SCR Processes 288
4.11.1 Mitsui Toatsu Process 288
4.11.2 Sumitomo Heavy Industries SCR Process 291
4.11.3 Nippon Kokan Iron Ore Catalyst Process 298
4.11.4 Mitsubishi Kakoki Kaisha (MKK) Process 300
4.11.5 Kawasaki Heavy Industries 301
4.11.6 Asahi Glass Products 302
4.12 References 303
5. SELECTIVE NONCATALYTIC REDUCTION (SNR) 304
5.1 Introduction 304
5.2 Laboratory Tests at Okayama University 308
5.3 Studies by CRIEPI 312
5.4 Large Scale Test of SNR at Chita Station, Chubu Electric.. 326
5.4.1 Introduction 326
xxxi
-------
CONTENTS (Continued)
5.4.2 Basic Study by Chubu Electric 326
5.4.3 Basic Study by MHI 333
5.4.4 Large Scale Plant Tests 337
5.4.5 Economics of the Chita Plant 345
5.5 Mitsui Petrochemical, Chiba Plant 345
5.6 Other SNR Plants 346
5.7 Comparison of SCR, SNR and the Combined Process 347
5.8 Evaluation 349
5.9 References 351
6. OTHER DRY PROCESSES 352
6.1 Introduction 352
6.2 Reaction of Activated Carbon with NOx 353
6.2.1 Introduction 353
6.2.2 NOX Adsorption and Desorption 355
6.2.3 Catalysts for Ammonia Reduction 358
6.2.4 Simultaneous Removal of SOX and NOX 359
6.3 Sumitomo Heavy Industries Carbon Process for Simultaneous
Removal 361
6.3.1 Introduction 361
6.3.2 Operation of the 2000 Nm3/hr Pilot Plant 362
6.3.3 Cost Estimation 364
6.3.4 Evaluation 369
6.4 Unitika Activated Carbon Process for Simultaneous Removal. 370
6.4.1 Introduction 370
6.4.2 Process Description 370
6.4.3 Cost Estimation 372
6.4.4 Evaluation 373
6.5 Shell Copper Oxide Process (SFGD Process) for Simultaneous
Removal 373
6.5.1 Process Description 373
6.5.2 Evaluation 377
xxxii
-------
CONTENTS (Continued)
6.6 Ebara Electron Beam Radiation Process for Simultaneous
Removal 381
6.6.1 Process Description 381
6.6.2 Evaluation 384
6.7 Other Dry Processes 384
6,7.1 Tests on Catalytic Decomposition of NO 384
6.7.2 NOX Recovery by Molecular Sieve 385
6.8 References 386
7. WET PROCESSES FOR NOX AND SIMULTANEOUS NOX AND SOX REMOVAL 387
7.1 General Description 387
7.1.1 Major Processes and Plants 387
7.1.2 Oxidizing Agents 389
7.1.3 Oxidation-Absorption and Absorption-Oxidation
Processes 390
7.1.4 Oxidation-Absorption-Reduction and Absorption-
Reduction Processes (Simultaneous Removal) 391
7.2 MHI Wet Processes 395
7.2.1 Introduction 395
7.2.2 Nitric Acid By-Production Process 395
7.2.3 Simultaneous Removal Process (Oxidation-Absorption-
Reduction Process) 397
7.2.4 Evaluation 402
7.3 IHI Oxidation-Absorption-Reduction Process (Simultaneous
Removal) 403
7.3.1 Introduction 403
7.3.2 Process Description 403
7.3.3 Economics 408
7.3.4 Evaluation 408
7.4 Sumitomo-Fujikasui Process (Moretana Process) 410
7.4.1 Introduction 410
7.4.2 Cl02-CaC03 Process 410
7.4.3 Evaluation 413
xxxiii
-------
CONTENTS (Continued)
7.5 Chisso Engineering Absorption Reduction Process (CEC
Process) for Simultaneous Removal 414
7.5.1 Introduction 414
7.5.2 Process Description *. 414
7.5.3 Economics 418
7.5.4 Evaluation 418
7.6 Asahi Chemical Process 420
7.6.1 Introduction 420
7.6.2 Process Description 420
7.6.3 Evaluation 425
7.7 Kureha Chemical Wet Simultaneous Removal Processes 425
7.7.1 Introduction 425
7.7.2 By-Product Gypsum and Ammonia Process 426
7.7.3 By-Product Sulfur and Ammonia Process 429
7.7.4 Evaluation 432
7.8 Chiyoda Thoroughbred 102 Process 433
7.8.1 Introduction 433
7.8.2 Process Description 434
7.8.3 Economics 437
7.8.4 Evaluation 439
7.9 Other Wet Processes 439
7.9.1 Permanganate Processes 439
7.9.2 Equimolecular Absorption 440
7.9.3 Others 440
7.10 References 441
xxxiv
-------
FIGURES
Number Page
1 Systems for SCR catalyst installation ..................... viii
2 Combined and simultaneous removal systems (figures show
temperatures, °C) ......................................... x
3 Suitable processes for various gas compositions ........... xvi
4 Investment cost in battery limits of NOX and SOX removal
plant s [[[ xviii
5 Investment cost in battery limits per unit capacity of
NOX and SOX removal plants ................................ xix
6 Annualized operation costs for NOx and SOX removal plants. xx
1-1 Comparison of U.S. and Japanese ambient NOa standards and
yearly average concentrations of 6 Japanese stations ...... 2
1-2 Comparison of U.S. and Japanese ambient SOa standards and
yearly average concentrations of 15 Japanese stations ..... 2
1-3 Comparison of yearly average NOa concentrations in major
cities of the world ....................................... 3
1-4 NOX emissions from Japanese stationary sources controlled
by regulation in 1975 ..................................... 5
1-5 Contribution of NOX sources to ground level
concentrations . ..................... . ...... ............... 7
1-6 1975 ambient NOz concentrations and the decrease by NOx
abatement for 80 areas of the 120 city areas ........... . .. 9
1-7 Estimated total amount of NOX emissions from road traffic
in the Tokyo Bay Area .............................. . ...... 13
1-8 Investment costs for SCR denitrif ication plants estimated
by MITI [[[ 21
1-9 Total investment costs required between 1976 and 1985 to
reach different N02 concentrations ........................ 23
1-10 Annual costs in 1985 needed to reach different N02
concentrations (daily averages) ........................... 23
1-11 Energy consumption required in 1985 to reach different
-------
FIGURES (Continued)
Number
1-12 Investment cost (I) and annual cost (A) for each case 27
1-13 Ambient NOa concentrations in 1975 and 1985 with
countermeasures a, b, c and d 30
1-14 Flowsheet of NR process (20 t/day) 40
1-15 Decrease of N in coal by heating (retention time is 2
hours) 41
1-16 NOX abatement of flue gas from sintering machine by N
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
2-17
2-18
Effect of NFK-TRW Burner on NOX
NOX reduction by divided-flame type burner and other
Effect of self-recirculation type burner on NOX reduction.
XB Burner
NOX emission levels from a low-NOx burner more advanced
Dual-air register type burner for use with pulverized
NOX emission levels of the burner shown in Figure 2-17
41
50
51
53
54
55
55
56
57
58
58
60
60
61
61
62
62
63
63
2-19 Atomizing nozzles in off-stoichiometric combustion-type
low-NOx burner for oil , 65
xxxvi
-------
FIGURES (Continued)
Number Page
2-20 Effect of low-NOx atomizer shown in Figure 2-19 on NOX
emissions ................................................. 65
2-21 Schematic of SRG burner ................................ ,. . . 66
2-22 Effect of SRG burner on NOX reduction .................. ... 66
2-23 TCG burner ................................................ 67
2-24 Two-stage combustion for small boiler ..................... 69
2-25 Flow pattern in reversely turned firing. . . ................ 73
2-26 Cost performance of NOX control techniques applied to
boilers ............................................ . ...... 80
3-1 Effect of oxygen on NOX removal (SV, 10,000 hr"1 ; NOX,
100 ppm; NH3/NOX, 1.0) .................................... 89
3-2 Criteria for catalyst for clean gas ....................... 90
3-3 Typical operation data of SCR for an oil-fired boiler at
350-400°C using granular catalyst ......................... 94
3-4 Results of clean gas treatment at different boiler loads
(NH3 /NOX = 1.0) ........................................... 96
3-5 Deactivation of catalyst on y-AlzOa carrier by SOX ........ 98
3-6 Weight increase in AlaOs (pre-calcined at 600-1100°C)
heated in gas containing SO 2 at rate of 2°C/min ........... 101
3-7 Comparison of NOX removal from inlet gas with several
TiOa based catalysts (calcined at 550°C for 5 hours, 1 mm
size) [[[ 102
3-8 Comparison of NOX removal from inlet gas with several
catalysts (calcined at 550°C for 5 hours, 1 mm size) ...... 104
3-9 Results of life tests with T 10 2 -based catalysts ........... 105
3-10 Effects of an ammonia converter on NOX removal from inlet
gas with ammonia decomposition catalysts .................. 106
3-11 Types of reactors ................ . ..... . .................. 110
3-12 Moving bed reactor ........................................ 112
3-13 Change of pressure drop with an intermittent moving bed... 113
3-14 Cross sections of parallel flow catalysts (actual sizes).. 114
-------
FIGURES (Continued)
Number Page
3-16 Example of a fixed bed reactor with honeycomb type
catalyst (Ishikawajima-Marima Heavy Industries; sizes are
in mm) 121
3-17 Honeycomb and reactor produced by Catalyst and Chemicals
Ind. Co 126
3-18 NOX removal at 375°C (NO = 160-173 ppm, NH3/NO = 1) 127
3-19 NOX removal at 340°C (NO = 176-184 ppm, NH3/NO = 1) 128
3-20 NOX removal at 300°C (NO = 163-178 ppm, NH3/NO = 1) 129
3-21 Durability of NOX removal catalyst for exhaust gas of
high-sulfur oil burning boiler 130
3-22 Formation ratio of S03 (Oa in gas = 1-2 percent) 134
3-23 Relation of Oa and S03 concentrations (heavy oil, S =
1-3 percent) 134
3-24 Relationship of SOs concentration to particulate content
(heavy oil, S = 1-3 percent; Oa in gas = 0.3-3 percent)... 135
3-25 NH3-H2SOit system 135
3-26 Formation temperatures of NH^HSOit 137
3-27 Formation temperatures of NH^HSOif (NH3 + S03 + H20 t
NHUHSCM 137
3-28 Formation temperatures of ammonium bisulfate and sulfate
(1) 138
3-29 Formation temperatures of ammonium bisulfate and sulfate
(2) 138
3-30 Schematic of Ljungstrom type air preheater 141
3-31 Temperatures of gas and heating elements in a commercial
Ljungstrom heat exchanger 142
3-32 Photographs of a Ljungstrom air preheater element 144
3-33 Results of tests on flue gas with a full dust load 145
3-34 Deposits in air preheater tubes (numbers show temperatures
of gas, air and tube) 146
3-35 Corrosion tests on steels 151
3-36 SCR systems used in cost estimation 154
3-37 Investment cost of SCR by direct process (System A) 156
xxxviii
-------
FIGURES (Continued)
Number Page
3-38 Investment cost of SCR by direct process (System B) 157
3-39 Investment cost of SCR with gas reheating and heat
recovery (System C) 158
3-40 Average plant costs per capacity for Systems A, B and C... 160
3-41 Average plant costs for Systems A, B and C 161
3-42 Annual cost of SCR by direct process (System A) 162
3-43 Annual cost of SCR by direct process (System B) 163
3-44 Annual cost of SCR process with gas reheating and heat
recovery (System C) 164
3-45 Average annual costs of SCR with Systems A, B and C 165
4-1 Flowsheet of the denitrification process at the Higashi
Nihon Methanol Company 170
4-2 Operation data of the HNM plant and two other SCR plants.. 173
4-3 Flowsheet of two Sumitomo Chemical commercial plants for
dirty gas 175
4-4 NOx removal efficiency vs. reaction temperature for
Catalysts A and B 177
4-5 SOX resistivity of Catalysts B and C 178
4-6 Apparatus for a test of the forced deterioration of a
catalyst by SOs 182
4-7 NOX removal efficiency of Noxnon 203, Noxnon 303 and
Noxnon 403 after forced deterioration of the catalyst by
S03 183
4-8 Structure of Noxnon 500 184
4-9 Pressure drop over 1 m of packed bed (Noxnon 500) 185
4-10 Initial efficiency of Noxnon 500 at various temperatures.. 186
4-11 Initial efficiency of Noxnon 500 at various NHs/NOx mole
ratios 186
4-12 Efficiency of Noxnon 500 as a function of time 186
4-13 Flowsheet of the Yokkaichi plant of Shindaikyowa
Petrochemical 190
4-14 Hitachi Zosen systems assumed for a cost estimate 192
4-15 Estimated cost for SCR by Systems I-a and I-b 197
xxx ix
-------
FIGURES (Continued)
Number Page
4-16 Estimated cost of SCR followed by FGD (System Il-a) 198
4-17 Estimated cost of FGD followed by SCR (System Il-b) 199
4-18 Flow diagram of the direct system for boiler flue gas at
Hitachi, Ltd 204
4-19 Catalyst size versus dust deposition for oil-fired dust... 206
4-20 SV versus NOX removal and NHs leak (ring type catalyst,
15 mm diameter; 350°C; NH3/NO, 1.0; inlet NOX, 250 ppm)... 206
4-21 Effect of NH3/NOX mole ratio on the performance of a ring
catalyst (15 mm diameter; SV, 6200 hr"1; 350°C; inlet
NOX, 250 ppm) 207
4-22 Effect of temperature on the performance of a ring
catalyst (15 mm diameter; 350°C; SV, 6200 hr"1; NH3/NO,
1.0; inlet NOX, 250 ppm) 207
4-23 Characteristic curve of space velocity versus NOX removal
efficiency by a plate catalyst 208
4-24 Reaction temperature versus NOx removal for a plate
catalyst 209
4-25 NH3/NOX mole ratio versus NOX removal for a plate catalyst
at 350°C, LV = 5.9 m/sec 209
4-26 Structure of parallel plate reactor 210
4-27 Test results for ring catalysts used on gas and oil fired
burners 212
4-28 Tests of a moving bed reactor with a 2000 Nm3/hr capacity
for coal-fired boiler flue gas after a hot electrostatic
precipitator 213
4-29 Flowsheet of the Kainan SCR plant 215
4-30 Flowsheet of the Chita SCR plant 217
4-31 Construction of a parallel passage reactor 221
4-32 The effect of temperature on the activities of catalysts
JP 501 and JP 102 (NO, 200 ppm; NH3, 240 ppm; 02, 3
percent; H20, 10 percent; SV, 20,000hr~1) 223
4-33 The effect of NH3/NO mole ratio on the activities of
catalysts JP 501 and JP 102 (380°C; SV, 20,000 hr-1;
NO, 200 ppm; S02, 1500 ppm; 02, 3 percent; N2, 10
percent) 223
xl
-------
FIGURES (Continued)
Number Page
4-34 The change in ,NOx removal efficiency with reaction
temperature and time for catalyst JP 501 224
4-35 Paranox process for Kashima Oil, 50,000 Nm3/hr 226
4-36 Paranox process for Fuji Oil, 70,000 Nm3/hr 228
4-37 Results of JGC testing of a high temperature catalyst at
Nagoya Works, Nippon Steel, for flue gas after an
electrostatic precipitator 231
4-38 Results of JGC testing of a high temperature catalyst at
Nagoya Works, Nippon Steel, for flue gas before an
electrostatic precipitator 232
4-39 Results of the third run of JGC testing of a low
temperature catalyst at Nagoya Works, Nippon Steel, for
flue gas after an electrostatic precipitator 233
4-40 Flowsheet of Chiba Plant, Mitsui Petrochemical 237
4-41 Flowsheet of Chiba Plant, Ukishima Petrochemical 239
4-42 Structure of a reactor with a tubular catalyst 240
4-43 Methods for packing of tubular catalysts 242
4-44 Effect of catalyst packing method on NOX removal and
pressure drop (fixed gas velocity, varying catalyst
volume 243
4-45 SV versus NOX removal for two diameters of tubular
catalyst (350°C, separate packing) 245
4-46 Performance of the tubular catalyst with 20 mm diameter
(14 mm inner diameter) at 350°C and SV of 4000 hr"1 246
4-47 Flowsheet of the SCR plant built by MHI at the Chiba
Plant, Fuji Oil 250
4-48 Packing of a honeycomb catalyst at the Chiba Plant, Fuji
Oil 251
4-49 MHI's tests with parallel plate catalyst at 5 pilot
plants 253
4-50 Flowsheet of the pilot plant at Yokosuka Station, with
plate catalyst 255
4-51 Gas temperature versus denitrification efficiency for the
moving bed reactor at Shinnagoya Power Station 260
xli
-------
FIGURES (Continued)
Number Page
4-52 SV value versus denitrification efficiency for the moving
bed reactor at Shinnagoya Power Station 260
4-53 NHa/NOx mole ratio versus denitrification efficiency and
reactor outlet ammonia concentration for the intermittent
moving bed reactor at Shinnagoya Power Station 261
4-54 Relation between boiler load and denitrification
efficiency for the intermittent moving bed reactor at
Shinnagoya Power Station 262
4-55 NHa/NOx mole ratio versus denitrification efficiency
versus reactor outlet ammonia concentration for the
honeycomb catalyst at Taketoyo Power Station 263
4-56 Catalyst life test results for the honeycomb catalyst
at Taketoya Power Station 264
4-57 Standard flowsheet for the boiler denitrification
equipment at 5 large IHI plants for combustion gas and
from heavy oil 266
4-58 Plant equipment layout (example) of 5 large IHI plants.— 267
4-59 Flowsheet of the Kurabo (Knorka) SCR process 271
4-60 Structure of the Kurabo moving bed reactor 272
4-61 Relationship of NHs/NOx mole ratio to outlet NOx, NHs
concentrations in the Kurabo moving bed reactor 273
4-62 Results of the Kurabo-Niigata process at a pilot plant
for coke oven flue gas 278
4-63 Flowsheet of Kureha SCR process for treating flue gas
discharged from a conventional FGD system 281
4-64 Flowsheet of pilot plant at Kakogawa Works, Kobe Steel.... 285
4-65 Initial activity of Kobe Steel's vanadium catalyst
(SV - 10,000 hr'1) 286
4-66 Initial activity of Kobe Steel's iron catalyst
(SV = 5000 hr"1) 286
4-67 NOX removal efficiency for the first 1000 hours of
operation at the Kakogawa pilot plant with vanadium
catalyst in a moving bed 287
4-68 Change of pressure drop for the first 1000 hours of
operation at the Kakogawa pilot plant with vanadium
catalyst in a moving bed 287
xlii
-------
FIGURES (Continued)
Number
Page
4-69 Performance of catalyst MTC-102 on a flue gas from LPG
burning [[[ 289
4-70 SV versus NOX removal for the catalyst MTC-102 on a flue
gas from LPG burning ...................................... 289
4-71 Performance of catalyst MTC-104 on the flue gas from
heavy oil burning ......................................... 290
4-72 SV versus NOX removal for catalyst MTC-104 on the flue
gas from heavy oil burning ................................ 290
4-73 Flowsheet of Sumitomo Heavy Industry SCR process for
boiler flue gas ........................................... 293
4-74 NOX removal versus temperature for the SHI process
(NH3/NO ratio = 1.0) ...................................... 295
4-75 NOx removal versus NHa inlet concentration versus
leakage for the SHI process at 300°C ...................... 295
4-76 Stacking of three SHI reactors to treat 186,000 Nm3/hr
of gas [[[ 296
4-77 Flowsheet of NKK process for treating the flue gas from
an iron-ore sintering machine ............................. 299
4-78 NOX removal efficiency of MKK catalyst used for SC-2-rich
gas at Nippon Yakin ....................................... 300
4-79 NOX removal efficiency of KHI catalysts used for flue gas
from oil-fired boilers ......... . .......................... 301
5-1 Results of laboratory tests of SNR by five research
organizations .................................. . .......... 307
5-2 Effects of H2 and EzO concentrations on NOx removal rate
and NHs conversion rate at a reaction time of 0.1-0.13
seconds, according to tests at Okayama University ......... 309
5-3 Effects of NHa and Ha concentrations and reaction
temperature on the NOx removal rate and the NHs conversion
rate with a reaction time of 0.1-0.13 second, according
to tests at Okayama University ............................ 310
5-4 Effect .of 02 concentration on the NOX removal rate and
: the NHs conversion rate with a reaction time of 0.1-0.13
second, according to tests at Okayama University .......... 311
5-5 Effect of Ha concentration on decomposition of NHs and
formation of NO with a reaction time of 0.1-0.13 second,
-------
FIGURES (Continued)
Number
5-6 Effect of reaction time, temperature and gas composition
on NOx removal, according to tests at Okayama University.. 314
5-7 Decomposition of NHs and formation of NO of different
gases as different temperatures, according to tests at
Okayama University 315
5-8 The effects of the NHs/NO mole ratio and temperature on
NOX removal, according to CRIEPI studies 316
5-9 Decomposition of NH9, calculated by CRIEPI 318
5-10 Effect of temperature on NHs conversion and NO formation,
calculated by CRIEPI 318
5-11 Effect of Ha on NHs conversion and NO and NHa formation,
calculated by CRIEPI 319
5-12 Furnace and ammonia injection nozzles for SNR by CRIEPI
(sizes are in mm) 320
5-13 The effect of cooling the NHs injection nozzle on a small
CRIEPI furnace 322
5-14 NOX removal by NHs injection at different points on a
small CRIEPI furnace 323
5-15 NOX removal efficiency for the flue gas from the burning
of several fuels in a small CRIEPI furnace 324
5-16 Effect of NH3/NO mole ratio and H2 addition on NOx
removal for the flue gas from oil burning in a small
CRIEPI furnace 325
5-17 Schematic of experimental apparatus for Chuba Laboratory
tests of SNR 327
5-18 NOX reduction versus reaction temperature with Chuba
Electric's SNR 328
5-19 NOX reduction versus residence time in reaction zone
with Chuba Electric's SNR 329
5-20 NOx reduction versus reaction temperature (effect of Ha
addition) with Chuba Electric's SNR 330
5-21 NOX reduction versus reaction temperature (effect of CHi,
addition) with Chuba Electric's SNR 331
5-22 NH9/NO mole ratio and NOX removal with Chuba Electric's
SNR 332
xliv
-------
FIGURES (Continued)
Number Page
5-23 Inlet and outlet NHs at different temperatures with
Chuba Electric' s SNR 332
5-24 Laboratory data (effect of gas temperature) for the
MHI-MCI Joint study at Chiba Station. 334
5-25 Laboratory data (effect of residence time) for the MHI-
MCI joint study at Chlba Station 335
5-26 A calculation result by simulation on practical boiler.... 336
5-27 Locations of nozzles in the No. 2 Boiler at Chita Station. 338
5-28 NHs injection nozzle in the No. 2 Boiler at Chita Station. 339
5-29 Schematic flow of boiler water circulation for nozzle
cooling in the No. 2 Boiler at Chita Station 340
5-30 Concept of nozzle cooling systems on the No. 2 Boiler
at Chita Station. 341
5-31 Effects of NHs injection position and load on NOx
reduction rates on the operation of the No. 2 Boiler at
Chita Station 342
5-32 Data on the designed operation conditions and anticipated
results at different operation loads of the No. 2 Boiler
at Chiba Station (Mitsubishi Heavy Industries) 343
5-33 Comparison between test results and predicted performance
of the No. 2 boiler at Chiba Station 344
6-1 Desorption of NOX adsorbed by activated carbon (6 mm in
diameter) by washing with water at different temperatures. 356
6-2 Desorption of NOX adsorbed by activated carbon by heating
at different temperatures 356
6-3 Removal efficiencies of activated carbon 360
6-4 Flowsheet of a 2000 NmVhr pilot plant using the SHI
carbon process for simultaneous removal of SOX and NOX.... 363
6-5 Continuous operation tests of the SHI carbon process for
simultaneous removal of SOX and NOx 365
6-6 Effect of temperature on simultaneous removal of SOX and
NOX with carbon (SOz, 150-250 ppm; NOX, 150-200 ppm; NHs,
370 ppm; SV - 1000 hr"1) 366
xlv
-------
FIGURES (Continued)
Number Page
6-7 Effect of space velocity on simultaneous removal of SOX
and NOX with carbon (S02, 150-200 ppm; NOX, 150-200 ppm;
NH3 , 370 ppm; 220°C) 366
6-8 Effect of NHs added on simultaneous removal of SOX and
NOx with carbon (NOX, 150-200 ppm; S02, 150-250 ppm; SV,
1000 hr'1 ; 220°C) 367
6-9 Effect of S02 on simultaneous removal of SOx and NOX
with carbon (NOX, 150-200 ppm; NHa , 370 ppm; SV, 1000
hr'1 ; 220°C) 367
6-10 Flowsheet of the Unitika process for simultaneous removal
of SOX and NOX 371
6-11 NOX slip> percent of intake, versus loading, CuSOi*/
(CuO + CuSOO mole ratio, for a commercial SFGD acceptor/
catalyst 376
6-12 NOX conversion versus NHs/NO mole ratio for a commercial
SFGD acceptor (SYS) 378
6-13 Unconverted NOX as a function of catalyst bed length for
a commercial SFGD acceptor (SYS) - upper and lower limits. 379
6-14 Performance of Shell FGD reactor at SYS, instantaneous
S02 and NOX slip (92) 380
6-15 Flowsheet of the 1000 Nm3/hr electron beam radiation
process pilot plant of Ebara Manufacturing Company 382
6-16 Removal ratios of SOX and NOX versus total beam for the
electron beam radiation process used at the Ebara
Manufacturing Company pilot plant 383
7-1 NO absorption in EDTA-Fe(II) liquor (0.01 mole/liter) 392
7-2 EDTA-Fe(II) concentration and NO absorption at 50°C 393
7-3 Flowsheet of Tokyo Electric-MHI oxidation absorption
process 396
7-4 MHI simultaneous removal process 399
7-5 Flowsheet of the 1HI process 404
7-6 Effect of CaClz and NaCl concentration on NOX removal
efficiency in the IHI process 406
7-7 Effect of pH on SOX and NOX removal efficiency in the IHI
process 407
xlvi
-------
FIGURES (Continued)
Number Page
7-8 Investment and operation costs , 409
7-9 Flowsheet of the Sumitomo Fujikasui ClOa-CaCOs process.... 411
7-10 Flowsheet of the CEC process 415
7-11 Decomposition rates of intermediates in the CEC process... 417
7-12 Flowsheet of the sodium process of Asahi Chemical 421
7-13 Solubilities of sodium and potassium imidodisulfonates as
a function of temperature 424
7-14 Block flow diagram of the Kureha NDSN gypsum and ammonia
by-product process for simultaneous removal of SOX and
NOX 427
7-15 Flowsheet of the Kureha KDSN sulfur and ammonia by-
production process for the simultaneous removal of SOX
and NOx 430
7-16 Schematic process flow diagram of the Chiyoda Thoroughbred
102 process 435
xlvii
-------
TABLES
Number
1 PRACTICAL NOV REDUCTION (NOV AND NH3, PPM) vi
X s\
1-1 AMBIENT N02 CONCENTRATIONS AND EMISSIONS IN 1975 FROM
STATIONARY SOURCES IN FOUR REGIONS 6
1-2 AMBIENT N02 CONCENTRATIONS 8
1-3 STANDARDS FOR AUTOMOBILE EMISSIONS IN JAPAN 11
1-4 STANDARDS FOR AUTOMOBILE EMISSIONS IN THE UNITED STATES... 11
1-5 CLASSIFICATION OF NOV ABATEMENT METHODS 14
X
1-6 NOX EMISSION STANDARDS FOR STATIONARY SOURCES (PPM) 17
1-7 NOX EMISSION STANDARDS FOR EXISTING OIL-FIRED BOILERS
WITH AND WITHOUT FGD SYSTEMS (PPM) 18
1-8 NOX ABATEMENT NEEDED TO REDUCE AMBIENT NOX CONCENTRATION
TO DIFFERENT LEVELS IN 1985 20
1-9 METHODS OF ABATEMENT 20
1-10 PRACTICAL AMOUNT OF FLUE GAS TO BE TREATED 22
1-11 NOX ABATEMENT (PERCENT) FOR STATIONARY SOURCES IN 1985 26
1-12 NOX EMISSIONS IN 1985 AND THE PERCENT OF INVESTMENT COSTS
FOR EACH COUNTERMEASURE 28
1-13 INVESTMENTS FOR NOX ABATEMENT BY INDUSTRIES (BILLIONS OF
YEN) 28
1-14 TEST AND COMMERCIAL PLANTS FOR UTILITY BOILERS (LARGER
THAN 280,000 NM3/HR) 35
1-15 BOILER AND NOX ABATEMENT DATA FROM CHITA STATION 36
1-16 DENITRIFICATION PLANTS OF STEEL INDUSTRY 43
2-1 TOTAL NUMBER OF BOILERS INSTALLED BY 1974 AND TOTAL NUMBER
OF BOILERS WITH COMBUSTION MODIFICATION (CM) IN 1977 74
2-2 BOILER NOX EMISSION LEVELS BEFORE AND AFTER COMBUSTION
MODIFICATION (CM) 76
2-3 NUMBER OF BOILERS WITH COMBUSTION MODIFICATION TECHNIQUES. 77
2-4 EFFECTS OF COMBUSTION MODIFICATION TECHNIQUES 78
xlviii
-------
TABLES (Continued)
Number Page
2-5 ATTAINABLE NOX EMISSION LEVELS (PPM) INDICATED BY BURNER
AND BOILER MANUFACTURERS 78
2-6 INVESTMENT COST FOR COMBUSTION MODIFICATION 80
2-7 CM TECHNIQUES USED FOR OIL HEATING FURNACES 82
2-8 NOX EMISSION LEVELS ATTAINABLE WITH SRG BURNER FOR OIL
HEATING FURNACES 83
2-9 EFFECTS OF CM TECHNIQUES ON NOX REDUCTION WITH METAL
HEATING FURNACES 84
2-10 CM TECHNIQUES USED IN METAL HEATING FURNACES 84
2-11 BURNER MANUFACTURERS' TEST RESULTS FOR METAL HEATING
FURNACES 85
3-1 SCR PLANTS (LARGER THAN 10,000 NM3/HR) 92
3-2 CHANGES IN PHYSICAL PROPERTIES OF A1203 CARRIERS BY
CALCINATION 100
3-3 REACTORS FOR DUSTY GAS TREATMENT Ill
3-4 SPECIFIC SURFACE AREA OF CATALYSTS 116
3-5 COMPARISON OF CATALYSTS AND REACTORS 117
3-6 RESULTS OF CORROSION TESTS PERFORMED BY HITACHI ZOSEN 149
3-7 RESULTS OF CORROSION TESTS PERFORMED BY SUMITOMO CHEMICAL. 150
4-1 DATA FOR OPERATION OF SCR PLANTS FOR CLEAN GAS 171
4-2 OPERATION DATA OF SCR PLANTS FOR DIRTY GAS 174
4-3 ESTIMATED COST FOR CLEAN GAS TREATMENT IN 1977 FOR
EXISTING BOILERS, INLET GAS 350°C, NOX 230 PPM 179
4-4 MAJOR SCR CATALYSTS DEVELOPED BY HITACHI ZOSEN 181
4-5 OPERATION PARAMETERS OF MAJOR PLANTS CONSTRUCTED BY
HITACHI ZOSEN 188
4-6 SCR PLANTS BUILT BY HITACHI LTD. (LARGER THAN 10,000
NM3/HR) 202
4-7 CATALYSTS AND REACTORS SELECTED BY HITACHI, LTD. FOR
SEVERAL GASES 205
4-8 DESIGN AND OPERATION DATA OF COMMERCIAL SCR PLANTS
(JGC PROCESS) 225
4-9 COMPOSITION OF THE GAS FROM THE IRON-ORE SINTERING MACHINE
AT NAGOYA WORKS, NIPPON STEEL 229
xlix
-------
TABLES (Continued)
Number Page
4-10 COMPOSITION AND SIZE OF THE PARTICIPATES IN THE GAS FROM
THE IRON-ORE SINTERING MACHINE AT NAGOYA WORKS, NIPPON
STEEL 229
4-11 CONDITIONS OF JGC TESTING AT NAGOYA WORKS, NIPPON STEEL... 230
4-12 OPERATING PARAMETERS OF MITSUI ENGINEERING & SHIPBUILDING
CO. SCR PLANTS 238
4-13 ESTIMATED SCR COST OF MITSUI SHIPBUILDING PROCESS WITH
TUBULAR CATALYST (1977) 247
4-14 MAJOR INSTALLATIONS OF THE PILOT PLANT AT YOKOSUKA
STATION, TOKYO ELECTRIC 256
4-15 LARGE-SCALE SCR PLANTS BY IHI 265
4-16 PLANT COST IN 1978 (MILLIONS OF YEN, $5000) FOR THE
KURABO KNORCA PROCESS 274
4-17 ANNUALIZED SCR COST FOR THE KURABO KNORCA PROCESS, 275
4-18 INVESTMENT AND ANNUALIZED COSTS (MITSUI TOATSU CHEMICAL).. 292
4-19 SCR COST FOR FLUE GAS FROM OIL-FIRED BOILER, ESTIMATED
BY SHI IN 1976 297
5-1 LARGE SNR INSTALLATIONS (THERMAL DENOX) 306
5-2 COMPARISON OF SCR, SNR AND COMBINED SYSTEMS FOR AN INLET
NOX CONCENTRATION OF 200 PPM ; 348
6-1 NOX ADSORPTION CAPACITY OF ACTIVATED CARBON (MG NOX/G
ACTIVATED CARBON) 355
6-2 DESORPTION OF NOX IN REDUCING GAS 357
6-3 EFFECT OF ADDITION OF BASE METAL COMPOUNDS TO CARBON ON
NOX REDUCTION EFFICIENCY (NO, 2000 PPM; NH3, 3000 PPM;
SV, 3000 HR *) 358
6-4 ESTIMATED COST FOR THE USE OF A FERRITE CATALYST FOR
SIMULTANEOUS REMOVAL OF SOX AND NOX VERSUS THE USE OF
SCR FOR THE FLUE GAS FROM A 500 MW COAL-FIRED BOILER 368
6-5 ESTIMATED COST FOR THE UNITIKA PROCESS FOR SIMULTANEOUS
REMOVAL OF 1000 PPM SOX AND 300 PPM NOX BY ACTIVATED
CARBON 374
7-1 MAJOR PLANTS FOR NOX REMOVAL FROM FLUE GAS BY WET
PROCESSES 388
7-2 SPECIFICATIONS FOR THE TEST PLANT (TOKYO ELECTRIC-MHI
PROCESS) 397
-------
TABLES (Continued)
Number Page
7-3 PILOT PLANTS BY MHI WET PROCESSES 398
7-4 SPECIFICATIONS OF 2000 NM3/HR TEST PLANT FOR SIMULTANEOUS
REMOVAL OF SOX AND NOX BY OXIDATION-ABSORPTION-REDUCTION.. 401
7-5 ESTIMATED OPERATING COSTS (CEC PROCESS FOR OIL FIRED
BOILERS) 419
7-6 CAPITAL INVESTMENTS AND OPERATING COSTS FOR THE
THOROUGHBRED 102 PROCESS 438
7-7 REQUIREMENTS FOR WASTEWATER TREATMENT (CHIYODA 102
PROCESS) 438
11
-------
CONVERSION FACTORS AND ABBREVIATIONS
CONVERSION FACTORS
The metric system is used in this report. Following are some factors
for conversion between metric and English systems:
1 m (meter) = 3.3 feet
1 m3 (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) =6.19 barrels
Gas flow rates are expressed in Nm3/hr (normal cubic meters per hour).
1 Nm3/hr = 0.59 standard cubic foot per minute
L/G ratio (liquid/gas ratio) is expressed in liters/Nm3.
1 liter/Nm9 = 7.4 gallons/thousand standard cubic feet
YEN-DOLLAR CONVERSION
Wherever possible in the report where costs in yen are presented, the
equivalent cost in U.S. dollars is also presented. The exchange rate used
to make these conversions was 200 yen/$ and this figure should be used when
converting cost data presented in graphs since the graphical data is
presented in terms of yen.
lii
-------
ABBREVIATIONS
FGT •" Flue gas treatment
FGD • Flue gas desulfurizatlon
CM « Combustion modification
SV • Space velocity
SCR • Selective catalytic reduction
SNR » Selective noncatalytic reduction
ESP » Electrostatic precipitator
kW = Kilowatt
kWh = Kilowatt hour
MW = Megawatt
LNG = Liquified natural gas
L/G - Liquid-to-gas ratio
Nm3/hr • Normal cubic meters per hour (0°C, 101,325 Pa)
liii
-------
SECTION 1
NOX CONCENTRATION, REGULATION AND ABATEMENT IN JAPAN
1.1 INTRODUCTION
In Japan, ambient SOa concentrations have decreased remarkably due to
the use of low-sulfur fuels and flue gas desulfurization. As a result,
recent efforts for air pollution control have been concentrated on the
reduction of NOX (Figures 1-1 and 1-2).
Man-made emissions of NOx in Japan have totaled nearly 2 million tons
yearly. About 65 percent are derived from stationary sources, while in
large cities 60 to 70 percent are attributed to mobile sources. Yearly
average N02 concentrations in major Japanese cities are compared with those
in several major world cities in Figure 1-3. This figure shows that N02
concentrations in the Japanese cities do not exceed the U.S. air quality
(ambient) standard, but are much higher than the 1973 Japanese standard.
The U.S. standard is a yearly average of 0.05 ppm, while the 1973 Japanese
standard is a yearly average of about 0.01 ppm (a daily average of 0.02
ppm)—the most stringent in the world.
The 1973 standard was based principally on animal tests, which gave
results similar to those obtained in the United States. The large differ-
ence in the standards of the two countries is due to the difference in the
safety factors. Since compliance to the Japanese standard is voluntary, a
large safety factor was used for N02 to ensure public health and to prevent
photochemical smog. Conformance to this standard was to be attained within
5 years in most districts and within 8 years in polluted cities such as
Tokyo and Osaka, which have a daily average of 0.06-0.08 ppm N02. For the
-------
CM
o
0.06
0.04
0.02
U.S. Standard
1978 Standard
1973 Standard
1970 1972 1974
year
1976
Figure 1-1. Comparison of U.S. and Japanese ambient NC-2 standards and
yearly average concentrations of 6 Japanese stations.
&
0.
0.06
0.04
0.02
Japanese Standard
J I I 1 i i
J I I I I
1965
1970
year
1975
Figure 1-2. Comparison of U.S. and Japanese ambient SOz standards and
yearly average concentrations of 15 Japanese stations.
-------
0.08
Los Angeles
0.06
OL
a.
-------
polluted districts, the Environment Agency also set an interim standard to
be achieved in 5 years: a daily average of 0.04 ppm which is equivalent to
a yearly average of 0.02 ppm.
Great efforts have been made since then to reduce NOX emissions through
combustion control, change of fuel, and flue gas denitrification at many
stationary sources; and also through quite stringent regulations on auto-
mobile exhausts. It has been quite difficult, however, to achieve not only
the final goal, but also the interim standard in large cities with many
automobiles.
Recent studies by the Environment Agency have shown that the large
safety factor which resulted in the 1973 standard (a daily average of 0.02
ppm N02> may not be necessary to protect human health. On the basis of a
report by the Central Council for the Control of Environmental Pollution,
the Environment Agency promulgated a revised N02 ambient standard in July
1978—a daily average of 0.04 to 0.06 ppm.
The following sections will describe various aspects of NOx abatement,
including the background of the revision of the standard.
1.2 NOX EMISSIONS, N02 AMBIENT CONCENTRATIONS, AND ABATEMENT
1.2.1 NOx Emissions
Figure 1-4 shows NOX emissions from large stationary sources in Japan
in 1975; the sources, which were controlled by emission standards, emitted
a total of 920,000 tons NOX.* Total NOX emissions from all stationary
sources, including small uncontrolled ones, are estimated at 1 million tons.
NOX derived from mobile sources is roughly estimated at about 600,000 tons
yearly.
* A metric ton equals 2,200 Ibs.
-------
Others
180,000 t
(20%)
Utility Boilers
310,000 t
(34%)
Oil
Refinery
40,000 t (4%)
Total
920,000 t
(100%)
Ceramic
Industry
130,000 t
Chemical
Industry
110,000 t
(12%)
Steel Industry
150,000 t
(16%)
Figure 1-4. NOX emissions from Japanese stationary sources
controlled by regulation in 1975.l
-------
Table 1-1 shows NOX emissions from stationary sources in 4 regions.
Region A includes 6 subregions: the 5 largest cities (Tokyo, Osaka, Yoko-
hama, Nagoya and Kobe), including industrial areas in and around the cities,
and the largest industrial complex in the Chiba-Ichihara district. Region
A has the highest N02 concentrations, a daily average over 0.08 ppm; about
250,000 tons of NOX were emitted there in 1975. Region B includes 9 medium
size cities with total emissions of 70,000 tons and a dally average of 0.06-
0.08 ppm. Region C, which includes 32 subregions (mainly smaller cities),
has NOx emissions of 190,000 tons and a daily average of 0.04-0.06 ppm*
Region D contains the remaining 72 subregions and has 410,000 tons of
emissions; its ambient N02 concentration, however, is low because it is much
larger than the other regions.
TABLE 1-1. AMBIENT N02 CONCENTRATIONS AND EMISSIONS IN 1975
FROM STATIONARY SOURCES IN FOUR REGIONS
Region
A
B
C
D
Total
N02
(ppm, daily average)
Over 0.08
0.06-0.08
0.04-0.06
Below 0.04
NOX from Stationary Sources
(1000 tons)
250
70
190
410
920
(percent)
27.0
7.7
20.7
44.6
100.0
A - Largest Cities and Industrial Complexes
B - Medium Size Cities
C - Smaller Cities
D - All Others
Figure 1-5 shows the contribution of NOX sources to ground level N02
concentrations. The NOx sources include large, controlled stationary
sources, mobile sources, and small uncontrolled stationary sources in each
region. Over 50 percent of N02 is derived from mobile sources in regions
A, B, and C.
-------
Group of
Regions
C
D
Stationary
Sources
(controlled)
36.5%
20.0% X'
25.0% \
34.5%
Small Stationary
Sources (uncontrolled)
Mobile Sources
52.7%
10.8%
65.7%
54.2%
/ 20.
31.0%
34.5%
Figure 1-5. Contribution of W* sources to ground
level N02 concentrations.1
1.2.2 Ambient NO2 Concentrations and Abatement Standards
Ambient NO2 concentrations have been monitored in 1975 at 666 stations
scattered over Japan. (This figure does not include roadside stations to
monitor automobile exhaust gas.) Annual changes in ambient NO2 concentra-
tions at continuous monitoring stations are shown in Table 1-2. The concen-
trations have remained at almost the same level or have improved slightly
since 1973. The ambient NOa standards set in 1973 were satisfied at 4 of
the total 288 stations in 1973, at 25 of the 448 in 1974, and at 54 of the
666 in 1975.
-------
TABLE 1-2. AMBIENT N02 CONCENTRATIONS
Fiscal
Year
1968
1969
1970
1971
1972
1973
1974
1975
Average of Annual
Averages at 6
Monitoring Stations*
(ppm)
0.025
0.025
0.028
0.026
0.029
0.034
0.033
0.030
Average of Annual
Averages at 16
Monitoring Stations**
(ppm)
0.025
0.024
0.023
0.028
0.030
0.030
* Tokyo, Kawasaki, Osaka, Amagasaki, Ube, and Kitakyushu.
** Six stations in Tokyo, one each in Ichihara (Chiba), Kawasaki, Nagoya,
Osaka, Amagasaki, Himeji, Matsue, Kurashiki, Ube, Kitakyushu.
Figure 1-6 shows the daily average NOa concentrations in 80 areas of
the 120 city areas and the percent of NOx abatement required to reduce
ambient NOa concentrations to different levels. The abatement ratios were
calculated with two assumptions: 1) a proportional relationship between the
amount of NOX discharged and the ambient concentrations and 2) a background
N02 concentration of 0.01 ppm, including N02 from small stationary sources
not subject to regulations.
The ambient NOa concentration is high in large cities, and the daily
average exceeds 0.08 ppm in six areas: 1) Tokyo, 2) Yokohama and Kawasaki,
3) Osaka and Sakai, 4) Nagoya, 5) Kobe and Amagasaki, and 6) Chiba and
Ichihara (Region A of Table 1-1 and Figure 1-5). It is necessary in those
areas to reduce NOX by 80-90 percent to achieve the 1973 standard (0.02
ppm), 60-70 percent to achieve the interim objective (0.04 ppm), and 30-50
percent to attain the new standard (0.06 ppm). For areas in which the NOX
-------
0.11 0.10 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01
u
10
20
30
40
50
60
70
80
N
\
N
\
\
\
\
\
\
\
\
\
X
\
\
\
\ ^
\
\
^ \
\
\
\
•\
\
\
\
N
\
^ \
\\
X
\
\
\
\
N \
\
\
\
\ \
\
Ov
\ \
\ '
X
\
\
\
\
\ N
\
\
s, \
\ \
\
, \
V \
SN
\
\
\
\
\
, \
\ \
\
\
\
\\
oN
* \ \
^
\
\
\
\
\
\
\
\
\\
\ .
\ \
\ '
\\
V\V
'xV\
1
\
1
1
1
1
\
\
\
1
\
\
I I
\ \
\ \
\ \
\\ \
\\\1
2 AREAS | 4 AREAS I 9 AREAS I 30 AREAS I 33 AREAS 1 2 AREAS
N02 (ppm, in daily average)
Figure 1-6. 1975 ambient NOa concentrations and the decrease by NOx
abatement for 80 areas of the 120 city areas.
Remarks: 1) Of the 120 areas specially designated under the Enforcement
Ordinance for the Air Pollution Control Law, 80 areas having
more than 2 available stations are listed in this table.
2) The background-concentration is assumed to be 0.01 ppm
(including smaller'NOX sources that are not subject to
regulations).
-------
concentration range is 0.06-0.08 ppm, up to 30 percent abatement is needed
to achieve 0.06 ppm.
The Environment Agency has promulgated stringent regulations on fc
automobile emissions and stationary sources in order to reduce NOX as much
as possible.
1.3 REINFORCEMENT OF NOx EMISSION CONTROL FOR PASSENGER CARS2
In December 1976, the 1978 Control designed to strengthen the discharge
limitations for nitrogen oxides from passenger cars, trucks and buses was
announced. In response to this announcement, the Safety Standards for
Vehicles for Road Transportation were partly modified by the Ministry of
Transportation; the modified safety standards were to be enforced as of
April 1978 on new models and March 1979 on existing models of which pro-
duction is to be continued. These modified standards are shown in Table
1-3, along with the standards for 1973, 1975, and 1976; Table 1-4 presents
automobile emission data for the United States.
NOX discharge is stipulated to be less than 0.25 g/km, on the average,
by 10 modes test. Accordingly, permissible limits are set at 0.48 g/km, as
shown in Table 1-3. The regulations are much more stringent than those in
the U.S. (Table 1-4). At the same time, those in 11 modes are strengthened,
establishing the limits at 6.0 g/test (average discharge: 4.4 g/test).
These modified standards are expected to contribute to a 60-70 percent
reduction in NOx discharge compared with the current regulations and to a
92 percent reduction compared with the time when there were no regulations.
Incidentally, it has thus far been believed that performance and fuel
economy would decline measurably as the result of an attempt to reduce NOX
discharge. Strenuous efforts were made, however, to maintain and even to
improve fuel economy while preserving performance to an extent that no
problems should occur in practical use in spite of the modified standards.
10
-------
TABLE 1-3. STANDARDS FOR AUTOMOBILE EMISSIONS IN JAPAN*
1973
1975
1976
1978
HC
(g/km, 10 modes)
3.80 (2.94)
0.39 (0.25)
0.39 (0.25)
0.39 (0.25)
CO
(g/km, 10 modes)
28.00 (18.40)
2.70 ( 2.10)
2.70 ( 2.10)
2.70 ( 2.10)
NOX
(g/km, 10 modes)
3.00 (2.18)
1.60 (1.20)
0.84 (0.60)
0.48 (0.25)
* Figures show allowable limits; those in parentheses show average
emissions.
TABLE 1-4. STANDARDS FOR AUTOMOBILE EMISSIONS
IN THE UNITED STATES*
HC CO NOX
(g/km) (g/km) (g/km)
1978 0.94 9.4 1.25
1979 0.94 9.4 1.25
1980 0.26 4.4 1.25
1981 0.26 2.1 0.63
* Established by the Clean Air Act of 1977.
11
-------
As a result of these efforts, fuel consumption of automobiles which conform
to the 1976 Control has been improved to the extent that there is no notice-
able decline compared with those conforming to the 1974 Control. On the
contrary, improvements have been made in the case of some types of auto-
mobiles.
Fuel consumption rates have also been considerably improved in systems
where NOX discharge limits of 0.25 g/km were satisfied. Performance better
than that of automobiles conforming to the 1976 Control has been observed
in most of the cases. It should be noted that a new combustion method has
made the fuel consumption rates of some models even better than those of
automobiles in conformity with the 1973 Control.
It was decided that the regulations for trucks and buses should be
reinforced with the control measures presently available and within the
scope that could be achieved in a limited time period. In December 1976,
reinforced regulations were announced as the 1977 Control for trucks and
buses. This announcement was made simultaneously with the announcement of
the reinforced regulations for passenger cars. These tightened regulations
provide that the discharge of NOx from diesel-engined vehicles should be
reduced by approximately 30 percent as compared with the time when there
were no such regulations; the discharge from heavier gasoline-engined cars
and LPG-engined cars is to be reduced by approximately 47 percent.
It is estimated, however, that the total amount of NOX discharged from
from automobiles will remain at somewhere around 66 percent of that in 1975,
including a 10 percent reduction through traffic control, even in the year
1983 when the control measures are expected to exhibit their maximum effect
(see Figure 1-7).
12
-------
£
X
200 -
150 -
100
(1 No regulation
(2 No modification to 1976 Control
(3 0.25 g/km 1978 Control on Passenger Cars
(4) 0.25 g/km 1978 Control on Passenger Cars, 15%
reduction from 1977 on diesel-engined cars and
heavyweight gasoline-englned cars
1965
1970
1975
1980
1985
1990
YEAR
Figure 1-7-
Estimated total amount of IStOx emissions from
road traffic in the Tokyo Bay Area.2
13
-------
1.4 COUNTERMEASURES FOR STATIONARY SOURCES
1.4.1 Methods of NOx Abatement
Major methods of NOx abatement for stationary sources are classified
in Table 1-5.
TABLE 1-5. CLASSIFICATION OF NOX ABATEMENT METHODS
Reduction of
Formation
Combustion
Modification
Use of Low-
Nitrogen Fuel
Low Oxygen Combustion
Flue Gas Recirculation
Staged Combustion
Off-Stoichiometric Combustion
Low-N0x Burners
Change of Fuel
Nitrogen Removal from Fuel
NOX Removal
from Flue Gas
Dry Process
Wet Process
Selective Catalytic Reduction
(SCR)
Selective Noncatalytic
Reduction (SNR)
Oxidation Absorption
Oxidation Reduction
Absorption Reduction
Combustion modification is the most reasonable primary step of
abatement. (It will be described in Chapter 2.) Combustion modification
has been widely applied in Japan and has contributed to 10-60 percent of
NOX abatement for each source.
14
-------
The change from high-nitrogen to low-nitrogen fuel has also been
carried out to a considerable degree. Heavy oil, the major fuel in Japan,
contains 0.2-0.3 percent nitrogen, and the combustion gas contains 110-130
ppm of NOX under the best combustion control. A considerable number of new
power plants burn LNG (imported liquid natural gas), which yields only 35
ppm of NOx under the best combustion control. There is, however, a limit to
the supply of such clean fuel, and it is costly. Hydrodesulfurization of
heavy oil, which removes 70-90 percent sulfur, also removes 20-30 percent
nitrogen, but this does not reduce NOX substantially.
In order to remove NOx present in flue gas, many processes of denitri-
fication have been developed. Of these, selective catalytic reduction (SCR)
that uses NHs and a catalyst to reduce NOx to Na at 200-400°C has been the
most widely employed: it is simple, has a high (90 percent) NOX removal
efficiency, and does not produce by-products difficult to dispose of.
Although NOX can be reduced to Na by other reducing gases such as Ha, CO
and CHit, large amounts of these gases are consumed by reacting with the 02
in flue gas. The use of NHs is more suitable because it selectively reacts
with NOX (see Chapter 3).
Selective noncatalytic reduction (SNR, NHs reduction without a
catalyst) has been used in several plants. SNR is much simpler than SCR,
but the NOX removal efficiency does not exceed 50 percent and a considerable
amount of NHa is emitted (see Chapter 5).
Both SCR and SNR consume ammonia—0.8-1.2 moles of NHg to every mole of
NOX f°r SCR and 1.5-2.5 moles for SNR. It is possible that when SCR and
SNR are applied widely in many countries, ammonia will fall into short
supply; as a result, supplies of nitrogen fertilizer and food will be
affected. Therefore, it is desirable that processes that do not consume
ammonia be developed.
15
-------
Most wet processes do not consume ammonia. Some of them even convert
NOx to NHs. The wet processes, however, are not as simple as SCR or SNR
and moreover, present wastewater treatment problems (Chapter 7).
1.4.2 Emission Standards
The first regulations governing emission standards for nitrogen oxides
from stationary sources were originally established in August 1973. Large
boilers, heating furnaces, nitric acid manufacturing facilities and other
sources were subject to the regulations in establishing the emission
standards.
In 1975, standard values and expansion of the scope of facilities
subject to the regulations were enforced. These second-stage regulations
included medium-size facilities, as well as cement calcination furnaces and
newly built coke furnaces.
Third-stage regulations were enacted in June 1977 (see Table 1-6).
Their major points include expansion of the types and capacities of facili-
ties subject to regulation, and reinforcement of the standards for newly
built facilities. The standards went into effect on June 18, 1977 for
facilities built after this date. In regard to the existing facilities, the
regulations provide that the standards be applied as of May 1, 1980, and
in principle allow for a grace period of three years. Also, the standards
are expected to be applied as of October 1, 1980, to liquid combustion
boilers of between 5000 and 10,000 Nm3/hr emission gas; this takes into
consideration the period required for measuring NOz concentrations and other
factors.
Less stringent standards are applied to existing boilers with an FGD
system (Table 1-7). Boilers to be constructed, however, are subject to
uniform NOx regulations regardless of whether they are equipped with an FGD
plant.
16
-------
TABLE 1-6. NOX EMISSION STANDARDS FOR STATIONARY SOURCES (PPM)
Sources
Boiler (Gas-Fired)
Boiler (Coal-Fired)
Boiler (Oil-Fired)
Iron Ore Sintering
Machine
Alumina Calciner
Metal Heating Furnace
Oil Heating Furnace
Cement Kiln
Coke Oven
Waste Incinerator
Capacity,
(1000 NmVhr)
More than 500
100-500
40-100
10-40
5-10
Less than 5
More than 100
40-100
10-40
5-10
Less than 5
More than 500
100-500
40-100
10-40
5-10
Less than 5
More than 100
Less than 100
More than 10
More than 100
40-100
10-40
5-10
Less than 5
More than 100
40-100
10-40
5-10
Less than 5
More than 100
Less than 100
More than 100
Less than 100
More than 40
New Plants
1977
60
100
130
130,
150 1
150 b
400
400
400,
400°
400 b
130
150
150
150 b
180 b
180 b
220
230
220
100
130
130
150
180
100
100
130
150
180
250
350
170
170
250
1975
100
100
130
130
n
n
480
480
480
n
n
150
150
150
150
n
n
n
n
n
100
150
150
n
n
100
100
150
n
n
250
n
200
n
n
1973
130
130
130
n
n
n
480
480
n
n
n
180
180
180
n
n
n
n
n
n
200
200
170
n
n
170
170
170
n
n
n
n
n
n
n
Existing Plants „ .,_
1977
130
130
130
150
150 a
n
480
600
600
480 a
n
180
190
190
230
250 a
n
260
270
n
160
170
200
n
n
170
170
180
n
n
480
480
350
350
n
1975
130
130
130
150
n
n
600
750
750
n
n
230
230
190
n
n
n
n
n
n
220
220
n
n
n
210
210
n
n
n
n
n
n
n
n
1973 Gas (%)
170 5
170
n
n
n
n
750 6
n
n
n
n
230 4
230
n
n
n
n
n 15
n
n 10
220 11
220
n
n
n
210 6
210
n
n
n
n 10
n
n 7
n
n 12
''Must be met by October 1, 1980.
For plants to be constructed after September 10, 1977.
n • No regulation.
17
-------
TABLE 1-7. NOx EMISSION STANDARDS FOR EXISTING OIL-FIRED BOILERS
WITH AND WITHOUT FGD SYSTEMS (PPM)
Capacity (1000 Nm3/hr)
More than
1000 500-1000 40-500 10-40 5-10
Without FGD 180 180 190 230 250
With FGD 180 210 210 250 280
The third-stage emission standards may be the most stringent at the
present in the world; nevertheless, it is possible to achieve them with the
most advanced technology of combustion modification, including the use of
low NOX burners and the switching of fuels to some extent.
It is expected that the third-stage regulations will result in a
reduction of NOX emissions of approximately 8 percent as compared with those
of 1975 and a cumulative reduction of approximately 30 percent as compared
with the value at the time of the first regulations. Even with these
efforts, it is not possible in many regions to attain even the interim
objective of the 1973 standard, a daily average of 0.04 ppm.
1.5 COST ESTIMATES FOR NOX ABATEMENT FOR STATIONARY SOURCES AND LIMIT OF
ABATEMENT
1.5.1 Estimate by MITI3
The Ministry of International Trade and Industry (MITI) has made cost
estimates for NOx abatement to reach different ambient NOa concentrations
ranging from 0.02 to 0.07 ppm in daily average. For convenience, the whole
of Japan was divided into four regions:
18
-------
Region I - Large cities such as Tokyo, Osaka, and Nagoya
that account for one-third of Japan's total
population.
Region II - Forty-three industrial complexes that include
Kashima, Mizushima, and Kitakyushu.
Region III - Eighty-six local cities, including Sapporo.
Region IV - Other areas.
NOX abatement required for stationary sources to reach different N02
levels in 1985 and methods for the abatement are shown in Table 1-8 and
1-9, respectively.
It has been assumed that SCR was used for NOX removal froiji flue gas,
because it is the most practical method at present. The investment costs
for SCR plants with different capacities for clean and dirty gases are shown
in Figure 1-8.* The plants are for existing boilers and treat clean gas
after it has passed through a preheater or dirty gas after FGD. The invest-
ment costs are rather high because those were estimated with enough safety
margin to ensure trouble-free operation. Utility requirements for an SCR
unit with a capacity of treating 200,000 Nm3/hr of flue gas (66 MW equiva-
lent) are as follows assuming operation for 70 percent of the total hours
of a year:
Power: 7.67 x 10s kWhr/year
Fuel for gas heating: 3.07 x 103 kl/year
Ammonia: 251 tons/year (at 300 ppm NOX in gas, 0.9 mole
mole NOX for about 85 percent removal).
* Clean gas is gas that is free of SOX and particulates, but does contain
NOX; dirty gas contains SOX and particulates in addition to NOX.
19
-------
TABLE 1-8. NOx ABATEMENT NEEDED TO REDUCE AMBIENT N02 CONCENTRATION
TO DIFFERENT LEVELS IN 1985
N02 Daily
Average (ppm)
0.06
0.05
0.04
0.03
0.02
Abatement Required (percent)
I
43
55
66
77
89
II
2
22
41
61
80
III
11
29
46
64
82
IV
0
0
19
46
73
TABLE 1-9. METHODS OF ABATEMENT
Method
Abatement
(percent)
Investment Cost
M/Nm3/hr ($/Nm3/hr)
Combustion Modification
10-40
470 (2.30)
Fuel Switching
7-39
None for User
(Fuel Cost Increases)
Flue Gas Treatment
80-90 6500-18,000 (32.5-90) - Dirty Gas
3000-7500 (15-37.5) - Clean Gas
(As shown in Figure 1-8)
20
-------
•3
^^^ W
I
•r- O
n t
o
o
c
0 40 TOO 200 300 400
Capacity (1,000 Nm3/hr)
Figure 1-8. Investment costs for SCR denitrification plants
estimated by MITI.
Clean Gas:
Dirty Gas (I):
Dirty Gas (II):
From fuel gas, kerosene.
From heavy oil burning boiler, coke oven, etc.
From iron-ore sintering machine, coal-fired boiler, glass
furnace, cement kiln, etc.
21
-------
Catalyst cost was assumed at 3.5 million yen ($17,500)/m3 of catalyst
with an SV of 5000 hr"1 for dirty gas and 2.5 million yen ($12,500)/m3 with
an SV of 10,000 hr"1 for clean gas.
By flue gas treatment, about 90 percent of NOx in the gas can be
removed. However, it is not practical to apply SCR to numerous sources with
a capacity smaller than 40,000 Nm3/hr (13 MW equivalent). Even for larger
sources, many sources have no room to install an SCR unit because they have
already installed an FGD unit. The practical maximum amount of flue gas
that can be treated is shown in Table 1-10.
TABLE 1-10. PRACTICAL AMOUNT OF FLUE GAS TO BE TREATED
NOx Daily Average (ppm) 0.06 0.05 0.047
Gas To Be Treated (10s Nm3/hr) 48 84 120
In order to reach the ambient NOa concentrations below 0.06 ppm, fuels
must be switched from heavy to lighter oils for many of the NOX sources for
which SCR cannot be applied. Thus construction of ,plants to reduce the
nitrogen content of heavy oil is required. To attain a daily average of
0.47 ppm in 1985, 30 million kiloliters of heavy oil should be treated
during the year. This requires 26 units for heavy oil decomposition by
hydrogen addition, with a total capacity of 640,000 BPSD, and 9 units for
thermal decomposition (coking) with a total capacity of 220,000 BPSD.
Figure 1-9 shows the total investment costs needed between 1976 and
1985 to reach different NOa levels; Figure 1-10 presents the annualized
costs, including 7 years depreciation, required to reach different NOa
levels in 1985. As for the investment costs, about 0.4 trillion yen ($2
billion) is required to reach 0.06 ppm, about 1 trillion yen ($5 billion)
22
-------
2.0
1.0
« 1.5
90
CM
ro
VI
01
10
4J
O
1.0
0.5
0
Oil Refining
Flue GAS
Denitrlficatlon
Combustion
Modification
0.07 0.06 0.05
N02 (ppm)
>
CM
*o
I/I
O
O
TJ
O»
N
0.5
Oil Refining
Flue Gas
Denitrification
Combustion
_ Modifjcation.
0.07 0.06 0.05
N02 (ppm)
Figure 1-9.
Total investment costs required
between 1976 and 1985 to-reach
different N02 concentrations
(daily averages).
Figure 1-10,
Annual costs in 1985 needed
to reach different NOa concen-
trations (daily averages).
-------
to reach 0.05 ppm, and 2 trillion yen ($10 billion) to reach 0.047 ppm. Of
the 0.4 trillion yen investment required to reach 0.06 ppm, nearly 0.1
trillion yen is for combustion modification (CM) and about 0.3 trillion yen
for flue gas treatment (FGT). To reach 0.047 ppm, an investment of about
1 trillion yen is required for heavy oil decomposition in addition to 0.2
trillion yen for CM and 0.7 trillion yen for FGT. Further reduction will
require exceedingly large investments for heavy oil decomposition. The
annual costs are about 150 billion yen ($750 million) for 0.06 ppm, nearly
500 billion yen ($2.5 billion) for 0.05 ppm,. and about 1 trillion yen ($5
billion) for 0.047 ppm.
The energy requirements for NOy abatement are shown in Figure 1-11.
The energy requirements are an equivalent of 110,000 kl oil to reach 0.06
ppm, 250,000 kl to reach 0.05 ppm (mainly for flue gas treatment), and about
650,000 kl (including about 400,000 kl for heavy oil decomposition) to reach
0.047 ppm.
7.5
LO
O
o
+J
o
CJ
O)
5.0
2.5
Oil Refining
Flue Gas Denitrification
Combustion Modification
0.07 0.06 0.05
N02 (ppm)
Figure 1-11. Energy consumption required in 1985 to reach different
NOX concentrations (daily averages).
24
-------
These estimates indicate that cost escalates so rapidly that the
practical, attainable ambient NOz concentration is a daily average of 0.05
ppm, which is equivalent to a yearly average of 0.025 ppm.
1.5.2 Estimates by Environment Agency1
The Environment Agency (EA) also estimated the cost of NOx abatement.3
Japan was divided into 4 regions (A, B, C and D), according to daily average
NOa concentrations of over 0.08 ppm, 0.06-0.08 ppm, 0.04-0.06 ppm, and below
0.04 ppm, respectively (see Table 1-1). The EA assumed the following 4
cases for each region:
• Case a - Maintain the third-stage emission regulations
(Table 1-6).
Case b - Tighten the regulations for stationary sources to
achieve a daily average of 0.06 ppm in 1985.
• Case c - Tighten the regulations to achieve a daily average
of 0.04 ppm in 1985.
Case d - Take all possible, practical countermeasures for
stationary sources, e.g., combustion modification
and flue gas denitrification.
In Case a, NOx emissions in 1985 will increase by 20 percent over 1975
due to an increase in new NOx sources. The abatement ratios in other cases
are shown in Table 1-11.
25
-------
TABLE 1-11. NOX ABATEMENT (PERCENT) FOR STATIONARY SOURCES IN 1985
Region A Region B Region C Region D
Case b 40 17 0 0
Case c 64 50 21 0
Case d 76 66 47 0
For the abatement in 1985, the methods of abatement were assumed to be
combustion modification (an average abatement of 30 percent), flue gas
denitrification by selective catalytic reduction (90 percent abatement),
and selective noncatalytic reduction (50 percent abatement). The estimated
total investment costs between 1975 and 1985 and the annual costs in 1985,
including 7 years depreciation, fuel for gas reheating, etc., are shown in
Figure 1-12.
NOX emissions from stationary sources in 1985 and the percents of
investment costs for combustion modification, SCR and SNR in each case are
shown in Table 1-12. In Case a, 68 percent of the total investment is for
combustion modification; in Case d, however, 15 percent is for combustion
modification, while 75 percent is for SCR and 10 percent for SNR. This
means that all possible combustion modification is done in Case a, but in
other cases flue gas denitrification, particularly SCR, is increasingly
used.
The investment cost for each industry is shown in Table 1-13. The
power industry requires large investments, for example, 283.8 billion yen
in Case d. Tht total power generation capacity in Japan in 1985 is
estimated to be 200,000 MW, about half of which is thermal (steam) power.
26
-------
looon
01
o
«/>
- 500-
CQ
3
(I)
859
(I)
674
(I)
471
(I)
194
(A)
(A)
310
(A)
572
Case a
Case b
Case c
Case d
Figure 1-12. Investment cost (I) and annual cost (A) for each case.
27
-------
TABLE 1-12. NOX EMISSIONS IN 1985 AND THE PERCENT OF INVESTMENT
COSTS FOR EACH COUNTERMEASURE
NOX Emissions (1000 t)
Combustion Modification (Percent)
SCR (Percent)
SNR (Percent)
Total (Percent)
Case a
1108
68
18
14
100
Case b
800
22
52
21
100
Case c
679
19
65
16
100
Case d
600
15
75
10
100
TABLE 1-13. INVESTMENTS FOR NOX ABATEMENT BY INDUSTRIES
(Billions of Yen)
Power
Steel
Chemical
Cement
Glass
Oil Refining
Other
Case a
66.4
48.1
15.6
7.8
0.0
7.0
49.4
Case b
187.0
75.2
37.1
26.0
0.2
15.6
129.8
Case c
240.0
91.0
59.9
31.6
0.9
18.1
232.5
Case d
283.8
134.8
98.0
35.9
2.6
36.9
267.1
Total 194.3 470.9 674.0 859.1
28
-------
Figure 1-13 shows ambient NOz concentrations in each region in 1975
and in cases a, b, c and d for each region in 1985. In addition to the
control of stationary sources, passenger cars are controlled by the 1978
regulation and buses and trucks by the 1979 regulations. In case d in 1985,
the concentration is reduced to 0.03 ppm in Regions C and D but exceeds 0.04
ppm in Region A. To reduce NOX further, many plants for heavy oil composi-
tion are needed. Since the construction of many decomposition plants is not
practical due to high costs, attainment of the interim objective of the 1973
standard—a daily average of 0.04 ppm—may be practically impossible in
Region A.
This conclusion is similar to that obtained by MITI (Section 1.5.1),
except that EA's cost estimates for NOx abatement for stationary sources
are about half of those estimated by MITI. EA's cost estimate for SCR is
much lower than MITI's (Section 3.6.2). The low cost may be due primarily
to the recent progress in SCR by which flue gas from the boiler economizer
containing large quantities of SOx and fly ash can be directly treated.
It may be based partly on the recent unusually low costs offered by plant
constructors due to excessive competition. MITI's cost estimate for SCR
may be based on the actual high cost in the past, including heating
facilities for the cooled gas. The real cost for NOX abatement may be
between the costs estimated by MITI and EA.
1.6 REVISION OF AMBIENT N02 STANDARD
1.6.1 Consultation for Criteria and Guidelines for NOa That Has Adverse
Effects on Human Health
The ambient air quality standard for N02, established in 1973, was
based on limited data and information available prior to June 1972. The
standard allowed for a sufficient safety factor with a view to nipping
pollution problems in the bud. Considerable amounts of data were made
available through biological experiments; on the other hand, epidemiclogical
29
-------
(US
o
0.10 -
0.08 -
E
01
o
c
0.04 -
0.02 -
Region A
Region B
Region C
Region D
Figure 1-13. Ambient N02 concentrations in 1975 and 1985
with countermeasures a, b, c and d.1
-------
data were scarcely available in Japan or in other countries. Five years
have passed, and some scientific data and information have been obtained
since then.
A meeting of an expert committee of the World Health Organization (WHO)
was held in Tokyo in August 1976; the one-hour NOa value of 0.5 ppm was
established as the threshold. Using a safety factor of 3-5, WHO suggested
a one-hour NOz value of 0.10-0.17 ppm as a criterion.
Also, a population-based survey of the influence of combined air
pollution on health was carried out as part of an epidemiological study on
the adverse effects of ambient air pollution on the public health of
communities. The results were made public in January 1977.
The Basic Law for Environmental Pollution Control stipulates that
scientific consideration shall always be given to any environmental quality
standard. Since much more data and information became available in Japan,
as well as from other countries, more accurate criteria were to be estab-
lished. In light of this, the Central Council for the Control of Environ-
mental Pollution was consulted in March 1977 about criteria and guidelines
concerning the adverse effects of NOa on human health and the establishment
of acceptable as well as desirable levels.
1.6.2 New N02 Standard
In March 1978, the Council recommended the following NOa criteria: a
yearly average of 0.02 to 0.03 ppm and an hourly average of 0.1 to 0.2 ppm.
These values are close to those given by the OECD experts.
In July 1978, EA promulgated an amended NOz ambient standard, a daily
average of 0.04 to 0.06 ppm. In regions where the daily NOa average con-
centration exceeds 0.06 ppm, the concentration is to be reduced to 0.06 ppm
in 7 years. In regions where the NO2 concentration ranges between 0.04 and
31
-------
0.06 ppm, efforts shall be made to keep the NOz concentration from exceeding
the present level appreciably. In regions where the NOz concentration is
below 0.04 ppm, efforts shall be made to keep it from exceeding 0.04 ppm
appreciably.
It may be unusual for an air quality standard to prescribe a range of
concentrations. Presumably EA wanted to set the standard at 0.04 ppm, the
interim objective of the 1973 standard, because some of the local govern-
ments had been trying to reduce NOa concentrations to 0.04 ppm. This level,
however, has been practically impossible to attain in large cities where
most of the NOx is derived from mobile sources and numerous small uncontrol-
lable stationary sources. For human health, 0.06 ppm is considered accept-
able, but a lower level is obviously desirable. The new standard with a
range of concentrations has been set under such circumstances.
EA is going to set a total mass NOa regulation for regions in which
NOa concentration exceeds 0.06 ppm in order to reduce the concentration to
0.06 ppm within 7 years. A considerable number of flue gas denitrification
plants will be required to meet the total mass regulation.
Some of the local governments which had made efforts to attain 0.04
ppm are not satisfied with the new standard and plan to continue to strive
for that level. This effort will require very stringent NOx control of new
large plants; it may even necessitate the construction of flue gas denitri-
fication plants in less polluted regions.
EA has continued to use the Saltzman process as the official analytical
method for the measurement of ambient N0£ concentration. The factor for
the calculation of NOz concentration has been revised from 0.72 to 0.84.
This factor change has resulted in a slight relaxation of the NOz standard.
32
-------
1.7 NOX ABATEMENT BY INDUSTRIES
1.7.1 Introduction
In Japan, the largest stationary source of NOx is the power industry,
which annually emits 310,000 tons of NOX. Thus, this industry accounts for
34 percent of the total emissions from stationary sources (Figure 1-4). The
second largest source is the steel industry, which emits 150,000 tons (16
percent), and the third is the ceramic industry, which emits 130,000 tons
(14 percent).
The power companies have concentrated their efforts mainly on combus-
tion modification and recently constructed several large test plants and
commercial plants for denitrification.
The steel industry has been most active in developing flue gas denitri-
fication technology, because combustion modification is not useful for flue
gas from iron-ore sintering machines, the major gas source at steel works.
The ceramic industry has also conducted pilot plant tests in an attempt
to reduce NOX. Flue gases from cement kilns and glass furnaces are diffi-
cult to treat by selective catalytic reduction because they contain con-
siderable amounts of particulates and alkaline vapor.
Efforts for NOx abatement by these industries will be briefly reviewed
next.
1.7.2 NOX Abatement for Utility Boilers
Utility power companies have made concerted efforts for NOx abatement
mainly by combustion modification and change of fuels. Consequently, the
emission standards have been more than adequately met. For example, flue
gas from new LNG-fired boilers contains only 35-50 ppm of NOX, while the
33
-------
emission standard is 100 ppm. Flue gas from existing oil-fired boilers
contains 100-130 ppm of NOX.
However, local governments usually impose much more stringent regula-
tions for the construction of new power plants. This necessitates flue gas
denitrification facilities. The power companies have made extensive tests
on flue gas denitrification with many pilot and prototype plants.
Table 1-14 lists major flue gas denitrification test plants and
commercial plants of power companies. Since NOX concentration is reduced
to a low level by combustion modification, 40-80 percent NOX removal by flue
gas denitrification is sufficient.
Several new coal-fired utility boilers are under construction or in
the planning .stage. Usually coal-fired boilers are installed in locations
considerably distant from cities. Although the NOX concentration of flue
gas is reduced to 200-300 ppm by combustion control, some boilers may
require a flue gas denitrification plant.
Most of the plants use SCR (selective catalytic reduction). SNR
(selective noncatalytic reduction) has been used in one plant at Chita
station, Chubu Electric, but the NOX removal efficiency is only about 45
percent, with an outlet NHs of 20-30 ppm. Large scale tests of the SNR-SCR
combination have been started at three power stations by Companies B, E and
F (the names are confidential). Their aim is to increase the NOX removal
efficiency to 50-60 percent and to decrease the outlet NHa to 10 ppm by
placing a small amount of SCR catalyst in a duct between the boiler
economizer and air preheater.
34
-------
TABLE 1-14. TEST AND COMMERCIAL PLANTS FOR UTILITY BOILERS (LARGER THAN 280,000 Nm3/hr)
to
Ui
Capacity
Power Company
Chubu Electric
Chubu Electric
Chubu Electric
Chubu Electric
Kansai Electric
Company A
Company B
Kyushu Electric
Kyushu Electric
Company C
Company D
Company E
Company F
Chugoku Electric
Chugoku Electric
Tohoku Electric
Electric Power P.C.
Hokkaido Electric
Plant Site
Chita
Chita
Chita
Chita
Kainan
—
—
Kitakyushu
Kitakyushu
—
—
—
—
Kudamatsu
Kudamatsu
Niigata
Takahara
Tomakomai
Fuel
Low-S Oil
LNG
LNG
Low-S Oil
Low-S Oil
Low-S Oil
Low-S Oil
LNG
LNG
Low-S Oil
Low-S Oil
Low-S Oil
Low-S Oil
Low-S Oil
Low-S Oil
Low-S Oil
Coal
Coal
MW
375
700
700
700
115
156
175
600
600
375
156
350
156
350
700
600
250
350x!«f
1000
Nm3 /hr
1036
1910
1910
1920
300
490
550
1610
1610
1010
490
1020
490
1000
1900
1660
800
280
Plant
Constructor
MHIa
Hitachi, Ltd.
Hitachi, Ltd.
MHI
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
MHI
MHI
MHI
MHI
IHId
IHI
IHI
IHI
IHI
—
Hitachi, Ltd.
Process
SNRb
SCRC
SCR
SCR
SCR
SCR
SNR+SCR6
SCR
SCR .
SCR
SCR
SNR+SCR
SNR'+SCR
SCR
SCR
SCR
SCR
SCR
Completion
Feb 1977
Apr 1978
Sept 1978
Feb 1980
June 1977
June 1978
June 1978
July 1978
Dec 1978
Feb 1978
July 1978
Apr 1978
June 1978
Apr 1979
July 1979
Aug 1981
June 1981
Oct 1980
NOx Removal
CO
45
80
80
80
80
508
50g
80
80
40g
40S
508
60g
80
80
60
(80)h
(90)h
{"Mitsubishi Heavy Industries.
Selective noncatalytic reduction.
^Selective catalytic reduction. .
Ishikawajima-Harima Heavy Industries.
^Combination of SNR and SCR
One-fourth of the gas from a 350 MW boiler.
?The low removal efficiency Is due to the use
L>esign value.
of a small amount of plate catalyst in an existing duct.
-------
1.7.3 Chubu Electric's Chita Power Station
Chita Station, which is the third largest power station in Japan, has
a total capacity of 3300 MW and is the most advanced in NOx control. Data
on its boilers and denitrification processes are given in Table 1-15.
TABLE 1-15. BOILER AND NOx ABATEMENT DATA FROM CHITA STATION
No. 1
Boiler 375
Capacity (MW) 375
Boiler Maker MHIa
Operation Start 1965
Fuel H0°
NOX (PPm) 90-100
Denitrification Plant
Process
Constructor
Completion
NOX After
Denitrification (ppm)
NHs Emitted (ppm)
No. 2
375
375
MHI
1966
cod
110-140
SNR
MHI
1977
60-80
20-30
No. 3 No. 4
500 700
500 700
MHI MHI
1967 1974
HO CO
90-100 95-100
SCR
MHI
1980
20
Below 10
No. 5
700
700
BHb
1978
LNG
35-50
SCR
Hitachi
1978
7-10
Below 5
No. 6
700
700
BHb
1978
LNG
35-50
SCR
Hitachi
1978
7-10
Below 5
Mitsubishi Heavy Industries
Babcock-Hitachi
Heavy oil
Crude oil
Under an agreement with local authorities (Chita City and Aichi Pre-
fecture), Chubu Electric keeps NOX in flue gas from oil-fired boilers below
100 ppm at Chita Station. The combustion modification technologies used
36
-------
include staged combustion, flue gas recirculation, low oxygen combustion,
low NOX burners, and fuel switching to low nitrogen oil containing about
0.1 percent or less N. The flue gas NOX concentration before the combustion
modification was about 400 ppm. Therefore, about 75 percent reduction has
been achieved by the modification.
The No. 5 and No. 6 LNG-fired boilers went into operation in 1978. In
order to obtain permission from local authorities for the construction of
the two new boilers, Chubu Electric, had promised to use LNG for fuel and to
install SCR units to minimize NOx emission. In addition, Chubu Electric had
promised to install an SCR unit for the No. 4 oil-fired boiler.
Flue gas from the new LNG-fired boilers contains only 35-50 ppm of NOX,
which is treated by SCR to reduce NOX below 10 ppm. The ammonia content of
the flue gas leaving the reactor is kept below 5 ppm. The SCR units were
constructed by Hitachi Ltd. Recently construction of the SCR unit for the
No. 4 boiler was started to reduce NOX from 100 ppm to 20 ppm by the MHI
process. Ammonia concentration at the reactor outlet will be kept below
10 ppm.
In addition, SNR has been applied to the flue gas from the No. 2 boiler
to reduce NOX from 110-140 ppm to 60-80 ppm (40-45 percent removal) with
20-30 ppm of ammonia in the treated gas.
Chubu Electric plans to install several new oil-fired boilers in other
power stations. All of the new boilers will have an SCR plant to remove
80 percent of NOX.
1.7.4 Efforts by the Steel Industry1*
Most of the NOX from the steel industry is derived from iron-ore
sintering machines, which use coke, while the rest comes from boilers, coke
ovens, heating furnaces and so on.
37
-------
In 1973, the Japan Steel Federation established a fund amounting to
3 billion yen ($15 million) for promoting studies on NOx, including its
formation mechanism, changes in the atmosphere, health effects, and abate-
ment technologies by combustion modification, flue gas denitrification, and
nitrogen removal from coke. Out of the fund, grants have been given to
universities for basic studies and to process developers for constructing
and operating pilot plants. In addition, the Federation has organized an
NOX study team for its own study of NOX abatement with a fund of 3.6 billion
yen ($18 million).
In 1976 and early 1977, Kawasaki Steel completed two commercial SCR
plants at the Chiba Works; this was done in accordance with an agreement
reached with Chiba Prefecture and Chiba City for pollution control under the
most stringent regulations. Kawasaki Steel signed the agreement because it
was constructing a new large blast furnace, and the agreement was needed to
start the operation of the blast furnace.
The first commercial SCR plant was constructed to treat 500,000 Nm3/hr
of gas from coke ovens owned by Kawatetsu Chemical, a subsidiary of Kawa-
saki. The second plant is for treating 750,000 Nm9/hr of flue gas from
iron ore sintering; the gas contains a considerable amount of particulates
and SO3 and is much more difficult to treat than coke oven gas. Gas from
sintering is first treated by wet FGD process and then by a wet electro-
static precipitator for further cleaning. Next it is heated to 400°C for
SCR by passing through a heater, heat exchanger, and another heater.
Although the plant has been operated with few problems, the investment and
operation costs are so high that such a process cannot be used widely.
In their search for better ways of NOX abatement, recipients of the
Federation's fund and the NOx study team have made extensive tests on flue
gas denitrification and also on nitrogen removal from coke.
38
-------
1-7.5 Nitrogen Removal from Coke1*
Most of the NOx in flue gas from iron ore sintering is derived from
coke, which usually contains about 1 percent nitrogen. Tests have been
carried out to reduce the nitrogen content by electrically heating coke up
to 1700°C. Following batch tests, a continuous test plant with a capacity
of treating 20 tons/day of coke was completed in August 1975 at the Kashima
Works, Sumitomo Metal. It has been operated since then in cooperation with
Sumitomo Heavy Industries.
A flowsheet of the NR process is shown in Figure 1-14. Pulverized
coke is preheated in a two-stage fluidized bed heater and then introduced
into an electrical furnace where it is heated up to 1700°C by the electric
current through the coke itself. The heated coke is then cooled in a
fluidized bed cooler by a circulating gas and new makeup nitrogen. The hot
gas discharged from the cooler is sent to the preheaters through cyclones.
The gas leaving the preheater is cooled and partly recycled to the fluidized
bed cooler, while the rest is sent to a stack through a baghouse.
The relation between the heating temperature achieved and N content of
the treated coke is shown in Figure 1-15. About 70 percent of the N in the
coke is removed by heating to 1700°C. By removing 70 percent of the N, the
NOx in the flue gas is also reduced by about 70 percent, as illustrated in
Figure 1-16; this indicates that most of the NOx is produced from the
nitrogen in the coke.
The pilot plant was completed in August 1975 and put into operation in
September. The largest problem has been the degradation of refractories
by local overheating and reaction with coke ash. Tests to solve the
problem have been carried out in a smaller test furnace with a capacity of
2 t/day.
39
-------
GAS COOLER
COKE
FLUIDIZED BED
PREHEATERS
COKE
FINES
Figure 1-14. Flowsheet of NR process (20 t/day).
-------
LL>
o
o
o
o
1.5
1.4-
1.2-
1.0-
0.6-
0.4-
0.2-
c/"
Untreated
1,400 1,600 1,800
HEATING TEMPERATURE (°C)
Figure 1-15. Decrease of N in coke by heating (retention time is 2 hours),
lOO
oo
S so H
^ 60 H
u_
o
z 40-
LU
5
§ 20-
x
o
rl
,700°C
rl
°1,600°C
,500'C
oi,400°C
20 40 60 80 100
N REMOVAL FROM COKE (%)
Figure 1-16. NOX abatement of flue gas from sintering machine
by N removal from coke.
41
-------
It is estimated that the process requires a large amount of electric
power, which is equivalent to 30 kWhr per ton of the sintered iron ore,
even when smooth operation is achieved. It is also necessary to analyze
the composition of the gas discharged from the furnace and to establish a
way for its treatment. At the present stage, it is not certain whether or
not the process can be commercialized.
1.7.6 Flue Gas Denitrification by the Steel Industry**
Gas from the iron ore sintering machine, which has a temperature of
about 150°C, usually contains 150-300 ppm of NOX (virtually all as NO),
150-400 ppm of SOX (mostly SOa), 200-300 mg/Nm3 of particulates containing
ferric oxide, and potassium chloride vapor.
The gas is not very suitable for most denitrification processes. For
SCR, the gas has to be heated by a heat exchanger and a heater from 150°C
to 300-400°C. Moreover, the particulates and potassium chloride vapor tend
to contaminate the catalyst. Noncatalytic ammonia reduction is not
applicable because of the much higher temperature requirement of 950-1000°C.
For wet processes providing simultaneous removal of NOx and SOx, the SOa/NOx
mole ratio is too small to reduce NOX by the effect of SOa as a reducing
agent. Among the NOx removal processes, SCR at a low temperature of 250-
300°C and simultaneous removal of SOX and NOX by activated carbon at 200-
250°C seem relatively suitable.
With the Federation's fund, 4 SCR pilot plants were constructed and
operated using the processes developed by JGC Corp. (parallel passage
reactor), Mitsubishi Kakoki Kaisha (MKK; Sarc catalyst), Sumitomo Heavy
Industries (activated carbon), and Kurabo-Niigata Iron (moving bed reactor).
The pilot plants are located in steel works. They have a capacity of
treating 1000-2000 Nm3/hr of flue gas from iron-ore sintering machines or
coke ovens. In addition, many pilot plants have been operated by the NOX
study team and also by each steel producer (Table 1-16).
42
-------
TABLE 1-16. DENITRIFICATION PLANTS OF STEEL INDUSTRY
-t-
U)
Plant Owner
Kawasaki Steel
Kawasaki Steel
*3
Kawatetsu Chemical
Kawasaki Steel
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Steel
Nippon Kokan
Sumitomo Metal
Sumitomo Metal
Sumitomo Metal
Kobe Steel
Kobe Steel
Kobe Steel
Nakayama Steel
Plant
Site
Chiba
Mizushima
Chiba
Chiba
Yawata
Kagoya
Yawata
Kimitsu
Keihin
Amagasaki
Kashima
Wakayama
Kakogawa
Kakogawa
Kakogawa
Funamachi
Process
SCR
SCR
SCR
SCR
SCR
SCR
EBd
SCR
SCR
Wet(Na)
Wet(Ca)
Carbon
Wet
SCR
SCR
SCR
Process
Developer
Hitachi Zosen
Mitsubishi Kakoki
Hitachi Ltd.
Hitachi Zosen
Nippon Steel
JGC
Ebara
JGC
Nippon Kokan
Sumitomo-Fuj ikasui
Sumitomo-Fuj ikasui
Sumitomo H.I.
Kobe Steel
Kobe Steel
Kobe Steel
Kurabo-Niigata
Gas Treated
Nm3/hr
5,000
1,500
500,000
750,000
1,600
2,000
3,000
150,000
15,000
62,000
25,000
2,000
10,000
1,000
1,000
2,000
Source
SMb
SM
coc
SM
SM
SM
SM
CO
SM
Boiler
SM
SM
SM
SM
CO
CO
Completion
1974
1975
1976
1977
1976
1976
1977
1978
1976
1973
1976
1975
1976
1976
1976
1975
Subsidiary of Kawasaki Steel.
^Sintering machine.
"Coke oven.
Electron beam radiation.
-------
1.7.7 NOX Abatement by Other Industries
The oil, petrochemical, and chemical industries have installed about
30 commercial SCR plants (Table 3-1) and 5 SNR plants (Table 5-1) for the
denitrification of flue gases from boilers and heating furnaces.
The ceramic industry has also emitted a considerable amount of NOx.
Cement kilns and glass furnaces are two major sources of NOx from the indus-
try. A rotary kiln for cement calcination has a wide range of hot zones
and gives off 400-800 ppm of NOx even when fired with oil. Pilot plant
tests have been carried out by the Cement Industry Association in an attempt
to reduce NOX emissions by flue gas denitrification using SCR, but the tests
were not successful because the gas has a high particulate loading and
contains an appreciable amount of alkaline compounds which affects the
catalyst. A new type of cement kiln will be installed in new plants. The
new type kiln has a suspension preheater for which more than half of the
fuel is used; this results in a considerable reduction of the hot zone area.
The NOX concentration in the flue gas is reduced to about 150 ppm.
Flue gas from a glass furnace contains a considerable amount of sodium
sulfate vapor, which deposits on the SCR catalyst and lowers its activity.
Either the alkaline vapor has to be removed in advance, or the catalyst has
to be washed to remove the sulfate. Tests to reduce NOX to Nz by reducing
gases without a catalyst have also been carried out. Nonetheless, flue gas
denitrification continues to be a problem for the gas from glass furnaces.
Tail gas from nitric acid plants contains 500-2000 ppm of NOX. Sodium
scrubbing to produce sodium nitrite as a by-product, catalytic reduction to
convert N0>< to N2, and absorption using zeolite (molecular sieve) have been
commercially applied. Treatment of the tail gas is not difficult because
the gas is clean, containing no SOX and particulates.
44
-------
1.8 REFERENCES
1. Air Preservation Bureau, Environment Agency. Cost Effects of NOX
Abatement. April 1978.
2. Environment Agency. Recent Countermeasures for Air Pollution Control
in Japan. (After OECD Review of Japanese Environmental Politics in
1976.) Japan, 1977. (In English.)
3. MITI, Industrial Pollution Control Association. Countermeasure of NOX
Pollution. Japan, December 1977.
4. Japan Steel Federation. Present Status of Development of NOX Abatement
Technology by the Steel Industry. April 1977.
45
-------
SECTION 2
NOX ABATEMENT BY COMBUSTION MODIFICATION (CM)
2.1 CLASSIFICATION OF COMBUSTION MODIFICATION TECHNIQUES
NOX emissions from combustion processes are formed by two mechanisms:
the oxidation of nitrogen in air at high temperatures, and the combustion
of various nitrogen compounds contained in fuel. The former is called
"thermal NOX," and the latter "fuel NOX."
Formation of thermal NOX can be reduced effectively by combustion
modifications, such as decreasing the oxygen concentration in the combustion
regions; shortening the residence time of combustion gases in high tempera-
ture zones; or lowering the flame temperature. Reduction of fuel NOX emis-
sions is achieved by decreasing the oxygen concentration in the combustion
reaction zone and switching to a low-nitrogen fuel.
Combustion modification techniques used widely in Japan for NOX control
can be classified into four categories:
Change of operating conditions
1) Low-excess-air combustion
2) Change in contacting of fuel with air in combustion chamber
3) Reduction of heat load in combustion chamber (reduction of
output power)
4) Lowering of air preheating temperature
• Modification of burner design
46
-------
• Modification of combustion system design
1) Reduction of heat load in combustion chamber
(increase of furnace volume)
2) Staged combustion
3) Flue gas recirculation
4) Water or steam injection
Other methods
1) Change of fuel
2) Modification of firing
Combinations of the above techniques are also in practical use.
Although operating conditions can be changed relatively easily in existing
installations, the changes usually reduce the NOX emissions only slightly
and often cause operating difficulties. On the other hand, modifications
of the burner and combustion system design are promising control techniques.
Fewer problems are encountered when design modifications are incorporated
in new plants; relatively large reconstruction costs are required when the
modifications are applied to existing installations.
The principles and characteristics of NOX control techniques are
described in the following section.
2.2 CHANGE OF OPERATING CONDITIONS
2.2.1 Low-Excess-Air Combustion (LEA)
Low-excess-air combustion, a means of preventing low-temperature corro-
sion caused by SOs in large boilers, is also useful in NOX abatement. LEA
combustion reduces the formation of thermal NOX and fuel NOX by decreasing
the oxygen concentration in combustion regions. There it has been applied
47
-------
to large boilers burning high-nitrogen fuels, such as grade B (0.08-0.35
percent N) or grade C (0.20-0.35 percent N) heavy oil. On the other hand,
the technique is not widely used with small boilers. The reasons are:
• Small boilers usually burn low-nitrogen fuels such as kerosene,
grade A heavy oil (0.005-0.08 percent N), or gas. Thus, there
is no need to attempt fuel NOX reduction.
• 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.
Recently, LEA combustion has been applied to large metal heating
furnaces as a means of NOX abatement and fuel conservation.1
2.2.2 Improvement of Fuel/Air Mixing Conditions in Combustion Chamber
It has been reported that NOX is reduced considerably through a slight
adjustment of the air register, for example, a change in the vane angle.
The reduction may be due to the change in mixing condition in the combustion
chamber. Few existing installations use this technique.
Reduction of NOX by a change in mixing conditions is achieved either
with very good or poor mixing. With good mixing, the combustion rate
increases and temperature in the reaction zone becomes high. Yet thermal
NOX formation can be reduced because the residence time in the high tempera-
ture region can be shortened. When the mixing condition is poor, the com-
bustion rate slows down and the flame temperature drops. This results in
the reduction of both thermal and fuel NOX emissions in the same way as in
the staged-combustion type burner. (The latter is discussed in section
2.3.5.)
48
-------
Changing the air system makes it necessary to readjust the air/fuel
ratio. An improper adjustment may increase the formation of unburned
products such as soot, carbon monoxide, and hydrocarbons, or may cause
unstable combustion.
2.2.3 Reduction of Heat Release Rate in the Combustion Chamber
(Reduction of Output Power)
Generally, when the heat load in the combustion chamber decreases,
the flame temperature is lowered and thermal NOX emissions are reduced.
The use of this method in existing installations, however, results in a
decrease in fuel efficiency and output power. Therefore, this method is
regarded only as an emergency measure.
2.2.4 Reduction of Air Preheating
Lowering of the air preheating temperature reduces the flame tempera-
ture and consequently the formation of thermal NOX. This method is seldom
used in existing installations because it also reduces fuel efficiency.
Moreover, its application is limited to large boilers with air preheaters,
for which better control techniques are available.
2.3 MODIFICATION OF BURNER DESIGN
2.3.1 Introduction
This method reduces NOX with the use of an improved burner, the
so-called low-NOx burner (LNB)- Low-N0x burners developed in Japan are
mostly applied to boilers and various kinds of furnaces. They may be
classified into five types:
Good-mixing
• Divided-flame
49
-------
Self-recirculation
• Staged-combustion
Combination
Large low-NOx burners developed so far have capacities of 5 x 10
kcal/hr for boilers and 2 x 107 kcal/hr for heating furnaces which are
large enough for commercial use.
2.3.2 Atomizing Type Burners
Atomizing type low-NOx burners are useful in reducing the formation
of thermal NOX by shortening the residence time of combustion gas in high
temperature regions. They have little effect on reducing fuel NOX, however,
because of the high oxygen concentration in the initial stage of combustion.
An example of an atomizing type low-NOx burner is illustrated in
Figure 2-1. This burner, developed by Nippon Furnace Kogyo Kaisha Ltd.,
is called the NFK-TRW burner because it has been commercialized under a
technical license agreement with Civiltech Corp. (U.S.) in a joint venture
with TRW Inc.
ATOMIZING
AIR OR STEAM
COMBUSTION
AIR
DEFLECTOR
ATOMIZED FUEL
COMBUSTION AIR
Figure 2-1. NFK-TRW Burner.
50
-------
In this burner, atomizing of fuel with air is achieved by impinging
the air stream at nearly right angles against the jets of fuel injected
radially outward through a large number of shaped ports. The air and fuel
are then further mixed by means of a deflector, which functions as a flame
holder. The mixture of fuel and air produces a thin, conical flame accord-
ing to the flow pattern. Due to the concical shape, a maximum radiation
surface is produced, and heat from the flame is rapidly removed. Therefore,
the residence time of combustion gas in the high temperature regions is
shortened, and thermal NOX is reduced. The NO concentrations produced
using the burner at various operating conditions are shown in Figure 2-2.
Burners which can burn either oil or gas have been commercialized. They
reduce not only NOX, but also soot emissions. In addition, they allow the
heat load in the combustion chamber to be maintained as high as before,
thereby increasing fuel efficiency. These burners are applicable only to
boilers, however, because of their flame shape. The pressure drop at the
burner throat is relatively high.
CVJ
o
O
UJ
o
ox
80
60
40
20
0
100% LOAD
1.0 1.2 1.4 1,6
EXCESS AIR FACTOR
Figure 2-2. Effect of NFK-TRW Burner on NOX.
(Fuel, grade A fuel oil; air temperature,
25°C; maximum fuel flow rate, 380 £/hr)
51
-------
2.3.3 Divided-Flame Type Burners
In the divided-flame type low-NOx burner for heavy oil, the flame is
divided into several independent small flames by means of grooved nozzle
tips. This prevents mutual interference between small flames. Hence, the
flame temperature and the residence time in the high temperature regions
decrease, resulting in thermal NOx reduction.
The burner, developed by Ishikawajima-Harima Heavy Industries Ltd.
(IHI), is shown in Figure 2-3. Its effect when combined with other NOX
control techniques is illustrated in Figure 2-4. The two curves indicate
the practical upper limit and lower limit. NOX reduction with this burner
decreases with an increase in the effects, of other control techniques. For
example, the reduction ratio due to the effect of the burner ranges from
18-42 percent without other CM techniques and from about 10-30 percent when
40 percent reduction is achieved by other techniques.
2.3.4 Self-Recirculation Type Burners
The configuration of the self-recirculation type burner is shown in
Figure 2-5. This burner has been developed by Daido Tokushuko Co. under
a technical license agreement with Caloric Gesellschaft fur Apparate Bau
(West Germany) . "*
It reduces both thermal and fuel NOX formation in such a manner that
partially combusted hot gas is reintroduced into the initial stage of combus-
tion by the fuel jet and atomizing air or steam. Thus, both vaporization of
fuel and firing in the low-oxygen-concentration zone are accomplished in the
initial stage of combustion.
This burner's effect on NOX reduction is illustrated in Figure 2-6.
This type of burner is effective not only for NOX reduction, but also for
soot suppression by hot combustion gas recirculation. It is used in metal
heating furnaces that burn oil and gas.
52
-------
Fuel Injection Port
— Grooves
Direction of Fuel Spray
Nozzle Tip
Details o
Fuel Injection Port
Figure 2-3. Divided-flame type burner.
53
-------
Other Means
None
Atomizing
Pressure Steam
v o
Flue Gas Redrculatlon (A)-
Staged Combustion (B) •
Water Injection (C)
AJ + (c;
A + B
A
_ _ _ __ A
V
Oi
C CO
o
•ou
-------
Ignition Plug
Combustion A1r
Figure 2-5. Self-recirculation type burner.
150
Q.
Q.
04
O
O
•p
•o
0)
4J
O
i.
O
CJ
100
50
Gun Type
Rotary Type
Oil Pressure Typ-e
Steam Atomizing Type
a
lea
i,
-------
A burner designed exclusively for gas firing is shown in Figure 2-7.
This burner developed by Sanray Reinetsu Co., Ltd., is called the XB burner.
Several fuel gas injection nozzles are installed circularly to shorten the
flame. NOX levels are presented in Figure 2-8. A similar type of burner
has been developed by Chugai Ro Kogyo Co., Ltd.
RECIRCULATED
COMBUSTION GAS
No. 1 PILOT BURNER
No. 2 PILOT BURNER
RECIRCULATED
COMBUSTION GAS
Figure 2-7- XB Burner.
56
-------
CM
o
ui
o
o
LU
DC
Of.
O
o
X
o
80
60
40
20
0
CITY GAS FIRING, NOX
• • A
CITY GAS FIRING/02%
0 2 4 6 8 10 X TO5
THERMAL INPUT, kcal/hr
o
H-1
£
LU
O
(NJ
O
Figure 2-8. Effect of XB burner on NOX emission levels.
2.3.5 Staged-Combustion Type Burners
Staged-combustion type burners are divided roughly into a two-stage
combustion group and an of f-stoichiometric combustion group. Burners in
the two-stage combustion group are classified according to the equivalence
ratio in the first stage into fuel-rich and air-rich. The fuel-rich type
is used usually for two-stage combustion, and the air is added in the second
stage; fuel-rich type burners are shown in Figures 2-9, 2-11, and 2-13. In
the air-rich type, fuel is injected in the second stage; an example of this
type of burner is presented in Figure 2-15.
Figure 2-9 illustrates the APOC burner developed by Tokyo Gas Company
for firing town gas.5 A more advanced burner of this type has a catalyst
at the outlet of the preliminary combustion region for further reduction of
NOX. Figure 2-10 shows the emission levels with this burner. HCN and NHa
are likely to play a significant role in NOX formation in the second stage.
The catalyst decreases this problem by reducing HCN and NHs together with
NOX in the first stage. The advanced type is not yet used commercially.
57
-------
GAS OR
GAS PRE-
MIXED WITH
AIR
FIRST STAGE
SECOND STAGE
u:
II
AIR
Figure 2-9. APOC burner.
Figure 2-10.
40 60 80 100
PREMIXING RATIO
(AIR/AIR + FUEL), *
NOX emission levels from a low-NOx burner more
advanced than the one shown in Figure 2-9.
(Fuel, methane; thermal input, 95,000 kcal/hr;
NOX concentration corrected to 0 percent 02)
58
-------
The FH burner illustrated in Figure 2-11, used for LPG, heavy oil or
kerosene fuels, is also of the fuel-rich type.6 Figure 2-12 shows its effect
on NOX reduction. This burner, developed by Chugai Ro Kogyo Co., Ltd., is
used in the heating furnace. Similar types of burners have been developed
by many manufacturers because of their relatively simple configuration and
the capability of fuel NOX reduction. These manufacturers include Mitsubishi
Heavy Industries, Ltd., Kawasaki Heavy Industries, Ltd., Osaka Gas Company,
Sanray Reinetsu Co., Ltd. and Rozai Kogyo Co., Ltd.
The fuel-rich type burner shown in Figure 2-13 was designed for use
with pulverized coal by Babcock Hitachi Co., Ltd.; it is called the dual-
air register type burner and is used in utility boilers. This burner is
effective for NOX reduction, as shown in Figure 2-14. IHI has developed
a similar type of burner.
o
An air-rich type two-stage combustion burner is shown in Figure 2-15.
This burner, developed by Tokyo Gas Company for gas firing, is called the
TGS burner. The effect of the TGS burner on NOX reduction is illustrated
in Figure 2-16. In the first stage, the flame temperature is decreased due
to high-excess-air combustion; in the second stage, the oxygen concentration
is lowered due to low-excess-air combustion. These decreases in flame tem-
perature and oxygen concentration provide NOx reduction. It is unlikely,
however, that complete combustion can be accomplished for oil or coal firing.
The basic structure of the off-stoichiometric combustion (OSC) type
burner for oil firing is presented in Figure 2-17. This burner has been
developed by MHI for boilers using grade A heavy oil or kerosene. The fuel
sprayed in the upper and lower compartments is evaporated by preheated air
and premixed with air resulting in combustion. In the middle compartment,
a normal diffusion flame is produced. Thus, off-stoichiometric combustion
is realized. To reduce NOX further, recirculated flue gas is mixed with air.
MHI guarantees less than 70 ppm of NOX with this system (see Figure 2-18).
59
-------
PRIMARY AIR
FUEL GAS
SECONDARY AIR
Figure 2-11. Schematic of FH burner.
UJ
on
i
UJ
n
UJ
(-
Q.
a.
n
CVJ
o
--a
-
0
i , i
ULI
O
UJ
g;
0
0
X
0
I4UU
1350
i "^nn
i ouu
•ann
JUU
250
200
150
100
50
n
1 ,
- t 8
i i
^-''*"" / ~~~-
CONVENTIONAL BURNER
"™^^W— — ^^^
*• -- _
NOV REDUCTION RATIO '
X
—
_
FH BURNER
\f
. C
— o — - —
1 1
100-
80-
t
60-
i
1 2
02 CONCENTRATION IN FLUE GAS, %
I
1.00
1.05 1.10
EXCESS AIR FACTOR
1.15
O
UJ
OL
X
o
Figure 2-12. FH burner's effect on NOX reduction.
60
-------
VENTURI PLUG
TERNARY
AIR REGISTER
PILOT BURNER
OIL BURNER
SECONDARY
AIR REGISTER
PULVERIZED COAL
AND PRIMARY AIR
COOLING
AIR
SECONDARY TERNARy
1 AIR PULVERIZED COAL
SECONDARY
VANE
INJECTION PORT
Figure 2-13. Dual-air register type burner for
use with pulverized coal.
z
o
M
CO
CO
i
Pw
o
o
M
H
90 100 110 120
RATIO OF AIR FLOW RATE
IN FIRST STAGE
(100Z - stoichiotnetric)
Figure 2-14.
Effect of low-NOx burner shown in Figure 2-13.
(Boiler capacity, 810 tons/hr; nitrogen content
in fuel, 0.89-1.21 percent)
61
-------
PREMIXED
FUEL GAS
AND AIR
FUEL GAS
Figure 2-15. Schematic of TGS burner.
cu
P.
cv|
o
8
W
ei
§
u
o
z
150
100
50
CONVENTIONAL
TWO-STAGE COMBUSTION
TYPE BURNER
TGS BURNER
I
0.4 0.8 1.2 1.6
EXCESS AIR FACTOR IN
FIRST STAGE
Figure 2-16. Effect of TGS burner on NOX reduction.
(Fuel, methane; thermal input, 2 x 105
kcal/hr; total excess air factor, 1.10;
furnace temperature, 1,200°C)
62
-------
AIR RICH PREMIXED FLAME ^
DIFFUSION FLAME J
AIR RICH PREMIXED FLAME B
UPPER COMPARTMENT
MIDDLE COMPARTMENT
LOWER COMPARTMENT
BURNER GUN
Figure 2-17. Basic structure of OSC
type burner.
CORRECTED TO
02, ppm
ou
60
40
20
1 1 1 1
CN
1.0 1.2 1.4
EXCESS AIR FACTOR
Figure 2-18. NOX emission levels of the
burner shown in Figure 2-17
with propane.
(Air temperature, 250°C:
thermal input, 3.5 x 10 kcal/hr;
flue gas recirculation ratio, 15
percent)
63
-------
Figure 2-19 shows the arrangement of the fuel injection holes of
atomizing nozzles in an off-stoichiometric combustion burner for oil firing;
this OSC burner was developed by Volcano Company. ° The main feature of
its atomizer is that not all of the fuel-injection holes are of 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. Since several different arrange-
ments 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. NOX levels produced by such a
burner are presented in Figure 2-20. This burner tends to increase soot
emissions when excess air is low, because some of the flame zones are defi-
cient in oxygen. A similar type of burner has been developed by Babcock
Hitachi Co., Ltd.
2.3.6 Combination Type Burners
There are several low-NOx burners which combine the above-mentioned
methods. For example, the burner illustrated in Figure 2-21 combines two-
stage combustion and self-recirculation of hot combustion gas. It is called
the SRG burner, and was developed by Nippon Furnace Kogyo Kaisha, Ltd.2
This burner is adopted mainly in oil heating furnaces and boilers. The
effect of the burner on NOX reduction is shown in Figure 2-22. Figure
2-23 illustrates a similar type of burner (the TCG) developed by Osaka
Gas Company. l
64
-------
OIL-
STEAM-
Figure 2-19.
Atomizing nozzles in off-stoichiometric
combustion-type low-NOx burner for oil.
400 >
s
Q.
0.
o* 300
200
100
I
60
30
LOAD, %
100
Figure 2-20. Effect of low-NOx atomizer shown in
Figure 2-19 on NOX emissions.
(Boiler capacity, 55 t/hr; fuel, grade
C heavy oil; air temperature, 280°C)
65
-------
.GASIFICATION GAS>
I CO, H2 RICH '
RE-COMBUSTION GAS,
GASIFICATION REACTION
C + 02—C02
C02 + C=*
C + H20^±(
Qnlln + mH,(
COMBUSTION PRODUCTS
SECONDARY AIR
FUEI
QnHn
Figure 2-21. Schematic of SRG burner.
.
a.
•
eFZSO
S 200
° 150
-. 100
50
NOX vs AIR RATIO. SRG BURNER
AND TA'NDEH BURNER
B FUEL OIL 105 t/hr
g 60
g *^ 40
I £ 20
1.0 1.1 1.2 1.3 1.4 1.5
AIR RATIO. \
NOX REDUCTION RATE. SRG vs TANDEM
qt 1 L
1.0 1.1 1.
2 1.3 1.4 1.5
AIR RATIO, \
Figure 2-22. Effect of SRG burner on NOX reduction.
66
-------
RECIRCULATED
COMBUSTION GAS
GAS
FOR
IGNITION
PRIMARY
AIR GAS
FOR
IGNITION
PRIMARY AND
SECONDARY AIR
TERNARY
AIR
Figure 2-23. ICG burner.
67
-------
2.4 MODIFICATION OF COMBUSTION SYSTEM DESIGN
2.4.1 Reduction of Heat Release Rate in the Combustion Chamber
(Increase of Furnace Volume)
Manufacturers currently are designing boilers and furnaces with heat
release rates about 10 to 20 percent lower than those of conventional ones.
The new boilers and furnaces are larger and more easily adaptable to NOX
control techniques.
2.4.2 Staged Combustion
There are two major categories of staged combustion:
1) Two stage combustion (TSC)*
In two-stage combustion, about 70 to 90 percent of the stoichiometric
air is admitted into the first stage. The additional air is then added
downstream to obtain complete combustion. With this technique, both thermal
and fuel NOX are reduced by fuel-rich (low-oxygen concentration) combustion
in the first stage.
Two stage combustion is classified into four types, depending on the
locations of second-stage air ports:
a) On the furnace wall above the burners.
b) Top burners for air injection only.
c) Side or rear walls of the furnace.
d) On the circumference of the burners.
Large boilers installed in power plants employ type a or b. Type b,
called quasi-two-stage combustion, is used in units to which type a cannot
be applied. Types c and d are used in medium- and small-capacity boilers.
*The number of stages seldom exceeds two.
68
-------
Type c is not popular for retrofitting water tube boilers because it requires
considerable revisions. Type d, which is illustrated in Figure 2-24, can be
readily applied to small installations with a single burner because it calls
for minor revisions. Two-stage combustion cannot be applied to installations
with furnace dimensions that cannot accommodate greater flame lengths. Also
two-stage combustion tends to increase the amount of unburned products; thus
the unconsumed air increases in the flue as with normal combustion. When the
air injection ports in the second stage are suitably located, further NO
reduction is obtained.12
FUEL-
0X2
OF SECOND-
STAGE AIR
FIRST-STAGE SECOND-STAGE
AIR AIR
Figure 2-24. Two-stage combustion for small boiler.
2) Off-stoichiometric combustion (OSC)
This method has an effect similar to that of two-stage combustion; it
is readily adapted to medium- and small-size boilers with several burners,
to which two-stage combustion is not easily applied.
In this technique, the fuel-rich burners and air rich burners or air
ports are suitably located. The fuel-rich portions reduce the oxygen con-
centration, and the air-rich portions decrease the temperature by means of
69
-------
the excess air; thus, NOX formation is reduced. Combustion in the fuel-rich
portions is completed upon mixing due to the excess air in the air-rich
portions.
The location of fuel-rich and air-rich burners or air ports is decided
after a systematic test. Often, it is effective to place the air-rich
burners or air ports in the central upper parts of the furnace walls, or
in the regions of highest heat release.
2.4.3 Flue Gas Raeirgulation JFGR)
In flua gas rseireulation, part of the flue gas is mixed with combustion
air and sent into the furnace. Thus, NOX is reduced through the decrease of
flame temperature resulting from the increase of the total heat capacity of
combustion gas. Thermal NOX formation can therefore be reduced.
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 condi-
tions and decrease the soot formation. However, the decrease of flame tem-
perature alters the distribution of heat transfer and lowers the fuel
efficiency of existing installations.
A recirculation fan and additional duct work are required to implement
flue gas recirculation. As a result, installation cost is considerably
higher, and more installation space is required than for the other methods.
Therefore, this method is not used with small boilers or furnaces.
2.4.4 Water or Steam Injection
Water or steam injection reduces thermal NOX emisssions mainly by
decreasing the flame temperature. There are two injection methods:
70
-------
1) Injection into the combustion chamber. (This includes
increasing the steam flow rate in the atomizer.)
2) Mixing water with the fuel (emulsification).
Injection ports close to the burners are effective for the first
method. Thus, burners with water injection ports are adopted. The upper
limit of the injection rate is about 5 kg/101* kcal.
With the us© of this technique, formation of soot decreases, so it is
possible to reduce the excess air. Therefore, the decrease in thermal effi-
ciency may not be great.
There are two kinds of emulsified fuel: water droplets dispersed in
oil (W/0 type) and oil dispersed in water (0/W type). The former type can
be used for NOX reducing fuel because the resulting emulsion viscosity is
lower than that of the 0/W type reducing the pumping power requirements.
With the use of emulsified fuel oil, the atomized oil droplets are
separated into very fine particles by sudden expansion of water contained
in the oil droplets during the initial stage of combustion; this increases
the contact area of fuel with air and also increases the combustion speed.
These facts are likely to contribute to the reduction of thermal NOX forma-
tion. Since water or steam is effective for suppression of soot formation,
excess air can be reduced, and the decrease in thermal efficiency is not
great. The increase of water concentration in the flue gas is small, so
that its influence on metal corrosion can be neglected.
In the fuel emulsion method, usually a slight amount of emulsifying
agent is used. Recently, a method that requires no agent has been developed
through improvement of the mixer.
71
-------
The price of a fuel emulsion system is about 50 x 105 to 60 x 105 yen
($0.25 x 105 to $0.30 x 105) for a capacity of 1 kl/hr, and the cost of an
emulsifying agent is about 100 yen ($0.50) per kiloliter.
Since it does not reduce fuel NOX, this method is useful with low
nitrogen oils.
2.5 OTHER METHODS
2.5.1 Change of Fuel
Except for special cases, NOX concentrations are related to the
nitrogen content of the fuel, which decreases in the following order:
solid fuels (coal and coke), liquid fuels (petroleum fuel oils), and
gaseous fuels (town gas, LNG and LPG).
Emissions from gas burning decrease in the following order: LPG (CsHs
and CijHio), LNG (CHi,), and town gas (synthetic gas with a heating value of
about 5,000 kcal/Nm3).
Hydrodesulfurization of heavy oil, which reduces sulfur content from
2-4 percent to 0.1-0.7 percent, also removes 20-30 percent of nitrogen,
which is usually present in a concentration of 0.2-0.3 percent in the oil.
2.5.2 Modification of Firing
It is well-known that tangential firing systems give lower NOX emissions
than front and opposed firing systems, because the flame temperature is
lower. This lower temperature is caused by better heat transfer resulting
from a larger flame volume.
72
-------
In small boilers, reversely turned firing is used to reduce NOX
emissions. This type of firing is illustrated in Figure 2-25. NOX reduc-
tion is obtained by recirculation in the flue tube.
SMOKE TUBE
FUEL
ISMOK
(T
FLUE' TUBE
D
AIR
Figure 2-25. Flow pattern in reversely turned firing.
2.5.3 Combination Techniques
Combinations of combustion modification techniques can increase the
reduction of NOX emissions and, therefore, are used widely. However, a
combination of techniques based on the same suppression principles, such
as the combination of two-stage and off-stoichiometric combustion, is not
efficient.
73
-------
2.6 COMMERCIAL APPLICATION OF COMBUSTION MODIFICATION (CM)13'1^
2.6.1 Boilers
Table 2-1 shows the total number of boilers in use in 1974 and the
total number of boilers using CM techniques in 1977- Most of the boilers
with a capacity over 500,000 Nm3/hr (167 MW equivalent) are utility boilers,
while smaller ones are industrial boilers.
TABLE 2-1. TOTAL NUMBER OF BOILERS INSTALLED BY 1974 AND TOTAL NUMBER OF
BOILERS WITH COMBUSTION MODIFICATION (CM) IN 1977
Capacity
(1.000
Nm /hr)
Above 500
100-500
40-100
10-40
5-10
Below 5
Total
Existing Boilersa
Total
(A)
154
331
407
2,683
3,696 )
72,313 t
79,584
With CM
(B)
153
155
72
41
18
439
B/A
(percent)
99.3
46.8
17.7
1.5
0.02
0.6
New Boilers
with CMb
22
30
9
24
11
96
Total
Boilers
with CM
175
185
81
65
29
535
aBoilers installed by 1974.
^Boilers installed between 1975 and 1977.
The NOX emission standards were first applied for large stationary
sources and have gradually been extended to smaller ones. As a result, the
CM application ratio (B/A) is higher with larger boiler capacity.
The number of small boilers using CM may not be very reliable because
the research was done mainly for large installations.
74
-------
NOX emission levels of boilers before and after the use of CM are
given in Table 2-2. The NOX reduction ratio ranges from 20 to 50 percent
with different fuels. New boilers tend to give lower NOX emissions than
do existing boilers, except for boilers using grades A and B oil. The
figures for the latter may not be very reliable because of the limited
number of boilers investigated.
The types of CM techniques used for boilers are presented in Table 2-3
(low-excess-air combustion is not listed). In general, the low-NOx burner
(LNB), two-stage combustion (TSC) and a combination of two-stage combustion
and flue gas recirculation (FGR) are used widely. Over 70 percent of large
boilers (over 500,000 Nm3/hr) use TSC and a combination of TSC and FGR.
Many small boilers which have flue gas flow rates under 100,000 Nm3/hr use
low-NOx burners.
Table 2-4 gives the effects of CM techniques on existing boilers.
The combination of two methods (TSC and FGR) and the combination of three
methods (LNB, TSC and FGR) give high efficiencies of 42-44 percent. The
combination of TSC and FGR is used for a large number of boilers. The
reduction ratio with a low-NOx burner is 27 percent and is equal to that
of two-stage combustion. Low-N0x burners and two-stage combustion are also
used widely. Off-stoichiometric combustion (OSC) requires a low investment
cost and gives a relatively low NOX reduction.
Table 2-5 shows typical NOX levels attainable by CM techniques for
large boilers (above 500,000 Nm3/hr), as indicated by burner and boiler
manufacturers. There are a few manufacturers who indicate values 20-30
percent lower than those shown in Table 2-5. Even lower concentrations
may be attained for smaller boilers. Comparison of Table 2-5 with Table
2-2 shows that further NOX abatement may.be achieved by CM to a considerable
extent for many NOX sources.
75
-------
TABLE 2-2. BOILER NOX EMISSION LEVELS BEFORE AND AFTER COMBUSTION MODIFICATION (CM)
Fuel Existing Boilers
Before CM After CM
Reduction
Max. Min. Mean Max. Min. Mean Ratiob
Coal 800 250 560 650 200 446 20.4
Grade C oil 600 100 273 320 80 176 35.5
and crude oil
Grade A and 290 50 134 150 30 77 42.5
B oils
Naphtha, 310 70 233 230 30 117 49.8
kerosene and
by-product oil
Gaseous fuel 250 50 153 170 20 101 34.0
New Boilers (with CM)
Max . Min . Mean
—
240 60 162
150 50 107
110 30 87
130 30 75
In ppm, corrected to 4 percent QZ for oil, 5 percent Oz for gas, and 6 percent for coal.
'Reduction ratio = 100 x [(mean before CM-mean after CM)/mean before CM].
-------
TABLE 2-3. NUMBER OF BOILERS WITH COMBUSTION MODIFICATION TECHNIQUES
Capacity
(1,000 m
Over
500
100-500
40-100
10-40
Below
10
Total
/hr)
Existing
Hew
Subtotal
Existing
Hew
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Existing
Hew
Subtotal
Existing
New
Subtotal
aLNB = Low HOX burner;
EF = Emulsified fuel.
LHBa
4
0
4( 2)
38
1
39(21)
33
2
35(43)
11
7
18(28)
5
9
14(48)
91
19
110(21)
ISC =
TSC
35
6
b 41(23)
21
5
26(14)
4
1
5( 6)
10
4
14(22)
0
0
0( 0)
70
16
86(16)
Two-stage
FGR
21
0
21(12)
8
2
10 ( 5)
3
1
4( 5)
2
5
7(11)
2
0
2( 7)
36
8
44 ( 8)
combustion;
OSC
0
1
K 1)
12
1
13 ( 7)
13
0
13(16)
1
0
K 2)
0
0
0( 0)
26
2
28( 5)
FGR = Flue
LNB +
SI FGR
0 0
0 0
0(0) 0(0)
1 6
0 0
Id) 6(3)
1 2
0 1
1(1) 3(4)
3 0
0 0
3(5) 0(0)
1 1
0 1
1(3) 2(7)
6 9
0 2
6(1) 11(2)
gas recirculation
TSC +
FGR
73
12
85(49)
15
0
15 ( 8)
2
3
5( 6)
2
5
7(11)
1
0
K 3)
93
20
113(21)
; OSC =
LNB + TSC
+ FGR
7
0
7( 4)
8
16
24(13)
0
1
K 1)
0
0
0( 0)
0
0
0( 0)
15
17
32 ( 6)
EF
0
0
0(0)
0
0
0(0)
3
0
3(4)
2
0
2(3)
0
0
0(0)
5
0
5(1)
LHB +
TSC
4
0
4( 2)
25
1
26(14)
3
0
3( 4)
2
0
2( 3)
1
1
2( 7)
35
2
37( 7)
Off-stolchiometric combustion;
LNB +•
OSC
2
0
2(1)
13
0
13(7)
0
0
0(0)
2
0
2(3)
0
0
0(0)
17
0
17(3)
SI = Steal
Other
7
3
10 ( 6)
8
4
12 ( 6)
8
0
8(10)
6
3
9(14)
7
0
7(24)
36
10
46 ( 9)
Subtotal
153
22
175(100)
155
30
185(100)
72
9
81(100)
41
24
65(100)
18
11
29(100)
439
96
535(100)
n injection;
Figures in parentheses are proportions (percent) in all types.
-------
TABLE 2-4. EFFECTS OF COMBUSTION MODIFICATION TECHNIQUES
Type of CM Technique
OSC
LNB
TSC
FGR
LNB + OSC
LNB + TSC
TSC + FGR
LNB 4- TSC + FGR
Other
Mean Reduction
Ratio (percent)
18
27
27
32
25
34
42
44
41
Number of Boilers
with CM Techniques
24
80
72
25
17
32
92
16
31
TABLE 2-5. ATTAINABLE NOX EMISSION LEVELS (PPM) INDICATED BY
BURNER AND BOILER MANUFACTURERS*
Fuel Existing Boilers New Boilers
Grades C and B Oil 130-160 110-120
(0.15-0.2 percent N)
Gas 100-130 50
Coal 350 300
^Typical values for large boilers, corrected to 4 percent Oz for oil,
5 percent 02 for gas, and 6 percent Oz for coal.
78
-------
The investment costs for CM techniques are shown in Table 2-6. In
general, countermeasures including flue gas recirculation are expensive,
while a low NOX burner and off-stoichiometric combustion are less expensive.
Figure 2-26 illustrates cost performance, that is, the cost of equipment
for every percent of reduction ratio, calculated from Tables 2-4 and 2-6.
Generally, an expensive countermeasure gives a high NOX reduction ratio.
A low NOX burner (LNB) is advantageous over two-stage combustion (TSC)
because of the lower cost for the same reduction ratio.
2.6.2 Oil Heating Furnaces
Figure 2-27 shows NOX emission levels of oil heating furnaces using
CM. In the same manner as boilers, new furnaces have lower emissions than
those of existing ones, and NOX emissions decrease with low-nitrogen fuels.
CM techniques for oil heating furnaces are shown in Table 2-7. Low-N0x
burners are used widely for both new and existing furnaces regardless of
capacity. Fewer CM techniques have been applied to oil heating furnaces
than to boilers. The combination of two-stage combustion and flue gas
recirculation, used widely for boilers, has not been adopted. The reasons
are:
1) The great majority of oil heating furnaces are natural-draft
units; therefore, extensive reconstruction is required for
the application of the above combination.
2) Low-N0x burners used in oil heating furnaces combine self-
recirculation and two-stage combustion.
79
-------
TABLE 2-6. INVESTMENT COST FOR COMBUSTION MODIFICATION
Control Technique
Cost (yen/Nm3/hr)
LNB + TSC + FGR
LNB + FGR
TSC + FGR
LNB + TSC
TSC
LNB
OSC
657 ($3.29)
491 ($2.46)
475 ($2.38)
480 ($2.40)
292 ($1.46)
129 ($0.65)*
26 ($0.13)
Mean 467 ($2.34)
* Including change of burner tip.
s-
>>
r—
(O
O
0)
S.
^^"^
S-
J=
•*>
E
"— -
ft
eu
o
c
S-
dl
Q.
-I-J
(/I
O
O
16
14
12
10
8
6
4
2
n
LNB + TSC + T + GR
O
LNB + TSC
0
TSC
O
•
•
LNB
O
OSC
0
1 1 1 I 1
10 20 30 40
Reduction Ratio, %
60
Figure 2-26. Cost performance of NOX control techniques
applied to boilers.
80
-------
NO
Grade C Oil
•5
\
New ,_^ Existing
Furnace — ^ Furnace
1
50 70 90 110 130 150 170 NOX ppm
c
=3
-------
TABLE 2-7. CM TECHNIQUES USED FOR OIL HEATING FURNACES
oo
Capacity
(1,000
Above
100
40-100
10-40
Below
10
Total
'h.EA-
D ._
Nm /hr)
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Low-excess-air
LEA
1
1(14)
3
3( 9)
5
5( 9)
2
2(13)
11
11(10)
combustion
CF SI
2
2(29)
1 2
1
2( 6)
3
3( 5)
1
1(7)
6 3
1
7( 6) 3(3)
; CF = Change of
LEA + LEA + CF + FGR + FGR +
LNB FGR SI LNB RHLb RHL LNB
2
1 1
3(43) 1(14)
83 1
11 3
19(58) 3(9) 1(3) 3(9)
22 4 1
17 3
39(70) 4(7) 1(2) 3(5)
5
7
12 (80)
37 3 4 1 1
36 16
73(66) 3(3) 4(4) 1(1) 1(1) 1( 1) 6(5)
fuel; SI = Steam injection.
LEA + CF Subtotal
+ LNB
5
2
7(100)
18
1 15
33(100)
1 36
20
1(2) 56(100)
8
7
15(100)
1 67
44
1(1) 111(100)
-------
Table 2-8 shows NOx levels attainable with the burner for natural
draft oil heating furnaces.
TABLE 2-8. NOX EMISSION LEVELS ATTAINABLE WITH SRG BURNER
FOR OIL HEATING FURNACES*
NOX Emission Levels
Fuel (ppm corrected to 6 percent Oa)
Grade A Oil 70-90
Kerosene 50-60
Gas 50-60
*Heat release rate: 4 x 104 to 8 x 1Q1* kcal/m3 hr
Furnace temperature: Below 900°C
o
The investment cost is about an average of 760 yen ($3.80)/Nm /hr
for all countermeasures, and about 880 yen ($4.40)/Nm /hr for a low-NOx
burner.
2.6.3 Metal Heating Furnaces
The effects of CM techniques on NOX reduction with metal heating
furnaces using different fuels are presented in Table 2-9. As with boilers
and oil heating furnaces, NOX emission levels are low with metal heating
furnaces using low-nitrogen fuels such as naphtha, kerosene and gaseous
fuels, while reduction ratios are high.
CM techniques used in metal heating furnaces are shown in Table 2-10.
Low-N0x burners, especially the staged-combustion type, are adopted widely.
Steam injection (SI) is used for 10 furnaces with a capacity between 10,000-
100,000 Nm3/hr. This technique is regarded as an emergency measure. In
existing small furnaces, emulsified fuels are used because of the ease of
adoption.
83
-------
TABLE 2-9. EFFECTS OF CM TECHNIQUES ON NOX REDUCTION WITH
METAL HEATING FURNACES
Fuel
Before CM
(ppm)
Grade C Oil
Naphtha and Kerosene
2
Gaseous Fuel
Grade C Oil + Naphtha
Grade C Oil + Gaseous Fuel
Corrected to 11
Example of high
percent
02.
228
171
295
138
200
After CM
(ppm) 1
152
62
87
105
138
Reduction
Ratio
(percent)
33
64
71
24
31
NOX level before CM.
TABLE 2-10. CM
TECHNIQUES USED
IN METAL
Capacity
1,000
Over
100
40-100
10-40
Under
10
Total
Nm Vhr
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
Existing
New
Subtotal
LNB
1
0
1(50)
3
0
3(38)
2
1
3(18)
2
1
3(50)
8
2
10(31)
TSC
0
0
0( 0)
0
0
0( 0)
0
0
0( 0)
1
0
1(17)
1
0
K 3)
SI
0
0
0( 0)
2
0
2(25)
7
1
8(47)
0
0
0( 0)
9
1
10(31)
EF*
0
0
0( 0)
0
0
0( 0)
1
0
K 6)
1
0
1(17)
2
0
2( 6)
HEATING
LNB +
TSC
0
0
0( 0)
0
0
0( 0)
2
1
3(18)
0
0
0( 0)
2
1
3( 9)
FURNACES
Others
0
1
1(50)
2
1
3(38)
1
1
2(12)
1
0
1(17)
4
3
7(20)
Subtotal
1
1
2(100)
7
1
8(100)
13
4
17(100)
5
1
6(100)
26
7
33(100)
*EF = emulsified fuel.
84
-------
The types of applicable CM techniques are limited due to the high flame
temperature requirement and the influence of flames on heated materials.
Other techniques, such as low-excess-air combustion, change of fuel,
and reduction of heat release rate by furnace reconstruction are also used.
Table 2-11 gives burner manufacturers' test results for metal heating
furnaces. Due to the demand by steel makers, the present target of develop-
ment is about 50 ppm at 11 percent Oa for low-nitrogen fuels such as grade
A heavy oil, kerosene and gaseous fuels. Reduction of burner size is desired
to ease the application to existing furnaces.
The investment cost for CM ranges from 20 to 340 yen ($0.10 to $1.70)7
Nm3/hr and averages about 210 yen ($1.05)/Nm3/hr. The cost of low-NOx
burners is about 320 yen ($1.60)/Nm3/hr.
TABLE 2-11. BURNER MANUFACTURERS' TEST RESULTS FOR
METAL HEATING FURNACES*
NOX Emission Levels
Fuel (ppm corrected to 11 percent
Grade C Oil 90
Grade A Oil 50
Gas 50
*Furnace temperature: about 1300°C.
85
-------
2.7 REFERENCES*
1. Hatanaka, T., NOX Control Techniques for Metal Heating Furnaces.
Journal of Industrial Public Nuisance Control Association, Vol. 14,
No. 6, 1978. p. 11.
2. Matsuo, M., Low-N0x Burner Developed by Nippon Furnace Kogyo Kaisha,
Ltd. Heat Management and Pollution Control, Vol. 26, No. 12, 1974.
p. 53.
3. Tsuji, S., IHI Low-N0x Burner. Heat Management and Pollution Control,
Vol. 26, No. 3, 1974. p. 21.
4. Masuda, N., Low-N0x Burner Developed by Daido Steel Co. Heat Management
and Pollution Control, Vol. 26, No. 3, 1974. p. 25.
5. Yamagishi, K., et al., Low-N0x Burner Developed by Tokyo Gas Co.
Journal of the Japan Society of Mechanical Engineers, Vol. 77, No.
663, 1974. p. 225.
6. Okada, U., Development of Low-N0x Burner, Chugai Technical Report.
76 WAL 1087, 1976.
7- Masuko, S., et al., Low-N0x Burner for Pulverized Coal Firing. The
Thermal and Nuclear Power, Vol. 29, No. 6, 1978. p. 29.
8. Sakurai, K., NOX Reduction in Gas Firing. Industrial Heating, Vol. 14,
No. 2, 1976. p. 43.
9. Takahashi, Y., et al., Study on a Method of Predicting NO* Emissions
from Boilers and Development of Extremely Low-NOx Burner. MHI Technical
Report, Vol. 13, No. 2, 1976. p. 40.
*A11 are in Japanese.
86
-------
10. Kanamori, S., Low-NOx Burner Developed by Volcano Co. Heat Management
and Pollution Control, Vol. 26, No. 12, 1974. p. 47.
11. Yoshida, A., Operation of Boilers Adopting Low-N0x Burners. Heat
Management and Pollution Control, Vol. 30, No. 6, 1978. p. 31.
12. Tsuji, S., NOX Control Techniques in Utility Boilers. Heat Management
and Pollution Control, Vol. 28, No. 5, 1976. p. 31.
13. Report of Nitrogen Oxides Investigation Committee. Environmental
Agency, January 1977 and April 1978.
14. Report of Research Committee of Nitrogen Oxides Control Techniques.
Japan Ministry of International Trade and Industry, May 1977.
87
-------
SECTION 3
GENERAL DESCRIPTION OF SELECTIVE CATALYTIC REDUCTION (SCR)
3.1 INTRODUCTION
3.1.1 Basic Reactions
Ammonia selectively reacts with NOX in flue gas at elevated
temperatures, as shown in equations 3-1 and 3-2. Actually a small amount
of oxygen is needed for the reaction as shown in Figure 3-1. Therefore,
the reactions are often expressed by equations 3-3 and 3-4. However, the
requirement of oxygen might not be stoichiometric and with some SCR
catalysts the reactions seem to be better expressed by equations 3-5 and
3-6.
4NH3 + 6NO = 5N2 + 6H20 (3-1)
8NH3 + 6N02 = 7N2 + 12H20 (3-2)
4NH3 + 4NO + 02 = 4N2 + 6H20 (3-3)
4NH3 + 2N02 + 02 = 3N2 + 6H20 (3-4)
8NH3 + 10NO + 02 = 9N2 + 12H20 (3-5)
12NH3 + 8N02 + 02 = 10N2 + 18H20 (3-6)
In addition, the following reactions between ammonia and oxygen can
take place at higher temperatures:
88
-------
4NH3 + 302 = 2N2 + 6H20
4NH3 + 502 = 4NO + 6H20
4NH3 + 402 = 2N20 + 6H20
(3-7)
(3-8)
(3-9)
80
760
£40
20
250 300 350 400 450
CATALYST TEMPERATURE, «C
Figure 3-1. Effect of oxygen on NOX removal (SV, 10,000 hr"1;
NOX, 100 ppm; NH3/NOX, 1.0).
Virtually all of the NOX in combustion gas is present as NO. The
optimum temperature for the reaction of NOX and NH3 without a catalyst is
in a narrow range of 950-1000°C. The reaction rate is low below this
temperature, whereas undesirable reactions (equations 3-7, 3-8, 3-9)
readily occur at higher temperatures. The use of a catalyst lowers the
reaction temperature to 150-450°C and prevents the undesirable reactions.
As shown in Figure 3-2, many catalysts have been developed with an optimum
reaction temperature of 300-400°C, so that the boiler economizer effluent
at 300-400°C can be directly treated.
89
-------
~ IUU
_)
>
i 80
LU
ce
ox 60
z
Vy " ^
/ \
. / \
B/ ^
\
\ E
X ^VS^X.
^^ ^^^s^ ^^^fc
"^•x^x^S.
'*•'*¥/<<
i 1 "^TrS
200
\iS N. N
/< NF V
/ \ ^
N
\
\
X
^
>^ss->_ ***" — ^
7j^42^;s2i$^2^4*^^ i i
300 400 500
20 1.
1— Q.
10 3 .
1 w .^4 ^
o Si"
REACTION TEMPERATURE, °C
A: Cr203-Al2
B: Pt-AT203
w • 1" C W ™ f\ 1 f) w
03 D: Fe203-AT203
E: Fe203-Cr203-Al203
F: V0Oc-Cr00,-AUO,
Figure 3-2. Criteria for catalyst for clean gas.1
Base metals have been used as the catalyst. Alumina was widely used in
the early stage of development as a catalyst carrier, but it is poisoned by
SOX in the gas. Therefore, TiOa-based, SOx-resistant catalysts are now
used; they will be described in Section 3.2.
Particulate plugging has been the most serious problem for the
catalyst, which usually is of granular or pellet type. To cope with this
problem, tubular, honeycomb and other specially shaped catalysts and
parallel flow type reactors have been developed. These catalysts and
reactors will be described in Section 3.3.
3.1.2 Major SCR Plants
SCR has been considered the most useful way to remove NOx because of
its simplicity and its removal efficiency of over 90 percent. Following
90
-------
numerous tests with many pilot plants by many process developers, over 40
commercial SCR units have been constructed. Table 3-1. shows commercial
plants and several pilot plants with a capacity of over 10,000 Nm3/hr (3.3 MW
equivalent). Over half the commercial plants use granular catalyst in a
fixed bed to treat clean gas from the combustion of LNG, LPG, kerosene, or
off gas from the petrochemical industry. The rest of the plants of reactors,
to use various types, treat semi-dirty gas from the combustion of low-sulfur
oil or dirty gas from the combustion of high-sulfur oil; these plants also
include those which treat dirty gas from coke ovens and iron ore sintering
machines.
At the early stage of dirty gas treatment, wet-process flue gas desul-
furization was applied in advance for gas cleaning (Shindaikyowa P.C. and
Kasasaki Steel shown in Table 3-1). The reason for this was that SCR cata-
lysts were apt to be affected by SO3 and dust in the gas. However, this
system is expensive because it requires a large amount of energy to heat the
gas cooled by the desulfurization. Since SO -resistant catalysts and reac-
X
tors free from dust plugging were developed later, SCR has been applied in
many plants to hot dirty gas without gas cleaning.
3.1.3 Operating Parameters
For 90 percent NO removal using granular catalyst, a space velocity
X
(SV)* of 5000-10,000 hr"1 and an NH3/M) mole ratio of 0.95-1.2 have been
X
used for dirty gas (see Figure 3-3). For clean gas, however, an SV of
10,000-20,000 hr 1 has been applied because a highly reactive catalyst which
may be poisoned by SO can be used. The SV is usually smaller with parallel
flow type reactors which are described in Section 3.3.2.
*Space velocity equals the ratio of the volume of gas passing through the
reactor in an hour at 0°C and atmospheric pressure to the volume of the
catalyst bed.
91
-------
TABLE 3-1. SCR PLANTS (LARGER THAN 10,000 Nm3/hr)
vo
N>
Process Developer
Hitachi, Ltd.
Mitsubishi H.I.
Ishikawaj ima H.I.
Hitachi Zosen
User
Kawatetsu Chemical
Chiyoda Kenzai
Kansai Paint
Nisshin Steel
Nisshin Steel
Kansai Electric
Chubu Electric
Nippon Oils & Fats
Nisshin Steel
Company A
Company B
Sumitomo Chemical
Osaka Gas
Tokyo Electric
Fuji Oil
Kyushu Electric
Company C
Company D
Chubu Electric
Chubu Electric
Aj inomoto
Company E
Company F
Chugoku Electric
Chugoku Electric
Tohoku Electric
Idemitsu Kosan
Shindaikyowa P.C.
Toshin Steel
Kawasaki Steel
Nippon Satetsu
Plant
Site
Chiba
Kaizuka
Amagasaki
Amagasaki
Amagasaki
Kainan
Chita
Amagasaki
Sakai
—
—
Sodegaura
Takaishi
Yokosuka
Sodegaura
Kokura
—
—
Chita
Taketoyo
Kawasaki
—
—
Kudamatsu
Kudamatsu
Niigata
Chiba
Yokkaichi
Hyogo
Chiba
Chiba
Gas Source
Coke Oven
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
SMm
Furnace
Fuel
COG, BFGa
HO(HS)e
Kerosene
HO(LS)g
HO(LS)
Crude Oil
LNG
HO
Kerosene
HO(LS)
HO(LS)
HO(LS)
LNG
HO(LS)
HO(LS)
LNG
HO(LS)
HO(LS)
HO(LS)
Crude Oil
HO(LS)
HO(LS)
HO(LS)
Crude Oil
HO(LS)
HO(LS)
CO
HO(HS)
Kerosene
Coke, Oil
HO
Capacity
(Nm3 /hr)
500,000
15,000
16,000
20,000
19,000
300,000
2,000,000x2
20,000
30,000
490,000
550,000
300,000
15,000x2
40,000
200,000
1,690,000x2
1,010,000
490,000
1,920,000
20,000
180,000
960,000
480,000
1,000,000
1,900,000
1,660,000
350,000
440,000
71,000
762,000
10,000
Gas Type of
Pretreatment Reactor
ESPb, HC
None
Hh
H
H
None
None
None
None
None
None
ESP, HE1
None
None
None
None
None
None
None
None
None
None
None
None
None
None
H
ESP, DS1, HE, H
None
DS, ESP, HE, H
None
IMBd
1MB
FBf
FB
FB
FB
FB
1MB
FB
FBP3
FBP
1MB
FB
FBPJ
FBHk
FB
FBP
FBP
FBH
FBH
FBH
FBH
FBH
FBH
FBH
FBH
FB
FB
FB
FB
FB
Completion
Nov 1976
Oct 1977
1977
Aug 1977
July 1977
1977
Apr 1978
Apr 1978
Dec 1978
June 1978
June '-978
Sept 1976
Dec 1976
Mar 1977
Jan 1978
Oct 1978
Feb 1978
July 1978
Feb 1980
Apr 1977
Jan 1978
Apr 1978
June 1978
Apr 1979
July 1979
Aug 1981
Oct 1975
Nov 1975
May 1976
Nov 1976
Dec 1976
(Continued)
Coke oven gas, blast furnace gas
bElectrostatic precipitator
^Heating
Intermittent moving bed
eHeavy oil (high sulfur)
Fixed bed
''Heavy oil (low sulfur)
hHeater
exchanger
JFixed bed parallel plate
kpixed bed honeycomb
Desulfurization
""Sintering machine
-------
TABLE 3-1. (Continued)
VO
OJ
Process Developer
Sumitomo Chemical
Mitsui Toatsu
Mitsui
Engineering
Sumitomo Chem.
Engineering
Mitsubishi Kakbki
J. G. Corp.
Asahi Glass
Kobe Steel
Kurabo
User
Sumitomo Chemical
Higashinihon M.
Nihon Ammonia
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Maruichi Kokan
Sumitomo Chemical
Sumitomo Chemical
Mitsui Toatsu
Osaka Pet. Chem.
Mitsui Toatsu
Nishinihon M.
Idemitsu Pet. Chem.
Mitsui Pet. Chem.
Ukishima Pet. Chem.
Toho Gas
Toho Gas
Toho Gas
Toho Gas
Nippon Yakin
Toho Gas
Toho Gas
Toho Gas
Toho Gas
Kashima Oil
Fuji Oil
Nippon Steel
Asahi Glass
Kansai N.K.
Kurabo
fHeavy oil (high sulfur)
Electrostatic precipitator
cFixed bed
Heavy oil (medium
sulfur)
Plant
Site
Sodegaura
Sodegaura
Sodegaura
Niihama
Anegasaki
Anegasaki
Sakai
Sodegaura
Sodegaura
Takaishi
Takaishi
Takaishi
Takaishi
Ichihara
Chiba
Chiba
Sorami
Sorami
Sorami
Sorami
Kawasaki
Sorami
Kawasaki
Chita
Chita
Kashima
Chiba
Kimitsu
Amagasaki
Hirakata
Gas Source
Boiler
Furnace
Furnace
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Boiler
Boiler
Furnace
Boiler
Coke Oven
Furnace
Coke Oven
Boiler
Fuel
HO(HS)a
LPG
LPG
LPG
LPG
LPG
LPG
HO (MS)"!
HO (MS)
Off Gas
Off Gas
Off Gas
Off Gas
Off Gas
CO
HO(LS)
Naphtha
Naphtha
Naphtha
Naphtha
HO(HS)
Naphtha
Kerosene
Kerosene
Kerosene
HO (MS)
CO
COG
COG
HO (US)
eHeavy oil (low sulfur)
Heat exchanger
gHeater
"Fixed bed
Capacity
Gas Type of
(Nm3 /hr) Pretreatment Reactor
30,000
200,000
250,000
200,000
200,000
100,000
10,000
240,000 Wet
300,000
87,000
91,000
170,000
363,000
300,000
200,000
220,000
31,000x2
23,000
23,000
19,000
14,000
19,000
10,000
30,000
43,000
50,000
70,000
150,000
70,000
104,000
30,000
ESPb
None
None
None
None
None
None
ESP, HE , Hg
ESP
None
None
None
None
None
ESP
None
None
None
None
None
H8
None
None
None
None
None
None
None
None
None
None
FBC
FB
FB
FB
FB
FB
FB
FB
IMB1
FB
FB
FB
FB
FB
FB
FBT"
FB
FB
FB
FB
FB
FB
FB
FB
FB
ppj
PP
PP
1MB1
1MB
CMBk
Completion
July 1973
May 1974
Jan 1975
Apr 1975
Apr 1975
Mar 1976
Oct 1976
Feb 1976
Sept 1976
Jan 1977
June 1977
Oct 1977
Oct 1975
Apr 1978
Oct 1977
Dec 1977
June 1978
July 1978
July 1978
Oct 1978
Nov 1978
Oct 1977
Oct 1978
Nov 1975
July 1976
Mar 1977
Aug 1977
Aug 1975
Intermittent moving bed
^ Parallel passage
Continuous
moving bed
tubular catalyst
-------
100 p
90
80
« 70
>
Oi
60
50
40
m
30 =
20
(0
-------
The amount of leak ammonia (NHa in reactor effluent) is minor when the
NH3/NO mole ratio is smaller than 0.8; it is considerable when the ratio is
xs
larger than 1.0. The use of a large SV reduces the amount of catalyst, but
lowers the NO removal efficiency and increases leak ammonia. This is illus-
A
trated in Figure 3-3.
The economizer outlet temperature often fluctuates between 270 and 400°C
with the boiler load. For clean gas treatment, a drop in the gas temperature
does not lower the NOX removal efficiency when an adequate catalyst is used
because the decrease in gas volume can compensate for the lowering of the
reaction rate (Figure 3-4). With dirty gas, the situation is similar above
about 300°C; below this temperature, however, ammonium bisulfate tends to
deposit on the catalyst to lower its activity. For some boilers using
sulfur-containing fuel, an auxiliary burner or a bypass system is used to
take a portion of the hot gas before it reaches the economizer. In this way
the gas temperature is maintained so that no deposit of the bisulfate is
formed. The gas heating system is not necessary if the low-temperature
period lasts for only several hours, because ammonium bisulfate is removed
when the temperature is raised to 350eC.
Low-temperature catalysts reactive at 150-200°C have been developed for
gases such as those from iron-ore sintering machines and coke ovens. Also,
it is not possible to modify the duct system of some existing boilers to
obtain hot gas before the air preheater. Since heating of those gases to
above 300°C is costly, requiring a heat exchanger, heater, and much energy,
low-temperature catalysts are advantageous. A catalyst developed by Kureha
Chemical is reactive at 150°C but is poisoned seriously by SOx and requires
nearly complete desulfurization of the gas in advance. It will be discussed
in Section 4.9. A catalyst developed by JGC works at 200°C and is resistant
to SOy. Although ammonium bisulfate deposits on it below 250°C and reduces
its activity, occasional heating of the contaminated catalyst to above 350°C
removes the bisulfate, This catalyst will be discussed in Section 4.4.
95
-------
(0
>
o
0)
a:
100
90
80
270
4000
8
Inlet NOV
}\
50-130 ppm
20
10
a>
300
330 350 (°C)
6000 9000 12000 (SV"1)
2 (02%)
1/4 2/4 3/4 4/4
Boiler Load
Figure 3-4. Results of clean gas treatment at different
boiler loads (NH3/NOX = l.O).1
96
-------
3.2 COMPOSITION AND BEHAVIOR OF CATALYSTS
3.2.1 Introduction
At the early stage of development of SCR, base metal catalysts based on
an alumina carrier were used predominantly. As a heat-resistant catalyst
carrier, Y~Al2Os has been most popular because of its large surface area and
high activity. The surface area of different carriers is generally as shown
below:
Y - A1203 > Ti02 > Zr02 > MgO > a - A1203 > Si023 (3-10)
A catalyst based on ferric oxide has also been used. However, alumina and
ferric oxide have been found to be poisoned by SOV in flue gas. Pretreat-
X
ment of flue gas by wet process desulfurization to remove most of SO has
X
been carried out in a few commercial plants; this system is very costly,
however, because a large quantity of fuel is needed to reheat the gas to a
temperature suitable to SCR.
Since a few years ago, titania (Ti02)-based catalysts have become popu-
lar for dirty gas treatment because of titania's resistance to S0_. The SO^
X X
resistance of the catalyst carrier is shown below:1
Ti02 = Si02 > a - A1203 > r)-A!203 > Y~A.1203 (3-11)
The catalyst metals also tend to react with S0x, particularly with S03.
Some of the metals lose activity when reacted with S03. Unlike these metals,
vanadium compounds are resistant to S03. Therefore, many of the SOX~
resistant catalysts are based on Ti02 and V20s and are fairly stable against
gases containing up to about 2000 ppm of SO of which about 100 ppm is
present as S03. Flue gas from high-sulfur coal contains larger amounts of
S02 but little S03. Currently, catalyst producers guarantee a life of one
year for catalysts used to treat such dirty gas.
97
-------
The SCR reactor effluent contains a small amount of ammonia. In order
to reduce the ammonia, ammonia reduction catalyst has been developed by
several producers.
3.2.2 Behavior of A1203 Carrier
Most alumina carriers are in the form of y-A!203, which has a large
surface area and high activity. The alumina in y form, however, readily
reacts with S03 to form A12(SOO3 to reduce surface area and activity (see
Figure 3-5). Even though a gas may contain S02 and no SO3, a portion of S02
is oxidized on the catalyst surface to SO3 and reacts with Al20s. By heat-
ing the poisoned carrier to 800°C, A12(SO^)3 is decomposed to regenerate
A1203 and to recover the activity.
A1203 + 3S03 = A12 (SOO3
A1203 + 3S02 + 3/202 = Al2CS0lf)3
A12(S04)3
800°C
A1203 + 3S02 + 3/202
(3-12)
(3-13)
(3-14)
O)
O)
3
t/1
-------
Since this thermal treatment is costly, however, a more stable form of
S0x~resistant carrier is desired. Calcination of Y-A1203 at 1100°C produces
a-A!203, which is fairly resistant to SO but has a smaller surface area and
X
lower catalyst activity. The change of properties of A1203 by calcination
will be described next.2
Three kinds of commercial Al20a carriers in y form (two cylindrical and
one spherical) were calcined at temperatures between 600 and 1200°C. The
changes in physical properties are summarized in Table 3-2. A1203 was
sintered above 750°C resulting in a decrease in pore volume and surface area.
The Y type was converted to a type at 1100°C or above. The surface area of
the alumina calcined at 1100°C is only one-fifth that calcined at 600°C.
Calcination at 1150°C or above further reduced the surface area and catalyst
activity remarkably. Figure 3-6 shows the weight increase in A1203 due to
reaction with S02; A1203 calcined at different temperatures was heated at a
rate of 2°C/min. in a gas containing either 0.1 percent S02 with 5 percent
02 or 2 percent S02 with 10 percent 02. The weight increase was larger in
A1203 calcined at lower temperatures. The addition of CuS04 remarkably
promoted the increase. It seems that a portion of S02 is oxidized to SOs
on the M203 surface and the SOs reacts with A1203 to form A12(SOM.)3; CuSOif
promotes this oxidation.
The a type Al20s prepared by calcination at 1100°C showed only a small
weight increase. It may not be feasible, however, to prepare an a type
Al20s carrier of good quality in a large amount, because precise control of
the calcination temperature at 1100°C is needed. Below 1100°C the conver-
sion from Y to a is not sufficient and above 1100°C the pore volume and
surface area decrease remarkably.
3.2.3 Ti02-Based Catalysts2
Ti02 is not affected by the SOx in flue gas. TiOz, anathase, not only
is a good carrier, but also shows catalytic activity by itself. Figure 3-7
99
-------
TABLE 3-2. CHANGES IN PHYSICAL PROPERTIES OF A1203 CARRIERS BY CALCINATION'
o
o
Calcination
Sample Temp. (°C)
f 600
A
(Cylindrical) <
750
900
1050
1100
1150
1200
V
{600
750
900
1100
c <
(Spherical)
600
750
900
1050
1100
. 1150
Average
Diameter
3.20
3.18
3.00
2.85
2.78
2.69
4.83
4.81
4.78
4.38
3.67
3.67
3.60
3.52
3.36
3.19
Size (mm)
Height
3.20
3.16
3.00
2.83
2.76
2.65
4.81
4.82
4.79
4.30
Apparent
Density
(g/cm3)
1.30
1.32
1.58
1.84
1.95
2.12
1.37
1.43
1.43
1.79
1.52
1.45
1.54
1.65
1.93
2.21
Pore
Volume
(cm3/g)
0.50
0.49
0.37
0.32
0.26
0.21
0.46
0.42
0.44
0.30
0.38
0.42
0.37
0.34
0.25
0.19
Surface
Area
(m2/g)
150
157
131
86
32
13
7
194
188
123
29
187
183
144
92
25
7
Average
Pore Size
(angstroms)
130
150
170
400
800
1200
95
90
140
410
81
92
100
150
400
1100
-------
trt
o
c
O)
300
400 , 500 600
Heating Temperature (°C)
Marks
e
o
•
A
A
X
Calcination
temp.(°C)
900
900
600
900
1050
1100
Additive
CuSOi*
CuSOit
None
None
None
None
composition
S02(%)
2
0.1
2
2
2
2
02(%)
10
5
10
10
10
10
Figure 3-6. Weight increase in AlaOa (pre-calcined at 600-1100°C)
heated in gas containing SOz at rate of 2°C/min.2
101
-------
100
80
" 60 h
ru
1
"x 40
i
20
250 300 350 400 450
Reaction Temperature (°C)
500
Mark
0
*
Q
*
T
A
Composition (%)
FeiOs T10z Additive
50
40
40
0
0
0
50 None
50 10 MoO,
50 10 WO,
1 00 None
100 None
100 None
Method of preparation
NHS was added to Fe1(SOit)J-Ti(SO^)i soln.
NH9 was added to Fe2(SCU}3-T1(SO,,)2-
(NHil)6Mo;021t solution
NH3 was added to Fe2(SOO>-T1(S002
-(N^)ioWuOm solution
Hydrolysis of T1(SOt,)2 solution
NH| was added to T1(SO*h solution
NajCOs was added to T1(SOJZ solution
Figure 3-7. Comparison of NOX removal from inlet gas* with several
Ti02 based catalysts (calcined at 550°C for 5 hours,
1 mm size).2
* The gas contained 500 ppm NO, 667 ppm NHs, 1000 ppm SOz, and 10 percent
02; the gas flow was 1000 Ncm3/hr and the SV, 110,000 hr"1.
102
-------
shows the results of laboratory tests on TiOa and TiOa-based catalysts in
small granules. TiOa (anathase) prepared by hydrolysis of TiCSOOa an<*
calcination at 550°C gave 90 percent NOx removal at 500°C. Catalyst pre-
pared by sodium carbonate precipitation gave poor results, possibly because
of an adverse effect produced by a small amount of sodium contained in the
product TiOz. Catalyst prepared by NHS precipitation gave better results
than did that from sodium precipitation.
catalyst yielded over 90 percent NOx removal at 450 and
500°C. Addition of Mo03 or WOs improved the efficiency at lower tempera-
tures. The SV used for the tests is much larger than that used commer-
cially. Use of an SV of about 10,000 with the TiOa-based catalysts will
result in over 90 percent removal above 350°C.
3.2.4 Vanadium Catalysts
Figure 3-8 shows the results of laboratory tests on vanadium catalysts
with TiOa and AlaOs (Y type). The TiOa-based catalysts gave a high NOx
removal efficiency between 350 and 400°C and were superior to the AlaOa-
based catalysts. Catalysts with 10 percent VaOs on a TiOa carrier showed
the highest efficiency.
Figure 3-9 shows the results of tests to determine catalyst life made
by the Government Chemical Research Institute in Tokyo, with VaOs and FeaOj
catalysts on a Ti02 carrier and flue gas containing 1000 ppm of SOa. The
NOx removal efficiency, which was nearly 90 percent with both catalysts at
the beginning, did not decrease with the VaOs catalyst, but it decreased to
about 78 percent with the Fe20s catalyst in 40 days.
One of the problems with the SCR catalysts is the tendency to oxidize
a portion of SOa in flue gas to SOa which would cause the ammonium bisulfate
problem and also might affect the environment. The poisoning of SCR
catalysts by SOa may also be due to the formation of SO 3 on the catalyst
103
-------
>
-------
lOOr
90
>
QL
80
70
oo °oo
oo o
10 20 30 40
Reaction Time, (day)
SV: 10000 hr"1
NO: 200 ppm
NH3: 200 ppm
S02: 1000 ppm
02:
H20:
Temp: 350°C
o Fe203 - Ti02
• V205 - Ti02
Figure 3-9. Results of life tests with Ti02-based catalysts/
105
-------
It is possible that the vanadium catalyst has a relatively large
capability for oxidation; vanadium has been used as oxidation catalyst to
convert SOa to SOs in sulfuric acid plants. The presence of ammonia, a
reducing agent, may depress the oxidation considerably. Further studies
are needed on this oxidation and its prevention.
3.2.5 Ammonia Decomposition Catalysts
Ammonia decomposition catalysts have been produced by several com-
panies. Results of tests on the catalysts with clean gas are shown in
Figure 3-10. The SCR rea,ctor had an NOx removal efficiency of about 90
percent and the reactor effluent contained about 10 ppm of NHs. By adding
an ammonia converter, the NHa concentration was reduced to below 3 ppm and
the NOX removal efficiency increased to over 97 percent; the decomposition
catalyst not only decomposes ammonia, but also promotes the reaction of NOx
and NHs in the gas.
j IOC
-------
On the other hand, the decomposition catalyst and the converter have
the following disadvantages: 1) Not only the investment cost, but also the
pressure drop of the gas (and consequently the power requirement) is
increased. 2) Most of the ammonia decomposition catalysts are subject to
poisoning by SOX. 3) The catalysts tend to convert a portion of SOz to
SOa. Most of them are oxidizing catalysts and the absence of ammonia is
favorable to the oxidation.
The amount of ammonia decomposition catalyst is about a half that of
the SCR catalyst. Therefore, the use of the decomposition catalyst is not
popular. Process users prefer to reduce leak NHs to below about 5 ppm by
reducing the NHs/NOx mole ratio to 0.9 to obtain an NOX removal efficiency
of 80-85 percent. When a higher NOx removal efficiency is needed, a
larger amount of SCR catalyst and a mole ratio of about 0.95 may be used.
3.2.6 Problems and Further Studies Required
SCR catalysts have been considerably improved. Currently most cata-
lysts have a life of over 2 years when used on clean gas, 1-2 years on semi-
dirty gas, and 1 year on dirty gas. The catalysts, however, still present
the following problems which remain to be studied and solved.
1) Many dirty gases at a high temperature contain appreciable
amounts of compounds in the vapor phase, such as vanadium in
oil-fired boiler flue gas and alkaline compounds in the gases
from a glass melting furnace, cement kiln, and iron-ore
sintering machine. In the liquid or solid phase, those com-
pounds tend to deposit even on catalysts of the parallel flow
type, on which solid particles in the gas do not deposit.
(This will be discussed further in Section 3.3.) The deposits
affect the catalysts in various ways. Vanadium rarely lowers
the catalyst activity, but can change the optimum reaction
107
-------
temperature. Alkaline compounds lower the activity remarkably.
Although most of the alkaline compounds can be removed by
water-wash, the catalyst is usually degraded to some extent.
2) Fine particulates smaller than about 1 micron tend to enter
the small pores of the catalyst carrier, reducing the activity
slowly. This "blinding" can not be avoided even in the parallel
flow type reactors.
3) Most of the SCR catalysts oxidize below about 2 percent of SO2
to SOs. The oxidation ratio may be higher when the gas contains
little ammonia.
4) The ammonia decomposition catalyst tends to be poisoned by SOX.
Moreover, it might convert S02 to SOa to a considerable degree.
5) Most of the catalysts use heavy metals. Caution is needed in
disposing of the spent catalyst.
3.3 CATALYST SHAPE AND REACTOR
3.3.1 Introduction
Particulate plugging has been the largest problem for catalysts. For
clean gas containing less than about 30 mg/Nm3 of particulates, a simple
type of reactor, with granular or ring tube catalyst in a fixed bed, can be
used without appreciable problems. Particulate-rich gases may be treated
by an electrostatic precipitator (ESP) to reduce the particulate content
to 30 mg/Nm3. However, since about 400°C is suitable for SCR, a hot ESP is
needed which is considerably more expensive than a cold ESP. The hot ESP
is useful for low-sulfur coal whose fly ash is not efficiently caught by a
cold ESP. A hot ESP, however, is not always highly efficient for
108
-------
particulates from high-sulfur fuels. Therefore, it is sometimes desirable
to apply SCR to a praticulate-rich gas.
For this purpose, moving bed and parallel flow type reactors have been
developed. (See Figure 3-11 and Table 3-3.) The moving bed reactor uses a
granular catalyst, which is charged from the top of the reactor and is moved
down intermittently or continuously while the gas is passed through the
catalyst layer in a cross flow; this is shown in Figure 3-12. The granular
catalyst discharged from the bottom of the reactor is screened to remove
particulates and is returned to the reactor. If needed, the catalyst is
heated to 700-800°C to eliminate sulfur compounds and carbon before it is
returned to the reactor.
The moving bed can treat gases containing less than about 300 mg/Nm3
of particulates, such as flue gas from an oil-fired boiler. For gases
containing 30-100 mg/Nm3 of particulates, an intermittent moving bed may be
suitable to control pressure drop (Figure 3-13), while a continuous moving
bed may be preferable for gases with 100-300 mg/Nm3 of particulates. Flue
gas from a coal-fired boiler containing about 20 g/Nm3 of particulates must
first be treated with an electrostatic precipitator or other means to reduce
the particulate content to below 300 mg/Nm3.
On the other hand, the parallel flow type reactor uses a fixed bed of
different types of catalyst (honeycomb, plate, and tube) or a parallel
passage device (Figure 3-14). The gas passes through a clearance between
the parallel layers of the catalyst, and the reactor is expected to be able
to handle gas from a coal-fired boiler even without previous dust removal.
In Japan, several intermittent moving bed reactors have been in
commercial use for flue gas from oil-fired boilers and coke ovens. Two
relatively small commercial plants (17-23 MW equivalent) with parallel
passage reactors for oil-fired industrial boilers and two medium-size
109
-------
Sort of
flue gas
Dirty gas
Clean gas
(LNG, LPG,
and naphtha
firing)
Rsactor
type
(A)
Fixed bed
type
(Parallel)
Flow
(B)
moving bed
type
(C)
Fixed bed
type
Catalyst
shape
Honeycomb-
type
catalyst
Pellet
type
catalyst
Pellet
type
catalyst
Principal
structure
Catalyst
L^ :>ii-
, , ;J;
'&£*
Flue g
^'
' •: • •*;••• .
fc, S | * '
' * ^ ,tf
^^^****^
as
& u Catalyst
'-s -r^ .•
1:S
'"^l A
gas ' (L.JJ
Catalyst
ffc»
• ' * *
-~^5
Flue ^
^ra-^
:-, !••• """
^4 ^i
gas
Figure 3-11. Types of reactors.
110
-------
TABLE 3-3. REACTORS FOR DUSTY GAS TREATMENT
Process
Type Developer
Moving Bed Hitachi Ltd.
Hitachi Ltd.
Hitachi Ltd.
Mitsubishi H.I.
Kobe Steel
Kiirabo
Asahi Glass
Honeycomb Mitsubishi H.I.
Mitsubishi H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Ishikawajima H.I.
Parallel Passage JGC
JGC
JGC
Parallel Plate Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Hitachi, Ltd.
Hitachi, Ltd.
Tube Mitsui Engineering
User
Kawatetsu Chemical
Chiyoda Kenzai
Nippon Oils & Fats
Sumitomo Chemical
Katisai H.K.
Kurabo
Asahi Glass
Fuji Oil
Chubu Electric
Chubu Electric
Ajinomoto
Company E
Company F
Chugoku Electric
Chugoku Electric
Tohofcu Electric
Kashlma Oil
Fuji Oil
Nippon Steel
Tokyo Electric
Company C
Company D
Company A
Company B
Dkishima Pet. Chen.
Plant
Site
Chiba
Kaizuka
Araagasaki
Sodegaura
Amagasaki
Hirakata
Sodegaura
Chita
Taketoyo
Kawasaki
Kudamatsu
Kudamatsu
Hiigata
Kashlma
Chiba
Kimitsu
Tokosufca
—
—
—
Chiba
Gas Source
Coke Oven
Boiler
Boiler
Boiler
Coke Oven
Boiler
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Coke Oven
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Fuel
COG/KFGa
HO(BS)b
HD
HD(tS)e
COG
HO(HS>
H»(tS)
BOOS)
Crude Oil
H9(LS>
BOCLS)
EO(LS>
Crude Oil
HO(LS)
HO(LS)
CO
COG
BO(LS)
M>
HOCtS)
HO(LS)
HO(LS)
BOCtS)
Capacity
Ota'/hr)
500,000
15,000
20,000
382,000
104,000
30,000
70,000
200,000
1,920,000
20,000
180,000
960,000d
480,000d
1,000,000
1,900,000
1,66O,OOO
5O,OOO
70,000
150,000
40,000
1,010,000
490,000
490,000
550,000d
22O.OOO
Completion
Nov 1976
Oct 1977
Apr 1978
Sept 1976
Aug 1977
Aug 1975
Jan 1978
Feb 1980
Apr 1977
Jan 1978
Apr 1978
June 1978
Apr 1979
July 1979
Aug 1981
Nov 1975
July 1976
Mar 1977
Mar 1977
Feb 1978
July 1978
June 1978
June 1978
Apr 1978
fcoke oven gas/refinery off gas.
bHeavy oil (high sulfur).
"JHeavy oil (low sulfur).
Combination with selective noncatalytic reduction.
-------
CATALYST
GAS INS-
MOVING BED
REACTOR
CONVEYOR
VIBRATING
SCREEN
CATALYST
PARTICULATES
Figure 3-12. Moving bed reactor.
112
-------
Moved
;
Fixed II Fixed
Moved
I I Fixed"
Moved
160
120
80
40
'13.4
(mmH20/day)$
10 20
Test Period (day)
30
Figure 3-13. Change of pressure drop with an intermittent moving bed.
113
-------
Honeycomb
(Ceramic)
(Grid Type)
PUDC
OOQC
Honeycomb
(Ceramic)
(Hexagonal)
Honeycomb
(Metal)
(Wave Type)
Plate (Ceramic)
Plate (Metal)
Tube (Ceramic)
!P1P§§%S^«wS
Parallel Passage
Figure 3-14. Cross sections of parallel flow catalysts (actual sizes)
114
-------
commercial plants (a 63 MW one using a tubular catalyst and a 73 MW one
with a honeycomb catalyst) were put into operation. Many larger plants
with honeycomb and plate catalysts started operation recently. (See Table
3-3.)
3.3.2 Moving Bed and Parallel Flow Reactors
Moving bed reactors have the following characteristics, advantages
and disadvantages as compared with parallel flow type reactors:
1) Catalysts can easily be replaced during the operation of the
denitrification plant.
2) A considerable portion of particulates in the gas is removed
by the moving bed. For example, 100 mg/Nm3 at the inlet can
be reduced to 30 mg/Nm3. Most of the collected particulate is
separated from the catalyst by screening.
3) The moving bed catalyst is granular and is considerably cheaper
than that for most of the parallel flow type reactors. The
catalyst can be packed easily in the reactor.
4) The moving bed catalyst is packed fairly densely and has a large
surface area per packed volume (Table 3-4).
5) A small gas velocity (0.5-1.5 m/sec) is used to maintain a
small pressure drop (Table 3-5). Therefore, for treating a
large amount of gas, a special reactor design is needed to
obtain a large gas flow area and to keep a uniform gas
flow (Figure 3-15).
115
-------
TABLE 3-4. SPECIFIC SURFACE AREA OF CATALYSTS
Dimension (mm)
Granule (Sphere) Diameter
4
XXX 6
OpO 8
^-^ 10
Tube* a b
10 7
a i H b 2° u
^w_<^ >f 25 17'5
1 ^cc-jc^y
JJ 23.1
Honeycomb a b
^k. 9 7
<:viF>f% 10 7
J^*-^\b 14 10
^ ' 20 14
a b
II l| 1! 7 5
20 14
Parallel Plate Thickness Clearance
1 5
1 8
3 8
10 12
Specific Surface
Surface Area
Catalyst Volume
1500
1000
750
600
1350
900
677
543
410
1866
1204
908
602
1819
1204
909
601
2000
2000
250
200
Area (m2/m3)
Surface Area
Packed Volume
1100
740
555
444
405
270
203
163
124
437
387
278
193
556
388
277
194
320
220
125
91
* Optimum separate packing (Section 4.5).
116
-------
TABLE 3-5. COMPARISON OF CATALYSTS AND REACTORS*
Catalyst Size (mm)
Diameter
Thickness
Opening
Q
Gas Velocity (m/sec)
Bed Depth (m)
SV (1000 hr-1)d
Pressure Drop (mmHaO)
Moving Honeycomb
Bed (Metallic)
4-8b
0.5-1
4-8
0.5-1.5 2-6
0.2-0.6 1-2
5-10 5-8
40-80 30-80
Honeycomb ,
Tube
(Ceramic)
1.5-3
6-20
5-10
1.5-5
4-8
40-140
Parallel
Passage
7-106
7-10
5-10
3-5
3-5
80-140
Parallel
(Ceramic)
8-10
8-14
5-10
4-6
1.5-3
80-160
Plate
(Metallic)
1
5-10
4-8
2-5
2-5
60-120
NOX removal at an NH3/NOX ratio of 1.1 and a temperature of 350-400°C for dirty gas.
Ring type catalysts with larger diameters are also used.
^Velocity at 350-400°C in open column (superficial velocity).
Gas volume (Nm3/hr)/catalyst bed volume (m ).
thickness of envelope.
-------
Flue gas
Inlet
(From econ-
omizer)
Flue Gas
Outlet
(To A1r
Preheated)
Catalyst
Layer
Conveyor
Figure 3-15. Moving bed type reactor for a large amount of gas.
118
-------
6) It is not easy to treat gases containing more than 200 mg/Nm3
of partlculates. Partlculate removal is needed in advance
for dusty gas.
7) Attrition of catalyst and mechanical trouble with the catalyst
transportation system can occur. Attrition is minimized by
the use of a hard spherical catalyst.
8) Very fine particulates may enter the small pores of the
catalyst and gradually reduce its activity. , Although the pore
blinding would occur also in the parallel flow type catalysts,
the problem might be more serious with the moving bed.
On the other hand, the parallel flow type reactor has the following
common characteristics, advantages, and disadvantages:
1) Gases with a large amount of particulate, such as coal-fired
boiler flue gas containing 20 g/Nm3 of fly ash, can be treated,
using parallel flow.
2) A catalyst with large openings (gas passing channels over about
7 mm) is safe for dust plugging. With such catalyst, a large
gas velocity, normally 5-10 m/sec, is used in order to give a
slight plug flow and to raise the NOX removal efficiency
(Table 3-5).
3) A large gas passing area (necessary for the moving bed) is not
required, while the depth of the bed is much deeper.
4) The large gas velocity and the deep bed tend to give a large gas
pressure drop (80-140 rnmHaO) for some parallel flow reactors in
order to attain a high NOX removal. By using a honeycomb with
thin walls and small openings (4-5 mm) and a moderate gas velocity
119
-------
(2-3 m/sec), a high NOx removal efficiency is obtained with a
small pressure drop (30-50 mmHaO) but the possibility of dust
plugging is greater than with a honeycomb with larger openings.
5) It is difficult to replace the catalyst without stopping plant
operation.
6) It is necessary to provide regular spacing of the catalyst
during replacement. Usually, tube, plate or honeycomb catalyst
is regularly spaced in containers about Imxlmxlmin
size, and the containers are placed regularly in the reactor
(Figure 3-16). Packing of the catalyst and setting of the
containers requires considerable time and labor.
7) Catalyst erosion by particulates can occur at a large gas velocity,
particularly when good, regular packing is not achieved.
8) In a large reactor it is not very easy to distribute gas
uniformly to each catalyst container; this differs from
granular catalyst in a moving and fixed bed, where the bed
itself helps to distribute gas uniformly.
9) The parallel flow type catalyst may be placed in a duct
between the boiler economizer and air preheater without
installing a separate reactor.
3.3.3 Comparison of Parallel Flow Type Catalysts and Reactors
Parallel flow type catalysts and reactors with different characteris-
tics have been developed (Tables 3-4 and 3-5). For commercial use, the
most suitable one should be selected on the basis of such characteristics
as removal efficiency required, composition of the gas, properties and
amount of particulates.
120
-------
Catalyst layer
Unit module
ODD.
ODD
Figure 3-16. Example of a fixed bed reactor with'honeycomb type catalyst
(Ishikawajima-Marima Heavy Industries; sizes are in mm).
121
-------
Of the parallel flow type catalysts and reactors, the parallel plate
may be the simplest and the least vulnerable to dust plugging. The plates
are made of either ceramic or metallic material; they are 50-100 cm square
with thicknesses of about 10 mm with ceramic and 1 mm with metallic
material. The plates are positioned in a basket in parallel with 5-10 mm
clearance (Figure 3-14), and the baskets are placed in the reactor. The
ceramic plate may not be highly efficient because it must be made fairly
thick to maintain enough strength; generally only a thin surface of the
catalyst (about 0.5 mm) is actually used for the reaction.
The metallic plate with about 1 mm thickness may be fairly efficient
when used with a small clearance (see Tables 3-4 and 3-5). It is made in
two ways:1* chemical treatment of the surface of a special alloy plate,
and cementing powdered catalyst on a metal plate surface. With the latter
method, the desired property of catalyst is easily obtained, but it is
necessary that the catalyst powder should not come off the metal.
Hitachi Ltd. (Section 4.3) and Mitsubishi Heavy Industries (Section
4.6) have conducted tests with plate catalysts that they have produced.
The plate catalysts may be best suited for plants for which a moderate NOx
removal efficiency of 50-70 percent is sufficient.
The parallel passage reactor was originally developed by Shell and
further developed for SCR by JGC. The reactor uses a small size catalyst
packed in thin envelopes made of metal gauze; these envelopes are placed
in parallel as described in Section 4.4.
This reactor is more efficient than the ceramic parallel plate reactor
because it has a larger catalyst surface area as shown in Figure 3-14; it
may cost more, however, because it requires special envelopes (Table 3-5).
One of the possible problems is the deposit of a small amount of dust on
the gauze. JGC has developed a sand-blast system to remove the deposit.
122
-------
Tubular catalysts have been tested by several companies, including
Kawasaki Heavy Industries and Mitsui Engineering. The latter commercialized
the catalyst for the Chiba plant of Ukishima Petrochemical (Section 4.5).
Many tubes are regularly placed so that gas flows in parallel both inside
and outside >of the tubes (see Figure 4-42). Mitsui Engineering has found
an optimum way to pack the tubes separately in which maximum efficiency is
obtained at a relatively low pressure drop. At the Ukishima Petrochemical
plant, tubes with 33 mm outer and 20 mm inner diameters and 1000 mm length
are used. Tubes with thinner walls (for example, 20 mm outer and 14 mm
inner diameters) have been produced which have better efficiency. Tubes
with even smaller diameters and thinner walls should be more efficient,
but may be too weak for commercial use.
The honeycomb type catalyst may be stronger than the tubular for a
given wall thickness; the production of a large honeycomb catalyst is not
easy, however. The largest ceramic honeycomb currently available commer-
cially is 150 x 150 x 500 mm. A thin-wall honeycomb catalyst with small
openings can be highly efficient and has been used at a commercial SCR plant
constructed by Mitsubishi Heavy Industries for Fuji Oil (Section 4.6).
Recently, Ishikawajima-Harima Heavy Industries started operations with
honeycomb catalysts at large plants for utility boilers (see Section 4.7).
A larger size honeycomb can be made with a metallic material by
activating the surface chemically. The honeycomb is easy to handle and has
been used by Hitachi Zosen for pilot plant tests (Section 4.2). Since the
reactive activated surface is very thin, erosion by particulates should be
prevented. Among different types of honeycombs as shown in Figure 3-14,
hexagonal type may be more efficient than others which tend to collect a
small amount of dust at the corners of the channels. It may also have a
little less strength, however. For treating flue gas from a coal-fired
boiler with a full dust load, a parallel flow catalyst with fairly large
openings is needed to prevent dust plugging. Therefore, a honeycomb
catalyst shape is not always the best choice.
123
-------
3.4 COMMERCIAL HONEYCOMB CATALYSTS
3.4.1 Introduction
Honeycomb type SCR catalysts have been produced and tested by many
companies. Most of the catalysts use a ceramic material such as TiOa or
related compounds, while the rest are made of metals. As mentioned pre-
viously, the largest size ceramic honeycomb catalyst currently available
is 150 x 150 x 500 mm with a wall thickness of about 2 mm. The honeycombs
are usually packed in a container measuring about 1 x 1 x 1 m. Smaller
honeycombs are easier to produce but are more difficult to pack regularly.
A few companies, including Hitachi Zosen and Mitsubishi Heavy Indus-
tries, have been working on metal-based honeycomb catalysts made by acti-
vating the surface of a special alloy. By using metal, much larger
catalysts (0.5 x 0.5 x 0.5 m with a thin wall about 1 mm thick) can be
produced. The metal-based catalysts are much easier to handle than the
smaller ceramic ones. On the other hand, the activated surface of the metal
is very thin (less than 0.2 mm usually), and it is possible that the
catalyst loses activity due to erosion by particulates. This is particu-
larly true when the catalyst is used for the treatment of flue gas from
a coal-fired boiler.
Further improvements are expected on the honeycomb catalysts to make
them very useful for the treatment of gases with high particulate loading.
3.4.2 Honeycomb Catalysts Produced by Catalyst and Chemicals Ind. Co.
Catalyst and Chemicals Ind. Co. is one of the largest catalyst pro-
ducers in Japan and has produced several kinds of honeycomb catalysts for
SCR. Results of tests with one of the catalysts are shown below.
124
-------
The honeycomb has a cross section of 96.6 x 96.6 ram with a length of
300 mm (see Figure 3-17). The channels are square with an opening of 6.6
mm and wall thickness of 2 mm. For the tests, 4 honeycombs in a block were
placed in 6 separate layers, so that the total cross section was 193.2 x
193.2 mm and the total depth of the layers was 1800 mm. The flue gas was
obtained by burning heavy oil and had the following composition:
NO: 160-180 ppm S02: 250-300 ppm
COa: 7-10 percent Oz: 7-10 percent
H20: 7-9 percent Particulates: 70 mg/Nm3
The gas volume used for the test was 700 Nm3/hr, and the NHs/NO mole
ratio was 1.1. The results are shown in Figures 3-18, 3-19 and 3-20. At
375°C, 90 percent NOX removal was achieved at an SV of about 10,000 hr"1.
The efficiency remained unchanged at gas velocities between 5.8 and 9.5
m/sec, but was a little higher at 12.2 m/sec. Theoretically, a larger gas
velocity should give more gas turbulence and larger NOx removal efficiency.
At 340°C the efficiency was slightly lower and showed no difference at gas
velocities between 5.3 and 12.3 m/sec. About 90 percent NOx removal was
obtained at an SV of 8000 hr"1. At 300°C, the efficiency was considerably
lower. The removal efficiency was nearly 90 percent at an SV or 5500 hr"1.
The catalyst may be useful in obtaining 90 percent NOX removal for a
boiler which gives gas from the economizer at 375-400°C at full load (SV =
10,000 hr"1) and at 300°C or above at less than half load (SV smaller than
5000). Honeycombs (150 x 150 x 500 mm) made by Catalyst and Chemicals Ind.
Co. have been used by Ishikawajima-Harima Heavy Industries for large-scale
tests with utility boilers.
3.4.3 Honeycomb Catalysts Produced by NGK Insulators Ltd.
("
NGK Insulators Ltd., the largest ceramic insulator producer in Japan,
has produced honeycomb catalysts for various uses. Figure 3-21 shows the
125
-------
Flue Gas
Honeycomb
Cross Section of Honeycomb
(Actual Size)
1
Structure of Test Reactor
Figure 3-17,
Honeycomb and reactor produced by
Catalyst and Chemicals Ind. Co.
126
-------
100
SV (hr'1)
70,000 30,000 20,000 15,000 12,500 10,000 9,000 8,000 7,000
6,000
N>
as
>
O
01
ee
x
o
50
0.5
1.0
10VSV (hr)
Superficial Gas
Velocity (m/sec)
6 12.2
A 9.5
D 7.6
0 5.8
1.5
5,000
-O
2.0
Figure 3-18. NOx removal at 375°C (NO = 160-173 ppm, NH3/NO = 1).
-------
NJ
CO
SV (hr"1)
70,000 30,000 20,000 15,000 12,500 10,000 9,000 8,000 7,000
6,000
1.0
10VSV (hr)
5,000
LV .(Superficial Velocity)
O 12.3 m/sec.
Figure 3-19. NOX removal at 340°C (NO = 176-184 ppm, NH3/NO = 1)
-------
70,000 30,000 20,00015,00012,500
SV (hr'1)
10,000 9,000 8,000
7,000
100
(O
o>
50
NJ
VO
6,000
1—
5,000
0.5
Superficial
Velocity (m/sec)
O 12.4
A 10.2
O 8.5
0 7.5
X 5.3
1.0
10VSV (hr)
1.5
2.0
Figure 3-20. NOx removal at 300°C (NO = 163-178 ppm, NH3/NO = 1)
-------
2.
>.
LU
U
Lu
Lu
LU
1—
U
oc
X
;£
100 -i
90-
80-
70-
60-
50-
^^^^^^^^^L^°**^^>OOrvJ'uiSv^j%rtrfWV*^^
^P
**S^^
9-, *• A ^^ ™"3
1 1 1 1
o
OQ
t—
<*
b
oc.
.20 S
£
-log
<
2
-0 z
U LU
1000 2000 3000
TEST PERIOD (HR)
4000
-200 I
"is°r
-100 |
-50 "
a:
a.
Test condition-
catalyst
hydraullc-dia
of channels
dimensions of
catalyst
space velocity
gas velocity
gas temperature
NH,/NO (mol/mol)
•j X
N6K honeycomb-
shaped catalyst
7mm
55mm d1a. x
2000mn length
10000 hr"1
12.6 m/s
350-360 °C
1.0
Gas
NO
soz
S03
k °2
co2
H20
dust
N2
composition
200-250ppm
900-lOOOppm
30ppm
4-5vo1 .%
10-llvol.*
8-1 Ovol. X
40-90mg/Nm3
balance vol .%
Figure 3-21.
Durability of NOX removal catalyst for exhaust gas
of high-sulfur oil burning boiler.
130
-------
results of a life test with an SCR honeycomb catalyst at 350-360°C with
flue gas from an oil-fired boiler containing 900-1000 ppm of SOa, 30 ppm of
SOs, and 40-90 mg/Nm3 of particulates. An NHs/NOx mole ratio of 1.0 was
used at an SV of 10,000 hr-1 with a gas velocity of 12.6 m/sec. The honey-
comb has a diameter of 55 mm and a total length of 2000 mm.
During the first 2000 hours of the test, NOx removal efficiency slowly
dropped from 98 to 94 percent and leak ammonia (NHs at reactor outlet)
increased from 5 to 14 ppm. Neither the efficiency nor leak ammonia
changed between 2000 and 4000 hours. The pressure drop was kept constant
at 160 mm HzO during the test, indicating no particulate plugging.
A honeycomb catalyst produced by NGK has been used at Fuji Oil Co.
for treating gas from an industrial boiler (Section 4.6).
3.4.4 Honeycombs Made by Other Producers
Sakai Chemical Industries has produced honeycomb catalysts for SCR, the
largest being 150 x 150 x 500 mm with opening sizes ranging from 5 to 10 mm.
The catalysts have been tested by Ishikawaj ima-Harima Heavy Industries in
a large-scale test with a utility boiler. No data have been disclosed yet.
Hitachi Zosen has also produced metallic honeycombs; they will be
described in Section 4.2.
There are several other companies working on honeycomb catalysts for
SCR.. Kobe Steel has succeeded in producing a fairly large ceramic honeycomb
with a very thin wall (0.5 mm).
131
-------
3.5 PROBLEMS OF AMMONIUM BISULFATE FORMATION
3.5.1 Introduction
Among the many processes for NOX removal, selective catalytic reduction
and noncatalytic reduction have been considered the most promising; both
use ammonia as the reducing agent. Considerable progress has been made on
the technology for these processes, and many of the problems have been
almost solved. The largest problem left unsolved may be the formation of
ammonium bisulfate in heat exchangers of air preheaters.
S03 , NHs , and HzO present in a hot flue gas combine to form ammonium
bisulfate, NHitHSO^, on cooling.
NH3( . + S03( v + H2°(g)-< - > NHijHSOifv (3-15)
The bisulfate would form at about 210°C or below from a typical gas leaving
the reactor containing about 10 ppm of both SOs and NHs . A simple calcula-
tion shows that 10 ppm each of SOs and NHs in gas from a 500 MW boiler
accounts for 2 tons/day of the bisulfate. The amount is sufficient to cause
serious problems, such as plugging and corrosion in an air preheater or
heat exchanger. Serious problems have been encountered in a few SCR facil-
ities in Japan. The problem has not been serious in facilities which
either discharge hot gas at 220°C or above without recovering much heat or
burn a very low-sulfur (0.2 percent) oil; the amount of bisulfate deposit is
small. For large boilers using medium- or high-sulfur fuel and requiring
enough heat recovery, the bisulfate problem may be important.
It should be noted that although the formation of ammonium bisulfate
causes a problem, it helps reduce SOX emissions. Flue gas usually contains
a small amount of SOs (considered to be present as HzSOi* mist), which is
far more hazardous than S02. The S03 is not captured well even with FGD.
With the addition of ammonia, ammonium bisulfate and sulfate are formed and
132
-------
can be captured fairly easily by the heat exchanger and an electrostatic
precipitator.
3.5.2 Formation of SOs
The relationship of the percent of SO 3 in total SOX to the sulfur
content of heavy oil is shown in Figure 3-22. The relationship of 80s and
Qz concentrations in flue gas from the combustion of heavy oil is shown in
Figure 3-23. Those figures indicate that the percent of SOs in SOx
increases with the 02/SOX ratio of the gas. Therefore, low-oxygen com-
bustion to reduce NOx also helps to reduce SOs. The use of less oxygen,
however, tends to increase the particulate content; as Figure 3-24 shows,
particulates increase as SOs decreases. Usually 1.5-2.5 percent of total
SOX is present as SOs in flue gas from a high-sulfur oil and about 3 percent
in gas from a low-sulfur oil under good combustion conditions. It is said
that SO 3 concentration in gas from coal is lower than that in gas from heavy
oil because the former gas contains less vanadium which acts as a catalyst
in the reaction forming SOs and moreover, it contains much fly ash which
absorbs SOs to a considerable extent.
It should also be noted that the catalyst for selective catalytic
reduction of NOx tends to oxidize a small portion of S02 to SOs • The
oxidation ratio differs with the catalyst composition, temperature, and S02,
02 and NHs concentrations. Usually the oxidation ratio is below 2 percent
but a considerable amount of SOs is formed in S02-rich gas.
3.5.3 Formation of Ammonium Bisulfate and Sulfate
Figure 3-25 shows the melting point of an NHg-I^SOii system. The
melting point of ammonium bisulfate, NHijHSOit, is 147°C. The melting point
is lowered by addition of a small amount of ammonium sulfate but further
addition raises the melting point significantly. A mixture of the bisulfate
133
-------
01234
Sulfur in heavy oil (!8)
Figure 3-22. Formation ratio of SOa (02 in gas = 1-2 percent).1
100
70
50
30
1
£ 15
8 10
0.3 0.5 1.0 2.0 3.0
0* (%)
Figure 3-23.
Relation of 02 and SOa concentrations
(heavy oil, S - 1-3 percent).5
134
-------
400
200
100
50
20 40 60 80
Figure 3-24. Relationship of SOs concentration to particulate
content (heavy oil, S = 1-3 percent; Oz in gas «
0.3-3 percent).5
a
•H
0
Bu
60
e
150
100
0.8 1.0 1.2
NHS/H3SO«, mole ratio
o By melting point
measurement
x By solidifying
temperature
measurement
Figure 3-25.
system.
135
-------
and sulfuric acid has a lower melting point, is very acidic and is highly
corrosive.
The formation temperatures of liquid ammonium bisulfate from the
gaseous phase are shown in Figures 3-26 and 3-27. The figures indicate that
the bisulfate liquid forms at about 210°C from a gas containing 10 ppm each
of NHa and S03 , and at about 240 °C from a gas containing 100 ppm each of
NHa and SOs , provided that a sufficient amount of water vapor is present in
the gas. For a typical SCR reactor effluent containing about 10 ppm of
and 20 ppm of 80s, the formation temperature is about 215°C.
Figure 3-26 also indicates that solid ammonium sulfate can form from
liquid ammonium bisulfate and ammonia at relatively low temperatures in the
presence of excessive ammonia.
Figures 3-28 and 3-29 illustrate the formation temperatures of the
bisulfate and sulfate. The formation temperatures of the bisulfate agree
fairly well with those shown in Figures 3-26 and 3-27, but those of ammonium
sulfate are considerably higher than those indicated in Figure 3-26. For
example, Figure 3-28 shows that ammonium sulfate forms at 230° C from a gas
containing 10 ppm of SOs and 30 ppm of NHs . This temperature might be too
high, because the sulfate decomposes to bisulfate and ammonia at a high
temperature according to Figure 3-26.
Jumpei Ando, the author of the present report, has found that a double
salt of the sulfate and bisulfate, (NHOsH(SOO 2 , forms when a flue gas
containing SOs is cooled in the presence of excessive ammonia. Studies on
the formation and decomposition of the double salt should solve the above
discrepancy.
Ammonium bisulfate deposits on SCR catalysts even at 300-330°C. This
may be due to high concentrations of NHs and also SOs on the catalyst
136
-------
o
9
I
240
220
200
180
0 50 100
NH3 in gas (ppo)
Figure 3-26. Formation temperatures of
1000
Figure 3-27.
10 100
S03, ppm
Formation temperatures of
(NH3 + S03 + H20
gas gas gas liquid
1000
137
-------
CX
Q.
500
100
50
10
5
1
Figure 3-28. Formation temperatures of ammonium bisulfate and sulfate (1) . '
1 5 10 50 100 500
(ppm)
§. 0.5
Q.
0.1
0.05
0.01
0.1 0.5 1
0.01
Figure 3-29. Formation temperatures of ammonium bisulfate and sulfate (2).
138
-------
surface, when a portion of SOa in the gas is oxidized to SOs on the catalyst
surface.
3.5.4 Reaction of Ammonium Bisulfate
Ammonium bisulfate is acidic with a pH below 1, and the melt is highly
corrosive. Ando has found that the bisulfate melt reacts with iron to form
ferrous sulfate, which combines with ammonium bisulfate to form a newly
identified double salt — FeSCH
Fe + NH^HSOif + h 02 = FeSO^ + NH3 + H20 (3-16)
FeSOit + NHi»HS
-------
a deposits at a low temperature. Ammonium sulfate and the
double salt may not be corrosive because of the higher pH and a high melting
point. The pH of the ammonium sulfate solution is about 5 and of the
double salt solution below 1. The deposit of the sulfates also tends to
hinder the gas flow.
It seems that various deposits form under different conditions:
temperature, NHa/SOs ratio of inlet gas, degree of corrosion of the pre-
heater material, and degree of reaction with fly ash. The differences in
composition result in the differences in deposit properties—corrosiveness,
difficulty in removing by soot blowing, etc.
Nakabayashi and Mouri9 have found that NHitAl(SOit)2 and NHitFe(SO^)2
formed in an air preheater from a coal-fired boiler used for SCR tests.
The aluminum compound is formed by the reaction of NHifHSOit with Al20s in
fly ash from coal.
3.5.5 Deposit in Heat Exchanger and Air Preheater
Figure 3-30 shows a model of a Ljungstrom air preheater (or heat
exchanger) which has rotating elements. Gas leaving the SCR reactor at
300-420°C is introduced from one side of the element and cooled to 150-200
°C, while air (or cold gas) is introduced from the other side and heated
to 250-320°C.
Figure 3-31 shows temperatures of the gas and elements in a Ljungstrom
heat exchanger used at a commercial SCR plant. In this plant, flue gas from
an industrial boiler using high-sulfur oil is first treated by wet process
FGD. The gas is then introduced into the Ljungstrom heat exchanger where
it is heated from 55 to 310°C, while the gas leaving the SCR reactor at 410
°C is cooled to 170°C. After FGD, the gas contains about 150 ppm of NOX and
150 ppm of SOX (20-30 ppm of SOs); the gas leaving the reactor contains
10-20 ppm of NHa with about 20 ppm of NOX and 150 ppm of SOX.
140
-------
Gas Inlet
Air Inlet
Figure 3-30. Schematic of Ljungstrom type air preheater.
141
-------
500
400
300
UJ
as
I
200
100
I
I LOW
• TEMPERATURE
ZONE
INTERMEDIATE
ZONE
500
400
300
200
100
Figure 3-31. Temperatures of gas and heating elements in a
commercial Ljungstrom heat exchanger.
142
-------
Deposits, mainly FeSCvNHifHSOit with some carbon dust, formed on the
elements below 210°C and resulted in an increase in gas pressure drop.
(Figure 3-32 is a photograph of a Ljungstrom air preheater before and after
plugging.) Although soot blowing has reduced the deposits, the heat
exchanger has required a water wash every few months. The water wash pro-
motes corrosion because the wash liquor is" acidic with a pH of about 1.
Electric Power Development Co. (EPDC) conducted tests on the deposits
in an air preheater using flue gases from a coal-fired boiler containing
1400 ppm of SOX and 300 ppm of NOX; the flue gases had either a full par-
ticulate load or a slight particulate load after passing through a hot
electrostatic precipitator. The gases (9000 Nm3/hr) at 280-300°C were
passed through a Ljungstrom type air preheater to cool to 1408C, while air
at 10-40°C was heated to 260-270°C. Ammonia was injected into the gases at
concentrations of either 20 ppm or 50 ppm.
The results of tests on flue gas with a full particulate load are
shown in Figure 3-33.9 Ammonia at a concentration of 20 ppm was injected
for the first 480 hours and then at 50 ppm for 3240 hours (a total of 3720
hours). The pressure drop of the gas, which was 123 mmfcO initially,
increased appreciably with the injection of 20 ppm of NHs. With 50 ppm of
NHa, the pressure drop increased rapidly and reached 185 rninHzO at the end
of the test. Deposits were formed in the low temperature and intermediate
temperature zones of the heating elements.
After particulate removal in the flue gas by a hot electrostatic
precipitator, the deposits formed much more rapidly, resulting in a
remarkable increase in pressure drop. This shows that the fly ash of coal
has a cleaning effect.
EPDC has made further tests with low particulate gas and has succeeded
in reducing the deposit considerably. EPCD presumes that with high
143
-------
Figure 3-32. Photographs of Ljungstrom air preheater element.*
*Above: Before plugging.
Below: After plugging by bisulfate, sulfate, and dust.
(A coin 21 mm in diameter has been inserted for size comparison.)
144
-------
(rnnAg)
I
U C
soo
1OOO
1500
2000
2500
3000
3500
4000
Test Air Preheater Running Time (Hr)
Figure 3-33. Results of tests on flue gas with a full dust load.
-------
particulate gas the deposit may be reduced sufficiently to allow plant
operation.
Figure 3-34 presents an example of a multitube type air preheater used
for SCR. Air passes down the outside of the tubes, while hot gas leaving
an SCR reactor goes up inside the tubes. The hot gas is at a temperature of
about 370°C and contains 1500 ppm of SOX (possibly 40-50 ppm of SOs) and
about 5 ppm of NHs . The tube surface temperatures are between those of the
hot gas and the air as shown in the figure. It has been found that a
deposit forms inside the tube below 190°C.
Bulk Air
Temperature
Bulk Gas
Temperature
80 150
220
Bulk Air
Temperature
80
Bulk Gas
Temperature
t
£^_220T
100 170
120 190
140 210
160 230
* t
240
260
280
300
t
(Deposit s)>
^
170 100 170
190 120 190
210 140 210
230 160 230
t 1 1
Air Gas Air
\
240 1
260
280
300
t
Gas
Tube Surface Temperatures
Figure 3-34. Deposits in air preheater tubes (numbers show
temperatures of gas, air and tube).
146
-------
The amounts of deposit in the tubes are small and do not appreciably
increase the pressure drop, presumably for the following reasons:
• Ammonia in the gas is kept at a low concentration.
• The gas velocity in the tubes is about 10 m/sec, in
comparison with 2-3 m/sec in the Ljungstrom type heat
exchanger.
The outlet gas temperature is 200°C, as compared with 165°C
for the Ljungstrom heat exchanger. The lowest temperature
of the tube is 150°C, while the Ljungstrom element
temperature is 110°C.
It is possible that when the gas is further cooled for better heat
recovery, the amount of the deposit increases considerably. It is also
likely that the liquid phase formed in the tube from a gas containing
less ammonia is highly acidic (a mixture of ammonium bisulfate and sulfuric
acid) and is very corrosive.
There is a possible way to prevent deposits in such a tube type air .
preheater. When the tubes are placed vertically as shown in Figure 3-34
and the hot gas containing SOx and NHs is passed downward, ammonium bisul-
fate liquid will form in the tube. This liquid will drop off because the
lowest temperature of the tube is 150°C, a little above the melting point.
In this case, however, ferrous sulfate or the double salt may deposit if
the tube material corrodes.
3.5.6 Corrosion Tests
Hitachi Zosen performed corrosion tests with the commercial Ljungstrom
type heat exchanger for SCR, on 5 materials: SS-41, SS-41 coated with
polyimide resin, S-ten, SUS 316, and SUS 316 JIL. Test pieces of the
147
-------
materials, 120 x 50 x 3 mm, were placed at 6 different points in the
Ljungstrom and exposed to gas for 95 days. Since considerable deposits
formed on the pieces, they were washed with an acidic solution and then
weighed. The results are summarized in Table 3-6.
The corrosion rate was larger with the samples placed in colder posi-
tions due to larger amounts of acidic deposits. SUS 316 and SUS 316 JIL
were fairly resistant to the corrosion. It may be difficult, however, to
manufacture intricate structures of the air preheater elements with these
materials. The resin-coated material showed a relatively small corrosion
rate at 58-59°C, but the rate was much higher above 100°C because a portion
of the resin tended to come off at the higher temperature.
Sumitomo Chemical did corrosion tests with SUS 304, NAR-F, SUS 316,
and mild steel. Test peices of those materials were dipped in different
concentrations of NHitHSO^ solutions at 40°C and 80°C for 20 hours. The
corrosion rates are shown in Table 3-7-
Results of corrosion tests on SUS 304 and SUS 316 with NHifHSOt and
HaSOif are compared in Figure 3-35. The corrosion rates with NHitHSCK solu-
tion at 60-90 percent concentration are about equal to that of HzSOit.
3.5.7 Countermeasures for the Bisulfate Problem
The ammonium bisulfate problem may be minimized by using the following
countermeasures:
1) Reduction of SOs or NHs or both in the SCR reactor effluent.
2) Application of soot blowing.
3) Use of simple air preheater configurations such as tube type
to increase the benefits of soot blowing.
148
-------
TABLE 3-6. RESULTS OF CORROSION TESTS
PERFORMED BY HITACHI ZOSEN
Test Piece
Number
/
Stationary •<
Parts
Cold Gas Inlet
(58-59°C)
Cold Gas Outlet
(Warmed Gas)
(325-333°C)
Hot Gas Inlet
(410°C)
Hot Gas Outlet
(Cooled Gas)
(162-165°C)
t
Rotating Parts ,
(Elements)
Cold End
(105-120°C)
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Hot End
(364-380°C)
1
2
3
4
5
1 - SS-42
2 - SS-41 resin coated
3 - S-ten
4 - SUS 316
5 - SUS 316 JTL
Corrosion Rate
(mm/year)
0.424
0.012
0.518
0.0098
0.0094
0.013
0.032
0.012
0.011
0.011
0.0055
0.021
0.0037
0.0102
0.011
0.192
0.186
0.184
0.0035
0.0043
0.392
0.334
0.390
0.013
0.0048
0.110
0.198
0.127
0.0096
0.0051
149
-------
TABLE 3-7. RESULTS OF CORROSION TESTS PERFORMED BY SUMITOMO CHEMICAL
in
O
Temperature
Material °C
Mild Steel
SUS 304
STJS 316
NAR F
40
80
40
80
40
80
40
80
Corrosion Rates (g/m2/hr) at Different Solution Strengths
(percent) of NHnHSOi,
10
69.70
282.10
2.89
17.50
0.01
0.42
0.00
0.04
30
114.00
333.70
9.51
127.20
0.27
1.19
0.01
0.84
50
148.70
330.20
18.40
284.60
0.17
0.31
60
17.10
339.20
11.20
287.90
0.38
0.49
80
24.70
41.10
13.40
95.80
2.29
6.06
-------
1,000
100
00
x*^
01
10
g
1 1
n
14
O
0.1
SUS 304
SUS 316
NH<.HSO«.(80°C)
SUS 304
A A
(80°C)
60 70 80
Concentration(Z)
90 60 70 80 90
Concentration (Z)
98
Figure 3-35. Corrosion tests on steels.'
-------
4) Use of heat exchangers in which the SCR effluent passes
downward at a fairly large velocity to blow down the deposits.
Soot blowing may also work efficiently for such a heat
exchanger.
5) Use of corrosion resistant material for low-temperature parts
where the deposit forms.
When the gas contains a considerable amount of SOs, reduction of
in the gas decreases the amount of deposit but makes it highly corrosive.
The reaction product may not be easily removed by soot blowing. On the
other hand, when the gas contains much more NHs than SOa, ammonium sulfate
and the double salt of ammonium sulfate and bisulfate form. They are less
corrosive, have a high melting point, and may be relatively easily con-
trolled by soot blowing.
For flue gas from low-sulfur fuel, the ammonium bisulfate problem can
be largely solved by the above countermeasures. There is no large commer-
cial plant yet that treats SOy-rich gas by SCR and recovers the heat
sufficiently to lower the flue gas temperature to 160°C or below. Further
studies are needed for flue gas treatment from high-sulfur fuel.
It is the experience of Electric Power Development Co. (EPDC) and
other companies that the ammonium bisulfate problem is less serious with
flue gas from coal containing a large amount of fly ash. This may be due to
two reasons:
1) When the gas is cooled, ammonium bisulfate deposits on fly
ash in the gas and leaves the heat exchanger with the fly ash.
2) Even when a small amount of deposit forms on the heat
exchange material, the deposit tends to be removed by
erosion by fly ash.
152
-------
It has also been observed that the use of a hot electrostatic
precipitator to eliminate the fly ash in the gas increases the bisulfate
problem. Therefore, a parallel flow type reactor should be used to minimize
the bisulfate problem with particulate-rich gas.
3.6 COST OF SCR (INVESTIGATION BY ENVIRONMENT AGENCY)
3.6.1 Assumptions for Cost Estimation
The Environment Agency has investigated the costs of flue gas denitri-
fication mainly through hearings with process developers and users partly.
Three systems for new and existing oil-fired boilers were chosen for the
estimation of SCR costs; they are designated as A, B and C in Figure 3-36.
System A is for a new boiler. Flue gas from a boiler economizer at
350-400°C is injected with ammonia and directly introduced into an SCR
reactor without bypass. The gas leaving the reactor is passed through an
air preheater for heat recovery. System A is the simplest and most
economical of the three. It is used at the Chiba plant of Ukishima
Chemical (see Section 4.5).
System B consists of an existing boiler that has been modified by the
addition of ducts and a fan. In this system, gas that has passed through
the boiler economizer is sent to an SCR reactor by a fan. Then the gas
leaving the reactor is sent to an existing air preheater.
System C is for an existing boiler which cannot be modified by install-
ing System B. In System C, the gas leaving the air preheater and electro-
static precipitator at about 150°C is passed through a gas-gas heater and
a fan and is further heated by another heater to 350-400°C. The gas
leaving the reactor is passed through the gas-gas heater and sent to a
stack. This system is the most expensive because it requires a gas-gas
heater, another heater, and much energy for the heating.
153
-------
i 1
(A)
(B)
(C)
I I FACILITIES FOR COST ESTIMATION
ESP
| REACTOR |
I I
AIR
PREHEATER
STACK
Figure 3-36. SCR systems used in cost estimation.
154
-------
The equipment cost for System A consists of costs for a reactor,
ammonia storage, an ammonia vaporizer, electrical instrumentation for SCR,
and an initial charge of catalyst. System B includes costs for a duct and
fan in addition to the plant cost for System A. System C includes all the
costs for System B plus those for a gas-gas heater and another heater.
For cost estimating purposes, the flue gas composition was assumed to
be 200 ppm of NOX, 1500 ppm of SOX, and 200 mg/Nm3 of particulates at 380
°C. The NOX removal efficiency was assumed to be 80 percent.
To calculate the annual SCR cost, the following estimates were used:
Fixed Costs— 7 years depreciation
10 percent interest/year (maintenance, 3 percent;
insurance, 2 percent; and overhead, 5 percent
of investment cost)
Variable costs— Catalyst life 2 years
Annual operation 8000 hours
Power 12 yen/kWhr ($0.06/kWhr)
Ammonia 80 yen/kg ($0.40/kg)
Steam 2 yen/kg ($0.01/kg)
Kerosene for gas heating 32 yen/kg ($0.16/kg)
3.6.2 Equipment Cost and Annual Cost
Figures 3-37 - 3-39 show equipment costs in battery limits for Systems
A, B and C with different capacities ranging from 40,000 to 1,000,000 Nm3/
hr (13-330 MW equivalent). The plant costs were proposed by five vendors.
One of them presented a relatively high cost, while the other four presented
considerably lower costs. The high cost is close to the actual cost for
plants already in operation. The lower costs may be assumed to be based on
155
-------
Ui
1.2
1.0
o 0.8
g
l-f
S 0.6
i
I 0.4
0.2
0 100 200
500
GAS TREATED (1000 Nm3/hr)
800
1000
Figure 3-37. Investment cost of SCR by direct process (System A).;
-------
Ui
1.4
1.2
^ 1.0
UJ
ji-
ii.
o
S 0-8
3 0.6
o
111
y 0.4
0.2
100 200
500
GAS TREATED (1000 Nn|3/hr)
1000
Figure 3-38. Investment cost of SCR by direct process (System B).;
-------
2.5
1000
GAS TREATED (1000 Nm3/hr)
Figure 3-39. Investment cost of SCR with gas reheating and
heat recovery (System C).3
158
-------
extensive advances which permit cost reduction, but some of them could be
a little optimistic. Two of the curves in Figure 3-39 show a larger
inclination above 600,000 Nm3/hr than between 200,000 and 500,000 Nm3/hr.
This is due to the use of one reactor below and two reactors above 600,000
Nm3.
The average costs of the five vendors for the three systems are com-
pared in Figures 3-40 and 3-41. The average cost for a 200-300 MW plant
is about 2800 yen/kW ($14/kW) for System A, 3700 yen/kW ($18.50/kW) for
System B, and 6300 yen/kW ($31.50/kW) for System C. The difference between
System A and System B is due to the additional ducts and the fan, and that
between System B and System C is due to the heat exchanger (gas-gas heater),
heater and duct.
Figures 3-42 - 3-44 show annual SCR costs for the 3 systems. The
costs differ considerably from vendor to vendor. Figure 3-45 shows the
average of the 5 vendors' costs for the 3 processes. For a 200-300 MW size
plant, the cost is about 0.3 yen/kWhr ($0.0015 kWhr) for System A, 0.4 yen/
kWhr ($0.002/kWhr) for System B, and 0.7 yen/kWhr ($0.0035/kWhr) for System
C. The latter cost is the highest due to the high investment cost and fuel
requirement for gas heating.
3.6.3 Factors that Affect Cost
System A is simple and has the lowest cost, which is only one-fourth
that for flue gas desulfurization. (FGD entails about 15,000 yen/kW in
investment and 1.4 yen/kWhr in annual costs.) The cost of System C is only
about half that of FGD. There are, however, several factors that may
increase the cost for SCR as shown below:
1) Energj Requirement for System A - System A includes no fan.
The fan for the boiler, however, needs increased capacity and
energy to compensate for the pressure drop by the SCR system.
159
-------
2.0
CO
CO
o
CJ
1.8 h
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
100
3APACITY (MW)
200
300
200 400
600
CAPACITY (1000 Nm3/hr)
(a)
Figure 3-40. Average plant costs per capacity
for Systems A, B and C.
400
800 1000 1200
160
-------
100
CAPACITY (MW)
200
300
(c)
s.
.c
oo
o
o
o
>-
o
400
10
6 ^
c
-------
100
200
300
400 500 600
GAS TREATED (1000 Nm3/hr)
700
800
900
1000
Figure 3-42. Annual cost of SCR by direct process (System A).
* Heavy oil, about 9500 kCal/1.
-------
CO
o
o
o
UJ
5* 1
0
0 40 100
200
400 500 600
GAS TREATED (1000 Nm3/hr)
700
800
900
1000
Figure 3-43. Annual cost of SCR by direct process (System B).
* Heavy oil, about 9500 kCal/1.
-------
I-1
ON
•P-
0
0 40 100
200
300 400 500 600
GAS TREATED (1000 Nm3/hr)
800
900
1000
Figure 3-44. Annual cost of SCR process with gas reheating and heat recovery (System C).
* Heavy oil, about 9500 kCal/1.
-------
100
CAPACITY(MW)
200
300
400
1.2
1.0
•St.
r>
(C)
(B)
(A)
0.8
0.6
o>
>>
<
I
0.4
0.2
0
1200
200 400 600 800 1000
CAPACITY (1000 Nm3/hr)
Figure 3-45. Average annual costs of SCR
with Systems A, B and C.
165
-------
When this cost is taken into account, the cost for System A
may increase by 10-20 percent.
2) Boiler Load Fluctuation - With utility boilers which undergo
a considerable load fluctuation, the economizer outlet gas
temperature may occasionally drop below 300-330°C, resulting
in the deposit of ammonium bisulfate on the catalyst. When a
high sulfur fuel is used, it may be necessary to maintain the
gas temperature above 300-330°C; this can be achieved either
by installing an auxiliary burner or a bypass to mix a
portion of the hot gas before it goes through the economizer
with the gas after the economizer.
Such facilities increase the investment cost. The operating
costs may not increase appreciably because most of the heat
required to raise the gas temperature can be recovered by
the air preheater.
3) Ammonium Bisulfate Problem - The ammonium bisulfate problem
described in Section 3.5 may increase the cost somewhat. The
problem, however, may not be appreciable with flue gas from
a coal-fired boiler because ammonium bisulfate can be removed
with the ash.
4) Catalyst Life and Disposal - The catalyst life of 2 years
assumed for the cost estimation may not always be achieved
for dirty gas rich in S0x» particulates and catalyst poisons.
Disposal of the spent catalyst may add some cost.
5) Combination with FGD - When FGD is applied after SCR,
accumulation of ammonia in the scrubber liquor and the need
for ammonia removal may constitute a cost-increasing factor.
166
-------
Although there are several factors that may increase SCR costs, there
is also the possibility of cost cuts through further improvement of the
catalyst and reactors. More complete and accurate SCR cost data will be
available in the near future from the operation of many large plants for
dirty gases.
167
-------
3.7 REFERENCES
1. Atsukawa, M., et at., Development of NOX Removal Processes with
Catalyst for Stationary Combustion Facilities. Mitsubishi Technical
Bulletin, No. 124, May 1977. In English.
2. Kasaoka, S., Present Situation of Fundamental Studies and Techniques
on Dry Reduction Removal of Nitrogen Oxides—Catalytic and Non-Catalytic
Methods. Nenryo Kyokaishi, No. 56, Vol. 599, 1977- pp. 145-167-
3. Report on NOX Abatement Technology. Air Preservation Bureau, Environ-
ment Agency, April 1978.
4. Atsukawa, M., et al., Development of Parallel Flow Type NOx Removal
Systems Using Plate Type Catalyst. Mitsubishi Juko Giho, Vol. 15, No.
2, 1978.
5. Ando, J., Fundamental Problems of Sulfur Oxides and Flue Gas Desulfuri-
zation. Industrial Pollution Control, Vol. 12, No. 9, 1976.
6. By Sumitomo Chemical.
7. By Mitsubishi Heavy Industries.
8. By Hitachi Zosen.
9. Nakabayashi, Y., and K. Mouri, Tests of NHs/SOx Compound Deposit
Problems on Air Preheater at Coal Firing Boiler. EPDC TPD-D 52-001,
September 1977. In English.
10. Operating Experiences of High Temperature Non-Catalytic DeNOx Process
at No. 2 Unit of Chita Thermal Power Station. Central Technical
Research Laboratory, Chubu Electric. In English.
168
-------
SECTION 4
SCR PROCESSES AND PLANTS
4.1 SUMITOMO CHEMICAL SCR PROCESSES
4.1.1 Introduction
In 1968 and 1972, Sumitomo Chemical, one of the largest chemical and
petrochemical companies in Japan, constructed two nonselective catalytic
reduction process plants for the removal of NOX from the tail gas of nitric
acid plants. Sumitomo Chemical has further developed the chemistry and
technology for the selective catalytic reduction of NOX in flue gas, tested
more than one thousand catalysts, and constructed several commercial SCR
units. In 1974, they completed the world's first commercial SCR unit for
combustion gas treatment.
For the commercial units, Sumitomo Chemical uses granular catalysts in
fixed and moving beds. The SCR units treat flue gas from industrial boilers
or from heating furnaces at ammonia and methanol plants of Sumitomo Chemical
and related companies.
4.1.2 Commercial Plant for Clean Gas Treatment
The first commercial SCR plant began operation in May, 1974. It treats
200,000 Nm3/hr of clean flue gas from the burning of LPG in a reformer at
the Sodegaura plant, Higashi Nihon Methanol Co. (HNM). A flowsheet of the
denitrification process at the HNM plant is shown in Figure 4-1. The flue
gas containing about 200 ppm NOX (Mainly NO) at about 320°C is injected with
ammonia at an NHs/NOx mole ratio of 1 and introduced into a reactor
169
-------
3*
REACTOR
FAN
REFORMER
BYPASS
HEAT
.EXCHANGER
FAN
AIR
STACK
FAN
Figure 4-1. Flowsheet of the denitrification process at the Higashi Nihon Methanol Company.
-------
containing a base-metal catalyst. About 90 percent of the NOX is reduced
to NZ by reaction with ammonia. The treated gas passes through a fan and
heat exchanger to a stack.
The data for the standard operation of the HNM plant and two other
SCR plants for clean gas treatment are shown in Table 4-1.
TABLE 4-1. DATA FOR OPERATION OF SCR PLANTS FOR CLEAN GAS
Major Product
Fuel
Gas for SCR (Nm3 /hr)
Inlet Gas Composition
Oa (percent)
SOx , part iculat e
NOX (ppm)
Reaction Temperature (°C)
Load Fluctuation (percent)
Space Velocity (hr"1 )
NH3/NOX Mole Ratio
NOx Removal (percent)
Leak Ammonia (ppm)
Pressure Drop (mmHaO)
Reactor
Total System
Plant Completed
Higashi Nihon
Methanol (HNM)
Methanol
(800 t/day)
LPG
200,000
7
None
200
320
60-100
7000
1.0
Over 90
Below 10
100
250
May 1974
Plant Owner
Nippon
Ammonia
Ammonia
(1550 t/day)
LPG
250,000
6
None
200
320
60-100
6000
1.0
Over 90
Below 10
100
250
March 1975
Sumitomo
Chemical
Ammonia
(810 t/day)
Off Gas
200,000
5
None
200
320
60-100
10,000
1.0
Over 90
10-20
50-60
100
March 1975
171
-------
The catalyst is granular and has a compressive strength of 12 kg/
granule. It is placed in a fixed bed.
The results of tests made at the HNM plant are shown in Figure 4-2.
They agree well with those of basic studies.
All of the plants for clean flue gas treatment have been operated
trouble-free. Two plants have been operated for 3 years; Nippon Ammonia's
plant has been operated for only about 10,000 hours due to the lack of
ammonia production requirements. The original catalyst is still reactive
and there is no plan to replace it.
4.1.3 Dirty Gas Treatment
A pilot plant and two commercial plants of Sumitomo Chemical at
Sodegaura treat the dirty gas from boilers burning medium-sulfur oil.
Operation data for these plants are shown in Table 4-2. A flowsheet of
the two commercial plants is shown in Figure 4-3.
Both of the commercial plants are for existing boilers with a flue gas
temperature of 150°C. Therefore, a heat exchanger (gas-gas heater) and an
oil-fired heater were installed to heat the gas to 320°C. The No. 1 boiler
has no electrostatic precipitator. A highly efficient wet electrostatic
precipitator was installed to eliminate not only particulates but also SO3
mists which affect the a-AlaOa based catalyst. The reactor is of the fixed
bed type. The No. 2 boiler has an electrostatic precipitator. A moving
bed reactor, designed and manufactured in cooperation with Mitsubishi Heavy
Industries, has been installed for the No. 2 unit.
In both units, gas heated to 320°C by the heat exchanger and heater is
injected with ammonia, routed to the reactor to remove over 90 percent of
the NOX, and passed through the heat exchanger before being sent to a
common stack.
172
-------
100
90
80-
(A)
• 330"C
0300»C
0.7,- 0.8
0.9 1.0 T7T
HH3/NOX MOLE RATIO
100
3
90
80
(B)
250
^-T—«— •
NH3/NO-1.2
300 3SO
REACTION TEMPERATURE. *C
400
20r r
10
(C)
Q300*C
O1-8*
0.3 O TTO 171 1.2 T.3
NiyNOg MOLE RATIO
Figure 4-2. Operation data of the HNM plant and two other SCR plants,
173
-------
TABLE 4-2. OPERATION DATA OF SCR PLANTS FOR DIRTY GAS
Gas for SCR (Nm3/hr)
Fuel
Load Fluctuation
(percent)
Stack Height (m)
Inlet Gas Composition
02 (percent)
SOx (ppm)
NOX (ppm)
Particulates After ESP
(mg/Nm3 )
FGD Unit
Space Velocity (hr"1)
Temperature (°C)
NOX Removal (percent)
NH3/NO Mole Ratio
Leak Ammonia (ppm)
Type of Reactor
Pressure Drop (mmt^O)
Reactor
Total System
Plant Completed
Pilot
30,000
Oil
(S=0.7%)
60-100
70
6
400
200
5-20
None
5000
320
Over 90
1.0
10-20
Fixed Bed
-
-
July 1973
Commercial
No. 1
240,000
Oil
(S-0.7%)
60-100
140
6
400
200
5-10
Scheduled
5000
320
Over 90
1.0
10-20
Fixed Bed
200
500
March 1976
No. 2
300,000
Oil
(S-0.7%)
60-100
140
6
400
200
10-20
Scheduled
5000
320
Over 90
1.0
10-20
Moving Bed
200
500
October 1976
174
-------
Ui
Figure 4-3. Flowsheet of two Sumitomo Chemical commercial plants for dirty gas.
-------
The commercial plants are on boilers at ammonia plants. Since ammonia
has been in oversupply and ammonia production has been stopped or slowed,
the boilers and SCR plants have been operated very little since their
completion.
4.1.4 New Catalysts
Sumitomo Chemical has continued extensive tests on the SCR catalyst.
After screening tests with several thousand samples, new catalysts have
been produced. Catalyst A is based on y-AlzOa and is an improved type of
the catalyst used for heating clean gas at three commercial plants since
1974. Figure 4-4 indicates that nearly 90 percent NOX removal is obtained
around 350°C at a space velocity of 20,000 hr"1.
Catalyst B is based on TiOa, more effective than A (see Figure 4-4),
and resistant to SOx. Virtually no decrease in activity was observed in
two years' test with dirty gas (Figure 4-5). Catalyst C was developed
earlier for use with dirty gas. It is based on a-AlaOa but which showed a
considerable decrease in activity after a year of use. A catalyst similar
to C has been used at the two commercial plants for dirty gas (Table 4-2).
Sumitomo Chemical is going to use Catalyst B commercially because it is not
only S0x-resistant, but also far more effective than Catalyst C. Ninety
percent removal efficiency is obtained at a space velocity of 15,000 hr"1
while Catalyst C requires a space velocity of 5000 hr"1. Catalyst B may
also be used for clean gas. Although Catalyst B costs more than Catalyst
A, Catalyst B is more efficient.
4.1.5 Economics
The HNM plant cost 250 million yen ($1,250,000) in 1973. About 800
kW are required for a blower to compensate for the pressure drop of the
gas through the reactor.
176
-------
100
80
a 60
40
X
o
20
CATALYST - B
250
CATALYST - A
SV 20,000 HR
NH3/NOX 1.0
-1
300 350
TEMPERATURE °C
400
CATALYST - A : Based on tf-
CATALYST - B : Based on TiO
Figure 4-4. NOX removal efficiency vs. reaction temperature
for Catalysts A and B.
177
-------
100
90
>-
o
X 80
70
x
o
60
50
CATALYST - B
CATALYST - C
SOX 600-800 ppm
TEMP. 350 C
mi i NOX 0.9
SV CAT. -B 15,000 HR'1
CAT. -C 5,000 HFT1
_L
0.5 1.0 1.5
LAPSE OF TIME, YEARS
CATALYST - B : Based on Ti02
CATALYST - C : Based on a-A!203
2.0
Figure 4-5. SOX resistivity of Catalysts B and C.
178
-------
The plant of Nihon Ammonia (250,000 Nm3/hr of clean gas) cost 600
million yen ($3,000,000) in 1975. The cost is equivalent to 7200 yen/kW
($36/kW). The costs estimated in 1975 for plants with different capacities
for clean gas treatment are shown in Table 4-3.
TABLE 4-3. ESTIMATED COST FOR CLEAN GAS TREATMENT IN 1977 FOR
EXISTING BOILERS, INLET GAS 350°C, NOX 230 PPM
Capacity (Nm3/hr)
Initial Cost (1000 yen, $5)
Plant Cost in Battery Limit
Initial Charge of Catalyst3
Annual Cost (1000 yen/yr, $5/yr)
Depreciation
Interest
Catalyst
Ammonia
Power
Other
Total Annual Cost
Q
Annualized Cost
Yen/103 Nm3 ($0.005/103 Nm3)
Yen/kl oil ($0.005/kl)
Yen/kWhr ($0.005/kWhr)
100,000
280,000
15,000
25,200
15,400
5,000
8,400
42,000
1,400
97,400
1,220
1,700
0.39
200,000
430,000
30,000
38,700
23,700
10,000
16,800
64,000
2,200
154,000
970
1,360
0.32
400,000
650,000
60,000
58,500
35,800
20,000
33,600
97,000
3,300
248,200
780
1,090
0.25
SV = 20,000 hr-1
For 10 years
C8000-hour annual operation
The plant cost for treatment of dirty gas may be more than double the
cost for clean gas. A highly efficient electrostatic precipitator, heat
exchanger, and heater may be required.
179
-------
4.1.6 Evaluation
Sumitomo Chemical pioneered, the development of SCR chemistry and
technology. Its success in operating the commercial plants, including the
world's first, and in producing S0x-resistant and ammonia decomposition
catalysts has encouraged other companies to develop SCR catalysts and
processes. The good agreement of the results of fundamental studies with
plant operation shows the reliability of the chemistry and technology
developed by Sumitomo Chemical.
The plant cost, however, is relatively high presumably because the
plants were constructed with extra safety to ensure long-term trouble-free
operation. The newly developed TiOa-based catalyst has outstanding fea-
tures. Use of the new catalyst will not only increase the NOx removal
efficiency but also reduce the costs.
4.2 HITACHI ZOSEN (SHIPBUILDING) PROCESS
4.2.1 Introduction
Hitachi Zosen (Hitachi Shipbuilding and Engineering Co.), one of
Japan's largest shipbuilding and engineering companies, has been very
actively developing SCR technology for dirty gases. After tests of more
than 10 pilot plants for various gases with different types of reactors
and catalysts, Hitachi Zosen constructed 5 commercial plants including 3
large ones (Table 3-1). At the 3 large plants constructed in 1975 and
1976, the gases at 55-160°C are heated to about 400°C by a heat exchanger
and an oil-fired heater before being sent to SCR. The heating is costly,
requiring large investment and fuel oil costs.
Hitachi Shipbuilding recently began to develop honeycomb type cata-
lysts, both ceramic and metallic, by which dirty flue gas discharged from
a boiler economizer at 350-400°C can be treated directly. They have made
180
-------
pilot plant tests with the metallic honeycomb catalyst for flue gas from
oil- and coal-fired boilers and other sources. A pilot plant funded by
the United States EPA will be constructed in the United States by Hitachi
Zosen with the metallic honeycomb catalyst for coal-fired boiler flue gas.
4.2.2 Catalysts
Table 4-4 lists major catalysts developed by Hitachi Zosen.
TABLE 4-4. MAJOR SCR CATALYSTS DEVELOPED BY HITACHI ZOSEN
NOy Removal Catalysts (Noxnon)
200 Series - 203, 204, 205, 206, 207, 208
- Cylindrical or tablet, for clean gas
300 Series - 303, 304, 305, 306, 307, 308
- Cylindrical, spherical, and tablet, for dirty gas
400 Series - 403, 404
- Ceramic honeycomb, for dirty gas
500 Series - Standard type, low-temperature type and abrasion-
resistant type
- Metallic honeycomb, for dirty gas
Decomposition Catalysts (Ammonon)
300 Series - 301, 303
- Spherical, for clean and dirty gases
500 Series - 501, 503
- Metallic honeycomb, for dirty gas
181
-------
is used as a catalyst carrier for clean gas, while SiOa,
, Ti02 , or a special alloy is used for dirty gas. For catalyst
metals, V, Mn, Fe, Co, Ni, Cu, Cr, W, Mo, etc. are used.
Forced tests of the deterioration of the catalysts by SOs have been
carried out. Figure 4-6 shows the apparatus for the tests. Sulfuric acid
is heated to generate 80s gas. The SOs gas is diluted to 3 percent concen-
tration (30,000 ppm) with nitrogen gas and passed through a bed of catalyst
(3 mm granules) at 400-420°C with a space velocity of 4000 hr"1 for periods
of from 10 to 60 minutes. The catalysts which have been exposed to the
80s gas are then used for SCR tests at 400-420°C with a space velocity of
10,000 hr'1.
CATALYST
ELECTRIC
HEATER
98% H2S04
nrrrrrrm
ELECTRIC HEATER
Figure 4-6. Apparatus for a test of the forced deterioration
of a catalyst by 80s.
182
-------
As shown in Figure 4-7, Noxnon 203, a catalyst based on y-AlaOs, was
seriously affected by the 80s-containing gas, while Noxnon 303 and 403
showed only a slight decrease in activity after the first ten minutes and
virtually no further change. The NOx removal efficiency of Noxnon 403
(about 60 percent) may be low because the space velocity is too large for
the honeycomb type catalyst.
O-
303
403
203 (on If- A1203)
10 20 30 40 50 60
FORCED DETERIORATION TIME (WIN.)
Figure 4-7. NOx removal efficiency of Noxnon 203, Noxnon 303 and Noxnon 403
after forced deterioration of the catalyst by 80s.
The structure of the metallic honeycomb (Noxnon 500) is shown in
Figure 4-8. The catalyst is produced by etching and activating the surface
of special alloy plates with a thickness of 0.5-1.0 mm. About 0.1-0.2 mm
of the surface of the plates is activated. A large honeycomb with a small
pitch (opening) can be made. One standard size is 0.5 x 0.5 x 0.5 m; larger
honeycombs can be made.
183
-------
Figure 4-8. Structure of Noxnon 500.
To specify the relation between gas volume and amount of catalyst,
area velocity (AV) is used instead of the space velocity, SV, for granular
catalysts.
AV =
Gas rate (Nm3/hr)
Apparent surface area of catalyst (m2)
The apparent surface area of the catalyst is measured assuming that the
surface is flat. An AV of 8 m /m *hr is equivalent to an SV of about
5000 hr
-i
184
-------
The metallic honeycomb causes a small pressure drop, as shown in
Figure 4-9. Since a bed depth of 1.5 m is sufficient to maintain 90
percent NOx removal with a honeycomb with 4 mm pitch at a gas velocity of
about 5 m/sec at 400°C, the pressure drop of the gas by the honeycomb for
actual use is only about 40 mm
o
CNJ
100
~ 50
is
O CO
LU •
ce. en
= o
GO r-
to—•
LU
on
0.
20
10
5
2
1
I
1 2 5 10 20
GAS VELOCITY (m/sec.)
(log. scale)
Figure 4-9. Pressure drop over 1 m of packed bed (Noxnon 500)
The NOX removal efficiency of the Noxnon 500 catalyst as a function of
temperature, NHa/NOx mole ratio and time obtained in a pilot plant with a
flue gas from an oil-fired boiler (200 Nm3/hr) containing 150 ppm of NOx,
400 ppm of SOX and 50-100 mg/Nm3 of particulate is shown in Figures 4-10 -
4-12. Reaction temperature of about 400°C or above provides greater than
95 percent NOX removal. Figure 4-11 indicates that over 90 percent of the
NOX was removed even at the NHa/NOx mole ratio of 0.8. These results
indicate that the action of the metallic honeycomb may be better expressed
185
-------
100
90
CM
O
80
350
AV = 10 m3/m2 X hr.
NH3/NOX -1.0
400
450
REACTION TEMPERATURE (°C)
Figure 4-10. Initial efficiency of Noxnon 500 at various temperatures.
<
o
X
o
100
80
60
NOV REMOVAL
400 - 410°C
AV = 10 m3/m2 X hr.
0 0.2 0.4 0.6 .0.8 1.0 1.2 1.4 1.6
NH3/NOX MOLE RATIO
Figure 4-11. Initial efficiency of Noxnon 500 at various
NHs/NOx mole ratios.
ox
100
90
80
70
AV = 8 m3/rn2 X hr.
400 - 410°C, NH3/NOX =1.0
5000
0 1000 2000 3000 4000
TEST PERIOD (hr.)
Figure 4-12. Efficiency of Noxnon 500 as a function of time.
186
-------
by Equation 4-1 rather than by 4-2, which has been assumed for ceramic-
based catalysts.
10NO + 8NH3 + 02 = 9N2 + 12H20 (4-1)
4NO + 4NH3 + 02 = 4N2 + 6H20 (4-2)
Figure 4-12 shows that the removal efficiency at an AV of 8 m3/m2*hr
was about 98 percent at first and dropped to about 90 percent after 2000
hours. No further decrease was observed.
When the metallic honeycomb becomes less reactive from contamination
or from possible erosion by particulates, it can be reactivated by washing
or by etching to remove the contaminated surface.
Hitachi Shipbuilding also has produced ammonia decomposition catalysts,
spherical catalyst for clean gas and metallic honeycomb for dirty gas.
Supposedly, 90 percent of NHs in the reactor outlet gas can be decomposed
by the catalyst, but no details have been published.
4.2.3 Commercial Plants
Hitachi Shipbuilding constructed three large commercial SCR plants for
dirty and semi-dirty gas treatment. These plants use a granular catalyst in
a fixed bed (Table 4-5). The Chiba plant, Idemitsu Kosan, can treat 350,000
Nm3/hr flue gas from a CO boiler and a furnace. This gas contains 230 ppm
N0x» 50-80 ppm S02, and 20-50 mg/Nm3 particulates. The plant began opera-
tion in October 1975.
The gas at about 160°C is heated in a heat exchanger to nearly 320°C
and then to 400°C by a heater. It is then treated by SCR to remove more
than 90 percent of the NOx- The gas discharged from the reactor is cooled
to 260-270°C in the heat exchanger.
187
-------
TABLE 4-5. OPERATION PARAMETERS OF MAJOR PLANTS
CONSTRUCTED BY HITACHI ZOSEN
Completed
Plant Site
Gas Source
Capacity (Nm3/tir)
Load Factor (percent)
Pretreatment of Gas
Reactor Inlet
NOx (ppm)
SOx (ppm)
Particulates (mg/Nm3)
Oz (percent)
Reactor Type
Reaction Temperature
NOX/NH3 Ratio
Catalyst No.
Space Velocity (hr-1)
NOX Removal (%)
Pressure Drop by SCR
Reactor (mmH20)
Catalyst Life
Idemitsu
Kosan
Oct 1975
Chiba
FCC-CO
Boiler and
Furnace
350,000
50-100
Heating
230
50-80
20-50
2.3
Fixed Bed
400
1.0
204
5,000
93
170
1 year
Shindaikyowa
Petrochemical
Nov 1975
Yokkaichi
Oil-Fired
Boiler
440,000
50-100
EP*, FGD,
Heating
150
80-130
30-100
3.2
Fixed Bed
410-420
1.0
304
10,000
80***
160
1 year
Kawasaki
Steel
Nov 1976
Chiba
Iron-Ore
Sintering
Machine
762,000
70-100
EP, FGD
WEP**,
Heating
200-300
5-20
3-10
11.2
Fixed Bed
380-390
1.0
304
4,000
95
50
1 year
Electrostatic precipitator.
** Wet electrostatic precipitator.
*** Leakage in heat exchanger reduced overall removal.
188
-------
The plant was operated for a short time without appreciable trouble
but has been shut down partly because the NOx concentration of the gas
dropped to about 100 ppm. The drop in NOX concentration was due to a
decrease in the boiler load. Another reason for the shutdown was the high
operation cost requiring a large amount of fuel for heating the gas.
The second plant, Yokkaichi plant of Shindaikyowa Petrochemical, can
treat 440,000 Nm3/hr flue gas from an oil-fired boiler (Figure 4-13). The
gas, which contains about 1500 ppm SOa, 150 ppm NOX, and 150 mg/Nm3 particu-
lates, is first treated by the Wellman-Lord-MKK flue gas desulfurization
process to reduce the level of 862 to about 100 ppm and the dust level to
about 50 ppm (20 ppm of the dust is a sodium salt derived from the mist of
the scrubber liquor). The treated gas at 55-60°C is heated to 330-340°C
in a Ljungstrom type heat exchanger and to 410-420°C by a heater. It is
then subjected to SCR to reduce NOX from about 150 ppm to 20 ppm. The gas
is cooled to 160-170°C in a heat exchanger. Because land space is limited,
there is no fan between the heater and the heat exchanger. The lack of a
fan allows inlet gas to leak into the outlet stream. Because of this
leakage, the NOX concentration varies from 20 to 30 ppm through the
exchanger.
The plant began operation in November 1975 and has been operated for
6000-7000 hours yearly, mainly because the boiler was shut down. There has
been no serious trouble with the SCR unit. . A catalyst on an alumina carrier
was charged initially and later replaced with an improved catalyst, Noxnon
304 on a silica carrier resistant to SOX. Major problems with the plant
have been the requirement of a large amount of energy for gas heating and the
deposit of ammonium bisulfate in the heat exchanger. Several soot blowings
a day and an occasional water wash are required for controlling the deposit.
The third plant was installed for Kawasaki Steel at Chiba. It can
treat about 800,000 Nm3/hr flue gas from an iron-ore sintering machine.
189
-------
400*C
AIR PREHEATER
BOILER
\
T
AIR
150'C
FGD
55-60'C
REACTOR
HEAT EXCHANGER
(LJUNGSTROM)
160-170'C
T s;
^ z.
330-340'C
HEATER
/ \
410-420'C
STACK
1
NH3
400-410°C
Figure 4-13. Flowsheet of the Yokkaichi plant of Shindaikyowa Petrochemical.
-------
The gas contains 200-300 ppm NOX and 300-400 ppm S02 particulates. If is
first treated with an electrostatic precipitator to reduce the particulate
level to 70 mg/Nm3, and then by the MHI lime-gypsum flue gas desulfurizatlon
process to reduce the S02 concentration to about 10 ppm. The gas is then
passed through a wet electrostatic precipitator to reduce the particulate
concentration to below 10 mg/Nm3. The cleaned gas is passed through a heat
exchanger, heater and subjected to SCR. The plant began operation in
November 1976 and has operated smoothly since then. The high degree of gas
cleaning prior to SCR is helpful but both investment and operation costs
seem high. Ammonium bisulfate is not an appreciable problem because of the
low SOs content of the gas; 80s is caught partly by FGD and predominantly
by the wet electrostatic precipitator.
4.2.4 Economics
The costs estimated by Hitachi Zosen for treating gas from oil-fired
boilers with different capacities (33-500 MW) using a low-sulfur oil (S
below 0.5 percent) and high sulfur oil (S = 3 percent) are shown below. The
following systems with each of the two fuels were assumed for the estimation
(Figure 4-14).
I. Low-sulfur oil (no desulfurization)
a) SCR is applied to the gas which leaves a boiler economizer at
about 400°C. After the SCR, the gas is passed through an
air preheater and an electrostatic precipitator and sent to
a stack at 200°C.
b) The gas leaving an air preheater and electrostatic precipitator
at 150°C is heated to 400°C by a heat exchanger and an oil-
fired heater and then subjected to SCR. The gas leaving the
SCR reactor at 400°C is passed through a heat exchanger and
then sent to a stack at 220°C.
191
-------
FACILITIES FOR COST ESTIMATION
I-a
BOILER
#2224
AIR
PREHEATER
'400
I X
ZOO'C
J REACTOR J
ESP
STACK
I-b
400*C
AIR
PREHEATER
BOILER
150-C
150'C
ESP
REACTOR
220'C
330*C
STACK
HEATER
400'C
HEAT
"EXCHANGER
I
Figure 4-14. Hitachi Zosen systems assumed for a cost estimate.
192
-------
FACILITIES FOR COST ESTIMATION
BOILER
3 4ob°C~
I AIR
PREHEATER
400°C
150"C
REACTOR
ESP
"I
HEATER
55-60°C
130 C
STACK
400°C
AIR
PREHEATER
BOILER
ESP
150'C- = ! 150'C
U,
130°C
400 *C
HEAT
EXCHANGER.
330"C
55-60°C
STACK
REACTOR
HEATER
400°C
Figure 4-14. (Continued).
193
-------
II. High sulfur oil (FGD)
a) SCR is applied to the gas which leaves a boiler economizer
at about 400°C. After the SCR, the gas is passed through
an air preheater and an electrostatic precipitator. It
then proceeds to an FGD system using the limestone/gypsum
process. The gas is then reheated to 130°C and led to a
stack.
b) The gas leaving an air preheater and electrostatic precipitator
at 150°C is first subjected to FGD by the limestone/gypsum
process. The gas, cooled to 55-60°C, is heated by a heat
exchanger and heater to about 400°C and subjected to SCR. The
gas leaving the reactor at 400°C is passed through the heat
exchanger to cool to 130°C and is then sent to a-stack. The
system is similar to that of the Yokkaichi plant, Shindaikyowa
Petrochemical (Figure 4-13), except that better heat recovery
is assumed.
The assumptions for the cost estimation are shown below:
a) NOX content of gas
150-180 ppm for low-sulfur oil
260-300 ppm for high-sulfur oil
b) NOx removal efficiency—90 percent
c) SOx removal efficiency—90 percent
d) Number of scrubbers and reactors
One scrubber and one SCR reactor up to 1,000,000 Nm3/hr of
gas (330 MW equivalent). Two scrubbers and two reactors are
needed for larger capacity.
e) Catalyst cost—1.5 million yen for 1 MW ($75,000/MW)
194
-------
f) Catalyst life—1 year
g) Operation hours—8000 hours yearly
h) Depreciation—7 years
i) Interest—10 percent per year
j) Maintenance cost—3 percent of investment per year
k) Tax and insurance—2 percent of investment per year
1) Labor
For SCR only (low-sulfur oil)—1 person/shift/reactor
For SCR and FGD (high-sulfur oil)
- Below 330 MW—3 persons/shift plus 5 persons in daytime
- Above 330 MW—3 persons/shift plus 7 persons in daytime
- Yearly salaries—3 million yen per person ($15,000/person)
m) Fuel oil for reheating—25 yen/10,000 kcal ($0.125/10,000 kcal)
n) NH3—80 yen/kg ($0.40/kg)
o) CaCOs—3 yen/kg ($0.015/kg)
p) Power—12 yen/kWhr ($0.06/kWhr)
q) By-product gypsum—no value
Investment and annualized costs for the application of SCR to gas from
low-sulfur oil burning are shown in Figure 4-15. The investment for system
I-a, which treats gas from a boiler economizer, includes a reactor, ammonia
storage and vaporizer, a fan to compensate for a pressure drop of 110
195
-------
, and Instruments for measurement and control. System 1-b, which
treats gas from an air heater, includes a heat exchanger and heater in
addition to the equipment of System I-a.
The investment cost for System I-a with a capacity larger than 300 MW
is about 4000 yen/kW ($20/kW). Costs for System I-b are almost double
because of the added cost of the heat exchanger and heater. There would be
no cost advantage of scaling the system above 330 MW, because two trains
would then be needed. The annualized costs at 300 MW are about 0.4 yen/
kWhr ($0.002/kWhr) for I-a and about 0.8 yen/kWhr ($0.004/kWhr) for I-b.
Figure 4-16 shows the costs for SCR followed by FGD (System Il-a).
SCR accounts for about 20 percent of the total investment and 25 percent of
the total annualized costs. The SCR costs are virtually the same as those
shown for System I-a of Figure 4-15. The total investment cost for 300 MW
is about 24,000 yen/kW ($120/kW) and the total annualized cost of 300 MW
is about 1.65 yen/kWhr ($0.008/kWhr)-
The costs for FGD followed by SCR (System Il-b) are shown in Figure
4-17. For a 300 MW plant, the total investment cost is 28,000 yen/kW
($140/kW) and the total annualized cost is 1.9 yen/kWhr ($0.009/kWhr). SCR
accounts for about 35 percent of the investment and nearly 50 percent of
the annualized costs. The Yokkaichi SCR plant, Shindaikyowa (150 MW
equivalent) is said to have cost about 2 billion yen ($10 million) in 1976
including the heat exchanger, while Figure 4-17 shows that a similar plant
costs 1.5 billion yen ($7.5 million). Recent improvements in technology
have resulted in the considerable cost reduction.
The annualized cost of FGD as shown in Figure 4-16 is about 20 percent
higher than that shown in Figure 4-17 because the former includes the cost
for reheating the gas after FGD to 130°C. When the reheating temperature
is 80°C, the cost difference is 6 percent.
196
-------
4 (-
z 3
o
1 2
I
- 2.0
GAS FROM ECONOMIZER (I - a)
GAS FROM AIR PREHEATER (I - b)
ANNUALIZED (I - fa)
ANNUALIZED (I - a)
I
I
1.5
o
o
1.0
8
ee.
o
o
LU
0.5
100
200 300
CAPACITY (MW)
400
500
Figure 4-15. Estimated cost for SCR by Systems I-a and I-b.
197
-------
§
HH
d
1—4
X
m
41V
8
u
J I
-L
2.5
2.0
t.
1.5 «
.0
0.5
TOO
200 300
CAPACITY (MW)
400 500
Figure 4-16. Estimated cost of SCR followed by FGD (System II-a).
198
-------
12
10
g
*f*
i a
I
•s
I
£ 6
CD
= 4
-i3.0
100
200 300
CAPACITY (MW)
2.5
2.0
1.5
1.0
0.5
400 500
Figure 4-17. Estimated cost of FGD followed by SCR (System Il-b),
199
-------
Hitachi Shipbuilding has also made a cost estimate for the treatment
of flue gas from a coal-fired boiler. Roughly speaking, the investment
cost is about 15 percent higher and the annualized cost nearly 20 percent
higher than those for the treatment of gas from an oil-fired boiler with
an equal SOX concentration. The higher cost is due largely to the larger
gas volume (about 20 percent more for coal than for oil for an equal MW) and
the much higher NOx concentration (500-600 ppm for coal).
4.2.5 Evaluation
When Hitachi Shipbuilding constructed large commercial plants in 1975
and 1976, neither SOx-resistant catalysts nor parallel flow type reactors
had been well developed; pretreating the gas with FGD was the only feasible
method for controlling SOx.
Comparison of Figures 4-16 and 4-17 indicates that the total annualized
cost for SOX and NOX removal by System Il-b is about 16 percent higher than
that for removal by System II-a. Lowering the gas reheating temperature
in System Il-a to 80°C will cut the annualized cost by about 10 percent, so
that the difference in cost between System Il-a and System Il-b will be
about 25 percent.
An ammonium bisulfate deposit in the air preheater will be a problem
with System Il-a. In System Il-b, SOs not sufficiently removed by FGD will
be a problem for the heat exchanger. This problem has been experienced at
the Yokkaichi plant, Shindaikyowa. The soot blow and possible water wash
required may add some cost. The problem may not be serious for Systems I-a
and I-b because of the low sulfur content of the fuel and the high gas
temperature in the air preheater and heat exchanger.
The metallic honeycomb developed by Hitachi seems promising because
of the large size which eases packing, the low pressure drop which saves
200
-------
energy, and the high NOX removal efficiency at low NH3/NOX ratio. On the
other hand, the catalyst may have the following drawbacks:
1) Erosion by fly ash may lower the activity since the activated
surface is very thin (about 0.15 mm). A moderate gas velocity
should be used to avoid the erosion.
2) The catalyst is most efficient at 400°C or higher. A heating
device may be required for existing boilers whose economizer
outlet temperature fluctuates between 300 and 400°C.
Further study is desired to produce large SOx-resistant honeycombs
reactive between 300 and 400°C. Such a type of catalyst is under develop-
ment by Hitachi Zosen. The U.S. EPA recently awarded a contract to Hitachi
Zosen for pilot scale (approximately 0.5 MW) study of the metal honeycomb
catalyst. Hitachi Zosen has subcontracted Chemico Air Pollution Corp. to
build and operate the unit which is to be sited at a coal-fired boiler at
Georgia Power Company's Mitchell Station. The tests will provide informa-
tion for further evaluation.
4.3 HITACHI LTD. PROCESS
4.3.1 Introduction
Hitachi Ltd., one of the largest boiler and machine makers in Japan,
has studied SCR processes at more than 10 pilot plants. They can treat
200-4000 Nm3/hr of various gases, including gas from a coal-fired boiler.
Hitachi Ltd. had constructed by April 1, 1978, 8 commercial plants with
capacities larger than 10,000 Nm3/hr, including very large plants for two
700 MW LNG-burning boilers (Tables 3-1 and 4-6). In addition, Hitachi Ltd.
recently constructed 3 prototype plants for oil-fired utility boilers
(Table 4-6).
201
-------
TABLE 4-6. SCR PLANTS BUILT BY HITACHI LTD. (LARGER THAN 10,000 MM3/HR)
Plant Owner
(Site)
Kawatetsu Chem.
(Chiba)
Chiyoda Kenzai
(Kaizuka)
Kansai Electric
(Kainan)
Nisshin Steel
(Amagasaki)
Kansai Paint
(Amagasaki)
Chubu Electric
(Chita)
Nippon Oils & Fats
(Amagasaki)
Company A
Company B
Company C
Gas Source Capacity Gas Pre-
(Fuel) NmVhr (MW) treatment
Coke Oven
(COG)
Boiler
(Heavy Oil)
Boiler
(Heavy Oil)
Boiler
(Heavy Oil)
Boiler
(Kerosene)
Boiler
(LNG)
Boiler
(Heavy Oil)
Boiler
(Kerosene)
Boiler
(Heavy Oil)
Boiler
(Heavy Oil)
500,000
(165)
15,000
(5)
300,000
ft of 450)
20,000
(7)
16,000
(5)
1,910,000x2
(700x2)
20,000
(7)
30,000
(10)
482,600
(175)
466,000
(156)
ESP
Heating
None
None
Heating
Heating
None
None
Heating
SNRb
None
Reactor Catalyst Reaction
Type Shape Temp,°C
Vertical
(Movable)3
Vertical
(Movable)
Horizontal
(Fixed)
Vertical
(Movable)
Vertical
(Movable)
Horizontal
(Fixed)
Vertical
(Movable)
Vertical
(Movable)
Parallel
Flow
Parallel
Flow
Ring 335
Ring 335
Ring 385
Ring 335
Ring 335
Pellet 350
Ring 350
Ring 280
Plate 355
Plate 385
SV NOX
Hr"1 Removal Completion
6200 95 Oct 1976
6000 70 Sept 1976
80 June 1977
6000 90 July 1977
6000 90 Oct 1977
80 Mar 1978
6000 90 Apr 1978
6000 90 Sept 1978
50C July 1978
30 July 1978
.Intermittent moving.
Selective noncatalytic reduction.
CTotal of SNR and SCR.
-------
Hitachi Ltd. has studied many types of catalysts, including the Rashig
ring type and the plate type catalysts being used for large-scale tests
with utility boiler flue gas.
4.3.2 Catalysts and Reactors
A flowsheet of the SCR system for boiler flue gas is shown in Figure
4-18.
Combinations of types of catalyst and reactor selected by Hitachi Ltd.
for different gas compositions are shown in Table 4-7. The Raschig ring
type catalyst not only has a large surface area and is highly effective;
it is less vulnerable to dust plugging because of its large size. Its
diameter is about 15 mm, compared to about 5 mm for most granular catalysts.
The dust deposition ratio (dust deposits on catalyst bed/percent of total)
depends upon the size of the catalyst as shown in Figure 4-19. The ring
is strong enough for use in a moving bed. Test results with the ring
catalyst are shown in Figures 4-20 - 4-22 for NOX removal and NHa leakage
as a function of SV and NHs/NOx ratio.
Hitachi Ltd. has developed a plate catalyst for particulate-rich gas.
A powdered catalyst is cemented on a thin metal plate; total thickness is
about 1 mm. The plates are placed in a basket in parallel and the baskets
are placed in a reactor. Test results with the catalyst are shown in
Figures 4-23 - 4-25 for NOX removal as a function of SV, temperature, and
NHs/NOx ratio.
Figure 4-26 shows an example of the structure of a large reactor with
the parallel plate catalyst. The catalyst plates are placed in baskets,
each measuring about Imxlmxlm, and the baskets are placed in the
reactor in 4 layers. The gas passes through the baskets downwards so that
the dust is readily blown down to keep the catalyst surface clean.
203
-------
S3
o
VAPORIZER
ECONOMIZER
BOILER
REACTOR
AIR
/K
AIR
HEATER
ESP
STACK
70-1332-1
Figure 4-18. Flow diagram of the direct system for boiler flue gas at Hitachi, Ltd.
-------
TABLE 4-7. CATALYSTS AND REACTORS SELECTED BY HITACHI, LTD.
FOR SEVERAL GASES
Gas
Fuel
Reactor
Catalyst
Clean
LNG
Fixed Bed
Pellet
Sphere
Semi-Dirty
Kerosene
Low-S Oil
Fixed Bed
Pellet
Ring
Sphere
Plate
Movable Bed*
Pellet
Sphere
Ring
Dirty
High-S Oil
Movable Bed*
Ring
Fixed Bed
Plate
* Intermittent moving.
205
-------
2 loo
g
§
80
5 60
§ 40
0 W0 24 6 8 10 12 14 16 18 20
DIAMETER OF GRANULAR CATALYST (mm)
Figure 4-19. Catalyst size versus dust deposition for oil-fired dust.
RING TYPE
CATALYST ^
02468
SV (x 1000, hr'1)
Figure 4-20. SV versus NOX removal and NHa leak (ring type catalyst,
15 mm diameter; 350°C; NH3/NO, 1.0; Inlet NOX, 250 ppm).
206
-------
£5.
I
s
on
o*
TOO
80
60
40
20
0.8 0.9 1.0 1.1
NH3/NO¥ MOLE RATIO
/\
60
40
20
0
1.2
§.
CL
a
2
Figure 4-21. Effect of NH3/NOX mole ratio on the performance of a ring
catalyst (15 mm diameter; SV, 6200 hr-1; 350°C; inlet
NOX, 250 ppm).
Figure 4-22.
DC
100
80
60
40
300 320 340 360 380
REACTION TEMPERATURE («C)
400
60
40
20
0
CO
Effect of temperature on the performance of a ring catalyst
(15 mm diameter; 350°Cj SV, 6200 hr^1; NH3/NO, 1.0; inlet
NOX, 250 ppm).
-------
PLATE TYPE CATALYST
o
o
LU
ox
TOO
80
60
40
20
GAS
REACTION TEMPERATURE 350 C
NH3/NOX =1.2
LV 5.9 m/sec.
SPACE VELOCITY SY (hr"1)
CATALYST ELEMENT
10 X103
Figure 4-23. Characteristic curve of space velocity versus
NOX removal efficiency by a plate catalyst.
208
-------
100
80
~ 60
_j
o
S 40
ox
20
SV 2500 hr'1
NH3/NOX 1.2
LV 5.9 m/sec.
300
325
350
375
400
TEMPERATURE (eC)
Figure 4-24. Reaction temperature versus NOX removal
for a plate catalyst.
100
80
~ 60
o
I 40
ox
20
SV = 2,500
-&
SV = 5,000
SV = 10,000
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
NH3/NOX MOLE RATIO
Figure 4-25. NHs/NOx mole ratio versus NOX removal for a plate
catalyst at 350°C, LV = 5.9 m/sec.
209
-------
Figure 4-26. Structure of parallel plate reactor.
210
-------
4.3.3 Pilot Plant Tests
Some of the results of pilot plant tests with gases from gas- and
oil-fired boilers are shown in Figure 4-27. These plants use the ring type
catalyst in a fixed bed. About 97 percent of the NOx was removed from LNG
boiler flue gas at 350°C, with an SV of 10,000 hr"1 and an NH3/NOX mole
ratio of 1. Increasing the SV to 20,000 hr-1 lowered the NOX removal
efficiency to 85 percent.
For flue gas from low sulfur oil, the removal efficiency was about 97
percent at an SV of 10,000 hr"1 , 94 percent at 15,000 hr"1, and 90 percent
at 20,000 hr . An S0x-resistant catalyst was used for the test and showed
virtually no degradation after 7000 hours of use.
For flue gas from high sulfur oil, 90-95 percent NOX removal efficiency
was obtained at an SV of 6200 hr"1 with an NH3/NOX mole ratio of 0.95. The
efficiency increased slightly at an SV of 4500 hr~ and decreased to about
83 percent at an SV of 6200 hr"1 with an NH3/NOX mole ratio of 0.83. The
NHa leakage was kept below 5 ppm by keeping the NH3/NOX mole ratio below
1.0, and thus the deposit of ammonium bisulfate in the air preheater was
reduced and could be removed by soot blowing.
Figure 4-28 shows the results of tests with a moving bed reactor and
the ring type catalyst for flue gas from a coal-fired boiler. The gas
contained about 30 mg/Nm3 particulates after passing through an electro-
static precipitator. The total test period was 6000 hours. The NH3/NOX
mole ratio was kept at 1.1 for over 5000 hours and then changed to 1.0.
The reaction temperature was maintained at 350°C for over 5500 hours and
then changed to between 280-350°C. Pressure drop of the gas through the
reactor was about 70 mmH20 initially. When it reached 90-130 mmH20, 20 to
100 percent of the catalyst was replaced to lower the pressure to 70-90
NOX removal efficiency was kept above 90 percent during the tests
211
-------
LNG boiler
N3
M
N)
100
90
Low-S 011 Boiler
[o-o-o-o
D
100
90
80
(
High-S Oil Boiler
i i
1 1000 2000 3000 4000 5000
i i
6000 7000
OPERATION PERIOD (Hours)
Circled figures show times when SV and HN3/NO mole ratio were changed.
1. SV 10,000-*- 20,000 hr'1 2. SV 10,000-* 15,000 hr'1
3. SV 15,000—>-20,000 hr"1 4. SV 6,200 ——4,500 hr"1
5. SV 4,500—»• 6,200 and the mole ratio 0.95-*. 0.83
Figure 4-27. Test results for ring catalysts used on gas and oil fired boilers.
-------
NH3/NOX Hole Ratio 1.1:1.0
| 1.0:1.0 |
350 C
280<-350°C—. 350°C
N3
l->
U>
X ?
o o
to O3C
a:
ui a E
a: E
100
90
80
201
100
0
20% Moving
Stop and Start Test
100% Moving
1000
2000 3000
OPERATION HOURS
4000
5000
6000
Figure 4-28. Tests of a moving bed reactor with a 2000 Nm3/hr capacity for coal-fired
boiler flue gas after a hot electrostatic precipitator.
-------
except when the temperature was lowered to 280*C. Efficiency dropped to
85 percent at the lowered temperature.
4.3.4 Kainan Plant. Kansai Electric
The plant treats 300,000 Nms/hr of flue gas which is one-fourth of the
total gas from an existing 450 MW oil-fired utility boiler burning oil
containing 0.2 percent sulfur. The gas leaving the boiler economizer con-
tains about 100 ppm of SOa, 2-3 ppm of SCh, 100 ppm of NOX and 20 mg/Nm3 of
particulates at 330-380°C. Ammonia diluted with air is injected in the flue
gas and sent to a reactor with two compartments each with a capacity of
treating 150,000 Nm3/hr of the gas (Figure 4-29). For test, two different
kinds of Raschig ring type catalysts, a high quality one and a low cost
one, both with a size of about 15 mm, have been used. One of the compart-
ments contains the high quality catalyst and the other contains the low
cost catalyst. The bed depth is about 30 cm and the SV about 5700. The
gas temperature fluctuates from 380°C at full load (450 MW) to 330°C at
minimum load (150 MW).
An NHs/NOx mole ratio of 1.07 has been used to achieve an overall
removal of 80 percent of NOX with about 10 ppm of leak ammonia. It is
estimated that over 90 percent NOx removal efficiency is obtained at that
mole ratio if the high quality catalyst is used in both beds.
The pressure drop of the SCR unit which has 40 mmH20 initially went
up to 165 mmH20 after 11 months' continuous operation, when the boiler was
subjected to scheduled shutdown for annual maintenance. The increased
pressure drop did not upset operations because the booster fan (induced
draft fan) installed for the SCR unit has a capacity of 200 mrnHaO. During
the boiler shutdown period of one month, the catalyst was screened to remove
particulates and then replaced in the beds.
214
-------
ESP
10
t-*
Ul
ECONOMIZER
I BOILER I
NH,?
AIR
FER
|fc
*-
i
J
*j
i
REACTOR
STACK
70-1333-1
Figure 4-29. Flowsheet of the Kainan SCR plant.
-------
The SCR unit was connected to the boiler in 1977 during a two-month
scheduled shutdown period for annual maintenance. Oil-fired boilers are
normally shut down annually—for two months every other year and one month
in between. Since the modification was made in a shutdown period, the
boiler operation was not hindered.
4.3.5 Chita Plant. Chubu Electric
Chubu Electric installed full-scale SCR plants for two new LNG-fired
boilers, each with a capacity of 700 MW. A flowsheet of the plant is shown
in Figure 4-30. Flue gas (1,900,000 Nm3/hr) from the economizer of the
700 MW boiler is injected with air-diluted ammonia and is sent to 2 parallel
reactors. Each reactor, measuring 10.5 x 11.5 x 12.0 m, has four 11 cm-deep
layers of pellet catalyst. The flue gas passes one of the layers. The SV
is 20,000 hr-1. The gas temperature ranges from 350°C at 700 MW load and
290°C at 250 MW load.
The inlet NOX concentration is only about 35 ppm because of good
combustion control. By adding 0.8 mole of NHa for each mole of NOX, 80
percent of the NOX .is removed. NHs leakage is 5 ppm or below. The plants
have been operated without trouble since start-up in April 1978.
4.3.7 Economics
The investment costs, including the initial charge for the catalyst,
was 1.7 billion yen ($85 million) in 1978 for the two SCR plants (total
1400 MW) at Chita station and 1.2 billion yen ($60 million) in 1977 for the
Kainan plant. The large difference in the cost per kilowatt—1210 yen
($6.05) at Chita and 10,670 yen ($53.35) at Kainan—is due to the following:
1) Chita has clean gas (SV 20,000 hr"1) while Kainan has
dirty gas.
216
-------
Figure 4-30. Flowsheet of the Chita SCR plant.
217
-------
2) Chita has a new boiler while Kainan has ah existing boiler.
3) Chita is much larger.
4) Kainan plant has various equipment for the operation of
the test equipment.
It is roughly estimated that the investment cost for a full-scale SCR
plant for an existing 450 MW oil-fired boiler is about 3 billion yen (6700
yen/kW or $34/kW). The cost should be considerably lower for a new boiler.
The catalyst life is estimated at 2 years for dirty gas and longer for clean
gas.
Operation of SCR plants is simple and is controlled from inside the
boiler control room.
4.3.8 Evaluation
Hitachi's ring type catalyst is not only highly reactive and SOX-
resistant but is also less vulnerable to dust plugging and strong enough to
withstand use in a moving bed. The smooth operation of the commercial
and prototype plants shows the reliability of the Hitachi process.
The plate catalyst may be most suitable for preventing particulate
plugging. Compared with metallic honeycombs produced by chemical treatment
of the surface of a metallic alloy, larger varieties of compositions of
plate catalysts may be produced by cementing catalyst powder on metal plate.
Although 90 percent NOx removal is achieved with the plates, they may be
better suited for moderate NOX removal efficiency, e.g. , 70-80 percent
for new boilers. For existing boilers, 30-50 percent removal may be achieved
simply by placing a small amount of the catalyst plates in the duct between
the boiler economizer and the air preheater, without installing a reactor
or fan and by using less than 0.7 mole NHs for each mole of NOX to minimize
218
-------
the NHa leakage. A combination of combustion modification and the simple
type SCR with an NOX removal efficiency of 50-70 percent may be one of the
best ways for NOx abatement for existing boilers.
Hitachi's pilot plant test with flue gas from a coal-fired boiler has
shown that by reducing the particulates content to about 200 mg/Nm3 with a
hot electrostatic precipitator, a moving bed reactor with the ring catalyst
works well. Further improvement of plate catalysts, however, may make them
advantageous over a moving bed reactor.
Hitachi is going to construct a commercial SCR plant with plate catalyst
for a new coal-fired boiler of Hokkaido Electric (Table 1-14), and plans to
use a hot electrostatic precipitator because it can efficiently remove the
fly ash from the domestic low sulfur coal to be used for the boiler.
When a hot electrostatic precipitator is used, the prevention of
ammonium bisulfate deposits in the air preheater is important. Experience
indicates that the sweeping effect of a large amount of fly ash in coal-
fired flue gas helps to reduce the deposits.
4.4 JGC PROCESS (PARANOX PROCESS)
4.4.1 Introduction
JGC Corporation (former Japan Gasoline Co.), one of Japan's largest
engineering and plant construction companies, constructed a Shell process
flue gas desulfurization unit at Yokkaichi for Showa Yokkaichi Seikiyu
(SYS). They used the parallel passage reactor developed by Shell. JGC
has produced SCR catalysts and applied them to the reactor for SCR, and
constructed two relatively small commercial SCR units for Kashima Oil and
Fuji Oil to treat flue gas from heavy-oil burning gas. In addition, JGC
has tested in a pilot plant the use of a newly developed low-temperature
catalyst for the treatment of flue gas from an iron-ore sintering machine.
219
-------
Recently JGC constructed a large test plant for Nippon Steel to treat flue
gas from a coke oven.
4.4.2 Parallel Passage Reactor
A parallel passage reactor, developed to treat dust-rich gas without
particulates clogging, has the following structure (Figure 4-31):
1) A thin rectangular box made of metallic gauze is filled
with the catalyst. This box is called an "envelope" and
forms the catalyst layer.
2) These envelopes are placed in parallel at narrow intervals.
The flue gas flows through the interval space along the surface
of the envelopes. This interval space is called the "gas
channel."
3) For ease of handling, a set of envelopes is placed in a large
square box called a "unit cell." The top and bottom are open
so that the flue gas can flow up and down through the box.
4) The unit cells are stacked in a reactor shell to make a
Parallel Passage Reactor. Flue gas can go through the
reactor either upward or downward.
Since the gas passes through the gas channel parallel to the catalyst
layers, dust plugging is avoided, except that a minor deposit of dust on
the gauze can occur. JGC has developed a "sand blast" system to remove
the deposit.
220
-------
REACTOR
FLUE GAS
/// L\\
\
YjVlV
x'Yjy'
N/ V/ \
\~XLX
1 *\\ /p /
X
X
» /
/.
/ \
V
/ V
X
X
\ /
t ^
^ \
\ /
A
\ ^x /
Xv Xv
r \U \
•r
V
ENVELOPE
FLUE GAS
CATALYST LAYER
DISTRIBUTOR
GAS CHANNEL
UNIT CELL
Figure 4-31. Construction of a parallel passage reactor,
221
-------
4.4.3 Catalyst
The catalyst Is a small cylinder, 1 torn In diameter and 3-5 mm in
length. Great mechanical strength is not needed because the catalyst is
packed in an envelope. Therefore, the catalyst can be porous to provide
high reactivity.
Figures 4-32 - 4-34 show the NOX removal efficiencies in laboratory
tests with a high-temperature catalyst JP 102 and a low-temperature catalyst
JP 501. For the tests, the catalysts were packed in an ordinary packed
bed (not parallel passage) through which gas was passed. Since the
catalysts are small and reactive, high efficiency was obtained at a large
SV of 20,000. The catalyst JP 102 is active above 350°C and is suitable
for boiler flue gas, while JP 501 is active even at 250°C and suits low
temperature gases.
Figure 4-34 indicates that the activity of the low-temperature catalyst
is stable at 300°C but is lowered at 250°C and 230°C. The decrease in
activity is caused by the deposit of ammonium bisulfate on the catalyst
below 270°C. By heating the contaminated catalyst above 350°C, ammonium
bisulfate is removed by decomposition and catalyst activity is restored.
4.4.4 Commercial Plants2
Design and operation data of two commercial plants are listed in
Table 4-8.
The first commercial plant at Kashima Refinery (Figure 4-35) was
completed in November 1975 and can treat 50,000 Nm3/hr of flue gas from a
heating furnace burning heavy oil containing 90-130 ppm of NOX, 310-680
ppm of SOx, and 20-40 mg/Nms of particulates. Flue gas at 215-225°C is
heated by an inline heater to 380-395°C, injected with ammonia at an
NHs/NOx mole ratio of 1.0-1.2, and introduced into a parallel passage
222
-------
100
£- 90
80
70
200
250 300 350
TEMPERATURE (°C)
400
450
Figure 4-32. The effect of temperature on the activities of catalysts
JP 501 and JP 102 (NO, 200 ppm; NHs, 240 ppm; 02, 3
percent; HzO, 10 percent; SV, 20,000 hr-1).
TOO
75
50
JP 102
1.0
NH3/NOX MOLE RATIO
2.0
Figure 4-33.
The effect of NHs/NO mole ratio on the activities of
catalysts JP 501 and JP 102 (380°C; SV, 20,000 hr"1;
NO, 200 ppm; SOa, 1500 ppm; 02, 3 percent; Nz, 10
percent).
223
-------
NO:
NH
3:
200 ppm
220 ppm
02: 3%
10%
100
S02: 220 ppm ^: Balance
Temp. 300 "C
50
NJ
N5
UJ
CXL
X
O
250 °C
25
0
50
100
PERIOD (Hours)
150
200
Figure 4-34.
The change in NOX .removal efficiency with reaction temperature
and time for catalyst JP 501.
-------
TABLE 4-8. DESIGN AND OPERATION DATA OF COMMERCIAL SCR PLANTS (JGC PROCESS)
N>
Ln
Kashima Plant, Kashima Oil
(Heating Furnace Flue Gas)
Flue Gas
Volume (Nm3/hr)
Temperature (°C)
NO (ppm)
SOx (ppm)
Particulates (mg/Nm3)
Oa (percent)
HzO (percent)
Reaction Temperature
NH3/NOX Mole Ratio
Space Velocity (hr l )
NOX Removal Efficiency (percent)
Pressure Drop in Reactor (rnrnHaO)
Design
50,000
220-250
60-250
370-850
Below 50
5-11
11-16
380-420
1.2
4000
Above 90
200
Operation
49,000-51,000
215-225
90-130
310-680
20-40
7-11
380-395
1.0-1.2
4000
95-98
140-180
Chiba Plant, Fuji Oil
(FCC CO Boiler Flue Gas)
Design
70,000
380-420
180
630
100
3
10
380-420
1.2
4000
90
160
Operation
78,000-81,000
365-410
160-220
340-660
60-70
2
13
385-405
1.0-1.2
4000
93-96
115-125
-------
STACK
AIR
PREHEATER
FLUE GAS2:
REACTOR
HEATER
FAN
FUEL
GAS
I
AIR
M. P. STEAM f-
EVAPORATOR
Figure 4-35. Paranox process for Kashima Oil, 50,000 Nm3/hr.
226
-------
reactor to remove 95-98 percent of NOX at an SV or 4000 hr 1. The ammonia
concentration of the reactor effluent is about 10 ppm.
The plant was operated continuously without problems for about 16
months but has been inoperative most of the time since then. Because of
the low operating load, the total emission of NOx of the refinery has
decreased so that denitrification has become unnecessary.
Another commercial plant was constructed for Fuji Oil at Chiba
(Sodegaura) Refinery (Figure 4-36) and began operation in July 1976. The
flue gas is discharged from a boiler economizer at 360-410°C and has the
composition shown in Table 4-8. A gas-fired heater with a capacity of
3.6 x 106 kcal/hr has been installed but is used only for limited periods
because the gas temperature is close to the desired range of 380-420°C.
Ammonia is diluted with steam and injected into the gas. The unit has
an open-bypass system with a blower with a capacity of 80,000 Nm3/hr.
Usually, a portion of the gas leaving the reactor is circulated through
the bypass and introduced into the reactor together with the gas from the
boiler. The total gas volume and the pressure drop in the reactor are kept
almost constant so boiler operation is not adversely affected regardless of
the fluctuation of the gas volume from the boiler. By using 1.0-1.2 moles
of NHa for each mole of NOX, 93-96 percent of the NOX is removed with leak
NHs (NHa in the reactor outlet gas) ranging from 50 to 30 ppm according to
the mole ratio.
The unit was operated without serious problem for nearly 2 years. A
small amount of particulates tends to adhere to the metal gauze of the
elements and the unit has a sand blast system to remove the particulates.
The sand blast system, however, has seldom been used. The catalyst itself
has been kept relatively clean and the life is estimated at over 2 years.
The floor space occupied by the SCR unit is about 250 m2. The plant is not
operated currently for the same reasons as with the above Kashima plant.
227
-------
AIR
PREHEATER
FLUE GAS £
STACK
REACTOR
HEATER
7 V
FAN
n
FUEL AIR
GAS
M. P. STEAM £
AMMONIA GASTr
Figure 4-36. Paranox process for Fuji Oil, 70,000 Nm3/hr.
228
-------
4.4.5. Pilot Plant Tests with Sintering Machine Flue Gas3
JGC has conducted pilot plant tests at the Nagoya Works, Nippon Steel,
sponsored by the NOX Fund of the Steel Federation. The pilot plant can
treat 2000 Nm3/hr of flue gas from an iron-ore sintering machine with the
composition shown in Table 4-9. The gas contains less than 250 mg/Nm3
participates before an electrostatic precipitator (ESP) and less than 50
mg/Nm3 after the ESP. The composition and size distribution of the particu-
lates are shown in Table 4-10.
TABLE 4-9. COMPOSITION OF THE GAS FROM THE IRON-ORE SINTERING
MACHINE AT NAGOYA WORKS, NIPPON STEEL
NOX
02
H20
150-250 ppm
15-17 percent
10-12 percent
SOX
C02
Particulates
Below 200 ppm
5-6 percent
Below 250 mg/Nm3
TABLE 4-10. COMPOSITION AND SIZE OF THE PARTICULATES IN THE GAS FROM THE
IRON-ORE SINTERING MACHINE AT NAGOYA WORKS, NIPPON STEEL
Fe (as FeaOs) 77-86 percent C About 5 percent
Si02 4-7 percent AlaOs 1-3 percent
CaO 4-8 percent
Minor Components - Mg, Mn, Zn, Pb, K, Na, etc.
Size (y)
Below 10
10-60
60-100
Above 100
Distribution
Before ESP
1
4
15
80
(percent)
After ESP
6
9
40
45
229
-------
High and low temperature catalysts are used for both the gases before
and after the ESP as shown in Table 4-11.
TABLE 4-11. CONDITIONS OF JGC TESTING AT NAGOYA WORKS, NIPPON STEEL
Test
No.
1
2
3
4
Test
Hours
1500
700
1200
6700
Gas
Obtained
After ESP
Before ESP
After ESP
Before ESP
Catalyst
Type
High Temp.
High Temp.
Low Temp.
Low Temp.
Reaction
Temp. (°C)
390
390
300
290-300
SV
(hr-1)
3150
3150
3620
3620
NH3 /NOX
Mole
Ratio
1.2-1.5
1.2-1.5
1.2-1.5
1.2-1.5
Since the flue gas was at about 160°C, the gas was heated to the
desired temperature by a heat exchanger and a heater. The results of tests
by No. 1 - No. 3 are shown in Figures 4-37 - 4-39.
During tests No. 1 and No. 2, there were short interruptions due to
the shutdown of the sintering machine. Tests No. 3 and No. 4 were carried
out continuously for 1200 and 6700 hours, respectively. In tests No. 1 and
No. 2, about 95 percent of the NOX was removed at a pressure drop of 100-
150 mraHaO. The pressure drop did not increase in test No. 1, but it did
increase about 10 mmHzO about 700 hours into test No. 2. In test No. 3,
NOX removal efficiency decreased from 98 to 96 percent and the pressure
drop increased from 105-110 to 120-130 mmH20 after 1200 hours.
In test No. 4, 95-96 percent of the NOX was removed at 300°C, while at
290°C the initial NOX removal efficiency of 93 percent decreased to 90
percent after 1200 hours. By heating up the flue gas to 380°C for two
hours, the NOX removal efficiency recovered to 93 percent. The pressure
drop increased from 130 to 180 mmHaO after 5300 hours and stayed at 180
mmHzO. The plant operation was trouble-free.
230
-------
10
u>
2500,
2000
Soj
= I500-
Ef=^ 1000
•—•
450
a-350
S
>-3eo
ZOO
tg~
100
«S? 90-
80
200
400
600
800
1000
1200
1400
1600
end
200
400
600
800
1000
1200
1400
1600
Start Nay 7
Hours
(Note) V shows the test-stop period caused by
operation-stop of sintering furnace etc.
(HR)
1st RUN
Figure 4-37. Results of JGC testing of a high temperature catalyst at Nagoya Workss
Nippon Steel, for flue gas after an electrostatic precipitator.
-------
CO
2500
2000-
1500-
1000
450
400-
350-
300
200
150-
100 -
50 -
0
100
90-
80
100
200
300
400
500
600
700
800
END
100 200
2nd RUN
300
400
500
600
700
800 (HR)
Figure 4-38.
Results of JGC testing of a high temperature catalyst at Nagoya Works,
Nippon Steel, for flue gas before an electrostatic precipitator.
-------
Ni
U)
00
~ 150H
o
CM
g 100-
°- 50-
100
90
I
400
600
800
1000
1200
200
400 600
TEST PERIOD (HR)
800
1000
1200
Figure 4-39. Results of the third run of JGC testing of a low temperature catalyst at Nagoya Works,
Nippon Steel, for flue gas after an electrostatic precipitator.
-------
A considerable amount of dust from the sintering machine and some
ammonium bisulfate deposited in the heat exchanger but they could be
removed by soot blowing.
4.4.6 Prototype Plant for Coke Oven Flue Gas
JGC constructed at the Kimitsu Works of Nippon Steel a prototype plant
which can treat 150,000 Nm3/hr of flue gas from a coke oven containing about
200 ppm of NOX and 50 ppm of SOX. The gas at about 200°C is treated by a
low-temperature catalyst in a parallel passage reactor at an SV of 4000.
The NOX removal efficiency is about 95 percent initially and decreases to
about 90 percent in about a week due to the deposit of ammonium bisulfate
on the catalyst. The flue gas is then heated above 350°C by a heater to
decompose and remove the ammonium bisulfate on the catalyst and to regain
the high NOx removal efficiency. Details of the plant operation have not
been disclosed yet.
4.4.7 Economics
JGC claims that the investment cost for the SCR plant for flue gas
from a 33-150 MW oil-fired boiler is 7500-15,000 yen/kW ($37.50-$75/kW)
excluding site preparation and the control room. Actually, the Chiba plant,
Fuji Oil (23 MW equivalent), cost 490 million yen ($2.45 million) including
the 40 million yen ($200,000) initial charge for the catalyst. The catalyst
life is estimated at 2 years. Power requirement for the fan of the plant
is 250 kW. Operating cost is reported to be 8200 yen/hr ($41/hr) including
ammonia and electrical power.
4.4.8 Evaluat ion
The catalysts developed by JGC are S0x-resistant and highly efficient.
The commercial and test plants constructed by JGC for SCR have had little
problem achieving over 90 percent NOX removal. JGC appears to have
234
-------
established know-how for SCR application of the parallel passage reactor
originally developed by Shell. The process is one of the most reliable
processes commercially proven for dirty gas treatment and may be applicable
also for flue gas from a coal-fired boiler without prior fly ash removal.
On the other hand, the parallel passage reactor has the following
drawbacks: 1) relatively large pressure drop, 2) relatively small SV, and
3) relatively high cost.
Moving bed reactors have also been used for treating flue gas from
oil-fired boilers. Tubular and honeycomb type catalysts are beginning to
be used for boiler flue gas and gas with higher particulate content.
Further comparison of the operation of the parallel passage reactor with
the other types of catalysts should be made in the future.
The low-temperature catalyst developed by JGC may be more economical
for low-temperature gases than the conventional SCR catalysts which are
active at 300-400°C, although intermittent gas heating to 350°C is needed
to remove ammonium bisulfate.
4.5 MITSUI ENGINEERING AND SHIPBUILDING PROCESS
4.5.1 Introduction
Mitsui Engineering and Shipbuilding Co. (formerly Mitsui Shipbuilding
Co.), one of the largest engineering and shipbuilding companies in Japan,
has studied SCR in cooperation with Mitsui Petrochemical Co., and con-
structed a commercial plant for both Mitsui Petrochemical and Ukishima
Petrochemical. Both use a fixed bed reactor.
The first plant, built for Mitsui Petrochemical in Chiba, can treat
220,000 Nm /hr of flue gas from a boiler burning a sulfur-free oil by
using an alumina-based catalyst placed in a special type of parallel flow
235
-------
reactor. The second plant, for Ukishima Petrochemical in Chiba, uses a
titania-based tubular catalyst and can treat 220,000 Nm3/hr of flue gas
from boilers burning heavy oil.
Mitsui has made extensive tests on tubular type catalysts at a pilot
plant at Tamashima Power Station, Chugoku Electric.
4.5.2 Chiba Plant, Mitsui Petrochemical
A flowsheet of the Chiba Plant of Mitsui Petrochemical is shown in
Figure 4-40. The plant was installed for an existing boiler which dis-
charges flue gas through an air preheater and an electrostatic precipitator
at 150°C. The gas is heated by a gas-gas heat exchanger and an oil-fired
heater to 350°C. The operation parameters are shown in Table 4-12. Over
90 percent of the NOX in the gas is removed at an NHs/NOX mole ratio of 1
and an SV of 2500 hr"1. The gas contains virtually no SOX with a small
amount of particulates. Plant operation was trouble-free, but is currently
stopped because of the high cost of fuel used for gas heating.
4.5.3 Chiba Plant, Ukishima Petrochemical
The Chiba Plant of Ukishima Petrochemical began operation in July 1977.
It can treat 220,000 Nm3/hr of flue gas from a new industrial boiler burning
a 0.6 percent sulfur oil. A flowsheet is shown in Figure 4-41. Operation
parameters are shown in Table 4-12. The plant has no bypass. The flue gas
from the boiler contains 150 ppm NOX, about 300 ppm SOX, and 100-150 mg/Nm3
particulates. This gas at 350-400°C is injected with ammonia and is
introduced into a reactor packed with a tubular catalyst. The catalyst,
based on Ti02, is resistant to SOz and SOs and has a 35 mm outer diameter,
20 mm inner diameter, and 1000 mm length. The catalyst is packed in a
Imxlmxlm cubic container as shown in Figure 4-42. The reactor has
3 layers of containers; total packed height is 3 meters.
236
-------
to
u>
PREHEATER
AIR
I
GAS-GAS
HEAT EXCHANGER
[BOILER
|
1^)
1
7 X"
STACK
Figure 4-40. Flowsheet of Chiba Plant, Mitsui Petrochemical.
-------
TABLE 4-12. OPERATING PARAMETERS OF MITSUI ENGINEERING
& SHIPBUILDING CO. SCR PLANTS
Mitsui
Petrochemical Co.
Ukishima
Petrochemical Co.
3
Capacity (Nm3/hr)
Gas Composition
NOx (ppm)
SOv \PpttO
Particulates (mg/Nm3)
Catalyst and Reactor
Catalyst Carrier
Catalyst Shape
SV (hr-1)
Temperature (°C)
NHs/NOx Mole Ratio
NOX Removal (percent)
Total Pressure Drop (imnHzO)
Leak NHs (ppm)
Operation Start
Plant Cost (10s yen)
Denitrification Cost
(yen/kWhr)*
200,000
190
None
20-50
A1203
Granule
2600
350
1.0
90
October 1975
220,000
150
300
100-150
Ti02
Tube
4000
350-400
1.0
90
130
Below 10
July 1977
260
0.33
* Including 7 years depreciation.
238
-------
to
CO
VO
STACK
Figure 4-41. Flowsheet of Chiba Plant, Ukishima Petrochemical.
-------
N>
-P-
O
PACKAGE
.
4
-X
Figure 4-42. Structure of a reactor with a tubular catalyst.
-------
The superficial gas space velocity in the reactor is 7-8 meters with an
actual gas velocity of 10-12 meters through the catalyst. About 90 percent
of the NOx can be removed by using 1 mole of ammonia for each mole of NOX.
The ammonia leakage is kept below 10 ppm. The pressure drop through the
system is 130 mmHaO, including 100 mm for the reactor and 30 mm for the ducts.
To cope with the drop of the gas temperature at reduced load, a system
has been installed to bypass a portion of the hot gas before passing
through the economizer.
The boiler supports an ethylene plant and has been frequently shut down
due to the oversupply of ethylene. The SCR plant operation has been
trouble-free except that ammonium bisulfate tends to deposit in the air
preheater (Ljungstrom type). To cope with the problem, NHs concentration
at the reactor outlet is maintained below 5 ppm by using a little less than 1
mole of ammonia for each mole of NOx, a heating element with a simple element
is used, and steam blowing has been applied. Two steam nozzles, each with a
capacity of 1 ton/hr of steam of over 10 kg/cm2 pressure, have been used
intermittently. The steam consumption is equivalent to 0.3 percent of the
steam generated by the boiler.
4.5.4 Tests on Tubular Catalysts
Mitsui Engineering has extensively tested tubular catalysts at a
pilot plant.
One of the most important findings was that a much higher NOX removal
ratio and a smaller pressure drop were obtained by separate packing of the
catalyst (Figure 4-43, B) than by contact packing (Figure 4-43, A) as
shown in Figure 4-44. With contact packing, most of the gas passes inside
the tubes where catalyst surface area is much smaller than outside the
tubes. Tests have shown that best results (optimum efficiency) are obtained
when the tubes are arranged in a separate position so that the ratio of gas
241
-------
(A) CONTACT PACKING
(a) Diamond
(b) Square
(B) SEPARATE PACKING
(a) Diamond
(b) Square
o
o
Figure 4-43. Methods for packing of tubular catalysts.
242
-------
100
90
80
70
NOX REMOVAL
(Separate Packing)
PRESSURE DROP
Separate Packing)
300
o
CM
200
100
1000
2000
3000 4000
SV (hr'1)
5000
6000
Figure 4-44.
Effect of catalyst packing method on NOX removal and pressure
drop (fixed gas velocity, varying catalyst volume).7
243
-------
flow inside and outside the tubes is equal to the ratio of the inside and
outside surface areas. The diamond and square packings (see Figure 4-43, B)
perform comparably at the same gas flow ratios.
The practical best dimensions for the tube were found to be 20 mm
outside diameter and 14 mm inner diameter (Figure 4-45). The tube has a
wall thickness of 3 mm. By using a tube of this size, 97 percent of the
NOX was removed at 350°C with an NHs/NOx mole ratio of 1.0. Ammonia leakage
was below 5 ppm (Figure 4-46). A thinner wall reduces the catalyst weight
but makes the tube too weak for the 1000 mm length.
4.5.5 Economics
The Ukishima Petrochemical plant cost 260 million yen ($1.3 million)
in 1975. Table 4-13 shows Mitsui Engineering's 1977 estimate of plant cost
and annualized cost, including 7 years depreciation for 5 different
capacities ranging from 13 to 330 MW and 3 different configurations, A, B
and C.
Configuration A is similar to that of the Ukishima Petrochemical plant
and the cost includes that for a SCR reactor, ammonia storage, ammonia
injection and mixing devices, and equipment for measurement and control. In
addition, Configuration B includes a fan. Configuration C is for existing
boilers and is similar to that of the Mitsui Petrochemical plant and further-
more includes a heater and a heat exchanger. The table also shows the space
requirement for each plant.
The cost for B is 10-30 percent more than that for A while that for C
is very high. For example, the plant cost for a 330 MW boiler (1,000,000
Nm3/hr) is 3270, 4360, and 7910 yen/kW ($16.35, $21.80 and $39.55/kW) for
A, B, and C, respectively; the annualized cost is 1300, 1770, and 3480 yen/
kl ($6.50, $8.85 and $17.40/kl) oil or 0.30, 0.41, and 0.81 yen/kWhr
($0.0015, $0.0020, and $0.0040/kWhr) for A, B, and C, respectively. The
244
-------
100
95-
90-
85-
3000
4000
5000
6000
Figure 4-45.
SV versus NOx removal for two diameters of tubular
catalyst (350°C, separate packing).
245
-------
TOO
90-
80-
i
70-
Leak
NH3
a 20-
0.7 0.8 0.9 1.0 1.1 1.2 1.3
NH3/NOX MOLE RATIO
10-
Figure 4-46. Performance of the tubular catalyst with 20 mm diameter
(14 mm inner diameter) at 350°C and SV of 4000 hr-1.
246
-------
TABLE 4-13. ESTIMATED SCR COST OF MITSUI SHIPBUILDING PROCESS WITH TUBULAR CATALYST (1977)
Investment Cost
(106 yen. $5000)
Annualized SCR Cost
(yen/kl, $0.005/kl)
Space Needed
(m2)
103
Nm3/hr
40
100
200
500
1000
MW
13
33
66
178
330
A
153
228
335
596
1080
B
168
264
407
800
1440
C
288
489
792
1520
2610
A
2690
1910
1620
1350
1300
B
3130
2370
2080
1850
1770
C
5790
4700
4140
3690
3480
A
140
180
300
540
900
B
160
240
420
740
1200
C
380
680
970
1900
3000
to
STACK
HEATER
-------
high annualized cost for C is due largely to the fuel requirement for gas
heating. The space requirement for C is more than triple that for A because
of the need for a heat exchanger and heater.
4.5.6 Evaluation
The parallel flow type reactor at the Mitsui Petrochemical Plant (No.
1) was operated smoothly, but has a low space velocity because of the special
type of reactor.
The process at the Ukishima Petrochemical plant, which uses a tubular
catalyst, has the following advantages: 1) The catalyst has no dust
plugging problems when used for flue gas from an oil-fired boiler. 2)
Optimum separate packing gives a high NOX removal efficiency. Although
the SV value is not very large, the amount of catalyst required is rela-
tively small. The SV may be much larger when a tube with 20 mm diameter
is used. 3) The plant cost is low. 4) The catalyst is based on titanium
and may be resistant to gas from a high sulfur fuel.
On the other hand, a high gas velocity in the reactor is needed with
the tubular catalyst. This high velocity causes a fairly large pressure
drop as in other parallel flow type reactors. The catalyst may also be
useful for flue gas from a coal-fired boiler with a full dust load. For
particulate-rich gas however, considerable erosion might occur when the
superficial gas velocity exceeds 10 m/sec.
The optimum tubular catalyst, 20 and 14 mm in outer and inner diam-
eters, has a thickness of 3 mm. It is desirable to produce a catalyst with
a thinner wall which has less weight and yet has a high efficiency and
strength. For this purpose, a honeycomb type with a fine structure may be
preferred. Tubular catalyst may be suitable for dust-rich gas by which a
honeycomb type catalyst with a fine structure may be readily plugged up.
248
-------
4.6 MITSUBISHI HEAVY INDUSTRIES PROCESS (MHI PROCESS)
4.6.1 Introduction
Mitsubishi Heavy Industries, one of the largest machine producers and
plant constructors, has extensively tested SCR catalysts and operated many
types of pilot plants for treating various gases, including coal-fired
boiler flue gas. For clean gas treatment, MHI constructed two relatively
small commercial plants for Osaka Gas and two large ones for Kyushu
Electric, one of which went into operation recently (Table 3-1). Those
plants use a granular catalyst in a fixed bed. For dirty gas treatment,
MHI constructed a plant for Fuji Oil with a honeycomb catalyst and is going
to construct a large plant (700 MW) for Chubu Electric. In addition, MHI
conducted pilot plant tests with a plate type catalyst. Two large plants
with a plate type catalyst began operation recently on utility boilers.
Details of some of the pilot plant tests have been reported.^ No
details of the commercial plants for clean gas from LNG burning have been
disclosed, but the clean gas treatment is easy and may not pose any problem.
The following sections will describe the Fuji Oil plant which uses a honey-
comb catalyst and outline other plants for the treatment of dirty gas.
4.6.2 Chiba (Sodegaura) Plant, Fuji Oil
The SCR plant at the Chiba Refinery, Fugi Oil, can treat 193,000 Nm3/hr
of flue gas from an industrial boiler which burns 0.4 percent sulfur oil
containing about 200 ppm S02, nearly 120 ppm NOX, and 15 mg/Nm3 particu-
lates. The plant, constructed by MHI, uses a honeycomb catalyst. The
catalyst unit is cubic and 15 x 15 x 15 cm in size. The catalyst is placed
in a duct after the boiler, so that no separate reactor is needed (Figure
4-47).
249
-------
AIRt *•
CATALYST
REACTOR
-tXJ
ESP
HXK
STACK
Figure 4-47. Flowsheet of the SCR plant built by MHI
at the Chiba Plant, Fuji Oil.
As shown in Figure 4-48, 75 units are packed in package A and 50 units
in package B. Thirty-six packages (4 x 9) are placed in a layer and the
catalyst bed consists of 3 layers of packages A and a layer of packages B.
Thus, there are 9900 units in total with a total bed depth of 1.65 m.
The gas at 350-400°C is passed through the catalyst bed at a super-
ficial velocity of about 9 m/sec and an SV of 6400 hr
-i
Pressure drop of
the gas through the bed is 120 mmfoO at full load and 50 mm at 65 percent
load. By using 0.75-0.8 mole of NHs for each mole of NOX, 75-80 percent of
the NOX is removed with leak NHs (NHs in the reactor effluent) less than
1 ppm. At 0.9 mole of NHs, the removal efficiency reaches 90 percent but
the NHs in the reactor effluent increases to 2-3 ppm. The plant is normally
operated at the low NHs ratio of 0.75-0.8 to depress the NHs emission. This
prevents an ammonium bisulfate deposit in the heat recovery system. (The
250
-------
K)
(_n
I-1
75 cm
X
10 cm
PACKAGE B
-75
X XX
75 W////;
x X XXyy
X" x X r^ X X /
f '
45 cm
PACKAGE A
7
X XX XXX
-TI- X X X X X
75 cm X Xyyy
x / xx x
X
GAS
1
GAS
Figure 4-48. Packing of honeycomb catalyst at the Chiba Plant, Fuji Oil.
-------
plant has an economizer for heat recovery replacing the usual air
preheater.)
The plant was commissioned in January 1978, was run for about a month
until the boiler was shut down, then resumed operation in March and has
since been operated trouble-free.
The catalyst life is estimated at 2 years. Since neither a separate
reactor nor a separate fan is needed, the plant cost is mainly for the
catalyst. The very low NHs emission which gives virtually no ammonium
bisulfate problem as well as the low investment and operation costs, are
the advantages of the process.
4.6.3 Parallel Plate Catalyst
MHI has extensively tested a parallel plate catalyst at 5 pilot plants
as shown in Figure 4-49. The catalyst plates are produced from either
ceramic (castable) or metallic material. The metallic catalyst has a
thickness of about 1 mm, and the ceramic one about 8 mm.
The metallic plate catalyst is made by the following two ways: 1) a
special alloy plate is chemically treated to dissolve some of the components
and make the plate porous and reactive, or 2) powdered catalyst is cemented
onto the surface of a metal plate.
The catalyst made by the first method can have a high reactivity but
is apt to be poisoned by a high concentration of SOX. The poisoning is due
to the formation of metal sulfates which may plug the pores and reduce the
reactive surface ares. Selection of the alloy and the method of chemical
treatment are important. By the second method, poisoning can be prevented
by using a catalyst known to be S0x-resistant. A possible problem is removal
of the catalyst powder from the metal plate due to the difference in thermal
252
-------
N3
Ul
UJ
Test Site
Research Center
(Hiroshima)
Research Center
(Hiroshima)
Company A
Tokyo Electric
(Yokosuka)
Company B
Fuel
Low-S Oil
High-S Oil
High-S Oil
Low-S Oil
Coal
Capacity and Test Period
1975
i
1976
20(
1977
) Nm3/hr
500 Nm:
1978
/hr
i
2000 I
dm3 /hr
1 i
40,000
Nm3 /hr
i i
60(
) Nm3/hr
i
1979
^^^^.
^^^^M-
"'l^^k.
Major Objective
Catalyst screening
Catalyst life test
Test with actual
boiler flue gas
Cooperation with
Tokyo Electric
Test with actual
boiler flue gas
Figure 4-49. MHI's tests with parallel plate catalyst at 5 pilot plants.
-------
expansion coefficients between the metal and the catalyst and also due to
the possible corrosion of the metal surface by SOX. Use of proper materials
and a proper method of cementing are important.
The 'ceramic plate is made by casting the catalyst material. Activity
and SOX resistance of the ceramic plate are similar to those of granular
catalysts. A problem with this catalyst is that it must be of a certain
thickness to maintain adequate strength; this thickness results in a small
surface area per unit volume of catalyst (Table 3-5). Based on screening
tests and life tests at MHI's Hiroshima Research Center, plate catalysts
with suitable qualities have been produced and used at three pilot plants
with flue gases from high-sulfur and low-sulfur oils and coal (Figure 4-49).
4.6.4 Cooperative Tests with Tokyo Electric Using Plate Catalyst
Cooperative tests with Tokyo Electric Power have been carried out since
March 1977 at Yokosuka Power Station. Plate catalysts were used for 40,000
Nm3/hr of flue gas from a boiler burning 0.2 percent sulfur oil. A flow-
sheet of the pilot plant is shown in Figure 4-50. The major equipment
installations are shown in Table 4-14.
During the tests, the catalyst plates were placed in a duct between
the boiler economizer and the air preheater, particularly on existing
boilers. In one test, ceramic catalyst plates, 1 m x 1 m with 10 mm
thickness, were installed with 14 mm clearance. About 90 percent NOX
removal was attained with 1 mole NHs for each mole of NOx at 300-380°C.
Gas velocity was 7-8 m/sec and SV about 1000 hr"1. Ammonia leakage was
kept below 20 ppm and the ammonium bisulfate problem in the air preheater
was not serious. There has been neither dust clogging nor catalyst
poisoning.
254
-------
to
Ui
AIR;
T
NH3
REACTOR
400-500°C
BOILER
AIR
1
AIR
.PREHEATER
7 \
300-350'C
7 \
FLUE GAS
ECONOMIZER
I
AIR
AIR
PREHEATER
STACK
ESP
Figure 4-50. Flowsheet of the pilot plant at Yokosuka Station, with plate catalyst.
-------
TABLE 4-14. MAJOR INSTALLATIONS OF THE PILOT PLANT
AT YOKOSUKA STATION, TOKYO ELECTRIC
Equipment
Description
Reactor
Induced Fan
Catalyst Bed
Capacity
Pressure
Power
3 x 2 x 8 m
2000 Nm3/min
500 mmH20
250 kW
Air Preheater
Type
Rotating
Gas Analyzer
NOX
NOX and SOX
NH3
Chemical
Luminescence (2)
Ultraviolet
Absorption (1)
Chemical Luminescence
Ion Electrode
Computer
Type
MELCOM 350-7
256
-------
4.6.5 Large Plants for Utility Boilers
Large SCR plants constructed by MHI for treating flue gas from utility
boilers were shown in Table 1-14. Full-scale plants for two new 600 MW
LNG-fired boilers began operation in 1978.
Two full-scale SCR plants, 350 MW (1,012,000 Nm3/hr) and 156 MW
(490,000 Nm3/hr), were constructed by MHI and also began operating in 1978.
The plants use a plate catalyst for the flue gas from oil-fired boilers.
The catalyst plates are placed in a duct between the boiler economizer and
air preheater to remove about 40 percent of the NOx. The pressure drop is
small and is covered by the existing fan. The plants aim at NOx removal at
a minimum cost. No details of the operations have been disclosed yet.
MHI has obtained an order from Chubu Electric Co. for the construction
of a full-scale SCR plant for the Chita No. 4 boiler (700 MW oil-fired
boiler burning 0.2 percent-sulfur oil) to reduce NOX from 100 ppm to 20 ppm
with ammonia leakage of about 10 ppm. The plant will be completed by
February 1980. The existing fan for the boiler has an excess capacity of
100 mmHzO. It is likely that a honeycomb catalyst will be used to give a
pressure drop of less than 100 mmHaO so that no additional fan would be
needed. The investment cost is estimated at 2.1-2.8 billion yen ($10.5
million-$14 million) or 3000-4000 yen/kW ($15-20/kW).
4.6.6 Evaluation
MHI has extensively studied NOX removal not only by SCR but also by
noncatalytic reduction (Section 5.4) and wet processes including simul-
taneous removal with SOX (Section 7.2), in attempts to establish reliable,
economical processes. The use of a honeycomb catalyst at the Chiba plant,
Fuji Oil, appears quite successful: a honeycomb will probably be used for
the 700 MW oil-fired boiler of Chubu Electric. The low investment cost of
257
-------
the plant (3000-4000 yen/kW) is due to its simplicity—just adding a reactor
without increasing the fan capacity.
The plate catalyst aims at further simplification of the process. A
certain thickness, however, is needed for the ceramic plate for maintaining
enough strength; therefore a much larger catalyst weight and bed volume is
required than for the honeycomb. Metal based thin plates may be preferred
for commercial use. The plate catalysts may be best suited for plants which
require a moderate NOX removal ratio of 60-80 percent or which treat the gas
after selective noncatalytic reduction to reduce NHa and to increase NOX
removal efficiency.
4.7 IHI PROCESS
4.7.1 Introduction
Ishikawajima-Harima Heavy Industries (IHI), one of the largest con-
structors of boilers and ships, has tested SCR systems treating flue gases
from boilers burning gas, oil, and coal at many pilot plants; it has con-
structed and will construct several large plants with honeycomb catalysts
mainly for oil-fired utility boilers. Large-scale tests with a combined
system of SNR and SCR have also been carried out. Catalysts by a few
producers, including Catalyst and Chemicals Inc. and Sakai Chemical, have
been used.
4.7.2 Pilot Plant Tests at Power Stations
Extensive pilot plant tests have been performed jointly with Chubu
Electric Power at the Shinnagoya Power Station. There are 2 moving bed
reactors with capacities of 6400 and 1000 Nm3/hr and a reactor with a honey-
comb catalyst (800 Nm3/hr). Also at Taketoyo Power Station is a larger
pilot plant (20,000 Nm3/hr) with a honeycomb catalyst.
258
-------
Figures 4-51 - 4-54 show test results, NOx removal versus temperature,
SV, NHa/NOx ratio, and boiler load for the intermittent moving bed type
reactor at Shinnagoya Power Station. The flue gas was from a low-sulfur-
oil fired boiler. About 95 percent NOX removal efficiency was obtained with
NH3 leakage below 10 ppm at 350°C. An NH3/NOX mole ratio of 1.0 and an SV
of 8000 hr"1 were used (Figure 4-53). Figure 4-54 indicates that the NOX
removal efficiency increased at lower load despite the temperature drop
because of the decrease in SV. The catalyst was not affected by ammonium
bisulfate even at 270°C at the lowest load because of the low SOX concentra-
tion of the gas.
Figures 4-55 and 4-56 show the results of tests with a honeycomb
catalyst at Taketoyo Power Station. The flue gas resulted from combustion
of crude or heavy oil and contained 150-200 ppm of SOX. Ninety percent NOx
removal with less than 10 ppm of NHs leakage was obtained at 360°C with an
SV of 6000 hr~ and NHs/NOX mole ratio of 1.0. The pressure drop was only
about 40 mmHaO and did not increase appreciably during 8000 hours of con-
tinuous testing. The NOX removal efficiency also did not change appreciably
during the period.
IHI has been operating 3 pilot plants, each with the ability to treat
1000 Nm3/hr of flue gas from a coal-fired boiler at Isogo Power Station,
EPDC. Honeycomb catalysts have been used in two of the plants for flue gas
with full particulate load and also for flue gas after fly ash removal by
a hot electrostatic precipitator. The tests have been carried out jointly
with EPDC and the results have not been disclosed yet. Supposedly there is
no erosion of the catalyst except the inlet edge by fly ash at full dust
load when the superficial gas velocity is kept below about 5 m/sec (actual
gas velocity through the honeycomb below 8 m/sec).
259
-------
Dentrification
Efficiency (%) _
O 0 0
^
•o-o-
^^
-° 8__ ^
SV = 8.000 hr"1
NHa/NOx-1.0
300 350 40
Figure 4-51.
Gas temperature
Gas temperature versus denitrification efficiency for the
moving bed reactor at Shinnagoya Power Station.
Denitrifi cation
Efficiency (%)
at co c
o o c
~ "{
" -<
-_
^c
Gas
temperature 350 °C
NH3/NOx = 1.0
5,000
SV value (hr"1)
10.000
15.000
Figure 4-52.
SV value versus denitrification efficiency for the
moving bed reactor at Shinnagoya Power Station.
260
-------
1UU
C«-
O of
•H »-•
4J
10
3&
3 §80
M-H
•W U
•^ -H
G) >M
Q H
o--^(
^^^^^
cy^^
^s^
erf'ft.M.,™
_(
o -9"" "".
o— • — — o— —
SV=aOOO hr-1
Gas temperature 350°C
Inlet NOx con-
centration 130 ppm
O**^*
^^^ ***
O"^**^
> "
,
g
a
.^
•^
4J
(0
1 •
1
O
U
(0
•H
10
on jj
•6V ^^
O
o
10 ii
iw
-------
100
90
30
g 70
60
SO
I 3°
I 20
n 3.
i £ o
LS CRUDE OIL SOX-50VO ppm
N0x>50f130 ppm
Dust-20v60mg/N m3
NH3/NOX mole ratlo-1.0
DENTRIFICATION
EFFICIENCY
SV VALUE
GAS TEMPERATURE
REACTOR OUTLET NHs CONCENTRATION
_L
1/4 L 2/4 L 3/4 L
BOILER LOAD
4/4 L
8000
7000
6000
5000
4000
3000
2000
1000
0
400
350
300
250
200
Figure 4-54.
Relation between boiler load and denitrification
efficiency for the intermittent moving bed
reactor at Shinnagoya Power Station.
262
-------
100
80
0>
60
c
o
10
IJ
SV = 6000 hr"1
Gas Temperature
360°C
Inlet NOX Concentration
180 ppm
o
d)
O
O
o
PO
(1)
s.
o
20
10
0.8 1.0
NH3/NOX Mole Ratio (mole/mole)
1.2
Figure 4-55. NHs/NOx mole ratio versus denitrification efficiency versus
reactor outlet ammonia concentration for the honeycomb
catalyst at Taketoyo Power Station.
263
-------
100
£ 80-
50-
CAPACITY
(Nm3/hr
20,000
FUEL
CRUDE
OIL AND
HEAVY
OIL
SV VALUE AND
THE LIKE
SV=6,OOOhrM
360'C
NH3/NOx mole
ratio 1.0
tx
o
a:
o
LU
a:
tft
C/1
LU
o:
Q_
a:
o
h^
CJ
<
y
40 j
30-
20-
10-
^_Q_— 0 O 0 O O O 0 0 0 O 0 ° ° O^
II li ii it
1000 2000 3000 4000 5000
OPERATION TIME (hrs)
6000
7000
8000
Figure 4-56.
Catalyst life test results for the honeycomb catalyst
at Taketoyo Power Station.
-------
4.7.3 Large-Scale Plants
Table 4-15 lists large-scale test plants and commercial plants
constructed by IHI.
TABLE 4-15. LARGE-SCALE SCR PLANTS BY IHI
User
Ajinomoto
Company A
Company B
Chugoku
Electric
Chugoku
Electric
Tohoku
Plant
Site
Kawasaki
—
—
Kudamatsu
Kudamatsu
Niigata
Reactor
Type
Moving Bed
Fixed Bed*
Fixed Bed*
Fixed Bed
Fixed Bed
Fixed Bed
Gas Treated
(Nm3 /hr)
200,000
1,020,000
453,000
1,000,000
1.900.000
1.660.000
Completion
January 1978
April 1978
June 1978
April 1979
July 1979
August 1981
* Combination with selective noncatalytic reduction.
The Ajinomoto plant uses a. granular catalyst in a moving bed to treat
flue gas from an industrial boiler which burns grade C heavy oil. The other
5 plants are for power companies and use a honeycomb catalyst in a fixed bed
to treat flue gas from utility boilers which burn low-sulfur oil. All of
the plants treat flue gas from a boiler economizer at 350-400°C as shown in
Figures 4-57 and 4-58.
The plants for companies A and B are for existing boilers and use the
combined SNR and SCR system to remove 50-60 percent of NOX. Ammonia is
injected into the boiler at 850-1000°C to remove 25-30 percent of NOX. A
small amount of catalyst is placed in the duct between the boiler economizer
and the air preheater for further reduction of NOx and for reduction of
265
-------
NJ
AMMONIA
INJECTION
BOILER
AMMONIA
INJECTION
DENITRIFICATION
REACTOR
OENITRIFIOATION
REACTOR
FAN
REGENERATIVE
AIR PREHEATER
REGENERATIVE
AIR PREHEATER
FAN
>AIR
-JAIR
STACK
EVAPORATOR
LIQUID AMMONIA
Figure 4-57. Standard flowsheet for the boiler denitrification equipment at 5 large IHI plants
for combustion gas from heavy oil.
-------
Denitrifi cation reactor
Side View
Plan
Figure 4-58.
Plant equipment layout (example)
of 5 large IHI plants.
267
-------
ammonia leakage to about 10 ppm. Although details of the plants have been
kept confidential, the system aims at NOX reduction without increasing the
fan capacity.
The plant for company C is under construction. It should remove 80
percent of NOX. The Kudamatsu plant, Chugoku Electric, is also under con-
struction and should also remove 80 percent of the NOX from the total
2,900,000 Nm3/hr of flue gas from two boilers which burn naphtha and low-
sulfur oil. The structure of the reactor for the Kudamatsu Plant is similar
to that shown in Figure 3-26. The Niigata Plant, Tohoku Electric, is being
designed for 60 percent NOX reduction. NHs leakage will be kept at 10 ppm
or below at those plants.
4.7.4 Economics
No economic data have been disclosed. However, it is said by a
catalyst dealer that the cost of a 150 x 150 x 500 mm honeycomb catalyst
is about equal to that of granular catalyst for a unit volume, although it
is higher for a unit weight.
4.7.5 Evaluation
The honeycomb catalyst used by IHI is efficient. It removes 90 percent
of NOX with 10 ppm of NHs leakage at an SV of 6000 hr"1 and with a pressure
drop of only 40 mmH20. And, the honeycomb catalyst costs about the same as
other catalysts for dirty gas for a unit volume. Therefore, for a dust-rich
gas, the honeycomb catalyst can be advantageous over the granular catalyst,
which requires a moving bed. The honeycomb catalyst seems resistant to SOX
and free from dust clogging. Such desirable features of the catalyst and
proper design of the reactor may have led IHI to construct the large SCR
plants for flue gas from utility boilers.
268
-------
The test plants for companies A and B remove 50-60 percent of NOX by
a combined system of SNR and SCR, while the commercial plant to be con-
structed at Niigata for Tohoku Electric is designed to remove 60 percent
of NOX with SCR only. Use of only SCR may be preferred if there is enough
space to install a reactor or to enlarge a portion of the duct so that a
catalyst can be inserted. With SCR only, ammonia consumption is much less,
process control is easier, a higher removal efficiency is attainable when
needed, and the investment cost including the catalyst may not be much
different from the cost of the combined system.
4.8 KURABO PROCESS (KNORKA PROCESS)
4.8.1 Introduction
Kurabo is one of the largest textile companies with an engineering
division. It has constructed many FGD plants with sodium and ammonia
scrubbing. Kurabo has pioneered the development of the SCR process for
dirty flue gas without precleaning from oil-fired boiler economizers at
350-400°C. In 1974, it completed a 5000 Nm3/hr pilot plant with a moving
bed reactor packed with an alumina-based catalyst and a thermal regeneration
system for the catalyst. A larger pilot plant (30,000 Nm3/hr) has been
operated since 1975.
In 1976, Kurabo made tests with a particulate-rich gas prepared by
adding coal fly ash to oil-fired boiler flue gas. Recently, Kurabo
developed an S0x-resistant catalyst based on titanium oxide which does not
require thermal regeneration. Tests to minimize the ammonia in the reactor
effluent have been carried out by using a microcomputer system.
Kurabo has licensed the process to TRW, Inc. of the United States.
269
-------
4.8.2 Process Description
A flowsheet of the Kurabo process is shown in Figure 4-59. The reactor
for the 30,000 Nm3/hr pilot plant has three elements capable of treating
10,000 Nm3/hr each (Figure 4-60). In each element a spherical catalyst on
an alumina support (carrier), 5 mm in diameter, moves slowly and contin-
uously downward. The gas passes through the elements in a cross flow.
Since the thickness of the element, or the catalyst layer, is fairly small
and the gas velocity through the catalyst is relatively low, the pressure
drop in the reactor is kept below 100
An SV of 7000-10,000 is normally used. The flue gas contains about
1600 ppm SOa , 280 ppm NOx, and nearly 100 mg/Nm3 particulates. More than
90 percent of the NOx and about 80 percent of the particulates are removed
in the reactor at 35 0-400° C with an NHs/NOx mole ratio of about 1.0. An
example of the operating data is shown below:
NOX concentration (ppm) 280 (inlet) 18 (outlet)
Dust content (mg/Nm3) 94 (inlet) 17 (outlet)
Gas volume (Nm3/hr) 24,000
SV (hr-1) 8,000
SOa concentration (ppm) 1,650
Gas temperature (°C) 400
Pressure drop in reactor (rnrnHjO) 65
Ammonia consumption (liters/min.) 106.5
The catalyst discharged from the reactor is screened to remove
particulates, heated to 800°C to decompose aluminum sulfate and to reacti-
vate the catalyst, and is then returned to the reactor. One cycle requires
about 100 hours, therefore only a small portion of the catalyst is regen-
erated at one time. The catalyst has a high strength of more than 10 kg
per granule; loss by crushing is less than 5 percent of the total catalyst
used yearly. After more than 10,000 hours' operation the catalyst activity
270
-------
to
-4
AMMONIAC ([
AIR
PREHEATER
TO FGD
CATALYST
ELEVATOR
(FOR AL203-BASE([
CATALYST)
Figure 4-59. Flowsheet of the Rurabo (Knorka) SCR process.
-------
ELEMENT
ro
^j
S3
CASING
Figure 4-60. Structure of the Kurabo moving bed reactor.
-------
has shown a slight decrease; catalyst life is estimated at over 16,000
hours.
The relationship of the NH3/NOX mole ratio to the NOX removal effi-
ciency and ammonia leakage (ammonia concentration of the gas discharged
from the reactor) is shown in Figure 4-61.
300
~ 200
oo
*100
0.5
1.0
1.5
NH3/NOX MOLE RATIO
Figure 4-61. Relationship of NH3/NOX mole ratio to outlet NOX,
concentrations in the Kurabo moving bed reactor.
Tests have also been carried out with a Ti02-based catalyst that is
resistant to SOX, and requires no thermal regeneration. Reactivity is
adequate at 350°C and life is estimated at 2 years. Use of the T102
catalyst may be more economical than that of AlzOs-based catalysts.
4.8.3 Economics
The investment and annualized costs for SCR, as estimated by Kurabo
for the TiOa-based catalyst, are shown in Tables 4-16 and 4-17. The
investment cost ranges from 4100 to 9300 yen/kW ($20.50 to $46.50/kW) for
273
-------
to
TABLE 4-16. PLANT COST IN 1978 (MILLIONS OF YEN, $5000)
FOR THE KURABO KNORCA PROCESS
Capacity, (Nm3/hr)
Item
Reactor, Heater, Etc.
Ducts and Pipings
Instrumentation
Transportation and Installation
Foundation
Subtotal
30,000
33
16
19
5
7
82
100,000
80
30
28
12
15
165
500,000
252
96
78
37
46
509
Catalyst 10.5 35 175
Total 92.5 200 684
-------
TABLE 4-17. ANNUALIZED SCR COST FOR THE KURABO KNORCA PROCESS*
Ui
Item
Depreciation
Interest
Ammonia**
Power (12 yen/kWhr)
Catalyst (2.5 x 106 yen/m9)
Steam (2 yen/kg)
Labor
Maintenance, etc.
Total
30,000
106 yen/yr
($5000/yr)
11.89
3.70
6.85
7.17
5.78
—
2.00
1.20
38.59
Nm9/hr
yen/kl oil
($0.005/kl)
756
235
435
455
367
—
127
76
2451
100,000
106 yen/yr
($5000/yr)
25.71
8.00
22.80
19.96
19.25
0.19
2.00
2.50
100.41
Nm3 /hr
yen/kl oil
($0.005/kl)
490
153
435
380
367
4
38
48
1915
500,000
106 yen/yr
($5000/yr)
87.94
27.36
57.12
84.27
96.25
0.95
4.00
7.60
365.49
ftuVhr
yen/kl oil
($0.005/kl)
335
104
218
322
367
2
15
30
1393
* C grade heavy oil (No. 6 heavy oil, 0.78 kl/10,000 Nm3; NOX, 280 ppm; 8400 hours annual operation; average
yearly load, 80 percent; gas temperature, above 350°C (upstream of air preheater); requires no auxiliary
burner; catalyst, TiOz-support catalyst.
**160 yen/kg ($0.80/kg) in cylinders for small amounts, 80 yen/kg ($0.40/kg) in tanks for large amounts.
-------
30,000-500,000 Nm3/hr, equivalent to 10-167 MW. The annualized cost ranges
from 1390 to 2450 yen/kl ($6.95 to $12.25/kl) oil or from 0.31 to 0.54 yen/
kWhr ($0.0016 to $0.0027/kWhr) assuming 8400 hours' operation in a year.
4.8.4 Recent Studies
Kurabo has studied reducing the ammonia in reactor effluent without
reducing the NOX removal efficiency by using microcomputer control. The
computer system not only controls the NH3/NOX ratio based on the automatic
analysis of inlet and outlet NOX and NHa concentrations but also corrects
possible temporary errors of the NOX analyzer due to corrosion by SOX or
other problems. With the control, the NHs/NOx mole ratio is always kept
at slightly below 1.0 to maintain an NOX removal efficiency of over 90
percent and an NHs leakage of below 10 ppm.
Kurabo has tested the effects of halogens in flue gas on catalysts.
The TiOa-based catalyst was much more stable than the alumina-based catalyst
against attack by chlorine and fluorine in the gas.
Kurabo has designed a larger reactor element with a unit capacity of
50,000 Nm3/hr. About 40 elements will be needed for a 500 MW coal-fired
boiler. To treat the coal-fired boiler flue gas, Kurabo intends to use a
multicyclone with 80-90 percent particulate removal ahead of the reactor to
reduce the fly ash to 2-3 grams/Nm3. While the amount of fly ash is still
sufficient to cause plugging, Kurabo has found a way to pass the ash through
the moving bed without causing clogging. Since the ash is not sticky, it
is blown off by flue gas at a certain velocity and by the movement of the
catalyst granules in the reactor.
4.8.5 Pilot Plant for Coke Oven Flue Gas (Kurabo-Niigata Process)3
Kurabo has licensed their process for coke oven flue gas to Niigata
Engineering Co., which constructed a pilot plant at the Funamachi Works,
276
-------
Nakayama Steel. This plant can treat 2000 Nm3/hr of flue gas from a coke
oven. The gas composition fluctuates every 20-30 minutes over a wide
range—NOX: 35-400 ppm, SOX: 80-250 ppm, 02: 3-15 percent, C02: 18.5-
20.5 percent, H20: 9-11 percent, and below 35 mg/Nm3 of particulates. The
following 6 kinds of catalysts have been tested:
NKN-1 NKN-2 NKN-3^6*
Shape
Size (mm)
Composition
Crushing Strength (kg/granule)
Spherical
4-6
Cu-Al203
8-10
Spherical
3.5-4.5
Cu-Al203
10-15
Various
4-6
Various
Various
* Including SOx-resistant and low-temperature catalysts.
The flowsheet of the plant is similar to that of Kurabp's pilot plant
(Figure 4-59) except that a gas-gas heat exchanger and a heater are
installed to heat the gas from 180°C. Test results with the catalysts
NKN-1 and NKN-2 are shown in Figure 4-62. NKN-2 is smaller and gives a
higher efficiency but causes a larger pressure drop; the pressure drop in
the reactor was 28-35 mmH20 with NKN-1 at an SV of 10,000 hr"1 and 35-45
mmH20 with NKN-2 at an SV of 7000 hr"1.
Since the gas does not contain much SOX and particulates, the catalysts
were not noticeably affected. For example, use of the catalyst NKN-2 for
1100 hours without moving the bed gave neither an appreciable increase in
pressure drop nor a decrease in NOx removal efficiency. The results of
tests with the other catalysts (NKN-3 - NKN-6) have not been disclosed.
These catalysts include those based on Ti02 which are not only SOx-resistant
but are also efficient at a relatively low temperature of 300-350°C.
277
-------
TOO
90
i
80
350 360
NKN-2
SV • 7,500 hr"1
NHa/NOx • 1.0
380
Temperature (*C)
400
100
90
80
NKN-2
NKN-1
400°C
NHj/NOx = 1.0
5000
7500
SV(hr~l)
10000
100
90
80
NKN-2
NKN-1
400*C
SV • 7,500 hr'1
0.9
0.95
NHj/NOx Mole Ratio
1.0
Figure 4-62.
Results of the Kurabo-Niigata process at a pilot plant
for coke oven flue gas.
278
-------
4.8.6 Evaluation
Kurabo, as a pioneer in dirty gas treatment with a moving bed reactor,
has considerably improved the plant operation and the catalyst. The micro-
computer system that reduces ammonia leakage to 1-2 ppm may have a con-
siderable significance. Further development of the system is desired so
that it can control a big reactor with many elements, each of which may
perform differently.
When the multicyclone is used for coal-fired boiler flue gas, another
particulate removal facility is needed after the air preheater. The total
cost of the multicyclone and the cold dust removal facility may be less
than that of a hot electrostatic precipitator for dust removal before the
reactor. Moreover, the considerable amount of fly ash passing through the
air preheater will reduce the ammonium bisulfate problem in the preheater
(Section 3.5).
The advantage of the Kurabo process is enhanced when the process is
used with an ammonia-scrubbing FGD process such as that developed by Kurabo.
Most ammonia discharged from the reactor is caught by the scrubber and
utilized for SOa recovery. In other scrubbing processes ammonia will
accumulate in the scrubber liquor and may cause a problem. Even when used
with ammonia scrubbing, it may be desirable to lower the level of ammonia in
the reactor effluent so that the ammonium bisulfate problem in the air
preheater and the plume problem in the scrubber can be minimized.
4.9 KUREHA PROCESS
4.9.1 Introduction
Kureha Chemical Industries is a medium-sized chemical company credited
with developing many new petrochemical and electrochemical processes which
have been used in many countries. Kureha recently developed a new SCR
279
-------
catalyst reactive at 150°C, the usual temperature of the boiler flue gas
after the air preheater. The catalyst is therefore designed for application
to existing boilers with which it is difficult to obtain gas at 400°C for
SCR with conventional catalysts.
The new catalyst, however, is-poisoned by SOX. SOX also causes an
ammonium bisulfate deposit which lowers the activity of the catalyst.
Therefore, flue gas containing SOX will first have to be treated to reduce
the SOx content to 1 ppm. Sodium scrubbing and a wet electrostatic precipi-
tator are needed for the purification as discussed below. A 5000 Nm3/hr
pilot plant was put into operation in February 1977.
4.9.2 Process Description
Catalysts usually have a carrier such as AlaOs or TiOa. The Kureha
catalyst consists of base metal compounds and does not have such a carrier.
The catalyst is a cylinder 5 mm in diameter and 10 mm tall. It costs about
1.5 million yen/ton ($7500/ton), no more than other SCR catalysts. The
catalyst is used in a fixed bed, with a superficial gas velocity of 0.75
m/sec and an SV of 5000 hr~ . It removes about 90 percent of NOX with
1.0-1.1 moles of NHa for each mole of NOX.
To reduce SOX in flue gas to less than 1 ppm, Kureha's sodium acetate
FGD process may be useful but the process has not yet been commercialized.
Conventional FGD processes may reduce the SOX content of the scrubber
effluent to about 10 to 200 ppm. Kureha has reduced the SOX concentration
to 1 ppm by sodium scrubbing and a wet electrostatic precipitator, as shown
in Figure 4-63.
The purified gas is heated to 120-125°C by a heat exchanger and to
150°C by an indirect steam heater, injected with ammonia, and sent to the
reactor.
280
-------
STACK
N>
00
FLUE
150»C
GAS
FGD
sox
10-200
55-60-C
S
t T
\
/
/
»>
SOv
80*C
loom
i5» K ESP
CRUBBER
HEAT
EXCHANGER
•« 1
HEATER
25^1^ 5T
m
\
150
3
•c
S
1
i
60
1!
icai
ii
!g
REACTOR
Figure 4-63.
Flowsheet of Kureha SCR process for treating flue gas
discharged from a conventional FGD system.
-------
4.9.3 Economics
Kureha estimates a total investment cost of about 12,000 yen/kW
($60/kW) including sodium scrubbing and a wet electrostatic precipitator.
The SCR unit itself costs about 3000 yen/kW ($15/kW). The following raw
materials and utilities are required for removal of 99 percent of the SOX
and 90 percent of the NOy from 100,000 Nm3/hr of flue gas from an oil-fired
boiler, after passing through an FGD system, containing 50 ppm SOX, 200 ppm
NOX and 100 mg/Nm3 particulates. A catalyst life of 1 year is assumed.
Raw Material Utility
NaOH 24.4 kg/hr Electricity 293 kW
NHa 14 kg/hr Process Water 1.0 tons/hr
Catalyst 22 tons/year Steam 2.1 tons/hr
The annualized cost, including 7 years' depreciation, is about 0.85
yen/kWhr ($0.0042/kWhr) for dirty gas.
4.9.4 Evaluation
Kureha1s catalyst has outstanding activity at low temperatures but its
commercial application is limited because it is poisoned by SOx in dirty
flue gas.
Compared with the other FGD-SCR combination system used in the
Yokkaichi plant, the Shindaikyowa Petrochemical system (4.2), the Kureha
process requires less energy for the gas heating (25-30°C temperature rise
as compared with 80-90°C at Yokkaichi) and has no ammonium bisulfate
problem in the heat exchanger. On the other hand, the Kureha process
requires sodium scrubbing and a wet electrostatic precipitator. These clean
the gas but add considerably to the total cost. The by-product sodium
282
-------
sulfite from the FGD process may be a disposal problem where there is no
demand for the sulfite.
The Kureha process should be used with a highly efficient FGD process
such as the Kureha sodium acetate process. A wet electrostatic precipitator
is needed even with an efficient FGD process because a small portion of SOs
is not caught by FGD.
If a catalyst is found which is S0x-resistant, reactive at 150°C, and
does not require gas purification, the Kureha process can be more widely
used. In this case, ammonium bisulfate will deposit on the catalyst to
lower its activity but the bisulfite can be removed by occasionally heating
the gas to 350°C as is done by JGC (4.4.6).
Because it can be applied without any pretreatment, Kureha's catalyst
is useful for cooled clean flue gas such as that from existing boilers which
burn natural gas. For the gas leaving the economizer of a new natural gas-
burning boiler, SCR with other catalysts at 300-400°C might be preferred
because a much larger SV, up to 20,000 hr"1, could be used.
4.10 KOBE STEEL PROCESS
4.10.1 Introduction
Kobe Steel, one of the largest steelmakers and plant constructors in
Japan, has tested SCR with moving beds at two pilot plants (1000 Nm3/hr
each). They also constructed a prototype plant for Kansai Netsukagaku
(Table 3-1). Kobe also has operated a SCR pilot plant with a honeycomb
catalyst (10,000 Nm3/hr).
Details of the prototype plant have not been disclosed. The operation
data of the pilot plants will be described below.
283
-------
4.10.2 Pilot Plant Test with Moving Bed3
Since early 1976 Kobe Steel has operated two pilot plants with catalyst
in a moving bed. Each can treat 1000 Nm3/hr of flue gas. One is on a coke
oven at the Amagasaki Iron Works and the other is on an iron-ore sintering
machine at the Kakogawa Works. Since the flue gases are discharged at
150-200°C, they are heated to 250-400°C by a heat exchanger and a heater.
A flowsheet of the Kakogawa plant is shown in Figure 4-64. Two
catalysts, a vanadium catalyst 8-10 mm in diameter and an iron catalyst 4-6
mm in diameter, have been used at the plant. Their initial activities are
shown in Figures 4-65 and 4-66. The results of 1000 hours' operation of
the pilot plant with the vanadium catalyst are shown in Figures 4-67 and
4-68. The flue gas contains 160-195 ppm NOX, 160-350 ppm SOX, and 20-40
mg/Nm3 particulates. Over 90 percent of the NOX was removed at 350°C with
an NH3/NOX mole ratio of 1.2 and an SV of 10,000 hr-1. The catalyst bed is
20 cm thick. The pressure drop in the reactor was normally between 40 and
60 mmHaO and increased to above 100 mmH20 late in the test period. The
pressure drop was lowered to 50 mmHaO by increasing the catalyst moving
rate.
4.10.3 Pilot Plant Test with Honeycomb Catalyst
Since September 1975, Kobe Steel has occasionally operated a pilot
plant which can treat 10,000 Nm3/hr of flue gas from a boiler burning coke
oven gas, blast furnace gas, and a low sulfur oil. The flue gas contains
30-50 ppm NOX and 150-180 ppm SOX with 20-30 mg/Nm3 particulates. The gas
leaving the boiler economizer at 320-370°C is injected with ammonia and
introduced into a reactor with a honeycomb catalyst in a fixed bed. The
density of the honeycomb is about 0.5 ton/m3; about 0.7 m3 of the catalyst
is placed in the reactor. The gas volume fluctuates between 7000 and 14,000
Nm3/hr so that the SV ranges between 10,000 and 20,000 hr"1. The NOX in
the gas is reduced to 5 ppm at NHa/NOX mole ratio of 1.2 and a 1.5-3 m/sec
284
-------
REACTOR
FLUE GAS £-
N>
00
Ui
HEAT
.EXCHANGER
HEATER
NH
CATALYST
ELEVATOR
PARTICULATES
STACK
Figure 4-64. Flowsheet of pilot plant at Kakogawa Works, Kobe Steel.
-------
100
S 90
1 80
S. 70
x
z 60
50
NHs/NO = 1.2
NHs/NO - 0.8
330 350 370
Temperature (°C)
Figure 4-65. Initial activity of Kobe Steel's vanadium catalyst
(SV = 10,000 hr-1).
100
S 90
! 80
2 70
iX60
sn
Illl J/ I1W 1 . t-
" ^^ ' ^ NHs/NO = 0.8
•
i i i
340 360 380 400
Temperature (°C)
Figure 4-66. Initial activity of Kobe Steel's iron catalyst
(SV • 5000 hr"1).
286
-------
N>
00
-vl
100
80
70
60-
«
1
0
NH3/NO =1.2 SV = 10,000 hr"1 Temp = 350°C
100 200 300 400 500 600
Test Period (hr)
700
800
900 1000
Figure 4-67. NOX removal efficiency for the first 1000 hours of operation at the
Kakogawa pilot plant with vanadium catalyst in a moving bed.
o
-------
gas velocity through the catalyst (actual velocity) . The pressure drop in
the reactor ranges from 30 to 50
The catalyst is S0x-resistant. The reactor has a soot blower to
eliminate dust deposited on the catalyst. The plant has been operated
trouble-free with occasional soot blowing. Seventy-five kW of power are
required for the flue-gas fan. The SCR unit cost 70 million yen in 1975.
Kobe Steel recently succeeded in producing a ceramic honeycomb catalyst
with a thin wall (0.5 mm) and has conducted pilot plant tests.
4.10.4 Evaluation
The moving bed reactor and the catalyst appear to have been working
well, but they have not been in operation long enough for precise evalua-
tion. The honeycomb catalyst appears promising because of the high effi-
ciency and a small pressure drop. Particularly, the thin wall honeycomb
has a very large surface area per catalyst weight ratio and should be highly
effective. Further tests are desired to determine the applicability of the
honeycomb to larger plants and the catalyst life in gases richer in SOX
and dust .
4.11 OTHER SCR PROCESSES
4 . 11 . 1 Mitsui Toatsu Process
Mitsui Toatsu Chemical, one of the largest chemical companies in Japan,
has developed two types of granular SCR catalysts, MTC-102 and MTC-104. The
MTC-102 catalysts are for clean gas with less than 5 ppm SOX and are
reactive at low temperatures as shown in Figures 4-69 and 4-70. The MTC-104
catalysts are resistant to SOx (Figures 4-71 and 4-72) and are highly
capable of ammonia decomposition. Virtually no NHs is detected in the
reactor outlet when the NHs /NOX mole ratio is kept below 1.2. To treat
288
-------
*«
TOO
9°
80
70
60
50
180
200
SV = 5,000
220
240
260
Temperature (*C)
280
300
Figure 4-69.
Performance of catalyst MTC-102 on a
flue gas from LPG burning.
OJ
cc
100 H
80
60
(at 240'C)
5,000
10,000
15,000
Figure 4-70.
SV (hr"1) -
SV versus NOX removal for the catalyst MTC-102
on a flue gas from LPG burning.
289
-------
ro
>
Oi
DC
100
80
SV
= 5,000
60
320
Figure 4-71.
340
360
380
Temperature (°C)
Performance of catalyst MTC-104 on the
flue gas from heavy oil burning.
100
10
>
o
-------
flue gas from heavy oil burning, Mitsui Toatsu has developed a simple
particulate removal facility in which a granular catalyst is used in a
fixed bed.
Seven recently constructed commercial plants for clean gas use the
Mitsui catalysts. (Five of the larger ones are shown in Table 3-2). All
seven have been operated without trouble. Mitsui Toatsu Chemical made the
basic design of the plants; detailed design and construction were by Toyo
Engineering, a related company, or by Ishikawajima-Harima Heavy Industries.
Mitsui Toatsu Chemical has obtained an order from the USSR to construct an
SCR plant for the treatment of 100,000 Nm3/hr of flue gas from a fertilizer
plant which uses nitric acid to decompose phosphate ore.
Investment and annualized costs are listed in Table 4-18. Investment
cost is fairly high, but plant operation is quite reliable and requires
minimal labor.
4.11.2 Sumitomo Heavy Industries SCR Process
Sumitomo Heavy Industries (SHI) has constructed and operated a proto-
type FGD plant with activated carbon in a moving bed, and also constructed
a pilot plant in 1976 at the Kokura Works, Sumitomo Metal, which can treat
500 Nm3/hr of flue gas from an iron-ore sintering machine. Using this
experience, SHI has developed an SCR process with a moving bed.
A flowsheet of the process for boiler flue gas is shown in Figure 4-73.
A catalyst is moved downward in the reactor, discharged from the reactor,
screened to remove particulates and returned to the reactor. The catalyst,
based on ferrite, is a 7 mm sphere, has a surface area of over 10 m /gram
and has a crushing strength of more than 15 kg per granule. No attrition
has been observed during the operation of the moving bed.
291
-------
TABLE 4-18. INVESTMENT AND ANNUALIZED COSTS
(MITSUI TOATSU CHEMICAL)
Capacity
(Nm3 /hr)
(MW)
Fuel
Gas Temperature (°C)
NOX Removal (percent)
Inlet NOX (ppm)
Inlet SOX (ppm)
Inlet Particulates (mg/Nm3)
Investment Costa
(millions of yen, $5000)
(yen/kW, $0.005/kW)
Annualized cost (yen/hr, $0.005/hr)
Fixed Cost
Variable Cost
Total Cost
Total Cost
(yen/t fuel, $0.005/t fuel)
(yen/kW, $0.005/kW)
Clean Gas
200,000
72
LPG
250-350
90
100
1
1
800
11,000
15,000
11,000
26,000
1690
0.36
Dirty Gas
200,000
68
Heavy Oil
350
90
100
300
200
1000
14,710
18,750
12,600
31,350
2005
0.43
Including catalyst.
3For 8000 hours; yearly operation.
292
-------
BOILER
BLOWER
]VAPORIZER|]-
CATALYST
MOVING
BED
REACTOR
VIBRATING
SCREEN
NH3
STACK
CATALYST
ELEVATOR
PARTICULATES
Figure 4-73.
Flowsheet of Sumitomo Heavy Industry SCR process
for boiler flue gas.
293
-------
Figure 4-74 shows results of tests with a gas containing about 200 ppm
NOX, 450 ppm SOx, and 150 mg/Nm3 particulates. About 95 percent of the NOx
was removed at 350°C, an NH3 /NOx mole ratio of 1.0, and an S/V of 5000 hr"1 .
Relations between the inlet NH3 concentration and the removal ratio and
ammonia leakage is shown in Figure 4-75. The gas for the tests shown in
Figure 4-75 contained about 130 ppm NOX, 450 ppm S02, and 150 mg/Nm9 dust.
Normally an NH3/NOX mole ratio of 1.0 is used at 350°C and an SV of 5000
hr'1 to remove more than 90 percent of the NOX and to keep ammonia leakage
below 10 ppm. Usually 70-80 percent of the particulates in the flue gas
is also removed by the catalyst bed.
It had been found that the activity of the ferrite catalyst is
decreased by the presence of SOX below 350°C. Sulfate deposits on the
catalyst but it can be removed and activity restored by heating the catalyst
at 700°C. The catalyst, with repeated thermal regeneration, is estimated
to have a life of about 1 year. Recently SHI developed an S0x-resistant
catalyst based on TiOa which requires no thermal regeneration.
Based on the experience of operating the prototype FGD plant, SHI has
designed a standard size reactor which can treat about 62,000 Nms/hr gas.
By stacking three reactors, ae shown in Figure 4-76, 186,000 Nms/hr of gas
can be treated. For a 500 MW coal-fired boiler, ten such combined units
may be used. A hot electrostatic precipitator will be installed to reduce
dust content 99 percent from 20 to 0.2 gram/Nm3 , which is an acceptable
limit for the reactor.
Plant operating costs for NOX removal are estimated by SHI as shown in
Table 4-19. The plant cost ranges from 9600 to 13,000 yen/kW ($48 to $65/
kW), depending on the plant size. The denitrification cost, including 7
years' depreciation, is 0.59-0.71 yen/kWhr ($0.0029-$0.0035/kWhr) for 8320
hours' yearly operation. The costs do not include the electrostatic
precipitator.
294
-------
TOO
80
60
Si 40
X
20
250 300 350 400 450 500
Temperature ('C)
Figure 4-74. NOX removal versus temperature for the SHI process
(NH3/NO ratio = 1.0).
60 TOO 140 180
NH3 Inlet (ppm)
Figure 4-75. NOX removal versus NHs inlet concentration versus
NH3 leakage for the SHI process at 300 °C.
295
-------
CATALYST
REACTOR
REACTOR
REACTOR
SCREEN Cr-
XJ
XI
XI
XI
XI
XJ
XI
XI
XJ
XI
XI
XI
XI
XI
XI
XI
CATALYST
ELEVATOR
PARTICIPATES
Figure 4-76.
Stacking of three SHI reactors to treat
186,000 Nm3/hr of gas.
296
-------
TABLE 4-19. SCR COST FOR FLUE GAS FROM OIL-FIRED BOILER,
ESTIMATED BY SHI IN 1976a
Capital Cost (millions of yen, $5000)b
Annual Cost (millions of yen, $5000)
Fixed Costc
Material Cost
Other
Total
Annualized Cost
Yen/kl Oil ($0.005/kl)
Yen/kWhr ($0.005/kWhr)
Flue
200,000
(65)
850
166
161
57
384
3250
0.71
Gas, Nm3/hr
400,000
(130)
1500
281
321
93
695
2940
0.64
(MW)
700,000
(230)
2210
431
562
142
1135
2750
0.59
aCatalyst cost: 3000 yen/kg ($15/kg)
Catalyst life: 1 year
Power cost: 10 yen/kWhr ($0.05/kWhr)
NH3 Cost: 39.7 yen/kg ($0.20/kg)
NOx concentration: 200 ppm
Including civil engineering .work (foundation of reactor).
«
7 years' depreciation, 10 percent interest.
95 percent operation per year (8320 hours).
297
-------
4.11.3 Nippon Kokan Iron Ore Catalyst Process3
Nippon Kokan (NKK), one of the largest steel producers, has developed
a process that uses a certain kind of iron ore as SCR catalyst. NKK has
performed pilot plant tests with 15,000 Nm3/hr of flue gas from an iron-ore
sintering machine (Figure 4-77). The flue gas contains about 200 ppm NOX,
300 ppm SOX, and 1.6 g/Nm3 particulates. It is passed through an electro-
static precipitator and an FGD unit with the NKK ammonia-lime system. A
portion of the desulfurized gas is used for the test after it has been
heated by a heat exchanger and a heater to 350-400°C. Since the gas still
contains considerable amounts of SOX, particulates, and entrained mist, a
moving bed reactor is used.
By using 1.2-1.5 moles of NHs for each mole of NOX and an SV of 4700
hr"1, 80-90 percent of the NOX is removed. The NHa leakage is 50 ppm at
350°C and almost zero at 400°C; the iron ore also functions as an ammonia
decomposition catalyst at 400°C or above.
Tests have shown that if the catalyst bed is riot moved, the catalyst
plugged up in a few weeks; the bed operated smoothly if it is moved. The
iron ore catalyst reacts with SOX to form a powdery sulfate. The iron ore
catalyst discharged from the reactor is screened and returned to the
reactor. The iron ore fines containing the sulfate are returned to the
sintering machine.
Iron ore is cheap and the process is useful for steel plants. Since
the iron ore contains small amounts of potassium and chlorine and the gas
from the sintering machine often contains an appreciable amount of alkaline
chloride vapor which tends to deposit on the SCR catalyst even in a parallel
flow type reactor, the iron ore catalyst may be more suitable for treatment
of such dirty gas. It need not be regenerated and is returned to the
sintering machine when contaminated. The largest disadvantage of the
process is that a considerable amount of energy is required for heating the
298
-------
N>
VO
VD
150'C
.UE
AS
ESP
FAN
PARTICIPATES
Figure 4-77.
Flowsheet of NKK process for treating the flue gas
from an iron-ore sintering machine.
-------
the gas to 350-400°C. With a low-temperature catalyst such as the one
developed by JGC (4.4.6), the energy requirement would be considerably
lowered.
4.11.4 Mitsubishi Kakoki Kaisha (MKK) Process
Mitsubishi Kakoki Kaisha (MKK) has constructed several small commercial
SCR plants as shown in Table 3-2. MKK initially used a ring shaped "Sarc"
catalyst produced by Santetsu Kagaku from ferric oxide. This catalyst,
however, is poisoned by S0x-rich gas when used below 450°C. In the
Kawasaki plant, Nippon Yakin, an improved type of catalyst, Sarc-SL, made
from FeaOa, TiOz and an additive, has been used. The NOX removal perfor-
mance is shown in Figure 4-78. The catalyst is actually used at around
350°C and an SV of 3000 hr"1 to remove 90 percent of the NOX in flue gas
from a high sulfur oil. In other plants, a TiOg-based ring shaped catalyst
produced by Catalyst and Chemicals (Section 3.4.2) has been used at an SV
of 7000 hr
-i
No details of the plants have been disclosed.
to
>
-------
4.11.5 Kawasaki Heavy Industries
Kawasaki Heavy Industries has used various types of catalysts—sphere,
pellet, plate, tube, and honeycomb—on SCR with flue gases from oil and
coal fired boilers. Some of the results for flue gases from oil-fired
boilers are shown in Figure 4-79. It is said that a tubular catalyst gave
the best result for flue gas from coal with a full particulate load, but
no details have been disclosed.
100
90
80
70
60
Honeycomb
Sphere
Tube
300 350 400
Temperature (°C)
Figure 4-79.
removal efficiency of KHI catalysts used for
flue gas from oil-fired boilers.
Note: NOX = 200-300 ppm, SOX = 800 ppm,
Particulates =200 mg/Nm3 for honeycomb
and tube, 20 mg/Nm3
NH3/NO = 1.0
for sphere,
301
-------
4.11.6 Asahi Glass Process
Asahi Glass Co., the largest glass producer in Japan, has tested SCR
in a prototype plant which can treat 70,000 Nm3/hr of flue gas from a glass
melting furnace. The gas contains 500-700 ppm NO and 300-500 ppm SOa. A
granular catalyst and intermittent moving bed developed by Asahi Glass have
been used.
Since the gas contains not only solid particulates but also a consider-
able amount of sodium sulfate vapor which tends to condense on the catalyst
and poison it, a catalyst carrier made of sintered MgO (periclase) is used.
The periclase is fairly resistant to the poisoning. The carrier is quite
strong and undergoes virtually no attrition in the moving bed. On the
other hand, it has a relatively small surface area and is limited to a low
SV. By treating the gas with about 1.1 mole of NHs for 1 mole of NOx, 90-95
percent of the NO is removed at 350-400°C and an SV of 1500-2000 hr"1
The catalyst has been water-washed once every month or two to remove
sodium sulfate and particulates, and has been reused after the base metal
catalyst component has been added. The life of the carrier is about 1 year.
Although over 90 percent of the NOX can be removed, the process is
troublesome and costly due to interference by sodium vapor. This inter-
ference can not be avoided even with a parallel passage reactor or honeycomb
type catalyst. There is presently no plan to install a commercial plant
for removing the NOX from flue gas of a glass melting furnace.
302
-------
4.12 REFERENCES
1. Inaba, H., Catalytic Reduction with Ammonia. Kaikan to Sochi, February
1977. In Japanese; same as Reference 8 of Chapter 3.
2. Miyata, K., Faranox Process Flue Gas Denitrification Plant. Haikan to
Sochi, February 1977. In Japanese.
3. Sams as Reference 4 of Chapter 1.
4. Atsukawa, M., ®t al., Development of NQX Removal Processes with
Catalyst for Stationary Combustion Facilities. Mitsubishi Juko Giho,
Vol. 13, No. 2, 1976. In Japanese.
5. Atsukawa, M., et at., Development of Parallel Flow Type NOx Removal
Systems Using Plate Type Catalyst. Mitsubishi Juko Giho, Vol. 14,
No. 2, 1978. In English.
6. Faucett, H. L., J. D. Maxwell, and T. A. Burnett, Technical Assessment
of NOX Removal Processes for Utility Application. EPRI AF-568, EPA-
600/7-77-127, November 1977. In English.
7- Removal of NOX by Dry Process. Mitsui Zosen Giho, No. 100, 1978. In
Japanese.
303
-------
SECTION 5
SELECTIVE NONCATALYTIC REDUCTION (SNR)
5.1 INTRODUCTION
It has long been known that NHs rapidly reacts with NO at around 1000
°C to form N2 and H20. The reaction is an undesirable side reaction in the
production of NO from NHa by catalytic oxidation at a nitric acid plant
(Equation 5-1). Recent studies have shown that a small amount of 02 is
necessary for the reaction. Therefore, the reaction may be expressed by
Equation 5-2:
6NO + 4NH3 = 5N2 + 6H20 (5-1)
4NO + 4NH3 + 02 = 4N2 + 6H20 (5-2)
Nippon Kokan (NKK) in 1970 applied for a patent for a selective non-
catalytic reduction (SNR) denitrification process in which ammonia is
injected into flue gas above 600°C without a catalyst. Since the reaction
of NHs with NO had been known, NKK claimed the use of a refractory structure
to promote the mixing of NHs with flue gas. Rapid good mixing is a key
point of the process.
A few years later, Exxon and Mitsubishi Chemical Industries (MCI)
independently started working on a noncatalytic ammonia injection process
and both applied for a patent, Exxon a little earlier than MCI. Exxon's
patent includes the use of water-cooled ammonia injection nozzles to
304
-------
minimize the thermal decomposition of ammonia prior to the reaction with
NOX and also the use of hydrogen to lower the reaction temperature. All
of the patents are now pending.
MCI, however, has made large scale tests at its own plants after
obtaining a license from Exxon. Table 5-1 presents major SNR plants.
Mitsubishi Heavy Industries (MHI) worked with MCI and recently completed
a large test unit at Chita Power Station of Chubu Electric which can treat
flue gas from a 375 MW utility oil-fired boiler.
Tonen Technology, a subsidiary of Toa Nenryo, has also obtained a
license from Exxon, conducted large scale tests, and constructed medium-
size commercial SNR plants (Table 5-1).
All of those units remove 40-60 percent of NOX using 1-2 moles NHa to
1 mole NOX. In addition, Mitsubishi Kakoki Kaisha has built many small SNR
units (3000-6000 Nm3/hr) to treat flue gas from heating furnaces. Although
the removal ratio is not as high as by SCR, SNR is simple, less expensive,
and useful for gas sources for which a high NOX removal ratio is not needed.
Laboratory tests on SNR have been carried out by many researchers.
Some results of the tests are compared in Figure 5-1. All tests indicate
a maximum removal efficiency between 900 and 1050°C. Over 80 percent
efficiency was obtained in laboratory tests in which NO and NHs were mixed
and reacted in a silica tube. Tests by Exxon and MCI (curves 1 and 2) gave
particularly high efficiency presumably because of the use of a large amount
j
of NH3 by Exxon and relatively long reaction time by MCI. Tests by CRIEPI
(curve 5) using a small oil-fired furnace (700-800 Nm3/hr of flue gas) gave
maximum 45 percent removal which is close to the efficiency obtained at
large scale plants.
Factors affecting NOX removal efficiency, reaction mechanism, and
operation of large scale plants will be described.
305
-------
TABLE 5-1. LARGE SNR INSTALLATIONS (THERMAL DENOX)
Process
Supplier
MCIa
MHIb
TT°
TT
TT
TT
TT
TT
Plant
User Site
MCI Mizushima
Chubu Electric Chita
Mitsui Chiba
Petrochemical
Tonen Sekiyu Kawasaki
Kagaku
Tonen Sekiyu Kawasaki
Kagaku
Kyokyto Chiba
Petroleum
Toa Nenryo Kawasaki
Toa Nenryo Kawasaki
Gas
Source
Industrial Boiler
Utility Boiler
Industrial Boiler
Industrial Boiler
Industrial Boiler
Pipistill Furnace
CO Boiler
Pipistill Furnace
Gas Treated
Fuel (Nm3 /hr)
Low-S Oil 540,000
Low-S Oil 1,036,000
Naphtha 120,000
Low-S Oil, 423,000
Gas
Low-S Oil, 423,000
Gas
Low-S Oil, 160,000
Gas
314,000
Low-S Oil, 254,000
Gas
Date of
Start Notes
1975 For test
Feb 1977 Continuous
operation
Dec 1975 Occasional
operation
Oct 1976 Emergency
use
July 1977 Emergency
use
Oct 1977 Emergency
use
Mar 1978 Emergency
use
Mar 1978 Emergency
use
Mitsubishi Chemical Industries.
Mitsubishi Heavy Industries.
Tonen Technology.
-------
CO
M
100
80
60
CO
I 40
x
o
55
20
(5)
(5)
700
900 1100
Reaction temperature (°C)
1300
Researcher
(1) Exxon1
(2) MCI
(3) Chubu
Electric
(4) Okayama
University2
(5) CRIEPI3
Reactor
Silica Tube
Silica Tube
Silica Tube
Silica Tube
Oil Furnace
NO
(ppm)
826
200
200
500
225
NH3
(ppm)
2540
400
240
667
225
02
(*)
4.6
2.0
1.0
1.0
1.0
Reaction
Time
(sec)
0.13-0.17
0.4
0.2
0.1-0.13
0.11
Figure 5-1. Results of laboratory tests of SNR
by five research organizations.2
307
-------
5.2 LABORATORY TESTS AT OKAYAMA UNIVERSITY
S. Kasaoka, Okayama University, has performed extensive laboratory
tests on the noncatalytic reduction process.2 In the tests, he used a 13
mm diameter silica tube placed in a tubular furnace, and passed a composite
gas through the tube at a rate of 1 normal liter per minute. Retention
time of the gas at reaction temperature was 0.10-0.13 second. Both the
inlet and outlet NO and NHs concentrations were measured.
Figure 5-2 shows the removal rate of NO and the conversion rate of NH3
when a gas containing 500 ppm NO and 667 ppm NH3 (NH3/NO mole ratio = 1.33)
with or without Ha or HaO was treated at temperatures between 450 and 1000
°C. For a gas without Ha, the optimum reaction temperature was 950-1000°C.
At 1000°C, about 80 percent of NOx and 95 percent of NH3 were removed,
indicating that less than about 100 ppm NOx and 30 ppm NH3 were present in
the outlet gas. Addition of Ha lowered the optimum reaction temperature
as claimed by Exxon and also tended to lower the NOx removal efficiency.
Presence of HaO lowered both NOx removal and NH3 conversion rates.
Figure 5-3 shows that the use of a larger amount of NH3 gives a higher
NOX removal ratio and larger amount of ammonia leakage (NH3 in outlet gas).
At 1000°C with 1000 ppm NH3 (NH3/NOX ratio 2) and without H2, the NOX
removal rate exceeded 90 percent and the NH3 conversion rate was 94 percent
(NH3 leakage was about 60 ppm). Reaction at -lower temperatures using Ha
gave a smaller NOX removal rate with more NH3 leakage.
Figure 5-4 shows the effect of Oa concentration on NOX removal and NH3
conversion. The optimum Oa concentrations for NOX removal were about 3
percent for a gas with 5 percent Ha and about 1 percent for a gas with 1
percent Ha. The optimum 02 concentration for a gas without Ha was below
1 percent (not shown in Figure 5-4)- Excessive Oa concentration lowered
the removal efficiency although it also lowered the amount of NH3 leakage.
308
-------
-------
100
5* 80
I
I 50
UJ
< 40
o
5
OS
20
NH3
500 1000 1500 2000
INLET NH3(ppm)
MARKS
0
A
9
O
D
INLET
NO(ppm)
500
500
500
500
500
GAS COMPOSITION
H2(ppm)
0
1000
3000
5000
50000
02 (%)H20 («) N2
5 10 BALANCE
5 10
5 10
5 10
5 10
REACTION
TEMP (°C)
1000
800
750
700
600
Figure 5-3. Effects of NHs and H2 concentrations and reaction temperature
on the NOX removal rate and the NHs conversion rate with a
reaction time of 0.1-0.13 second, according to tests at
Okayama University.
310
-------
1
£
oi
100
80
60
40
20
•- 0-0 — •••"
5 10
INLET Og CONCENTRATION
15
MARK
o
A
O
0
D
INLET GAS COMPOSITION
NO (ppm) NH3 (ppm) H2 (ppm)'
500 667 0
500 667 1000
500 667 5000
500 667 10000
500 667 50000
H20 (%)
10
10
10
10
10
REACTION
TEMP. ( C)
1000
800
700
650
600
Figure 5-4.
Effect of Oz concentration on the NOx removal rate and the
NHs conversion rate with a reaction time of 0.1-0.13 second,
according to tests at Okayama University.
311
-------
Figure 5-5 illustrates the thermal decomposition of NHs in a. gas
without NO. When Hz was absent, NHs decomposed about 800°C forming a small
amount of NO. HZ promoted the decomposition of NHs and the formation of
NO at a lower temperature. These results explain the effects of HZ shown
in Figure 5-2.
Figure 5-6 indicates the relationship of the reaction time to NOX
removal. Under appropriate conditions (950°C without Hz and 750°C with
3000 ppm Hz), about 70-80 percent of the NOX was removed within 1 second.
A much longer reaction time was required at a lower temperature. In an
actual boiler, the retention time of the gas in a suitable temperature
range (for example, 950-1050°C without H2 and 650-750°C with 3000 ppm H2)
may be less than 0.2 second. Therefore, it is important that NHs be
introduced in a correct temperature zone and be mixed with the gas rapidly.
Figure 5-7 illustrates the relationship of reaction time to NHs
conversion and NO formation for gases with different compositions at
different temperatures. (NO was not present at first.) Under conditions
suitable for NOx removal, 950-1000°C without Hz, NHs decomposed rapidly; the
amount of NO formed was small and did not increase at prolonged reaction
time. At 700°C and 650°C in the presence of HZ, the NO concentration
increased with prolonged reaction time.
5.3 STUDIES BY CRIEPI
The Central Research Institute for Electric Power Industry (CRIEPI)
has studied SNR to characterize the theoretical relationships of the
associated reactions. Since equilibrium constants and reaction rate
constants of many of the reactions related to NOX reduction are already
known, Criepi calculated from those data the reaction rates for a gas con-
taining 200 ppm NO, 1 percent 02, and 11 percent HzO. The results are
shown in Figure 5-8. The gas composition is similar to that of flue gas
from an oil-fired boiler.
312
-------
MARKS
O
A
IUU
w 80
o
h-
^ 60
o
i— i
Of.
LU
z1 40
o
o
CO
z
20
0
6
/
/ NH3
r i
I
I /
/ NHX •
I
30 800 10
3UU
400
300 |
Q
LU
Of.
200 £
0
100
0
00
TEMPERATURE (°C)
INLET GAS COMPOSITION
NH3(ppm) H2(ppm) 02(%) H20(%) N2
667 3000 5 10 Balance
667 0 5 10 Balance
Figure 5-5. Effect of Ha concentration on decomposition of NHs and
formation of NO with a reaction time of 0.1-0.13 second,
according to tests at Okayama University.
313
-------
X
o
80
60
40
20
2 4
REACTION TIME (sec.)
Marks
D
A
O
O
Composition
NO(ppm) NH
500
500
500
500
of Inlet Gas
3(ppm) MX)
667 0
667 0
667 3000
667 3000
02(X)
5
5
5
5
H20 (X) N2
10 Balance
10
10
10
Reaction
Temp. (°C)
950
850
700
650
Figure 5-6. Effect of reaction time, temperature and gas composition
on NOX removal, according to tests at Okayama University.
314
-------
2
o
I—1
00
O
O
- 60 £
LU
o
O
246
REACTION TIME (sec.)
Mark
D
A
O
o
Inlet
NH3 (ppm)
667
667
667
667
Gas Composition
H^ppm)
0
0
3000
3000
Op («)
5
5
5
5
H20 (%) N2
10 Balance
10 ..
10 ,,
10 n
Reaction
Temp. (°C)
1000
950
700
650
Figure 5-7. Decomposition of NHs and formation of NO of different gases
at different temperatures, according to tests at Okayama
University.
315
-------
200
I
p.
\M/
00
cd
M
0.04 0.08 0.12
Reaction time (sec.)
0.16
0.20
Figure 5-8. The effects of the NHa/NO mole ratio and temperature on
NOx removal, according to CRIEPI studies.*
* Calculated for a gas containing 200 ppm NO, 1 percent 02 and 11 percent
, similar to a gas from an oil-fired boiler.
316
-------
Figure 5-8 indicates that 95 percent NOX removal can be attained at
1027°C (1300°K) in 0.1 second with an NH3/NO mole ratio of 1.5 and in 0.07
second with a ratio of 2.0. Those calculations were made assuming an ideal
gas mixing which cannot be actually attained, particularly in large scale
plants. Figures 5-9 and 5-10 show the conversion of NHa and formation of
NO. Figure 5-10 indicates that the concentration of NO formed at 1027°C
is only about 2 ppm and hardly increases at longer reaction time while much
more NO is formed at higher temperatures.
Figure 5-10 does not agree well with Figure 5-7. This indicates that
NO concentration exceeded 10 ppm within a second at 1000°C. The discrepancy
may be due to the difference in 02 concentrations—1 percent in Figure 5-10
and 5 percent in Figure 5-7—as well as to the difficulty of instantaneous
and uniform heating and mixing of a gas.
Figure 5-11 (A-C) illustrates the effects of Hz on the conversion of
NHs and formation of NO and NHs. The theoretical calculation shows that
small amounts of NH2, an unstable gas, form at the beginning of NHs con-
version and that H2 promotes the conversion of NH3 and the formation of NO.
CRIEPI assumed that the following reactions would occur:
NH3 + k 02 = NH2 + h H20 (k = 1.2 x 10~3 at 700°C) (5-3)
NO + NH2 = N2 + H20 (k - 7.9 x 1025 at 700°C) (5-4)
NH3 + NO + k 02 = N2 + 1*8 H20 (5-5)
H2 + 02 = H20 + 0 (k - 1.8 at 700°C) (5-6)
0 + 2NH3 - 2NH2 + H20 (5-7)
NH3 + 2% 0 = NO + llg H20 (k - 9.4 x 1039 at 700°C) (5-8)
317
-------
200
100
ft
>»••
n
0 0.04 0.08 0.12 0.16 0.20
Reaction time (sec.)
Figure 5-9. Decomposition of NHs, calculated by CRIEPI.
I
a.
200
150
100
50
NH, 1027'C
NH9 1127°C
NO 1227°C
_NO 1127°C
"NO ~QZ7'C
50
40
30
20 5
10
0
•%
i
8
12
16
20
Reaction tine (nilisecond)
Figure 5-10. Effect of temperature on NHs conversion and
NO formation, calculated by CRIEPI.
318
-------
200
100
(A) 1027'C
Hi neat
s
8 12 16 20
40
20
I
0 i
200
100
200
100
*
a"
(B) 1027'C
Hs 200ppm
40 ^
20 «&
0 9
8 12 16 20
(C) 1027'C
Ha 300ppm
MO
2/3 H«
40
20
3 12 16 20
Figure 5-11. Effect of HZ on NHs conversion and NO and
formation, calculated by CREIFI.
319
-------
BOILER TUBES
BURNER
V
1
FURNACE
AIR HEATING TUBES
(.A) OUTLOOK OF FURNACE
(I) (ID (HI)
NH3 INJECTION POINTS
COOLING WATER OUT
COOLING WATER OUT
GAS FLOW *-
I)
(ID (HI)
(1). *
(2)
FOR
TEMPERATURE
, MEASUREMENT
COOLING WATER IN
(B) DETAIL OF NHjINJECTION
POINTS (I,II,III) AND
TEMPERATURE MEASURING
POINTS (1,2.)
(VIEW OF OPPOSITE SIDE
OF THE FURNACE SHOWN IN A)
COOLING WATER IN
NH3 INJECTION NOZZLES
(3 NOZZLES WITH TOTAL 19 HOLES)
(CROSS SECTION OF FURNACE)
(a) REFRACTORIES
(b) COOLING WATER
Figure 5-12. Furnace and ammonia injection nozzles for
SNR by CRIEPI (sizes are in mm).
320
-------
NHa reacts only slightly with NO when Oa is not present. It is
assumed that reaction 5-3 occurs to form NHs, which reactions with NO to
form Na as shown by equation 5-4. The overall reaction is shown by
equation 5-5. The equilibrium constant k of reaction 5-3, however, is very
small at 700°C and thus reaction 5-5 does not proceed at 700°C.
When Ha is added, reaction 5-6 occurs to form an oxygen radical which
readily reacts with NHa to form NH2. The NH2 thus formed readily reacts
with NO. The equilibrium constants for reactions 5-6, 5-7 and 5-4 are
large enough for the reactions to proceed at 700°C. A portion of the oxygen
radical reacts with NHs to form NO (reaction 5-8). Therefore the overall
NOx removal efficiency tends to be lowered by the addition of Ha as already
shown in Figure 5-2.
CRIEPI has made tests with a small oil-fired furnace about 1 meter in
diameter, burning 50 liters/hr of oil, giving 700-800 Nm3/hr gas as shown
in Figure 5-12. Three ammonia injection points, I, II and III were used
for tests. Three nozzles were placed in each point as shown in B and C of
the figure. The three nozzles have a total of 19 holes with 1.1 mm diameter
through which NHs or a mixture of NHa and Ha was injected. The injection
point I gives about 0.1 second reaction time of NHs with NOX while the
points II and III give less reaction time. The gas velocity at the
injection points was about 4.6 m/sec at 1000°C.
Figure 5-13 shows the results of preliminary tests on the effect of
water cooling of ammonia injection nozzles. By injecting 1 mole NHs to
1 mole NOX from the injection point I, about 45 percent NOX removal was
attained when water-cooled nozzles were used but the results were poor
without water cooling. This indicates that a large portion of NHs was
decomposed at above 700°C in the tubes before it was mixed with flue gas.
Water-cooled nozzles were used for other tests.
321
-------
s-e
-------
50
40
30
20
10
III
Figure 5-14.
700 800 900 1000 1100 1200
Temperature(° C)
NOX removal by NHs injection at different points
on a small CRIEPI furnace.*
* Inlet NOX = 225 ppm
NH3/NOX ratio = 1
02 = 1 percent
323
-------
0 100 200 300 400 500
INITIAL NOV (ppm)
Figure 5-15. NOX removal efficiency for the flue gas from the burning
of several fuels in a small CRIEPI furnace.
324
-------
MARK
0
A
D
•
A
•
MOLE
NH3/NO
1
1
1
2
2
2
RATIO
H2/NO
0
1
2
0
1
2
800 900 1000 1100
GAS TEMPERATURE (°C)
1200
Figure 5-16.
Effect of NHa/NO mole ratio and Ha addition on NOx removal
for the flue gas from oil burning in a small CRIEPI
furnace.
325
-------
CRIEFI detected minor amounts of cyanate in the gas after SNR. When
1 mole NHs was used to 1 mole NOx, the amount of cyanate was close to that
without ammonia injection. The amount of cyanate increased with the amount
of ammonia but was still very low and may not be an environmental problem.
5.4 LARGE SCALE TEST OF SNR AT CHITA STATION, CHUBU ELECTRIC
5.4.1 Introduction
In 1977, Chubu Electric Installed a full scale SNR test unit at Chita
Station for the existing No. 2 boiler which burns a low sulfur oil. Before
the installation, basic studies were made by Chubu Electric and also by
MHI (Section 4.6), the constructor of the unit. The unit is the world's
first full scale SNR plant for use on a utility boiler.
5.4.2 Basic Study by Chubu Electric^
Chubu Electric has made extensive laboratory tests on SNR using an
apparatus as shown in Figure 5-17. Some of the test results are shown in
Figures 5-18 - 5-21. Maximum NOX removal was obtained at 950-1000°C
(Figure 5-18). When 240 ppra of NHs was used for 200 ppm NO, 80 percent of
the NO was removed at 1000°C with a reaction time of 0.2 second (Figures
5-18 and 5-19). The addition of Ha or CHi* shifted the optimum reaction
temperature from 1000°C to about 800°C (Figures 5-20 and 5-21).
Figures 5-22 and 5-23 show the effects of temperature and NHa/NO mole
ratio on NOX removal efficiency and NHs leakage. At 1100°C and 1200°C,
both NOx removal efficiency and NHa leakage are low. It is likely that a
portion of the NHs is converted to NO at the high temperatures.
326
-------
NOf
OJ
10
GAS FLOW
CYLINDERS
HUNIDIFIER
SAMPLING
TUBE FOR
RESIDUAL NH3
HEAT TRACED
LINE -150'C
MIXING &
PREHEATING
CHAMBER
(100-200'C)
REACTION ZONE
FLOW
BYPASS
Figure 5-17. Schematic of experimental apparatus for Ctebu Laboratory tests of SNR.
-------
00
800
900
1000 1100
TEMPERATURE (*C)
12QO
CONDITIONS
NO 200ppm
NH3 240 ppm
02 1%
C02 15%
H20 11%
Figure 5-18. NOx reduction versus reaction temperature with Chubu Electric's SNR.
-------
u>
ro
vo
IUU
s 80
| 60
0
•o
Si 40
ox
Z 20
0
C
XJ
/^
/^
/6
^^
_J
CONDITIONS
NO 200 ppm
NH3 240 ppm
02 1%
C02 15%
H20 11%
0.05 0.10 0.15 0.20
Residence Time (seconds) at 1000°C
Figure 5-19. NOX reduction versus residence time in reaction zone
with Chubu Electric's SNR.
-------
loo r
u>
o
700
800
9QO 1000
Temperature (°C)
1100
1200
CONDITIONS
NO
NH3
02
C02
H20
Reaction
time
200 ppm
200 ppm
1*
15%
11%
0.2 sec
Figure 5-20. NOX reduction versus reaction temperature (effect of H2 addition)
with Chubu Electric's SNR.
-------
to
800
900 1000
Temperature (°C)
1100
1200
CONDITIONS
NO 200 ppm
NH3 200 ppm
02 1%
C02 15%
H20 10%
Reaction
time 0.2 sec
Figure 5-21. NOX reduction versus reaction temperature (effect of CH^ addition)
with Chubu Electric?s SNR.
-------
II
h
X
§
0 0.5 1.0 1.5 2.0 2.5
NHs/NO mole ratio
(Initial NO 200ppm, 08 1Z, CO, 15Z, HaO 11Z)
(Reaction time 0.2 sec.)
Figure 5-22. NHs/NO mole ratio and NOx removal with Chubu
Electric's SNR.*
* Initial NO, 200 ppm; 02, 1 percent; C02, 15 percent; H20, 11 percent;
reaction time, 0.2 second.
30
n
as
z
to
-------
5.4.3 Basic Study by MHI5
MHI has studied SNR jointly with Mitsubishi Chemical Industries (MCI),
one of the largest chemical companies in Japan, and constructed a large-
scale test plant for Chubu Electric at Chita Station. Figures 5-24 and
5-25 show some of the results of laboratory tests. To design the large
plant, MHI assumed the 3-major reactions and determined the reaction rate
constants of the three: ki, k2, and k3.
ki
4NO + 4NH3 + 02 >• 4N2 + 6H20 (5-9)
k2
4NH3 + 502 >• 4NO + 6H20 (5-10)
4NH3 + 302 >• 2N2 + 6H20 (5-11)
k1(NO)a(NH3)3(02)^1 - k2(NH3)6(02)Y2 (5-12)
ki(NO)a(NH3)3(02)Yl + (k2 + k3)(NH3)6(02)Y3 (5-13)
The concentrations of NO and NH3 during and after the reactions in a
boiler can be calculated from equations 5-12 and 5-13. An example of the
calculation, assuming the treatment of flue gas in a large scale utility
boiler, is shown in Figure 5-26. The figure shows gas temperatures and NOX
removal efficiencies at distances within 1800 mm from the NH3 injection
point. The flue gas first passes an open space (0-400 mm) at about 1050°C
in 0.04 second, and then passes the heat transfer zone (400-1800 mm) at
1050*-920°C in 0.1 second. NH3 is mixed with the flue gas in an open space
and in a part of the heat transfer zone. Further mixing is accomplished
in the dispersion reaction zone. NH3 rapidly reacts with NOx in the
dispersion reaction zone.
333
-------
100
80
| 60
I
I 40
NO 200 ppm
t 0.4 see
NHS/NOM 2, Oi 2%
±
Reduction, %
— NHS Reduction,
=3=z~
600 700 800 900
Gas Temperature (*G)
1000 1100
Figure 5-24. Laboratory data (effect of gas temperature) for the
MHI-MCI joint study at Chita Station.
334
-------
I
-------
UJ
o
t-
4->
-
60
50
30
20
10
Open Space
(r = 0.04)
Heat Transfer Zone (r = 0.10 sec)
1200
600
70 210 350 490 700
Jet Mixing Zone
800 1000 1200 1400 1600^
Distance In Gas Flow Direction (inn)
1800
Dispersion Reaction Zone
Figure 5-26. A calculation result by simulation on practical boiler.
-------
5.4.4 Large Scale Plant Tests'*'5
The large scale SNR tests have been performed on the No. 2 boiler at
Chita Station which can generate 1225 tons/hr of steam and 375 MW of power.
The total gas volume exceeds 1,000,000 Nm3/hr. The boiler burns a low
sulfur oil (0.2 percent S). The gas contains about 100 ppm SOx, of which
about 3 ppm is S03, and 110-140 ppm NOX, virtually all of which is NO. The
operation load of the boiler varies from 100 to 30 percent (375 to 113 MW).
To cope with the temperature fluctuation due to the boiler load, 15 water-
cooled ammonia injection nozzles were placed in two locations, front and
rear, as shown in Figure 5-27. The structure of the nozzles is shown in
Figure 5-28. The nozzle cooling systems are shown in Figures 5-29 and 5-30.
As shown in Figures 5-31 and 5-32 A, the gas temperature at 150-375 MW
(40-100 percent load) ranges from 900 to 1150°C at the front nozzles and
800-1030°C at the rear ones. NHa is injected from the rear nozzles at 290-
375 MW (78-100 percent load, 930-1030°C), from both nozzles at 225-290 MW
(60-70 percent load), and only from the front nozzles below 60 percent load
(below 1000°C). When the boiler load drops below 50 percent, no NHa is
usually added because the amount of NOx emission can be within the regula-
tion.
Operation design conditions for 40 percent NOX removal with less than
20 ppm of NHs leakage are shown by the shadowed lines in Figure 5-32 B, C,
and D. A NHs/NOx mole ratio of 1.5 at full load and smaller ratios at
lower load may be used.
Figure 5-33 compares the calculated NOX removal efficiencies and NHj
leakage with those observed in actual plant operation. The calculated
efficiency and NHs leakage are shown by lines. The removal efficiency was
a little lower when the initial NOX concentration was 110 ppm than when it
was 140 ppm. In both cases the predicted performance ?and the actual
operating data agreed well for NOx removal. Actual NHs leakage was a
337
-------
REAR SIDE
NOZZLE
FRONT SIDE
NOZZLE
SECONDARY
SUPERHEATER
PRIMARY
SUPERHEATER
BURNERS
Figure 5-27. Locations of nozzles in the No. 2 Boiler at Chita Station.*
* Boiler Type: Mitsubishi - CE Control, Radiant Reheat Divided.
Evaporation: 1225 T/H, 375 MW
Pressure: 176 atg.
Temperature: 571/540.6°C
338
-------
Figure 5-28. NHa injection nozzle in the No. 2 Boiler
at Chita Station.
339
-------
c
t T
MAIN DRUM
(
\
^DE
i
1EA
r
1
I
J)EI
R
1 II II 1
BWCP
C)
1EADE
LU I/)
O LU
•-i co
i—
O —1
0£. <
Lu 3
oo
LU LU
O CO
D>h-
-------
TO DRUM
NOZZLE HANGING
TUBES
NOZZLE
COOLING TURFS
Figure 5-30. Concept of nozzle cooling systems on the
No. 2 Boiler at Chita Station.
341
-------
o
— 1200
5 noo
(0
I 1000
OJ
ITl
cs
900
800
Point®
60
40
c
o
u
T3
Ol
Of
ox 20
NH9 Injection at PointQ)
A
only front
nozzles
(pointQ)
NHS Injection at Point®
both nozzles
~~ t -*-
only rear nozzles
(point®)
»
100
200
300
375
Load (MM)
Figure 5-31. Effects of NHa injection position and load on NOx reduction
rates on the operation of the No. 2 boiler at Chita Station.
342
-------
Ga«
H"
s
o
»-•
l
o
l
o
800
)
g
BO,
§
20
4!
80
60
40
20
-(P)
23
SO
73
100
3 !
2 9
4.
0.3
23
30
73
100
23
30
73
100
N 1 N !\ \
2\ °'5\ l\ 2-
V Sv \ ^
Sx \ \
Design
X
23 50 73
Boiler load (Z)
100
Figure 5-32. Data on the designed operation conditions and anticipated
results at different operation loads of the No. 2 boiler
at Chita Station (Mitsubishi Heavy Industries).
343
-------
70 r
60
50
40
30
20
10
TEST
O RESULTS (BASE NOX HO ppm)
X " ( " 110 " )
ppm
70
60
50
40
30
20
10
ppm
0.5 1.0 1.5 2.0
NH3/NOX MOLE RATIO
2.5 3.0
Figure 5-33.
Comparison between test results and predicted performance
of the No. 2 boiler at Chita Station.
344
-------
little larger than the calculated amount. An NOX removal efficiency of
about 40 percent has been obtained with about 30 ppm of NHs leakage by
using 1.5 mole of NHs for each ntole of NO.
Operation was begun in February 1977 and has been virtually trouble-
free except for deposits of ammonium salts with particulates in the air
preheater. NOX removal has been 40-45 percent. At the beginning of the
operation a considerable amount of ammonium salt deposits formed. These
increased the pressure drop and necessitated a water wash for their removal
after 2 months' operation. The problem has been mitigated by increasing the
soot blowing capacity. It is estimated that the water wash is required
about twice a year. The stack usually gives no plume, but in winter a
considerable plume was observed when outlet NHs exceeded about 40 ppm.
5.4.5 Economics of the Chita Plant
The SNR unit at the Chita plant cost 600 million yen ($3 million) in
1977 (1600 yen/kW, $8/kW). MHI has begun construction of an SCR plant for
the existing No. 4 boiler of Chita Station which burns 0.2 percent sulfur
oil. It will remove over 80 percent of the NOx- The plant cost is
estimated at about 3500 yen/kW ($17.50/kW). Therefore, the investment cost
for SNR is a little less than half that for SCR. On the other hand, the
NOX removal efficiency of SNR is about half that of SCR while ammonia
consumption and leakage are much higher. The water wash required for the
air preheater increases the cost of SNR.
5.5 MITSUI PETROCHEMICAL, CHIBA PLANT
The SNR unit for the industrial boiler which can generate 120 tons/hr
of steam (flue gas 120,000 Nm3/hr, 40 MW equivalent) has been operated
occasionally since 1976 at the Chiba plant, Mitsui Petrochemical. The plant
has 5 boilers. One of the boilers was chosen for SNR because it has a
relatively large volume per capacity and has enough space for ammonia
345
-------
injection nozzles. The space available is in the temperature range of
about 700°C to 800°C, a little lower than the optimum reaction temperature
(850-1000°C) for NOx reduction by NHs. Therefore, HZ has been used with
NHs to obtain optimum NOx reduction at 700-800°C. Actually three nozzles
have been placed in the temperature range and a total of 2 moles NHs and
8 moles Ha has been injected for each mole of NOX. Steam has been used for
diluting the injection; air must be avoided because of the presence of Ha-
Usually 35-40 percent of the NOx is removed and there are 10-15 ppm of NHs
leakage.
At the beginning of the operation, the boiler used 0.8 percent sulfur
oil and leak NHs ranged from 20 to 30 ppm to achieve about 45 percent NOX
removal. The air preheater became plugged within a few weeks by the
ammonium bisulfate deposit. Since then, sulfur-free fuel, either gas or
oil (S = 0. 02 percent or below), has been used for the boiler. Operation
has been trouble-free.
Mitsui Petrochemical recently installed an SCR unit for a larger boiler
(200,000 Nm3/hr) which burns low sulfur oil (S = 0.2-0.6 percent) to reduce
more than 90 percent of the NOx with NHs leakage below 5 ppm. The air
preheater has no dust plugging problem because of the low NH$ leakage.
Mitsui Petrochemical estimates that the overall cost to remove an equal
amount of NOX is lower with SCR than with SNR.
5.6 OTHER SNR PLANTS
As shown in Table 5-1, there are five relatively large SNR units con-
structed in 1976 or later. Those are all for oil or petrochemical companies
and have been used for emergency when a photochemical smog warning is issued
or when the total NOX emission from the refinery or the petrochemical plant
exceeds the regulation. Those units can be operated continuously if needed
and are capable of removing 45-57 percent of NOx by adding 1-2 moles of
together with Ha to 1 mole of NOx.
346
-------
5.7 COMPARISON OF SCR, SNR AND THE COMBINED PROCESS
Table 5-2 shows estimated NOx removal efficiency and ammonia leakage
when flue gas containing 200 ppm NOX is treated by SCR, SNR and the combined
process.
It is possible to attain 95 percent NOx removal efficiency by SCR when
sufficient amounts of NHa and catalyst are used (No. 1 of Table 5-2).
However, optimum operation may be represented by Nos. 3 and 5 which attain
80-85 percent NOX removal with 5 ppm of NHa leakage using a smaller amount
of NHa and catalyst, because the investment cost as well as consumption of
power and ammonia is low and ammonium bisulfate problem is reduced. When
lower removal efficiency is adequate, both catalyst and NHa may be further
reduced. If increasingly stringent regulations are expected, the reactor
vessel and the fan with sufficient capacities may be installed and the
amounts of catalyst and NHa may be increased to meet the regulations.
SNR with 1-1.5 moles NHa for 1 mole NOX (200 ppm in flue gas) may give
40-50 percent removal efficiency with 30-40 ppm NHa leakage. Use of 2-3
moles of NHa may give 55-60 percent efficiency with 50-80 ppm of NHa leakage
which may be unacceptably high. Placing a small amount of parallel flow
type SCR catalyst in a boiler duct at 300-400°C will reduce NHa leakage and
increase NOx removal efficiency (Nos. 22, 23, 32, 34, and 42). By using
about 30 parts or less catalyst, the increase in the pressure drop may be
below 30 mm HaO and may not require increasing the capacity of the fan.
The combination of SNR and SCR with the small amount of catalyst may be
useful for existing boilers which have no space to install an SCR reactor
or an additional fan. The NOX removal efficiency may be further increased
by adding a small amount of NHa before the catalyst bed (Nos. 23, 24, 34,
35 and 44). This method, however, may complicate the process control.
It is possible by the combined system to attain 80 percent or more
NOX removal with 10 ppm of NHa leakage when more than 50 parts of the
347
-------
TABLE 5-2. COMPARISON OF SCR, SNR AND COMBINED SYSTEMS
FOR AN INLET NOX CONCENTRATION OF 200 PPM
No.
1
2
3
4
5
6
7
8
9
10
11
12
21
22
23
24
31
32
33
34
35
41
42
43
44
51
52
53
54
NHa to
Boiler
(ppm)
0
0
0
0
0
0
0
0
0
0
0
0
200
200
200
200
300
300
300
300
300
400
400
400
400
600
600
600
600
Boiler
Outlet (ppm)
NOX NHa
200
200
200
200
200
200
200
200
200
200
200
200
120
120
120
120
100
100
100
100
100
90
90
90
90
80
80
80
80
0
0
0
0
0
0
0
0
0
0
0
0
30
30
30
30
40
40
40
40
40
50
50
50
50
80
80
80
80
NHs to
Reactor
(ppm)
210
190
180
195
170
185
160
175
140
180
155
135
0
0
50
50
0
0
0
30
35
0
0
0
20
0
0
0
0
Catalyst
(Relative
Amount)
100
100
90
80
80
70
70
60
60
50
50
40
0
20
30
50
0
30
50
30
50
0
20
40
50
0
20
40
70
Reactor
Outlet (ppm)
NOx NHs
10
20
30
20
40
30
50
40
70
40
60
80
120
100
60
50
100
70
60
50
40
90
60
50
30
80
30
20
10
10
5
5
10
5
10
5
10
5
15
10
10
30
10
15
10
40
10
5
15
10
50
20
10
10
80
30
20
10
NOX
Removal
(%>
95
90
85
90
80
85
75
30
65
80
70
60
40
50
70
75
50
65
70
75
80
55
70
75
85
60
80
90
95
348
-------
catalyst are used (Nos. 35, 44, and 54). However, NHs consumption is high
and the catalyst may cause too great a pressure drop to be used in the
boiler duct without increasing the fan capacity. If a separate reactor or
an additional fan is needed, SCR only may be preferred to the combined
system.
5.8 EVALUATION
SNR is the simplest method for removing NOX from flue gas and is useful
for NOX removal up to about 50 percent. It has, however, the following
disadvantages:
1) For many existing boilers, space for ammonia injection is not
available in a suitable temperature range.
2) SNR has a limited NOX removal efficiency particularly for large
boilers due to the difficulty of rapid uniform gas mixing and
the limited reaction time in the suitable temperature range.
3) Both consumption and NHa leakage are high.
The practical NOx removal efficiency for utility boilers may be about
40 percent when 20 ppm of NHa leakage is allowed and 30 percent when NHs
leakage is kept at about 10 ppm. With flue gas containing a considerable
amount of SOx, the problem of fouling the air preheater with ammonium
bisulfate may be serious when NHa leakage is high.
To obtain 50-70 percent NOX removal efficiency with 10 ppm of NHs
leakage, a small amount of catalyst should be used at 300-400°C. The
combination system may be useful for boilers which have no room for a
separate SCR reactor and have a small excess fan capacity (below about 30
rnmHaO).
349
-------
SNR may give a better than average NOX removal efficiency for some
furnaces in which flue gas has a relatively large retention time at 850-
1000°C.
Because of the low investment cost, SNR may be particularly suitable
for 40-50 percent NOX reduction for emergency, as has been needed in some
districts in Japan.
350
-------
5.9 REFERENCES
1. Exxon, patent applied for.
2. Kasaoka, S., Kagakukojo, 20-7, 1975 and Nenryokyokaishi, 56-3, 1977.
In Japanese.
3. Sato, M., &t at. CRIEPI report No. 27626, 1977. In Japanese.
4. Operating Experiences of High-Temperature Noncatalytic De-N0x Process
at No. 2 Unit of Chita Thermal Power Station. Central Technical
Research Lab., Chubu Electric. Report for IERE, 1978. In English.
5. Noncatalytic NOX Reduction Process Applied to Large Utility Boiler.
MHI Report. November 1977. In English.
351
-------
SECTION 6
OTHER DRY PROCESSES
6.1 INTRODUCTION
The following three dry simultaneous SOX and NOX removal processes have
been developed:
1) Activated carbon process in which ammonia reduces NOX
to Na,
2) Copper oxide process in which ammonia reduces NOx to Na,
3) Electron beam radiation process in which ammonia is
used to produce ammonium sulfate and nitrate.
With processes 1) and 2), SOX is absorbed by either activated carbon
containing a metal catalyst or copper oxide supported by an alumina carrier.
Sulfuric acid or sulfate is formed, while NOX is reduced to Nz by ammonia.
In both processes, concentrated SOz is recovered by heating the carbon or
by reducing copper sulfate with hydrogen.
With process 3), flue gas is mixed with ammonia and exposed to an
electron beam. Powdery ammonium sulfate and nitrate are obtained as by-
products.
The reaction temperature is 220-250°C for 1), 350-400°C for 2, and
about 100°C for 3). The advantage of these processes is that they remove
352
-------
SOX and NOX simultaneously in a dry state and require no reheating of the
treated gas. On the other hand, process 1) requires a large amount of
carbon, while 2) requires hydrogen for regeneration and 3) requires a high-
intensity electron beam.
Processes 1 and 3 have been tested in pilot plants which can treat
1000-5000 Nm3/hr. Process 2 has been applied commercially to the flue gas
(120,000 Nm3/hr) from an industrial boiler. All three processes can remove
80-90 percent of SOX and NOX.
SOX and NOX can also be removed by a combined system of SCR and FGD.
The combination of wet FGD followed by SCR is expensive; it requires a large
amount of energy for heating the gas after FGD. Recent improvements of SCR
catalysts have made it possible to use SCR first on dirty gas without
cleaning, and then use FGD. The combination seems less expensive although
there is no large commercial plant yet using it. Operation of large plants
with this combination system and the dry simultaneous removal process will
determine which of the two is better.
Catalytic decomposition of NOX to Na and QZ has been studied by a few
companies but so far does not seem practical because of the low decomposi-
tion ratio and the short life of catalysts.
As a dry process for NOX removal, the molecular sieve adsorption-
desorption process has been used for treating the tail gas from a nitric
acid plant of Nissan Chemical.
6.2 REACTION OF ACTIVATED CARBON WITH NOX
6.2.1 Introduction
Activated carbon has the following functions useful for removing NOX:
353
-------
1) Adsorption and desorption - It adsorbs N02 below 100°C and
desorbs it above 100°C.
2) Oxidation - It promotes the oxidation of NO to N02 below 100°C.
3) Catalytic reduction - Above 100°C, it promotes the following
reactions:
6NO + 4NH3 > 5N2 + 6H20 (6-1)
4NO + 4NH3 + 02 >• 4N2 + 6H20 (6-2)
6N02 + 8NH3 > 7N2 + 12H20 (6-3)
2NO + 2CO >• N2 + 2C02 (6-4)
2NO + 2H2 > N2 + 2H20 (6-5)
2NO + C > N2 + C02 (6-6)
N02 + 2CO > % N2 + 2C02 (6-7)
The N02 adsorbing capacity is not large enough to treat a large amount
of gas, as will be explained below. Carbon has been used in several small
wet-process NOX removal plants as the catalyst for oxidation of NO to N02.
The use is also limited to small amounts of gas at low temperatures.
For treating a large amount of flue gas above 100°C, carbon is useful
as an NOx reduction catalyst. Addition of metal compounds increases the
activity. Although the efficiency is less than that of SCR catalysts
(Chapters 3 and 4), carbon can remove SOX simultaneously.
354
-------
6.2.2 NOx Adsorption and Desorption
Table 6-1 shows the NOX adsorption capacity of activated carbon under
different conditions. NOz is adsorbed by carbon fairly quickly but NO is
not. NO, however, is oxidized to NOz in the presence of Oz by the catalytic
reaction of carbon and the resulting NOz is adsorbed. The adsorption
capacity is fairly large at low temperature and humidity. Since flue gases
from boilers and furnaces have higher temperatures and humidities, the
carbon may not be useful as an NOX adsorbing agent for those gases.
TABLE 6-1. NOX ADSORPTION CAPACITY OF ACTIVATED CARBON
(mg N0x/g activated carbon)
Relative Humidity, percent
14
20
35
40
60
75
Adsorption Temperature
30°C 50°C
120.0
110.2
104.0 95.0
75.4
67.7
70°C
56.8
36.8
NOX adsorbed by carbon is desorbed either by washing with water or by
heating. Water-washing produces a dilute nitric acid. Higher temperatures
are more favorable for the washing, as shown in Figure 6-1. The wet carbon
has a poor NOX adsorption ability and should be dried before reuse for
adsorption. However, wet carbon is effective for SOX adsorption.
Figure 6-2 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 6-8 to form NO
355
-------
o
CD
1.0
x 0.8
0.6
5 0.4
O
O
0.2
70°C
20 40
WASHING TIME, rain
60
Figure 6-1. Desorption. of NOX adsorbed by activated carbon (6 iran in
diameter) by washing with water at different temperatures,
1.0
0.8
ox
z 0.6
CD
*•«
~ 0.4
o
2 0.2
g
0 10 20 30 40
HEATING TIME, m1n
Figure 6-2. Desorption of NOX adsorbed by activated carbon
by heating at different temperatures.
356
-------
and C02. When the gas is heated above 450°C, a larger amount of carbon is
consumed by reaction 6-9 to form N2 and C02.
2N02 + C >• 2NO + C02 (6-8)
2N02 + 2C >• N2 + 2C02 (6-9)
Heating the carbon in a stream of reducing gas lowers the carbon con-
sumption substantially and increases the conversion of NOx to N2 (Table
6-2).l
TABLE 6-2. DESORPTION OF NOx IN REDUCING GAS
Conversion Ratio of
Adsorbed NOX into N2
Gas Temperature, on Desorption,
°C percent
H2
CO
NH3
He
(Inert Gas)
200
400
600
200
400
600
200
400
600
200
400
600
10
68
100
12
70
100
100
100
100
0
5
70
357
-------
6.2.3 Catalysts for Ammonia Reduction
Among the reducing gases mentioned previously, ammonia is the most
useful for flue gas treatment. Unlike other reducing gases, ammonia reacts
with NOX selectively and is consumed only slightly by Oa in the flue gas.
The effect of carbon on the reaction between NOX and NHs is further in-
creased by adding a small amount of base metal compounds (Table 6-3).l
Copper and vanadium have been found most effective. Takeda Chemical has
commercially produced activated carbon catalysts with metallic components.
TABLE 6-3. EFFECT OF ADDITION OF BASE METAL COMPOUNDS TO CARBON ON NOx
REDUCTION EFFICIENCY (NO, 2000 ppm; NH3, 3000 ppm; SV,
3000 hr"1)
NOx
Metal 110°C
None 38
Ti
Cr 55
Mn 50
Fe 52
Co 63
Ni
Cu 91
V 80
Mo
W
Reduction Efficiency,
150°C
44
65
70
67
67
75
67
99
88
70
65
percent
250°C
78
95
88
90
98
100
100
The efficiency shown in Table 6-3 is obtained at an SV of 3000 hr"1.
An SV of about 1000 hr"1 has been used commercially, considerably less than
the 5000-10,000 used in SCR (Chapters 3 and 4). It is possible to increase
the reactivity of carbon by adding larger amounts of metal components.
358
-------
However, because of local overheating from the heat of reaction, this
addition can result in the combustion of carbon.
Hitachi Ltd. has found that activated carbon treated with ammonium
bromide is effective even at 100°C for NOx reduction by ammonia.2 The low-
temperature activity is useful for energy savings, but at the low tempera-
ture ammonium sulfate and bisulfate will deposit on the carbon.
6.2.4 Simultaneous Removal of SOx and NOx
Activated carbon has been used commercially as an adsorbent for SOz .
As shown in Figure 6-3, a lower temperature is suitable for SO 2 adsorption
while a higher temperature favors NOX removal by selective catalytic
reduction. The optimum temperature for simultaneous removal of SOX and NOX
is 220-230°C. NOX is converted to Na while SOa is adsorbed by the carbon
to form HaSOit and NH^HSO^. The ratio of NHi^HSOit to HaSCH increases as the
temperature decreases. (NHOaSOi, also forms at lower temperatures.
S02 + % 02 + H20 - > HaSO^ (6-10)
HaSOi, + NH3 - > NH^HSOi* (6-11)
+ 2NH3 - > (NHOaSOi* (6-12)
The formation of NH^HSOi* and (NHOaSOi* not only increases ammonia consump-
tion but also lowers catalyst activity.
The EzSOn, NHitHSOif, and (NHOaSOi* contained in the carbon can be
removed by water-washing resulting in a dilute solution of the compounds.
Although the solution can be used for fertilizer production, a large amount
of energy is required to concentrate it.
359
-------
100
90
80
70
100 150 200 250
Temperature (°C)
300
350
Figure 6-3. Removal efficiencies of activated carbon.
Thermal regeneration has been applied to the processes tested in pilot
plants. The carbon, which contains sulfuric acid and the sulfates, is
heated at about 350°C in an inert gas to remove those compounds by the
following reactions:
+ C
S02 + CO + H20
(6-13)
2H2SO
-------
The activated carbon used in Japan for SOx and NOx removal is of high
quality and free from the burning problem. Low cost carbon usually contains
a considerable amount of ash containing alkali and alkali earth elements
which promote burning. To produce high quality carbon, the ash components
are reduced to less than 1 percent. However, it costs nearly 2 million
yen/ton ($10,000/ton) while the carbon for SOz removal costs 800,000 yen/
ton ($4000/ton). About 1000 tons of carbon are required for treating flue
gas from a 300 MW boiler. The cost of carbon for simultaneous removal is
estimated to be lowered to 1,250,000 yen/ton ($6250/ton) by mass production
if demand is sufficient. Further cost reduction is desired before the
process can be applied commercially.
6.3 SUMITOMO HEAVY INDUSTRIES CARBON PROCESS FOR SIMULTANEOUS REMOVAL
6.3.1 Introduction
In 1971, Sumitomo Heavy Industries (SHI, formerly Sumitomo Ship-
building), one of the largest shipbuilding and plant constructing companies,
constructed a prototype FGD plant which could treat 175,000 Nm3/hr of flue
gas from an oil-fired boiler at Sakaiko Station, Kansai Electric. The plant
used an activated carbon process with moving beds and thermal regeneration.
The plant was operated well for about 4 years but the operation was given
up because of the high cost of carbon consumption.
Using the moving bed technology, SHI has developed a simultaneous SOx
and NOX removal process and constructed a pilot plant which can treat 2000
Nm3/hr of the flue gas from an iron-ore sintering machine at the Wakayama
Works, Sumitomo Metal. The plant was constructed with funds of the Japan
Steel Federation and was operated in 1976. SHI also constructed and
operated a smaller pilot plant (500 Nm3/hr) for the selective catalytic
reduction of a gas from an iron-ore sintering machine. This plant had a
similar moving bed.
361
-------
6.3.2 Operation of the 2000 Nm3/hr Pilot Plant3
After passing through an electrostatic precipitator, the flue gas from
the sintering machine of the 2000 Nm3/hr pilot plant contains 150-200 ppm
NOX, 150-250 ppm S02, 12-15 percent 02, and 80-150 mg/Nm3 of particulates
at about 150°C. In some of the tests, S02 was added to the gas to increase
the concentration to 350-480 ppm.
A flowsheet of the pilot plant is shown in Figure 6-4. The carbon is
cylindrical with a diameter of 4 mm, a height of 6 mm, and a crushing
strength of 10 kg/granule. The gas is heated to 200-230°C by an oil-fired
furnace, ammonia is injected and introduced in a moving bed reactor. The
carbon catalyst moves slowly downward in the reactor while the gas is passed
through the carbon bed in a cross flow. About 90 percent of the S02, 85
percent of the NOx, and 75 percent of the particulates were removed with an
NHs leakage of 5 ppm and a pressure drop of 200 mmH20 under the standard
conditions shown below:
Reaction Temperature 220°C
SV 1000 hr"
NH3 370 ppm
The carbon leaving the moving bed reactor contains H2SOit and
as was described in Section 6.2.4 (reactions 6-10 - 6-12). The carbon is
discharged from the reactor, screened to remove dust, and sent to a
regenerator which is also a moving bed reactor. The carbon again moves
slowly downwards while a hot inert gas at 350°C is passed through the carbon
bed in a parallel flow to decompose the H2SOit and NHitHSOi* and to purge
concentrated S02 by reactions 6-13 - 6-16 presented in Section 6.2.4. The
inert gas is produced by incomplete combustion of propane and contains
1-1.5 percent CO, 9.5-10 percent C02, about 0.8 percent H2, and 0.5 percent
02. The purged concentrated S02 can be used to produce sulfuric acid or
elemental sulfur. The regenerated carbon is returned to the first reactor.
362
-------
FLUE GAS
CO
o>
CO
AIR
NH3 VAPORIZER
70-1349-1
Figure 6-4. Flowsheet of a 2000 Nm3/hr pilot plant using the SHI carbon process
for simultaneous removal of SOX and NOX.
-------
The result of a 2300-hour continuous test is shown in Figure 6-5. The
test was carried out under the above-mentioned standard conditions except
for the temperature fluctuation. At 210-230°C, 90-97 percent of the S02 and
85-90 percent of the NOx were removed. At the beginning and the end of the
test, the temperature was low (180-190°C) and the NOx removal efficiency was
low (40-50 percent), while SOz removal efficiency was maintained above 95
percent. Pressure drop of the gas was maintained between 180 and 210 mmHzO
throughout the test.
Tests at different conditions are shown in Figures 6-6 - 6-9. The
reactor temperature of 220°C was required for 85 percent NOX removal and
230°C was required for nearly 90 percent removal at an SV of 1000 hr"1
(Figures 6-6 and 6-7). Increasing the amount of ammonia raised not only NOx
removal efficiency but also SOa removal efficiency because much of the SOz
is removed by the formation of NHitHSOii and ammonia leakage increased
(Figure 6-8). On the other hand, an increase in the SOa concentration at a
fixed NHa concentration resulted in a slight decrease in the removal
efficiencies (Figure 6-9). The standard NHs concentration (370 ppm) is
approximately 1 mole NHg to 1 mole NOx plus 1 mole NHs to 1 mole S02.
Based on the 2300 hours of operation, the carbon consumption is
estimated to be 113 percent for a year (8320 hours of operation) with an
SOa concentration of 150-250 ppm. The NOX and SOX removal efficiencies
were maintained by replacing the carbon at this rate. The plant operation
has been virtually trouble-free.
6.3.3 Cost Estimation
Based on the operation of pilot plants, SHI has estimated the cost for
the simultaneous removal of SOx and NOx and also for NOX removal by SCR
with a ferrite catalyst. These estimates are for flue gas from a 500 MW
coal-fired boiler containing 500 ppm SOa and 600 ppm NOX (Table 6-4).
364
-------
LOW TEMPERATURE TEST
UJ
mn
"-" 80-
$ 60-
o
ui 40
Of.
•z.
°0
TEMPERATURE
100
S e°"
^ 4°
§ 40-
UI
* 20-
CM J
o 0
00 "(
r- -i
..,»-* *«-.<->..« ^-* t t • . . • • j«^» • • » -• /*~*
^-_~_^_^^.
y^ v
» V
2do ' 400 600 ' 800 1/WO ' 1.2^0 ' 1.4bo ' 1,600 ' 1.800 ' 2J030 ' 2^3 ' 2.400
190 1 1
"*" 1 1
s^^—
i 200 ' 4o'o ' 6&o eio i.obo ' ubo ' i.4bo ' i.<{oo ' i.ebo 2^00 2^00 ' 2.400
TEST PERIOD (hr)
Figure 6-5. Continuous operation tests of the SHI carbon process
for simultaneous removal of SOX and NOX.
-------
100
80
; eo
40-
20-
180 190 200 210 220 230
Reaction Temperature (°C)
Figure 6-6. Effect of temperature on simultaneous removal of SOX and NOX
with carbon (S02, 150-250 ppm; NOX, 150-200 ppm; NHs, 370
ppm; SV = 1000 hr"1).
100
80-
60-
•
40-
20-
n
-d D D—
— T— -^
D S02
o NOX
500 1000 1500 2000 2500
Space Velocity (hr"a)
Figure 6-7- Effect of space velocity on simultaneous removal of SOx and
NOX with carbon (S02, 150-200 ppm; NOX, 150-200 ppm; NHs,
370 ppm; 220°C).
366
-------
100
•
80-
£ 60-
r—
I 40
° SO,.
oNOx
ANH9
300 400 500 600
NH3 (ppm)
Figure 6-8. Effect of NHs added on simultaneous removal of SOx and NOx
with carbon (NOx, 150-200 ppm; SOz , 150-250 ppm; SV,
1000 hr"1; 220°C) .
IUU
80-
* 60 "
*g
| 40"
20-
n
— n .,n
D S02
o NOX
0 100 200 30d 400 500
S02 (ppm)
Figure 6-9. Effect of SOz on simultaneous removal of SOx and NOx with
carbon (NOX, 150-200 ppm; NHs, 370 ppm; SV, 1000 hr"1;
220°C).
367
-------
TABLE 6-4. ESTIMATED COST FOR THE USE OF A FERRITE CATALYST FOR
SIMULTANEOUS REMOVAL OF SOX AND NOX VERSUS THE USE OF
SCR FOR THE FLUE GAS FROM A 500 MW COAL-FIRED BOILER*
Reaction Temperature
SOa Removal (percent)
NOX Removal (percent)
Space Velocity (hr"1)
Catalyst Charge (tons)
Catalyst Makeup (tons/year)
Electric Power (kW)
NH3 (kg/hr)
LPG (kg/hr)
Steam (tons/hr)
Cooling Water (m3/hr)
Maintenance Cost
Operator
Costs (millions of yen, $5000)
Plant Cost Including Site
Preparation and Control Room
Initial Catalyst Charge
Total
Simultaneous Removal
(Activated Carbon)
220°C
95
90
1000
1560
3000
3100
1050
470
1.0
5.0
2 percent/year
of plant cost
2 men/ shift
5045
1930
6975
SCR
350°C
—
90
5000
354
295
3300
760
—
0.6
1.0
2 percent/year
of plant cost
2 men/shift
3033
867
3900
* 1,845,000 Nm3/hr of gas containing 500 ppm S02 and 600 ppm NOx.
368
-------
The investment cost for SCR is 7800 yen/kW ($39/kW), but the cost for
the simultaneous removal is 13,950 yen/kW ($69.75/kW). Simultaneous removal
is considerably less costly than the combination of SCR and conventional
FGD; a plant by the combined system usually costs about 20,000 yen/kW
($100/kW).
Three thousand tons of carbon catalyst would be consumed per year at
a cost of 3.75 billion yen ($18.75 million) per year, assuming a catalyst
cost of 1.25 million yen/ton ($6250/ton). For 7500 hours of operation per
year, the cost of the carbon would be 1 yen/kWhr ($0.005/kWhr), while
depreciation of the plant cost for 7 years would be approximately 0.2 yen/
kWhr ($0.001/kWhr).
6.3.4 Evaluat ion
The simultaneous removal process is a simple modification of the
activated carbon FGD process with thermal regeneration. The FGD process is
fairly costly but the simultaneous removal process looks fairly promising
because NOx is removed by simple addition of ammonia. The greatest dis-
advantage of simultaneous removal is the enormous consumption of carbon,
which is more expensive than the carbon for FGD. Since the consumption of
carbon as well as that of inert gas and ammonia increases with the 862
content of the gas, the process should be used for flue gas containing a
relatively small amount of SOa, for example, a gas from burning a low-
sulfur fuel. This gas would contain less than about 400 ppm of SOx but
would still require desulfurization. Even in this case, reduction of the
carbon cost is desired.
The reactor can treat gases with particulate contents of up to about
200 mg/Nm3. For flue gas from a coal-fired boiler, an electrostatic pre-
cipitator is needed ahead of the reactor to treat either the gas leaving the
boiler economizer at about 400°C or the gas leaving an air preheater at
220-230°C. An additional air preheater may be needed after the reactor to
369
-------
recover the heat of the reactor effluent. There seems to be little problem
with ammonium bisulfate in the air preheater after the reactor because more
than 95 percent of the SOx is caught by the carbon and the ammonia leakage
is low.
6.4 UNITIKA ACTIVATED CARBON PROCESS FOR SIMULTANEOUS REMOVAL
6.4.1 Introduction
The engineering division of Unitika Co., one of the largest textile
companies in Japan, has developed a dry simultaneous removal process for SOx
and NOX which uses activated carbon. Since November 1975, they have oper-
ated a pilot plant which can treat 4500 Nm3 /hr of the flue gas from a glass
melting furnace.
6.4.2 Process Description
The gas from the glass melting furnace contains a considerable amount
of sodium sulfate vapor which contaminates the carbon catalyst. Therefore,
the flue gas is first washed with a water spray to remove most of the
sodium sulfate. The gas cooled to 60°C contains about 300 ppm S02, 200 ppm
NOX, and 5-10 mg/Nm3 particulates.
The gas is heated to 250° C, injected with ammonia, and introduced into
an adsorption tower (Figure 6-10) . The adsorption tower has four compart-
ments (reactors), each of which has a fixed carbon bed. One bed is re-
generated while the other three operate in series to remove SOa and NOx .
About 400 ppm NHs is added to the gas at about 230°C and the gas mixture is
led into three compartments. About 90 percent of the NOX and SOa is
removed .
The carbon that has absorbed SOz is heated to 350°C by an inert gas
produced by incomplete combustion of propane. This gas contains 1-1.5
370
-------
INERT GAS
GAS COOLER
\
FLUE GAS
60«C
HEATER
250°C
FAN
300-350*C
NH
H3
-------
percent CO, 9.5-10 percent COz, about 0.8 percent Ha and 0.5 percent Oa«
The recovered concentrated SOa gas can be used for sulfuric acid production
or other purposes. Ammonium sulfate and bisulfate which tend to form on
the carbon are decomposed to SOz and N2 during regeneration.
The activated carbon, produced by Takeda Chemical for simultaneous
removal, is in the form of cylinders with a 5 mm diameter and 8-10 mm
length. Carbon consumption is estimated at about 50 percent of the charge
per year.
The main design and operating parameters are as follows:
Tower Height 17 m
Carbon Bed Thickness 1 m
Pressure Drop 100 mmHaO (reactor)
250 mmH20 (total)
Absorption Time for One Cycle 3 days
Regeneration Time for One Cycle 12 hours
SOz in Gas From Regeneration Step 5-10 percent
Space Velocity (SV) 700-1500
Superficial Gas Velocity 0.1-0.2 m/sec
At an early stage of operation, the reaction temperature was about
180°C. At that temperature a large amount of ammonium bisulfate and sulfate
formed on the carbon and caused plugging. Use of a 230°C temperature solved
the problem.
6.4.3 Cost Estimation
The investment cost includes the initial charge of carbon (1.25
million yen/ton, $6250/ton), and civil engineering work. The plant cost
for the 500,000 Nm3/hr plant is equivalent to 18,000 yen/kW ($90/kW).
372
-------
The operating cost was calculated for an industrial boiler with a
flue gas temperature of 200°C operated for 95 percent of the total hours.
The cost includes 7 years of depreciation and 10 percent interest.
6.4.4 Evaluat ion
The Unitika process is similar to the SHI process (Section 6.3) except
that it uses a fixed bed. For flue gas containing little particulate
matter, a fixed bed may be preferred because of the smaller consumption of
carbon. The estimated yearly carbon consumption, 50 percent of initial
charge for flue gas containing 1000 ppm of SOX, is small compared with
that by SHI (previous section), 113 percent of the charge for flue gas
containing 150-250 ppm of SOX. The estimate might be a little optimistic.
Both Unitika and Sumitomo seem to use similar qualities of carbon and inert
gas, and the difference in the carbon consumption may be too big to be
attributed to the reactor types. Further studies may be needed on carbon
consumpt ion.
The process requires a considerable amount of fuel for heating the
flue gas from 200 to 250°C. For the flue gas from utility boilers which
normally is 150°C, the use of fuel to heat it to 250°C is too costly. Two
air preheaters may be needed for a utility boiler; one between the boiler
economizer and the reactor to cool the gas to 220-230°C and the other after
the reactor for further heat recovery.
6.5 SHELL COPPER OXIDE PROCESS (SFGD PROCESS) FOR SIMULTANEOUS REMOVAL
6.5.1 Process Description**
The Shell process uses copper oxide supported on granular alumina in
two parallel-passage reactors. The process was originally used only for
SOX removal. The reactions involved in the process are shown be-low:
373
-------
TABLE 6-5. ESTIMATED COST FOR THE UNITIKA PROCESS FOR SIMULTANEOUS
REMOVAL OF 1000 PPM SOX AND 300 PPM NOX BY ACTIVATED
CARBON
Capacity. 1000 Nm3/hr
30
100
500
Investment Cost (10s yen, $5000)
Major Units
Accessories
HzSOit Plant
Total
Operating Cost (yen/kl oil, $0.005/kl)
Depreciation and Interest
Electric Power
Fuels*
Ammonia
Other**
Total (yen/kl oil, $0.005/kl)
(yen/kWhr, $0.005/kWhr)
248
41
111
400
3700
650
2500
1200
1500
9550
2.12
587
87
226
900
2890
650
2500
1200
1400
8640
1.92
2048
252
700
3000
2440
650
2500
1200
1400
8190
1.82
* Kerosene for gas heating from 200 to 250°C and for heating for
regeneration, and LPG to produce inert gas.
** Mostly for carbon.
374
-------
CuO + SOa + Jg 02 " CuSOit (acceptance) (6-17)
CuSOit + 2H2 » Cu + S02 + 2H20 (regeneration) (6-18)
Cu + % 02 « CuO (oxidation) (6-19)
Flue gas at 400°C is introduced into one of the reactors, where SOx
reacts with copper oxide to form copper sulfate (reaction 6-17). Then
hydrogen is introduced into the reactor to convert the copper sulfate to
metallic copper and concentrated SO2 which is used for sulfur or sulfuric
acid production (reaction 6-18).
The two reactors are used alternately for acceptance and regeneration.
Between the acceptance and regeneration, steam is injected to the reactor
to purge the remaining flue gas or hydrogen. This eliminates any possi-
bility of explosion. At the beginning of the acceptance flow, copper is
oxidized by the oxygen in the flue gas and then reacts with SOa (reaction
6-19).
The Yokkaichi plant, Showa Yokkaichi Seikiyu (SYS), has treated flue
gas (120,000 Nm3/hr) from an oil-fired boiler since 1973. The plant reduces
SOx from about 1200 to 120 ppm. Since 1975, ammonia has been introduced
into the reactor during the acceptance period to reduce NOX. The process
utilizes CuO and CuSOif as catalysts for the following selective reduction:
4NO + 4NHs + 02 • 4Na + 6H20 (6-20)
Figure 6-11 shows the relation between NO* in the reactor and the
loading (a)—the mole ratio of CuSO(t/(CuO + CuSOO—-when different amounts
of NHs were used. The NOx in the reactor effluent decreases (NOX removal
efficiency increases) as CuSOi* increases, indicating that CuSOi* works
better than CuO as a catalyst.
375
-------
100
80
60
SO
40
30
20
*
0* 10
8
6
5
4
3
2
1
"**J
*
3
^>»i
V^
SvS
gsjv
w
"*«-
w
Sfc^
OJyV
\\
^
CONDITIO
BED LI
-•"
•*«te:
Nn
^^i
Sx
\'
V
N
MS:
:NGTH
mft_
\
X
\
X
%
t
\
\
1
Im.
TEMPERATURE 400 °C
REYNOLDS NUMBER 2000
+ NORMAL SYS OPERATION
NH3
"MOI
~ ^
~ E
0
\
\
/NO
.AH RATIO ~
0.3
0.6
0.8
0.9
1.0
\
IwC
\
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.
LOADING I Q I
Figure 6-11.
NOX slip, percent of intake, versus loading,
CuSOit/(CuO + CuSOit) mole ratio, for a
commercial SFGD acceptor/catalyst.
376
-------
Figure 6-12 shows the relationship of NOx conversion to NHs/NO mole
ratio at 0.4 mole SOa/mole Cu (0.4 mole CuSCH 4- 0.6 mole CuO). The con-
version rate reached 70 percent when 1 mole of NHs was used for each mole
of NO.
Figure 6-13 shows the relation between the catalyst bed length (depth)
and NOx slip when CuSOi» and a sufficient amount of NHa (1.1-1.5 mole NHs
for each mole of NO) were used. A bed 4 meters deep can remove over 95
percent of the NO under such conditions.
Performance of the commercial SYS reactor is shown in Figure 6-14. At
the beginning of acceptance, a considerable portion of copper is present as
metal. This metal is not only inefficient for SOx removal but also tends
to oxidize NHs to NOX. Cu is rapidly oxidized to CuO which readily reacts
with SOX. Ammonia is injected after Cu is converted to CuO.
The S02 removal efficiency reaches a maximum in about^ZO minutes and
then decreases as most of the CuO is converted to CuSOi*. NOX removal
efficiency continues to increase because CuSOij is a better catalyst than
CuO. With an acceptance time of about 120 minutes, the overall removal
efficiency reaches 88 percent for SOx, 50 percent for NOx with an NHs/NO
mole ratio of 0.71, and 65 percent for NOx with a NHs/NO mole ratio of 0.99.
The Yokkaichi paint has removed about 90 percent of S02 and 40 percent
of NOX to meet local regulations. Leak ammonia present in the treated gas
is only about 1 ppm at 40 percent removal. Operability of the plant which
was not high for the first few years has been increased through improve-
ments. Continuous operation for nearly one year was recently achieved.
6.5.2 Evaluation
The process is a desirable dry simultaneous removal process although
it consumes considerable amounts of hydrogen and steam. Nearly 70 percent
377
-------
§
oc
U!
o
u
lx
1UU
an
yu
on
ou
in
/u
an
tv
Q
£
f\
A
/
/
g
t
i
t
1
£*
• o o />
/.•
'
o
f>
1.5 METER BENCH S
2000 SO2
4 % O2
~GL~ 0.4 mol SO2 / m
T(°C) GHSV
O 350 8000
4UU 4UUU
O 400 8000
A 450 8000
CALE UNIT
10! Cu
1.0 2.0
NH3/NO MOLE RATIO
Figure 6-12.
NOX conversion versus NHs/NO mole ratio for a
commercial SFGD acceptor (SYS).
378
-------
100
10
*
»
J
\
\
\
^
N
x
\
\
\
^
1
\
\
\
CONO
anno
CuA
NH3
PERf
• PERf
AFFf
OUT
1
>
N
l\
\
\
ITIONS:
C
SCuSO4
/NO 1.1 ~1.S
MAL EXPEC
ORMANCE
ORMANCE
CTEO BY
SIDE FACTC
\
f\
\
new
IRS
*,
\
3 4
BED LENGTH, (METERS
Figure 6-13.
Unconverted NOx as a function of catalyst bed length
for a commercial SFGD acceptor (SYS) - upper and
lower limits.
379
-------
480
i i
AT REACTOR INLET1.
FLOW • 137,000 NmVh
SOt * 1260 ppmv
NOx • 293 ppmv
REACTOR BED LENGTH • 4 METER
AT NH. /NO-0.0
20 40 60 60 100 120
ACCEPTANCE TIME, MlN
Figure 6-14. Performance of Shell FGD reactor at SYS,
instantaneous S02 and NOx slip (92).
380
-------
NOX removal with about 90 percent SOa removal is achieved by using an NHs/
NO mole ratio of 0.99. To obtain a higher NOX removal efficiency, a larger
reactor may be needed: a prolonged acceptance time increases the NOX
removal efficiency but lowers the SOX removal efficiency, and the use of
excessive ammonia would increase the ammonia leakage.
The process may be suitable for flue gas with a moderate SOz concen-
tration. An SOa-rich gas would require frequent switching between accep-
tance and regeneration which would cause the consumption of a large amount
of hydrogen and steam and a decrease in NOX removal efficiency.
The parallel-passage reactor is virtually free from dust plugging
and may be useful for the flue gas from coal firing as well as from oil
firing. The U.S. EPA recently awarded a contract for pilot scale (approxi-
mately 0.5 MW) studies of the Shell simultaneous removal process. Universal
Oil Products (UOP) will operate the test unit which is to be sited at a
coal-fired boiler at Tampa Electric Power Company's Big Bend Station in
Florida.
6.6 EBARA ELECTRON BEAM RADIATION PROCESS FOR SIMULTANEOUS REMOVAL
q
6.6.1 Process Description
A process for simultaneous removal of SOx and NOX with the use of an
electron beam has been developed by Ebara Manufacturing Co. jointly with
the Japan Atomic Energy Research Institute. It has been tested by Nippon
Steel in a pilot plant which can treat 3000 Nm3/hr flue gas. A flowsheet
of the 1000 Nm3/hr test plant of Ebara Manufacturing Co. is shown in
Figure 6-15. The larger pilot plant of Nippon Steel went into operation
in late 1977 to treat flue gas from an iron-ore sintering machine.
A flue gas at about 100°C to which a small amount of ammonia has been
added is exposed to electron beam radiation. A powdery product consisting
381
-------
AMMONIA
COOLER
OD
K>
BOILER
CYCLONE
ESP
I
STACK
ELECTRON
BEAM
ACCELERATOR
REACTOR
ssss/sss/ss
S S S // SS/SSSS/SSSS
Figure 6-15. Flowsheet of the 1000 Ito3/hr electron beam radiation process
pilot plant of Ebara Manufacturing Company.
-------
of ammonium nitrate and sulfate is formed by the radiation in 1-5 seconds
and is caught by an electrostatic precipitator. Although the reaction
mechanism is not clear yet, the OH radical seems to play an important role.
Figure 6-16 shows test results of the process with flue gas at about
90°C containing 240 ppm NOX, 220 ppm SOX, 13 percent C02, 4 percent 02, 9
percent H20 and 0.1-0.2 g/Nm3 particulates to which 700 ppm NH9 was added.
Although nearly 90 percent of the NOX and SOX was removed with an electron
beam of 5 Mrad, about 80 percent removal with about 2 Mrad seems practical.
100
tt 80
8
60
40
I
2 3
Total beam (Mrad)
Figure 6-16,
Removal ratios of SOX and NO* versus total beam for the
electron beam radiation process used at the Ebara
Manufacturing Company pilot plant.
383
-------
6.6.2 Evaluat ion
Energy consumption is claimed to be not more than that of other
simultaneous removal processes but the investment cost seems high. Many
large electron beam generators are needed for a large-scale plant because
the electron beam has a limited range. Another drawback of the process is
that flue gas must be cooled to about 100°C; the removal efficiency is lower
at the usual flue gas temperature of 150°C.
The process may suit gases at moderate flow rates and relatively rich
in NOx and low in S02 because the nitrogen content of the by-product is low
when the gas is rich in S02- The by-product may be useful for use as
fertilizer. When flue gas from a coal-fired boiler is treated, fly ash
should be removed in advance to make the by-product ammonium salts useful
for fertilizer.
6.7 OTHER DRY PROCESSES
6.7.1 Tests on Catalytic Decomposition of NO
The National Research Institute for Pollution and Resources has been
testing catalytic decomposition of NO at 400-700°C jointly with Catalyst and
Chemicals Industries.
Many kinds of catalysts based on aluminum and zeolite impregnated with
alkalis, bases, and noble metals have been tested. Platinum and palladium
(Pd) have shown an appreciable catalytic effect on the decomposition of
NO to Na and 02 as shown below:
384
-------
Decomposition Ratio, percent (SV = 3000)
Pt-Al203 (Pt 0.5 percent)
Pd-Al203 (Pd 0.5 percent)
Pt-Zeolite (Pt 0.5 percent)
Pd-Zeolite (Pd 2.0 percent)
400°C
4.2
0.0
5.1
4.5
500°C
3.7
0.0
5.4
6.1
600°C
9.0
10.0
10.9
11.0
700°C
10.0
27.0
17.6
—
The decomposition ratio, however, was not high. Even for a Pd
-i
catalyst at an SV of 500 hr and a temperature of 700°C, the ratio was only
34 percent. Moreover, the catalysts are poisoned by many of the components
of flue gas.
6.7.2 NOX Recovery by Molecular Sieve
Nissan Chemical has installed an NOX recovery plant with the PuraSiv
N process developed by Union Carbide Corp., U.S. It treats 22,000 Nm3/hr
of flue gas from a nitric acid plant (160 tons HNOs/day) containing 1000
ppm of NOX• The plant was commissioned in May 1976 and has since been
operated successfully. It reduces the NOx concentration to below 50 ppm
by molecular sieve adsorption. The adsorbed NOX is thermally desorbed and
returned to the nitric acid plant. By using the recovery system, the yield
of the nitric acid plant has increased by 1 percent. The process seems
very useful for treating tail gas from nitric acid plants.
385
-------
6.8 REFERENCES
1. Ninomiya, N., Simultaneous Removal of NOx and SOx by Activated Carbon.
Report of Takeda Chemical, 1975. In Japanese.
2. Seki, M., et at., Ammonium Hallide Activated Carbon Catalyst to
Decompose NOx in Stack Gas at Low Temperature Around 100°C. American
Chemical Society, Chicago, August 1975. In English.
3. Present Status of Development of NOx Abatement Technology by the
Steel Industry. Japan Steel Federation, April 1977.
4. Nooy, F. M., and J. B. Pohlenz, Nitrogen Oxides Reduction with the Shell
Flue Gas Desulfurization Process. The Second Pacific Chemical Engineer-
ing Congress, Denver, August 1977. In English.
386
-------
SECTION 7
WET PROCESSES FOR NOX AND SIMULTANEOUS NOX AND SOX REMOVAL
7.1 GENERAL DESCRIPTION
7.1.1 Major Processes and Plants
Major commercial and test plants for NOx and simultaneous NOX and SOX
removal by wet processes are listed in Table 7-1. The plants are relatively
small; the largest one has a capacity of 100,000 Nm3/hr (33 MW equivalent).
The major wet processes may be classified as follows:
1) Oxidation-absorption
2) Absorption-oxidation
3) Oxidation-absorption-reduction (simultaneous removal of
NOX and SOX)
4) Absorption-reduction (simultaneous removal of NOX and SOX)
Essentially all of the NOX in the combustion gas is in the form of NO,
which has poor reactivity and is not readily absorbed by most absorbents.
NO is oxidized in NOa in air, but the oxidation occurs slowly. In
processes (1) and (3) oxidizing agents are used to convert NO to NOa or
NaOs which are more readily absorbed. In the absorption oxidation process,
NO is absorbed in a solution containing an oxidizing agent such as sodium
387
-------
TABLE 7-1. MAJOR PLANTS FOR NOX REMOVAL 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-
Fuj ikasui
Sumitomo Metal-
Fuj ikasui
Sumitomo Metal-
Fuj ikasui
Sumitomo Metal-
Fuj ikasui
Osaka Soda
Shirogane
Chiyoda
Mitsubishi H.I.
Ishikawajiaa H.I.
Kureha Chemical
Chisso Corp.
Mitsui Engineering
Asahi Chemical
Type of
Process
Oxidation
absorption
Oxidation
absorption
Oxidation
absorption
Absorption
oxidation
Absorption
oxidation
Absorption
oxidat ion
Absorption
oxidat ion
Absorption
oxidation
Absorption
oxidation
Oxidat ion
reduction
Oxidat ion
reduct ion
Oxidation
reduction
Oxidation
reduction
Oxidation
reduction
Oxidation
reduction
Oxidation
reduction
Oxidation
reduction
Oxidation
reduction
Absorption
reduction
Absorption
reduction
Absorption
reduction
Absorption
reduction
Plant
Owner
Tokyo Electric
Tokyo Electric
EPDC
Ijima Metal
Nissan Chemical
Mitsubishi Metal
Kobe Steel
Kobe Steel
Hodogaya
Sumitomo Metal
Toshin Steel
Sumitomo Metal
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
Kakogawa
Koriyama
Amagasaki
Fuji
Osaka
Kashima
Amagasaki
Kawasaki
Kawasaki
Hiroshima
Yokahama
Nishiki
Goi
Chiba
Mizushima
Capacity,
Nm3 /hr
2,000a
100,000a
5,000a
1,800
3,000
4,000a
l,000a
50,000a
4,000a
62,000
100,000
39,000
25,000a
60,000
48,000
l,000a
2,000a
5,000a
5,000a
300a
150a
600a
Source
of Gas
Boilerb
Boilerb
Boiler0
Pickling
HNOS Plant
Boilerd
Furnace6
Furnace6
Furnace6
Boilerd
Furnace
Boilerd
Sintering
machine
Boilerd
Boilerd
Boilerd
Boilerd
Boilerd
Boilerd
Boilerd
Boilerd
Boilerd
Completion
Dec
Oct
Dec
July
Mar
Dec
Dec
Mar
Oct
Dec
Dec
Dec
Mar
Aug
Dec
Sept
Apr
1973
1974
1975
1973
1975
1974
1973
1976
1975
1973
1974
1974
1976
1976
1974
1973
1974
1975
1975
1974
1974
1974
By-Product
HNOs
HNOs
Gypsum,
Ca(NOs)a
NaNOa
NaNOs , NO
KNOa
Gypsum, Ha
Gypsum, Na
NaNOs , NaCl
NaNOs, NaCl,
NaaSOi,
NaNOs, NaCl,
NaaSOi.
NaNOs, NaCl,
NaaSOi,
Gypsum, CaCla ,
Ca, NO
NaNOs , NaCl,
NaaSOi,
NazSOi,, NaNOs
Gypsum,
Ca(NOs)a
Gypsum, NHs
Gypsum, Na
Gypsum, N2
(NHOaSOi,
Ha SO*, Na
Gypsum, Na
.Test plants; others are commercial plants.
Gas-fired boiler.
cCoal-fired boiler.
^Oil-fired boiler.
.Iron-ore sintering machine.
Metal-heating furnace.
388
-------
or potassium permanganate, while in the absorption reduction process NO is
absorbed in a liquor containing catalysts, as will be described in Section
7.1.4.
Little progress has been observed with wet processes in the past year.
Many of the pilot plants have discontinued operation for the following
reasons: 1) the processes are not as simple as selective catalytic
reduction, and 2) many of the processes by-produce nitrates or ammonia.
The Japanese fertilizer industry has suffered from excessive ammonia pro-
duction capacity. Moreover, Japanese farmers have used little nitrate
fertilizer because the major crop is paddy rice for which nitrate fertil-^
izers are not suitable.
Some of the wet processes, however, are capable of removing NOX and
SOX simultaneously and converting NOX to nitrogen or ammonia which can be
used for fertilizer in many other countries, in contrast to the dry pro-
cesses which consume a considerable amount of ammonia. Existing FGD plants
can be modified in order to remove NOx simultaneously by adding a few
vessels. It is desirable that the wet processes will be improved to pro-
duce a simpler system and more favorable economics.
7.1.2 Oxidizing Agents
Ozone (Og) and chlorine dioxide (C102) are used mainly for the
oxidation of NO in the gaseous phase. They oxidize NO to N02 within a
second but barely oxidize S02 to 80s . Ozone can oxidize NO to N20s when
added in excess.
NO + Os - >N02 + 02 (7-1)
2NO + 303 - >N205 + 302 (7-2)
Ozone is fairly expensive; it costs 330 to 400 yen per kilogram and
about 0.4 yen/kWh ($1.65 to $2.00/kg, $0.002/kWh) to treat a flue gas from a
389
-------
boiler containing 150 ppm of NO. A large scale ozone generator with a
capacity of 100 kilograms per hour of ozone is near completion, It can
treat about 230,000 Nm3/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 lower
than that of ozone, .but chlorine dioxide introduces hydrochloric and nitric
acids, which complicate the system.
2NO + C102 + H20 — >N02 + HNOj + HC1 (7-3)
Solutions of potassium and sodium permanganates, sodium and calcium hypo-
chlorites, and hydrogen peroxide have been used for oxidation in the liquid
phase, but these chemicals are also expensive.
7.1.3 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 pro-
cesses the NO is absorbed in a solution containing an oxidizing agent.
Usually NOX absorption occurs more slowly in the latter case because NO
must be absorbed in the liquor before it can be oxidized.
Most plants which use nitric acid for such processes as metal washing
emit a gas fairly rich in NOx (1000 to 10,000 ppm). However, the amount of
gas is not great (500 to 5000 Nm3/hr). In many of the plants, all or part
of the NO is oxidized to NOa, and the gas is absorbed in a sodium hydroxide
solution. Activated carbon is used in some of the plants as a catalyst for
NO oxidation by air. In other plants NOX is absorbed in a solution contain-
ing an oxidizing agent such as NaCIO or HaOa. In both cases the resulting
liquor, containing nitrate and nitrite, is sent to a wastewater treatment
system. Some plants use activated sludge treatment to reduce the nitrogen
content of the water. Such processes cannot be applied on a large scale
390
-------
because the treatment does not sufficiently remove nitrogen compounds from
the wastewater.
Tests were made in pilot plants to recover nitric acid for industrial
use (see Section 7.2.2) or to recover potassium or calcium nitrate for
fertilizer (see Sections 7.9.1 and 7.9.2). Those processes do not seem
promising because of the high cost and the limited demand for the by-
products.
7.1.4 Oxidation-Absorption-Reduction and Absorption-Reduction Processes
(Simultaneous Removal)
Since 1973 many oxidation-absorption-reduction and absorption-
reduction processes have been developed in which NOx and SOx are absorbed
simultaneously. In the oxidation-absorption-reduction process, NO is first
oxidized and then absorbed together with SOx by lime/limestone scrubbing.
In the absorption-reduction process NO is absorbed with SOX in a liquor
containing ferrous ion, which can form an adduct with NO. Usually EDTA
(ethylenediamine tetraacetic acid, a chelating compound whose present cost
in Japan is about 800,000 yen/t, $4000/t) is added to promote the reaction
of ferrous ion with NO (Figures 7-1 and 7-2).1
In both cases, various reactions occur in the liquor or slurry and
result in the reduction of NOX by S02 (or sulfite) to NH3 through imido-
disulfonic acid (HN(S03H)2), sulfamic acid (H2NS03H), or a salt of either
acid. The reaction sequence may be expressed in the following way.2
HONO
t
(HO)2NS03H
Reduction
T
>»HONHS03H
I
HON(SOsH)2
H2NS03H
T
HN(S03H)2
N(S03H)3
(7-4)
Hydrolysis
t
391
-------
CO
'o
s
§
2
10 30 50 70
LIQUID TEMPERATURE, °C
Figure 7-1. NO absorption in EDTA-Fe(II)liquor
(0.01 mole/liter).*
* Amount of liquor, 1 liter; gas flow, 1.5 liter/min; pH of liquor, 5.65;
initial NO concentration, 275 ppm.
392
-------
'o 2
ca
ae
o
to
ca
< 1
I
0 0.01 0.02
EDTA-Fe(II), mole/liter
Figure 7-2. EDTA-Fe(II) concentration and NO absorption
at 50°C.*
* Amount of liquor, 1 liter; gas flow, 1.5 liter/min; pH of liquor, 5.65;
initial NO concentration, 275 ppm.
393
-------
Although the chemistry is complex, the overall reactions may be
expressed simply in the following equations:
2N02 + 7S02 + 7CaC03 + 3H20 = JCaSO* + 2NH3 + 7C02 (7-5)
(oxidation reduction process, limestone scrubbing)
2NO + 5S02 + 8$fa + 8H20 = 5(NHit)2SO^ (7-6)
(absorption reduction process, ammonia scrubbing)
The reactions indicate that 3.5 moles of S02 is needed for 1 mole of
N02 and 2.5 moles of S02 is needed for 1 mole of NO. Since flue gas con-
tains 02 which oxidizes a considerable portion of S02 in a scrubber liquor,
4-5 moles of S02 is actually needed to attain 80 percent NOx removal.
A portion of the NOX can be reduced to N2 in an aqueous solution. The
reactions are complex but may be simply described as:
NH2S03H + N02~ - > N2 + HSOiT + H20 (7-7)
Tanaka2 has found that a compound Na2S03*2NO is formed when NO is
absorbed in an NaS03 solution. The compound is stable at a pH above 8,
but it decomposes to form Na2SOif and N20 at lower pH. It is possible that
in addition to N2 or NH3 , N20 also is formed in some of the wet processes.
Na2S03'2NO - >- Na2SO^ + N20 (7-8)
The presence of EDTA and ferrous ion prevents the formation of N20.
In some of the processes a considerable portion of the NOX remains in
the resulting liquor as nitrite and nitrate, which would cause a problem
in wastewater treatment.
394
-------
The advantage of such wet processes over dry processes is that they
can simultaneously remove SOz and NOx without the problems of particulates
and ammonium blsulfate. They have not yet been commercialized on a large
scale. Five relatively small commercial plants are in operation (Table
7-1).
7.2 MHI WET PROCESSES3
7.2.1 Introduction
Mitsubishi Heavy Industries (MHI), the leading FGD process developer
and FGD plant constructor, has developed two wet processes for NOX and
simultaneous SOx and NOx removal, both using ozone as the oxidizing agent.
The NOx removal process uses water absorption to produce nitric acid; it
was tested in cooperation with Tokyo Electric Power at Minamiyokohama Power
Station. The simultaneous removal process uses lime/limestone scrubbing;
it was tested at MHI's Hiroshima Research Institute.
7.2.2 Nitric Acid By-Production Process
A flowsheet of the process is shown in Figure 7-3. The flue gas
(100,000 Nm3/hr) from an LNG-fired boiler containing about 100 ppm NOX is
cooled with water spray, injected with ozone to oxidize NO to NaOs, and
then washed with water in an absorber. More than 90 percent of the NOX is
removed when more than 1.7 moles of ozone are used for every mole of NO.
An approximately 10 percent nitric acid solution is obtained which is con-
centrated to 60 percent for industrial use. The gas discharged from the
absorber contains a small amount of ozone, which is removed by treatment
with calcium sulfite slurry containing a small amount of catalyst. Major
facilities of the plant are shown in Table 7-2.
395
-------
u>
vo
o\
FLUE,
GAS1
OZONE
GENERATOR
NOX
REMOVAL
I
CONCENTRATOR^
DILUTE
HN03
PRESCRUBBER
COOLER
HN03
TANK
THICKENER
GYPSUM
70-1352-1
Figure 7-3. Flowsheet of Tokyo Electric-MHI oxidation absorption process.
-------
TABLE 7-2. SPECIFICATIONS FOR THE TEST PLANT
(TOKYO ELECTRIC-MHI PROCESS)
Material Structure Dimensions (m)
Prescrubber SUS 304L Packed Tower 4 x 4 x 11
NOx Absorber SUS 304L Packed Tower 4 x 4 x 19.5
Ozone Absorber SUS 304L Packed Tower 4 x 4 x 19.5
Mist Eliminator Polypropylene Chevron 2.6 x 1.8 x 3
Blower SS 60 (Impeller) Centrifugal 370 kW
Operation parameters are listed in Table 7-3 and compared to those for
the simultaneous removal plant. The nitric acid by-production process plant
was operated for about 2 years beginning in late 1974 with virtually no
problem. Ozone leaving the absorber in a small amount was entirely caught
by the sulfite slurry and gypsum of good quality was by-produced. The
process, however, is more costly than SCR which can be easily used for clean
gas. The plant operation has been shut down.
7.2.3 Simultaneous Removal Process (Oxidation-Absorption-Reduction
Process)
Another wet simultaneous removal process is a modification of the wet
lime/limestone process as shown in Figure 7-4. Flue gas from an oil-fired
boiler (2000 Nm3/hr) is first cooled to 60°C by a prescrubber. Ozone is
introduced into the flue gas prior to scrubbing. A water-soluble catalyst
in a lime/limestone slurry is added to promote the oxidation of NO. More
than 85 percent of the NOX and more than 95 percent of the S02 are removed
when more than 4 moles of S02 per mole of NOx are present in the gas. The
slurry discharged from the scrubber contains gypsum and a small amount of
unreacted lime or limestone and soluble nitrogen-sulfur (N-S) compounds
(reactions 7-9 - 7-11). The unreacted lime or limestone is treated with
397
-------
TABLE 7-3. PILOT PLANTS BY MHI WET PROCESSES
Gas Source
Capacity (Nms/hr)
03/NOX Mole Ratio
NOX (inlet, ppm)
NOX Removal (%)
S02 (inlet, ppm)
SO 2 removal (%)
Particulates (inlet, mg/Nm3)
Prescrubber Type
Absorber Type
pH of Scrubber Liquor
Ozone Remover Type
By-Product
Year Completed
Nitric Acid
By-Production
LNG-Fired Boiler
100,000
1.7-2.0
100
Over 90
None
—
Trace
Packed Tower
Packed Tower
Below 1
Packed Tower
HNOs , Gypsum
1974
Simultaneous
Removal
Oil-Fired Boiler
2000
0.9-1.0
150-250
75-92
500-1200
90-99
90-120
Spray Tower
Packed Tower
4-6
—
Gypsum
1975
398
-------
Ui
vo
vo
AIRI—-q D—I
OZONE
GENERATOR
THICKENER
PARTICIPATES
-. HIST
1 ELIMINATOR REHEATER
-•*-». IVAAI I 1
STACK
THICKENER
ABSORBENT HAKE-UP
TANK
GYPSUM
Figure 7-4. MHI simultaneous removal process.
-------
sulfuric acid to increase the purity of by-product gypsum; 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 reactor at 100-130°C to decompose more than 95 percent of the
N-S compounds (reactions 7-12 - 7-15) - A small portion of the nitrogen is
converted to N2 ; most of it is converted to NlUHSOii which is then treated
with lime to recover gypsum and gaseous ammonia. The gypsum and ammonia,
5-10 percent concentration, may be recovered and used (reaction 7-16) .
2N02 + Ca(OH)2 + CaSOs'h H20 + % H20 - •> (7-9)
Ca(N02)2 + CaSOif-2H20
Ca(N02)2 + 4(CaS03'%H20) + 2H20 - > (7-10)
2CaNOH(S03)2 + 3Ca(OH)2
CaNOH(S03)2 + CaS03«%H20 + 3/2 H20 - >• (7-11)
CaNH(S03)2 + CaSOit«2H20
2Ca[NOH(S03)2] + 2H20 - > (7-12)
Ca(NOH'HS03)2
Ca(NOH*HS03)2 + 2H20 - > (7-13)
2/3 N2 + CaS0lt'2H20 + 2/s
2Ca[NH(S03)2] + 2H20 - -> (7-14)
Ca(NH2S03)2 + Ca(HSO.t)2
Ca(NH2S03)2 + Ca(HSCK)2 + 4H20 - >• (7-15)
2CaS0lf-2H20
+ Ca(OH)2 - > (7-16)
NH3 + CaSOit«2H20
400
-------
Some of the operation data of the pilot plant are listed in Table 7-3.
Specifications of the plant facilities are shown in Table 7-4.
TABLE 7-4. SPECIFICATIONS OF 2000 NM3/HR TEST PLANT FOR SIMULTANEOUS
REMOVAL OF SOX AND NOX BY OXIDATION-ABSORPTION-REDUCTION
Prescrubber Spray Tower 0.7 m diameter, 7 m height
Absorber Grid Packed Tower 0.5 m diameter, 13 m height
Ozonizer Mitsubishi Electric 1200 g/hr*
(OS-1200)
Flue Gas Fan Capacity 5000 Nm3/hr Pressure, 600 mm H20
Centrifuge Basket Type 300 1/hr
* About half of the capacity is sufficient for the pilot plant.
The plant has been operated without appreciable problem. Nearly 1
mole of Os has been used for each mole of NO to obtain 85-90 percent NOx
removal efficiency. It is possible to obtain over 90 percent efficiency by
using a larger amount of 03. In this case, however, a small amount of
nitrate is formed in the liquor, which makes wastewater treatment difficult.
Unreacted ozone is entirely removed in the scrubber by the calcium sulfite
slurry. There is no leakage of Os from the system.
Because of the oxidizing effect of NOz and 0$, virtually all of the
sulfite is oxidized to gypsum in the scrubber. An oxidizing tower, which
is usually installed for conventional lime/limestone gypsum processes, is
not needed when the SOz content of flue gas is below 1200 ppm. For flue
gas richer in S02, an oxidizer may be needed but should be substantially
smaller than those in conventional FGD plants.
Gypsum derived from the sulfuric acid is less than 5 percent of the
total by-product gypsum. Sulfuric acid is not needed when a high purity
of gypsum is not required.
401
-------
Both capital and operation costs for a commercial simultaneous removal
process plant to treat flue gas from an oil-fired boiler containing about
150 ppm of NOx are estimated by MHI to be about 40 percent higher than
those for an SOz removal process plant. About 80 percent of the increased
costs are due to the ozone generator.
7.2.4 Evaluat ion
A desirable feature of the by-product nitric acid process is that NOX
is converted to a useful by-product. The process, however, seems fairly
costly and requires a large amount of ozone and many vessels. Moreover,
it may not be applied to S0x-containing gas. For NOX removal from clean
gas, SCR may be easier and less costly.
The simultaneous removal process is a fine process in which both NOX
and SOa are removed by the simple addition of ozone to the gas and a
catalyst to the scrubber liquor. It is easy to modify the existing lime/
limestone FGD plant for simultaneous removal.
A large disadvantage of the process is the high cost of ozone. The
process may be useful for flue gas with a relatively low NOX concentration
such as that from an oil-fired boiler, but may be too expensive for an NOX-
rich gas such as that from a coal-fired boiler. Recovery of the by-product
ammonia should be considered. The ammonia may be injected into the boiler
to reduce NOX. The simultaneous removal process may be applied to flue gas
from a coal-fired boiler if NOX concentration is reduced to 250 ppm by
combustion control and to below 150 ppm by injecting the by-product ammonia
into the boiler.
As in the case of conventional FGD processes, some wastewater would
have to be purged to prevent the accumulation of impurities from the gas
and lime or limestone. The recovery of catalyst from the wastewater may
present a problem.
402
-------
7.3 IHI OXIDATION-ABSORPTION-REDUCTION PROCESS (SIMULTANEOUS REMOVAL)
7.3.1 Introduction
Ishikawaj ima-Harima Heavy Industries (IHI), one of the largest
producers of boilers, ships, and machines, has developed a wet simultaneous
removal process with ozone oxidation and lime/limestone scrubbing to by-
produce gypsum which converts NOx to N2 . A pilot plant which can treat
5000 Nm /hr flue gas from an oil-fired boiler was operated in 1976 and
1977.
7.3.2 Process Description
A flowsheet of the IHI process is shown in Figure 7-5. Flue gas is
cooled, injected with ozone to convert NO to N02, and treated in a scrubber
with a lime/limestone slurry at pH 5-6 containing small amounts of CuCl2
and NaCl as catalysts. Over 80 percent of the NOx and over 95 percent of
the S02 are absorbed at an L/G of over 8 liters/Nm3 . Most of the absorbed
NOX forms calcium sulfamate, a soluble compound (reaction 7-20), while a
small portion of N02 is converted into various forms (reactions 7-21 -
7-24) .
S02 + Ca(OH)2 - > CaS03 + H20 (7-17)
S02 + CaC03 - > CaS03 + C02 (7-18)
S02 + CaS03 + H20 - >• Ca(HS03) (
2N02 + 5CaS03 + Ca(HS03) + 11H20 - >• Ca(NH2S03)2 + (7-20)
5(CaS0lt'2H20)
2N02 + 4CaS03 + 8H20 - >• N2 + 4(CaSOit'2H20) (7-21)
403
-------
AIR*
EXHAUST GAS
.p-
o
.p-
PARTICULATES
CaS04
Figure 7-5. Flowsheet of the IHI process.
-------
4N02 + 4CaS03 + 2H20 - >• Ca(N02)2 + Ca(N03)2 + 2Ca(HS03)2 (7-22)
2N02 + 3CaS03 + 6H20 - *• N20 + 3(CaSOit'2H20) (7-23)
Ca(N02)2 + Ca(NH2S03)2 + 2H20 - *• 2N2 + 2(CaS0lf'2H20) (7-24)
The relation between the catalyst concentration and NOX removal
efficiency is shown in Figure 7-6. The combination of CuCl2 and NaCl is
effective for NOX removal. Figure 7-7 shows the effect of pH on SOX and
NOX removal with and without the catalyst . A high pH increases SOX removal
but lowers NOX removal. A pH between 5 and 6 is optimum for simultaneous
removal .
The slurry leaving the absorber is treated with air bubbles if
necessary to complete the oxidation of CaSOs to gypsum. The gypsum slurry
is filtered to obtain a good quality by-product gypsum. Most of the liquor
discharged from the centrifuge is returned to the absorber; the rest is
sent to a thermal decomposition step after separation of copper by lime
addition.
For the thermal decomposition, the liquor is first evaporated and
dried to crystallize calcium sulfamate and nitrate. These are then calcined
with limestone in a rotary kiln at about 900°C to produce calcium sulfate
and nitrogen gas (reactions 7-25 and 7-26).
2Ca(NH2S03)2 + 4CaC03 + 2Ca(ES0^2 + 302 + 2H20 - >• (7-25)
2N2 + SCaSCK + 4C02 + 8H20
Ca(N03)2 + Ca(NH2S03)2 - > 2N2 + 20350^ + 02 + 2H20 (7-26)
The pilot plant was operated without appreciable problem. IHI also
developed a process to hydrolize the sulfamate to by-produce ammonia. In
405
-------
CONDITIONS
CONCENTRATION OF CaSOs: 5 wtX
pH OF LIQUOR : 5.5
N02/(NO + N02) : 0.95
NaCl 0.5 moles/a
NaCl 0.17 moles/
• i ••
\
No NaCl, CuCl2 only
0.5 1.0 1.5 2.0
Concentration of CuCl2 (moles/a)
2.5xlO"2
Figure 7-6. Effect of CaCl2 and NaCl concentration on NOx removal
efficiency in the IHI process.
406
-------
CONDITIONS
CONCENTRATION OF CaS08: 5 wtX
CONCENTRATION OF CuCl2: 0.01 mole/*,
CONCENTRATION OF Nad : 1 wtt
NOa/(NO + N02) : 0.95
nSOv(CaSOs+CuCl2+NaCl)
a*
at
u
M
I
100
80
60
40
20
\
8
'nSOv(CaS08 ONLY)
?\
nNO (CaSO,+Cuda+NaCl)
nNOx(CaS09 ONLY)
pH of Liquor
Figure 7-7. Effect of pH on SOX and NOX removal efficiency
In the XHI process.
407
-------
this case, calcium nitrate accumulates in the scrubber liquor and has to
be separated.
7.3.3 Economics
Figure 7-8 shows investment cost and operating cost for the simul-
taneous removal of SOX and NOX for 8000 hours of annual operation. The
figures include 7 years' depreciation. Roughly speaking, both investment
and operation costs are about 40 percent more than those of FGD only. The
figure also shows the incremental operating cost for NOx removal (the
difference between simultaneous removal and FGD). The operation costs may
be considerably higher for NOx-rich flue gas from a coal-fired boiler
because a much larger amount of ozone may be needed.
7.3.4 Evaluation
IHI made extensive fundamental studies of the IHI process. The pilot
plant was operated without appreciable problem and proved the reliability
of the process. However, there is no plan yet to install a commercial
plant based on the process. The major reason for this hesitance is the
current policy of power companies to use LNG or low-sulfur oil for new
boilers with selective catalytic reduction (SCR) of NOX without FGD.
For flue gas from high-sulfur oil, the IHI process might be more
economical than a combination of SCR and wet FGD. The process may not be
useful for flue gas from a coal-fired boiler because the large amount of
ozone required would make the process too costly.
408
-------
•510,000
8000
£ 5000
o
o
§ 4000
•i—
+J
rO
2 2000
Capacity (Nm3/hr)
100 200
300
200 400 600
Capacity (1000 Nm3/hr)
800
10
c
O)
M-
O
§
CO
o
CJ
V)
0 I
Figure 7-8. Investment and operation costs.*
* Inlet SOX, 1500 ppm; inlet NOX, 180 ppm; yearly operation, 8000 hours.
409
-------
7.4 SUMITOMO-FUJIKASUI PROCESS (MORETANA PROCESS)
7.4.1 Introduction
Sumitomo Metal, one of the largest steel producers, jointly with
Fujikasui Engineering, a pollution control plant engineering and construc-
tion company, has developed a wet simultaneous removal process which uses
C102 as the oxidizing agent and NaOH as the absorbent. They have con-
structed relatively small commercial plants using the process. The process
can remove over 90 percent of the SOX and NOX but it may not be suitable
for large plants. A large amount of NaOH is consumed and the waste liquor
containing sodium chloride, nitrate and sulfate must be treated.
Sumitomo and Fuj ikasui recently developed a modified process which
uses ClOz as the oxidizing agent and a CaCOs slurry containing a catalyst
as the absorbent. After tests with a small pilot plant, Sumitomo Metal
has operated since 1976 a larger pilot plant which can treat 25,000 Nm3/hr
of flue gas from an iron-ore sintering machine at the Kashima Works.
7.4.2 C102-CaCOs Process**
A flowsheet of the simultaneous removal process which uses C102 and
CaCOs is shown in Figure 7-9. The flue gas contains about 200-250 ppm NOX,
250-500 ppm SOX, 15-18 percent 02, and 30-80 mg/Nm3 particulates. The gas
(25,000 Nm /hr) is first treated with water in a Moretana tower, an improved
type of perforated plate tower, to remove most of the particulates and to
cool and humidify the gas. The gas is then injected with ClOz gas to
oxidize NO to N02 by the following reaction.
2NO + C102 + H20 >• N02 + HC1 + HN03 (7-27)
A stoichiometric amount of C102 is added to instantaneously oxidize
virtually all of the NO to N02. S02 in the gas is not oxidized by C102.
410
-------
GAS SOURCE£-
WATER5-
C102
THICKENER
SCRUBBER
V
A
1 f
TANK
CATALYS1
CATALYST
RECOVERY
MIST
ELIMINATOR
GYPSUM
TO WASTE
WATER TREATMENT
STACK
{CENTRIFUGE
SLUDGE
Figure 7-9. Flowsheet of the Sumitomo-Fujikasui C102-CaC03 process.
-------
The gas is then treated in another Moretana tower (scrubber) with a
limestone slurry containing a catalyst. The following reactions occur in
the scrubber.
CaC03 + S02 >• CaS03 + C02 (7-28)
2N02 + 4CaS03 Catalyst> N2 + 4CaSOi, (7-29)
2HN03 + 5CaS03 Catalyst> NZ + sCaSCH + H20 (7-30)
2HN03 + CaC03 >~ Ca(N03)2 + H20 + C02 (7-31)
2HC1 + CaC03 > CaCl2 + H20 + C02 (7-32)
The treated gas leaving the scrubber contains 60-80 ppm NOy, 25-50 ppm
SOX, 15-18 percent 02, and 20-40 mg/Nm3 particulates. The gas is passed
through a mist eliminator and led to a stack. The total pressure drop of
the gas in the system is 500 mmH20.
Most of the slurry leaving the scrubber is recirculated, but a portion
is centrifuged to separate gypsum, which is of high quality and useful for
cement and wallboard. About 150 liters/hr of the liquor from the centrifuge
are sent to a wastewater treatment system after most of the catalyst in
the liquor is recovered and returned to the scrubber. Limestone and a small
amount of makeup catalyst are added to the remaining liquor before it is
sent to the scrubber.
The waste liquor contains calcium chloride and nitrate. It can be
dried in a spray dryer to give a powdery product that may be useful for
production of an anti-freezing agent. It is also possible to treat the
liquor with ammonium sulfate to precipitate gypsum and obtain a liquor
containing ammonium chloride and nitrate that may be used as liquid
412
-------
fertilizer. The small amount of catalyst remaining in the product is not
toxic and the product has no harmful effect on the environment.
7.4.3 Evaluation
An advantage of this simultaneous S0x/N0x removal process is that a
plant can be constructed simply by adding a wastewater treatment system to
a conventional wet limestone scrubbing plant.
The pilot plant has removed about 80 percent of the NOX and over 90
percent of the SOX. It may be possible to remove 90 percent of the NOX and
over 95 percent of the S0>< by using two scrubbers in series. Removal of
over 90 percent of the NOx is difficult with other wet simultaneous removal
processes. It is possible with the Moretana process because half of the
NO is oxidized to form HNOa which is readily absorbed by the limestone
slurry; removal of 80 percent of the NOa formed by the oxidation results in
an overall removal of 90 percent.
On the other hand, the waste liquor contains large amounts of nitrate
and chloride. The use of these by-products is limited. Although the pro-
cess may be better than the ClOa-NaOH process, disposal of the waste liquor
or the by-product from the liquor may not be easy. The ClOz-NaOH process
not only consumes a large amount of NaOH but also produces a waste liquor
containing sodium chloride, sulfate and nitrate.
A considerable amount of ClOa has been used in Japan as a bleaching
agent. For treating a large amount of flue gas, CIOz should be produced at
the plant. Though this is not difficult, it does require an electrolytic
cell, NaOH and HC1, and may not be convenient for power plants.
Ozone can be used as the oxidizing agent in place of ClOa. Sumitomo
and Fujikasui have preferred ClOa to ozone because ClOa is considerably
less expensive than ozone. When ozone is used, the by-products of the
413
-------
process are gypsum and calcium nitrate. A considerable amount of the
nitrate may be used for fertilizer or it can be treated with ammonium
sulfate to produce gypsum and ammonium nitrate. For flue gas that is not
rich in NO, ozone may be used instead of C102 to simplify the system.
7.5 CHISSO ENGINEERING ABSORPTION REDUCTION PROCESS (CEC PROCESS) FOR
SIMULTANEOUS REMOVAL
7.5.1 Introduction
Chisso Engineering is a subsidiary of Chisso Corp., a large chemical
company which produces petrochemicals and fertilizers. Chisso has developed
a wet simultaneous removal process in which the gas is scrubbed with an
ammonia solution containing EDTA and the ferrous ion. A pilot plant which
can treat 300 Nm3/hr flue gas from an oil-fired boiler has been operated
since 1975. NOx in the gas is converted to NHs by the reducing effect of
SOa to produce ammonium sulfate. The reactions are complex but the overall
reaction may be simply expressed as follows:
2NO + 5S02 + 8NH3 + 8H20 = 5(NHit)2SOif (7-33)
The process has been licensed to Catalytic, Inc., U.S.A.
7.5.2 Process Description5
The flue gas is first subjected to water scrubbing for cooling and
removal of particulates and chlorine, and then is treated in a scrubber
with an ammonia solution at pH 6.0 (Figure 7-10). The main reactions are
shown below:
S02 + (NHif)2S03 + H20 > 2NIUHS03 (7-34)
NO + 3NHi»HSOs ——> NHOOsNHOz + ls(NH^)2S2Os + H20 (7-35)
414
-------
FLUE
NEUTRALIZATION &
CRYSTALLIZATION
(HH4)2S04
Figure 7-10. Flowsheet of the CEC process.
-------
+ %02 - >• (NHlt)2S206 + H20 (7-36)
(NHi,.)2S03 + %02 - >• (NHif)2SOif (7-37)
To obtain a high NOX removal ratio, about 5 moles of S02 are needed
for each mole of NO. Normally 70-85 percent of the NOx is removed with
more than 99 percent of the S02 in a scrubber with about 20 stages of
perforated plates. A portion of the scrubber liquor is subjected to air
bubbling to oxidize ammonium sulfite and bisulfite to sulfate and dithion-
ate. Then sulfuric acid is added to the liquor to lower the pH to 0.5.
The acidic liquor is cooled to 0 ^ -5°C by the latent heat of evaporation
of liquid ammonia. About 90 percent of EDTA is crystallized by the cooling.
The EDTA is centrifuged and returned to the absorber. The liquor discharged
from the centrifuge is then heated under 3 atmospheres of pressure at
120-130°C for several hours to decompose imidodisulfonate and dithionate.
Reactions 7-38, 7-39 and 7-40 occur because of the heating in the acidic
liquor (Figure 7-11) . A considerable portion of the remaining EDTA is also
decomposed in the decomposition step.
NH(S03NHif)2 + H20 - >• NH2S03NHlf + NH^HSOi, (7-38)
NHzSOsNHif + H20 - >• (NHif)2SOi» (7-39)
(NHO2S206 - >• (NHOaSCK + S02 (7-40)
S02 released by reaction 7-40 is returned to the scrubber. This is
done to maintain an adequate N0x/S02 ratio and to reduce S02 emissions. The
liquor obtained from the decomposition step contains ammonium sulfate and
bisulfate with small amounts of dithionate, EDTA and the ferric ion. The
liquor is neutralized with ammonia and concentrated in a circulation flash
evaporator after the separation of ferric hydroxide. The crystallized
ammonium sulfate is centrifuged; the mother liquor containing dithionate
416
-------
10
NH2S03/
S2°6
0.1
100 120 140
TEMPERATURE (°C)
Figure 7-11. Decomposition rates of intermediates in the CEC process.*
* -d(S2o| )/dt = k(S2Os")
-d(NH2SOl)/dt
417
-------
with small amounts of EDTA and chloride is returned to the catalyst
recovery step.
About 70 percent of the sulfur and 10-20 percent of the nitrogen in
the by-product ammonium sulfate are derived from the SOa and NOx of the
flue gas and the rest from sulfuric acid and ammonia added to the system.
7.5.3 Economics6
Chisso Engineering considers the plant cost for the process to be 20
to 30 percent lower than the total of wet process FGD and dry process
denitrification. The estimated operating costs are shown in Table 7-5.
The estimated costs are 7289 yen/kl ($36.4/kl) or 1.82 yen/kWhr
($0.0091/kWhr) for a plant which can treat 150,000 Nm3/hr (50 MW) and 6593
yen/kl ($33.0/kl) or 1.65 yen/kWhr ($0.0083/kWhr) for a 500,000 Nm3/hr plant
(166 MW), excluding the cost for gas reheating.
7.5.4 Evaluation
Production of ammonium sulfate from recovered NOX and SOa is an
advantage of the process. A major disadvantage may be the requirement of
a large scrubber with about 20 stages of perforated plates. The gas pres-
sure drop is about 30 inmHaO for each stage. Oxygen concentration of the
gas affects the NOX removal efficiency. For a gas containing 500 ppm NO
and 2500 ppm SOa, the NOX removal efficiency would be about 80 percent for
3 percent 02 and 70 percent for 5 percent Oa.
The plume problem common to ammonia scrubbing FGD processes has not
been encountered with the CEC process, presumably because of the unusually
high SOX removal efficiency of over 99 percent.
418
-------
TABLE 7-5. ESTIMATED OPERATING COSTS (CEC PROCESS FOR OIL FIRED BOILER)
Basis:
Utility
Electric Power
Steam
Cooling Water
Process Water
Plant Size
Fuel Oil Burned
Stream Time
NOx Absorption Ratio
Consumption
Per Hour
1800 kWhr
6000 kg
700 m3
10 m3
150,000 Nm3 (Dry)/hr (Oz , 4 percent)
12.5 kl/hr (Sulfur, 3 weight percent)
8000 hours /year
80 percent (200 ppra •* 50 ppm)
Unit Price Cost Per Kl
8.5 yen/kWhr ($0.042/kWhr) 1224 yen/kl ($
2.2 yen/kg ($0. Oil/kg) 1056 yen/kl ($
4.0 yen/m3 ($0.02/m3) 223 yen/kl ($
16.0 yen/m3 ($0.08/m3) 12 yen/kl ($
Oil
6.12/kl)
5.28/kl)
1.12/kl)
0.06/kl)
Total
2515 yen/kl ($12.575 kl)
Ammonia and Chemicals
2493 yen/kl ($12.465/kl)
Labor Cost,
10 men x 5 M yen/man-year
500 yen/kl ($ 2.50/kl)
Depreciation, Interest,
Maintenance, Taxes, Insurance
3660 yen/kl ($18.30/kl)
Recovered Ammonium Sulfate
1700 kg/hr x 15 yen/kg (Return)
1880 yen/kl ($ 9.40/kl)
Net Operating Cost
7289 yen/kl ($36.445/kl)*
* For a 500,000 NmVhr plant: 6593 yen/kl ($32.965/kl). Cost for flue gas reheating is not included above.
-------
The process would be useful where there is a good market for ammonium
sulfate and where the removal of 70-80 percent of the NOx is required.
Where there is no market for the sulfate, it should be decomposed either by
addition of lime to the sulfate liquor to precipitate gypsum or by the IFF
process to recover elemental sulfur. By these processes, ammonia is regen-
erated and returned to the scrubbing system.
A portion of the ammonia is derived from the NOX and must be discharged
from the system. One disposal method is to inject the ammonia into the
boiler at 900-1000°C for further reduction of NOX. With these processes,
refrigeration is needed for cooling the liquid in the catalyst recovery
step because latent heat of evaporation of liquid ammonia would not be
available.
7.6 ASAHI CHEMICAL PROCESS
7.6.1 Introduction
The environmental division of Asahi Chemical, one of the largest
chemical companies in Japan, has developed wet simultaneous removal pro-
cesses based on sodium and potassium scrubbing, both using EDTA and the
ferrous ion. The entire system was tested at a small pilot plant which
could treat 40-60 Nm3/hr of flue gas; the scrubbing and thermal decomposi-
tion steps were tested with larger facilities equivalent to 500-600 Nm3/hr.
The potassium process can remove over 90 percent of the NOX with over 99
percent of the SOX, but it seems costly. The sodium process, capable of
removing 80-90 percent of the NOX with 99 percent of the SOX, may be more
practical. The sodium process will be described below.
7.6.2 Process Description
A flowsheet of the sodium process is shown in Figure 7-12. Flue gas
is first treated by water scrubbing to remove particulates and chlorine
420
-------
NJ
WATER
I
r-WlIl)
PRESCRUBBER
PARTICULATES
& CHLORIDE
TREATMENT
FLUE GAS
MAKEUP.
Na2C03'
Na2S206.2H20
Na?S04>10H20
Figure 7-12. Flowsheet of the sodium process of Asahi Chemical.
-------
and is then introduced into a scrubber. In the scrubber is it treated with
a liquor at 60°C which has a pH of 6 and contains sodium sulfite, sulfate,
EDTA and the ferrous ion. By using about 10 stages of perforated plates
in a scrubber with a total pressure drop of 400-500 mmHzO and L/G of 15
liter/Nm9 , about 90 percent of the NO is removed when the 02 content of the
gas is below 5 percent and the S02/N0 mole ratio is above 3. An increase
in the 02 concentration lowers the NOX removal efficiency slightly.
Asahi assumes the reactions to be as follows: NO forms an adduct with
Fe-EDTA (equation 7-41) which reacts with sulfite and bisulfite in a "reduc-
tion tank," which is merely a holding tank to give sufficient reaction time
for the formation of N2 and sodium imidodisulfonate (equations 7-42 and
7-43). More than 99 percent of the S02 is also absorbed to form dithionate
and bisulfite (equations 7-44 and 7-45). Those reactions also occur in the
scrubber.
Fe2+ - EDTA + NO - -> Fe2+ - EDTA-NO (7-41)
Fe2+ - EDTA-NO + Na2S03 - - >• Fe2+ - EDTA + >* N2 + NaaSOif (7-42)
2Fe2+ - EDTA- NO + 5NaHS03 - > 2NH(S03Na)2 + NaHSOi* (7-43)
+ 2Fe2+ - EDTA + H20
Na2S03 + S02 + % 02 - > Na2S206 (7-44)
Na2S03 + S02 + H20 - > 2NaHS03 (7-45)
Na2S03 + % 02 - •>• NaaSOi* (7-46)
Most of the liquor discharged from the reduction tank is recycled to
the absorber; a portion is sent to an evaporator 'and the concentrated
liquor is then cooled to about 10°C by water and a refrigerator to crystal
lize sodium dithionate and sulfate. These salts are centrifuged, dried
422
-------
dried and calcined at about 400°C to decompose dithionate to sulfate and
SOa (reaction 7-47). The SOa and most of the sulfate are sent to a sulfate
converter. Most of the liquor discharged from the centrifuge is returned
to the scrubbing system. A portion of the liquor is treated with potassium
sulfate to precipitate potassium imidodisulfonate (reaction 7-48) which is
much less soluble than sodium imidodisulfonate (Figure 7-13) . The potassium
imidodisulfonate is separated by a centrifuge and calcined at about 500°C
to recover potassium sulfate and concentrated S02 (reaction 7-49) . The
sulfate is recycled and the SO 2 is sent to the sulfate converter.
Na2S206 - > NaaSCH + S02 (7-47)
2NH(S03Na) + KaSO* - > Na2SOif + 2NH(S03K) (7-48)
2NH(S03K) + h 02 - > KzSO* + S02 + N2 + H20 (7-49)
The sulfate conversion step and the subsequent steps are the same as
those in the sodium-limestone double alkali processes developed by both
Kureha-Kawasaki and Showa Denko and applied to many commercial FGD plants.
The reactions in the converter are as follows:
CaSOs + S02 + H20 - > Ca(HS03)2 (7-50)
Ca(HS03)2 + Na2SOit + 2H20 - >• 2NaHS03 + CaSOi^Hj-O (7-51)
The by-product gypsum slurry is centrifuged. The mother liquor is
sent to a stripper to decompose the remaining calcium bisulfite and to
recover S02 which is sent to the sulfate converter.
Ca(HS03)2 Heat> CaS03 + S02 + H20 (7-52)
The liquor discharged from the stripper contains sodium bisulfite with
some unreacted sodium sulfate and calcium sulfite. It is treated with
423
-------
70-
60-
50-
-------
limestone In a reactor to recover sodium sulfite which is returned to the
absorbing system. Calcium sulfite precipitated in the reactor is sent to
the sulfate converter.
2NaHSOs + CaC03 > Na2SOs + CaS03- ^HaO + \ H20 + C02 (7-53)
From the results of the tests with different gas compositions, it is
estimated that when a typical flue gas from a coal-fired boiler is treated,
about 93 percent of the absorbed S02 is converted to gypsum while about 7
percent is converted to sodium sulfate.
7.6.3 Evaluation
Although the Asahi Chemical process is not simple, it is a combination
of commercially proven processes and operation of the whole system may be
feasible. The vessels other than the prescrubber and the absorber are
small because only the separated sodium and potassium salts are treated in
them. The loss of EDTA must be very small because the salts are separated
from the liquor containing EDTA and the liquor is returned to the absorbing
step. Even when the thermal treatment and regeneration steps have to be
shut down, the scrubber system can continue operation for a while. Opera-
tion of a larger pilot plant will be needed for a reliable cost estimation
for the process.
7.7 KUREHA CHEMICAL WET SIMULTANEOUS REMOVAL PROCESSES
7.7.1 Introduction
Kureha Chemical is a medium-size chemical company credited with
developing many new processes in petrochemical and electrochemical fields
as well as in FGD. These processes have been licensed to many countries.
Kureha modified the sodium acetate FGD process and developed a simultaneous
removal process by which SOX is converted to gypsum while NOX is converted
425
-------
into NH3 . In cooperation with the Federation of Electric Companies, they
operated a pilot plant which could treat 5000 Nm3/hr of the flue gas from
an oil-fired boiler.
7.7.2 By-Product Gypsum and Ammonia Process
A flowsheet of the gypsum and ammonia by-production process is shown
in Figure 7-14. Flue gas containins SOX and NOx is treated in a packed
tower with a liquor at a pH of nearly 6. The liquor contains sodium
acetate, a catalyst (ferrous ion and EDTA), and a small amount of seed
crystals of gypsum. The following reactions occur in the scrubber:
S02 + 2CH3COONa + H20 >• Na2S03 + 2CH3COOH (7-54)
2ND + 5Na2S03 + 4CH3COOH > 2NH(S03Na)2 + 3Na2SO<, (7-55)
+ 4CH3COONa + H20
2N02 + 7Na2S03 + 4CH3COOH >• 2NH(S03Na)2 + 3Na2S0lf
+ 4CH3COONa + H20 (7-56)
A small amount of acetic acid formed by reaction 7-54 is volatilized
from the absorption section (packed tower) and is caught by the water
scrubber at an upper part of the absorber with perforated plates. Sodium
sulfite is also formed by reaction 7-54. NOX in the gas is reacted with
the sodium sulfite in the presence of catalysts to form sodium imidodi-
sulfonate and sulfate (reactions 7-55 and 7-56).
The liquor discharged from the scrubber is treated with lime in a
reactor. The following reactions occur in the reactor:
2CH3COOH + Ca(OH)2 > (CH3COO)2Ca + 2H20 (7-57)
Na2SOit + (CH3COO)2Ca + 2H20 > CaSO.*'2H20 + 2CH3rCOONa (7-58)
426
-------
to
S02 & NOX
ABSORBER^.
STACK
WATERf
V
FLUE GAS
NH3
DECOMPOSITION
Ca(OH)2
GYPSUM
REACTOR
TANK
NH3
STRIPPER
H2S04
CENTRIFUGE
HYDROLYSIS
TANK
GYPSUM
Figure 7-14. Block flow diagram of the Kureha NDSN gypsum and ammonia by-product process for
simultaneous removal of 50% and NOX.
-------
Acetic acid reacts with lime to form calcium acetate, which reacts
with sodium sulfate to precipitate gypsum (reactions 7-57 and 7-58). The
gypsum is centrifuged. Most of the liquor discharged from the centrifuge
is returned to the absorber. A portion of the liquor is acidified with
sulfuric acid and heated to 120-130°C under 2-3 atmospheric pressure to
decompose the imidodisulfonate (reaction 7-59). Ammonium bisulfate formed
by the decomposition is reacted with lime to precipitate gypsum and to
release ammonia. The ammonia is decomposed by catalytic oxidation when no
use can be found (reaction 7-61).
NH(S03Na)2 + 2H20
Na2SOi,
(7-59)
Ca(OH)2
(7-60)
4NH3 + 302
2N2 + 6H20
(7-61)
Operation data of the pilot plant are shown below.
NOX (ppm):
SOX (ppm):
02 (percent):
L/G:
Pressure Drop:
Raw Materials (kg/hr)
By-Product:
Utilities:
Electric Power:
Steam:
Process Water:
Inlet, 200; outlet, 30-40
Inlet, 1500; outlet, 1
2.1-4.2 (in and out)
28 liters/Nm3
200 mmH20 through the whole system
Ca(OH)2, 33
H2SOit, 13
CH3COOH, 0.2
NaOH, 0.1
Gypsum, 76 kg/hr
205 kW
60 kg/hr
250 kg/hr
428
-------
A scaling problem was encountered at an early stage of operation but
was solved by adding gypsum crystal seeds.
7.7.3 By-Product Sulfur and Ammonia Process
A-flowsheet of the sulfur and ammonia by-product process is shown
in Figure 7-15. Flue gas is first treated in a prescrubber with water
to remove particulates and chlorine and then introduced into a scrubber
(absorber). There it is treated with a. liquor containing potassium sulfite,
EDTA and the ferrous ion. More than 99 percent of the S02 and over 90
percent of the NOx are removed at an L/G ratio of 10 liters/Nm3 by the
following reactions:
K2S03 + S02 + H20 - > 2KHS03 (7-62)
2K2S03 + 2S02 + 02 - > 2K2S206 (7-63)
2NO + 4KHS03 + K2S03 - > 2NH(S03K)2 + K2SO^ + H20 (7-64)
The liquor leaving the scrubber at 55°C is evaporated under a vacuum
to crystallize potassium sulfate K2SOn, dithionate K2S2Oe, and imidodi-
sulfonate NH(S03K)2. The solids are filtered and sent to a hydration and
decomposition step where the following reactions occur at 280-300°C:
NH(SOaK)2 + 2H20 - > NHitHSOi* + K2SOit (7-65)
K2S206 - >• KaSO^ + S02 (7-66)
Most of the S02 formed by reaction 7-66 is sent to a Claus plant to
maintain the stoichiometric ratio of H2S and S02. The rest of the S02 is
returned to the prescrubber to promote the absorption of NOX.
429
-------
ELEMENTAL
SULFUR
Figure 7-15. Flowsheet of the Kureha KDSN sulfur and ammonia by-production process
for the simultaneous removal of SOX and NOx.
-------
Ammonium bisulfate formed by reaction 7-65 is treated with potassium
carbonate to release ammonia (reaction 7-67) .
K2C03 = KaSCK + NH3 + H20 + C02 (7-67)
Potassium sulfate is reduced by coal at 800°C to produce potassium
sulf ide and carbonate (reactions 7-68 - 7-70) .
+ 2C - > K2S + 2C02 (7-68)
+ 4C - > K2S + 4CO (7-69)
K2SOi, + 2C - > K2C03 + CO + S (7-70)
The solids are treated with C02 and steam at 400°C (dry carbonization)
to convert the sulf ide into carbonate and to release hydrogen sulf ide, which
is then sent to the Glaus furnace to produce elemental sulfur.
K2S + C02 + H20 - > K2C03 + H2S (7-71)
2H2S + S02 - >• 3S + 2H20 (7-72)
Potassium carbonate is dissolved in water (condensate from the vacuum
crystallizer) and the resulting liquor is filtered to remove coal ash.
Since reaction 7-71 requires a large amount of C02 gas, a portion of
the flue gas leaving the scrubber is treated with the potassium carbonate
liquor to recover C02 (reaction 7-73).
K2C03 + H20 + C02 = 2KHC03 (7-73)
The resulting liquor is reacted with a portion of the liquor leaving
the scrubber which contains potassium bisulfite to form potassium sulfite
431
-------
and to release C02• The potassium sulfite is returned to the absorbing
system; the C02 is used for reaction 7-71.
KHC03 + KHS03 > K2S03 + C02 + H20 (7-74)
Kureha has operated the whole system at a pilot plant which can treat
200 Nm3/hr of flue gas. A larger scale test (3000 Nm3/hr equivalent) was
carried out for the reduction of potassium sulfate.
7.7.4 Evaluation
The gypsum and ammonia by-production process is fairly simple and can
remove 99 percent of the SOx with 90 percent of the NOx, but it requires a
large L/G ratio of about 30 and a considerable amount of sulfuric acid. The
consumption of EDTA seems pretty large because the liquor containing EDTA
is treated in the hydrolysis step at 120-130°C. Separation of EDTA from
the liquor prior to the hydrolysis substantially lowers the consumption
but complicates the system.
The sulfur and ammonia by-product process is fairly complex, but it
may be the only process currently available for conversion of both SOX and
NOX to useful by-products. It can remove over 99 percent of the SOX with
over 90 percent of the NOX at an L/G ratio of 10. Consumption of EDTA
should be very small because the potassium salts are separated from the
liquor containing EDTA and all of the liquor is returned to the scrubber.
Consumption of potassium may not be large; however, since coal is used to
reduce potassium sulfate, some of the" potassium is discarded with the fly
ash that is also formed.
Although the process includes many operations for thermal treatment
and recovery, the vessels for these operations are much smaller than those
for scrubbing systems because only the separated potassium salts are
treated in them. Moreover, even when the thermal treatment and recovery
432
-------
operations must be shut down, the scrubbing system can continue to operate
for a considerable period.
The dry carbonization step developed by Kureha seems advantageous over
the conventional wet carbonization step in which various compounds other
than HzS and KzCOs are formed by side reactions.
When the ammoina by-product is not wanted, the process can be simpli-
fied because the hydrolysis and ammonia production steps are eliminated.
Further evaluation and a reliable cost estimation for both processes
would require continuous operation of a larger plant.
7.8 CHIYODA THOROUGHBRED 102 PROCESS
7.8.1 Introduction
Chiyoda Chemical Engineering and Construction Co. has modified the
Thoroughbred 101 FGD process to remove NOX simultaneously. Flue gas, after
being cooled in a prescrubber, is injected with ozone and is treated in a
scrubber with dilute sulfuric acid containing ferric sulfate in the same
way as with the 101 process. Most of the NOX is converted to HNOs while a
portion is converted to Na and a small portion to NzO. Absorbed SOz is
converted to HaSO^. The liquor is reacted with limestone to precipitate
gypsum. Since the waste liquor contains calcium nitrate, it is treated with
activated sludge to remove the nitrate. A pilot plant with a capacity of
1000 Nm3/hr was operated. By using 1.5 mole Os for every mole of NOX in
flue gas, 90 percent of the SOz and 80 percent of the NOX were removed.
433
-------
7.8.2 Process Description
The process flow diagram is given in Figure 7-16. The total process
consists of 6 sections: prescrubbing, ozone oxidation, absorption, S02
oxidation, gypsum production, and liquid waste disposal sections.
Prescrubbing —
Flue gas from a boiler is first scrubbed with water to be cooled
approximately to its adiabatic saturation temperature. A large portion of
the particulates in the gas and some of the sulfur trioxide are removed
in the prescrubber.
Ozone Oxidation —
Ozone is then injected into the gas to oxidize the NO to N02 and N20g .
S02 is not oxidized by ozone. The main reactions in this section are as
follows:
NO + 03 > N02 + 02 (7-75)
2N02 + 03 S1°W> N205 + 02 (7-76)
Absorption —
At the absorption stage, the gas is scrubbed to remove S02 , N02, N20s
and Os by an absorbent. The absorbent is a solution of dilute sulfuric and
nitric acids and ferric ions as catalyst; it is saturated with dissolved
oxygen. S02, N02, N20s , and Os are absorbed into the liquid phase.
434
-------
.p.
w
Ui
AIR?
MAKEUP
WATER e
COMBUSTION
FLUE GAS
PARTICULATES
BLEED TO
LIQUID WASTE
DISPOSAL
1 LIMESTONE
'CRYSTALLIZER
CENTRIFUGE GYPSUM
Figure 7-16. Schematic process flow diagram of the Chiyoda Thoroughbred 102 process.
-------
The reactions in the absorber are as follows:
S02 + H20 - > H2S03 (7-77)
2N02 + H20 - > HN02 + HN03 (7-78)
N205 + H20 - > 2HN03 (7-79)
Fe3 +
2H2S03 + 02 -££— >• 2H2SO.t (7-80)
2HN02 + 2H2S03 - > N20 + 2H2SOi> + H20 (7-81)
2HN02 + 3H2S03 - > N2 + 3H2SOif + H20 (7-82)
S02 Oxidation —
The absorber effluent liquid is pumped into a bubble column oxidizer,
where H2S03 is oxidized into sulfuric acid by air injected at the bottom
of the oxidizer.
The liquid, saturated with oxygen, is circulated back to the absorber.
Gypsum Production —
A portion of the absorbent is transferred to the crystallizer and
charged with limestone slurry to precipitate gypsum crystals. The by-
product gypsum crystals are continuously separated from the liquid by a
centrifuge. The mother liquor is returned to the absorption section.
Liquid Waste Disposal —
A small amount of mother liquor is continuously bled from the liquid
return stream to maintain a constant nitric acid concentration level and
436
-------
also to limit the accumulation of heavy metals and other undesirable salts
which come into the system from the flue gas, makeup water, and limestone.
The catalyst (ferric ion) is recovered by precipitation in the form of
Fe(OH)s by neutralization and reused. Nitrate ion, thereafter, can be
removed by biological treatment under anaerobic conditions. Nitrate is
reduced to nitrogen gas by adding methanol while the liquor is neutralized
by sulfuric acid injection. Reduction is performed in two stages.
Methanol is used in excess and removed biologically under aerobic conditions
before final purge.
7.8.3 Economics
Only ozone generators need be added to the Thoroughbred 101 plant for
the Thoroughbred 102 process. Almost all of the cost for NOx removal is
that for generating ozone. Overall power consumption for ozone generation
is 23-26 kWh/kg-03. The cost for ozone is 500-700 yen/kg-03 ($2.50-3.50/
kg-Os) for ozone generators with a 100-500 kg/h capacity at 70-90 percent
operation.
Table 7-6 shows the capital investment cost for the Thoroughbred 102
process and the removal costs for NOX and S02. The costs include those for
waste liquor treatment.
Requirements for the wastewater treatment system are shown in Table
7-7. When flue gas at 650,000 Nm3/hr is treated to reduce NOX from 120 to
30 ppm (75 percent removal), the amount of wastewater is 10 m3/hr. The
3300 ppm of NOT in the water is reduced to 30 ppm, which is about equal
to the concentration after city sewage water is treated by an activated
sludge process.
The operation cost including 25 percent of the investment cost at
6000 hours' yearly operation is 17,000 yen/hr ($85/hr) or 1700 yen/m3
($8.50/m3) of water and is equivalent to 360 yen/kl ($1.80/kl) oil.
437
-------
TABLE 7-6. CAPITAL INVESTMENTS AND OPERATING COSTS
FOR THE THOROUGHBRED 102 PROCESS
Basis of Calculation:
NO Concentration
S02 Concentration
NOX Removal Ratio
SOa Removal Ratio
Annual Operating Ratio
150 ppmv at 8 vol. percent moisture
1000 ppmv
80 percent
90 percent
90 percent
Flue Gas
Capacity
(Nm3/h)
Plant Sections
Capital Cost
(106 yen,
$5000)
Operating Cost*
(yen/kl fuel,
$0.005/kl)
300,000 Ozone Generator 1300
Absorber and Other Equipment 2500
3600
5600
450,000 Ozone Generator 1900
Absorber and Other Equipment 3500
3500
5400
750,000 Ozone Generator 3000
Absorber and Other Equipment 5000
3400
5000
* Operating cost includes fixed cost, actual operating cost as well as
depreciation allowance.
TABLE 7-7. REQUIREMENTS FOR WASTEWATER TREATMENT (CHIYODA 102 PROCESS)
Plant Cost:
Site Requirement
Requirements for Operation:
Electric Power
Methanol
H2S0lt
Fed 3
Labor
180 million yen ($900,000)
30 m x 18 m
80 kW
73 kg/hr
115 kg/hr
2.4 kg/hr
0.9 kg/hr
0.5 man per shift
438
-------
7.8.4 Evaluation
The Thoroughbred 102 process is simple and can remove about 80 percent
of the NOX simultaneously with over 90 percent of the SOz. The catalyst is
neither expensive nor hazardous to the environment. Although waste liquor
treatment is needed to remove the nitrate, the treatment does not cost
much.
The major disadvantage of the process is the requirement of a large
amount of ozone—1.5 moles for each mole of NO. The process should be used
for flue gas from a boiler with extensive combustion modification to reduce
NOX.
7.9 OTHER WET PROCESSES
7.9.1 Permanganate Processes
Nissan Engineering, a subsidiary of Nissan Chemical, developed a
sodium permanganate process and constructed 4 small commercial plants to
treat 100-2000 Nm /hr of the flue gases from plants using nitric acid.
Mitsubishi Metal, jointly with Mitsubishi Kakoki Kaisha and Nippon Chemical,
developed a potassium permanganate process and operated a pilot plant
capable of treating 4000 Nm3/hr of flue gas.
In both processes, NOX is absorbed by a permanganate solution to form
nitrate or nitrite, while the permanganate is converted to manganate or
manganese dioxide. The manganate or manganese dioxide is converted to
permanganate by chemical processes including electrolytic oxidation.
Although over 90 percent NOX can be removed, the regeneration of the
permanganate is costly. Treatment of the by-product liquor containing the
nitrate or nitrite is not easy. Moreover, when the gas contains SOz, the
439
-------
permanganate ig consumed by the SOa. Therefore, use of the process may be
limited to a small amount of gas containing no SOa.
7.9.2 Equimolecular Absorption
Kawasaki Heavy Industries tested magnesium scrubbing of flue gas
(5000 Nm3/hr) from a coal-fired boiler. NOa, equimolecular to NO in flue
gas, is added to the gas and the gas is subjected to magnesium scrubbing to
form magnesium nitrite and sulfite. The liquor is oxidized by air and
treated with sulfuric acid to generate NOx in an amount equal to that of
inlet NO. The NOX is air-oxidized to NOz and injected into flue gas. The
resulting liquor is treated with lime to precipitate gypsum. Calcium
nitrate is also obtained as a by-product. The process was abandoned
because of the low NOX removal efficiency and the difficulty in disposing
of the calcium nitrate.
7.9.3 Others
The processes by Shirogane and Osaka Soda (Table 7-1) are similar to
the sodium process by Sumitomi-Fujikasui (Section 7.4.1) except that
Shirogane uses ozone as the oxidizing agent. The Shirogane process may
not be suitable for large plants because of the consumption of the oxidizing
agent and sodium hydroxide and the problem of the wastewater treatment.
Kobe Steel, Hodogaya, and Mitsui Engineering have given up the
operation of the pilot plants shown in Table 7-1 or have been trying to
change the process.
440
-------
7.10 REFERENCES
1. Tanaka, I., M. Koizumi, and Y. Ishihara, Wet Process for Nitrogen
Oxides Removal from Flue Gases (Part 3). Denryoku Chuokenkyujo Hokoku,
CRIEPI Report, 275017, April 1976. In Japanese.
2. Audrieth, L. F., &t al.t Sulfamic Acid, Sulfamide, and Related Aquo-
Ammonosulfuric Acids. Symposium on the Chemistry of Liquid Ammonia
Solutions, American Chemical Society, Milwaukee, September 1938. In
English.
3. Atsukawa, M., et al.t Development of Wet-Type NOX Removal Processes.
Mitsubishi Juko Giho, Vol. 13, No. 2, 1976. In Japanese.
4. Present Status of Development of NOx Abatement Technology by the Steel
Industry. Japan Steel Federation, April 1977.
5. Sawai, K., and T. Gorai, Simultaneous Removal of SOz and NOx from Stack
Gas by Scrubbing (CEC Process). The Second Pacific Chemical Engineering
Congress, Denver, August 1977. In. English.
441
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-205
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
NOx Abatement for Stationary Sources in Japan
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Jumpei Ando
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Chuo University
Tokyo
Japan
10. PROGRAM ELEMENT NO.
INE624
11. CONTRACT/GRANT NO.
68-02-2161
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 AN.D PERIOD COVERED
Annual; 6/78 - 6/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES TERIj_RTp pr0jeCt Officer iS J.
2915. EPA-600/7-77-103b and EPA-600/2-76-013b
David Mobley, MD-61, 919/541-
are earlier related reports.
16. ABSTRACT TI^ repOrt describes current NOx air pollution problems in Japan and the
methods used there to control NOx emissions from stationary sources. The two most
common techniques in use are combustion modifications and flue gas treatment
(FGT). FGT technologies have been developed more extensively in Japan than in any
other country as a result of Japan's stringent NOx emission standards. The various
NOx control technologies are examined with respect to both principles of operation
and specific process applications. Currently the most popular (over 60 commercial
applications) FGT technology is selective catalytic reduction. This technology is
capable of 80-90% NOx control and has the lowest cost per kilogram of NOx remo-
ved. Recently catalysts have been developed that are resistant to particulate plug-
ging and appear to be applicable to coal-fired boilers in the U.S. Cost and commer-
cial performance data, based on Japanese installations, are presented in much grea-
ter detail than was previously available.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
cos AT I Field/Group
Pollution
Nitrogen Oxides
Combustion
Flue Gases
Catalysis
Reduction
Pollution Control
Stationary Sources
Japan
Combustion Modification
Flue Gas Treatment
Catalytic Reduction
13B
07B
21B
07D
07C
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
496
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
442
*U.S. GOVERNMENT PRINTING OFFICE: 1979-640- 013 3926REGIONNO.4
------- |