EPRI AF-568
EPA-600/7-77-127
November 1977
TECHNICAL ASSESSMENT
OF NOX REMOVAL PROCESSES
FOR UTILITY APPLICATION
by
HI Faucett, J.D. Maxwell,
and T.A. Burnett
Tennessee Valley Authority
Office of Agricultural and Chemical Development
Muscle Shoals, Alabama 35660
EPA Interagency Agreement No, 07-E721-FU
EPA Program Element No. INE624A
EPRI Project No. RP 783-1
EPA Project Officer: J. David Mobtey
EPRI Project Monitor: Donald P. Teixetra
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
ABSTRACT
A state-of-the-art review of the processes currently being developed for
the removal of nitrogen oxides (N0X) from power plant stack gas was conducted.
Included within the 48 processes discussed in the report are: dry NOX pro-
cesses, dry simultaneous N0x~SOx processes, wet NOX processes, and wet simul-
taneous NOx~SOx processes. The major sections in each technical evaluation
for each process included detailed process description containing a simplified
block flow diagram, the current status of development of the process, the raw
material and utility requirements, the published economics, the technical and
environmental considerations, and the overall advantages and disadvantages of
the process. In addition to this review of the current NOX flue gas treatment
(FGf) technology, eight of these processes were recommended as candidates for
preliminary economic analysis in the next phase of the study. The information
for this report was gathered during the first half of 1977 and work was com-
pleted as of July 1977. The report was prepared by the Tennessee Valley
Authority under a project cofunded by the U.S. Environmental Protection Agency
and the Electric Power Research Institute.
11
A
-------
CONTENTS
Abstract 11
Figures. ........ ..... V
Tables . x
Abbreviations and Conversion Factors ... xii
Executive Summary XW
Introduction ................ ... 1
General Chemistry of NOX ........ 7
Wet NOX Removal Processes
Introduction to the Wet Flue Gas Denitrification Processes. ... 17
Asahi Chemical Process - Wet, Absorption-Reduction
(NOX-SOX) 28
Chisso Engineering Process - Wet, Absorption-Reduction
(NOX~SOX). 37
Chiyoda Thoroughbred 102 Process - Wet, Oxidation-
Absorption-Reduction (NOX-SOX) 48
Ishikawajima-Harima Heavy Industries Process - Wet,
Oxidation-Absorption-Reduction (NOX-SOX) . ..... 58
Kawasaki Heavy Industries Magnesium Process - Wet,
Oxidation-Absorption (NOX-SOX) 69
Kobe Steel Process - Wet, Absorption-Oxidation (NOX-SOX) 78
Kureha Process - Wet, Absorption-Reduction (NO-jr-SOx) . 84
Mitsubishi Heavy Industries Process - Wet, Oxidation-
Absorption-Reduction (NOX-SOX) 93
Mitsui Engineering and Shipbuilding Process - Wet,
Absorption-Reduction (NOX-SOX) . . 102
MON Alkali Permanganate Process — Wet, Absorption-
Oxidation (NOX-SOX) 107
Moretana Calcium Process — Wet, Oxidation—Absorption-
Reduction (NOX-SOX) 113
Moretana Sodium Process - Wet, Oxidation-Absorption-
Reduction (NOX-SOX) 123
Nissan Permanganate Process - Wet, Absorption-Oxidation (NOX) . . 132
Pittsburgh Environmental and Energy Systems SCORe Process -
Wet, Absorption-Reduction (NOX-SOX) 137
Tokyo Electric-Mitsubishi Heavy Industries Process - Wet,
Oxidation-Absorption (NOX) ..... 144
Ube Industries Process - Wet, Oxidation-Absorption (NOX). .... 149
ill
-------
Dry NOX Removal Processes
Introduction to the Dry Flue Gas Denltriflcatlon Processes. . , . 151
Ebara-JAERI Process - Dry, Electron Beam Radiation (NOX-SOX). . . 163
Environics, Inc. - Eneron Corporation - Dry, SCR of NOX 172
Exxon Thermal Denox Process - Dry, Selective Noncatalytic
Reduction (NOX). . ..... 187
Exxon Research and Engineering Company - Dry, SCR (NOX-SOX) . . . 197
Foster Wheeler - Bergbau-Forschung Process - Dry, Adsorp-
tion (NOX-SOX) ..... 200
Hitachi, Ltd., Process - Dry, SCR (NOX) ...... 208
Hitachi Zosen Process - Dry, SCR (NOX). ...... 216
JGC Paranox Process - Dry, SCI (NOX) .......... 221
Kobe Steel Process - Dry, SCR (NOX) . 231
Kurabo Knorca Process - Dry, SCR (NOX) 238
Kureha Process - Dry, SCR (NOX) ........... 245
Mitsubishi Heavy Industries Process - Dry, SCR (NOX) 251
Mitsubishi Kakoki Kaisha Process - Dry, SCR (NOX) ... 266
Mitsubishi Petrochemical Process - Dry, SCR (NOX) ... 271
Mitsui Engineering and Shipbuilding Process - Dry, SCR (NOX). . . 278
Mitsui Toatsu Chemicals Process - Dry, SCR (NOX) 284
The Ralph M. Parsons Company Process - Dry, Nonselective
Catalytic Reduction (NOX-SOX) 289
Sumitomo Chemical Process - Dry, SCR (NOX) . 296
Sumitomo Heavy Industries Process - Dry, SCR (NOX-SOX) 307
Sumitomo Heavy Industries Process - Dry, SCR (NOX). . 317
Takeda Process - Dry, SGR (NOX-SOX) 323
Ube Process - Dry, SCR (NOX). 334
Unitika Process - Dry, SCR (NOX-SOX). .... ..... 339
Unitika Process - Dry, SCR
-------
FIGURES
Number
1 Effect of air preheat temperature on NOX emissions. .... 10
2 Effect of heat release rate on NOX emissions. 11
3 Effect of excess air on NOX emissions 12
4 Effect of fuel nitrogen on NOX emissions. .... 13
5 Effect of excess air on fuel nitrogen conversion 14
6 Classification system for wet NOX removal processes .... 18
7 Flow diagram of Asahi Chemical process . 29
8 Flow diagram of Chisso Engineering process . 38
9 Flow diagram of IFF S recovery process 47
10 Flow diagram of Chiyoda Thoroughbred 102 process 49
11 Chiyoda reaction crystallizer .... 51
12 Effect of NOX oxidation level on NOX absorption
efficiency for Chiyoda process .... 57
13 Flow diagram of Ishikawajima-Hariroa Heavy Industries
process 59
14 Effect of absorber L/G on NOX removal efficiency
for IHI process. . 65
15 Effect of ozone/NO mol ratio on the NOX removal
efficiency for IHI process ...... 67
16 Flow diagram of Kawasaki Heavy Industries process 70
17 Flow diagram of Kobe Steel wet process. 79
18 Flow diagram of Kureha wet process. 85
19 Flow diagram of Mitsubishi Heavy Industries
wet process 94
20 Flow' diagram of Mitsui Engineering and Shipbuilding
Company wet process. . 103
21 Flow diagram of MON Alkali Permanganate process 108
22 Flow diagram of Moretana Calcium process 114
23 Moretana plate tower 119
24 Flow diagram of Moretana Sodium process 124
25 Flow diagram of Nissan Engineering process 133
26 Flow diagram of Pittsburgh Environmental Energy
Systems process 138
27 Flow diagram of Tokyo Electric-Mitsubishi Heavy
Industries process ....... . . 145
28 Temperatures below which NH^HSC^ forms 155
29 Flow diagram of Ebara-JAERI electron beam radia-
tion process ......... ........ 164
30 Effect of beam Intensities on NOX and S02 removal
efficiencies for Ebara-JAERI process 165
v
-------
FIGURES (continued)
Number
31 Relationships between inlet gas concentrations, pressure,
and NOX and SOX removal efficiencies for Ebara-JAERI
process ...... ........... 168
32 The relation between the staying time of gas in the
reactor chamber and the removal of SQ2 or NOX for
Ebara-JAERI process , ...... . 170
33 Flow diagram of Environics, Inc. - Eneron Corpora-
tion process 173
34 Effect of temperature on NOX removal efficiency for
Eneron process 178
35 Effect of temperature and NH3:NOX ratio on removal
efficiency for Eneron process ............... 179
36 Optimum test conditions-utility pilot plant for
Eneron process. ..... . 180
37 Effect of inlet NH3 concentration on. NOX removal
.efficiency at various temperatures for Eneron process . . . 181
38 Effect of space velocity on NOX removal efficiency
for Eneron process. .................... 182
39 Effect of space velocity on removal efficiency for
Eneron process. 183
40 Effect of inlet NH3 concentration on NOX removal
efficiency and NH3 in effluent for Eneron process ..... 184
41 Effect of NH3 on NOX removal efficiency for Eneron
process ........ . 185
42 Flow diagram of Exxon Thermal Denox process. ........ 188
43 Effect of temperature on NO removal and NH3 concentra-
tion without H2 addition for Exxon Thermal Denox
process ................. 189
44 Effect of temperature on NOX removal and NH3 concen-
trations with H2 addition for Exxon Thermal Denox
process , 190
45 Effect of NH3INO mol ratio and various Q£ levels on
NO reduction for Exxon Thermal Denox process. . 192
46 Effect of NH3:NOX mol ratio on NOX removal efficiency
during commercial tests for Exxon Thermal Denox
process ...... .............. 193
47 Flow diagram of Foster Wheeler - Bergbau Forschung
dry adsorption process 201
48 Effect of space velocity on NOX removal efficiency
for FW-BF process ......... 206
49 Flow diagram of Hitachi, Ltd., process ...... . 209
50 Relationship between reaction temperatures and NOX
removal efficiency for Hitachi, Ltd., process ....... 212
51 Relationship between space velocity and NOX removal
efficiency for Hitachi, Ltd., process . .... 213
vi
A
-------
FIGURES (continued)
Number
52 Characteristic curve tof the effect of mol ratio of
NH3:NOX on NOX removal efficiency for Hitachi, Ltd.,
process ..,..,./.. 214
53 Flow diagram of Hitachi Zosen process 217
54 Flow diagram of JGC Paranox process 222
55 NOX removal efficiency, NOX concentration, and pressure
loss over 2,000-hr test period for JGC Paranox process. . •. 224
56 Effect of NO initial concentration on NOX removal
efficiency for JGC Paranox process. . 225
57 Effect of D£ concentration on NOX removal efficiency for
JGC Paranox process 226
58 Effect of H20 concentration on NOX removal efficiency
for JGC Paranox process 227
59 Effect of NHjiNO mol ratio on NOX removal efficiency
for JGC Paranox process 229
60 Effect of temperature on NOX removal efficiency for
JGC Paranox process 230
61 Flow diagram of Kobe Steel process 232
62 Catalyst reactor of KSL moving bed 235
63 Flow diagram of Kurabo Knorca process 239
64 Flow diagram of Kureha process ....... 246
65 Flow diagram of Mitsubishi Heavy Industries process. .... 252
66 Results from MHl's pilot-plant tests showing effect of
oxygen on NOX removal efficiency 255
67 Structure of Mitsubishi Heavy Industries' moving-bed
reactor 257
68 Pressure drop variation with MHI's intermittent moving
catalyst bed. ..... 258
69 Results of tests with fixed-bed reactor and dirty flue
gas after use of an ESP showing effects of temperature
• and space velocity on NOX removal efficiency for MHI
process 259
70 NOX removal efficiency and outlet gas NH3 concentra-
tion at different boiler loads for MHI process 260*
71 NOX removal efficiency and NH^ emission in relation
to NH3:NOX mol ratio and operating period for MHI
process .......... . . 261
72 Effect of space velocity on NOX removal efficiency
and NHj emission for MHI process. 262
73 Results of MHI's tests on NOX removal efficiency and
NH3 emission using an NH3 converter to reduce NH3
emission. 264
74 Deposit of NH3~SC>4 compounds in air reheater from
MHI tests 265
75 Flow diagram of Mitsubishi Kakokl Kalsha (MKK)
process 267
vii
-------
FIGURES (continued)
Number
76 Effect of 02 concentration on NOX removal efficiency
with MFC catalyst 272
77 Effect of NH3:NOX mol ratio on NOX removal efficiency
with MFC catalyst ........... 273
78 Effect of temperature on NOX removal efficiency
with various types of MFC catalysts 275
79 Relationship of NOX removal efficiency to temperature
and space velocity for various MFC catalysts . 276
80 Flow diagram of Mitsui Engineering and Shipbuilding
process ....... ..... 279
81 Flow diagram of Mitsui Toatsu Chemicals process . 285
82 Flow diagram of The Ralph M. Parsons Company process .... 290
83 Flow diagram of alternative Ralph M. Parsons process .... 294
84 Flow diagram of Sumitomo Chemical process 297
85 Relationship among inlet NH3:NOX mol ratio, NOx
removal efficiency, and exiting NH3 concentration
using the Sumitomo Chemical C-l catalyst 298
86 Relationship among inlet NH3:NOX mol ratio, NOX
removal efficiency, and exiting NH3 concentration
using an NH3 decomposition catalyst with the C-l
catalyst 299
87 Comparison of NOX removal efficiency with Sumitomo
Chemical's S0x-resistant D catalyst and C-l catalyst. . . . 300
88 Temperatures below which NlfyHSO^ forms ..... 302
89 Simplified flow diagram of the flue gas treatment
facility at Higashi Ninon Methanol plant 303
90 Result of life test for Sumitomo Chemical catalyst 305
91 Flow diagram of Sumitomo Heavy Industries process. ..... 308
92 Effect of temperature on NOX and S02 removal effi-
ciencies with SHI carbon-based catalyst 313
93 Effect of temperature and space velocity on NOX and
SC-2 removal efficiencies for SHI process. ......... 314
94 Effect of NH3 Injection rate on NOX and S02 removal
efficiencies for SHI process 315
95 Flow diagram of Sumitomo Heavy Industries process. ..... 318
96 Effect of temperature and S02 on NO removal effi-
ciency with SHI metal-oxide catalyst. . 320
97 Flow diagram of Takeda Chemical Industries, Ltd.,
process (regeneration by washing) . . 324
98 Flow diagram of Takeda Chemical Industries, Ltd.,
process (regeneration by heating) ..... 325
99 Relationship of 02 and H20 concentration in flue
gas to NOX removal efficiency for Takeda process. ..... 328
100 Relationship of linear velocity and pressure drop
for Takeda process. 329
A
-------
FIGURES (continued)
Number
101 Relationship of space velocity and NOX removal effi-
ciency for Takeda process 330
102 Relationship of temperature and NOX removal efficiency
for Takeda process. ............ 331
103 Flow diagram of Ube Industries process ... ... 335
104 Flow diagram of Unitika, Ltd., process ..... 340
105 Relationship of S(>2 removal efficiency to reaction
temperature and space velocity for Unltlka process. .... 344
106 Relationship of reaction temperature and space
velocity on NOX removal efficiency for Unitika
process ......... .... 345
107 Flow diagram of OOP process 351
108 Diagram of alternative methods used with UOP process
for the following sections: (1) regeneration gas
supply, (2) flow smoothing, and (3) workup section 353
109 Performance of Shell FGD reactor at SYS; instan-
taneous S02 and NOX slip. 359
110 NOX slip, percent of intake vs loading, CuSC>4/
(CuO + CuS04) mol ratio for a commercial Shell
FGD acceptor/catalyst with UOP process 360
111 Flow diagram of UOP process. 363
112 Unconverted NOX as a function of catalyst bed
length for UOP process 368
A-l Welsbach ozonator. . , 390
A-2 Ozone generation systems .................. 391
-------
TABLES
Number
S-l Current Flue Gas Denitrification Processes ........ xviii
S-2 Comparison of Dry NOX Removal Processes . xxii
S-3 Comparison of Wet NOX Removal Processes ......... xxv
S-4 Comparison of the Advantages of the Current Flue Gas
Denitrification Processes ..... ... xxviii
S-5 Comparison of the Disadvantages of the Current Flue
Gas Denitrification Processes xxlx
S-6 Processes Recommended for Further Study in Phase II ... xxxi
1 NOX Emissions in the United States 2
2 NOX Emissions from Selected Stationary Sources in
the U.S. in 1972. 3
3 NOX Emission Standards and Projected Research Objectives
for Large Fossil Fuel-Fired Boilers 3
4 Effects of Boiler Modifications to Reduce NOX Emissions
of Fossil Fuel Type . 5
5 Outline of the Detailed Technical Evaluation of Each
Process ..... 6
6 Concentration Ranges of NOX from Coal-Fired Power
Plants 9
7 Typical Flue Gas Compositions from Oil- and Coal-Fired
500-MW Boilers 15
8 Chemical Composition of Flyash from a Coal-Fired
Boiler . 16
9 Comparison of Wet NOX Removal Processes ......... 20
10 Advantages and Disadvantages of Absorption-Oxidation
Processes 21
11 Advantages and Disadvantages of Oxidation—Absorption
Processes 22
12 Advantages and Disadvantages of Oxidation-Absorption-
Reduction Processes ...... 24
13 Comparison of Oxidation-Absorption-Reduction Processes . . 25
14 Advantages and Disadvantages of the Absorption-
Reduction Processes 26
15 Comparison of Absorption-Reduction Processes 27
16 Comparison of Dry NOX Removal Processes . 153
17 Advantages and Disadvantages of SCR Processes 156
18 Process Conditions and Economics for Dry SCR NOX
Removal Processes ..... ... 157
19 Advantages and Disadvantages of Nonselectlve
Catalytic Reduction Processes 158
A
-------
TABLES (continued)
Number • Page
20 Process Conditions and Economics for Dry NOX
Removal Processes Other than SCR Type 159
21 Advantages and Disadvantages of Selective,
Noncatalytic Reduction Processes . 160
22 Advantages and Disadvantages of Adsorption Process .... 161
23 Advantages and Disadvantages of Radiation Process 162
24 Summary of Catalyst Performances: NO (about 300 ppm)
in N2 . 174
25 Results of Coal-Fired Bench-Scale Tests . 198
26 Results of Bench-Scale Tests with Synthetic Flue Gas ... 199
27 Bench-Scale Results on N0x-0nly Removal. ... 199
28 Actual or Designed Operating Conditions of NOX
Removal Pilot Plants by Mitsubishi Heavy Industries ... 254
29 Bench-Scale Test Conditions for Unitika (NOX)
Process , . . . 348
30 Processes Selected for Further Study in Phase II ..... 372
31 Status of Development of the Flue Gas Denitrifi-
cation Processes 375
A-l Summary of Capital Investment, Revenue Requirements,
and Energy Requirements for the Oxidation of Flue
Gas NO by 03 in 500-MW Oil-Fired and Coal-Fired
Plants. 392
8-1 Solubility of Various Gases in H20 , 396
xi
-------
ABBIEVIATIONS AM) CONVERSION FACTORS
ABBREVIATIONS
3
aft actual cubic feet kWh
bbl barrel 1
Btu British thermal unit Ib
°C degrees Centigrade k
cm centimeter m
dia diameter M
ESP electrostatic preclpitator mg
°F degrees Fahrenheit mln
FGD flue gas desulfurization mm
FGT flue gas treatment mol
ft feet MW
ft/sec feet per second MW equiv
g gram Nm3
G billion ppm
gal gallon SCI
gpm gallons per minute sft-Vhr
gr grain
hr hour sec
in. inch ., sv
kg kilogram '• vol
kl " kiloliter wt
kW kilowatt yr
kilowatthour
liter
pound
thousand
meter
million
milligram
minute
millimeter
mole
megawatt
megawatt equivalent
normal cubic meter
parts per million
selective catalytic reduction
standard cubic feet per hour
standard cubic feet per minute
second
space velocity
volume
weight
year
xii
-------
CONVERSION FACTORS
To convert from
To
Multiply by
British thermal unit
degrees Fahrenheit-32
dollars
feet
cubic feet
feet per minute
cubic feet per minute
gallons
gallons per minute
gallons per thousand actual
cubic feet <50°C)
grains (troy)
grains per cubic foot
megawatt (electric)
megawatt (electric)
mills per kilowatthour
pounds
standard cubic feet per
minute (32°F)
tons (short)
tons (long)
tone per hour
gram-calories 252
degrees Centigrade 0.5555
yen 0.00333
centimeters 30.48
cubic meters 0.02832
centimeters per second 0.508
cubic meters per second 0.000472
liters 3.785
liters per second 0.06308
liters per normal cubic meter 0.1608
grams 0.0648
grams per cubic meters 2.288
normal cubic meters per hour 0.000333
standard cubic feet per minute 0.000535
yen per kiloliter of oil 1248.a
kilograms 0.4536
normal cubic meters per hour 1.695
(0°C)
metric tons 0.90718
metric tons 1.016
kilograms per second 0.252
a. Assuming 300 yen/dollar, 144,000 Btu/gal of oil, and 9,000 Btu/kWh.
xiil
-------
Executive Summary
With the increasing application of particulate removal systems and
the expansion of the development and application of flue gas desulfurization
(FGD) systems, regulatory attention within the U.S. Environmental Protection
Agency (EPA) has begun to shift to controlling a third emission component
present in combustion flue gases, nitrogen oxides (NOX). Largely, this shift
is due to the inability of automobile manufacturers to develop a viable NOX
emission control technology for mobile sources without incurring significant
fuel economy penalties. More recently, the 1977 Clean Air Act Amendments
have increased the impetus for more stringent NOX regulations. Combustion
modification techniques currently under research by the Electric Power
Research Institute (EPRI) and SPA may provide sufficient control of the
power plant contribution to NO emissions. However, since the success of
these processes is unknown at this time and since the ultimate regulatory
posture to be adopted regarding NOX emissions la as yet undefined, alter-
native, though potentially less cost-effective, approaches cannot be ruled
out.
This report covering the first phase of a multiphase NOX removal process
study, is & state-of-the-art survey of all known flue gas denitrlfication
processes currently undergoing development in the U.S. and Japan. Because
low NOX emission limits have been emphasized in Japan in recent years, flue
gas denitriflcation technology is much more diversified in that country;
therefore, most of the processes included in the first phase of this study
were developed and are being marketed by Japanese companies. During the
preparation of this report, various Japanese and American companies were
contacted for pertinent technical and economic information outlining their
NOX removal processes. This information and the data obtained from other
sources were combined to describe the major technical and economic aspects
of each process. Major sections discussed for each process Include a pro-
cess description, the status of development, the reported economics, utility
and raw material requirements, technical and environmental considerations,
and the advantages and disadvantages. The 48 flue gas denitrification pro-
cesses currently listed in the literature are included in Table S-l and are
classified into types and subcategories which will be described later.
Since essentially no information is available on 6 of the processes, only
42 processes are described in the report. In addition to describing the
xiv
A
-------
TABLE S-l. CURRENT FLUE GAS DENITRIFICATION PROCESSES
Dry processes
Wet processes
Selective catalytic reduction
Asahi Glass8
Eneron
Exxon
Hitachi, Ltd.
Hitachi Zosen
JGC Faranox
Kobe Steel
Kurabo Knorca
Kureha
Mitsubishi Heavy Industries
Mitsubishi Kakoki Kaiaha
Mitsubishi Petrochemical
Mitsui Engineering and
Shipbuilding
Mitsui Toatsu
Nippon Kokana
Sumitomo Chemical
Sumitomo Heavy Industries*5
Takeda
Ube
Unltikab
Universal Oil Products-Shell
Nonselectlve catalytic reduction
The Ralph M, Parsons
Selective noncatalytlc reduction
Exxon Thermal Denox
Adsorption
Foster Wheeler-Bergbau Forschung
Radiation
Ebara-JAERI
Uncertain
Princeton Chemical Researcha
Absorption-reduction
Asahi Chemical
Chisso Engineering
Kureha
Mitsui Engineering and
Shipbuilding
Pittsburgh Environmental and
Energy Systems
Oxidation-absorption-reduction
Chlyoda Thoroughbred 102
Ishikawajina-Harima Heavy
Industries
Mitsubishi Heavy Industries
Moretana Calcium
Moretana Sodium
Osaka Sodaa
Shirogene3
Absorption-oxidation
Hodogayaa
Kobe Steel
MON Alkali Permanganate
Nissan Engineering
Oxidation-absorption
Kawasaki Heavy Industries
Tokyo Electric-Mitsubishi HI
Ube
a. Essentially no information available.
b. These companies have two dry processes:
(2) S02-NOX.
(1) N0x-only and
-------
various alternative processes, another major objective of this first phase
is to recommend processes for preliminary economic analysis as a. second
phase study.
COMPARISON OF DRY AND WET NOX REMOVAL PROCESSES
The flue gas denltrification processes can be initially separated into
two types, wet or dry, depending on whether or not the NOX is absorbed into
an aqueous solution. With a few notable exceptions the dry processes are
NOjj-only removal systems while in general the wet processes are simultaneous
sulfur dioxide (862) and NOX removal systems. Although there are a few
examples of wet N0x-only technology, these processes were originally developed
for treatment of nitric acid (HNC>3) plant tail gas and may not compete eco-
nomically with the much simpler dry N0x-only units. Therefore, most of the
initial development work on wet NOX removal processes has been in adapting
the existing wet FGD technology to simultaneously remove both S02 and NOX.
For a variety of reasons, including simplicity, more favorable economics,
and the fact that most existing power plants in Japan operate with low-sulfur
(S) oil as fuel with small amounts of particulate and S02 flue gas emissions,
the dry N0x-only removal processes seemed the most promising and were developed
first. The initial research for most of the wet processes was begun later
and, hence, the dry processes as a group have been more extensively tested
and are more commercially acceptable.
Although there are many different types of dry and wet processes, in
most cases the dry processes have the following advantages over the wet
processes.
1. Lower projected total capital investment and lower annual revenue
requirements
2. Simpler process with few equipment requirements
3. Higher NOX removal efficiency (>90%)
4. More extensive tests in large units (oil- and/or gas-fired boilers)
5, No waste stream generation
However these dry systems also have the following disadvantages.
1. More sensitive to inlet particulate levels
2. Requirement for ammonia (NHj) from either an energy-sensitive source
(natural gas) or more expensive coal gasification, methods _
3. Possible emission of NH3 and ammonium sulfates ./.(NH^^SO^/ and
bisulfates (M4HS04); precipitation of same may create fouling of
downstream equipment o
4. Relatively higher reaction temperatures (350-400 C) which must be
located in the power generation cycle before the air preheater or must
be attained by auxiliary heating after the preheater
The most critical of these disadvantages, particularly for the U.S.
utility industry with its heavy reliance on coal for power generation, is
the sensitivity of these processes to inlet particulate levels. However,
major research is now underway to develop methods to enable dry systems to
xvi
A
-------
handle flue gas with high particulate loading. Evaluations are being per-
formed on coal-fired flue gas; however, these tests have not been executed
on a very large scale.
Another disadvantage of the dry, selective catalytic reduction (SCR)
processes is that the ideal catalyst location may be in the region between
the economizer outlet and the air preheater inlet and, hence, the process
is intimately involved in the power generation cycle. Therefore, if prob-
lems of operating these processes occur, the adverse impacts on the basic
utility operations may be greater.
In addition to the above-mentioned disadvantages, the long-term supply
of NH3 for these dry NOX removal processes is a potential problem. NH3 is
the reducing agent for converting NOX to molecular nitrogen (N2) for the
SCR processes (which comprises nearly all of the dry processes and about
half of all the NOX removal processes—see Table S-l) and the selective non-
catalytic reduction process. With an NH3:NOX mol ratio of about 1:1 a
single 500-MW coal-fired power plant (600 ppm NOX in the flue gas) would
consume about 5950 tons/yr of liquid anhydrous NH3, In view of the con-
tinuing increase in the world's demand of NH3 and NH3~based fertilizers,
the availability of NH3 for larger numbers of these dry NOX removal units
warrants concern and further investigation.
The wet NOX removal processes have certain general advantages and dis-
advantages as compared with the dry systems. These major advantages include;
1. Simultaneous S02~NOX removal may be a potential economic advantage
2. Relatively insensitive to flue gas particulates
3. Higher S02 removal (>95%)
On the other hand the major disadvantages of these wet systems include:
1. More expensive processes due to the insolubility of NOX in aqueous
solutions
2. Formation of nitrates (N03~) and other potential water pollutants
3. More extensive equipment requirements
4. Formation of low-demand byproducts
5. Flue gas reheat required (however, if a wet S02 removal system were
used in series with a wet removal system for NOX only, the reheat
would have already been incorporated into the design)
6, Only moderate NOX removal
7. Application of some processes may be limited to flue gas with high
SOX:NOX ratio
The two primary disadvantages of the wet systems are the high capital
and operating costs and the formation of N03~-containing wastewater. The
generation of N03™ salts in most of these processes results in the need to
remove these salts from the effluent by either evaporation or biological
treatment.
xvii
-------
Another important factor which must be considered in the application of
certain wet NOX removal processes Is the minimum SOX:NOX ratios required.
It may not be feasible to operate wet NOX removal processes using flue gas
from Western U.S. coals, which characteristically possess low amounts of S,
since sufficient SOX:NOX ratios may not be achieved for adequate NOX removal.
DRY NO REMOVAL PROCESSES
The dry flue gas denitrlficatlon processes can be subdivided into five
major categories. These categories and the number of processes of each type
Included in this study are shown below.
_ Category _ Number of processes
Selective catalytic reduction 22
Nonseleetive catalytic reduction 1
Selective noncatalytlc reduction 1
Adsorption 1
Radiation 1
Table S-2 shows a general comparison of these various types of dry processes.
As is readily apparent from this list, the overwhelming majority of the
dry systems currently undergoing development are based on the SCR method
and use NHj ae the reductant. In this method the anhydrous NH3 Is injected
into the flue gas after the boiler economizer and the resulting mixture is
passed over a proprietary base-metal catalyst. The NH3 selectively reduces
the NOX to molecular N2 which then passes out of the NOX removal system
and into the boiler air heater.
The primary disadvantage associated with the SCR method when applied
to coal-fired flue gas is the sensitivity of the catalyst to the higher
particulate levels in the flue gas. Although most of these processes have
been designed to minimize the effects of dust, either through the type of
reactor used or shape of the catalyst particles, this development work was
done using heavy oil-fired flue gas. Additional detailed pilot-plant testing
on coal-fired flue gas will be needed and Is underway by some companies to
confirm the ability of these innovations to handle the higher particulate
loadings associated with coal combustion. Upon completion of these tests,
further development on even larger scale will be necessary.
One additional potential problem is the formation and precipitation of
ammonium bisulfate (NttyHSOjj) downstream from the reduction reactor, partic-
ularly in the boiler air heater. The NH^HSO^ formation is dependent upon
temperature and the NH3 and sulfur trioxlde (863) concentrations. It may
become necessary to decrease the mol ratio of N§3:NOX and thus decrease
the denitrification efficiency below 90% to prevent the formation of NlfyHSO^
The use of a catalyst for the decomposition of NHj would provide an alter-
native method of control though it would Increase complexity of capital
investment. The development of such catalysts is being conducted by some
companies.
xviii
A
-------
TABLE S-2. COMPARISON OF DRY NOX REMOVAL PROCESSES
tf
Dry NOX removal process type
Selective Nonselective
catalytic catalytic
Process characteristics3 reduction reduction
Simultaneous S02~NOX removal
Achieves moderate S02 removal
f*** Q c *y \
i ^ O jyb /
Achieves high NOX removal (>90%)
Operating conditions
Produces waste stream
Uses NH3
Forms NH4HS04
Operates with sensitivity to
particulates
Produces marketable byproduct
Current development status
Tested on coal-fired flue gas
Tested on pilot plant or larger
scale
-b x
X
X X
- -
X
X
X
-b X
_b
V _
Selective
noncatalytic
reduction Adsorption
X
X
-
X
X
X
X X
X
X
X X
Radiation
X
X
X
X
-•
-
-
-
-
X
a. An "X" indicates the process has this characteristic.
b. Depends on process.
-------
The nonselective catalytic reduction processes involve the injection
of a fuel or reducing gas into the radiant zone of the boiler to chemically
bind the excess oxygen (02) and thus minimize the formation of oxides of
S and M. For economical operation, this use of a reducing gas will be com-
bined with combustion modifications such as firing with a slightly substoich-
iotaetric amount of air to decrease the consumption of reducing gas as much
as possible. As this (^-deficient flue gas containing some S02 and NOX is
passed over a nonnoble metal catalyst, the S(>2 and NOX are reduced to
hydrogen sulfide (H2S) and molecular ^2- The H2S is then removed in a con-
ventional Stretford unit and the remaining flue gas is exhausted through
the stack. This type of process, in addition to simultaneously removing
SC>2 and NOX, produces elemental S as a marketable byproduct. The major
disadvantages Include the expense of the reducing gas and the possibility
of increased corrosion in the boiler due to the high temperature reducing
atmosphere.
In the selective noncatalytlc reduction processes the NH3 is injected
directly into the upper portion of the boiler to selectively reduce the
NOX to molecular $2- This procedure eliminates the need for any supplemental
equipment downstream and results in a process scheme with minimum capital
investment. Unfortunately it has the major disadvantages of requiring a
higher NI^NOx mol ratio and thereby potentially creating more problems
with NH4HS04 than catalytic processes, operating In a very narrow temperature
range, and yielding only low NOX removal (40-60%). There is also uncertainty
concerning the effects of flyash from coal-fired flue gas upon the NOX
removal efficiency of this process.
The dry adsorption processes are based on the use of activated carbon (c)
to adsorb both S02 and NOX from the flue gas. The adsorbent is regenerated
at a high temperature to yield a concentrated off-gas stream of S02 and
molecular $2* This SC>2 could be used to generate elemental S or byproduct
sulfuric acid (12804).
The disadvantages of this type process include low NOx removal effi-
ciency and high C loss.
The radiation process is unique in that the flue gas is bombarded with
an electron beam. This radiation converts the particulates , S02, and NOX
into a powdery, complex mixture which is then removed in an electrostatic
precipitator (ESP), The major disadvantages include high initial capital
investment, high annual revenue requirements, secondary waste disposal
problem, and low (80%) S02 removal efficiency.
WET NOX REMOVAL PROCESSES
The wet denitrification processes can also be subdivided into four
major categories. These categories and the number of processes of each
type included in this study is as follows :
xx
A
-------
Category Number of processes
Absorption-reduction 5
Oxidation-absorption-reduction 5
Absorption-oxidation 3
Oxidation-absorption 3
An overall comparison of these categories of wet processes is given in
Table S-3,
The absorption-reduction technology is based on using a water (H20)-
soluble ferrous-ehelating compound; e.g., Asahi Chemical Industry Company
uses ferrous ethylene diamine tetraacetic acid (Fe+^*EDTA) as a catalyst
to aid in the absorption of the relatively insoluble NOX. These ferrous
compounds have the ability to form complexes with the NOX and thus promote
the absorption of the NOX (primarily_NO). Once in the solution the NOX
can be reduced by the absorbed S02 /i.e., the sulfite (803™) ion__/ to form
molecular N2 or reduced N compounds.
Even with the use of this absorption catalyst, the NO is sufficiently
insoluble that the gas-liquid contact time must be increased; therefore,
high liquid-to-gas (L/C) ratios and low superficial gas velocities are
required. The L/G ratio in the absorber is normally about 15 1/Nm3 (93 gal/
kaft3 at 127°F) and the superficial gas velocity is in the range of 1-3
m/sec (3-10 ft/sec).
Unfortunately these absorption-reduction processes also have other
significant disadvantages in addition to the large absorption section. Most
of these processes have extensive equipment requirements and the NOX removal
efficiencies are also sensitive to the inlet flue gas composition, partic-
ularly the inlet 02 and S02 concentrations. These processes have only been
tested in bench-scale units for treating oil-fired flue gas,
The oxidation-absorption-reduction processes, on the other hand, are
in most cases simple modifications of commercially available FGD technology.
To use the conventional FGD scrubbers, the insoluble NO in the flue gas
must be chemically converted to increase its solubility. A gas-phase oxi-
dant, such as ozone (03) or chlorine dioxide (C102), is injected into the
flue gas to selectively oxidize insoluble NO to the more soluble N02« For
processes using H20-soluble absorbents and thus generating i^O-soluble
sulfites when absorbing the S02 from the flue gas, the simple injection of
a gas-phase oxidant is all that is necessary to convert a conventional FGD
system into a simultaneous S02~NOX removal system. If, however, the S02
absorbent is relatively insoluble in aqueous solution and forms insoluble
SOg**, for example limestone scrubbing, the simple addition of a gas-phase
oxidant will not yield good NOX removal. An H20-soluble catalyst must also
be present in the solution to supplement the gas-phase oxidant and give
good NOX removal efficiency.
Oxidation-absorption-reduction processes, then, are commercially
available FGD systems which have been modified to give both S02 and NOX
xx 1
-------
TABLE S-3.
COMPARISON OF WET NOX REMOVAL PROCESSES
Process characteristics3
Simultaneous S02~NOX removal
Achieves high S02 removal (>95Z)
Achieves moderate NOX removal
(>85Z)
Operating conditions
Requires absorption catalyst
Requires liquid-phase oxidant
Requires gas-phase oxidant
Requires large absorber
Requires flue gas reheat
Forms nitrate salts in wastewater
Requires specific range of flue
gas constituents
Current development status
Tested on coal-fired flue gas
Tested on pilot plant or larger
scale
Wet
Absorption-
oxidation
_
—
X
-
X
-
X
X
X
X
-
-
NOV removal
Oxidation-
absorption
*
XD
X
-
-
X
X
X
X
-
t-
xb
X
process type
Oxidation-
absorption-
r eduction
X
X
X
Xb
-
X
-
X
xb
&
-
X
Absorption
reduction
X
X
-
X
. -
-
X
X
-
X
Xb
xb
a. An "X" indicates the process has
b. Depends on process.
this characteristic.
-------
removal. Other advantages include good NOX removal efficiencies (85-90%),
excellent SC>2 removal efficiencies (>95%), and these processes are also
relatively Insensitive to the inlet flue gas composition.
However, these types of processes have two serious drawbacks. The most
serious is probably the cost of the gas-phase oxidant. For the rapid, se-
lective oxidation of NO, either 03 or C102 must be used. Both are extremely
expensive and must be generated onsite (5, 19). 63 is the more expensive of
the two but has the advantage of not releasing any additional pollutants
into the flue gas. C102, although costing only about one-fourth as much
as 03, unfortunately releases additional Cl~ and NOj" into the scrubbing
solution. The second major problem with the oxidation-absorption-reduction
processes, regardless of which gas-phase oxidant is used, is that they
generated a wastewater N0g~ stream which must be treated before it can be
released. Depending on the process involved, the percentage of the absorbed
NOX converted to W%~ salts ranges from 10-50%. The two most common waste-
water treatment methods for this soluble N0-j~ are evaporation and biological
denitrlfication.
Although the oxidation-absorption-reduction and absorption-reduction
processes have several major disadvantages, the absorption-oxidation and
oxidation-absorption processes have these same problems as well as several
additional drawbacks. The absorption-oxidation processes have similar
problems to those given for the absorption-reduction processes, i.e., forcing
the insoluble NO into the aqueous scrubbing solution and hence requiring a
large and expensive absorber, Also, these processes are further complicated
by the fact that they use a liquid-phase oxidant, either a permanganate or
hypochlorlte solution, to convert the absorbed NOX to N03~ salts. This con-
version of the absorbed NOX to highly soluble N0o~ salts is undesirable
since these NC>3~ salts are difficult to remove and cannot be released as
a wastewater stream. The use of an expensive liquid-phase oxidant prevents
the introduction of the ferrous-chelating compound to aid in the absorption
of NOX. For economic reasons the use of a liquid-phase oxidant also limits
these processes to NOX removal only since the oxidant would readily convert
the SOj™ Ion, formed from the absorption of S02, to sulfate (SO^**). Thus
the S02 must be removed before entering the NOX absorber to prevent the
excessive consumption of expensive oxidant.
The oxidation-absorption processes, unlike the other classes of wet
flue gas denitrification processes, have no common treatment mechanism.
These denitriflcation systems are similar only in the fact that they con-
tain an initial gas-phase oxidation stage and follow this with an absorp-
tion stage. The example processes range from the straightforward gas-phase
oxidation and absorption to the equimolar absorption of NO and N0£ using
a recycle NC^-rich stream.
In addition to their general class disadvantages, such as the formation
of wastewater NO,"" solutions and the requirement for a gas-phase oxidant,
each type of oxidation-absorption has a unique set of treatment problems.
The straightforward oxidation-absorption type uses a high oxidant:NOX mol
ratio and an S02~free flue gas source to recover the NOX as a weak HN03
xxiii
-------
solution. The excess oxidant must then be removed in a separate closed-loop
absorption section. The other type of oxidation-absorption process, equimolar
absorption of NO and N02» requires the generation of an N02-rlch recycle
stream and a two-stage NOX absorber. The latter is needed since this equi-
molar absorption mechanism is limited to a final outlet NOX concentration of
150-200 ppm. Further NOX removal (i.e., 90% overall removal) requires the use
of a gas-phase oxidant and a second absorption stage.
COMPARISON OF THE ADVANTAGES AND DISADVANTAGES OF THE INDIVIDUAL FLUE GAS
DENITRIFICATION PROCESSES
A detailed technical evaluation of each of the flue gas denitrlfication
processes included in this study was prepared from information already avail-
able in the public domain and from direct contacts with the various process
developers. After this information had been assimilated into several sections
including a process description, a block flow diagram, technical and environ-
mental considerations, and the current status of development, a list of
advantages and disadvantages for each of the individual flue gas denitrifica-
tion processes was prepared. Table S-4 contains a comparison of the various
advantages while Table S-5 shows the disadvantages of each of the flue gas
denitriflcation processes Included in this study. These advantages and dis-
advantages are based on information available at the present time and can
be expected to change as new information becomes available. Also, the stated
technology may not be optimum for example, the use of auxiliary heaters by
most of the dry processes may be circumvented in future designs.
CONCLUSIONS AND RECOMMENDATIONS
The most tested and advanced NOX flue gas treatment (FGT) method now
is SCR. Several commercial, oil-fired SCR denitrlfication units have been
operated and about half of the total denitrificatlon processes included in
this report are the SCR type. However, most of the overall FGT development
work performed to date has been with bench-scale and pilot-plant size opera-
tions on gas- or oil-fired flue gas. Also, there are several drawbacks to
be considered and solved in future development regarding both dry and wet
NOX removal processes. Therefore, before any FGT processes are feasible
for application to commercial, coal-fired power plants, further testing must
be done on pilot-plant and prototype units with coal-fired flue gas. It
should be noted, too, that the economics for the processes reported herein
may be subject to revision as the processes are developed further on a larger
scale.
A prime objective of this study, in addition to being a state-of-the-art
review of all NOX processes undergoing development, is to recommend processes
for further evaluation in the second phase (Phase II) of this NOX removal
process study. This second phase will be a preliminary economic analysis of
the processes selected during this first phase.
xxiv
A
-------
TABLE S-4. COMPARISON OF THE ADVANTAGES OF THE
CURRENT FLUE GAS DENITRIFICATION PROCESSES'
Advantage0
Removes NOx & S02 simultaneously.
Achieves (> 95%) S(>2 removal.
Produces potentially marketable
byproduct .
a. Elemental sulfur
b. Sulfuric acid
c. Gypsum
d. Other
flue gas on pilot-plant scale
CfSt greater scale.
Has been applied to flue gas from
a commercial oil-fired boiler.
Is a slight modification of a
commercially available FGD system.
Operates with full particulate
loading.
Operates with moderate particulate
loading.
Claims less than 10 ppm by volume
NHi in treated flue gas.
Claims full turndown capability.
Operates between 100 & 200°C, which
may require only negligible reheat.
Does not require chemical raw
materials.
Requires no additional post-boiler
processing equipment.
Uses catalyst which, when spent,
has other potential applications.
5 S? S
*"" o -o
~-< « O
•g j: s
S € 5
x
X
X
yd
X
X
X
Xe
X
X
X
X
X
X
Wet SO* Removal Processes
lahlkawajlma HI (OAR)
Kawaiakl HI (QA)
Kobe Steel (AO)
X
X
X
X
X
X
X
X
X
X
B M
1 3 1 3
5 1 3 3
• . •>
3 -< •* O
^ K £ £
X
X
X
X
X
X
X
X
X
X
X
X
X1
X
Ji£L
X
ll
* s. I
3 $ g
Z E S
X
X
X
X
X
X
xd
X
X
1
i
*"" M
B
X O
1 1
X
X
X
X
Xs
1 I
: g
s s
U M
x
x
X
Exxon (Thermal) (SNR)
Foiter Wheeler (A)
Hitachi, Ltd. (SCR)
X
X
X
X
X
x
X
X
X"
X
g
CO
X
X
X
B
.III
B • c en
Su
o> O •
£4 •£ X
. a o
8Z M 2
033
-1 * X X
X
X
X
X
ry NOx Removal Processes
g g |
g g 7 " g ~
X E S " S
-i ^ •».
•S -S « H .
11 is s .=
.0 0 « Ch
s » je S 5
X
X
X
X
X
X
I g
*-• in at at
8w 'x mm
O r-> *-* *-* "~?
m m c g
•3 t V S £ 2
1 g • i - ? 51
6!e£8^gg^
U4J *J-O*^^*J^'
<^ v4
-------
TABLE S-5. COMPARISON OF THE DISADVANTAGES OF THE
CURRENT FLUE GAS DENITRIFICATION PROCESSES3
&yx
3? v\
^x* ^B&k
Disadvantage11
Can only achieve a •"•*•"• NO^ removal
efficiency of c 70t
Forms secondary source of pollution
a. Vaatewater nitrate salts
b. Wi emissions (potential)
c. JH3 off-gas (151)
d. HS03 (251)
e. Absorbent-ash sludge
Requires significant amounts of energy
for the regeneration step
Has not been developed beyond the conceptual
design stage (for treating flue oas)
Has been tested only in a bench-scale unit
Has not been operated as an Integrated process
Has not been operated for a long-term
continuous period
Requires SQ2-free gas feed
Incorporates design features which may present
Incorporates questionable design features
a. Final-stage KEi scrubbing
b. L/G ratio appears much too low
o. Placing catalvst in air heater
d. Stacking of catalyst in reactor may
by required
Requires hot solids handling
Uses moving-bed reactor which increases
maintenance and catalyst attrition
Has low space velocity in the reactor ( c 5000 hr"1)
Has a low superficial gas velocity In the
absorbent (^ 10 ft/sec)
Has a high L/C ratio in the absorber f ^70 eal/kaft^)
Requires flue gas constituents within specific
ranges for high NO* removal
reaction with SO?
Requires relatively large NH3:NOX awl ratio for
equivalent HOx removal
Requires an expensive liquid-phase oxidant
Requires an expensive liquid-phase catalyst
Incorporates unique treatment methods to prevent
secondary sources of pollution (biological wastewater
denitriflcation)
|
X
X
X
X
X
X
X
X
ChlSBO
X
X
X
X
X
X
X
X
Chlyoda
X
X
Ishlfcawajlma-Harlm*
X
x
He
n:
X
X
X
X
t NT*
V
«j
m
1
X
X
X
X
X
x re
!
X
X
X
X
x
X
•ova
Httlubllhl HI
X
X
X
X
1 pr
Shlpbldg.
X
~x
X
X
X
X
X
oces
5
2
S
X
X
X
sea
Moretana Calcium
X
X
Moretana Sodium
X
X
Nissan
X.
X
X
X
X
s
X
X
X
X
X
X
Tokyo Electric -
HHI
X
X
•-t
*
1
X*
1
X
X
X
X
!
X
X
X
X
s
i
i
JJ
&
X
X
X
X
X
Hitachi, Ud.
Hitachi Zoian
u
X
S
co
I
X
X
D
Kurabo
X
X
ry H
1
X
T
0, r.
Mltlubllhl HI
X
X
BOOV
Httlubllhl KK
X
ll p
Hltiubliht
Patrochemlcal
roca
Hit §ul Engrg. &
Shlpbldg.
X
iiai
Hltaul Toatsu
X
X
i
X
i.
X
X
I
I
X
•4
K
H
X
I
fe
X
X
X
X
X
X
L
If
T
X
H
X
¥
K
I
__.
X
1
.6
•
e
X
X
X
1
t
4J
I
X
1
•e
i
UOP (KOx-only)
a. Based on the latest information, other advantages and disadvantages may bee
b. .An "X" denotes that the process has this characteristic.
c. Forms a powder composed of complex N-N-O-S compounds.
:ome apparent as more data are released.
-------
After a detailed technical review of the currently available flue gas
denitrification processes had been completed, eight processes were recommended
for further study based on the following criteria:
* Technical considerations
* Development status
• Representative sample
The reasons for basing the selection on the technical feasibility of the
processes are obvious in that to be considered the process must be able to
remove NOX from flue gas and also be potentially applicable to coal-fired
boilers. Other technical considerations involved in the selection process
included NOX removal efficiency, generation of waste streams, background
of process developers, and system complexity. Also, to rate the process
highly, some type of modification must have been added which would minimize
the effects of the increased particulates, NQX, and Cl~ associated with
coal combustion.
The second major selection criterion was the development status of the
process. The ideal status of development would be a process which had been
extensively tested in a commercial-scale unit (>50 MW) treating flue gas
from a coal-fired boiler. Since only one or two of the processes have been
extensively tested on coal-fired flue gas, the developmental status criterion
was reduced to simply the size of the test unit. For this study, the fol-
lowing classification system for the test units (and for comparison purposes,
the number of examples at each stage at the present time) was specified.
Size Number of processes
Commercial MW 550 5
Prototype 5< MW <50 10
; Pilot plant Q.5< MW <5 14
Bench scale MW £0.5 8
Conceptual - 4
Except for the Exxon Thermal process, which is a selective noncatalytic
reduction process, all of the commercial units currently operating are
based on the dry SCR technology. The dry SCR also represents four of the
prototype units with four others being various types of wet processes.
The pilot-plant units are primarily dry SCR processes with a few wet oxidation-
absorption-reduction and a few wet absorption-reduction processes, whereas
the bench-scale units include most of the wet absorption-reduction and a
variety of dry processes.
Since very few of the processes have been tested on coal-fired flue
gas and only one class of processes has reached the commercial stage of
development, the third and possibly the most important selection criterion,
a representative sample, was added. If only the previously mentioned criteria
were used, the preliminary economic analysis in Phase II would Include only
two broad classes of NOX removal processes, dry SCR and wet oxidation-
absorption-reduction. Thus, to provide a good comparison of the various
xxvii
-------
available and potentially available technologies, a third selection criterion,
the desire to obtain a representative sample of the alternative processes,
was included. For this reason the Asahi Chemical process, the Ishikawajima-
Haritna Heavy Industries process, and the MON Alkali Permanganate process
were chosen to represent the absorption-reduction, oxidation-absorption-
reduction, and the absorption-oxidation processes respectively. The other
five processes were selected based on both technical considerations and
development status. The eight processes chosen for preliminary economic
analysis are listed in Table S-6.
TABLE S-6. PROCESSES RECOMMENDED FOR FURTHER STUDY IN PHASE II
Process
Type of process (classification)
UOP Shell Copper Oxide
UOP Shell Copper Oxide
Hitachi Zosen
Kurabo Knorca
Moretana Calcium
Ishikawajima-Harlma Heavy Industries
Asahl Chemical
MON Alkali Permanganate
Dry Simultaneous S02~NOX
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Wet simultaneous
(Oxidation-absorption-reduction)
Wet simultaneous S02~NOX
(Oxidation—absorptlon-t-reduct ion)
Wet simultaneous SO£-NOX
(Absorption-reduction)
Wet NOX only
(Absorption-oxidation)
At first glance the recommendation of four dry SCR processes for
further study would seem to contradict the desire for a representative
sample; however 22 of the 42 processes included in this study are based on
the dry SCR technology. In addition, each of these four processes involve
variations in this SCR technology. For example, the UOP process can be
operated for either simultaneous S02~NOX removal or N0x-only removal and
uses a unique parallel passage reactor to minimize partlculate problems.
The Hitachi Shipbuilding and Engineering process (Hitachi Zosen), on the
other hand, uses a honeycomb-shaped catalyst, whereas the Kurabo Knorca
xxviil
A
-------
system utilizes a moving-bed reactor to minimize the effects of dust. Since
plugging by dust is expected to be a major problem with the dry SCR processes,
analyzing these three systems will allow a comparison of the various approaches.
The recommendation of two wet oxidation-absorption-reduction processes
would also appear to be contrary to the desire for a representative sample.
Upon closer examination, however, the Ishikawajima-Harima Heavy Industries
process is based on using 63 as the gas—phase oxidant while the Fuji Kasui-
Sumitotao Metals' Moretana Calcium process uses C102. These two oxidants,
although serving the same purpose, involve different economics and different
technical problems in their use.
The selection of these processes should, at the end of Phase II, allow
comparison of various types of wet simultaneous S02~NOX processes, dry
versus wet simultaneous SC>2-NOX processes, dry versus wet N0x-only processes,
and simultaneous S02~NOX processes versus N0x-only processes combined with
an FGD system.
In some cases there are several other processes similar to those chosen
for further study in the second phase. Their elimination should not imply
that these processes are inferior, but rather that these processes represent
similar technologies and as such do not possess significant differences from
the processes selected.
"xxtx
-------
A
-------
INTRODUCTION
Man-made nitrogen oxides (NOX) are classified as two types, depending
upon their source—stationary or mobile. Each source is responsible for
approximately half of the total NOX emissions. These stationary and mobile
sources can be further divided into the various groups listed in Table 1.
The estimated total amounts and the percentages of the total amount of man-
made NOX released from each source for the years 1970, 1972, and 1974 are
also shown. During this period the only two groups with increasing NOX
emissions were the mobile sources, particularly automobiles and trucks, and
the fuel combustion (boiler) group of stationary sources,
The NO., emission limits for the mobile sources were formulated in 1971
A
by the U.S. Environmental Protection Agency (EPA) and originally required
compliance by the 1976 model year. However, due to technical, economic, and
other considerations, these standards for automobiles have been delayed once
and are expected to be postponed again. Thus the brunt of decreasing the
man-made NOX emissions, at least for the near future, will be directed toward
the stationary sources. Since the fuel combustion group represents 90-95% of
all the emissions from stationary sources, this group appears the most likely
to undergo more stringent regulation.
The fuel combustion group can be split into other subgroups according
to the type of fuel used and the type of operation being performed. As one
would expect, the six fuel combustion sources emitting the largest quantities
of NOX are utility and industrial boilers burning coal, oil, or gas. These
six sources and the estimated amounts of NOX emitted for each are listed in
Table 2. The largest source of NOX in the fuel combustion group is coal-
fired utility boilers followed by oil-fired industrial boilers and utility
boilers; approximately 30% of all stationary source NOX is emitted by coal-
fired utility boilers.
Although NOX is present in relatively minor amounts in the flue gas
(<0,1% by volume of the total flue gas), a single 500-MW coal-fired power
plant releases approximately 10,800 tons/yr of NOX to the atmosphere
(assuming 600 ppm of NOX in the flue gas). This NOX in the flue gas is
generated during combustion from either the nitrogen-containing (N) compounds
in the fuel (fuel NOX) or the reaction of molecular nitrogen (N2) and oxygen
(©2) at high temperatures (thermal NOX). The portion of the total NOX
emissions from each of these sources is a function of many variables and is
relatively site-specific.
-------
TABLE 1.
NOX EMISSIONS IN THE UNITED STATES (39)
Annual NOX emissions
1970
Source
Mobile
Highway
Nonhighway
Subtotal
Stationary
Fuel combustion
Industrial processes
(noncombustion)
Solid waste
Miscellaneous
Subtotal
Total NOX emissions
Amount,
Mtonsa
6.9
2.4
9.3
10.1
0.6
0.3
0.1
11.1
20.4
% of total
NOX
33.8
11.8
45.6
49.5
2.9
1.5
0.5
54.4
100.0
1972
Amount,
Mtonsa
7.9
2.6
10.5
10.8
0.6
0.2
0.1
11.7
22.2
% of total
NO
35.6
11.7
47.3
48.6
2.7
0.9
0.5
52.7
100.0
1974
Amount,
Mtonsa
8.1
2.6
10.7
11.0
0.6
0.1
0.1
11.8
22.5
% of total
NOX
36.0
11.6
47.6
48.9
2.7
0.4
0.4
52.4
100.0
a. M = one million; tons are metric tons.
-------
With the passage of ths Clean Air Act of 1970, EPA formulated and
released New Source Performance Standards (HSPS) _fpr NO* emissions in
December 1971. The standards for large boilers /i.e., >250 MBtu/hr (M - one
million) (28 MW equiv}7> are shown in Table 3. These standards recognize
both the differing amounts of bound H in the fuels (i.e., ranging from 0,2
Ib NOx/MBtu for natural gas with no fuel N to 0.7 Ib NOx/MBtu for coal which
contains about IX fuel N), the size of the utility boiler, and also the best
available t&chnology for NO^ control.
TABLE 2. NOX EMISSIONS PROM SELECTED STATIONARY SOURCES
IN THE U.S. IN 1972 (95)
emissions __
Annual amount, % of stationary
Stationary source metric ton sources
Utility boilers
Coal fired
Oil fired
Gas fired
Industrial boilers
Coal fired
Oil fired
GAB fired
3,495,000
1,114,000
835,000
735,000
1,245,000
491,000
30,7
9.8
7.3
6.5
10.9
4.3
TABLE 3. NO* EMISSION STANDARDS AND PROJECTED RESEARCH OBJECTIVES
FOR LARGE FOSSIL FUEL-FIRED BOILERS
Present EPA standardL (40) Projected resaarcn objtctives (95)
Lb tfOx/MBtu '. 1980 . 1985 .
input j:o boilsr SQx, ppn H0*j ppa ^0^,_ ppat
Gaseous fuel
Liquid fuel
Solid fuel
0.2
0.3
0.7
150
22.1.
550
too
ISO
zoo
50
90
100
a. Expressed at H02*
b. Calculated *t 32 excess 02, dry basis,
-------
The NSPS may become more stringent since the amount of NQX released to
the atmosphere by utility boilers Is increasing. Also, the further develop-
ment of NOX control technology increases the probability of stricter NOX
emission regulations in the near future (see projected research objectives
in Table 3). This NOX technology includes both combustion modification,
which is receiving major attention in the U.S., and flue gas treatment (FGT),
which la being developed primarily In Japan.
Even at the present time, the amount of NOX released in-certain areas
of the U.S. is causing the ambient standard to be exceeded. Boiler modifica-
tions will probably receive the initial emphasis since these are considered
the least expensive methods of controlling NOX emissions. These techniques,
hovever, are limited to leaa than 502 reduction in NOX, primarily because
boiler modifications make their major impact by reducing the formation of
thermal NQX.
The percentage reductions in NOX possible with various types of boiler
modifications for the three types of fossil fuels are listed in Table 4.
These boiler modifications have their greatest impact on oil- and gas-fired
boilers, where most of the NOX is generated by the thermal mechanism, and
less of an impact on coal-fired boilers, vhere most of the NOX Is formed by
the fuel-N mechanism. Unfortunately, the modifications may also result in
unstable flame conditions, lover thermal efficiency for the toller, and
increased corrosion, inside the boiler. For reductions greato.r than
50% of the NOX emissions, which may become necessary in the future, FGT will
be required.
Since the costs of FGT processes are considered higher than for boiler
modification methods for NOX reduction, superficially ono could deduce that
a combination of these means would result in the least expensive system for
achieving NOX reductions of 70-90Z* However, if 90S NOX removal is required,
a single FGT system may be less expensive to achieve the full NOX removal as
compared with boiler modification techniques to achieve the first 502 of HO*
removal plus an FGT system for the remaining portion of NOX removal.
Although boiler modification techniques ate now receiving'primary
attention In the U.S., the FGT processes will certainly,become Increasingly
important. Therefore, the study has been undertaken to assess the develop-
ment of FGT technology currently being.considered both in the'U.S. and abroad.
The results of the first-phase activities of this NOX removal process
study are Included in this report. The initial phase has.a twofold purpose
(1) a state-of-the-art technical review of flue gas denitrificatlon processes
currently undergoing development work and (2) a screening process in which
selected FGT processes are proposed for preliminary economic evaluation ln^
Phase II of the study. The second phase will also recommend'eoma of, these.
processes for further study la Phase III, which will be a"detailed economic
analysis of the most promising FGT processes for future commercial application.
-------
TABLE 4, EFFECTS OF BOILER MODIFICATIONS TO REDUCE NQX EMISSIONS
BY FOSSIL FUEL TYPE (95)
NOy reduction! 7._
_
__ Soiler jaodlf ication _ ___ Gaa __ Oil ___ Coal
Prevention of thermal NOX
Flu« gas recirculation 60 20 -a
Reduced air preheat 50 40 -^
Steam or water Injection 60 40 -b
Prevention of both thermal and fuel NOX
Staged combustion 55 40 40
Low excess air 20 20 20
Reduced heat release rate 20 20 20
Combination of stage combustion, low excess
air, and reduced heat release rate 50 35 40
Prevention of fuel NOX
Change to fuel with lower % N -c 40 20
a. Not effective.
b. Not competitive.
c. Not applicable.
The main body of this Phase I report is composed of three sections; >i
short introduction into how NQx is formed during the combustion of fossil
fuels; a discussion of wet NOX removal; and finally an analysis of dry NOK
removal. A general overview of the various types of flue gas denitrification
systems, including a listing of all the FGT processes currently available
under the appropriate type of FGT, and a detailed technical discussion of
each of the flue gas denitrification processes is included in each of the
wet and dry NOX removal sections. The major parameters which are discussed
in the detailed technical evaluation of each process are shown in Table 5.
-------
TABLE 5, OUTLINE OF THE DETAILED TECHNICAL EVALUATION OF EACH PROCESS
Process description and principles of operation
Flow diagram
Evaluation of basic chemistry
Operating conditions and ranges
Status of development
Size of experimental facilities
Duration of testing with acceptable performances
Source of treated gas
Projected commercial availability
Background of process developer
Capability
Accessibility of process to U.S. market
Published economic data
Raw materials, energy, and operation requirements
Raw material consumption
Energy consumption Including reheat (steam, fuel, electricity)
Operating manpower
Maintenance
Technical support during routine operation
Latest reported operating conditions and estimated requirements for
typical 500-MW unit
Technical considerations
System complexity and process control capability
Sensitivity to inlet gas composition
Alternative treated gas sources
Comparative size of required equipment
Retrofit applicability
Turndown capability
Materials of construction
Environmental considerations
Sensitivity of removal efficiency to control atoichiometry and
operating conditions
Potential for removal of other pollutants
Potential problems with waste disposal
Interference of gas species and contaminants of process performance
Work hazards
Critical data gaps and poorly understood phenomena
Advantages and disadvantages
-------
GENERAL CHEMISTRY OF NQX
$2 and 02 can combine chemically to form one or more of seven oxides
of N: nitric oxide (NO), nitrogen dioxide (N02), nitrogen trioxide (NO-j),
nitrous oxide (N£0), dinltrogen trioxide (N^Qj), dinltrogen tetroxide (1*204),
and dinltrogen pentoxide ((1205)* For most combustion processes, particularly
fossil fuel combustion units, the only oxides of N (NOX) present in signifi-
cant concentrations in the flue gases are NO and NO2 with NO representing
90-95% of the total NOX from the combustion unit (38). The colorless NO in
the presence of air will, given sufficient time, oxidize to N02- N02 la a
reddish-brown, toxic gas which in the presence of certain hydrocarbons and
sunlight is a precursor for photochemical smog.
During the combustion of fossil fuels, NOX is formed by two mechanisms:
the thermal fixation of atmospheric N2 and the conversion of fuel-bound N.
In both cases the end result is primarily the formation of NO because the
residence time in most combustion units is too short for a significant
amount of oxidation of NO to N©2 to occur (25, 99). The reaction mechanisms
are complex and are described in more detail in other sources (24,, 188).
Simplified forms of the overall reactions are shown below.
Thermal fixation:
Fuel-bound N?
N2(g) * °2(g) * *eat + 2NO(g)
2°2(g) * heat + 2N°(g)
The combination of molecular N£ and 0£ by thermal fixation Is an equi-
librium reaction with the final concentration of NO primarily dependent on
the reaction temperature, but it also varies with the amount of free 02 and
molecular K2 available In the boiler. The conversion of fuel-bound N, on
the other hand, IB relatively independent of combustion temperature and forms
rapidly even at moderate temperatures. The conversion la primarily dependent
on the availability of free 62 above the combustion zone. During the com-
bustion of fossil fuels containing chemically bound N (oil and coal) , portions
of the resulting NO* will ba formed by both mechanisms. The exact amounts
arising from fuel-bound H and from the thermal fixation of N£ cannot be cal-
culated from theoretical considerations but vary with specific boiler operat-
ing conditions.
-------
tn the flow of the flue gaa through the boiler, the NOX is hypothesized
to be formed in the following manner. The NOX formed from fuel-bound N la
generated slightly above the actual combustion zone as the gases are beginning
to cool and free D£ is more readily available. At this stage the fuel N is
converted to NO with essentially no N02 -formed. As the flue gases pass
through the boiler the rate of conversion of fuel N rapidly declines to
essentially negligible levels. The amount of thermal NOX formed, on Che
other hand, increases aa the gases exit the combustion zone where both the
temperature and the 02 levels are still high. After leaving this zone v?here
NO reaches its maximum value, the flue gases begin to cool very slowly and
NO undergoes decomposition by the following equilibrium reaction.
Unfortunately thia equilibrium reaction has a high activation energy and
once the flue gases reach the section of the boiler where they begin to
undergo rapid cooling, the rate of this reaction drops to negligible levels.
No further decomposition of NO occurs and the equilibrium KO concentration
occuro. The equilibrium NO levels are "frozen" at the temperature where this
rapid cooling began.
However, as the NO passes through the cooler portions of the boiler, the
NO is gradually, but continuously, converted to N02 by the following reaction,
Although thannodynamically favored, the rate o€ formation of N02 in the
boiler is very slow end the time available is too short for this reaction
to reach equilibrium. The actual amount of N02 formed in any boiler ia site-
specific but typically the exiting flue gas contains less than 101 of the
total NOX as N02 (99).
From the previous discus a ion of the mechanism of NOX formation then,
the total amount of NOX formed In the specific boiler is primarily a function
of the flame temperature and tha free 02 concentration. The higher the flame
temperature and/or the higher the free Qj concentration, the higher the
resulting equilibrium concentration of 80 in the flue gas. Thus any boiler
or operating modifications which minimize either of these operating parameters
will decrease the total NO* emitted from the boiler. Operating conditions
which Increase flame temperature aad/or 02 concentration in the flue gas,
and thereby Increase the concentration of KOX in the flue gas, are . i fc 'lows:
* Preheating fuel or air (increases temperature)
* Increasing excess air ( ice r eases 02 content)
• Increasing heat rate (increases temperature)
* Increasing load (increases temperature)
Since preheating the fuel or the air, increasing the heat rate, or
increasing the boiler load all have the effect of Increasing the flame
temperature in the boiler, each will contribute to increasing tha equilibrium
concentration of NOX. figure* 1 and 2 graphically illustrate the effects of
8
-------
preheating the air and increasing the load. Modifying these operating
variables to decrease the NOX emissions from the boiler are not desirable
alternatives since not preheating the fuel or air will decrease the thermal
efficiency of the boiler and decreasing either the heat rate or the boiler
load wastes a portion of the boiler capacity.
Various methods have been developed to reduce the free D£ in the boiler.
Decreasing the amount of excess air to the boiler can have a large impact on
NOX emissions, particularly for coal-fired boilers, as can be seen in Figure
3. Redesigning the boilers to prevent the excess air from reaching the
hotter .areas of the boiler have also received considerable attention as
methods to reduce the formation of NOX. Boiler modifications of this type
currently undergoing development include staged combustion, flue gas recir-
culation, and rearrangement of the burners. The arrangement of the burners
and the type of firing can have a significant impact, as shown in Table 6,
on the concentration of NOX in the flue gas.
TABLE 6. CONCENTRATION RANGES OF NOX FROM COAL-FIRED POWER PLANTS (38)
Typical NOX
Type of firing concent rat ion, ppm
Vertical 225-310
Horizontally opposed 340-375
Spreader (stoker) 400-470
Tangential (corner) 420-500
Front wall 390-600
Cyclone 800-1200
One other method of reducing NOX emissions would be to switch to a
boiler fuel containing less fuel-bound N, As would be expected, natural
gas firing results in the lowest NOX concentrations since the only NOX formed
is thermal NOX. Oil-fired boilers generate slightly higher exit NOX levels
since oil contains some fuel-bound N. Coal-fired boilers produce the highest
NOX levels due to the relatively high N content of most coals. Although
only a small portion (<30%) of the N in the coal is converted to NOX, the
proportion of the total NOX emitted which comes from fuel-bound N can be as
much as 80% in some coal-fired boilers (24).
As the concentration of bound N in the fuel increases, the actual
conversion of this N to NOX decreases but the total NOX emitted from
the boiler increases. Figure 4 (5) shows this relationship between the NOX
concentration in the flue gas and the percent N in the fuel (for oil-fired
boilers). Similar relationships would probably exist for the conversion of
fuel N in coal. That is, the conversion of fuel N to NOX would decrease
with increasing fuel N concentration in the coal. The conversion of fuel
N is also dependent on the free-02 concentration, as shown in Figure 5 (47),
Increasing the 02 concentration in the boiler Increases the conversion of
fuel-bound N,
-------
1250
1000
a.
Q.
X
o
750
500
COAL
250
OIL
I
I
300
400 500 600
AIR PREHEAT TEMPERATURE, °F
700
Figure 1. Effect of Air Preheat Temperature on NOx Emissions (18).
-------
1000
750
a.
a.
500
250
I
I
I
20
30 40 50
HEAT RELEASE RATE . KBTU/FT'/HR
60
Figure 2. Effect of Heat Release Rate on NO* Emissions (18),
-------
1250
IOOO
K>
a 750
a.
50O
25O
COAL
I
10
2O
EXCESS AIR, %
30
40
Figure 3. Effect of Excess Air on NO, Emissions (18).
-------
400
350
300
a
a.
x
o
250
200
150
100
50
125 MW
POWER
PLANT
(OIL FIRED)
T
0,1 0.2
0.3
0.4
FUEL NITROGEN, %
Figure 4. Effect of Fuel Nitrogen on NOX Emissions (5).
13
-------
40
30
8
20
O
a:
to
0
1 I I I
10 20
30 40 50
EXCESS AIR, %
60 70
Figure 5. Effect of Excess Air on Fuel Nitrogen Conversion (47).
14
A
-------
The flue gas from fossil-fueled boilers contains several other pollutants
in addition to NOX. Although most of the processes under consideration were
tested on either gas- or oil-fired flue gas, several of these pollutants,
particularly hydrogen chloride (HC1), sulfur oxides (SOX), and particulates,
will have significant impacts on the operation of these NOX processes when
they are adapted for use on coal-fired boilers. A comparison of typical
flue gas compositions from both oil- and coal-fired boilers is presented in
Table 7. Concentrations of these pollutants are shown to be larger for coal
firing as compared to oil firing..
TABLE 7. TYPICAL FLU! GAS COMPOSITIONS F10M OIL- AND COAL-FIRED 500-MW BOILERS
Fossil fuel type
Component , vol %
N2
o2
C02
so2
303
NOX
HC1
HoO
Particulates (dry basis) , gr/sf t^
Oil*
73.60
2.54
11.96
0.13
0.0013
0.02
—
11.75
(60PF) 0.36
Coal°
73.76
4.83
12.31
0.24
0.0024
0.06
0.01
8.79
6.65
a. No. 6 fuel oil: 144,000 Btu/gal, 2.5% S
b. Coal: 10,500 Btu/lb, 3.5% S, 16% ash, 0
to SOX in flue gas.
, 0.1% ash.
.15% Cl; 95% S
in coal ie converted
Of these other pollutants, the increase in HC1 and particulates are
probably the most important factors to consider in switching these NOX
removal processes from oil to coal. The dry processes, since in general
they use catalytic reactors, are particularly sensitive to dust and plugging
and the catalyst base material may be subject to chemical attack. The flue
gas from a typical coal-fired boiler (10,500 Btu/lb, 16% ash, 500 MW) con-
tains about 20 times as much dust as an oil-fired boiler of the same size.
Hence, operating problems and equipment modifications in the reductor section
can be expected due to the increased particulate levels. These particulates
from a coal-fired boiler have the typical chemical composition given in
Table 8 and a particle size range of up to 21,5 y.
15
-------
TABLE 8. CHEMICAL COMPOSITION OF FLYASH FROM A COAL-FIRED BOILER (37)
Range of concentration,
Component % by wt
Silica (Si02> 38-58
Alumina (A1203) 20^40
Iron oxide (Fe203> 6-16
Lime (CaO) 2-10
Magnesia (MgO) 1-3.5
Potash (K20) 2-5.5
Sulfuric anhydride (S03) 0.5-2.5
16
A
-------
WET NO REMOVAL PROCESSES
INTRODUCTION TO THE WET FLUE GAS DENITRIFICATION P10CESSES
The wet flue gas denitriflcation processes were, In most cases, designed
to take advantage of technology already available from the previously developed
flue gas desulfurizatlon (FGD) systems. Except for the absorption-oxidation
processes (defined below) , the wet processes were all originally designed as
simultaneous S02-NOX removal systems. The absorption-oxidation processes
were originally developed for the treatment of tall gases from nitric acid
(HN03) plants and probably, due to the process chemistry, can economically
be used for NOX removal only.
The NOX formed during the combustion of coal is present in the flue gas
mainly in the form of NO (90-95% of the total NOX) with the remainder being
essentially N02- Unfortunately NO is relatively insoluble in aqueous solu-
tion and thus presents a formidable obstacle for these wet NOX removal pro-
cesses, N02> on the other hand, although not nearly as soluble in aqueous
solution as S02» is much more soluble than NO (see Relative Solubilities of
Various Gases in Appendix B) . The main problem associated with any wet NOX
removal process is to force the NOX, particularly the 10, which is present in
relatively low concentrations in the flue gas (600 ppm) into the scrubbing
solution where the NOX can be concentrated and more economically converted
into other forms.
The two common methods of removing the NOX from the inlet flue gas are:
dissolving the NOX in the absorbing solution without pretreating them or
using a gas-phase oxidant to convert the relatively insoluble NO, either
partially or entirely, to N02- Thus, the wet flue gas denitrificatlon pro-
cesses can be divided into two major classes, absorption or oxidation,
depending on whether or not the inlet flue gas is treated with a gas-phase
oxidant ,
Each of these major classes are further subdivided into two subsections
based on the further treatment of the absorbed HOX in the scrubbing solution.
One subsection in each of the two major classes includes those processes
which reduce the absorbed NOX, either partially or entirely, to molecular N2
or ammonia (NH3)-type compounds. The other subsection Includes all the other pro-
cesses in the class which do not reduce the absorbed NOX. Thus each of the
wet NOX processes can be listed under one of the following groups; oxidatlon-
abaorptlon-reduction, oxidation-absorption, absorption-reduction, or absorption-
oxidation. Figure 6 graphically shows this classification system. For example,
the, Moretana Calcium process first oxidizes the NOX with chlorine dioxide
(C102> (oxidation step) and once the NOX is absorbed into the scrubbing
solution (absorption step) , it is reduced to molecular N2 and
17
-------
INLET
FLUE GAS
00
CONTAINING
I NO, 1
YES
ARE
NITROGEN
COMPOUNDS
REDUCED
GAS
OXIDATION
ARE
NITROGEN
COMPOUNDS
REDUCED
OXIDATION -ABSORPTION-REDUCTION
OXIDATION- ABSORPTION
ABSORPTION - REDUCTION
ABSORPTION-OXIDATION
Figure 6. Classification System for Wet NO Removal Processes.
X
-------
(reduction step), Thus the Moretatia Calcium process would be listed
as a wet, oxidation-absorption-reduction process.
Each of these four types of wet flue gas denitriflcation processes is
explained in greater detail in this section under separate headings. Each of
the flue gas denitrification processes Included in this study is listed under
the specific subsection in which they belong. A detailed discussion of each
of these wet processes is given later in this section. A general comparison
of various operating parameters for each type of wet process is shown in
Table 9.
Absorption-OxidationProcesses
The absorption-oxidation processes were originally developed to treat
the tall gas from HN03 plants and later were adapted for FGT. These processes
absorb the relatively insoluble NO into an aqueous salt solution containing
a liquid-phase oxidizing agent. This oxidizing agent rapidly converts the
absorbed NOX into nitrate (NC>3~) salts which then must be removed from a
wastewater stream.
These absorption-oxidation processes have two fundamental problems
inherent to their basic process chemistry. The most critical problem is that
the liquid-phase oxidizing agent not only oxidizes the absorbed NO but also
readily oxidizes any absorbed S02- Thus for flue gas denitrification appli-
cations, the S02 must be removed in a separate absorber prior to entering the
denitrlfication section to prevent the excess consumption of the oxidizing
agent. These absorption-oxidation processes are not simultaneous S02~NOX
removal processes but are simply the second stage of a two-stage processing
scheme*--one stage for SO? removal and a second for denitrification.
The second fundamental problem is that these processes rely on absorbing
relatively insoluble NO into an aqueous salt solution. Although the absorption-
reduction processes, in which the reactions are catalyzed, have a similar
problem, the absorption of the NO is even more difficult in the absorption-
oxidation processes which do not use catalysts since the catalyst activity
would be destroyed by the liquid-phase oxldant. For this reason the absorber
for removing the NOX will be large with a high liquid to gas (L/G) ratio.
The three processes included in this study which are examples of this
type of denitrlfication scheme are the Mon Alkali Permanganate process, the
Nissan Permanganate process, and the original Kobe Steel wet process. The
development of the latter, the Kobe Steel process using calcium hypochlorite
£Ca(C10)2_/ has recently been terminated (43). One other absorption-oxidation
process, the Hodogaya process, has been reported (5) to be undergoing develop-
ment but no information has been made available.
A list of the advantages and disadvantages common to all of the absorption-
oxidation processes is given in Table 10.
19
-------
TABLE 9. COMPARISON OF WET NOX REMOVAL PROCESSES
NJ
o
Process characteristics3
Simultaneous S02~NOX removal
Achieves high 862 removal (>95%)
Achieves moderate NOX removal
(>85%)
Operating conditions
Requires absorption catalyst
Requires liquid-phase oxidant
Requires gas-phase oxidant
Requires large absorber
Requires flue gas reheat
Forms nitrate salts in wastewater
Requires specific range of flue
gas constituents
Current development status
Tested on coal-fired flue gas
Tested on pilot plant or larger
scale
Wet
Absorption-
oxidation
_
-
X
-
X
-
X
X
X
X
-
—
NO-jr removal
Oxidation-
absorption
Sb
X
X
-
-
X
X
X
X
-
•L
xb
X
process type
Oxidation-
absorption-
reduction
X
X
X
Xb
-
X
-
X
xb
xb
-
X
Absorption
reduction
X
X
—
X
-
-
X
X
-
X
Xb
Xb
a. An "X" indicates the process has
b. Depends on process.
this characteristic.
-------
TABLE 10, ADVANTAGES AND DISADVANTAGES OF ABSORPTION-OXIDATION PROCESSES
_ Advantage _ Disadvantage _
Achieves high NOX removal efficiency Requires two separate absorbers — one
for desulfurization and one for deni-
trification
Requires liquid-phase oxidizing agent
Requires exceedingly large denitrifi-
cation absorber and high absorber L/G
Converts absorbed NOjj to wastewater
N03~ salts
Requires flue gas reheat
Oxidation-Absorption Processes
The oxidation-absorption processes are a catch-all group which includes
all of the other oxidation processes which do not meet the criteria for
inclusion in the oxidation-absorption-reduction section. Thus an overview
of oxidation-absorption processes with generalized advantages and disadvan-
tages is rather difficult. There are, however, two types of oxidation-
absorption processes at the present time: (1) equimolar NO-N02 absorption
and (2) N02 or ®2®5 absorption. Although both of these types of processes
result in a byproduct N0j~ solution, the overall processes are substantially
different.
The equiraolar absorption processes such as those developed by Kawasaki
Heavy Industries and Ube Industries are based on the absorption of N£03
which is formed by the gas-phase reaction of NO and N02« Unfortunately the
NOX in power plant stack gas is mostly NO (>90%) and a N02~rich recycle
stream must be added to the flue gas to adjust the molar ratio of N02 and
NO to one. The source of this N02~rich recycle stream for the Ube Industries
wet process has not been Identified, but in the Kawasaki process the nitrite
(N02~) salts formed during the absorption of ^Oj are chemically decomposed
into an NO-rieh gas stream and N03~ salts. This NO-rich stream is then air
oxidized to N02 and reinjected into the flue gas prior to the absorber.
Thus for every 2 mols of NOX absorbed, i.e., 1 mol of NO and 1 mol of N02,
1 mol of N03~ salt is formed and 1 mol of NO is regenerated, oxidized to N02,
and reinjected into the flue gas for reabsorptlon.
This equimolar absorption mechanism is complicated by the fact that
when the NOX concentration in the flue gas falls below approximately 200 ppm,
the rate of formation and hence absorption of ^03 rapidly decreases to in-
significant levels. For 80-90% NOX removal the flue gas must be further
treated in another stage with a gas-phase oxldant to convert all the remaining
21
-------
NOX to N0£ prior to absorption. Although the use of an NC^-rich recycle
stream, at least for coal-fired boilers, substantially reduces the consumption
of expensive ozone (0^), the generation of this NC^-rich stream further
complicates an already complex process.
The oxidation-absorption processes based on NC>2 or ^05 absorption are
apparently earlier versions of the present day oxidation-absorption-reduction
systems. The only example of this class of oxidation-absorption process
included in this review is the Tokyo Electric-Mitsubishi Heavy Industries
(MHI) process, The NOX in the flue gas is converted to $205 using an excess
of 63. The 03 requirement, for the same inlet NOX concentration, is 50%
higher for the Tokyo Electric-MHI process than for the other oxidation pro-
cesses using 03. The resulting ^05 forms HN03 when absorbed into aqueous
solution and is removed and concentrated to form a 60% HNOj solution as by-
product. This desire to form only byproduct HNO
-------
much more soluble In aqueous solution than NO. Before entering the absorber
the flue gas is injected with a gas-phase oxidizing agent, either 03 or C102»
to convert the NO to N02-
This gas-phase oxidant rapidly and selectively oxidizes the NO and the
resulting N0£ is absorbed into the aqueous sulfite (S0-j~) solution. The S02
in the flue gas is simultaneously absorbed and forms either the 803™ or bi-
sulfite (HSOg") ion in the scrubbing solution. These S0^= ions are partially
oxidized to sulfate (SO^") during the reduction of the absorbed NOx and the
remaining 803" is oxidized by air to S0^= in the regeneration section and
removed as gypsum (CaSO^*2H20) . The absorbed NOX is at least partially
reduced and removed as a combination of N03~ salts, molecular $2, and NH3~
based waste stream.
The degree of reduction of the NOX depends on whether a catalyst is
present in the scrubbing solution to aid in the absorption of the NOX. The
Chiyoda Thoroughbred 102 process, which uses a weak sulfuric acid
solution, and the Moretana Sodium process do not use a catalyst for
absorption and approximately half of the absorbed NOX is reduced to molecular
N2 and the other half is converted to N03~ salts. The remaining oxidation-
absorption-reduction processes are based on limestone slurry scrubbing with
a proprietary catalyst added to the scrubbing solution to enhance NOX absorp-
tion. The addition of this catalyst apparently results ln_most of_the NOX
being reduced to complex N-S compounds such as sulfamine _/NH(S03)2~_/ salts.
Only relatively minor amounts of molecular N£ .and 1103" salts are formed in
these processes. These complex N-S compounds are then decomposed into various
3~ based byproducts depending on the process involved.
The use of a gas-phase oxidant decreases the size of the absorber and
lowers the absorber L/G ratio since N02 is relatively soluble in aqueous
solution. However, its use is also the major drawback for these processes
since both On and C102 are expensive and must be generated onslte. The use
of C102 also leads to high levels of chlorides (Cl~) in the circulating
solution and the resulting production of Cl salts in the wastewater. A
brief description of the technical and economic considerations Involved in
generating Oo or C102 is included in Appendix A.
The examples of oxidation-absorption-reduction processes considered in
this study include the Chiyoda Thoroughbred 102, the Moretana Sodium, the
Moretana Calcium, the Ishikawajima-Harima Heavy Industries, and the MHI.
Additional oxidation-absorption-reduction processes have been reported by
others (5) , but are not detailed in this study due to lack of available
information. These processes include those undergoing development by Osaka
Soda and Shirogane in Japan.
A list of the general advantages and disadvantages for oxidation-
absorption-reduction processes is given in Table 12. Table 13 gives a
simplified comparison of each of these oxidation-absorption—reduction
processes.
23
-------
TABLE 12. ADVANTAGES AND DISADVANTAGES OF
OXIDATION-ABSOHFTION-REDUCTION PROCESSES
Advantage Disadvantage
Removes NOX and SC^ simultaneously Requires an expensive gas-phase
oxidant
Produces a potentially marketable
byproduct (CaSO^-lH^jO) Forms an N-containing waste stream
Achieves high NOX removal rates (85-90%) lequires flue gas reheat
Achieves high S02 removal rates (80-90%)
Absorption-Reduction Processes
The absorption-reduction processes were apparently specifically developed
for the simultaneous removal of S02 and NOX from power plant flue gas without
using an expensive gas-phase oxidant. These processes are based on certain
ferrous (Fe ) chelating compounds which "catalyze" the absorption of the
relatively insoluble NO. Thus, the need for an expensive gas-phase oxidant
is eliminated. The NO is absorbed into the scrubbing solution and forms a
complex with the chelated compound. The S02 is simultaneously absorbed into
the scrubbing solution as the SO^* ion and reacts with the NO complex. The
NO is reduced to molecular N2» the Fe chelating compound is regenerated,
and the SOj™ ion is oxidized to SO^" in a single reaction. This SO^" ion is
generally removed as CaS04*2H20, but in one specific case the 50^" is removed
as ammonium sulfate
The absorption of NO requires a large absorber and a high L/G ratio in
the absorber since even with the use of the catalyst, NO is relatively
insoluble in the scrubbing solution. Due to the basic chemistry of these
absorption-reduction processes, they are sensitive to the inlet flue gas com-
position, particularly the levels of 0£, S02, and NOX. The molar ratio of
S02 to NOX in the flue gas must remain above approximately 2.5 for good deni-
trification efficiency and the flue gas 02 must remain as low as possible.
As the flue gas 02 level increases and/or the mol ratio of S02 to NOX decreases,
the NOX removal efficiency will decline.
The examples of reduction processes being considered in this study
include the Asahi Chemical, the Chisso, the Mitsui wet, the Kureha wet, and
the Pittsburgh Environmental and Energy Systems.
A list of the advantages and disadvantages common to all of the
absorption-reduction processes is given in Table 14 and a simplified com-
parison of each of these absorption-reduction processes is given in Table 15.
24
A
-------
r
TABLE 13. COMPARISON OF OXIDATION-ABSORPTION-REDUCTION PROCESSES
N3
Ul
Development
status, MW
S02 removal, Z
NOx removal, %
(inlet NOx level)
Longest continuous
operation
Absorbent raw
material (catalyst)
Oxidant
(oxidantiNOx ratio)
Byproducts
Waste disposal
(other than gypsum)
Absorber L/G ratio
Reported capital
investment, b $/kW
Reported revenue
requirements, b mills/
kWh
Unusual features or
unique problems
Chlyoda
Thoroughbred 102
Bench scale
(0.3)
95
80
(150 ppm)
-
His°4 _
/Fe2(S04)3_/a
°3
(1.5)
N2, CaS04'2H20
-
-
107
6.73
Reactor-
crystallizer
Biological
denitrification
Ishikawa j ima-Harima
Heavy Industries
Pilot plant
(1.6)
90+
80
(180 ppm)
3000 hr
CaC03
(CuCl2, NaCl)
°3
(1.0)
N2, CaS04'2H20
CaS04~CuCl2-NaCl
62-75
84
6.80
Thermal
decomposer
Mitsubishi
Moretana
Heavy Industries Calcium
Pilot plant
(0.6)
95+
80-90
(150 ppm)
720 hr
CaC03
(unknown)
03
(1.0)
NH3 , CaS04 • 2H2
10-15% NH3 gas
43-62
-
-
Thermal
decomposer
Prototype
(8.3)
95+
90
(200 ppm)
-
CaC03
(CuCl2,
proprietary additive)
C102
(0.55)
0 N2, CaS04'2H20
CaCl2-Ca(N03)2(aq)
20-45
-
-
"Moretana"
absorber
Moretana
Sodium
Prototype
(10-40)
95+
90
(200 ppin)
Normal plant
availability
NaOH
-
C102
(0.55)
N2, Na2S04
Na2S04, NaCl-
NaN03
25
134
8.94
"Moretana"
absorber
Evaporation
of wastewater
a. Oxidation catalyst.
b. 1976 dollars except
for Chiyoda which
are for 1977; for source of value see
detailed process descripti
.on.
-------
TABLE 14. ADVANTAGES AND DISADVANTAGES OF THE
ABSORPTION-REDUCTION PROCESSES
Advantage Disadvantage
Removes S02 arid NOX simultaneously Requires a large absorber and a high
L/G ratio in the absorber
Produces a potentially marketable
byproduct (CaS04"2H20) Requires flue gas reheat
Achieves a high S02 removal Involves a series of complicated
efficiency (>90%) processing steps
Requires specific range of flue gas
constituents
26
A
-------
TABLE 15. COMPARISON OF ABSORPTIOH-REDUCTTOU PROCESSES
Development
status, Mtf
S02 removal, Z
•O^ removal, Z
(inlet VOx level)
Longest continuous
operation
Absorbent raw
•aaterlal (catalyst)
Byproducts
« '
Haste disposal
(other than gypsum)
Absorber' L/C ratio
gal/kaft3
Reported capital
Investment,* $/kH
Reported revenue
requirements,*
mills/kWh
Unusual features
or. unique problems
Asahl Chemical
Bench scale
(0.02)
95+
80-85
(150-250 pp.)
1000 hr
"^Ss
(Fe ^EDTA)
85 * CaSO A • 2B20
Some CaS03'l/2H20
"Trace" Ca(H03)2faa\
90-100 *•"•'
127
7.4
Extremely
large absorber
Extensive solid
handling
Chlsco
Bench scale
(0.1)
95+
70-80
(200 pp.)
335 hr
RB3
(proprietary)
(HH3)2S04(aq)
Fe(OH)2-ash sludge
_
90-100
98
6.6
Extremely
large absorber
Possibility of
sulfitic pluae
Kureha
Pilot plant
(1.6)
95+
80-85
(200 pp.)
3000 hr
Acetic acid
(proprietary)
B2, CaS04*2H20
'
125-185
65
4.8
Extremely
large absorber
Possibility of
HH3 emissions
Mitsui
Bench scale
(0.05)
95+
85
—
—
(proprietary)
"2.
concentrated S02
-
_
_
-
Extremely
large absorber
Electrolytic
reduction of
Fe+5
PEBSTS SCOBe
Pilot plant
(1.0)
90
60
(400-900 pp.)
52 hr
FeS
H2, S
FeS- ash sludge
—
10-15
125
0.49
.
Coal-fired reducing
kiln
Extensive solids
handling
a. 1977 dollars except for Chlsso which are 1976; for source of value see detailed process description.
-------
AS AMI CHEMICAL PIOCESS - WET, ABSORPTION-REDUCTION (NOX~SOX>
Process Description and Principles of Operation (107)
This process developed by the Asahi Chemical Industry Company uses an
alkali solution (Na2SO^) to absorb both S0£ and NOX. Since NOX is relatively
insoluble in aqueous solutions, an Fe * chelating compound has been added to
catalyze the absorption of NOX. The five major sections of this system include
(1) preacrubbing, (2) absorption of both S02 and NOjj, (3) reduction of NOX,
(4) decomposition of dithionate (S20o™)» an<* O) CaSC^^I^Q production. The
Asahi flue gas denitrification system, as a whole, is an intricate processing
scheme made up of many simple processing units us shown in the block flow
diagram in Figure 7.
The flue gas from the air heater passes through a prescrubber which is
expected to remove 991 of both the particulates and the HC1 resulting from
the coal combustion. In addition* some of the circulating solution evaporates
to adiabatieally cool and hunidify the flue gas from 150 C (300°F) to 53°C
(127°F). The liquid effluent from the particulate scrubber drops to a holding
tank from which moat of the liquid Is recirculated to the prescrubber after
fresh makeup HjgO has been added. A small purge stream is removed and pumped
to a flyash thickener. The thickener overflow stream Is recycled to the
effluent holding tank. The thickener bottoms containing the flyash and the
Cl~ are removed, neutralized by limestone, and pumped to a waste disposal
pond. Thus, the flyash and Cl~ in the flue gas are removed and treated before
the flue gas reaches the absorber.
After leaving the prescrubber the flue gas enters a sieve tray absorber
where it flows countercurrently to a mixed Na salt scrubbing solution con-
taining Fe+2-EDTA (ethylene diamine tetraacetic acid) at a pH of 6.3. As the
flue gas passes through the lower portion of the absorber, the S<>2 is rapidly
absorbed into the solution and undergoes the following reactions.
802(g) - S02<*q> ™
N«2S03(aq) * S02(aq) + H2°
1/2°2(aq) * Na2S2°6(aq) + H2°
The NOX on the other hand la gradually absorbed over the entire
absorber and react* according to thi following equation to form an Fe*^ chelate
complex.
N°(aq) <8>
Fe+2-EDTA-NO(aq) (9)
28
-------
H,0
PRESCRUBBER-'J "P
ABSORBER
REHEAT
CLEAN
SULFUR
DIOXIDE i
STRIPPER|
* *
NJ
VO
ASH
THICKENER
DOUBLE
1 DECOMPOSITION
[REACTOR!
LIMESTONE
SLURRY
NEUTRALIZING
{REACTORS
PURGE TO
DISPOSAL POND
SODIUM
iSULFATE
CONVERTER
REDUCTION
REACTOR
NH{S03K)2I
CENTRIFUGE I
EVAPORATOR*®-
1
NH(S03K)2
DECOMPOSE®
COOLIMG
CRYSTALLIZER
DITHIONATE
CENTRIFUGE
I
CqC03
n
4
1
1
VTE
>SE
H2
R
0
— W
•if, — -
THICK
!
i
JCAL:
SULf
CENTR
L__
> \
CALCIUM
SULFITE
:UM
Na2S04
Figure 7. Flow Diagram of Asahi C3iemical Process.
-------
in addition to these primary reactions, other side reactions are occurring
simultaneously in the absorber. Both the Fe chelating compound and the
sodium sulfite (^2803) are readily oxidized in the scrubbing solution by 62
absorbed from the flue gas to form ferric EDTA (Fe"1"3 • EDTA) and sodium sulfate
(Na2SC>4), i.e.,
°
2(aq)
Na2S°3(aq) + 1/2°2(aq) * Na2S04(aq)
Fe+2.EDTA _p.2 Fe+3.EDTA, , (12)
(aq) (aq)
The flue gas upon leaving the absorber passes through a mist eliminator,
is reheated by mixing with the flue gas from the thermal decomposers, and
sent to the stack.
The liquid effluent from the sieve tray absorber is pumped to a reducing
tank where makeup soda ash (Na2C03) is added to replace Na+ ions lost in the
system and, indirectly, to convert sodium bisulfite (NaHSO^) to Na2S03- The
reactions occurring in the reducing tank include:
Fe+2.EDTA(aq) + N^SO^, + l/2«2(g) t (13)
Na9CO,, . + 2NaHSO,, .-»• 2Na9SO,, , + H90 + C
2 3(aq) 3(aq) 2 3(aq) 2
Most of the NO absorbed is thus converted to harmless N2 gas in the
reducing tank_and vented_to the atmosphere. In addition, some sodium imido-
dlsulfonate ^NH(S03Na)2_/ is formed and the Fe chelating compound oxidized
in the absorption column by 02 from the flue gas (see eq 12) is reduced by the
503° ion in this reducing tank. Host of the effluent from the reducing tank
is recycled to the absorber. The remainder (10-20% of the circulating solu-
tion) is pumped to an evaporator in the regeneration section to concentrate
the solution and then to a cooling crystallizer. In the crystallizer sodium
dithionate (Na2S20g*2H20) and sodium sulfate (Na2S04"lOt^O) crystals are
produced under vacuum at 10°C (50°F). These Na2S2Og*2H20 and 5132804-10H20
crystals are separated from the mother liquor in a centrifuge and sent to a
dryer (operating at about 120-150°C (250-300°F) in which the hydrated crystals
are converted to anhydrous Na+ salts. For example, the desiccation of the
crystals is given by the following reaction,
Na2S206-2H20(8)- Na^O^ + 2H20(g) + (15)
After being separated in the centrifuge, the mother liquor, containing
the NH(S03Na)2 salts, is split into two streams. Most of the mother liquor
is recycled to the reduction reactor while the smaller stream passes through
30
A
-------
tjve NH(SOgNa)2 treatment section. This section is necessary since soluble
NHCSOjNa^ Is formed In small quantities by a side reaction in the absorber
and must be removed to prevent Its buildup In the scrubbing solution. The
NH(S03Na>2 is converted to Insoluble potassium imldodlsulfonate ^.N
by reaction with potassium sulfate
K2S04(aq) * OT(S°3K)2(s)* + Na2S°4(aq)
The resulting crystals of NH(S03K>2 are separated In a centrifuge and
sent to a thermal cracker while the mother liquor, containing mainly dissolved
is recycled to the reducing tank. The thermal cracker, which is
heated Indirectly by flue gas from an oil-fired furnace, operates at 500°C
(932°P) and decomposes the NH(SC>3K)2 Into SC>2» $2, and potassium sulflte
(K2S03> and K2SC>4 by the following reaction.
ZHH(S03K>2(8) i 2S02(g)t + K2S04(8) + K2S03(s) + H20(g)t + M2(g)+ (17)
The concentrated S02 gas stream is sent to the Na2SO^ converter while
the solid i^SO^ is recycled to the NH(S03Na)2 reactor.
The anhydrous Na2S04 and Na2S20g salt mixture from the dryer is sent to
a thermal cracker where the Na+ salt mixture is indirectly heated to 300 C
(572 F) and decomposed into Na2S(>4 and SC>2 according to the following reaction.
Na2S2°6(s)* Na2S04(s) + S02(g)+ (18>
Depending on the S02 concentration in the flue gas, a portion of this
^2804 is removed as a byproduct, but most is sent to an $32804 converter
where it is reacted with calcium sulflte (CaSOj) at 40°C (1Q4°F) and a pH of
2 to produce CaSC>4*2H20. The SC>2 from both the Na2S20g and the NH(S03Na)2
crackers are also recycled to the $32804 converter where the following
reactions occur.
CaS03(s) + S02(g) + H20 - Ca(HS03)(aq) (19)
2H20(a)+ (20)
The reactor product stream is pumped to a centrifuge where CaS04'2H20
is removed as a solid byproduct. The centrate is sent to an S(>2 stripper,
operating at 95°C (203°F)» pH 4, and_a pressure of 100 mm Hg, where the HSO-j"
salts are partially converted to 803" and S02.
Ca(HSO,),., v •»• CaSO,'l/2H00., ,4- + SO,, f + 1/2H00 (21)
J 2(aqJ 3 2 (s) »^~-» »
2NaHS°3(aq) * Na2S°3(aq) + S°2(g)* + H2° (22)
31
-------
rhe: S0£ produced in the stripper Is recycled back to the Na2S04 converter
and the effluent containing the S03= is sent to the double decomposition
reactor where the Na2S03 is converted to CaSO^rl/2E20 by the following reac-
tion with makeup calcium carbonate
2NaHSO_, . + CaCO., . •*• Na,,SO,, s + CaSO '1/2H.,0, ,* + 1/2H.O + C00, .t (23)
3(aq) 3(s) 2 3(aq) 3 2 (s) 2 2(g)
The effluent from the double decomposition reactor is pumped to a
thickener where the bottoms, a CaSO-j" 1/2H20 slurry, are removed and centri-
fuged. The solid CaSO^ from the centrifuge is then reslurried with makeup
H20 and recycled to the Na2SC>4 converter. The thickener overflow and the
centrate from the centrifuge, both of which are essentially Na2SC>3 solutions,
are recycled as a single stream to the absorber,
S t a tus o f De ve logmen t
The Asahi process has only been tested, as an integrated system, in a
bench-scale unit treating 40-60 Nm^/hr (24-35 sft^/mln) of flue gas from a
residual oil-fired boiler containing 150-200 ppm NOX, 1250-1500 ppm SCU, and
3-6% 02- During a 1000-hr continuous test, this processing unit was able to
demonstrate 80-85% removal of NOX and more than 991 removal of S02- The
absorption, crystallization, and decomposition portions of this system have
been confirmed and scaleup data obtained in a bench-scale unit treating 500-
600 Nra^/hr. This process has not been tested on flue gas from a coal-fired
boiler.
Since this system has only been tested on a laboratory scale, the next
major development step should be a pilot-plant unit in the 1-10 KW range
before a prototype or commercial unit is constructed. In addition to pilot-
plant testing on flue gas from an oil-fired boiler, this process should
also be tested on flue gas from a coal-fired boiler to determine its ability
to handle the increased particulate, Cl~, and trace metal loadings associated
with coal combustion.
Background of Process Developer
Asahi is a relatively large Japanese chemical company with its primary
emphasis in petrochemicals, fibers, plastics, and basic chemicals. Asahi
has had considerable previous experience in both the design and development
of air pollution control systems, including the development of an FGD system
for power plant stack gas and a denitrification system for HN03 tail gas.
Their FGD process, based on a limestone slurry scrubbing solution, is being
used in a commercial unit treating 500,000 Nnrvhr (167 MW equiv) of flue gas
from an oil-fired boiler, Their denitrification system for treating HNOj
tail gas is an NH3~based selective catalytic reduction (SCR) process similar
to the dry, N0x-only processes. A pilot plant to test this dry process on
oil-fired flue gas was recently built but no results are yet available.
At the present time no American company has been licensed by Asahi to
market this process on a commercial scale in the U.S.; however, Asahi does
have a liaison office in New York City through which inquiries about this
32
A
-------
process are handled. Asahi has recently entered into a formal agreement with
the NUS Corporation, an engineering and environmental consulting firm in
Rdckville, Maryland, to advance the Asahi process through the pilot and proto-
type stages on coal-fired flue gas in the U.S.
PublishedEconomic Data
The published economics for this process should be considered as only
tentative, and subject to revision, since the entire process system is in an
early stage of development and has not been tested as an integrated unit in
a pilot plant. Asahi estimates (107) the total capital Investment (construc-
tion basis: U.S. and mid-1977 costs) for a system using their process to
treat flue gas from a 500-MW coal-fired unit containing 2200 ppm S02 and 600
ppm NOX to be $63.6 million ($127/kW). The estimated revenue requirement for
this simultaneous SO£-NOX system on the same basis is about 7.4 mills/kWh.
RawMaterials, Energy, and Operation Requirements
The raw material requirements given below are based on treating the
flue gas from a 5QQ-MW coal-fired boiler containing 2200 ppm S02 and 600 ppm
NO,,.
Material
Quantity required
Limestone
FeS04'7H20
EDTA
127,000 tons/yr (short tons)
930,000 Ib/yr
930,000 Ib/yr
16,500 tons/yr (short tons)
The associated utility requirements are:
Utility
Quantity required
Residual oil (cracking)
Steam (150 psig max)
Cooling water
Electricity
1,500 gal/hr
110,000 Ib/hr
2,000 gpm
27,000 kW
The only byproducts produced under these assumptions would be 600 tons/
day of CaS04'2H20, 70 tons/day of Na2S04, and two small wastewater streams.
One stream, which is purged from the S206= mother liquor for Cl~ control,
contains sodium chloride (NaCl) and the other is the flyash disposal pond
overflow.
The electrical requirements for this process using a sieve tray absorber
are said to represent approximately 5% of the generating capacity of the
boiler.
For the 500-MW system the operating manpower is estimated to be 19 men/
day and the annual maintenance requirements are estimated to be $2.54 million.
33
-------
TechnicalConsiderations
The Asahi process using Na2S03 scrubbing is a relatively complex system
involving many simple chemical processing steps. For example, there are six
solid-liquid separation stepss six chemical reactors, two thermal decomposers,
one crystallizer, and one dryer in the regeneration section. In addition,
there will probably be at least four trains of sieve tray absorbers for each
500-MW boiler. Further complicating this process are numerous solid-handling
systems and also relatively high temperature requirements. Although the
process vendor envisions sufficient surge capacity to effectively decouple
the various treatment areas, the overall system would still be relatively
complex and the process control would probably be more difficult relative to
other wet and dry processes.
Assuming that the partlculate prescrubbing system is adequately designed
for high Cl~ and particulate removal, this process could probably be applied
to flue gas from a coal-fired boiler. The SC>2 removal efficiency will be
relatively Independent of the inlet gas composition. The NOX removal effi-
ciency, however, is a function of the flue gas 02 and S02 concentrations.
The S03= ion which is formed when the SC>2 is absorbed into the scrubbing
solution is used to reduce both the absorbed NO to molecular N2 and the Fe+'
ion back to the Fe4^ form. The 803™ ion is also consumed by reaction with
D£ absorbed from the flue gas. Thus for flue gas containing 5% or less 02,
approximately 2.5 mols of S02 are required in the flue gas for each mol of
NOX to obtain an NOX removal efficiency of 80-85%. For each 1% Increase in
G£ concentration in the flue gas above 5%, the NOX removal efficiency will
decrease 1-2% due to the increased oxidation of both Na2SC>3 and the Fe+2 ion.
The relative insolubility of NO in the scrubbing solution requires a
very high L/G ratio in the absorbers, on the order of 10-15 1/Nnr* (65-100
gal/kft3) of gas (127 F) depending on the NOX concentration in the flue gas.
At an NOX level of 600 ppm in the flue gas, the L/G ratio in the absorber
will be about 15 l/Nm^ (90-100 gal/kft3). The size of the absorbers is
further increased by the low superficial gas velocity through the absorber
(approximately one-third the gas velocity in a limestone desulfurization
system). Thus, instead of four relatively small absorbers per 500 MW of
capacity, as for conventional limestone desulfurization systems, four large
scrubbers will be required for this simultaneous S02~NOX process. Asahi is
now assessing the use of packed-bed absorbers to obtain better gas-liquid
contact andthus decrease the size of the absorbers.
The relatively large differences in solubility of S02 and NOX in the
absorbing solution also lead to the unique situation in which the S02 is
almost completely absorbed in the lower portion of the absorber and the
upper part of the absorber is essentially a denltrlficatlon section. Although
some recycle SOj* is available in the upper portion of the absorber, most of
the NOX is converted in the reduction reactor. Thus relatively large quan-
tities of the Fe"**2 chelating compound are required to absorb and transport
the NO until it can be reduced by the absorbed S02 in the reducing tank.
Since the L/G ratio in the absorber is very high to facilitate NOX removal,
the SC>2 removal efficiency will be relatively insensitive to changes in flue
gas composition.
34
A
-------
Fortunately the equipment in the regeneration system will be much
smaller than for the absorption section, since only 10-20% of the circulating
solution pasees through the regeneration section at any given time. In fact
the equipment sizes will be the same order of magnitude as a conventional
double-alkali desulfurlzation system since only the SO-j™ and SO^* pass
through the regenerator system. The NOX is reduced to molecular N£ and
passes out of the system before reaching the regeneration section,
Since this is a wet simultaneous process, this system can be retrofitted
relatively easily onto power plants. The major consideration would be land
areas available nearby for both siting the la.'ge numbers of equipment and
disposing and/or storing the byproducts or wastes.
This process should have relatively good turndown capabilities since
the system contains four scrubber trains, any combination of which could be
shut down completely. However, during turndown the NO removal efficiency
will decline at least somewhat since typically, when the boiler load is
reduced, the airflow is not correspondingly decreased. Thus the flue gas
will contain a higher concentration of 02 and a lower concentration of SC>2.
Most of the equipment and piping will be epoxy, elastomer, or flake
glass-reinforced, polyester-lined carbon steel to prevent the corrosion and
erosion problems associated with wetted flue gas and the circulating slurries
and liquors. Exceptions will include the thermal crackers and the S02 strip-
per, which because of high temperature and/or low pH, will be constructed of
310 stainless steel. In addition the absorber and reducing tank will be
made from 316 stainless steel.
Due to the high partlculate and Cl*" removal in the prescrubbing section,
the purging of a smaller amount of the absorber scrubbing solution to control
the Cl~, and the additional washing of the CaS04'2H20 cake, this byproduct
CaS04*2H20 is claimed to be suitable for use in the manufacture of wallboard.
Its ability to displace naturally occurring CaSO^-Zl^O in the U.S. market is
questionable and thus the CaSO^'Zi^O will probably be used as landfill mate-
rial. Similarly the Na2S04 will also probably be disposed in a landfill,
although the process developer believes it also can be marketed as a byproduct.
The operation at relatively high temperatures (500°C) required in the
thermal crackers is a disadvantage for this process. A portion of this heat
(approximately 10-15%) is retained in the solids as sensible heat and is
lost in the surge volume storage of these calcined products. The remaining
heat is used in either drying the solids prior to cracking or reheating the
clean flue^ gas from the absorber prior to exhausting the gaa through the
stack. "*
Environmental Considerations
During a continuous test using a sieve tray absorber on residual oil-
fired flue gas (150-250 ppm NOxJ 1250-1500 ppm SO-^; and 3-6% 02) in a bench-
scale unit (0.2 MW equiv), 80-85% removal of NOX and more than 99% removal
35
-------
of SOX were attained. Host of the NOX is reduced to molecular N2 with "trace"
amounts converted to NOj*" and the SOX Is converted mainly to CaSO^'lI^O with
some NaSC> formed.
If a market for the byproduct CaS04*2H20 and Na2S04 can be found, the
only waste disposal problem for this process will be the flyash-Ca salt
solution from the prescrubber.
Critical Data Gaps and Poorly Understood Phenomena
At tlie present time no critical data gaps or poorly understood phenomena
other than those being investigated by Asuhi are readily apparent.
Advantages and Disadvantagea
The major advantages and disadvantages of the Asahi sodium scrubbing
are listed below. Other advantages and /or disadvantages may become more
apparent once the process is tested on flue gas from a coal-fired boiler.
Advantages
1. Removes NOg and SC>2 simultaneously
2. Achieves >95% S02 removal efficiency
3. Produces a potentially marketable byproduct (CaS04"2H20, N82S04>
4. ,-Qperates vith full particulate loadings (>7gr/sft3)
<*
D Isadvan t age s
1. Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal- fired flue gas
3. Has been tested only in a bench-scale unit
4. Has not been operated for a long-term continuous period
5. Uses significant amounts of stainless steel or exotic materials
for process equipment
6. Requires hot solids handling
7. Has a low superficial gas velocity In the. absorber (<10 ft/sec)
8. Has a high L/G ratio in the absorber (> 70 gal/kaft3)
9. Requires flue gas constituents within specific ranges for high
NOX removal
10. Requires an expensive liquid-phase catalyst
36
-------
CHISSO ENGINEERING PROCESS - WET, ABSORPTION-REDUCTION (NO -SO )
X X
Process Description and Principles of Operation (94,97)
_ The Chisso Engineering Company's process uses an alkali solution
^ammonium sulfite— -(NH^^SO^/ to absorb both S02 and NOX. This absorption
of the relatively insoluble NOX is "catalyzed" by a proprietary Fe chelate
which is added to the scrubbing solution. The Chisso process consists of
four major sections including (1) prescrubbing, (2) absorption of both S02
and NOX, (3) reduction of NOX and oxidation of S03=, and (4) byproduct
purification. The general outline of these processing sections can be seen
in the block flow diagram for the Chisso process given in Figure 8. One major
difference between the Chisso process and the other absorption-reduction pro-
cesses is that Chisso reduces this NOX to the ammonium (NH^ ) ion instead of
molecular N2 and thus forms (^4)2804 as a potentially marketable byproduct,
The flue gas from the air heater at approximately 150°C (300°F) is
passed through a prescrubber in which most of the particulates and Cl~ are
removed. As the flue gas passes through the prescrubber it is also adia-
batlcally cooled to approximately 53 C (127 F) and humidified by the evapora-
tion of H20 from the scrubbing solution. The scrubbing solution drops to a
holding tank from which most of the solution is recirculated through the
prescrubber. A small stream is removed and pumped to a flyash centrifuge
where the flyash is separated and purged from the system. The centrate is
recycled to the holding tank with a small portion removed and purged from
the solution to prevent the buildup of Cl~ and trace elements in the pre-
scrubbing solution.
After leaving the prescrubber the humidified flue gas enters the main
section of the tray absorber and flows countercurrently to an
chelate solution at a pH of 6-^7. As the flue gas passes through the lower
portion of the absorber, the S02i which is much more soluble than the NOX,
is rapidly absorbed into the scrubbing solution and undergoes the following
reactions.
S02(g)+S02(aq) (24)
S°2(aq) + (NH4)2S°3(aq) + H2°" 2• NO, N (26)
(g) (aq)
Fe+2'(PCC) , . + NO, »-»• Fe+2'NO, N (27)
'(aq) (aq) (aq)
where (PCC) denotes the proprietary chelating agent.
37
-------
i
BOILER
t
AIR
HEATER
t
AIR
00
CLEAN
•• FLUE GAS
EVAPORATIVE
CENTRIFUGE CRYSTALUZER
I 111
(NH4)2 S04
NHjCI
Figure 8. Flow Diagram of Chisso Engineering Process.
-------
This Fe"1"2 chelate-NO complex reacts with the_ ammonium bisulfite /.
to form either imldodisulfate /NH( 803^14 )£_/ or amidosulfate
For example, the reaction for the formation of NH(S03NH4)2 is:
Fe+2.(pcc)-NO, ^ + 2NH4HS03(aq) •*• M(S03NH4)2(aq) + Fe+3- (PCC) -OH, . (28)
, ^ 43(aq) •• 342(aq) e- -, .
In addition to these primary reactions, other secondary reactions are
occurring in the absorber including the oxidation of both the 50^ ion and
the Fe"*"2 chelating compound by flue gas ©2 absorbed into the scrubbing solu-
tion. The reactions for the oxidation of the SOj" ion include;
°2(aq) (29)
(30)
2NH4HS°3(aq) + 1/2°2(aq) * (M4)2S2°6(aq) + H2°
+2
The reactions involved in the oxidation and regeneration of the Fe
chelating compound are hypothesized to be the following:
Fe+2« (PCC) (flq) + 0 •> Fe+2- (PCC) .0 (32)
Fe+2-(PCC)'02(a . + 3Fe"*"2'(PCC)(a ^ + 2H20 -"AFe"1"3* (PCC)''OH( , (33)
Fe+3-(FCC)-OH(aq) +NH4HS03(aq)-
Fe+2- (PCC) + l/2(NH)S0 + HO (34)
One other secondary reaction occurring in the absorber is that any
remaining HC1 in the flue gas will react with the scrubbing solution to
form ammonium chloride (NH^Cl) , i.e.,
HC1, , + NHQ/ , +NH.C1, v (35)
(aq) 3(aq) 4 (aq)
The clean flue gas from the top of the absorber, after passing through
a mist eliminator, is reheated and discharged through the stack.
The scrubbing solution from the absorber drops to a holding tank from
which most of the solution la recirculated through the absorber after makeup
NH3 has been added. The remainder of the scrubbing solution enters the
regeneration section where the scrubbing solution is first pumped to an
oxidizing tower. Inside this tower either air or Q£ is passed counter-
currently to the NH4+ salt solution to convert the remaining S03= and
salts to 804* and 820^* by reactions (30) and (31).
39
-------
The liquid effluent from the oxidizing tower is pumped to a catalyst
recovery section, acidified with makeup H^SO^., and sent to a cooling crystal-
lizer to allow the catalyst to crystallize. The resulting slurry is pumped
to a centrifuge to recover the solid catalyst for recycle while the mother
liquor from the centrifuge is further processed in a thermal decomposer
operating at about 120-140 C (248-284 F) . Under these reaction conditions
the (NH4)2S206 is decomposed while NH(S03NH4>2 and NH2^SO3NH4) are hydrolyzed
into (1514)2504 and NH4HSO4 by the following reactions.
(36)
2H20 - (NH4)2S0 + NHHS0 (37)
(NH4>2S206(aq) * (NH4>2S04(aq) *
The S02 gas generated in the decomposition of the (NH4)2S20g is recycled
to the flue gas ducts for reabsorptitm.
The hot liquid effluent from the decomposer is pumped to a neutralization
reactor where NHj is injected. This conversion from an acidic to_a "basic
solution results in the precipitation of iron hydroxide /Fe(OH)2_/ from the
remainder of the solution and, indirectly, the conversion of NlfyHSC^ to
2NH3(aq) + 2H2° * ^^(s)4" + 2NH4+ (39)
NH4HS°4(aq) + ^(aq) + 2S°4(aq) (40)
The resulting slurry is pumped to a centrifuge where the Fe(OH)2 is
recovered from the solution. Most of the Fe(OH)2» along with the solid
catalyst separated earlier, is redissolved and recycled to the absorber.
A small portion of the"Fe(OH)2, however, is removed from the system as a
waste sludge to prevent the buildup of the ash in the circulating solution.
The liquid from the centrifuge, containing mainly ($114)2804, is sent to an
evaporating crystallizer, where the solution is concentrated, using indirect
steam heating, until (NH^^SO^ begins to crystallize. After cooling to allow
crystal growth, the resulting crystals are removed in a centrifuge and con-
veyed to a byproduct storage section. The mother liquor, containing mainly
(^4)2804, is recycled to the acidification reactor.
Status of Development
The Chisso NH-j scrubbing process has only been tested on a bench-scale
unit of 300 Nm3/hr (0.1 MW equiv) at the Chisso Petrochemical Company's Goi
plant. However, the entire process including both the absorption and regen-
eration sections was tested as a single unit. The flue gas was from an oil-
fired boiler and contained 1600 ppm S02, 200 ppm NO^ 4% Q^, and a dust
level of 0.05-0,1 g/Nm3 £0,0215-0.0430 g/sft3 (32°F)/. The longest continuous
operation of this bench-scale unit was said to be about 2 wk (335 hr) and
reportedly achieved 95% S02 removal and 80% NOX removal.
40
A
-------
Although this wet simultaneous process has not been tested on coal-fired
flue gas, Chisso believes that with the removal of some of the precipitated
Fe(OR)2~ash sludge from the system,, the ash levels in the circulating solu-
tion can be controlled at a moderate level and thus this process could handle
the increased particulate loading associated with coal combustion.
Since this process has only been tested on a laboratory scale, the next
major development step should be a pilot-plant unit in the 1-10 MM range.
Ideally this pilot plant would be able to test the applicability of this pro-
cess for treating flue gas from a coal-fired boiler, as well as flue gas from
an oil-fired boiler.
Background of ProcessDeveloper
Chisso Engineering Company is a subsidiary of Chisso Corporation, which
is a major fertilizer producer in Japan, Prior to the development of this
process, which is based on the original NH3 scrubbing system of Catalytic,
inc., Chisso apparently had very little previous experience in FGT of either
S02 or NOX, The development of this process was begun in 1973 with the lab-
oratory testing of both S02 and NOX removal efficiencies using an (NH^^SOj
scrubbing solution containing their special catalyst. The following year
bench-scale tests were completed to teat the regeneration system and also
the effects of other flue gas species, particularly 02, on the absorption
efficiencies, In 1975 bench-scale testing was begun at the Chisso Gol instal-
lation using the entire system as a single unit.
Chisso has recently licensed Catalytic to market, engineer, and construct
plants utilizing this process in the U.S.
Published Economic Data
Since this process has only been tested on a bench-scale unit, the follow-
ing published economic data should be considered as a preliminary estimate
and subject to revision, Chisso estimates (97) the total capital investment
for a system to treat 150,000 Ntrr/hr of flue gas from a heavy oil-fired
boiler containing 1,600 ppm SO? and 200 ppm NOX as 1,464 million yen (assumed
construction basis: Japan and 1976 costs). If 300 yen/$ and 3,000 Nm3/hr/MW
are assumed, this corresponds to a total capital Investment of $4,880,000 for
a 50-MW system or approximately $98/kW of installed capacity. The estimated
revenue requirements for this wet simultaneous S02~NOX system were estimated
as 7915 yen/kl of oil (95% desulfurizatlon and 80% dehitrification). Using
the same assumptions listed above this corresponds to a revenue requirement
of about 6.6 mills/kWh for simultaneous S02 and NOX removal.
Raw^Materials, Energyf and Operation Requirements
The following raw material requirements are estimated by Chisso based
on a plant treating 150,000 Nnr/hr (50 MW equiv) of flue gas. This flue gas
from an oil-fired boiler was assumed to contain 1600 ppm SC>2, 200 ppm NOX,
41 02, and 0.05-0.1 g/Nm^ of particulates. S02 and NOX removal efficiencies
were assumed to be 95% and 807, respectively.
41
-------
Material Quanti ty r_equired
Liquid NH3 425 kg/hr
H2S04 (98%) 340 kg/hr
Catalyst makeup $25/hr
The utility requirements using the same basis were estimated as;
Utilities Quantity, required
Steam
Cooling water
Proceas water
Electric power
6,000 kg/hr
700 m3/hr
10 m3/hr
1,800 kW
This electrical requirement represents 3.6% of the total generating capacity
of the boiler.
The byproducts produced include solid (NH/^SO/ containing some
and a small amount of Fe(OH)2~flyash sludge. For the assumptions listed
above, approximately 1700 kg/hr of (NH/^SO^ would be produced.
The estimated operating manpower, including maintenance personnel for
this system, would be two men per shift plus two men per day.
Technical Considerations
The Chlsso process, although very similar to other wet, absorption
reduction processes, particularly in the absorption section, is less complex
overall. However, many pieces of process equipment are required in the
regeneration section including: three centrifuges, two crystallizers, one
thermal decomposer, one oxidizing tower, and five assorted tanks or reactors.
In addition there will be at least four large tray absorbers for each 500-MW
plant. On the other hand there are relatively few requirements for solids-
handling equipment and also the temperatures In the regeneration section are
less severe than those in some of the other absorption-reduction processes.
Thus the overall process control requirements, although much more complex
than the dry SCR processes, will probably be similar in complexity to the
other wet, absorption-reduction processes.
The process developers suggest that, since this is a wet process, the
increased particulate and Cl~ loadings associated with coal combustion
will not present significant problems in adapting their process to treat
flue gas from a coal-fired boiler. The only modifications envisioned by
Chisso to prevent the buildup of these particulates In the circulating solu-
tion is the addition of a prescrubber and associated process equipment to
remove most of the ash before the flue gas reaches the absorber and also a
small purge stream to remove some of the ash- contaminated Fe(OH)2 sludge
from the circulating solution. The Cl" will be removed primarily in the
prescrubber with a minor amount purged as an NI^Cl contaminant in the
byproduct stream.
42
A
-------
The Chlsso NH-j scrubbing process has been tested on a bench-scale unit
treating flue gas containing 1600 ppm S02, 200 ppm NOX, 4% Oo, and 0.05-0.1
g/Nrn3 of dust. During a 2-wk continuous test, this process was able to main-
tain 95% SC>2 and 80% NOX removal efficiency. The removal efficiency for S02
is relatively Insensitive to both the operating condition in the absorber and
the inlet gas conditions because the L/G ratio required in the absorber to
remove the relatively insoluble NO is much higher than that required for FGD
only. The NOX removal efficiency, however, is dependent on both the S0£ and
the D£ content of the flue gas. The reasons for this dependency are the same
as were given for the other absorption-reduction processes, that is, the S02»
once absorbed into solution as the SOj ion, is used to reduce the absorbed
NOX. When either the Q£ content in the flue gas is high or the SO^ concentra-
tion in the flue gas is low, lees of the SO^*8 ion is available to reduce the
NOX, For a flue gas stream containing 2400 ppm S02, 600 ppm NOX, and 4.5%
02, only about 80% NOX removal can be expected.
Since the overall absorption section for the Chisso process Is very
similar to that for the other absorption-reduction processes, many of the
parameters and conclusions for the Chisso absorption system will probably be
similar to those listed for the other absorption-reduction processes. The
L/G ratio In the absorber will probably range from 10-15 1/Nm3 (65-100 gal/
kaft3 at 127 F) depending on both the concentration of the NOX in the flue
gas and the removal efficiency required. For flue gas containing 600 ppm
NO-jj and L/G ratio in the upper portion of this range, probably approaching
15 1/Nm3 (100 gal/kaft3 at 127°F)S would be required. In addition the
superficial gas velocity through the absorber will probably be in the range
of 1-2 m/sec (3-7 ft/sec). These two parameters indicate that the absorbers
for the Chisso process will be much larger than conventional limestone desul-
furization scrubbers.
The large difference in the solubilities of S02 and NO leads to the
unique situation in the absorber where the S02 is rapidly stripped from the
flue gas in the lower portion of the absorber while the NO ia more gradually
removed throughout the absorber. This occurrence is one of the primary
reasons why the Fe+^ chelating compound is a major constituent of the
scrubbing solution. Since the Fe"*"^ chelating compound forms a complex with
the absorbed NO, it not only stimulates further NO absorption by the scrub-
bing solution, but also binds the NO and transports It to the lower portion
of the absorber where a higher concentration of absorbed SC>2, i.e., the
ion, is available to reduce the NO.
Another major reason for the high concentration of Fe+ chelating com-
pound in the scrubbing solution is that a substantial portion of the Fe
ion is oxidized during passage through the absorber by 02 absorbed from the
flue gas and is not available to absorb NO, (The NO apparently combines
with the Fe+2 ion in a complex form which the Fe" ion is not capable of
forming.) Thus, a relatively high concentration of the Fe+^ chelating com-
pound is required to furnish enough reaction sites for acceptable NO removal.
43
-------
Since thia is a wet removal system the major consideration in retro-
fitting this system to existing coal-fired boilers will be the land necessary
for the process equipment and for a landfill area for both the flyash and a
minor amount of Fe(OH)2-aah sludge.
The process should have relatively good turndown capabilities since the
absorption system contains four scrubber trains, any combination of which
could be shut down completely. However, during turndown the NOX removal
efficiency will decline at least somewhat since in general when the boiler
load la reduced the airflow is not correspondingly decreased. Thus, the flue
gas will contain a higher concentration of (U and a lower concentration of
The incoming flue gas ducts should be constructed of Inconel 625 to pre-
vent corrosion while most of the remaining process equipment and piping should
be elastomer-lined carbon steel to prevent corrosion problems associated with
circulating Cl~* solutions. The only major piece of equipment not elastomer
lined should be the decomposer which will be constructed of glass-lined steel.
Environmental Considerations
The Chlsso process, during a test on oil-fired flue gas (1600 ppm S02,
200 ppm NOX, and 4% 02> In a laboratory bench-scale unit (0.1 MW equiv), was
able to achieve 951 S02 removal and 80% NOX removal. The absorbed NOX is
reduced to NH3 while the S02 is absorbed as SOg" and oxidized to SQ^85. Then,
by simply combining the NH3 and the SO^*, byproduct (NH4)2S04 Is formed. As
was stated earlier, the 862 removal is insensitive to changes in the inlet
gas composition or the absorber operating conditions. The NQX removal effi-
ciency is, however, sensitive to both the 02 and the S02 content of the flue
gas. This could present a problem for flue gas from a boiler fired with low-
S coal.
Critical Data Gaps and Poorly tinders tood Phenomena
The major critical data gaps for the Chisso process include the following:
1. L/G ratio in the absorber
2, Scrubbing solution composition
3. Operating conditions and stream flows in the process equipment in the
regeneration section
4. Marketability of byproduct (NH4>2S04 in the U.S.
5. Method of preventing sulfitlc plume from the stack
The poorly understood phenomena would include the actual series of reac-
tions involved in the absorption and reduction sections. Although hypothe-
sized reactions are given by Chisso for the absorption and reduction of NOX,
others have given different reactions for these steps. Also, their method
to eliminate the formation of aulfltic plume is not specified and in previous
studies (106) this has been a significant problem.
44
A
-------
Advantages and Disadvantages
The Chisso NH3 scrubbing process has the following advantages and
disadvantages over other wet processes as well as the dry NOX processes.
Advantages
1. Removes NOX and S0£ simultaneously
2. Achieves > 951 SC>2 removal efficiency __ _
3. Produces a potentially marketable byproduct ./.(NH^
4. Operates with full particulate loadings (>7 gr/sft->)
Disadvantages
1, Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Has been tested only in a bench-scale unit
4. Has not been operated for a long-term continuous period
5, Incorporates design features which may present significant process
control problems
6. Uses significant amounts of stainless steel or exotic materials
for process equipment
7. Has a low superficial gas velocity in the absorber (<10 ft/sec)
8. Has a high L/G ratio in the absorber (>70 gal/k-aft3)
9. Requires flue gas reheat for plume buoyancy
10. Requires flue gas constituents within specific ranges for high NOX
removal
11. Requires an expensive liquid-phase catalyst
Process Modifications Suggested by Cooperating Companies to Overcome
Some TechnicalDisadvantages (94)
To overcome two of the main drawbacks of the basic Chisso NH3~scrubbing
process—the sulfitic plume frorn the absojrber and the formation of a rela-
tively unmarketable byproduct ^/(NH^^SO^/—Catalytic has modified the orig-
inal absorber and also added the Institut Francais du Petrole (IFF) S recovery
process to the original Chisso scheme.
By redesigning the original tray-type absorber and converting it to a
multistage unit with Individual liquid reclrculation for each stage, the
operating conditions are expected by Catalytic to be carefully controlled
such that the characteristic sulfitic plume is eliminated, As the flue gas
is passed through each successive stage of the tower, the concentration of
NH3 in the scrubbing solution is gradually decreased from stage to stage.
As a final treatment step, the final stage in the absorber contains a weak
circulating acid solution to remove most of the remaining NH-j and thus mini-
mize the NHj emissions from the system. This use of progressively decreasing
NH3 concentrations in the scrubbing solutions from the first to the last
stage in the absorber, coupled with the decreasing S02 concentration in the
flue gas and the acid scrubbing in the final stage, is said to eliminate the
sulfitic plume normally associated with Nt^-scrubbing processes such as the
original Chisso system.
45
-------
The second major modification of the original Chisso system Is the elimi-
nation of the byproduct (NH^^SC^ recovery section. As an alternate scheme,
the SO^"* solution from the Fe(OH)2 centrifuge Is pumped to an IFF recovery
section (see flow diagram in Figure 9) . The NH4+-salt solution is first _
pumped to an evaporator to remove the excess H20 and then the remaining S0^~
is combined with molten S and sent to an S0^~ reducer operating at 370 C
(700°F) where the S0^= is converted to NH3, t^O, and S02- These gaseous
products are reacted with a reducing gas over a proprietary catalyst to form
hydrogen sulfide (H2S) , i.e.,
2S00/, x + 6CO, , + 2H00, . + 2H0S, , + 6CXL, . (41)
2(g) (g) 2 (g) 2 (g) 2(g)
2SO_, . + 6H. , N -*• 2H.S/ N + 4H-0, , (42)
2(g) 2(g) 2 (g) 2 (g)
The resulting product stream is cooled to 160°C (320°F) and enters a liquid
Glaus unit where molten S produced by the following reaction is removed as
byproduct.
2H2S(g) + S°2(g)" 3S(1) + 2H2°(g) (43)
The off-gases of the Glaus unit, primarily H20 and NH3, enter an aqueous
stripper where the NH3 is removed prior to exhausting the waste gas. The
aqueous NHg solution recovered in the stripper Is concentrated by steam
distillation and recycled to the NH3 solution storage tank.
This IFF system has been tested on a 30-MW prototype unit at a power
plant near Paris, France, while the absorber has been tested at Cataly tic's
2-MW-equiv engineering optimization unit operating on a coal-fired boiler at
Calvert City, Kentucky.
The increased capital Investment and revenue requirements associated
with the absorber have not been released but the costs for the IFF process
for a 500-MW coal-fired boiler (3.5% S coal, 90% S02 removal, 7000 hr/yr)
have been estimated (15) as approximately $40/kW of generating capacity
and 2.0 mills /kWh respectively.
46
A
-------
H80
t
r I
(NH4)2S04 (Aq) •> EVAPORATOR
(CATALYTIC
S02,NH3 | S02 HjS,SO2,
J REACTOR]
REDUCING
GAS
RECOVERY —^
RECYCLE TO
ABSORBER
ELEMENTAL
SULFUR
STEAM
DISTILLATION
I I
TO AMMONIA (Aq)
STORAGE
Figure 9. Flow Diagram of IFF S Recovery Process.
-------
CHIYODA THOROUGHBRED 102 PROCESS - WET, OX1DATION-ABSORPT10N-
KEDUCTION (NOX-SOX)
Process Description and Principlea of Operation (35, 36, 57)
The Chiyoda Thoroughbred 102 process has been developed by the Chiyoda
Chemical Engineering and Construction Company, Ltd. This process uses 03 for
the gas-phaee oxidation of NO to N0£ and a weak H2S04 solution to absorb both
the S0£ and the fK^. The Chiyoda Thoroughbred 102 process is basically
similar to the other wet, oxidation-absorption-reduction processes in that
it consists of six distinct sections Including (1) prescrubbing, (2) gas-phase
oxidation of NO, (3) absorption of both S02 and NOX, (4) oxidation of SQ^
and reduction of absorbed NOX, (5) byproduct CaS04*2H20 production, and
(6) wastewater treatment. The basic outline of these sections can be seen
in the block flow diagram for this process in Figure 10,
The flue gas from the air heater is passed through a particulate pre-
scrubber, which removes 90% of the partlculates and essentially all of the
HC1 from the gas, also adiabatlcally humidifies and cools the gas from 150 C
(300°F) to 53 C (127°F). The scrubbing solution from the base of the pre-
scrubber drops to a particulate prescrubber holding tank from which most of
the solution is recirculated through the prescrubber after makeup 1^0 has
been added. The remainder of the scrubbing solution is removed as a purge
stream to prevent the buildup of Cl~ and flyash in the circulating solution.
This purge stream is pumped to a centrifuge where the flyash is removed and
the centrate is sent to the base of the absorber.
The gas passes through a mist eliminator at the prescrubber exit and
is sent to the absorber through a flue gas duct. In this duct the gas Is
Injected with a weak (<5%) O^-air mixture in an amount such that the resulting
mol ratio of 03 to NO in the flue gas is approximately 1.5 (for 80% NOX
removal). This O^-air mixture is generated onsite using a battery of corona-
discharge 03 generators. The NO is rapidly and selectively oxidized by the
Og to N0£.
N0(s)+°3(g) "N°2(g)+°2(g)
By positioning these 03 injection nozzles to give good mixing of the
gases, the following undesirable side reaction Is minimized.
2N°2(g) + °3(g) -
Following this Oj oxidation In the flue gas ducts, the flue gas passes
countercurrently to a weak (2-3%) 12804 acid solution in the absorber. The
is absorbed into the solution and undergoes the following reactions.
S°2(g) *
S02(aq)+H20"H2S03(aq) (47)
48
A
-------
< '
HEA
-------
NOX on the other hand is absorbed into the solution as the ditner mole-
cule (^04) and undergoes the following reactions.
2N02(g) * N2°4(g)
N2°4(g) + N2°4(aq)
N2°4(aq)+H2° +ra°2<.q) + HN°3(aq) (50)
2HN°2(aq) + 2H2S°3(aq) * N2°(g)f + 2H2S°4(aq) + H2°
2HN00, v + SH-SO,,,, ^-^ K«, .t + 3H_,SO, , , + H»O (52)
2(aq) 2 3(aq) 2(g) 2 4(aq) 2 v '
In addition to these primary reactions occurring in the absorption
section of the vessel, other secondary reactions occur simultaneously.
N2°(aq) + 2FeS°4(aq) + H2S04(aq)+ N2(g) * + Fe2(S04)3(aq) + H2° (53)
H2S°3(aq) + l'202(aq) H2S04(aq)
Fe4"2, % + 1/20./ , + H0SO~, . * Fe"1"3, s + H-SO. , s (55)
(aq) 2(aq) 2 3(aq) (aq) 2 4(aq)
The scrubbed flue gas is passed through a mist eliminator, reheated for
plume buoyancy, and sent to the stack.
The liquid effluent from the absorber is pumped to an oxidation tower
where any remaining dissolved S02 is converted to 112804. The oxidation
catalyst (Fe ) is also regenerated during its passage through the oxidation
tower (reaction 55) . Most of the liquid solution is recycled to the absorber
while a small bleed stream is removed and is initially sent to a holding tank
in the byproduct recovery section. From this holding tank the weak H2SC>4
solution is pumped to a unique reaction crystallizer patented (44) by Chiyoda
(see Figure 11). This solution ie pumped into the middle section of the
crystallizer and pushed up by an air stream injected below the H2S04 solution.
Near the top of the crystallizer makeup limestone is added to the H2S04 and
the following reaction occurs.
CaCO,, , + H,,SO, ( , + HUO * CaSO, '2E~Q, A + C00 , N t (56)
3(s) 2 4(aq) 2 4 2 (s) 2(g)
The CaSO/*2H20 crystals formed by this reaction are suspended in the
upward flowing air-solution mixture and overflow into the outer portion of
the annular crystallizer. Air is also injected into the outer region of the
crystallizer to suspend the large number of small CaS04*2H.2<) crystals In the
mother liquor and allow these crystals to coalesce and grow Into larger
crystals. These crystals will remain suspended in the crystallizer until
they have reached the predetermined size and overcome the force of the air
passing through the solution. After these crystals have separated and
50
A
-------
EXIT
AIR
LIMESTONE SLURRY
CRYSTCLLIZER
OVERFLOW
AIR
SULFURIC ACID *»t
SOLUTION FROM *
ABSORBER
AIR
GYPSUM SLURRY
Figure 11. Chlyoda Reaction Grystallizer (44).
51
-------
settled to the bottom of the crystallizer, they are removed in a bottoms
slurry and pumped to a centrifuge where the CaSO^'ZI^O is separated and sent
to a byproduct storage area. The mother liquor from the CaSC^'ll^O centri-
fuge and the overflow from the reaction crystalli2er are pumped to a thickener
where the bottoms containing mainly undersized CaSO^'Zl^O crystals, are
recycled to the crystallizer to serve as "seed" crystals. Most of the thick-
ener overflow stream is recycled to the top of the absorber after makeup
catalyst has been added, while a small amount is purged from the system to
prevent the buildup of N0j~ ions.
A catalyst recovery section has recently been added to this purge stream
to minimize the loss of the catalyst in the wastewater stream. The neutrali-
zation of the purg^e stream results in the precipitation of the catalyst as a
hydroxide J_(.^H)~_J and thus allows its removal in a centrifuge and its recycle
to be redissolved. The remaining liquid in this purge stream from the catalyst
recovery section is pumped to a biological treatment section where, with the
addition of methanol (CH30H) -t the N0j~ is reduced to molecular N2 in two
stages. Both stages are biological with the first stage being anaerobic and
the final stage aerobic.
*
Status of Development
The Chiyoda Thoroughbred 102 process was developed by adding an 03
injection stage to their more advanced FGD system, the Thoroughbred 101 pro-
cess. The former process was initially tested in a bench-scale unit treating
200 Nm3/hr (0.06 MW equiv) of flue gas from a heavy oil-fired boiler in 1973.
A larger bench-scale unit designed to treat 1000 Nnr/hr (0.3 MW equiv) of
flue gas from a heavy oil-fired boiler was started up in 1975 and was able
to maintain 90-98% desulfurlzation and 80% denitrification on flue gas con-
taining 500-2000 ppm S02 and 80-250 ppm NOX. The longest period of continuous
operation for this bench-scale unit has not been revealed. Plans for further
development, including acaleup to a pilot plant or prototype unit, also have
not been published.
This process has not been tested on flue gas from a coal-fired unit
although Chiyoda Thoroughbred 101 process (the FGD system), which is very
similar to thia denitrification process, is undergoing testing in a proto-
type unit (20 MW) on a coal-fired boiler at the present time.
Chiyoda has recently formed a wholly owned subsidiary, Chiyoda Inter-
national Corporation in Seattle, Washington, to market their FGD system in
the U.S. Chiyoda has made no plans yet to market their Thoroughbred 102
simultaneous S02~NOX removal process in this country and only limited amounts
of information are available from them.
Background of Process Developer
Chiyoda is a well-known Japanese company with design and construction
projects around the world. They have been involved in air pollution control
since the early 1970's, first in the development of an FGD process and later
modifying this process to simultaneously remove both SC>2 and NOX. The 13
prototype and commercial FGD plants built by Chiyoda for treating flue gas
52
A
-------
from heavy oil-fired boilers range in size from 20-350 MW. During 1975 a
23-MW prototype plant was started up and operated to test the ability of the
Chlyoda FGD process to treat flue gas from a coal-fired boiler at the Scholz
Steam Plant of Gulf Power.
Chiyoda's desulfurization process (Thoroughbred 101) was converted to
simultaneously remove SC>2-NOX in 1975 by simply adding an 03 injection
system to the desulfurization pilot plant. This modified desulfurization
process was renamed the Thoroughbred 102 process and tested first in a bench-
scale plant (0.06 MW equiv) and later in 1975 in a large (0.3 MW equiv) bench-
scale unit.
Published Economic Data
The following capital investment and revenue requirements for the
Ghiyoda Thoroughbred 102 processes are baaed on treating 750,000 Nm /hr of
oil-fired flue gas containing 1,000 ppm S(>2 and 150 ppm NOX and on S02 and
NOX removal efficiencies of 90 and 80% respectively. Under these conditions
the total capital investment has recently been estimated (36) as 8000 x 10"
yen and the revenue requirement as 8400 yen/kl of oil (assumed basis: Japan
and 1976 costs). If 300 yen/$ and 3000 Nm /hr/MW are assumed, this corres-
ponds to a capital investment of $107/kW of generating capacity and a revenue
requirement of 6.73 mills/kWh. In addition to these total costs, Chiyoda
has estimated the capital investment and revenue requirements for the 03
generating system as $40/kW and 2,72 tnills/kWh respectively. Thus the cost
of generating 0^ represents approximately 38% of both the capital investment
and also the revenue requirements for this system.
RawMaterials, Energy, and Operation Requirements
The major raw material requirements for the Chiyoda Thoroughbred 102
process are limestone for the neutralization reaction to produce byproduct
CaSO^-2H20 and 03 for the conversion of NO to NC^. However, since the 0-j
will be generated onsite with only the consumption of electricity and air,
it will not be considered as a raw material requirement but as an energy
requirement. The only other raw material needjed for this process is the
oxidation catalyst, ferric sulfate ^62(804)3_/, which is lost in the by-
product CaSO^"2H20 stream. No data have been published on the quantities
of the raw materials required for this process.
The major energy requirements for this process include fuel oil for
reheating the flue gas and electricity, mainly for 63 generation but also
for operating the process equipment. The only other utility requirements
would be for cooling E^O and process air. No data are available on these
utility requirements.
Teehnical Gone iderations
The Chiyoda Thoroughbred 102 process uses 03 for the gas-phase oxidation
of NO and a weak I^SO^-serubbing solution to absorb the resulting N02- This
system appears to require approximately the same number and type of process
equipment as other oxidation-absorption-reduction processes and hence the
53
-------
process control problems will be very similar to these other processes. Since
Qg cannot be stored but must be generated onsite, the major process control
problem will be the control of the 0-j generation and injection system. If
the 03 is injected in less than stoichiometric amounts, the NOX removal
efficiency will decline and, if Oj is injected in more than stoichiometric
amounts, the annual revenue requirements for this system will increase sharply
since the cost of generating 0/j is one of the major economic factors assoc-
iated with this process.
NOX removal efficiency ia relatively insensitive to the inlet gas compo-
sition. In the case of the Thoroughbred 102 process, this insensitivity is
due primarily to the fact that only approximately 50% of the absorbed NOX is
reduced by the 803" ion to molecular N2. The remainder of this absorbed NOX
leaves the system as NC>3~ salts without undergoing reduction by the SO-,™* ion.
The absorber and oxidizer operating conditions have not been specified
but the primary concern will probably be the costs associated with circulating
these large volumes of solution between the absorber and the oxidizer. Chiyoda
has recently designed a combined absorber-oxidizer for their Thoroughbred 101
FGD process and its adaption for use in the Thoroughbred 102 simultaneous
S02~NOX process would substantially reduce these costs.
Although the Thoroughbred 102 process has not been tested on flue gas
from a coal-fired boiler, Chiyoda does not expect any serious technical prob-
lems in modifying this system to handle the increased particulates and Cl~
associated with coal combustion. The major reason for this confidence is
their previous experience in modifying their Thoroughbred 101 FGD system to
handle coal-fired flue gas at a 23-MW prototype unit which has been operating
at the Scholz Steam Plant for the past year. The only major modification was
the separation of the prescrubber section from the absorber section, that is,
the prescrubber at the Scholz plant is a closed-loop operation with ash and
Cl~ removed from the prescrubbing solution and the "cleaned" solution recycled
back to the prescrubber instead of being pumped to the absorber as was pre-
viously done in oil-fired flue gas applications.
The adaptation of this process to treat coal-fired flue gas will, however^
create some economic problems. The increase in inlet NOX concentration in :
the flue gas in converting from oil to coal will cause a proportionate Increase
in the amount of 03 required and thus substantially increase the annual revenue
requirements for this process,
The CaSO^*2H20 produced as a_Jbyproduc£ will contain numerous salt impur*-
ities including calcium nitrate /Ca(N03)2_/ and 162(804)3. These salt impur™
ities will however be present in lesser amounts in the CaS04'2H20 cake from
the Chiyoda process than in the CaS04«2H20 from the other processes since
Chiyoda has developed a unique reaction crystallizer which yields a
CaS04'2H20 product containing large crystals that are much easier to wash
and centrifuge. For this reason the CaSO^^l^Q produced in this process,
without further processing, is much more likely to find a market in the U.S.
than the CaS04*2H20 produced by the other processes.
54
A
-------
It should be pointed out that the 162(804)3 is present in the scrubbing
solution to catalyze the oxidation of the S0^= ion in the scrubbing solution
during its passage through the oxidizer and is not used to aid in the absorp-
tion of N02.
The reasons behind Chiyoda's decision not to actively market this process
have not been published but is probably a combination of the following reasons.
The process is still in the early pilot-plant development stage. Secondly,
the NO^" wastewater problem will be very difficult to treat in a large-scale
unit and finally, the relatively low denitrification efficiency of the Chiyoda
process as compared with the other NOX removal systems. The reasons behind
the apparent maximum NOX removal rate of only 80% are not clear at the present
time, but they are probably related to the selection of an l^SO^ solution to
scrub the flue gas.
The purge stream from the regeneration section containing the N(>3~ will
be treated in a two-stage biological treatment scheme. An excess of Cl^OH
is added to the first stage under anaerobic conditions and the effluent from
this reactor is pumped to the second stage where, under aerobic conditions,
the excess CHjOH is removedt The NO-j™ is converted to molecular N£ which is
then exhausted. The remainder of the purge stream is pumped to a disposal
pond.
Since this is a wet simultaneous process, retrofitting this system on
existing power plants should be relatively easy if sufficient land is avail-
able for the siting of the process equipment and holding ponds. The major
modification would be removing the existing "cold" electrostatic preeipitators
(ESP) and installing a common plenum to connect the boilers to the four
scrubber trains. The use of this common plenum and four scrubber trains will
allow this system to be easily turned down by simply closing off any combina-
tion of scrubber trains. The ozone generation system can also be easily
turned down since it is made up of a battery of smaller ozone generators.
The process equipment and piping for this system will be primarily made
up of either stainless steel or elastomer-lined carbon steel depending on
whether the equipment is in the absorption or regeneration section. The
absorber section containing the circulating weak H2S04 solution will require
316 stainless steel while those pieces of equipment in contact with the
circulating CaS04"2H20 slurry will require elastomer-lined carbon steel,
The incoming flue gas ducts, of course, will be constructed of Inconel 625
to prevent the corrosion associated with treating high-temperature, S-
containing flue gas.
Environmental^ Considerations
During the operation of a pilot plant treating 1000 Nm3/hr (0.3 MW equiv)
of flue gas from an oil-fired boiler (normal composition: 1500 ppm S02» 150
ppm NOX, 4% 02), the S02 removal efficiency was typically 90-98% while the
NO removal efficiency was approximately 80%. The S(>2 is absorbed as S0>>~,
oxidized to SO^™, and recovered as CaSO^'2H20. The NOX is oxidized by 63,
absorbed into the solution, and removed from the system.as either molecular
N£ or N03~ in the wastewater purge stream.
55
-------
The NOX removal efficiency for the Chiyoda process is dependent on the
amount of 03 injected as shown in Figure 12. As can be seen from this figure,
from an economic point of view the optimum NOX removal rate for the Chiyoda
process is approximately 60%. Since one mol of 03 oxidizes 1 mol of NO to
N02> this 60% NOX removal occurs at an 03 to NOX ratio of 0,5. Since the
generation of Oj is extremely expensive, economically it would seem pointless
to approximately triple one of the major operating costs to increase the NO
removal efficiency from 60 to 80%. For high NOX removal requirements, i.e.,
greater than 80%, this process would not be recommended,
The 03 used to oxidize the NO in the flue gas is a significant work
hazard associated with this process. This gas is a very strong oxidant and
can be dangerous if workers are exposed to even low concentrations. The
circulating 1*2804 solution, although relatively weak, is also a hazard for
workers coming into contact with it.
Critical Data Gaps and Poorly Understood Phenomena
Included under critical data gaps and the preliminary engineering
assumptions used to overcome these data gaps are;
1. The raw material and energy requirements.
2. The portion of the absorbed NOX converted to $03" salts in the waste-
water. Assumed to be 50%.
3. The operating conditions in the absorber and the oxldizer.
4. Oxidation of 1^803 in absorber? In oxidizer?
5. The minimum ratio of S0£ to NOX in flue gas for 80% NOX removal
without requirement for makeup sulfurous acid (12803)?
6, Maintenance aad manpower requirements. •
Advantages
The major advantages and disadvantages of the Chiyoda Thoroughbred 102
process are listed below.
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >95% SO, removal efficiency
3. Produces a potentially marketable byproduct (CaSO^'Zt^O)
4, Is a slight modification of a commercially-available FGD system
5. Operates with full particulate loadings (>7 gr/sft3)
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Has been tested only in a bench-scale unit
3. Uses significant amounts of stainless steel or exotic materials
for process equipment
4. Requires flue gas reheat for plume buoyancy
5. Requires an expensive gas-phase oxidant
6. Incorporates unique treatment methods to prevent secondary sources of
pollution (biological wastewater denitriflcatlon)
56
A
-------
100
90
80-
70
60
50
8
«
i* 3°
20
10
0
I
NO*
INLET GAS
NO PPMV 600
NOtPPMV
S02 PPMV 1300
TEMPERATURE: 55°C
ABSORBER COLUMN;
60MM DIA.xIQOOMM
HIGH
II I I II
1.5
VALUE OF X IN NO,
Figure 12. Effect of NOX Oxidation Level oti
NOX Absorption Efficiency for Chlyoda Process (113)
57
-------
ISHIKAWAJIMA-HARIMA HEAVY INDUSTRIES PROCESS -WET, OXIDATION- ABSORPTION-
REDUCTION (NO -SO,,)
X «•
Process Description and Principles of Operation (11, 60, 116, 117)
In the Ishikawajima-Harima Heavy Industries (IHI) process the NO is
oxidized to N02 by injecting 03 into the flue gas stream. This N0£ is then
absorbed into the solution to form an equiinolar aqueous mixture of N0£~ and
NOg". Most of these N0£~ and NOj are reduced to molecular N2 and complex
N-S compounds by the 803™ ions formed from the absorbed S02« A small amount
of these N03~ salts is removed as a byproduct resulting from evaporation of
a waatewater purge stream. The flue gas S02 is absorbed into solution as
SOg™ ions and oxidized to SO^ ions in an oxidizer. These SO^^ ions are
removed in a centrifuge as byproduct
The overall system is very similar to the other wet, oxidation-absorption-
reduction processes, in that it consists of six major sections (1) prescrubbing,
(2) gas-phase oxidation, (3) absorption of both S02 and NOX, (4) oxidation of
SOg" and reduction of absorbed NOX, (5) byproduct CaS04'2H20 recovery, and
(6) wastewater treatment. A more detailed outline of the IHI process is
shown in Figure 13.
The flue gas from the air heater passes through a prescrubber which
removes most of the particulates and essentially all of the Cl~. The flue
gas is cooled from 150°C (300°F) to 53°C (127°F) and humidified by the evapo-
ration of i^O from the scrubbing solution. Most of the liquid effluent from
the particulate scrubber is recycled to the prescrubber, while part is purged
to a centrifuge where the flyash is removed. The centrate, although not
specifically mentioned by IHI, is probably recycled to the prescrubber holding
tank after a small purge stream has been removed for Cl~ control.
As the cooled gas passes through the connecting ducts , 0^ in a weak 0^-
alr mixture, at approximately 80% of the stoichiometric amount required, is
injected to selectively oxidize the NO to N02 by the following reaction:
NO, % +-0Q/ % -*-NO,,,, v + 0,,, N (57)
(g) 3(g) 2(g) 2(g)
The Oj injection system must be properly designed to give good mixing
characteristics and thus prevent the following undesirable side reaction:
2N°2(g) + °3 * N2°5(g) + °2(g) (58)
Following oxidation, the flue gas passes countercurrently to a lime or
limestone slurry in a Turbulent Contact Absorber (TCA) . This slurry at a pH
of 4-6 also contains copper chloride (CuC^) and NaCl as catalysts at a con-
centration of about 0.005 molar to aid in the absorption of N02- The S02
is absorbed from the flue gas and undergoes the following liquid-phase reac-
tions:
58
A
-------
f
DOll CD
bmi_tK
f
AIR
HEATER
AIR
j
I
* P
1
PF
.
A
'\
RESC
H20
• i
?ESCF
HOL
TA
J
Ul
CLEAN
GAS
OZONE
-GENERATOR
ASH
CENTRIFUGE-
^J PURGE f
ASH
SLURRY | lilMESTONE
*~PREPARATION-* PREPARATION *~CoC03
AIR
\
1 GYPSUM
CENTRIFUGE
1
I
GYPSUM
1
, NEUTRALIZATION .
REACTOR,
1
N2
H20
CoS04
CaClg
ITHERMAL
DECOMPOSER
Figure 13. Flow Diagram of Ishikawajima-Harima Heavy Industries Process.
-------
S02(g)+S02(aq) (59)
S0», % + CaCO^f , + 1/2H00 -»• CaSO ' 1/2H00, ,. 4 + C00 , ,.-}• (60)
2(aq) 3(s) 2 3 2 (s) 2(g)
S°2(aq) + CaS03'1/2H20(s) + 1/2H2° * Ca(HS03)2(aq)
The NOo Is absorbed and undergoes the following reactions In the
scrubber solution.
2N02(g) ^° <62)
N204(g)*N2°4(aq)
0)., % + 4CaSO ' 1/2H,,0, N + 6HLO -*N0, ,t+ 4CaSO, '2H70., ^ + (64)
2 4(aq) 3 I (8) 2 2(g) 4 / (s)
(65)
Ca(N°3)2(aq) + ^^ (66)
Ca(HSO,)0x .* -*- ,,,x ,
J z(a<|) J / (s)
In addition to these primary reactions, the other secondary reactions
are occurring in the absorbing solution. Q£ is also absorbed from the flue
gas and oxidizes this SOj™ and 504" by the following reaction.
2, , -f 3/2H2<> •* CaS04'2H20, ,4- (67)
The NOX, including some ^€5, can also undergo the following reactions.
N^O,,, % + 3CaSO,'l/2H00,, , + 3/2H00-> N00., J + 3CaSO. '2H.O., % (68)
2 4(aq) 3 2 (s) 2 2 (g) 4 2 (s)
N,0,x , + 2CaSO,'l/2H00,, . ->Ca(NO-)0,, '+Ca(HSO,)0/ . (69)
2 5(aq) 3 2 (s) 3 2(aq) x 3 2(aq)
Thus the S(>2» absorbed to form the S0.,= ion, Is used to reduce the N(>2
to molecular N£ and to complex N-S compounds. Approximately 80% of the
absorbed NOX is converted to calcium sulfamates /Ca(NH2S03)2_/ and most of
the remainder to molecular N£. According to the process developer, byproduct
NOj -N02~* salts represent only a small amount (<10%) of the total absorbed
NOX.
After passing through a mist eliminator, this cleaned flue gas is
reheated to provide plume buoyancy and is then exhausted through the stack.
60
A
-------
From the absorber, the solution flows into a holding tank where small
amounts of fresh catalyst (as required) and either slaked lime or limestone
are added. Most of this slurry is recycled to the absorber while a small
(1-2% of the circulating solution) stream is sent to an oxidizing tower in
the regeneration section. Any CaSO-j- 1/2H20 remaining in this purge stream
is converted to CaSO^ as it passes countercurrently to an air stream in the
oxidizing tower. This CaSO^ slurry from the oxidizing tower is then eentri-
fuged to remove byproduct
The final treatment step for the centrate, which contains the complex
N-S compounds and the N0j~ salts, has not been firmly established as yet.
IHI offers three alternative methods for treating this system (1) evaporation
and atmospheric thermal decomposition of the N-S compounds, (2) pressurized
thermal decomposition, or (3) chemical decomposition method. Each of these
treatment methods would form molecular N2 but would be expensive.
At the present time IHI is apparently favoring the evaporation-
decomposition method for removing the NOj" and Ca(NH2 863)2 from the waste-
water stream. The liquor from the 0*5804*21120 centrifuge is first neutralized
with limestone slurry to prevent the formation of NH^ and pumped to a thick-
ener. The bottoms stream from the thickener, containing mostly excess lime-
stone slurry, is removed and recycled to the absorber. The thickener overflow
is pumped to a steam evaporator to concentrate the solution before entering
the thermal decomposer. At the higher temperatures in the decomposer the
Ca(NH2S03)2 and the N03~ salts are decomposed to molecular N2 and CaS04'2H20.
The decomposition of the Ca (ML 803) 2 is given by the following reaction.
2(HH,SO,)0Ca, , + 4CaCO,, \ + 2Ca(HSO, )„ ., x + 300, N + SH-O •>
2 3'2 (aq) 3(aq) v 4'2(aq) 2(aq) 2
2N0, J + 8CaSO,'2BLO, J + 4CO., A (70)
2(g) 4 2 (s) 2(g)
The off-gas may contain some S02 in addition to the N£ and C02 and will
be recycled to the flue gas ducts. The remaining slurry from the decomposer
containing mainly CaS04*2H20 will be centrifuged and the centrate will be
purged.
Status of Development
The Initial development work for the IHI process was done on a bench-
scale unit treating simulated flue gas. Bench-scale testing was begun in
mid-1974 in a 500 Nm3/hr (0.16 MW equiv) unit treating flue gas from an oil-
fired boiler. A larger pilot plant treating 5000 Nm3/hr (1.6 MW equiv) of
flue gas from an oil-fired boiler, containing 1150 ppm S02 and 180 ppm NOX,
has been operating since September 1975. Over a long-term (3000 hr) contin-
uous operation the S02 and NOX removal efficiencies in this pilot plant
averaged 90 and 80% respectively.
It has recently been reported (95) that IHI is testing their process on
a 27,000 Nm3/hr (9 MW equiv) prototype unit at Tokyo Shibaura Electric Com-
pany near Tokyo. Additional information has not been published.
61
-------
iHI is now working toward reducing the costs associated with the genera-
tion of Og and also the development of an effective but inexpensive method of
treating wastewater from their process.
This process has not been tested on flue gas from a coal-fired boiler.
The next development step for the IHI process would be to demonstrate the
long-term reliability of this system in their prototype unit. The ability of
this process ito handle coal-fired flue gas also needs to be demonstrated,
Background of Process Developer
IHI Is a relatively large Japanese shipbuilding and metal fabricating
company which entered the air pollution control field in the early 1960's.
IHI is a leading manufacturer and marketer of ESP, TCA, and complete FGD units
in Japan. The total SC>2 removal plant capacity built by IHI in the last 5
yr is approximately 10 million Nm /hr (3300 MW equlv) with. about 501 baaed
on the wet lime-llmestone-CaSO^*2H20 recovery system.
No American company has yet been licensed by IHI to market their process
in the U.S. However IHI has a. subsidiary, Ishikawajima-Harima Heavy Indus-
tries, Inc., in New York City (branch offices in San .Francisco and Houston),
through which inquiries about this process are handled.
PubllBhed Economic Data
The following total capital investment and annual revenue requirements
are based on economic data provided by IHI for treating flue gas from an oil-
fired boiler (1500 ppm S(>2, 180 ppm NOX, and 0.1 g/Nm3 of dust). Assuming
SC>2 and NOX removal efficiencies of 90 and 80%, respectively, the total
capital investment for a 500,000 Nnr*/hr (167 MW equiv) unit is estimated
(117) at 4,200 million yen and the revenue requirement is estimated at 8,200
yen/kl of oil (construction basis; Japan and 1976 costs). If 300 yen/$ and
3000 NnrVhr/MW equiv are assumed, these values correspond to a capital invest-
ment of $84/kW of installed capacity and revenue requirement of 6,8 mllls/kWh«
Raw Material , Energy j
Operation Requirements
Assuming the same basis as was given above for the published economic
data, i.e., 500,000 Nm^/hr of oil-fired flue gas containing 1,500 ppm S02 and
180 ppm NOX and removal efficiencies of 90% for S0£ and 80% for NOX, the raw
material requirements would be;
Raw material
Quantity
CaC03
H2S04 _
Slaked lime _/Ca(OH)
Additives
3,5 tons/hr (short tons)
«_
352 lb/hr
154 lb/hr
62
A
-------
The utility requirements under these same conditions would be:
'"Utility _ (Quantity
Electricity 8,400 kW
Fuel oil 4,675 Ib/hr
Steam 9,460 Ib/hr
H-O 240 gpm
This quantity of electricity represents approximately 5% of the generat-
ing capacity of the boiler. For treating flue gas containing 2400 ppm S0£
and 600 ppm NOX, the electrical requirements have been estimated (117) as
about 9.2% of the boiler generating capacity.
The operating personnel needed for the IHI processes are nine men per
day.
The maintenance requirements for this process have not been published,
Technical Considerationa
The IHI simultaneous S02~NOX removal process, since It Is very similar
to the other wet, oxidation-absorption-reduction processes, will have the
same process control problems. The major control problem will be the 03
generation and injection system since 0^ is extremely reactive and cannot
be stored economically. As the amount of 0-j injected falls below the stolch-
lometric amount required, the NOX removal efficiency gradually drops from
80-85% until the NOX removal efficiency reaches approximately 5% when no Og
is injected. Increasing the amount of Og Injected above the atolchiometric
amount sharply increases the operating costs for this system, since the cost
of 03 generation represents a substantial portion of the operating costs
for this system.
The S02 removal efficiency is insensitive to the inlet flue gas compo-
sition and will remain over 90% under reasonable absorber operating conditions,
The overall design of the system is very similar to a conventional FGD system
with a L/G ratio in the absorber of about 10-12 1/Nra3 (60-75 gal/kaft3 at
127°F).
The NOX removal efficiency is also relatively insensitive to the inlet
flue gas composition provided that the mol ratio of S02 to NOX in the flue
gas remains greater than 2,5. Although not mentioned by IHI, this ratio was
assumed since this process, at least in the absorption step, is nearly iden-
tical with other oxidation-absorption-reduction processes, which operate
under similar conditions and apparently by the same mechanisms. This rela-
tively high mol ratio is required since the absorbed S02, in the form of the
SOg™ ion, is used to reduce the absorbed NOX. When the S02~NOX mol ratio
in the flue gas falls below 2.5si, insufficient amounts of the SQ^*3 ion are
available to reduce the NOX and the NOX removal efficiency will drop below
80-851.
63
-------
louring bench-scale testing, a spray tower was used as an absorber and
the NOX removal efficiency as a function of various absorber parameters aad
scrubbing solutions was studied. Initially a comparison of Na2S03 versus
CaSC>3 was made with the result being that the NOX removal efficiency of a 5%
Na2SO-j solution was approximately 70%, whereas the efficiency of a 5% CaS03
slurry was only about 20%, (Although not mentioned by IHI, this large
difference in NOX removal efficiency is probably due to the differences in
the solubility of this S0^m in the scrubbing solution.) However, by simply
adding a small amount (0.01-0.02 aol/1) of CuCl2 mixed with either NaCl or
calcium chloride (CaCl2) (0,17 mol/1) to the scrubbing solution, the NOX
removal efficiency can be sharply increased to 80-90%. From these studies
by IHI it is apparent that the addition of an Oj generation and an H^O-soluble
catalyst to a conventional limestone FGD system will result in a feasible
simultaneous S02~NOX system.
One of the most important operating parameters for this process is the
L/G ratio in the absorber. From preliminary pilot-plant tests conducted by
IHI, Figure 14 shows a comparison of the NOX removal efficiency as a function
of the L/G ratio in the absorber. For 80-90% NOX removal an absorber L/G
ratio in the range of 10-12 1/Nnr (60-75 gal/kaft3 at 127°F) will be required.
Although this process has not been tested on flue gas from a coal-fired
boiler, the only two major technical considerations in adapting this system
for treating flue gas from a coal-fired boiler would be the addition of a
closed-loop prescrubber section to remove the increased partlculate and Cl~
loadings associated with coal combustion and secondly, what type of H20
treatment system would be required to prevent the discharge of excess N com-
pounds in the wastewater. Although IHI estimates only a small amount of NOj*
is formed, that statement is based on NOX concentrations of 100-200 ppm instead
of the 600 ppm or higher NOX levels which are associated with coal-fired flue
gas. This, of course, also leads to economic considerations in converting
this process to treating coal-fired flue gas since a wastewater treatment
section will probably be required and this will increase the costs for the
process. In addition, tripling the NOX levels will approximately triple the
03 generation costs and, since the cost of 03 is one of the major operating
costs associated with this process, this high concentration of NOX will also
increase the annual operating costs,
Recently published information for their 5000 Nm /hr pilot plant suggests
that IHI has modified the regeneration section of this process to eliminate
this potential wastewater problem. Apparently after the byproduct CaSO^-Z^O
has been removed in the centrifuge, the remaining liquid is evaporated at
90°C (194°F) to form a solid mixture of 03(^2803)2, CaS04, CuCl2f and CaCl2.
This mixed salt product is then conveyed to a thermal decomposer operating
at 900°C (1650°F) where the Ca(NH2S03)2 is decomposed into CaS04 and molec-
ular K£. Detailed information on this new regeneration scheme has not been
made available.
As with the other wet simultaneous removal processes, retrofitting this
system on existing power plants would be relatively easy if sufficient land
is available nearby for siting the various process equipment. The major
modification to the existing plant would be the installation of a common
64
A
-------
too
3*
*i
o
z
UJ
o 80
u.
U-
yj
60
O
Z
40
I
1
8 10 12 14
ABSORBER L/G, l/Nm3
16
Figure 14. Effect of Absorber L/G on NOX
Removal Efficiency for IHI Process (60).
65
-------
plenum from the boilers or the ESP to feed the four scrubber trains on a
500-MW boiler. The existing cold ESP could be removed and replaced with
the venturi-type prescrubbers to remove the Cl~ as well as the flyash from
the flue gas. The use of a common plenum and four absorber trains will allow
this system to cycle with the boiler by simply closing off any combination
of trains. The Oj generation section can also be easily turned down since
this section is made up of a battery of small Oj generators, any combination
of which can be shut down independently.
Most of the process equipment and piping for this process will need to
be elastomer-lined carbon steel to prevent the corrosion and erosion problems
associated with handling circulating limestone slurries and Cl~ solutions.
Exceptions to this elastomer lining include the flue gas ducts and the thermal
decomposer. The flue gas ducts will be constructed of Inconel 625 to prevent
both 12804 mist corrosion and the oxidation associated with the use of 03.
The decomposer will probably require either stainless or glass-lined steel
due to the extreme reaction conditions required to decompose the N-S compounds.
Environmental Considerations
Although this process has not been tested in a large pilot plant, the
developers believe this process can remove more than 90% of the SC>2 and 80%
of the NOX from oil-fired flue gas containing 1000 ppm S02 and 200 ppm NOX.
The absorbed S02 is oxidized to S0^~ and is removed as CaSO^^I^O. Approxi-
mately 80% of the absorbed NOX is reduced to CaCN^SO^ with most of the
remainder converted to molecular $2* However, some of the absorbed NOX will
be converted to 63(1*03)2 which Is very soluble in aqueous solution. These N
compounds, i.e., both CaCNlE^SOj^ and Ca (1103)2, can then be thermally decom-
posed to CaSGA*2H2Q and N£ to prevent a significant wastewater problem.
The M)x removal efficiency is dependent on the amount of 0«j injected as
can be seen in Figure 15. As the amount of 03 Injected decreases below the
stoichiometric requirement, the HOX removal drops below 85%. A complete
failure of the Oj generation section will result in only approximately 5% of
the NOX being removed (i.e., only the NOX already present as N02 in the flue
gas). On the other hand, a complete failure of the 63 generation section
will not affect the S02 removal efficiency.
The 03 used to oxidize the NO in the flue gas is a very strong oxidizing
agent and hence, even at low concentrations, it presents a significant work
hazard for this process.
Critical DataGagsand Poorly Understood Phenomena
The critical data gaps for this process included:
1. Stream compositions and operating conditions in each piece of
process equipment
2. The maintenance requirements for the IHI process
3. Reactions and mechanisms in the decomposer
66
-------
100
80
8*
o
z
UJ
o 60
u.
u.
LU
o
2
UJ
Q£
x
O
40
20
0.2
0.4
0.6 0.8
03 / NO MOL RATIO
1.0
Figure 15. Effect of Ozeme/NO Mol Ratio on the NOX
Removal Efficiency for IHI Process (60).
67
-------
Advantages and Disadyan.tage8
The major advantages and disadvantages of the IHI process are
listed below.
Advantages
1. Removes NOX and S02 simultaneously
2. Produces a potentially marketable byproduct (CaSO^-ZI^O)
3. Is a slight modification of a commercially available FGD system
4. Operates with full particulate loadings (>7 gr/sft^)
Disadvantages
1. Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Uses significant amounts of stainless steel or exotic materials
for process equipment
4. Requires flue gas reheat for plume buoyancy
5. Requires flue gas constituents within specific ranges for high
NO,, removal
A
6. Requires an expensive gas-phase oxidant
68
-------
KAWASAKI HEAVY INDUSTRIES MAGNESIUM PROCESS - WET, OXIDATION-ABSORPTION
(HQX-SOX)
Process Description and^ Principles^ of Operation (65, 66)
This process, which was developed by Kawasaki Heavy Industries,__is a wet
"simultaneous" SQ2-NOX process using a magnesium hydroxide /Mg(OH)y /•alurry
and 03 injection. This absorber is unique in that it is divided into three
sections, one section for desulfurization, the second for denitrification
based on equimolar NQ-N02 absorption, and the third for further denitrifica-
tion based on 63 injection and N02 absorption. Without the third stage of
absorption using 03, NOX reduction below 150-200 ppm is difficult to attain,
while with 03 injection 80-90% denitrification can be achieved. The overall
efficiency depends on the OjsNO ratio in the third section; the higher the
Q3:NO ratio, the higher the overall removal efficiency. Since an NOX
removal efficiency of 90% is desirable, only the three-section absorber with
03 Injection will be considered. In addition to 90% NOX removal, this
process is aleo able to remove more than 95% of the S02 in the flue gas.
The Kawasaki process consists of five major sections including (1) gas-
phase oxidation, (2) absorption of S02 and NOX, (3) regeneration and
oxidation ojf recycle NO, (4) CaS04'2H20 production, and (5) regeneration of
absorbent /Mg(OH)2_/- The overall process outline is shown In the flow
diagram in Figure 16.
The flue gas from the air heater passes countercurrently to an Mg(OH>2
slurry in the first, or desulfurization, section of the absorber. As the
flue gas passes through this section, the gas is cooled and humidified as the
S02 is stripped from the flue gas and subsequently enters the absorbing
solution. During this absorption, the S02 undergoes the following reactions:
S02(g) + S02(aq) (71)
Mg(OH)2(s) + S02(aq) + 5H20 •>• MgS03'6H2Q(s)4- (72)
MgS03'6H20(s) + S02(aq) + Mg(HS03)2(aq) + 5H20 (73)
Mg(OH)2(s) + Mg(HS03)2(aq) + 4H20 + 2MgS03'6H20{s)4- (74)
The liquid effluent from this desulfurization section is pumped to a
liquid holding tank while the flue gas enters the first of two denitrifica-
tion sections in the absorber.
In the first denitrification stage of the absorber, the flue gas is
mixed with a recycle N02 stream to adjust the mol ratio of NO to N02 to 1. As
the flue gas passes countercurrently to an Mg(OH)2 slurry, equimolar amounts
of NO and N02 are absorbed and converted according to the following
react ions:
69
-------
CLEAN
FLUE GAS
REACTOR
CRYSTALUZER
PURGE
Ga(NOs)2
Figure 16. Flow Diagram of Kawasaki Heavy Industries Process.
-------
N2°3(g) •* N2°3(aq)
M*(OH)2(aq) + N2°3(aq) * M^(N°2)2(aq) + H2° <77>
As the concentration of NO in the flue gas decreases below about 200 ppm,
the rate of reaction (75) decreases and becomes negligible and the remaining
NQ2 is absorbed by reactions (79) , (80) , and (81) , shown "below. Further NOX
reductions below 150-200 ppm require the use of 63 and a separate dentrifi-
cation stage. The liquid effluent from this first denitrification stage is
sent to the absorber holding tank and mixed with the effluent from the
desulfurization section, while the flue gas containing 150-200 ppm NOX passes
overhead and is injected with 03 to oxidize NO to N02 before entering the
second denitrification section.
The flue gas then passes counter currently to an Mg(OH)2 slurry in the
second denitrification stage of the absorber where the N02 undergoes the
following reactions :
2N02(g) .* N2°4(g) (79)
N2°4(g) * N2°4(aq) <80>
2N2°4(aq) + 2M*(OH)2(s) * I*(H03)2(an) + ^^(aq) + 2H2°
The liquid effluent from this denitrification stage is split into two
streams— one stream is recycled to feed liquid solution to each section of the
absorber while the other enters the holding tank below the absorber and is
mixed with the effluent from each of the other two sections of the absorber.
The scrubbing solution in the holding tank thus contains both dissolved
N0»~ and N0j~ salts and a slurry of S0-j= and S0,= and is pumped to the
regeneration section,
The first stage of the regeneration section consists of a thickener to_
separate the soluble magnesium nitrites /Mg(N02)2_/ and nitrates /Mg(NO^)y / ,
which are removed in the thickener overflow, from the insoluble 803" and
SO/,*, which are taken out in slurry form from the bottom of the thickener.
The thickener overflow containing the mixed N02~ and N03~ salts is pumped to
a reactor where the N02~ is decomposed by the addition of 12804 according
to the following reactions
71
-------
2MgS04(aq) + Mg(N03)2(aq) + 4NO(g)t -h 2^0 (82)
This NO, which Is insoluble in solution, la removed as a gas stream and
is passed through an oxidation tower where the NO is oxidized in air to form
""(E) + 1/2°2(g) * 21)°2
This N02 is then injected into the flue gas prior to the first denltri-
flcation section of the absorber to adjust the ratio of NO:N02 to 1:1 and
thus facilitate NOX absorption by reactions (75), (76), and (77).
The liquid effluent from the N02~ decomposition reactor, containing
magnesium sulfate (MgS04) and Mg(N03)2 is mixed with the 803" slurry bottoms
from the thickener and sent to a second oxidizing tower where an air stream
is used to oxidize the magnesium sulfite (MgS03"6H20) to
l/202(g) - MgS0 + 60 (84)
The MgSO^ removed from the oxidizing tower is further treated with
*n a double decomposition reactor to produce CaSO^^E^jO and
Ca(N°3)2(aq) + ^S°4(aq) + 2H2° * CaSV2H2°(s)* + ^ <85)
The resulting CaS04"2H2Q slurry is pumped to a centrifuge where the
Is removed and sent to a byproduct storage area.
The centrate, containing mainly Mg^Qj)^, is converted in a second
double decomposition to insoluble Mg(OH)2 by reaction with a makeup slurry
of Ca(OH)2-
Ca(OH)2(8) * Ca(N03)2(aq) + M*2(s) * (86)
After separating this Mg(OH)_ as a slurry product from the bottom of a
second thickener, the Mg(OH)2 Is recycled back to the absorber. The over-
flow from the thickener, containing mainly Ca(N03)2» is split Into two
streams: one for recycle to the CaS04*2H20 production reactor and the other
to a byproduct recovery section. It Is hoped by the developers that, by
concentrating this Ca (1103)2 byproduct stream, it can be sold as a low-grade
liquid fertilizer. If this €3(1103)2 cannot be sold as is, the developers
suggest decomposing this Ca(N03)2 into CaO and NO which will then be reduced
with NH3, i.e.,
^"•Vztaq) J ^(s) + 2NO(e) + 3/202(8) <87)
* 5N2(S) + 6H2°(g) (88)
72
A
-------
Although this series of process steps Is technically feasible, the high
temperatures and complicated process equipment requirements would make this
process expensive.
Other possible solutions to the CaCNO^^ problem would be either
evaporation and disposal in a landfill or biological treatment in a series of
holding ponds. These would be expensive alternatives and hence disposal of
the 03(^3)2 as a liquid fertilizer, whether sold as a byproduct or given
away, would be a more suitable solution.
Status of Development
The absorption section of the Kawasaki simultaneous S02~NOX process was
first tested in a bench-scale unit treating 100 Nm3/hr (0.03 MW equiv) of
flue gas from an oil-fired boiler. However, high denltriflcation efficiencies
were not attainable using only equlmolar absorption and therefore testing
was begun in March 1974 using 03 for the gaa-phase oxidation of NO.
This resulting process was scaled up to treat 5000 Nm-Vhr (1.6 MW equiv)
of flue gas from a coal-fired boiler in June 1975. Very little data have
been published concerning the operation of this pilot plant.
Background of Frocess Developer
Kawasaki la a large, diversified Japanese company with extensive
previous experience in air pollution control technology. This involvement
began in 1968 with the start of research aimed at the development of an
FGD system. Initial work on an FGD system included studying an Mg scrubbing-
regeneration process, an Na scrubbing process, and a Ca scrubbing process.
Approximately seven prototype units using the Na scrubbing system and two of
the Ca-based scrubbing systems were completed between 1973 and 1975.
In 1971 Kawasaki entered into an agreement with Kureha Chemical Industry
Company for the Joint development of the Na2S03-CaS04'2H20 double-alkali
FGD system. Three commercial units based on this technology are now
operating and two more commercial units are under construction.
The development work for their Mg-CaSO, '2HLO double-alkali FGD process
was begun in 1972 with scaleup to a pilot-plant unit In early 1975. Two
commercial plants based on the Mg-CaS04.2H20 system were completed and began
operating in 1976.
No American company has been licensed by Kawasaki to market their
process in the U.S. However, Kawasaki does have a liaison office in
New York City through which Inquiries about their process are handled.
Published Economic Data
Neither the estimated total capital investment nor the annual revenue
requirements for the Kawasaki simultaneous S02-N0x process has been
published.
73
-------
Raw Mivcerialsr Energy, jmd_ _0p_er_ation RequirementB
The major raw material requirements for" this process will include lime-
stone (or lime) to remove the 804= as byproduct CaS04'2H20, 12804 for the
decomposition of Mg(N03)2» and lesser amounts of Mg(OH)2 to replace Mg losses
throughout the system. The quantities of these raw materials required have
not been published.
The major energy requirements for this system will include fuel oil for
flue gas reheat and electrical energy for both 0, generation and process
equipment operation. Other utilities would include cooling and makeup
These utility requirements have not been published.
The operating manpower and maintenance requirements for this system have
not been made available.
Technical Conside ra t ions
The Kawasaki process, with its extensive equipment requirements and 03
generation section, will be one of the more complex processes and hence one
of the most difficult to control. The major control problems will be the
absorber since it is apparently split into three separate scrubbing sections
in series with three different liquid feed streams and each section
operating under different absorption conditions. Controlling the Oj system
will also be relatively easy since the amount of 03 required in the second
denitrlfication section will be relatively constant regardless of the inlet
flue gas composition.
The S(>2 removal efficiency is insensitive to the inlet flue gas
composition since the first section of the absorber is essentially a con-
ventional Mg FGD scrubber and it is followed by two other absorber sections
using an Mg slurry. Thus the desulfurizatlon efficiency should be con-
sistently over 95%.
The NOx removal efficiency should also be insensitive to the inlet flue
gas composition since the NO is not reduced by the SQ^* ion as in other wet
processes but are simply oxidized and removed as NOV" ions in the wastewater.
As mentioned above, from the limited amount of information available at
the present time, it appears likely that each of the three sections of the
absorber operate under slightly different conditions. The desulfurization
section of the absorber operates at an L/G ratio of 5 l/Nm^ (31 gal/kaft^ at
127°F) and a pressure drop of 150 mm H^O to give an S0£ removal efficiency
of 95%. The second, or equimolar NO-N02 absorption section, operates at an
L/G ratio of about 10 1/Nm3 (65 gal/kaft3at 127°F) and removes 40-50% of the
incoming NOX. The third, or NQ2 absorption section, also operates at an L/G
ratio of about 10 1/Ntn3 but the NOj^ removal efficiency in this section is a
function of the 03; NO ratio. Under the L/G ratio specified above and an
03:NO ratio of 1.4, approximately 85% of the remaining NOX can be removed.
Therefore, an overall NOX removal efficiency of about 90% can be attained.
74
A
-------
At the present time the actual design of the absorption section is not
particularly clear especially for larger prototype units. Since the three
sections of the absorber are physically separated from each other and each
section is estimated to be approximately the same size as a conventional
Mg FGD scrubber, when this process reaches a commercial size more than one
absorber in each train may be necessary. The increased cost of such a system
would probably make this process very expensive t
The CaS04'2H20 byproduct stream would be contaminated with various
and 803" salts and hence would probably not be marketable in the U.S. where
high-quality natural CaS04"2H20 is readily available. The cost of upgrading
this CaSO^'2H20 would further increase the cost of this already expensive
process and therefore most of this CaSO^^I^O will probably be used as a
landfill material,
The wastewater purge stream containing a relatively high N0j~ concen-
tration will require some sort of treatment, either biological treatment
or evaporation to prevent secondary water pollution. Kawasaki has not
specified as yet what type of H20 treatment will be used.
Since this is a wet process, retrofitting this system on a conventional
coal-fired power plant will be relatively simple providing sufficient land is
available nearby for both the extensive process equipment requirements and a
wastewater treatment section. The major modification would be the removal
of the cold ESP and the installation of a common plenum to distribute the
flue gas from the boiler to the four scrubber trains. This design will also
allow the removal system to cycle with the boiler by simply closing off
various combinations of scrubber trains.
Most of the process equipment and piping will be elastomer lined to
prevent the corrosion and erosion problems associated with the circulation of
Mg slurries. The only exceptions to this elastomer lining are the flue gas
ducts, which will be made of Inconel 625, and the decomposition reactor
which will be made of stainless or glass-lined steel.
Environmental Considerations
During pilot-plant (1.7 MW equlv) testing on flue gas from a coal-fired
boiler containing 1000 ppm S02 and 200-400 ppm NO , the Kawasaki Mg process
was able to maintain 95% S02 removal and approximately 50% NOX removal at an
L/G ratio of 10 1/Nm3 (85 gal/kaft3 at 127°F) without the addition of 03. If
is added to the third stage absorber to oxidize NO to NO,, the NOX
removal can be increased to 80% removal at an OjsNO ratio of 1.3 and 90%
,,
f
removal at a ratio of 0:NO of 1.4.
The S02 absorbed from the flue gas is air-oxidized to S04= and removed
as CaS04'2H20. The equimolar mixture of NO and N02 is absorbed into the
slurry to form N02~ salts, converted to N03~ salts, and removed as Ga(NOg)2.
Unless this weak Ca(N03>2 solution can be sold as a low-grade fertilizer,
75
-------
further processing, either by evaporation to solid Ca(N<>3)2 or denitrif ication
by biological treatment, will be required to prevent secondary water pollution
from this stream.
Since this process does not use the S0£ absorbed from the flue gas to
reduce NOX, there is no theoretical minimum S02!NQX ratio in the flue gas as
in most of the other wet simultaneous removal processes.
The generation and use of Oj in this process presents a significant
hazard of workers in this plant. Qg is a very strong oxidizing agent and can
have significant effects on the health of workers even in relatively low
concentrations .
Critical Data Gaps and Poorly Understood Phenomena
Listed below are the major data gaps for the Kawasaki process.
1. Operating conditions for each piece of equipment
2. Stream compositions throughout the system
3. Estimated capital investment and revenue requirements
4. law material, energy, and operation requirements
5. Conversion In the various oxidlzers and reactors
6. Type and diagram of particular removal system for coal-fired
applications _ _
7. Type and diagram of wastewater j^CaCNOj^- \ / treatment section
The entire absorber section is not well understood at the present time.
The single absorber in the pilot plant is apparently internally split into
three sections with countercurrent gas-liquid flow, but when this pilot
plant is scaled up to a prototype or commercial unit three separate absorbers
connected in series may be required for each scrubber train. For the four
scrubber trains on a 500-MW plant this would mean a total of 12 absorbers the
size of a conventional Mg FGB scrubber. Further complicating this section is
that the total liquid effluent from the first two sections of the absorber
apparently passes through the regeneration section.
Advantages and^JDiaadvantagea
By virtue of the fact that this process is a wet simultaneous SC>2-NQX
process using modified gas-phase oxidation, it has certain advantages and
disadvantages, as listed below, over other types of flue gas denitrif ication
systems.
Advantagga
1. Achieves >95% S02 removal efficiency
2. Produces a potentially marketable byproduct (CaSO^^l^O)
3, Has been tested on coal-fired flue gas on pilot-plant or greater
scale
4. Operates with full particulate loadings (>7 gr/sft^)
76
A
-------
Disadvantageg
1, Forms secondary source of pollution (wastewater N0~~ salts)
2. Has not been operated for a long-term continuous period
3. Incorporates design features which may present significant process
control problems
4. Uses significant amounts of stainless steel or exotic materials for
process equipment
5. Has a high L/G ratio in the absorber (>70 gal/kaft3)
6. Requires flue gas reheat for plume buoyancy
7. Requires an expensive gas-phase oxidant
77
-------
KOBE STEEL PROCESS - WET, ABSORPTION-OXIDATION (NOX-SOX)
Process Description and Principles of Operation (5, 43)
In addition to their dry, SCR process, Kobe Steel, Ltd., has also been
developing a wet process to remove NO . Although very little information
has been published as yet, this NOX process is -a wet, absorption-oxidation
process; that is, the NQX is absorbed into the scrubbing solution as NO and
oxidized by Ca(ClO)2 to form NOg"* salts. The use of a liquid-phase oxidant,
however, requires two separate absorbers for desulfurization and denitrifiea-
tion to prevent the reaction of Ca(C10)2 with S02 and thus the excessive con-
sumption of this expensive oxidizing agent. , The S02 absorbed in the separate
FGD scrubber is oxidized to SO^™ and removed as byproduct CaSO^-IH^O.
The Kobe Steel CAL~NOX process is composed of six major sections including
(1) prescrubbing, (2) S02 absorption, (3) NOX absorption, (4) chlorine (Cl2)
absorption, (5) NOV~ decomposition, and (6) CaSO,^^!^ recovery. A more
detailed outline of this process can be seen in the block flow diagram in
Figure 17,
Although not mentioned by Kobe Steel the flue gas from a coal-fired
boiler will probably require a closed-loop prescrubbing section to remove
the increased particulates and Cl~ levels associated with coal combustion
before entering the absorption section. In addition to removing 90% of the
particulates and essentially all of the Cl"% the prescrubber adiabatically
humidifies and cools the flue gas from 150°C (300 F) to 53°C (127°F). The
scrubbing solution from the base of the prescrubber flows to a holding tank
from which most of the solution is recirculated through the prescrubber after
makeup H^O has been added. The remaining portion of the solution is removed
as a purge stream to prevent the buildup of Gl~" and ash and is probably
centrifuged to remove the ash from the Cl~ solution. This Cl~ solution would
then be mixed with the purge stream from the absorption section before final
treatment.
After passing through a mist eliminator, the flue gas passes counter-
currently to a CaCl2-lime slurry as in a conventional lime FGD system. The
S02 is readily absorbed and undergoes the following reactions.
S°2(g) * S°2(aq)
Ca(OH)2(aq) 4 S02(aq) * CaS03.l/2H20(g)* + 1/2H20 (90)
CaS03-l/2H20(8) + S02(aq) + 1/2H20* Ca(HS03)2(aq) (91)
(92)
The scrubbing solution flows to a holding tank from which most of the
liquid is recycled to the absorber. The remaining portion of the scrubbing
solution is pumped to the Cl2 absorber.
78
A
-------
r
BOILER
s
t
AIR
HEATER
t
AIR
i
I
1
1
1 i
PRESCRUBBER
C
H20
i
HOLDING
TANK
i
,
ASH
ENTRIFUG
i
ASH
Co(C
-. Dl IO
E
,
-
ABSORBEF
)H)g
r
HOLDING
TANK
i
1
GE
J
/
t
Co
DE
]
f
NOX
ABSORBER
(010)2
« 1
i
• n
HOLDING
TANK
1
> N2
t
NITRATE
[COMPOSE
*
C
• 1
CHLORINE
ABSORBER
I .
HOLDING
TANK
1 '
ir
ENTRIFUGI
FLUE GAS
1
GYPSUM
m
Figure 17. Flow Diagram of Kobe Steel Wet Process
-------
fhe flue gas from the FGD absorber is sent to the denitrification
absorber where it passes countercurrently to a Ca slurry containing
The NO is absorbed and undergoes the following liquid-phase reactions:
N0(g) * N0(aq) (93)
2NO,, . + Ca(ClO),, v + &„, x •*• Ca(NO.J,, , + C10, vt (94)
(aq) 2(aq; 2(aq; 3 2(aq) 2(g;
The liquid effluent from the absorber drops to a holding tank from which
most of the solution is recycled to the NOK absorber after fresh Ca(C10)2
solution has been added. The remaining liquid is removed as a purge stream
and sent to a decomposer where, it is claimed that, the N0-j= is chemically
decomposed and reduced to molecular N2- The decomposition method and reac-
tions have not been disclosed.
The "clean" flue gas from the NOX absorber is sent to a Cl2 absorber in
which the flue gas passes countercurrently to the purged scrubbing solution
from the FGD absorber. This solution, containing both a mixture of CaSOj-
and CaS04'2H20 and dissolved CaCl2» absorbs the Cl2 released when Ca(ClO)2
oxidizes the NO and converts it to CaCl2, i.e.,
C12(g) * C12(aq)
Ca(OH)2(s) 4- CaS03'l/2H20(-) + Cl2(aq) + 1/2H20 +
"2H2°(s)* (96)
The flue gas is passed through a mist eliminator, reheated, and sent to
the stack.
The liquid effluent from the Cl2 absorber drops to a holding tank and
most of it is recirculated back through the absorber. A small purge stream
is removed and pumped to a centrifuge where a byproduct CaSO^-2H20 cake is
removed. The centrate, consisting mainly of a CaCl2 solution, is split into
two streams: one is recycled to the FGD scrubber after makeup Ca(OH)2 is
added and the other is purged from the system to prevent the buildup of Cl~
salts in the circulating solution. The final treatment for this solution,
is any, has not been specified.
Status of Development
The Kobe Steel absorption-oxidation process has been tested on a pilot
plant treating 1000 Nm^/hr (0.3 MW equiv) of flue gas from an iron-ore sinter-
Ing plant. Detailed information concerning other operating parameters, such
as flue gas composition, denitrification efficiency, and length of continuous
operation, has not been revealed. A larger pilot plant to treat 50,000 Nm^/hr
(16.7 MW equiv) is under construction at Kobe Steel's Kakogawa Mill but no
further details have been published.
80
A
-------
This process has not been tested on flue gas from a coal-fired boiler.
Recently Kobe Steel has stopped the development of this process and
initiated studies into an entirely new wet denitrification process. However
no information has been published as yet (43),
Background of Process Developer
Kobe Steel is a large Japanese corporation heavily involved in the iron
and steel industry. With the coming of the strict air pollution regulations
in Japan in the early 1970's, Kobe Steel entered the air pollution control
field to solve the S(>2 emission problems from their plants. Their CAL-SOX
FGD system was developed and has been in commercial operation (50-120 MW equiv)
since April 1976,
In response to increasingly strict NOX regulations, Kobe Steel began the
development of both a dry SCR process using NH^ and a wet absorption-oxidation
process for NOX removal.
No American company has been licensed to market this technology in the
U.S. However, Kobe Steel does have a liaison office in New York City through
which inquiries about this process are handled.
Published Economic Data
Neither the estimated capital investment nor the annual revenue require-
ments has been published.
Raw Material_f Energy, and Operation Requirements
The major raw material requirements for this process are lime for makeup
slurry for the FGD absorber and Ca(C10)2 for the scrubbing solution in the
NOX absorber. No Information has been made available on the quantities of
lime and Ca(C10)2 required for this process.
The major energy requirements are fuel oil for reheating the flue gas
and electricity for operating the process equipment. No information has been
published on the quantities of energy required for this process.
The operating manpower and maintenance requirements have not been pub-
lished.
Technical Considerations
The Kobe Steel wet process appears to be a relatively complex process
with many pieces of process equipment including three absorbers and one
prescrubber per train in the absorption section. The use of the liquid-phase
oxldant would appear to be the most difficult process control problem but
this would not be as difficult as using a gas-phase oxidant since Ca(C10)2
is relatively stable and could be stored In bulk quantities.
81
-------
AS is typical for absorption-oxidation processes, under normal circum-
stances, the NOX removal efficiency is relatively insensitive to the inlet
gas composition. The NOX Is absorbed and chemically converted to N(>3~ with-
out reacting with 862. This conversion to N03~ salts does require dissolved
Q£ in the scrubbing solution but under normal boiler conditions enough 02 is
present in the flue gas for this reaction,
The desulfurization absorber is fed a lime slurry containing as its
major constituent dissolved CaC^. The operating conditions for this absorber
have not been published but they would probably be the same order of magnitude
as those in a conventional lime FGD system.
The operating parameters for the denitrif ication absorber and the Cl2
absorber have not been published and hence only general comments can be made
concerning their operation. Since NO Is relatively insoluble in aqueous
solutions, the NOX absorbers must be extremely large with a high L/G ratio
and low superficial gas velocity. The. Cl2 absorber would probably be much
smaller since Cl2 is relatively soluble in aqueous solution but no operating
parameters have been assumed.
The CaSO^-2H20 produced as a byproduct will contain relatively high con-
centrations of CaCl2 and CaSOg'l/Zl^O and hence the market value of this
CaS04'2H20 will probably be relatively low. The expense required to further
purify this CaSO^-Zl^O would,probably not be Justified and thus the CaSO^-Zl^O
would probably be used as a landfill material.
Since this is a wet process, retrofitting this system on existing power
plants should be relatively easy if sufficient land is available for siting
the extensive process equipment required. The major modification would be
the installation of a common plenum to connect the boilers to the four
scrubber trains. This use of a common plenum will also allow this system
to follow the boiler availability by simply closing off any combination of
scrubber trains.
Except for the flue gas ducts which will be made of Inconel 625, the
process equipment and piping for this system will be made of elastomer-lined
carbon steel to prevent the pitting corrosion associated with circulating
Cl~ solutions.
Environmental Considerations
No information is available on the sensitivity of either desulfurization
or denltrificatlon efficiencies to stoichiometry and operating conditions in
the absorption section. The desulfurization section is essentially the Kobe
Steel CAL-SOX process which in other pilot plants has demonstrated >90% S02
removal. The 862 is absorbed to form SO^*2, oxidized to S0^~, and removed as
CaSO^-2H20. The NOX is absorbed as NO, oxidized to form N03~, and chemically
decomposed to yield molecular NO.
82
A
-------
Critical Data Gapsand Poorly Understood Phenomena
Since only a minimal amount of information is available on the Kobe
Steel wet process, detailing all of the critical data gaps and poorly under-
stood phenomena would require an extensive list. Therefore, only the major
data gaps have been listed below.
1. Stream compositions
2. Equipment operating conditions
3. Wastewater treatment section
4. Section for chemical decomposition of £3(^3)2
5. Saw material, energy, and operation requirements
6. Capital investment and annual operating cost
7. NO absorber and Cl.2 absorber operating conditions
8. Pilot-plant results
9. Ca(ClO)2 section (onsite generation or bulk storage)
Since there are so many critical data gaps it is difficult to specify
any poorly understood phenomena other than the actual reactions involved in
the absorption of NOX. From previous studies the insolubility of NO, which
represents 90-95% of all NOX in the flue gas, has been overcome by either
oxidizing the NO to N02 before absorption or adding a complexing agent to the
scrubbing solution. Since this process oxidizes the NO after it has been
absorbed into solution and no solution additive is mentioned, the actual
mechanism of absorption is not clear.
Advantages and Disadvantages
The following list of advantages and disadvantages should be considered
as being tentative and subject to revision as more Information on the Kobe
Steel process becomes available.
Advantages
1. Produces a potentially marketable byproduct (CaSO
2. Operates with full particulate loadings (>7 gr/sft-*)
Disadvantages
1. Forms secondary source of pollution (wastewater N0j~ salts)
2. Has not been tested on coal-fired flue gas
3. Requires clean (S02~free) gas feed
4. Has a low superficial gas velocity in the absorber (< 10 ft/sec)
5. Has a high L/G ratio In the absorber (>70 gal/kaft^)
6. Requires flue gas reheat for plume buoyancy
7. Requires an expensive liquid-phase oxidant
83
-------
KUREHA PROCESS - WET, ABSORPTION-REDUCTION (NO -SO )
xx
Process Description and Principles of Operation (61, 75)
The Kureha Chemical Industry Company, Ltd., process is similar to other
absorption-reduction processes except that Kureha uses an Na2S03~acetic acid
solution to simultaneously remove both S&2 and NOX. The absorber for this
process is unique in that it la separated Into two compartments , with the
flue gas passing through each compartment. These compartments or stages
Include a simultaneous S02~NOX absorption section and an acetic acid removal
section.
The overall process can be divided into six major sections including
(1) prescrubbing, (2) absorption of both S02 and NOX, (3) acetate recovery,
(4) CaS04'2H20 production, (5) NH(S03Na)2 decomposition, and (6) ffl3 decom-
position. The basic outline of this absorber and the associated process
equipment la shown in the block flow diagram in Figure 18.
Although not mentioned by Kureha, the flue gas from the boiler air
heater probably enters a closed-loop prescrubbing section which is assumed
to remove 99% of the particulates and essentially all of the Cl~ from the
flue gas before entering the absorption section. In addition, the prescrubber
adlabatlcally humidifies and cools the flue gas from 150°C (30Q°F) to 53°C
(127 F) . The scrubbing solution drops into a holding tank from which most of
the solution is recirculated through the prescrubber after makeup 1^0 has
been added. The remainder of the solution is removed as a purge stream and
is pumped to a centrifuge to remove the flyash. Most of the centrate is
recycled to the prescrubber holding tank, while a small purge stream is
removed to prevent the buildup of Cl~ in the circulating solution.
The flue gas from the prescrubber enters the desulfurization section at
the bottom of a perforated plate absorber , As the gas passes countercurrently
to a sodium acetate (CHjCOONa) scrubbing solution, most of the 862 is rapidly
stripped from the flue gas and undergoes the following reactions.
S02(g) * S°2(aq)
2CH3COONa(aq) + S0 + H0 •+ NaS0 + 2CHCOOH (98)
Na2S03(aq) + S02(aq) + H2° * 2NaHS°3(aq)
In addition to these primary reactions, the Na2SC>3 also undergoes the
following secondary reactions.
Na2S03(aq) + 1/2°2(aq) * Na2S°4(aq)
The NOX, consisting primarily of NO, is relatively insoluble and is
absorbed over the entire length of the absorber. This NO absorption is
aided by an HoEJ-soluble catalyst which Is a proprietary Fe+2 chelating
compound. The NO is absorbed and undergoes the following overall reactions.
84
A
-------
1
BOILER
t
AIR
HEATER
1
AIR
00
CLEAN
FLUE GAS
AIR
Figure 18. Flow Diagram of Kureha Wet Process.
-------
(101)
2N02(g) -N° (102>
N2°4(g) + N2°4(aq) <103>
2NO, , + 5Na«S00/ , + 4CH0COOH, %-
(aq) 2 3(aq) 3 (aq)
2NH(S03Na)2(aq) + Na^^ -h 4CH3COONa(aq) + H20 (104)
2N2°4(aq) + 7Na2S°3(aq) + 4CH3CO°H(aq) *
2NH(SO,Na)0, . + SNa.SO. , % + 4CHQCOONa,, , + H.O (105)
3 2(aq) 2 4(aq) 3 (aq) 2
The liquid effluent flows to the combination absorber holding tank and
CaSO^*2H20 production reactor while the flue gas enters an acetic acid recov-
ery section. This recovery section has been added because as the scrubbing
solution passes through the S02~NOX absorber, some of the acetic acid is
vaporized. Thus to prevent the excessive loss of acetic acid from the system,
the exiting flue gas is passed countercurrently to an aqueous NH^ solution
in the final stage of the absorption section. The acetic acid is reabsorbed
and undergoes the following reaction.
CH-COOH, N + NH0/ N -f CIiLCOONH, / N (106)
3 (aq) 3(aq) 3 4(aq) v '
The flue gas is then passed through a mist eliminator, reheated, and sent to
the stack,
The liquid effluent from the acetic acid recovery section drops into a
holding tank from which most of the scrubbing solution is recycled after
makeup aqueous NH3 has been added. The remaining portion is pumped to the
production reactor for further processing.
In the CaSO^^t^O production reactor the absorbent is mixed with both
the purge from the acetic acid recovery section and a fresh limestone slurry
stream. The following reactions occur in the resulting mixture.
2CH3COOH(aq) + Ca(OH)2(aq) - (0^00) + 2 H0 (107)
, . 24fa . 2 -• , 3l(a , (108)
Most of the liquid from the combination SCU-NO^j absorber holding tank-
production reactor is recirculated through the absorber after
makeup acetic acid and catalyst have been added. A small purge stream from
this holding tank-reactor is removed and divided into two streams . The
smaller stream is sent to the NHj reactor while the remainder is pumped to a
centrifuge where the CaSO^^I^O crystals are removed as a byproduct stream.
The centra te from the CaSO^-2H20 centrifuge is split into two streams: one
86
A
-------
is recycled to the SC>2-NOX absorber and the other is pumped to a hydrolysis
reactor In the NHCSO^Na^ decomposition section. Makeup IL^SO^ is added to
this mixing reactor and the NH(S03Na)2 is hydrolyzed to NH^HSO^ and
by the following reaction.
M(S03Na)2(aq) + 2H20 ^ HH4HS04(a<|) + Na^^ (109)
The effluent from this hydrolysis reactor is recycled to the CaSO^-2H20 pro-
duction reactor.
In the NH3 reactor the small purge stream from the CaSO^^l^O production
reactor is mixed with a makeup lime slurry to neutralize the solution and
yield the following reaction.
NH.HSO. , . + Ca(OH)0, , •* NH-, , * + CaSQ,-21^0, ,4, (110)
k 4(aq) H '2(s) 3(g) 4 2 (s)Y
Any remaining ammonium acetate in this purge stream also reacts with the
lime slurry to yield NH^, CaSO/"2H20, and sodium acetate:
2CH3COONH4(aq) -f Na2S04(aq) + Ca(OH)2(g) -
(HI)
3, v 42, 3, ,
The liquid effluent from the NH3-Btripping unit is recycled to the
production reactor. The TOj off-gas from these reactions is stripped from
the solution by an injected air stream and sent to a catalytic reactor where
the NH3 is oxidized at 350-400°C (660-750°F) to form molecular N2 by the
following reaction.
4NH3(g) + 3°2(g) * 2N2(8) + 6H2°(g) (U2)
The off -gas, containing primarily molecular N2 and I^O, from this reactor is
recycled through the deacetating section of the absorber,
Status of Development
The Kureha simultaneous S02~NOX removal system was developed as a simple
modification of their original sodium acetate-CaSC»4 • 21^0 FGB process. Initial
bench-scale tests, treating 250 Nm^/hr (0.08 MW equiv) , were begun in 1973.
The simultaneous SC^-NOj, removal process was scaled up and operations begun
in early 1976 in a 5000 Nm^/hr (1.6 MW equiv) pilot plant at Kureha 's Nishiki
plant. The heavy oil-fired flue gas typically contained 1100-1700 ppm S02»
160-240 ppm NOX» 2-4£ 02, and 0.15 g/NmS of dust. During a 3000-hr (120-
day) continuous test, the system was able to demonstrate more than 99% S0£
removal and 80-90% NOX removal with negligible acetic acid emissions. Future
plans for this system have not been made available.
The Kureha system has not been tested on flue gas from a coal-fired
boiler.
87
-------
The next development step for this process would apparently be the
testing of a prototype unit in the 5-50-MW range. Also, this process's
ability to treat flue gas from a coal-fired boiler needs to be demonstrated.
Background of Process Developer
Kureha is a medium-sized Japanese chemical company with major interest
in fertilizers and petrochemicals as well as basic organic and inorganic
chemicals. Kureha entered the air pollution control field in 1967 with the
initial bench-scale testing of their ^2863 FGD system. After additional
pilot-plant work during 1968, several commercial plants based on this
FGD technology were completed during the early 1970 'a. This original Na
process was later modified in 1971 to Incorporate acetic acid and the new
system was renamed the sodium acetate-CaSO^-2H20 process. A bench-scale
unit (0,03 MW equiv) to demonstrate this improved FGD process was built and
began operating in 1973. Later a pilot plant using the sodium acetate-
CaSO^-2H20 system and treating 5000 Nnr/hr (1.6 MW equiv) of oil-fired flue
gas began operating in March 1975.
Initial investigation into the possibility of modifying the sodium
acetate-CaSO^'2H20 FGD system to simultaneously remove both S02 and NOX was
begun in late 1971. A bench-scale unit treating 250 Nnr/hr (0.08 MW equiv)
of oil-fired flue gas was completed in 1973 and continued operation until
mid-1974. A 5000 Nnr/hr (1.6 MW equiv) pilot plant was completed In early
1976 and is still operating.
Also, Kureha has recently begun the development of a dry SCR process
using NHj. The Initial research was finished in 1975 and a bench-scale unit
(0.1 MW equiv) began operating in early 1976. Scaleup to a pilot plant
(1.6 MW equiv) was completed in late 1976 and testing was scheduled to begin
in February 1977.
No American company has been licensed by Kureha to market this technology
in the U.S. However, Kureha has recently established a liaison office in New
York City through which inquiries about this process are handled.
Published Economic Data
The total capital investment and annual revenue requirements for the
Kureha wet reduction process have been estimated (75) as 9820 Myen and 4460
Myen/yr respectively. These estimates are based on treating flue gas from a
500-MW oil-fired boiler containing 1500 ppm S02, 200 ppm NOX, and 4% 02 with
a Japanese construction site and March 1977 costs. Assuming 300 yen/$ these
costs would be equivalent to approximately $65/kW for capital investment and
4.8 mills/kWh for the revenue requirements.
Raw tfaterjjjl^jjjnergyj T_and Operation Requirements
The following raw material, utility, and operation requirements are pre-
liminary estimates by Kureha and are based on a 500-MF oil-fired boiler treating
flue gas containing 1500 ppm S02, 200 ppm NOX, and 4% 02 (dry basis). The S02
88
-------
and NOX removal efficiencies were 99 and 871 respectively. Under these
conditions the raw material requirements would be: (All quantities in tons
are assumed to be metric tons.)
Raw material Quantity, kg/hr
Ca(OH)2 9,700
H2SC>4 3,900
Catalyst 100
Acetic acid 60
NaOH 60
The utility requirements would be:
Utility Quantity
Electricity .9,000 kW
Steam 14 tons/hr
H£0 66 tons/hr
Fuel oil 5.5 kl/hr
The only major byproduct under these assumptions would be 22.5 tons/hr of
The number of operators for the system would be two men per shift. The
maintenance requirements have not been made available.
Technical Considerations
Although the Kureha process does not appear to be as complex as the
other wet, absorption-reduction processes, it is still much more complex than
the dry, NOX removal processes. The process control requirements will be
similar to most of the other wet reduction processes with the major problem
being the large equipment requirements.
The NOX removal efficiency for the Kureha process will be sensitive to
both the 02 and the S02 concentration in the flue gas for the same reason
as the other wet, absorption-reduction processes. The catalyst for this
process is a Fe+^ chelating compound which forms a complex with the relatively
insoluble NO to aid the absorption process. As the 02 concentration in the
flue gas increases, more of the Fe"1"^ ion is oxidized to the Fe+^ ion by flue
gas 02 absorbed into the solution, and hence less of the chelating compound
in the Fe+2 form would be available to aid in the absorption of NO. In addi-
tion, the S02 in the flue gas, when it is absorbed in the scrubbing solution,
is used to reduce the absorbed NO , Thus, when the S02 levels are_low or the
02 levels in the flue gas are high, a larger proportion of the S03~ ion is
consumed by oxidation and less is available to reduce the absorbed NOX. The
S02 removal efficiency, on the other hand, would be relatively insensitive
to the flue gas concentration.
89
-------
'i'he packed-bed towers used to absorb the NO, even with the use of the
catalyst, will be much larger than the absorber required for their sodium
acetate~CaSO/»2H20 FGD system since the absorption mechanism for removing
the NO is similar to that in the other wet, absorption-reduction process.
The operating parameters for the 802-NOx absorber are an L/G of 20-30 1/Nnr*
(125-185 gal/kaft"* at 127°F) and a superficial gas velocity in the range of
1-3 ra/sec (3.3-10 ft/sec). The deacetating section of the absorber, which
contains perforated plates, operates at an L/G ratio of approximately 2.0
I/to3 (12.5 gal/kaft5 at 127°F).
The extreme differences in solubilities of S02 and NO leads to the
unusual situation where the S02 is rapidly stripped from the flue gas in
the lower portion of the absorber while the NO is gradually absorbed over
the entire length of the absorber. This large difference in solubility is
indirectly the primary reason for the excellent S02 removal efficiency and
the insensltivity of the S0£ removal efficiency to flue gas composition.
That is, in order to remove the NO, the absorber is overfieslgned for S02
removal. This large difference in the solubilities of S02 and NOX also
explains why such a relatively high concentration of the catalyst is required.
The catalyst is needed to form a complex with the absorbed NO and transport
the NOX to the lower portion of the absorber where there is a higher concen-
tration of the SOg" Ion available to reduce the NOX.
Retrofitting this system to a conventional coal-fired power plant
equipped with a cold ISP should be relatively easy if additional land is
available nearby. The land areas would be relatively large to accommodate
both the large equipment requirements in the regeneration section and the
limestone preparation section. The only modifications to the existing equip-
ment may be the removal of the cold ESP and the Installation of a common
plenum to feed the two scrubber trains.
The use of a common plenum would give the process a relatively good
turndown capability since a combination of the scrubber trains could be shut
down completely. The primary consideration during turndown will be flue gas
02 and S02 concentrations. In general when the boiler load is reduced, the
air is usually not reduced and hence with the higher 02 and lower S02 levels,
the NOX removal efficiency will decrease. The S02 removal efficiency should
not be affected during turndown.
Most of the equipment and piping for this process would be elastomer-
lined carbon steel to prevent corrosion and erosion problems associated with
the circulating solutions. The only exceptions to this lining would be the
flue gas ducts which will be constructed of Inconel 625 and the decomposition
reactor which will require glass-lined carbon steel,
Environmental Considerations
The S02 and NOX removal efficiencies for the Kureha simultaneous S02~NOX
removal system during pilot-plant testing on oil-fired flue gas were approxi-
mately 99 and 801, respectively, but the denitrlfication efficiency is depend-
ent on the inlet flue gas composition. For good NOX removal, i.e., 80%, a
minimum raol ratio of S02:NOX of about (2.5-3.0);! is required depending on the
90
A
-------
flue gas ©2 concentration. The higher the flue gas Cu levels the higher the
tnol ratio of SO?: NO must be for the same NO removal.
*• * x
The S02 ie absorbed as N32S03, oxidized to $32804, converted to
and removed as a byproduct. The NOX is absorbed, reduced to complex N-S com-
pounds, and decomposed to form molecular N2- Only a trace of N0^~ is expected
to be formed and hence will not present a significant waatewater problem.
The use of an NH^ solution to scrub the gas exiting from the absorber
removes essentially all of the acetic acid vaporized from the absorber
scrubbing solution, but the resulting possibility of NHj emissions could be
considered a secondary source of pollution.
The byproduct CaS0A*2H20 formed will probably contain too many impurities
to be sold in the U.S. and hence, could probably only be used as landfill
material,
Crltlca^ Datai Gaps an4 j?oorly^ JUgderstOjod Phenomena
The major data gaps for the Kureha simultaneous S02~NOX removal process
would include the following".
1 . Stream compositions
2. Pilot-plant results
Poorly understood phenomena would include primarily the actual reactions
involved in this process. Kureha has only listed some of the reactions
occurring in each phase of the process and many of these appear to be overall
reactions and as such, the basic mechanisms are not clear. This is partic-
ularly true in the NO absorption section where the most complex process
chemistry is occurring. . In addition, the conversion of the deacetating solu-
tion from a limestone slurry to an NHj solution, with its associated problems,
has not been explained.
Advantages and Disadvantages
The advantages and disadvantages for the Kureha sodium acetate-CaSO^-Zl^O
simultaneous St^-NQjj removal process are nearly identical with the advantages
and disadvantages of the other wet, absorption-reduction processes and are
listed below.
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >95% S02 removal efficiency
3. Produces a potentially marketable byproduct (CaSQ^^l^O)
4. Is a slight modification of a commercially available PGD system
5. Operates with full particulate loadings (>7 gr/sft^)
Disadvantages
1. Requires significant amounts of energy for the regeneration step
91
-------
2. Has not been tested on coal-fired flue gas
3. Incorporates questionable design features (final-stage NH3
scrubbing)
4. Uses significant amounts of stainless steel or exotic materials
for process equipment
5. Has a low superficial gas velocity in the absorber (< 10 ft/sec)
6. Has a high L/G ratio in the absorber (>70 gal/kaft3)
7. Requires flue gas reheat for plume buoyancy
8. Requires flue gas constituents within specific ranges for high NOX
removal
9. Requires an expensive liquid-phase catalyst
92
A
-------
MITSUBISHI HEAVY INDUSTRIES PROCESS - WET, OXIDATION-ABSORPTION-
REDUCTION (NOX-SOX)
Process Description and Pr in dp lea of Operation (5, 12, 21, 22)
Mitsubishi Heavy Industries (MHI) has developed a wet, oxidation-
absorption-reduction process for the simultaneous removal of S02 and NOX.
MHI uses Oj for the gas-phase oxidation of NO to N02, thereby overcoming the
problem of the relative insolubility of NO. Once the NOX is converted to
the more soluble N02> these oxides are absorbed, reduced by the SQ-f ion,
and recovered as NH3- The S02 is absorbed by the scrubbing solution to form
the S03= ion, oxidized to the S0^= ion, and removed as a byproduct CaSO^"2H20
stream.
The overall system is composed of five major sections including (1) pre-
serubbing, (2) gas-phase oxidation, (3) absorption of both S02 and NOX,
(4) byproduct CaSO^-2H20 production, and (5) N-S decomposition. The basic
outline for the MHI process is shown In the block flow diagram In Figure 19.
As In each of the other wet-scrubbing processes, the flue gas from the
air heater is first passed through a prescrubber which removes about 90% of
the particulates and essentially all of the Cl~ from the flue gas, and cools
the gas from 150°C (300°F) to 53°C (127°F) by adiabatic humldification. The
liquid effluent from the particulate scrubber drops into a holding tank from
which most of the liquid is recirculated through the scrubber after makeup
H^O has been added. Although not mentioned by MHI, it is assumed that the
remaining liquid in the holding tank is purged to either a centrifuge or
holding pond to remove the flyash and thus prevent the buildup of flyash in
the scrubbing solution.
After passing through a mist eliminator, the flue gas enters a duct
between the prescrubber and the absorber and, before entering the absorber,
is injected with a stolchiometric amount of 03 as a weak (< 5%) 03~air mixture.
This 03-air mixture is generated onslte using a battery of corona discharge
units, since 03 is a strong oxldant and is too unstable to be stored in
large quantities. The NO in the flue gas is rapidly and selectively oxidized
by 03 according to the following reaction.
N0(g)+°3 <113>
Proper positioning of the 03-air injection nozzles will eliminate
poor mixing of the gases and thus minimize the following undesirable side
reaction.
2N°2(g) + °3(g) * N2°5(g) + °2(g)
-------
VD
•C-
CLEAN
FLUE GAS
OZONE
^-GENERATOR
I LIMESTONE
PREPARATION
SECTION
ASH
CENTRIFUGE
THERMAL1
DECOMPOSER
SLURRY
PREPARATION
GYPSUM
Figure 19. Flow Diagram of Mitsubishi Heavy Industries Wet Process.
-------
The S(>2 is absorbed from the flue gas and undergoes the following
liquid-phase reactions .
S°2(g) + S°2(aq)
S°2(aq) * CaC03(8) + 1/2H2° " <**yi/ai20(j|)+ + 002(g)t (116)
S02(aq) + CaS03(aq) + H20 - Ca(HS03)2(aq) (117)
The NC>2, which is absorbed Into the scrubbing solution more slowly
than the SC^, undergoes the following reactions.
N2°4(g)
N2°4(g)
N2°4(aq) + Ca(OH)2(s) + CaS03'l/2H2°(s) + 1/2H2° *
Ca(N02)2(aq) + CaSV2H2°(s) (120)
Once both the S(>2 and N0£ have been absorbed, the N(>2~ formed by
reaction (120) are reduced by the HSOg" ion formed _in reaction^ (117) to
form complex N-S compounds, such as hydroxylamlne /NOH (803) £*_/ salts.
Ca(N02>2(aq) + 3Ca2(aq)*
(s)4' + H2° (12l)
These . complex N-S compounds are further reduced by the SO-j" ion in the
scrubbing solution to produce 1^1(803)2" salts by the following reaction.
Ca/NOH(S03)2_7( ,+
2_( , 42H204- (122)
In addition to these primary reactions, an Important secondary reaction
occurs In the absorber. Some of the SOg* ions formed in the scrubbing solu-
tion are oxidized to the SO^" ion by '02 absorbed from the flue gas, i.e.,
l/2H20(s) + l/20 + 3/2H0- CaSC2H0 + (123)
The flue gas from the absorber, after passing through a mist eliminator,
Is reheated for plume buoyancy and sent to the stack.
95
-------
The effluent slurry from the absorber, containing a mixture of dissolved
N-S compounds and insoluble CaSC>3*l/2H20 and CaSO^*2H20, drops into an
absorber holding tank. A small purge stream (10-20%) of the circulating
solution is removed from this holding tank and pumped to a neutralization
reactor in the regeneration section while the remaining effluent from this
holding tank is recirculated through the absorber after makeup limestone
slurry has been added. Much of the original limestone (or lime) in this
purge stream to the regeneration section has reacted with the absorbed SC>2
to form CaS03-l/2H20 which has further reacted with either 02 or the N0x
to form CaSO^-2H20. H2SC>4 is added to the neutralization reactor to convert
any remaining unreacted limestone to CaSO^'2H20 and, Indirectly, to_convert
any insoluble CaS03-l/2H20 to soluble calcium bisulfite £Ca(HS03)2_/ and
insoluble GaSO^Zi^O. These primary reactions occurring in the neutraliza-
tion reactor include:
Ca(OH)2(aq) + H2S04(aqr CaS°l>'2*2°(S)* <124)
.2H20(s)* + C02(g) + (125)
, , 2 , , 32, . (126)
This slurry from the neutralization reactor is pumped to a thickener
from which the bottoms, containing mainly CaSO* -21*20 are pumped to a centri-
fuge where the CaSO/-2H20 crystals are separated as a byproduct stream.
The mother liquor from the CaSO/-2H20 centrifuge is sent to the limestone
slurry preparation tank. As makeup limestone is added to this bisulfite
^/CaXHSOs) 2_/ rich liquor, the Ca(HS(>3)2 w*--^ ^e converted back into
CaS04'l/2H20 for recycle to the absorber.
( .
The thickener overflow, containing the H20 soluble N-S compounds, is
split into two streams: one for recycle to the limestone slurry preparation
tank and the other for further processing to remove these N-S compounds from
the system. This latter stream, which represents only 5-10% of the total
circulating solution, is first acidified with H2S04 until the pH is less
than 1 and then pumped to a thermal decomposer operating at 130-150 C
(265-300°F) . Under these reaction conditions, 95% of the N-S compounds are
hydrolyzed to NH^HSO^ during a residence time of 1-2 hr. These hydrolysis
reactions include:
0OH, % + 2KUSO, , .•*• N00, f + 2NH/HSO//, N + 3H00 (128)
2 (aq) 2 4(aq) 2 (g) 4 4(aq) 2
0OH, . + H.SO, / .-*• N_, . t+ NH.HSO., . + 3H»0
2 (aq) 2 4(aq) 2(g) 4 4(aq) 2
(129)
NH2S°3H(aq) + H2° -ra4HS0 (130)
96
A
-------
The NH4HS04 solution from the decomposer is pumped to a neutralization
reactor where this solution is mixed with a slaked lime slurry to yield
ammonium hydroxide (NH^OH) and
Ca(°H)2(S) + H2° -eaS04'2H20(s) + + SH4OH (131)
MHI has suggested three possible methods of treating this resulting
slurry to remove the NH40H from the system. These three hypothesized options
are:
1 . Increase the pH of the slurry until the NH^OH decomposes into NH-J and
then separate the gaseous NHj (10-15% by vol) from the N£ and ^0 gas
stream by an unspecified process,
2. Use high temperature to thermally decompose this NHj to molecular N2-
3. Reclaim the NH^OH in a stripping unit with makeup H20.
Once the NH^OH has been removed, the remaining GaSO^^^O slurry is
recycled to the limestone slurry preparation tank.
S tatus o f Development
Initial work on the MHI process for the simultaneous removal of S02 and
NOX was begun on a bench-scale unit in June 1974. This unit treated 200 NnrV
hr (0.06 MW equiv) of flue gas from a heavy oil-fired boiler until December
1975, Scaleup to a pilot plant treating 2000 Nm3/hr (0.6 MW equiv) of flue
gas from a heavy oil-fired boiler was completed in June 1975. During two
separate continuous month-long tests on this flue gas containing 800-1100
S02, 180-220 ppm N0%, 3-8% 02, and 0.09-0.12 g/Nm3 of dust, the MHI process
was able to maintain 95% S02 removal and 80-90% NOX removal. This process
has not been tested on flue gas from a coal-fired boiler.
At the present time plans are being formulated to construct a prototype
or commercial unit (33-67 MW equiv) but both the type of flue gas to be
treated and the site of the plant have not been specified.
Background of Process Developer
MHI has been Involved in the development of air pollution technology
since the completion of their first FGD system in 1972. Since then MHI has
completed approximately 50 desulfurization units with a total capacity of
almost 20 million Nm^/hr (6700 MW equiv).
The MHI wet simultaneous S(>2-NOX removal process, which has been under
development since 1974, is very similar to their commercial FGD system with
the only major modifications being the addition of an Og generation section
and an N1U recovery section.
No American company has been licensed by MHI to market their process
in the U.S., but MHI does have a liaison office in New York City, through
which inquiries about this process are handled.
97
-------
Published Economic Data
No estimates for either the total capital investment or the annual
revenue requirements for the MHI simultaneous S02~NOX process have been
published. The costs, however, for this process will probably be on the
same order of magnitude as the other Ca-based oxidation-absorption-reduction
processes.
Raw Materials, Energy,and Operation Requirements
The major raw materials required for the MHI process are limestone,
slaked lime, l^SO^, and Oo. Limestone and l^SO/ are common industrial raw
materials and readily available. The Oj, however, due to its extreme reac-
tivity will be generated onsite by corona discharge and hence is not a raw
material but an energy requirement.
The raw material requirements listed below were estimated by MHI for a
system treating 100,000 Nm^/hr (33 MW equiv) of flue gas from an oil-fired
boiler containing 1,500 ppm S02 and 150 ppm NOX. The S02 and NOX removal
efficiencies were assumed to be 95 and 90% respectively. (All quantities
measured in tons are assumed to be metric tons.)
Material Quantity
Limestone 0.58 ton/hr
Slaked lime 0.05 ton/hr
H2S04 0.08 ton/hr
Catalyst <0.8 kg/hr
Q
The energy requirements for treating 100,000 Nm /hr under the same
basis as above are:
Utility Quantity
Electricity 1750 kW
Industrial water 12.5 tons/hr
Steam 5.0 tons/hr
The operating manpower and maintenance requirements for this process
have not been published.
Technical Cons iderat ions
The MHI simultaneous S02~NOX process, using 0-j for the gas-phase oxtda~
tion of NO and a CaSO.,'1/2H20 scrubbing solution to absorb the resulting
N02» appears to require about the same amount and type of process equipment
as the other oxidation-absorption-reduction processes. As a result the
process control requirements for this process will be very similar to that
required for the other wet simultaneous FGD and denitrificatlon systems of
this type. The major process control problem will be the 03 generation
and injection system since the 03, once generated, is extremely reactive
and cannot be stored economically. If the 03 is injected in less than
98
A
-------
stoichlometric amounts, the NQX removal efficiency will decrease, and if it
is Injected in more than stoichiometric amounts, the revenue requirements
for this system would rapidly increase since the cost of 0^ is a major con-
stituent in the annual revenue requirements of this system.
The SC>2 removal efficiency is insensitive to the inlet flue gas compo-
sition since there are no interfering gas species and the absorber is essen-
tially an overdesigned FGD scrubber. With the higher L/G ratio in the absorber
required for denitrification, the SC>2 removal efficiency in this process is
expected to consistently remain above 90%.
The NC>2 removal efficiency for this process is relatively insensitive
to the inlet flue gas composition provided that the mol ratio of S02:N02
in the flue gas is >2,5, For most coal-fired boilers, the flue gas would
be expected to meet this requirement with the only possible exception being
flue gas from boilers fired with low-S western coal. This relatively high
mol ratio is necessary since the absorbed S02« the 803" ion, is needed to
reduce the absorbed NOX.
The absorber selected by MHI is a grid-packed tower in which the
CaSOj"1/2H20 slurry and the flue gas are passed cocurrently with each other.
Although the actual operating conditions in this absorber have not been
published, the absorber parameters are assumed to be similar to those in a
conventional limestone FGD system. The L/G ratio in the absorber would
probably be in the range of 7-10 1/Nm3 (44-62 gal/kaft3 at 127°F) or
slightly higher than a conventional limestone FGD system to accommodate
the increased difficulty in absorbing NOX. In addition, a proprietary 1^0
soluble catalyst has been Introduced into the scrubbing solution to aid in
the absorption of the NOX.
Although this process has not been tested on flue gas from a coal-
fired boiler, MHI does not expect any serious technical problems in con-
verting this system to coal-fired flue gas providing that a prescrubber
is inserted to remove the particulates and Cl~. The economics of operating
this system, however, will be drastically altered in that approximately
three times as much 63 will be required to oxidize the increased NOX levels
in the coal-fired flue gas. In addition, the electrical energy consumed
to generate the 63 will also increase sharply. In fact, in a previous
study (49) the 03 consumption alone is expected to require 9-10% of the
generating capacity of the boiler when the inlet NOX concentration is 600
ppm. Thus this process, as with the other oxidation-absorption-reduction
processes, would be very expensive unless combined with some type of boiler
modification system to reduce NOX levels below 600 ppm before reaching the
wet—scrubbing system.
As is common for the wet simultaneous removal processes, retrofitting
this system on existing power plants would be relatively easy if sufficient
land is available nearby for siting the various process equipment. The
major modification to the existing plant would be the installation of a
common plenum from the boilers or the ESP to feed the four scrubber trains
on a 500-MW boiler. The existing cold ESP would be removed and replaced
with the venturi-type prescrubbers to remove the Cl~ as well as the flyash
99
-------
from the flue gas, The use of a common plenum and four absorber trains will
allow this system to cycle with the boiler by simply closing off any combina-
tion of trains. The 63 generation section can also be easily turned down
since this section is made up of a battery of small 03 generators, any com-
bination of which can be shut down independently.
Most of the process equipment and piping for this process will need to
be elastomer-lined carbon steel to prevent the corrosion and erosion problems
associated with handling circulating limestone slurries and Cl~ solutions.
Exceptions to this elastomer lining include the flue gas ducts, the thermal
decomposer, and the neutralizing reactor. The flue gas ducts will be con-
structed of Inconel 625 to prevent both HoSO^ mist corrosion and the oxida-
tion associated with the use of Oj. The decomposer will require either
stainless or glass-lined steel due to the high temperature and acidity
required to hydrolyze the N-S compounds . The neutralizing reactor will
require stainless steel to prevent corrosion by the NHj atmosphere generated
in this reactor.
Environmental Considerations
During the initial 8 mo of operation in a small pilot-plant unit
treating 2000 Nm^/hr (0.6 MW equiv) of flue gas containing 1500 ppm SC^,
150 ppm NOX, 4.5% 02, and 0.1 g/Nm5 (0.04 gr/sft3 at 32°F) of dust, the
MHI process was able to maintain 90% NOX removal and better than 95% S02
removal. The S02 is absorbed, oxidized, and recovered as byproduct
A"2H20 while the NOX is absorbed, reduced, and recovered as a weak
(10-151) NH-j gas stream. At the present time MHI has not selected a method
of removing this NHj.
The NOX removal efficiency is, of course, dependent on the amount of
Oj injected. If the amount of 03 injected decreases below the stoichiometric
requirement, the NOX removal will drop below 90%. A complete failure of
the 03 generation section will result in only approximately 5% of the NOX
being removed. On the other hand a complete failure of the Oj generation
section will not affect the S02 removal efficiency and it will remain
above 95%.
The 03 used to oxidize the NO in the flue gas is a very strong oxidizing
agent and even at low concentrations presents a significant work hazard for
this process.
Critical Data Gaps and Poorly Understood Phenomena
Included are critical data gaps and the preliminary engineering
assumptions used to fill in these critical data gaps.
1. The L/G ratio and superficial gas velocity in the absorber (assumed
to be 44-62 gal/kaftj and 3-10 m/sec respectively)
2. The conversion of SOg™ to 804* in scrubber solution with flue gas
having 4% 02
3. The catalyst used for NOX reduction
100
A
-------
4, The operating conditions in the process equipment and stream
compositions
5. Maintenance and manpower requirements
6. Estimated total capital investment and annual revenue requirements
Included under poorly understood phenomena would be the sequence of
reactions involved in reducing the NOo^ salts to complex N-S compounds,
The series of reactions given by MHI Tsee reactions (121), (122), (128),
(129), and (130)_/ include apparent inconsistencies. For example, from other
processes calcium sulfamine ^psNBKSO^) 2_/ formed by reaction (122) given
below
Ca/NH(S03)2_7(aq) + CaS04-2H20(s) * (122)
Ca/IOH(S03)2_7/a ^
would more likely be calcium sulfamate ^Ca(NH2^°3)2 T> particularly since one
of the hydrolysis reactions given is the hydrolysis~~of the M^SOg" ion, i»e.,
NH2S03H + H20 + NH4HS04 (130)
Although not mentioned by the process developer, it is assumed that
at least some NO 3" salts would be formed when the NOX is absorbed. If
sufficient wastewater NOg" is formed the disposal of these salts could be
a problem.
Advantages and Disadvantages
As with the other wet simultaneous S02~NOX removal processes using gas-
phase oxidation, this process has certain advantages and disadvantages over
dry NOX processes and wet nonoxidation processes. These advantages and dis-
advantages are listed below.
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >95% S02 removal efficiency
3. Produces a potentially marketable byproduct (CaSO^'ZI^O)
4. Is a slight modification of a commercially available FGD system
5. Operates with full particulate loadings (>7 gr/sft-*)
1. Forms secondary source of pollution (15% NH3 off-gas)
2. Requires significant amounts of energy for the regeneration step
3. Has not been tested on coal-fired flue gas
4. Has not been operated for a long-term continuous period
5. Uses significant amounts of stainless steel or exotic materials for
process equipment
6. Requires flue gas reheat for plume buoyancy
7. Requires flue gas constituents within specific ranges for high NO
removal x
8. Requires an expensive gas-phase oxidant
101
-------
MITSUI ENGINEERING AND SHIPBUILDING PROCESS - WET, ABSORPTION-REDUCTION
Process Description and Principles of Operation (3, 5, 16)
This process Is In the early stages of development by Mitsui Engineering
and Shipbuilding Company, Ltd. , and very little information is available as
yet. It is a wet, absorption-reduction process which uses an Fe+^ chelate
solution to simultaneously absorb the 862 and the NOX from the flue gas. The
S02 Is absorbed and regenerated to form a concentrated S02 gas stream. The
NOX is similarly absorbed, regenerated, and reduced to molecular N£- The
Mitsui wet process consists of five parts including (1) prescrubbing, (2) an
NOX~S02 absorption section, (3) a reduction section, (4) a stripping section,
and (5) a stripped gas-treating section, A preliminary block flow diagram
for this process is shown in Figure 20.
Although not mentioned by Mitsui, the flue gas from the air heater is
assumed to pass through a closed-loop prescrubber section before entering
the absorber to remove the Increased amounts of particulates and Cl~
associated with coal combustion. In addition, some of the circulating solu-
tion is evaporated to adiabatically humidify and cool the flue gas from 150°C
(30Q°F) to 53°C (127°F) . The liquid effluent from the particulate scrubber
drops to a holding tank from which most of the liquid is reclrculated to the
prescrubber after fresh makeup H^O has been added. A small purge stream is
removed from the holding tank, neutralized, and pumped to a disposal pond.
The flue gas from the prescrubber is passed countercurrently to a
proprietary Fe^ chelate scrubbing solution. As the flue gas passes through
the lower portion of the absorber, the S02 is rapidly absorbed into the
solution and undergoes the following reactions with the oxidized Fe~*"2
chelate:
S02(g) + S°2(aq)
Fe+2.PCC(aq) + (> + 20 - 4Na (Fe.0H.PCC) (133)
Na_ (Fe+3.OH«PCC), % + S00, N -*• Na (Fe+3-PCC) -NaHSO_ , . (134)
2 (aq) 2(aq) 3(aq)
where PCC represents the proprietary chelatlng compound.
The NOX, mainly in the form of NO, is gradually absorbed over the entire
absorber and is hypothesized to react according to the following reactions:
N°(«) * N°(aq) (135)
N°(aq) + Na2(Fe+2-PCC) (-q) -> Ha^Fe^-PCC-NO) (aq) (136)
102
A
-------
r
REHEAT
SOg , NO I N0
CLEAN
FLUE GAS
Figure 20. Flow Diagram of Mitsui Engineering and Shipbuilding Co.
Wet Process
-------
The flue gas from the absorber Is passed through a mist eliminator, re-
heated, and exited through the stack.
The liquid effluent from the absorber drops to a holding tank from which
most of the solution is recirculated through the absorber. A small purge
stream is removed and pumped to a reduction reactor in the regenerator section
where the Ie+^ ion formed by reaction (133) is apparently reduced to the Fe"**
ion by electrolysis.
Na-de^-OH'PCC), „ + e" + H+ •+ Na, (Fe+2'PCC), , + H«0 (137)
2 (aq) 2 (aq) Z
Na(Fe"'"3-PCC) -NaHSO,, , -I- e~ + H+ •*• Na(Fe"l"2-PCC-H) -NaHSOq , % ,, ,„
3(aq) J(aq) (138)
The effluent from the reduction reactor is sent to a steam stripper where
an overhead stream containing the S02 and the NO is removed from the scrubbing
solution while the liquid solution is removed from the bottom of the stripper
and recycled to the absorber holding tank. Reactions occurring in the steam
stripper include;
Na0(Fe"l"2-NO*PCC)/ % * Na0 (Fe+ 2 • PCC) , . + NO, . t (139)
2 (aq) a 2 (aq) (g)
NaCFe^-PCOD'NaHSO-, , t Na-(Fe+2-PCC) , v -f H-0 -I- SO.,, ,f (140)
3(aq) A 2 (aq) 2 2(g)
The overhead gas stream is passed through a gas treatment section where
the NO is catalytically reduced to molecular N£ by the
2N°(8) + 2S°2(g) - N2(g) + 2S03(g)
The remaining gas stream containing primarily S07 can be used to produce
H2S04.
Status of Deyelopment
At the present time the Mitsui wet, absorption-reduction process is only
in the preliminary stages of development. A bench-scale unit treating 150
Nnr/hr (0.05 MW equlv) was recently completed and initial testing indicates
removal efficiencies of 95% for S02 and 85% for NOX. No other information
has been released.
This process has not been tested on flue gas from a coal-fired boiler.
Since this system has only been tested in a bench-scale unit, after
extensive testing to demonstrate the feasibility of the technology of this
reduction process, the next major development step would be a small pilot
plant in the 0.5-2.0 MW range.
104
A
-------
Background of Process Developer
Mitsui is a relatively large and diversified Japanese corporation which
entered the air pollution control field in the early 1970' s. In addition to
the development of their lime slurry FGD system, their early involvement in
air pollution control included a licensing agreement to build the Dowa
aluminum sulfate /A^CSO^^-CaSO^'lI^O FGD system.
In response to the adoption of stringent NOX emission controls in Japan,
Mitsui initially began the development of their dry SCR process using ^13,
This dry SCR process has reached the commercial stage of development.
Recently Mitsui began the development of this wet, absorption-reduction
process in addition to their original dry SCR process.
No American company has been licensed by Mitsui to market this process
in the U.S. However, Mitsui has a liaison office in New York City through
which inquiries about this process are handled.
Published Economic Data
No information on the economics of this process has been published.
Raw Material, Energy, and Operation Requirements
Essentially no information has been published for raw material, energy,
and operating requirements.
Technical ConsiderationB
Since little information is available, only very general statements can
be made about this process. For instance, it is a wet, absorption-reduction
process using an Fe* chelating compound in the scrubbing solution to absorb
NO. Many of the operating parameters and technical considerations in the
absorption section will be similar to those specified in the other absorption-
reduction processes. The regeneration section, however, is completely
different and hence no technical evaluation can be made now with the limited
amount of information available.
Environmental Considerations
Essentially no information is available to make any comments on the
potential environmental effects of this process.
Critical L_ Data Gaps and Poorly Understood Phenomena
Due to the lack of even general information concerning this process,
only the major data gaps will be listed.
1. Operating conditions in the process equipment
2. Stream flows and their composition
105
-------
3. Nature and description of the regeneration and byproduct treatment
section
4. Raw material, energy, and operating conditions
5. Economic data
6. Pilot-plant results
Advantages and Disadvantages
Although other advantages and disadvantages may become apparent after
additional data are released, the following list has been completed with the
information presently available.
Myanyagea
1. Removes NOX and S(>2 simultaneously
2. Achieves >95% S(>2 removal efficiency
3. Produces a potentially marketable byproduct (concentrated SQ2 gas)
4. Operates with full partieulate loadings (>7 gr/sft^)
Disadvantages
1. Requires significant amounts of energy for the regeneration step
2, Has not been tested on coal-fired flue gas
3. Has been tested only in a bench-scale unit
4. Has not been operated as an integrated process
5. Has not been operated for a long-term continuous period
6, Has a low superficial gas velocity in the absorber (<10 ft/sec)
7. Has a high L/G ratio in the absorber (>70 gal/kaft3)
8. Requires flue gas reheat for plume buoyancy
9. Requires an expensive liquid-phase catalyst
106
A
-------
MON ALKALI PERMANGANATE PROCESS - WET, ABSORPTION-OXIDATION (NOX-SOX)
Process Jtejcrlption and Principles of Operation (5)
The MON Alkali Permanganate process was Jointly developed by Mitsubishi
Metal Company, Mitsubishi Chemical Machinery (Mitsubishi Kakokl Kaisha) , and
Nippon Chemical Industrial Company. As is common with absorption-oxidation
processes, the desulfurization step takes place in a separate absorber before
the flue gas denltrification step and hence this process Is not a simultaneous
S02-NQ process. The Mon Alkali system uses a sodium hydroxide (NaOH)
scrubbing solution to remove the SOo in the FGD absorber and a potassium
hydroxide-based (KOH) scrubbing solution containing potassium permanganate
(IMn04) to absorb the NOX in the second absorber. The overall process
consists of six major sections (1) preserubblng, (2) S02 absorption, (3)
oxidation to 804* and formation of Na2SO^ in the FGD system, (4) NOX
absorption, (5) NOX oxidation with permanganate (Mn04~) , and (6) removal and
regeneration of the Mn04~ oxidant in the denitrification section. The basic
outline of this process Is shown in a block flow diagram in Figure 21.
Although not specifically mentioned, the process is expected to require
a closed-loop prescrubber section when treating flue gas from a coal-fired
boiler to remove the increased particulates and Cl~ associated with coal
combustion. In addition to removing 90% of the particulates and essentially
all of the Cl~ from the flue gas, the prescrubber also humidifies and cools
the gas from 150°C (300°F) to 53°C (127°F). The liquid effluent from the
prescrubber drops into a holding tank from which most of the solution is
reclrculated through the prescrubber after makeup I^O has been added. The
remaining solution from the holding tank is sent to an ash centrifuge where
the flyash is separated. Most of the centrate is recycled to the holding
tank but a small purge stream Is removed and pumped to a waste disposal pond
to prevent the buildup of Cl~ in the scrubbing solution.
The flue gas from the prescrubber is sent to a desulfurization scrubber
in which the gas passes countercurrently to an Na-based scrubbing solution.
The S02 is rapidly stripped from the solution and undergoes the following
reactions s
S°2(g) * S02(aq)
NaOH(aq) + S°2(aq) * Na^S0 (143)
Na2S°3(aq) + S°2(aq) + H2° * 2NaHS°3(aq) (144)
In addition to these primary reactions, an important secondary reaction occurs
in the absorber. Since the SOj" ion is present as the soluble Na salt, it is
easily oxidized by flue gas 02 absorbed into the solution, I.e.
107
-------
I
BOILER
t
AIR
HEATER
t
AIR
o
00
J^p
c
.
,
RESCRUBE
HgO
• i
ER
n
HOLDING
TANK
i
ASH
ENTRIFUG
1
ASH
E
L*
PL
i
^ma
JRG
>-»
r
E
E
CF
CE
i
sc
ABS(
taOH
I ,
r
>2
i
n
HOLDING
TANK
1
AIR
OXIDIZING
TOWER
t
AIR
\
HgO
4
VAPORATOf
YSTALLIZE
1
INTRIFUGE
?
R
I
K
C
E
|
N
ABSC
:OH
Ox
FLUE GAS
f
HOLDING
TANK
i
i
MnOg
ENTRIFUG E
- 1 <
KOH
4
KMn04
: ^ REGENERATOR — **
H20
4
VAPORATOF
~L
?
IMa2S04
Figure 21. Flow Diagram of MON Alkali Permanganate Process.
-------
°2(g)*02t«,)
Sa2S03(a,) + 1/2°2(aq) * " (146)
The scrubbing solution from the absorber drops Into a holding tank from
which most of the solution is recycled to the absorber after makeup NaOH has
been added. A small purge stream is removed from the holding tank and sent
to a regeneration section to remove the absorbed S02 as
Since essentially no Information Is available on the regeneration system,
the following has been assumed. The purged solution is probably pumped to an
oxidizing tower to convert the remaining S03= to S04= by reaction (146) .
Since Na2SOA is very soluble, the effluent from the oxidlzer will have to be
partially evaporated, cooled iii a crystallizer , and centrlfuged to remove
the byproduct Na2SO^. The mother liquor separated In the centrifuge is
recycled to the absorber holding tank,
The flue gas from the FGD section is then passed counter currently to a
KOH solution containing KMnQ4 in the second absorber. The NOX in the flue
gas is gradually absorbed over the length of the absorber and once absorbed
is readily oxidized by the Mn04~ ion to fora N03~ salts. This reaction
sequence is hypothesized to be as follows:
NO, s ^ NO, % (147)
(g) (aq)
NO, , + KMnO,, N ->• KNO,., , + Mn00, ,4- (148)
(aq) 4(aq) 3(aq) 2(s)
The flue gas from the NOX scrubber is reheated and sent to the stack
while the scrubbing solution drops into a holding tank. Most of this
solution is recycled to the absorber after makeup KOH and recycle KMnQ^ have
been added. A small purge stream is removed from the holding tank and sent
to the regeneration section where the manganese dioxide (Mn02) Is separated
in a centrifuge, converted into Mn04"~ by electrolysis, and recycled to the
absorber. The centrate containing the potassium nitrate (KNOg) is to be
marketed as a liquid fertilizer although it will probably be necessary to
concentrate this solution by evaporation before It is .sold.
Status of Development
Initial testing of the MON Alkali Permanganate process was begun in late
1972 in a bench-scale unit treating 300 Nm3/hr (0.1 MW equiv) of flue gas
from an oil-fired boiler. After approximately 2 yr of testing, a pilot
plant treating 4000 Nm^/hr (1.3 MW equiv) was completed. No information has
been made available concerning the results of these pilot-plant tests.
This process has not been tested on coal-fired flue gas.
The process developers theorize that the most promising use of this
process Is for treating off-gases from steel mills rather than power plant
109
-------
stack gases. The economics of this process for treating flue gas containing
large amounts of S02 are not competitive when compared with other simultaneous
S02~NOX processes. The high coat of this process under these conditions is
primarily due to the excessive consumption of
Since so little information about the results of the pilot-plant testing
Is available, it is difficult to project the next stage of development for
this process.
Background of Process Developers
The MON Alkali Permanganate process was developed jointly by Mitsubishi
Metal, Mitsubishi KK, and Nippon to convert the NOX in HN03 tall gas to KN03.
No Information is available on the background of Mitsubishi Metal but
Mitsubishi KK is a well-known Japanese manufacturing company. Nippon is a
Japanese chemical concern which produces alkali permanganates.
Mitsubishi Metal and Nippon apparently have not had any previous
experience in air pollution control. However, Mitsubishi KK has also been
involved in the development of a dry NH-j-based SCR process which is discussed
in more detail under the Mitsubishi HK process in the section, Dry NOX
Removal Processes.
Published Economic Data
No information has been published concerning the total capital in-
vestment or the annual revenue requirements for treating power plant stack
gas using the MON Alkali Permanganate process. However, for treating
200,000 Nm3/hr of off -gas from a steel mill containing 100 ppm S02 and
200 ppm NOX with 75% NOX removal and 100% S02 removal, the total capital
investment has been estimated (113) ae $5 million while the revenue require-
ments were estimated as $13-15 /kl of heavy oil. These costs are assumed
to be based on a Japanese construction site and mid-1974 dollars.
Raw Material f Energy , and Operation Requirements
No information has been published concerning the raw material, energy,
and operation requirements for this process. The major raw material would
probably be KOH and NaOH with lesser amounts of makeup limestone and KMnQ,^.
The major energy requirements would be fuel oil, for both flue gas reheat
and to evaporate the byproduct salt solutions, and electrical energy for
both regenerating the KMnO^ and operating the processing equipment.
Technical Considerations
Since very little information is available on the MON Alkali Permanganate
process, a detailed technical assessment of this process is impossible and
only general statements can be made.
As in the case of the other wet, absorption-oxidation processes, the
use of a liquid-phase oxidant, such as the IMn04 used in this process,
110
A
-------
brings about two major problems when applied to power plant flue gas.
1. Since S02 is also readily oxidized by the liquid-phase oxidant,
the S02 must be removed in a separate absorber to prevent excessive
consumption of the liquid-phase oxidant.
2. The NOjj absorber must be very large to remove the relatively in-
soluble NO from the flue gas.
Thus, from a technical viewpoint, this process is not a "simultaneous" S02~
NOX removal system but simply a combination of separate desulfurization and
denitriflcatlon systems.
The NOX and S02 removal efficiency of the MON Alkali Permanganate system
should be relatively Insensitive to the inlet flue gas composition. The
operating costs for this MnO/~ process, however, will be influenced by the
inlet flue gas composition since the amount of KMnC>4 required will be
proportional to both the inlet NOX and SC>2 concentrations.
The byproducts formed In this process, (i.e., Na2S04 and KNOj) although
potentially marketable, will have a difficult time displacing competing
natural sources. In particular, KNC>3 as a weak aqueous solution would have
very little demand as a liquid fertilizer.
Since this is a wet simultaneous process, retrofitting this system on
existing power plants should be relatively easy if sufficient land is
available nearby for the siting of the process equipment and holding ponds.
The major modification would be removing the existing cold ESP and in-
stalling a common plenum to distribute the flue gas to the scrubber trains.
The use of this common plenum and four scrubber trains will allow this system
to be easily turned down by simply closing off any combination of scrubber
trains .
The materials of construction required for this process have not been
specified.
Environmental Cong Ide rat ions
Essentially no information has been published listing the NOX removal
efficiency and the sensitivity of this removal efficiency to changes in the
flue gas composition and operating conditions in the absorbers.
The generation of N0-j~ salts in this process represents a significant
waste disposal problem since these extremely soluble N0g~ would be considered
a secondary pollutant. The treatment options for the wastewater stream
would be either desiccation and disposal in a landfill or some form of
biological wastewater treatment. However, neither of these options would be
economically attractive,
111
-------
Critical Data Gaps and Poorly UnderstoodPhenomena
There Is insufficient information available to discuss the critical data
gaps in detail or discuss the poorly understood phenomena and hence only the
following major critical data gaps are listed.
1. Equipment operating conditions and sizes
2. MtiO^"" regeneration section
3. Stream flow and compositions
4. Estimated capital and operating costs
5. Raw material, energy, and operation requirements
6. Pilot-plant data
Advantages and Disadvantages
Although sufficient information has not been published to enable a
complete listing of the technical advantages and disadvantages for this
process, the following advantages and disadvantages are apparent at the
present time.
Advantages
1. Achieves >90% NOX removal efficiency
2. Produces a potentially marketable byproduct (Na2SO/)
3. Operates with full partlculate loadings (>7 gr/sft*)
Disadvantages
1. Forms secondary source of pollution (wastewater NOo" salts)
2. Requires significant amounts of energy for the regeneration step
3. Has not been tested on coal-fired flue gas
4. Uses significant amounts of stainless steel or exotic materials
for process equipment
5. Requires flue gas reheat for plume buoyancy
6t Requires an expensive liquid-phase oxidant
112
A
-------
MORETANA CALCIUM PROCESS - WET, OXIDATION-ABSORPTION-REDUCTION (NOX-SOX)
Pro cess Des c r ip tion and Pr incIgle s of Operat ion (17, 56)
This process was jointly developed by Fuji Kasui Engineering Company
and Sumitomo Metal Industries as a modification to their original Na-based
simultaneous S02~NOX removal system. The process description and operation
is very similar to the original process with the major modification being
the replacement of Na+ ions by Ca"*"* ions and thus using lime or limestone as
a raw material rather than expensive NaOH. This process is also a wet,
oxidation-absorption-reduction system using C102 for the gas-phase oxidation
of NO to N02i Unfortunately, this gas~phase oxidation also results in the
formation of undesirable N0a~ and Cl salts in the scrubbing solution.
The overall process consists of five basic steps Including (1) pre-
scrubbing, (2) gas-phase oxidation, (3) simultaneous S02~NOX absorption,
(4) byproduct sludge removal, and (5) byproduct N03~ removal, A block flow
diagram detailing these various steps for the Moretana Calcium process is
shown in Figure 22.
The flue gas from the air heater passes through a prescrubber counter-
currently to an H20~based scrubbing solution which is assumed to remove 90X
of the particulates and essentially all of the HCl in the flue gas. As the
flue gas passes through the prescrubber it is also adiabatically cooled
from 150°C (300°?) to 53°C (127°F) and humidified by the evaporation of H20
from the scrubbing solution. The liquid effluent from the prescrubber drops
to a holding tank from which most of the solution is recirculated through
the prescrubber after makeup 1^0 has been added. The remaining liquid is
purged and pumped to a centrifuge where the flyash is removed. This flyash
is sent to a waste disposal area, while the centrate is recycled to the pre-
scrubber holding tank.
The flue gas, after passing through a mist eliminator, enters the flue
gas duct and is injected with CK^- This C102 is generated onsite and is
Injected at a rate of 10% in excess of the stoichiometric requirement. With
the proper design of the injection system, the C102 selectively oxidizes the
NO by the following reaction.
2NO, . -t- CIO., . + H00> + N00/ . + HNO-, , + HCl, % (149)
(g) 2(g) 2 (g) 2(g) 3(g) (g)
This gas-phase reaction occurs very rapidly and completely oxidizes the
NO to N02 and HN03.
The oxidized flue gas is then passed countercurrently to a limestone
slurry In a "Moretana" plate tower absorber. The limestone slurry at a pH
of approximately 5.5 is a mixture of Ca salts, mainly CaS03 and CaC03, and
a catalyst to aid in the reduction of NOX. As this limestone slurry passes
through the absorber, the S02 is stripped from the flue gas and the following
reactions occur in the scrubbing solution.
113
-------
r
BOILER
t
AIR
HEATER
t
AIR
PRESCRUBBER I »ABSORBER
CLEAN
FLUE GAS
LIMESTONE
PREPARATION'
ASH
CENTRIFUGE-
I f
EVAPORATOR
CaCO3
/CaCI2{s)
Ca(N03)2
•CaC03
NH4CI (aq)
NH4N03(oq)
Figure 22. Flow Diagram of Moretana Calcium Process.
-------
S°2(aq) (150)
CaC03(g) + 80+ eaS0 + C0+ (151)
CaS03(aq) + S02(aq) + H2° " Ca(HS03)2(aq)
At the same time the gases formed from the oxidation of NO are also
absorbed from the flue gas stream and undergo the following hypothesized
reactions.
N2°4(8) (153)
00, , , •* N00. , , (154)
2 4(g) 2 4(aq)
2N2°4(aq) + 2CaS°3(aq) + 2CaC°3(s) + 2H2°
aq B (155)
Ca(HS03)2(aq)
HN°3(aq) (157)
C02(g) , (158)
HC1, N •*• HC1, , (159)
(g) (aq)
2HC1/ N + CaC00/ v •+• CaCl™, , + CO-, ^ + H«0 (160)
(aq) 3(s) 2(aq) 2(g) 2 v
Secondary reactions also occur in the absorber; for example, 02
from the flue gas is absorbed into the scrubbing solution and causes the
oxidation of CaS03 to
Q/ N + 1/200<, , + 2H00 •*• CaSO. -2H00,, > (161)
3(aq) 2(aq) 2 4 2 (s)
However, this oxidation of the SO-~ ion in the Ca system is not as
serious a problem as in the Na-based systems, since CaS03 is relatively
insoluble in the scrubbing solution. Hence, most of the CaSO-j is not
present in the solution where it can be oxidized. After passing through a
mist eliminator, the cleaned flue gas is reheated and exited through the
stack.
The liquid effluent from the absorber drops into a holding tank from
which most of the slurry is recirculated through the absorber. The remaining
portion of the slurry is pumped to a centrifuge where the relatively insoluble
CaS03 and CaS04'2H20 are separated from the soluble CaCl2 and €3(1103)2. The
115
-------
solids removed in the effluent slurry centrifuge (the CaSOg and CaSO/*2H20)
are sent to a byproduct disposal area and dumped. The centrate containing
the soluble Ca salts is split into two streams, the largest of which is
pumped to the limestone slurrying section to transport makeup limestone and
makeup catalyst to the absorber holding tank while the second stream is purged
to a byproduct production section.
This mixed solution of CaCl2 and Ca(N(>3)2 removed to the byproduct
section can be treated by either simply drying the_solution _to produce CaCl2
or reacting the solution with ammonium carbonate ,/,(NH4)2C03_/ or (NH^^SO^
to produce a liquid fertilizer. The production of liquid fertilizer involves
pumping the Ca salt solution to a double decomposition reactor and mixing
with the salt solution where the following reactions occur.
CaCl0, x + (NH.KCO., . •*• 2NH/C1, , + CaCO., A (162)
2(aq) ^ 42 3(aq) 4 (aq) 3(s)
Ca(NO,),., , + (NH,),CO,, . + 2NILNO,,, , + CaCO,, ^ (163)
3 2(aq) 4' 2 3(aq) 4 3(aq) 3(s)
This liquid fertilizer may then be sold as is or further processed by
evaporation to be sold as a solid fertilizer.
Status of Development
Since this is a modified version of Fuji Kasui-Sumitomo's original
Na-based wet, simultaneous S02~NOX removal process, this process has not
been as extensively tested as the original Na-based system. Testing of
this Ca-based system was begun in 1975 in a bench-scale unit treating 500
Km /hr (0.16 MW equiv) of flue gas from an oil-fired boiler. After this
initial evaluation on a bench scale, testing has recently (1976) begun on
a 25,000 Nnr/hr (8.3 MW equiv) prototype unit treating flue gas from a
sintering furnace containing 400-850 ppm S02, 180-250 ppm NOX, 16-20% 02,
and 1-3 g/Nm of dust. In addition, bench-scale testing on synthetic
coal-fired flue gas was conducted after testing on sintering furnace flue
gas was initiated.
Although this process has not been tested on flue gas from a coal-
fired boiler, the developers believed that the use of a Moretana plate
tower as a prescrubber to remove the increased particulates and Cl~ assoc-
iated with coal combustion before these pollutants reach the absorber
would enable this process to treat flue gas from a coal-fired boiler.
The next stage of development would be either a pilot-plant or small
prototype unit to test the ability of this system to handle flue gas from
either an oil- or coal-fired boiler.
Background of Process Developers
In the past 20 yr Fuji Kasui has been active in the development of
pollution control equipment and processes. In the early 1970's Fuji Kasui
undertook the development of processes to remove S02 and NOX from combustion
exhaust gas and later began Joint research with Sumitomo for development of
116
A
-------
processes using the Moretana absorption tower. This Involvement began with
the design and engineering of the Moretana absorption tower and the later use
of this absorption tower both for a particulate removal and FGD applications.
One of these early applications was for the desulfurization of heavy oil-
fired flue gas at a Sumitomo plant in 1971. In May 1973 a prototype plant
(20 MW equiv) was started up using a CaO-scrubbing process for desulfurization
and a new test plant for simultaneous 862 and NOX removal was begun.
At the present time 12 desulfurization and 1 combination desulfurization-
denitrification plants (10-40 MW equiv) are already operating (although only
2 of these S02~NOX processes use the Ca-based system) and hence at the present
time Fuji Kasui-Sumitomo probably has more experience in flue gas denltrifi-
cation than any other company.
No American company has been licensed by Fuji Kasui-Sumitomo to market
their process in the U.S. However, Sumitomo does have a liaison office in
New York City through which inquiries about this process are handled.
Published Economic Data
The following estimated costs for the Moretana Calcium process are
based on treating 500,000 Nnr/hr of flue gas from a heavy oil-fired boiler
containing 1,500 ppm SC-2, 200 ppm NOX, 4% (>2, and 0.1 g/Nm3 of dust and a
desulfurization and denitrification efficiency of 99 and 90% respectively.
The total capital investment and the revenue requirements (assumed basis:
Japan and 1976 costs) for the above-mentioned plant are estimated (17)
by the developers to be 2500 million yen and 7009 yen/kl of oil res-
pectively. If 300 yen/$ and 3000 Nm /hr/MW are assumed, this corresponds
to an estimated capital investment of $50/kW of capacity and an estimated
revenue requirement of 5.78 mills/kWh.
Raw Materials, Energy, and Operation Requirements
Other than the crushed limestone for the absorbing solution which is
readily available, the major raw materials required are H2S04 and the
chemicals required to generate C102 (these chemicals are considered propri-
etary). In addition, lesser amounts of makeup catalyst will also be needed.
The quantities of these raw materials have not been published.
The energy requirements such as fuel oil for flue gas reheating, steam,
H20, and electricity have also not been made available.
The operating manpower and maintenance requirements for this process
have not been published.
Technical Considerations
The Moretana Calcium process uses C102 for the gas-phase oxidation
of NO and a CaS03 scrubbing solution, containing a dissolved catalyst to
aid in the absorption of the resulting NC<2. Since C102 cannot be stored
but must be generated onsite, the major process control problem will be the
control of the ClC>2 generation and injection system. If C102 is injected
117
-------
in less than stoichiometric amounts, the NOX removal efficiency will decline
proportionally and if it is injected in more than stoichiometric amounts,
excess C10£ will contribute excess Cl~ ions to the scrubbing solution and
will also increase operating expenses since the generation of C102 is one of
the major costs.
The use of C102 as the gas-phase oxidant also complicates the process
control scheme for another reason. Since C102 is generated by chemical reac-
tion and requires significant time periods (15-30 min) to adjust to changing
NOX levels in the flue gas, this process will have difficulty cycling with
changing flue gas NOX concentrations,
As is typical for most of the oxidatlon-absorption-redxiction processes,
the NOX removal efficiency is relatively insensitive to the inlet gas compo-
sition. In the case of the Moretana Calcium process this insensitivity is
due to several factors, the most important of which is the relative insolu-
bility of the CaSC>3 in the scrubbing solution. This insolubility prevents
most of the SOj™, formed by the absorption of S0£ from the flue gas, from
being oxidized to S0^~ by flue gas Q£ absorbed into the scrubbing solution
and hence more of the S0^= ion is available for reducing the NOX, A second
major reason for this insensitivity to flue gas composition is that only a
portion of the absorbed NOX is reduced by the SO^" to molecular N2« Thus
the relative insensitivity of this process to the flue gas composition is
due to the use of insoluble CaSOj which decreases the loss of the SQy* ion
by oxidation and also because of the small amount of SO^" required since
only part of the absorbed NOX is reduced to molecular N£«
Although the use of relatively insoluble CaSC>3 minimizes the oxidation
of the SO-j* ion in the scrubbing solution, this insolubility of the S0-j~
ioa necessitates the use of a dissolved catalyst to aid in the absorption
of the NOX. Apparently for proprietary reasons, the process developers have
not published detailed information concerning this catalyst. However, it is
probably a mixture of copper (Cu) ions and other proprietary additives. A
similar mixture is used in some of the other oxidation—absorption-reduction
processes and the mechanism by which it aids NOX removal is not well understood.
The Moretana plate tower is used for both the prescrubber and the
absorber in this process. The Moretana plate tower is very similar to a
sieve tray absorber except that it has no weirs or downcomers as can be seen
in Figure 23. As the liquid and gas pass countercurrently to each other in
the tower, there is an alternating flow of first liquid and then gas through
the perforations in each plate. As the height of liquid on a. plate increases,
the weight of the liquid overcomes the force of the gas pushing through the
perforations and the liquid displaces the gas and flows through the first
plate to the lower plate. As the liquid height on the first plate decreases,
the gas pressure forces the displacement of the liquid and the gas flows
through the perforations in the plate. This alternating pulsing of first
liquid and then gas through the plate gives the gas-liquid contact required
for simultaneous SQ2-N02 absorption and gives the Moretana plate tower
advantages over other types of absorbers. The developers claim this design
allows the Moretana plate tower to handle higher superficial gas velocities
which reduces the size of the absorber required in a particular situation.
118
A
-------
CLEAN
FLUE GAS
INLET
FLUE GAS
-"—
RECIRCULATING
CALCIUM SLURRY
SAMPLE PLATE
SCRUBBING SLURRY
TO HOLDING TANK
Figure 23. "Moretana" Plate Tower (110).
119
-------
This constantly and violently moving gas-liquid contact area Is also claimed
by the developers to prevent both the formation and the buildup of scale
Inside the tower and also the plugging associated with conventional FGD
systems using plate towers.
The operating parameters for the Moretana plate tower used as the
absorber In the Ca-based process are similar to those in their conventional
limestone FGD system, i.e., an absorber L/G ratio of 3-7 1/Nm3 (20-45 gal/
kaft at 127 F), and the superficial gas velocity of about 4 m/sec (13 ft/sec).
Although this process has not been tested on flue gas from a coal-fired
boiler, the developers do not expect any serious problems in converting this
system to handle the increased particulates and Cl~ associated with coal
combustion. The major reason for this confidence is the wide experience in
applying their FGD system, which is similar to their simultaneous S02~NOX
removal system. Their desulfurization system has been tested under various
operating conditions ranging from heavy oil-fired furnaces (1,400 ppm S02
and 0.2 g/Nitr5 dust), to a Rhodine salt incinerator (28,900 ppm 862 and 0.5
g/Nm3 dust), to a refuse incinerator (1-3 g/Nm^ dust). However, when treating
flue gas from a coal-fired boiler containing 600 ppm NOX in the flue gas,
the consumption of C102 would increase proportionately. This would cause a
significant change in the economics of this process since the cost of gen-
erating C102 is a substantial part of the annual revenue requirements for
this process.
Apparently at this stage in the development of Ca-based scrubbing
process, it has not been decided whether removing the Cl~ and NO^" salts
as a solid salt mixture or converting them to an NH3~based liquid fertilizer
is the best method of disposal of these salts. Since neither of these
byproducts has a large market demand in the U.S. and the manufacturing of
liquid fertilizer byproduct consumes (NH4)2C03, simply drying these salts
and placing in a landfill probably would be the best alternative.
The CaS04*2H20 sludge produced as a byproduct will contain numerous
salt Impurities including S031=, bisulfates (HS04~), Cl~, and NOj". Since
CaSO/-2H20 which does not contain these impurities is readily available In
the U.S., any CaSO^-2H20 formed will probably be used as landfill material
rather than being further processed for sale as a byproduct.
Since this is a wet simultaneous removal system, retrofitting existing
power plants should be relatively easy if sufficient land is available
nearby for the siting of the process equipment. The major modification
would be the installation of a common plenum from the boilers or the ESP
to the four scrubber trains on a 500-MW boiler. The existing cold ESP would
probably be removed and replaced with Moretana plate towers acting as pre-
scrubbers to remove the Cl~ as well as the flyash.
The use of a common plenum and four absorber trains will allow this
system to be easily turned down from 100% to 75, 50, or 25% by simply
closing off any combination of scrubber trains.
120
A
-------
Since the use of C102 as an oxidizing agent will inject relatively high
(2000 ppm) concentration of Cl~ ions in the scrubbing solution, most of the
process equipment and piping will need to be elastomer-lined carbon steel to
prevent pitting corrosion associated with circulating Cl~ solutions. Excep-
tions to this elastomer lining of process equipment and piping include the
inlet flue gas ducts which will be constructed of Inconel 625 to prevent
corrosion by the flue gas, and the absorber and absorber pumps which will be
made from 316L stainless steel.
Environmental Considerations
During bench-scale testing of this process on a 500 NmrVhr unit (0.16
MW equiv) treating flue gas from a heavy oil-fired boiler (1500 ppm S02» 200
ppm NOX» 4% Q£» 0.1 g/Nur* dust), 862 and NOX removal efficiencies of 95 and
90%, respectively, were attained. The 502. which is absorbed as the SOj"
ion, is oxidized to SQ^** either by reducing the absorbed NOX to molecular
N2 or by reacting with 62 absorbed into the solution. Approximately 301 of
this SQ3= to 504° oxidation is assumed to occur in the absorber with the
remainder occurring in the oxidizer of the CaS04'2H20 production section.
The absorbed S02 is thus recovered in the form of byproduct CaSO^'2H20 which
along with the flyash will probably be used as a landfill material. Some
of the absorbed NOX is reduced to molecular N2 by oxidation of the S0-j~ ion
while the remainder is removed as NG-j~" salts, either as a solid salt mixture
or as an ammonium nitrate (NH^NOj) solution.
The NOX removal efficiency is dependent on the amount of C102 injected.
If the Injection of C102 declines below the 10% excess rate specified by
the process developers, the NOX removal efficiency will drop below 90%. A
complete failure of the C102 generation equipment will result in only approx-
imately 5% of the NOX being removed. The S02 removal efficiency, which is
independent of the amount of C102 injected or the inlet flue gas conditions,
will remain >95%.
The C102 used to oxidize the NO in the flue gas is the only significant
work hazard associated with this process. This gas is a very strong oxidant
and can be dangerous if workers are exposed to even low concentrations.
CriticalDataGaps and Poorly Understood Phenomena
Listed below are the critical data gaps for the Moretana Ca-based
system. In addition preliminary engineering assumptions are listed for
some of these critical data gaps.
1. Stream compositions and operating conditions in each piece of
process equipment
2. Percent of NOX reduced to molecular N2 (assumed to be 50%)
3, Oxidation of 803™ ion in absorber
4. Raw material, energy, maintenance, and manpower requirements and
costs
5. Identity and concentration of catalyst (assumption: Cu ion and
some proprietary additives)
121
-------
Ihe sequence of reactions occurring in the scrubbing solution to reduce
the NOX to molecular N£ is not completely understood with Fuji Kasul-Sumitomo
apparently only providing overall reactions. The sequence is hypothesized by
others (33) to be the following: The NC»2 apparently forms the dimer molecule
and enters the solution in this form. From a literature survey the absorbed
dimer molecule forms both NC>2~ and N03 ions in the scrubbing solution. The
reduction of the NC>2~ ion to molecular % by the combined action of the HSO^"
ion and the dissolved catalyst, which is a major advantage for this process,
is not well understood and only the overall reaction is supplied by Fuji
Kasui-Sumitomo (reaction 156).
Advantages and Disadvantages
Since this is a wet simultaneous S02~NOX process, it has certain advant-
ages and disadvantages over the dry MOX processes and some over other wet
nonoxidation processes. The advantages and disadvantages are listed below.
Advantages
1. Removes NOX and S02 simultaneously
2, Achieves >95% SC»2 removal efficiency
3. Is a slight modification of a commercially available FGD system
4. Operates with full partlculate loadings (>7 gr/sft-')
Disadvantages
1. Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Incorporates design features which may present significant process
control problems
4. Uses significant amounts of stainless steel or exotic materials for
process equipment
5. Requires flue gas reheat for plume buoyancy
6. Requires an expensive gas-phase oxidant
122
A
-------
MORETANA SODIUM PROCESS - WET, OXIDATION-ABSORPTION-REDUCTION (NOX-SOX)
Process Description and Principles of Operation (5, 17)
The Moretana Sodium process jointly developed by Fuji KasuI-Sumltomo
uses C102 for the gas-phase oxidation of NO to N02 and an NaOH-NagSOj solution
to absorb both S02 and NQ2- The overall process Is made up of six major
sections Including (1) prescrubblng, (2) gas-phase oxidation, (3) S02 and
NOX absorption, (4) SOj™ oxidation, (5) byproduct Na2SO^ recovery, and (6)
wastewater treatment. A more detailed outline of this process can be seen in
the block flow diagram given in Figure 24.
Fuji Kasui- Sumitomo also has a Ca-based process which is based on
similar technology except limestone is used as the raw material. The primary
differences between these two processes are the operating costs Involved. For
this reason Fuji Kasui-Sumitomo recommend their Na-based system only for
relatively small installations treating <100,000 Nm^/hr (33.5 MW equiv) and
more specifically for those sites burning a gaseous or liquid fuel. Thus,
for most U.S. applications, the Ca-based process is more Important since it
will be used for large coal-fired power plants.
The use of C102 in the Moretana Sodium process to chemically oxidize
the NO overcomes the major problem of wet denltrification systems, i.e., the
relative insolubility of NO In aqueous solutions, and substantially reduces
the equipment requirements for the denitrification process. This is readily
apparent from a comparison of the block flow diagram for this process shown
in Figure 24 with the block flow diagrams given for the other types of wet
processes. The gas-phase oxidation of NO with C102, however, In addition to
the increased expense for gaseous C102» also results in the formation of
undesirable N03~ and Cl~ aalts in the scrubbing solution.
The flue gas from the air heater passes through a Moretana prescrubber
countercurrently to an NaOH solution. This solution is assumed to remove
90% of the particulates and essentially all of the HC1 from the flue gas. At
the same time the flue gas is adiabatically cooled from 150°C (300°F) to
53°C (127°F) and humidified by the evaporation of H20 from the scrubbing
solution. The reactions occurring in the prescrubber, due to the absorption
of 503 and HC1, would include the following:
S°3(g) * S°3(aq)
HC1(g) ^ HC1(aq)
S°3(aq) + H2° + 2NaOH(aq) * Na2S°4(aq) + 2H2° (166)
HCl(aq) + NaOH(aq) - NaCl(aq) + H20 (167)
123
-------
NJ
f
BOILER
>
t
AIR
HEATER
AIR
i
• '
^PRESCF
H20
rj
r.
?UBBf
-R
1
NoOH
1
EFFLUENT
HOLDING
TANK
i
CHLO
j DIO>
GENER
\
IV
r
RINE
rinp
ATOF
t
R
VASTE LIQUID)
r •
ABSORBER
(
EFFLUENT
HOLDING
TANK
i
V
EGENERATX
HOLDING
TANK
i
OXIDATION
REACTOR
i
CE
ASf
MTRIF
H
"UGE.
.1
N(V
k
'
i
OH
1
e
OR
1
CE
REHEAT
, HgO
^ f
^APORATOI
i •
COOLING
YSTALLIZE
i i
No2SO4
NTRIFUGE
FLUE GAS
^
•R
H20
t
* URVER
~T ~T
Na2S04 NaCI
NaNOa
Na2S04
ASH
Figure 24. Flow Diagram of Moretana Sodium Process.
-------
The liquid effluent from the prescrubber drops into a holding tank
where makeup H20 and makeup NaOH are added to replace these raw materials
consumed during passage through the prescrubber. Most of the solution is
recirculated through the prescrubber but a small portion is removed as a
purge stream to prevent the buildup of flyash and the Na salts in the pre-
scrubber solution. This purge stream is pumped to a holding tank in the
regeneration section for further processing.
The flue gas passes through a mist eliminator at the top of the pre-
scrubber and is injected with a weak (3-5%) C102 gas stream in the flue gas
duct prior to entering the absorber. The C102 is generated onsite and is
injected in the amount of 10% in excess of the stoichiometric requirement.
With a properly designed injection system, the C102 selectively oxidizes
NO by the following reactions
2NO, % + C10-, % + H.,0, ,. -*• N00, . + HNO,,, % + HC1, .
(g) 2(g) 2 (g) 2(g) 3(g) (g)
This gas-phase reaction is very rapid (0.3 sec) and essentially goes to
completion. The reaction is also very selective, in that none of the S02 is
oxidized by the C102-
Immediately after the injection of the C102» the flue gas passes
countercurrently to a slightly basic (pH of 7.2-7.4) Na-based scrubbing
solution in a Moretana plate tower. As the solution passes through the
absorber, the S02 is rapidly stripped from the flue gas and the following
reactions occur in the solution:
S°2(g) * S°2(aq)
2NaOH(aq) + S00,_,.N + Na0SO,,^N + HnO (170)
At the same time the gases formed from the oxidation of NO in reaction
(168) are also absorbed from the flue gas stream and undergo the following
hypothesized reactions;
N02(g) * N02(aq)
2N°2(aq> + 4Na2S03(aq) * 4Na2S°4(aq) +
™03(g) * M°3(aq)
2HN03(aq) + NaOH(aq) * NaN°3(aq) * H2° (174)
HC1(g) * HCl(aq)
HC1(aq) + NaOH(aq) * NaGl(aq) + H2° <176>
125
-------
In addition to these primary reactions, a secondary reaction occurs as
the flue gas passes through the absorber. Flue gas 02, which is readily
absorbed into the scrubbing solution, causes the following undesirable side
reaction:
Na2S°3(aq) + 1/2°2(aq) * Na2S°4(aq) (177)
This loss of SC>3= Ion by oxidation can be a substantial problem at higher
flue gas Q£ concentrations. After passing through a mist eliminator, the
flue gas from the top of the absorber is reheated and exits through the
stack.
The spent scrubbing solution Is removed from the bottom of the Moretana
plate tower and sent to an effluent holding tank. Most of this solution is
removed and recycled to the absorber after makeup NaOH has been added. The
remaining portion of the solution, containing mainly Na2S04 and Na?SO», is
pumped to another holding tank in the regeneration section where it is mixed
with the purge stream from the prescrubber. The liquid effluent from this
holding tank In the regeneration section and the liquid waste from the C102
generating system are sent to an oxidation reactor where air is injected to
oxidize any remaining SOo" to 804™. The solution from the oxidizer is
pumped to a centrifuge where the flyash is removed, and then on to an
evaporator where Indirect steam heating is used to concentrate the solution.
This concentrated solution is passed to a cooling crystallizer where Na2S04»
which is the largest constituent and the least soluble of the Na salts in
the solution, is precipitated and then recovered as a byproduct in a.
centrifuge. The mother liquor from this centrifuge is sent to a dryer where
the liquor Is desiccated and a mixture of Na salts is recovered.
Status of Development
This process, originally conceived and developed for dust removal and
deaulfurlzatlon In 1970, was modified In 1973 to simultaneously remove NOX
from flue gas. The only significant modification required to make this
change to a simultaneous SQ2~NQX removal system was the addition of the C102
generation and injection system. At the present time five prototype units
(10-40 MW) using the Fuji Kasui-Sumitomo Na-based simultaneous S02-NOX
removal process are treating flue gas from heavy oil-fired boilers (typically
containing 1500 ppm S02» 200 ppo NQX» 4% 02, and 0.1 g/Nm3 of dust), A unit
(20 MW equlv) at the Sumitomo plant in Amagasaki City has been operating
since 1973, at first only as a desulfurization unit and then later modified
for simultaneous 802-NOjj removal.
Although this process has not been tested on flue gas from a coal-fired
boiler. It has been applied to dust removal and desulfurization of flue gas
from sintering furnaces and refuse incinerators with dust loadings as high as
1-3 g/Nm3 CO.41-1.22 gr/sft3 at 60°F) which is approximately one-fifth the
dust levels expected from coal-fired boilers. It has also been used for
126
A
-------
desulfurlzatlon of the exhaust gas containing as much as 3% SOX. Once an
Initial prescrubber has been added to remove the increased levels of Cl~ in
the flue gas, this process would probably be capable of handling coal-fired
flue gas.
Background ofProcess Developer
In the past 20 yr, Fuji Kasui has been active, in the development of
pollution control equipment and processes. In the early 1970*s» Fuji Kasul
undertook the development of processes to remove 862 and NOX from combustion
exhaust gas and later began joint research with Sumitomo for development of
processes using the Moretana absorption tower. This involvement began with
the design and engineering of the Moretana absorption tower and the later use
of this absorption tower both for a particulate removal and FGD applications.
One of these early applications was for the desulfurization of heavy oil-
fired flue gas at a Sumitomo plant in 1971. In May 1973 a prototype plant
(20 MW equiv) was started up using a CaO scrubbing process for desulfurization
and a new test plant for simultaneous S02 and NOX removal was begun.
At the present time 12 desulfurization plants and 7 combination desul-
furization-denitrification plants (10-40 MW equlv) are already operating and
hence at the present time Fuji Kasui-Sumltomo probably have more experience
In flue gas denitrification than any other company.
No American company has been licensed by Fuji Kasui-Sumitomo to market
their process in the U.S. However, Sumitomo does have a liaison office in
New York City through which Inquiries about this process are handled.
Published Economic Data
The following estimated costs for the Moretana Sodium process are based
on 62,000 Nm^/hr of flue gas from a heavy oil-fired boiler, which contains
1500 ppm S02, 200 ppm NOX, 4% 02» and 0.1 g/Nm3 of dust, with a 99% S02 and
90% NOX removal efficiency respectively. The total capital investment and
the revenue requirements have been estimated (17) by Fuji Kasui-Sumitomo
(assumed basis—Japan and 1976 costs) as 830 million yen and 11,160 yen/kl
of oil respectively. If 300 yen/$ and 3000 Nm^/hr/MH are assumed, these
values correspond to an estimated capital investment of $134/kW of Installed
capacity and a revenue requirement of 8.9 mills/kWh.
Raw Materials ^Energy, and Operation Requirements
The major raw material requirements given below are based on treating
62,000 Itoi3/hr of oil-fired flue gas containing 1500 ppm S02 and 200 ppm
NOX. The desulfurization and denitrification efficiencies were assumed to
be 95% and 90% respectively.
Raw material Quantity
NaOH 204 kg/day
127
-------
The utility requirements on the same basis as listed above ares
Utility _ Quantity
Electricity 100 kW
Steam 32,500 kg/day
Fuel oil 290 kg/day
The estimated manpower to operate and maintain this system has not been
published.
Technical Considerations
The Moretana Sodium process, using ClOg for the gas-phase oxidation of
NO and an K^SO^ scrubbing solution to absorb the resulting N02» appears to
require fewer pieces of process equipment than the other oxidation processes.
As a result the process control for this system should be less difficult than
that required for the reduction processes. Since C102 cannot be stored but
must be generated onsite, the major process control problem for this process
will be the C102 generation and injection system. If the C102 is injected in
less than stoichiometric amounts, the NOX removal efficiency will decline and,
if injected in more than stoichiometric amounts, excess C102 will contribute
excess Cl~ ions to the scrubbing solution and increase operating costs for
both C102 generation and makeup NaOH.
The use of C102 as the gas-phase oxidant also complicates the process
control scheme for another reason. Since C102 is generated by chemical
reaction and requires significant time periods (15-30 mln) to adjust to
changing NOX levels in the flue gass this process will have difficulty
cycling with changing flue gas NOx concentrations.
The N02 removal efficiency is relatively insensitive to the inlet gas
composition providing that the S02 :NOX mol ratio is relatively high (a mol
ratio of 5 is required for 90% NOx removal). The major reason for the
sensitivity of this process to the S02 concentration in the flue gas is that
although only approximately 50% of the absorbed NOX is reduced by the 503™
ion to molecular N2» substantial amounts of 803™ (absorbed S02) in the
solution are lost due to oxidation by flue gas 02- For NOX removal effi-
ciencies of less than 90% and/or lower flue gas 02 levels (<4-5I) the
required S02!NOX raol ratio would be lower. Thus for coal-fired boilers,
particularly boilers fired with low-S western coal, high NOX removal effi-
ciencies may be difficult.
The Moretana plate tower, which was developed by Fuji Kasui-Sumitomo , is
used for both the prescrubber and the absorber in this process. The Moretana
plate tower is very similar to a sieve tray absorber except that it has no
weirs or downcomers as can be seen in Figure 23. As the liquid and gas pass
countercurrently to each other in the tower, there is an alternating flow of
first liquid and then gas through the perforations in each plate. As the
128
A
-------
height of liquid on a plate increases, the weight of the liquid overcomes the
force of the gas pushing through the perforations and the liquid displaces
the gas and flows through the first plate to the lower plate. As the liquid
height on the first plate decreases, the gas pressure forces the displace-
ment of the liquid and the gas flows through the perforations ir the plate.
This alternating pulsing of first liquid and then gas through the plate gives
the gas-liquid contact required for simultaneous SQ2-N02 absorption. In
addition, the developers claim this design allows the Moretana plate tower
to handle higher superficial gas velocities and thus reduces the size of the
absorber required in a particular situation.
For Fuji Kasui-Sumitomo ' s Ka scrubbing process, the Moretana plate
tower absorber is operated at an L/G ratio of 2.0-3.5 I/Km3 (10-25 gal/kaft3
at 127°F) and a superficial gas velocity of 3-5 m/sec (10-16.5 ft/sec). The
pressure drop through the Moretana plate tower operating under these condi-
tions (i.e., a superficial gas velocity of 4.5 m/sec and L/G ratio of 25) at
a pilot plant averaged 200 mm of H20.
The major disadvantage associated with this process is the extensive
energy requirement for the regeneration section. These high energy require-
ments, particularly the steam needed for the evaporators, are a direct
result of the oxidation of NO by C102« The conversion of NO to N02 results
in the formation of ML*3 salts when the N02 is absorbed into the scrubbing
solution. Further complicating the regeneration section is the formation
of Cl~ salts from the C102 used to oxidize NO. Not only is this NaCl salt
extremely soluble and thus difficult to remove, but It also consumes
expensive NaOH. Since the NOg" salts are also extremely soluble, the process
developers have suggested a two-stage desiccation scheme using steam
evaporators to recover these salts. Unfortunately with increases In primary
energy costs, the use of steam to evaporate this purge stream from the
regeneration section can only lead to high and steadily increasing operating
costs.
Recently the developers have begun tests using Og instead of 0102
the gas-phase oxidation of NO. Although the use of 03 eliminates the
problems associated with the formation of Cl~ in the scrubbing solution, 03
has two major disadvantages when compared with C102- First, 1 mol of ClO^
will oxidize twice as many mols of NO as 1 mol of 03 and secondly, the unit
cost of 03 is approximately four times as expensive as that of C102 system
for its wet simultaneous S02~NOX removal systems.
Although this process has not been tested on flue gas from a coal-fired
boiler, the developers do not expect any serious problems in converting this
system to handle the increased particulates and Cl~ associated with coal
combustion. The major reason for this confidence is the wide experience in
applying their FGD system which is very similar to their simultaneous S02~
NOX removal system. Their desulfurlzation system has been tested under
various operating conditions ranging from heavy oil-fired furnaces (1400 ppm
S02 and 0.2 g/Nra-* dust), to a Rhodlne salt incinerator (28,900 ppm S02 and
129
-------
0.5 g/Hm^ dust), to a refuse incinerator (1-3 g/NnP dust). However, the
increased levels of NOX in flue gas from a coal-fired boiler will sharply in-
crease the consumption of C102 and this will further aggravate the problem
of Cl~ salts production and its associated consumption of expensive NaOH.
Since this is a wet simultaneous removal system, retrofitting this
system on existing power plants should be relatively easy if sufficient land
is available nearby for the siting of the process equipment. The major
modification to the existing plant would be the installation of a common
plenum from the boilers or the ESP to the two scrubber trains on a 500-MK
boiler. The existing cold ESP would probably be removed and replaced with
Moretana plate towers acting as prescrubbers to remove the Cl~ as well as the
flyash.
The use of a common plenum and four absorber trains will allow this
system to be easily turned down from 100% to 75, 50, or 25% by simply
closing off any combination of scrubber trains.
Since the use of C102 as an oxidizing agent will inject relatively high
(2000 ppm) concentration of Cl~ ions in the scrubbing solution, most of the
process equipment and piping will need to be elastomer-lined carbon steel
to prevent pitting corrosion associated with circulating Cl~ solutions.
Exceptions to this elastomer lining of process equipment and piping Include
the inlet flue gas ducts, which should be constructed of Incoloy to prevent
corrosion by SOj in the flue gas, and the absorber and absorber pumps which
should be made from 316L stainless steel.
Environmental Ooneidera_tionis
Over a long-term continuous operation of their prototype units (10-40
MW equiv) on heavy oil-fired boilers treating flue gas containing 1500 ppm
S02, 200 ppm NOX, and 4% 02, S02 and NOx removal efficiencies of 95 and 90%,
respectively, were attained. The S02 is absorbed as an SOj™ ion and
partially oxidized to 804" in the absorber either by reducing NOX to
molecular N2 or by flue gas 02 absorbed into the solution. The remaining
803™ ion in the solution is converted to SO^™ by reaction with air in an
oxidizing tower. Part of the NOx (assumed 50%) is reduced to molecular $2
fay the SOj53 ion in the scrubbing solution while the remainder Is removed
from the solution as NO-j" salts.
The NOX removal efficiency is directly dependent on the amount of C102
injected. If the injection of C102 declines below the 10% in excess of
stoichiometric rate, the NOX removal efficiency will drop below 90%. A
complete failure of the C102 generation equipment will result in only
approximately 5% of the NQx being removed. On the other hand the S02
removal efficiency will remain well above 95% since it is independent of the
amount of C102 injected.
A potential NOj" wastewater problem has been eliminated by the decision
to evaporate the purge stream and recover the mixed Na salts as a solid. If
130
A
-------
a market cannot be found for the Na salts, the ultimate disposal of these
salts could become a problem.
The C102 used to oxidize the NO in the flue gas is a very strong
oxidizing agent and hence presents a significant work hazard for this process.
Critical Data G_aps__and Poorly Understood Phenomena
Listed below are critical data gaps followed in some cases by the pre-
liminary engineering assumptions used to overcome these data gaps.
1. Percent of NOX reduced to molecular N£ (assumed to be 50%)
2. Oxidation of Na2S03 in absorber with 4% D£ in flue gas
3. Operating conditions in the process equipment and the stream
compositions
4. Maintenance and manpower requirements
Included under poorly understood phenomena for this process is the
following;
The actual sequence of reactions occurring in the scrubbing solution
to reduce the NOX to molecular N£ is not completely understood. From
a literature review it is hypothesized that the N02 apparently forms
the diner molecule and enters the solution in this form. However,
the process developers claim that N02 enters the scrubbing solution
as N02 and forms only the N02~" Ion. The resulting absorbed NOX is
apparently reduced by the 803*" ion to form molecular N£.
Advantage s and Pis advantages
Since this is a wet simultaneous S02-NOX process, it has certain
advantages and disadvantages, as shown below, over dry NOX processes and
some over other wet nonoxidation processes.
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >95% S02 removal efficiency
3. Produces a potentially marketable byproduct (^2804)
4. Is a slight modification of a commercially available FGD system
5. Operates with full particulate loadings (>7 gr/sft^)
Disadvantages
1. Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Incorporates design features which may present significant process
control problems
4. Uses significant amounts of stainless steel or exotic materials
for process equipment
5. Requires flue gas reheat for plume buoyancy
6, Requires an expensive gas-phase oxidant
131
-------
NISSAN PERMANGANATE PROCESS - WET, ABSORPTION-OXIDATION (NOX)
Process Description and Principles of Operation (5)
The Nissan Permanganate process was developed by Nissan Engineering, a
subsidiary of Nissan Chemical Industries, Ltd. This process was originally
designed to treat HNO-j tail gas but with the installation of an FGD system
prior to the denitriflcation system, the process, from a technical standpoint,
can be adapted to fully treat power plant stack gas. The Nissan Permanganate
is divided into six major sections (1) prescrubbing , (2) S02 absorption, (3)
byproduct 804" formation, (4) NOX absorption, (5) NOX oxidation with Mn04~,
and (6) removal and regeneration of the MnO^". A block flow diagram of the
Nissan process is shown in Figure 25, The Nissan Permanganate process is very
similar to the MON Alkali Permanganate process as seen from a comparison of
Figure 21,
Although not mentioned by the process developers, this system is expected
to require a closed-loop prescrubber section when treating flue gas from a
coal-fired boiler to remove the particulates and the Cl~. In addition to re-
moving 90% of the particulafces and essentially all of the Cl~, this prescrubber
also humidifies and cools the gas from 150°C (300°F) to 53°C (127°F) .
The flue gas from the prescrubber enters a desulfurization system to
remove most of the 862 and thus prevent the excessive consumption of MnO^"
ion in the denitriflcation system. Since no specific FGD system has been
mentioned by Nissan Engineering, a limestone throwaway system was assumed.
The flue gas from the limestone scrubber enters the devitrification
scrubber and passes countercurrently to a KOH scrubbing solution containing
KMnO^. The NO is gradually absorbed over the length of the scrubber and
undergoes the following reactions:
N°(g) * N°(aq) (178)
N0(aq) + 2KOH
3NO(aq) 4 2KOH(aq) 4- J0ta0 - Mn0^ + 3KN0 + (180)
2N°2(g) * N2°4(g)
N2°4(g) + N2°4(aq) <182>
N2°4(aq) + ay^fcq) * 2KMn°4(aq) + 2KN02 (183>
However, the main reaction occurring in the neutral absorbing solution is
the direct oxidation of the NO by KMnO^, i.e.,
N0(flq) + KMn04(aq) * KN03(aq) + Mn0 , (184)
132
A
-------
i
OJ
OJ
i
BOILER
AIR
HEATER
AIR
CoC03-
PRESCRUBBER
NEUTRAUZATION
H
PURGE
I
LIMESTONE
'PREPARATION
MANGANATE
REACTOR
CLEAN
FLUE GAS
HN03(Aq)
ELECTROLYTIC
lOXICXZERi
Figure 25. Flow Diagram of Nissan Engineering Process.
-------
Gas from the scrubber Is reheated and sent to the stack. The scrubbing
solution dcops into a holding tank from which most of the solution is re-
cycled to the absorber after makeup KOH has been added. A small purge stream
is removed from the holding tank and pumped to a centrifuge in the regenera-
tion section where the precipitated Mn02 is removed from the KNG^ solution,
This KNO-j solution is sent to an electrolytic cell to produce a weak
HN(>3 solution (25-30%) and a mixed stream of KOH and KNC^. The final use of
this HNC>3 stream is not specified but the mixed K salt stream is reacted with
the earlier removed Mn02 according to the following reaction:
Mn°2(s) + 2KOH(aq) + l/202(aq) - O + (185)
The resulting potassium manganate (l^MnCh) is subjected to electrolytic
reduction to regenerate
H2° - ^(.q) + KOH(aq) + 1/2H2(g)t <186)
The regenerated K solution is pumped to the holding tank to be recycled to
the absorber!
Status of Development
Initial testing of the Nissan Permanganate process was begun in the
early 1970 'a. Since 1972 four small pilot plants (100-2000 Nra-Vhr) were
constructed and operated treating the tail gas from HNO^ plants.
This process has apparently not been tested on any type of power plant
flue gas.
Since so little information is available for this process, it is
difficult to project the next stage of development.
Background of Process Developer
Nissan Engineering, a subsidiary of Nissan Chemical Industries, Ltd.,
has developed this process using an alkali permanganate to convert the NOX
in HNO^ tail gas to an alkali nitrate.
Nissan Engineering's experience in air pollution control is apparently
limited to this process for HN03 tail gas with no background in treating
power plant flue gas,
Pub 1 i s jhed EC onomic Data
No information has been published on the total capital Investment or
the annual revenue requirements for the Nissan Permanganate process.
134
A
-------
Raw MaterjLalj Energy, and Operation Requirements
No information about the raw material, energy, or operating require-
ments for this process has been published. The major raw material for the
Nissan process would be makeup KMnO^ and KOH. The major energy requirements
would be electricity for both the electrolysis reactors and also operating
the process equipment and fuel oil for the flue gas reheat section.
Technical Considerations
See similar section for MON Alkali Permanganate process.
Environmental Consideratlong
Essentially no information has been made available concerning the NOX
removal efficiency and the sensitivity of this removal efficiency to changes
in either the flue gas composition or the operating conditions in the
absorber.
The generation of a weak HNOj solution represents a significant waste
disposal problem since there is essentially no market for this solution.
The generation of H/? during the electrolytic oxidation of l^MnQ^
represents a significant work hazard for this process due to its explosive
nature.
Critical Data Gaps and Poorly_J^derjB_tood^Phenomena
There is insufficient information available to discuss the critical
data gaps in detail or to discuss the poorly understood phenomena. Hence,
only the major gaps are included below,
1. Equipment operating conditions and sizes
2. MnO^~ regeneration section
3. Stream flows and compositions
4. Estimated capital and operating costs
5. Raw material, energy, and operation requirements
6. Pilot plant data
7. Ability to handle power plant flue gas
Advantages and D1sadvantages
Although sufficient information has not been published to enable a
complete listing of the technical advantages and disadvantages for the
Nissan Permanganate process, the following list has been compiled from the
available data,
Advantages
None apparent at the present time.
135
-------
Disadvantages
1. Forms secondary source of pollution (25% HN03)
2. Requires significant amounts of energy for the regeneration step
3. Has not been tested on coal-fired flue gas
4. Has not been developed beyond the conceptual design stage
5. Requires clean (S02~free) gas feed
6. Uses significant amounts of stainless steel or exotic materials
for process equipment
7. Requires flue gas reheat for plume buoyancy
8. Requires an expensive liquid-phase oxidant
136
A
-------
PITTSBURGH ENVIRONMENTAL AND ENERGY SYSTEMS SCORe PIOCESS - WET, ABSORPTION-
REDUCTION (NOX-SOX)
Process Description and Principles of Operation (55, 76)
The Pittsburgh Environmental and Energy Systems' SCORe (SULF-X Concurrent
Oxides Removal) process is a wet St^-NO^. removal process using a ferrous
sulfide (FeS) solution to absorb both S02 and NOx. This process is somewhat
different from the other wet, absorption-reduction processes in that the
sorbent material, FeS, is used to reduce both the S0£ to elemental S and the
NOX to molecular %. The sorbent is then regenerated using carbon (C) as a
reducing agent in a coal-fired kiln. The SCORe process consists of five
major sections including (1) particulate removal (ESP), (2) absorption of
S02 and NOX» (3) separation and reduction of absorbent, (4) S recovery, and
(5) ash-sludge disposal, A block flow diagram of this process is shown in
Figure 26,
After passing through a cold ESP, the flue gas enters a quenching tower
where a spray of recycle FeS slurry at a pH of 6 adiabatically humidifies and
cools the flue gas from 150°C (300°F) to 53°C (127°F). In addition some of
the S02 and NOX are absorbed into the slurry in the quenching section and
undergo the following reactions;
S°2(g) - S02(aq)
5FeS(g) + S02(aq) + 2FeO(s) -h 3FeS (188)
S°(g) - N° <189>
2FeS(s) + N°(aq) * 1/2N2(g) + Fe°(s) + FeS2(s)
The liquid effluent from this section drops into a. holding tank from which
most of the slurry is recirculated .
The flue gas leaves the quench section and enters the first stage of the
absorber where the gas passes cocurrently with a spray of makeup FeS slurry.
The S02 and NOx levels in the flue gas are further reduced by reactions (187)
through (190). The slurry from this first stage of the absorber drops into
the quench section holding tank to be recycled to the quench tower. The
flue gas is further stripped of both S(>2 and NOX in a second stage of the
absorber by contacting a spray of regenerated FeS slurry. The flue gas from
this second stage is passed through a mist eliminator, reheated, and exhausted
through the stack,
The spent slurry from the second stage of the absorber is pumped to a
thickener where a purge stream of flyash and FeS is removed from the bottom
137
-------
u>
oo
ABSORBER
SECOND
STAGE
Lr-1 J
FeS-ASH SLUDGE F«s
CT
HgO
CLEAN
FLUE
GAS
Figure 26. Flow Diagram of Pittsburgh Environmental Energy Systems Process.
-------
of the thickener. This bottom stream is sent to a filter where the solids
are removed and purged from the system. The filtrate from this ash filter
is combined with the thickener overflow and recycled to the absorber second
stage after makeup 1^0 has been added.
A purge stream from the quench tower holding tank is pumped to a
thickener in the regeneration section. The thickener bottoms are filtered
and sent to an indirectly fired regenerating kiln where coal is used both as
a fuel for heating the kiln and as a source of C to reduce the ferrous
disulfide (FeS2) formed during the reduction of both the SQ2 and the NO. The
primary reaction occurring in the kiln is:
FeS0, s + C, . + 00/ N -»• FeS, N + S, . 4- CO-, , (191)
2(s) (s) 2(g) (s) (g) 2(g)
The ferrous oxide (FeO) produced during the reduction of S02 and NOX is also
converted in this kiln according to the following reaction:
2FeO, . + 3FeS_, . + 2C, . + 0,,, N * 5FeS, v 4- S, , + 2C00, s (192)
(s) 2(s) (B) 2(g) (s) (g) 2(g)
The off-gas from the reducing kiln is passed through a cyclone to remove
some of the particulates and then is cooled to condense approximately 80% of
the S as a liquid byproduct. The remaining gas is sent through the kiln
firebox and then recycled to the flue gas duct to reenter the quench tower.
Recent tests are said (76) to indicate that the hot solids do not
agglomerate and hence crushing equipment is no longer needed. Although not
mentioned by the process developers, these hot solids (750°C) are probably
cooled in a surge bin before being sent to a reslurrylng tank. Feedstreams
to the reslurrying tank include the thickener overflow and filtrate from
the spent sorbent drying section. This regenerated FeS slurry is then
recycled to the first stage of the absorber.
Status of Development
The SULF-X process was originally conceived in mid-1974 and in late
1974 bench-scale testing was begun. Although originally developed for flue
gas desulfurizatlon, early laboratory and subsequent field tests are said
to have shown the capability of removing HOX- After absorption tests were
completed in March 1976, testing of the absorption section was begun in a
1-MW pilot plant treating flue gas from a coal-fired boiler at the U.S.
Army's Fort Benjamin Harrison power plant. The regeneration of the spent
sorbent, however, has not yet been demonstrated on a pilot-plant scale.
During a 52-hr (2-day) continuous test of the absorption section, it was
claimed that the S02 removal efficiency averaged 90% and the NOX removal
ranged from 60-70%. Pilot-plant testing was discontinued after only 2 mo
of operation.
139
-------
T\e next stage of development for this process would be to demonstrate
long-term continuous operation with the entire system coupled as a unit in a
small pilot-plant treating flue gas from a coal-fired boiler. At the present
time plans are being finalized for such a pilot plant to be built at a
Westlnghouse plant in western Pennsylvania.
Background of Process Deve lope r
Pittsburgh Environmental and Energy Systems, Inc. (PENSYS) , is a small
American company baaed in Pittsburgh with apparently very little previous
experience in air pollution control. Most of their experience has been
gathered during the development of their SCOEe process, originally for FGD
only, and later for simultaneous removal of S02 and NOX.
At the present time no company has been licensed to market this
technology and all inquiries about this process are handled by PENSYS offices
in Pittsburgh.
Published Economic Data
PENSYS had originally estimated (55) the total capital investment and
annual revenue requirements as $12,500,000 and $339,300, respectively,
(assumed basis: U.S, and 1976 dollars) based on a 100-MW coal-fired boiler
burning 3,5% S coal. The reported economics corresponds to a capital in-
vestment of $125/kW of generating capacity and a revenue requirement of 0.49
mllls/kWh. However, the regeneration system has recently been modified,
as described in this report, and hence this economic data need to be updated.
Although not specified by the process developer , the system was assumed to be
able to remove 90% of the S02 and 60% of the NOX from the flue gas containing
approximately 1000 ppm S02 and 600 ppm
Raw Material > Energy , and Operation Requirements (76)
The only major raw materials required for this process are FeS for
makeup slurry and coal for the reducing kiln, PENSYS offers three or four
sources of obtaining this FeS but the process vendors believe that the
pyritic S from coal waste would be the most economical source. The following
requirements have been estimated for a 100-MW coal-fired unit burning 3.5%
S coal (2400 ppm SO 2 and 600 ppm NOX) and assuming 90% of the S02 and 60%
removal of the NOX. (All quantities measured in tons are short tons.)
Raw material _ Quantity ___
FeS 820 tons/yr
Coal 15,450 tons/yr
The major utility requirements include fuel oil for reheating the clean
flue gas (which was not specified) , electricity for operating the process
equipment, and makeup H_0, Although the process developers indicate recovered
140
A
-------
heat from the regenerator could be used for reheating, without the installa-
tion of an additional ESP to sharply lower partlculate emissions from the
regenerator, it is uncertain whether this off-gas could be used as reheat.
The quantities required, using the same basis as given above, would be:
Utility Quantity
Electricity 1500 kW
Process water 265 gpm
Fuel oil (reheating) Unknown
The operating manpower and maintenance requirements for this process
(100 MSI) have been estimated by PENSYS as two operators per shift and
$239,000/yr respectively.
Technical Considerations
The SCORe process developed by PENSYS appears to be a simple processing
scheme that should be relatively easy to control. The process does not use
a gas-phase oxidant and appears to be a simple scrubbing system with few
pieces of equipment.
The KOX removal efficiency ie claimed to be relatively insensitive to
inlet flue gas composition. In contrast with the other wet absorption-
reduction processes, the eorbetit in the scrubbing slurry is used to reduce
both the S02 and the NOX and hence no minimum amount of S02 in the flue gas
is required. The S02 removal is also said to be insensitive to the inlet
flue gas composition. The claim that this process is insensitive to the
inlet particulate composition is not well understood since the combination
of FeS solids and flyash in the scrubbing slurry would probably make the
separation of the flyash difficult and lead to extensive losses of sorbent.
Sensitivity to another important flue gas constituent, ©2 is not considered
to be significant by the process vendors. However, it is well known that
the Pe"*"2 ion, which is important in the absorption of NOXt is readily
oxidized to the Fe"*"3 ion by flue gas 02 absorbed into the scrubbing solution.
Although most of the Fe+2 ±on ±B present in the slurry as a solid and hence
practically immune to attack by flue gas 02. some will be present in the
solution and will be oxidized to the Fe+3 ion.
The absorber uses cocurrent flow in each of the three stages for gas-
liquid contact. The absorbent slurry is apparently pumped under pressure
into the absorber and is sprayed downstream with the flow of the gas. The
L/G ratio in the absorber is apparently very low, on the order of 10-15
gal/kaft3 of gas. The relative Insolubility of NO, which makes up 90-95%
of the total NO^ in the boiler flue gas, normally requires an L/G of 10-15
I/Mm3 (60-75 gal/kaft3) for reasonable NOX removal; (See Technical
Considerations under the other wet absorption-reduction processest) The
fact that the FeS is relatively insoluble in aqueous solution and hence not
freely available for reducing both the S02 and the NOX would seem to indicate
that the L/G should be even higher than that for the other wet absorption-
reduction processes for the same NOX removal efficiency.
141
-------
A major disadvantage of the PENSYS process is the extensive energy
requirements in the regeneration section. The reducing kiln requires
temperatures of about 750°C (1385°F) to regenerate the spent absorbent.
Although some-of this heat is recovered by recycling the S condenser off-gas
to partially dry the incoming absorbent and preheat the air to the kiln, most
of the energy expended in the kiln is apparently lost to the system. For
example, the hot solids from the kiln are partially cooled in a surge bin and
then H20 quenched with none of this heat recovered.
Another potential problem concerns the purity of the byproduct S.
Although a cyclone has recently been added to remove some of the particulate
matter, the byproduct S from the PENSYS system may not be of the same quality
as that produced in some of the other processes.
Since this is a wet simultaneous removal process, retrofitting this
system on existing power plants should be relatively easy if sufficient land
is available nearby for siting of the process equipment. The major modifi-
cation to the existing plant would be the installation of a common plenum
from the existing cold ESP to the four scrubber trains on a 5QQ-MW boiler.
This use of a common plenum and four absorber trains would allow this system
to be easily turned down by simply closing off any combination of scrubber
trains.
Although the process developers claim that most of the process piping
and equipment can be made of carbon steel, for long-term operation the
materials will have to be rubber—lined carbon steel to prevent the erosion
and corrosion associated with circulating slurries and Cl~. The reducing
kiln will probably have to be made of stainless steel due to both the high
temperatures and the reducing atmosphere required to regenerate the spent
sorbent.
EnvironmentalConsiderations
During a 52-hr test of the 1-MW pilot plant treating flue gas from a
coal-fired boiler (1000 ppm S02 and 600 ppm NOx), the absorption section of
the SCORe process demonstrated 90% S02 removal and 60% NOX removal. However,
the data base on which the NOx removal efficiency was determined is limited
and heace further testing in the pilot plant is required to confirm the
ability of this process to remove NOX. The S02 is absorbed, reacted with the
FeS in the scrubbing slurry, and regenerated to yield byproduct elemental S.
The NOX is absorbed, reduced by feS, and exhausted as molecular $£•
The only waste stream generated by this process is flyash-FeS sludge
which will need to be purged to a landfill disposal site. The NO-n is
reduced to molecular N2 and hence no wastewater NOg13" stream is generated.
The S02 is recovered as elemental S and hence no waste limestone sludge is
generated.
142
-------
Critical Data Gaps and Poorly Understood Phenomena
The critical data gape associated with the SCQRe process developed by
PENSYS Include;
1. Size and operating conditions of process equipment
2. Stream compositions and flow rates
3. Long-term pilot plant data
4. Updated economic data including capital investment and revenue
requirements
Included under poorly understood phenomena would be the following;
1. The ability of this process to achieve the claimed NOX removal
efficiencies does not seem reasonable under the operating
conditions specified by the process developers.
2. The impact of the oxidation of the Fe"*"^ ion by flue gas 02 on
the absorption reactions and NOX removal efficiency.
3. The process equipment and piping being constructed from mild
carbon steel.
4. The low annual revenue requirements.
5. The recovery of marketable grade S with this process.
Advantages and Disadvantages
Since this is a wet simultaneous S02~NOX process, it has the following
advantages and disadvantages as compared to the dry NOX processes and the
other wet processes.
Advantages
1. Removes NOX and S02 simultaneously
2. Produces marketable byproduct (elemental S)
3. Has been tested on coal-fired flue gas on pilot-plant or
greater scale
4. Operates with full particulate loadings (>7 gr/sft^)
Disadvantages
1. Can only achieve a maximum NOX removal efficiency of <70%
2, Requires significant amounts of energy for the regeneration step
3. Has not been operated as an integrated process
4. Has riot been operated for a long-term continuous period
5. Incorporates questionable design features (absorber L/G appears
to be too low for NO removal)
6. Uses significant amounts of stainless steel or exotic materials
for process equipment
7. Requires hot solids handling
8. Requires flue gas reheat for plume buoyancy
143
-------
TOKYO ELECTRIC-MITSUBISHI HEAOT INDUSTRIES PROCESS - WET OXIDATION-
ABSORPTION (NOX)
Process Description and Principles of Operation (5)
The Tokyo Electric-Mitsubishi Heavy Industries (MHI) process is a wet
oxidation-absorption system developed jointly by Tokyo Electric Power and
MHI, This process was originally developed for treating clean flue gas
(no SOX) and for this reason would only be applicable for natural gas-fired
boilers. The NOX Is oxidized by Oj to ^05, absorbed In 1^0 to form 8-10%
HN03, and concentrated to a 60% acid solution in the byproduct recovery
section.
The Tokyo Electric-MHI system consists of four sections (1) gas-phase
oxidation of NOX, (2) absorption, (3) byproduct HN03 recovery, and (4) excess
03 absorption. The overall outline of this process is shown In the block
flow diagram in Figure 27.
The flue gas from the boiler is sent through a spray tower for adia-
batic cooling and humidifIcation and then injected with 03 before entering
an oxidizing tower. The NOX is rapidly and selectively converted to $205
by the following reactions,
N°(g) + °3(g) *• N°2(g) + °2(g) (193)
2N°2(g) + °3
-------
-p-
Ln
1
mm ' •
BOILER
j
AIR
HEATER
f
AIR
i
-4
] '
H2O
V • '
HOLDING
TANK
, f
' I.
.. OXIDIZING
i TOWER
t
OZONE
GENERATOR
T
AIR
.
V.
S
NOy
ABSORBER
i
i
m ., *
.n
HOLDING
TANK
(
t
1
^
HN03
HOLDING
TANK
1
V.
k
c
1
\
OZONE
ABSORBER
oSOj
* i
r
HOLDING
TANK
1
t
'
H20(B)
t
I
i
ENTRIFUG!
I
SLUDGE
'
i
» REHEAT
1
^
r i
EVAPORATOR
HNO
Figure 27. Flow Diagram of Tokyo Electric-Mitsubishi Heavy Industries Process,
-------
(0,6 MW equiv) of flue gas from a gas-fired boiler containing 90-220 ppm NOX.
More than 90% NOX removal was achieved. Other operating conditions and
results have not been made available.
In December 1974 a prototype unit at the same power plant began opera-
ting based on the same technology. This unit treated 100,000 Nm3/hr (33.3
MW equiv) of flue gas from the gas-fired boiler. Operating conditions and
results of this prototype testing have not been published.
This process has not been tested on either oil- or coal-fired flue gas.
The plans for the next stage of development have not been published.
Background of Process Developer
Tokyo Electric is one of the larger power companies in Japan. MHI has
had extensive experience including the research and development of several
FGD systems and two other flue gas denitrif ication systems.
Published Economic Data
No information has been published concerning either the total capital
investment or the annual revenue requirements for the Tokyo Electric-MHI
process.
Raw Matter jjU^. Energy, and Operation Requirements
No information has been published concerning the raw material, energy,
or operating requirements for this process. However, the major raw material
for the Tokyo Ilectric-MHI system would be makeup CaSOj'l/Zl^O. The major
energy source would be electricity, primarily for 03 generation, but also
to operate the process equipment.
Technical Considerations
The Tokyo Electric-MHI process is a relatively simple NOX removal pro-
cess with few pieces of process equipment. The major process control prob-
lem will be generating and injecting the 03 in proportion to the amount of
NOX passing through the system. Although not mentioned by the process
developers, the 03 is probably metered into the duct in proportion to the
measured flue gas rate and NQX concentration.
The NOX removal efficiency is relatively insensitive to the inlet
flue gas composition and is dependent only on the 03:NOX mol ratio.
The relative size and operating conditions in the two packed-bed
absorbers have not been made available but they should be similar in size
and operation to conventional limestone FGD absorbers.
The use of 03 to generate ^Og instead of simply oxidizing the NO to
will result in at least a 50% Increase in the consumption of 03
(typically 03;NOX mol ratio of 1.7) over that required for the wet simultaneous
146
-------
OX processes using Oj, Since the cost of Oj generation typically repre-
sents approximately 30-40% of both the total capital investment and annual
revenue requirements for the flue gas denitrification system, increasing the
03 consumption by more than 50% would seem to rule out the widespread use
of this process due to economic considerations.
This high mol ratio of 03:NOX results in the need for a second absorber
to remove any remaining €3 before releasing the flue gas through the stack.
A circulating CaSOj'1/2H20 slurry containing a proprietary catalyst is used
to absorb this 03 and generate a byproduct sludge containing both SOg* and
SOA= salts of Ca. -This byproduct sludge would be pumped to a waste disposal.
pond.
Further diminishing the usefulness of this process in the U.S. power
industry is the fact that it tan be used only on clean flue gas, i.e. , SOX
free. Although the process developers have not listed specific reasons for
requiring S02~free flue gas, it would apparently complicate the chemistry
of the acid solution and prevent the removal of pure HNOj as a byproduct.
Retrofitting this system on power plants would be relatively easy pro-
viding sufficient land was available adjacent to the power plant for siting
the process equipment. The Installation of a common plenum to interconnect
the four scrubber trains on a 50Q-MW bailer would also be required. With
the use of four trains, the NOX removal system could easily cycle with the
availability of the boiler.
The primary material used in the construction of the process equipment
and piping would be 304L stainless steel. The only exceptions would be the
oxidizer which will be plastic-lined stainless steel.
EnvironmentalConsiderations
During the operation of the prototype unit (33 MW equiv) at Minamyokohama
power station treating flue gas from a gas-fired boiler (100-200 ppm NOX),
this process was able to maintain better than 90% NOX removal. The NOX is
absorbed and removed as a weak HNQ-j solution.
The Tokyo Electric-MHI process is designed specifically for NOX removal
and does not remove any other pollutants.
The NOX removal efficiency is dependent on the amount of 0^ injected.
If the mol ratio of C>3:NOX in the flue gas declines below 1.7, the acid
bleed stream will contain a mixture of HN02 and 1^03. The percentage of acid
stream present as undesirable HNC^ will increase as the 03:NOX ratio decreases
until an equimolsr mixture of HN02 and HNOj is reached at an C>3;NQX ratio
of (1.0-1.1):!. As the (>3:NQX mol ratio drops below 1.1, the NOX removal
efficiency will decrease below 90%. A complete failure of the 03 generation
equipment will result In an NOX removal efficiency of approximately 5%,
The use of 03 presents a significant work hazard since it is a strong
oxidizing agent even at low concentrations.
147
-------
Critical Data Gapsand Poorly Understood Phenomena
Included under critical data gaps are the following:
1. Stream flows and compositions
2. Absorber operating conditions
3. Economic data
4. Pilot or prototype data and results
5. Nature and coat of oxidation catalyst
6. Maintenance and manpower requirements
Poorly understood phenomena would include the reason why this process
can only be used to treat clean flue gas.
Advantages and Disadvantages
The primary advantage of the Tokyo Electric-MHI wet oxidation-absorption
process is its high NOX removal while the major disadvantage is the high
consumption and hence cost of 03. Other advantages and disadvantages are
listed below.
Advantages
1, Achieves >90% NOX removal efficiency
2. Produces a potentially marketable byproduct (60% HNOj)
Disadvantages
1, Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Requires clean (S02~ and particulate-free) gas feed
4, Uses significant amounts of stainless steel or exotic materials
for process equipment
5. Requires flue gas reheat for plume buoyancy
6. Requires an expensive gas-phase oxidant
148
-------
USE INDUSTRIES PROCESS - WET OXIDATION-ABSORPTION (NOX)
Process Description and Principles^of Operation (5)
The Ube Industries wet NOX removal process was originally developed to
treat the tail gas from HNO^ plants. The absorption of NOX is based on the
injection of an N02~rlch stream to obtain an equimolar NO:N02 ratio in the
flue gas. Although this system may be feasible for an HNO^ plant, its
ability to treat power plant stack gas is highly questionable primarily
because there is no source of N02~rich gas readily available at a power plant,
Since Ube has not suggested any modified method for treating flue gas, this
process will not be evaluated further.
149
-------
LEAVE THIS PAGE BLANK
150
A
-------
DRY NQX REMOVAL PROCESSES
INTRODUCTION TO THE DRY FLUE GAS DENITRIFICATION PROCESSES
Various dry flue gas denitrlf icatlon methods have been developed because
of the need for integration of NOX removal with existing S02 removal systems
and the high cost and the adverse environmental consequences from the liquid
and solid waste products of wet denitrif icatlon processes. The major modes
of dry NOX removal which have been studied include (1) catalytic decomposi-
tion, (2) selective catalytic reduction (SCR), (3) nonselective catalytic
reduction, (4) selective noncatalytic reduction, (5) adsorption, and (6)
electron beam radiation.
Catalytic decomposition would appear to be the most attractive means of
eliminating NOX from flue gas. By this method, NO (90-951 of the total NOX)
is converted to N£ and Q£ in the presence of a catalyst as follows:
catalyst
1/2°2(g) (195)
Unfortunately, the reaction rates and NO removal efficiency have been quite
low on tests performed to date with known catalysts. Thus, this method is
generally considered impractical at present for flue gas denitrification
(1, 2, 34, 72) and therefore has not been included in this study.
One of the most promising means of NOX removal for combustion flue gas
involves SCR. This process uses NH3 to selectively reduce NOX to N2 an<*
H20 in the presence of a catalyst.
Another catalytic reduction process consists of nonselective NOX
reduction. The major reductants used are hydrocarbons (such as CIfy) , CO,
and R2* I*1 this system ample reductant must be used to reduce all the
oxldant components of the flue gas, which for a typical coal-fired flue gas
are primarily 02, SOX, and NOX.
Selective noncatalytic reduction of NOX incorporates a gas-phase homo-
geneous reaction of a selective reductant (such as Wiy) converting NOX to $2-
This mode of NOX removal differs from SCR in that much higher temperatures
are used and the need for a catalyst is obviated.
NOX adsorption has been studied in two different areas. For HNO-j plant
tail gas, which is free of SOX and large quantities of H20, NOX may be
adsorbed by molecular sieves, gelatinous silica, etc. The adsorbed NOX is
then released by heating and returned to the acid plant. This method,
151
Preceding page blank
-------
however, is generally not applicable for combustion flue gas containing SOX,
H£0, and dust; therefore, processes of this type are not included in this
study. Another process using a special activated char (developed for SOX
removal) is suitable for NOX removal from combustion flue gas containing
some particulates. The adsorbed NOX is converted to N£ in a high-temperature
char-regeneration vessel. One process of this type is included in this report,
A completely novel process for simultaneous S02~NOX removal from flue
gas incorporates flue gas bombardment with electron beam radiation.
Each of these types of dry NOX removal processes with the exception of
the catalytic decomposition type is discussed in the following section.
Table 16 summarizes a comparison of the different types of NOX removal. Also,
each of the denitrification processes included in this study is listed by
type in the following section. A detailed explanation of each of these dry
processes is presented in the final portion of this chapter.
SCR
The SCR processes take advantage of the reduction selectivity of NH-j for
NOX. The reactions of NEU with NOX are usually given in the process descrip-
tions as follows:
4NH3(g) + 6NO(g) -> 5N2(g) + 6H20(g) (196)
8NH3(g) + 6N02(g) -> 7N2(g) + 12H20 (197)
In most instances, however, the reactions may be better represented accord-
ing to the following:
4m3(g) + 4N°(g) + °2(g) - 4N2(g) + 6H2°(g) <198>
4NH3(g) + 2N02(g) + 02(g) -> 3N2(g) + 6H20(g) (199)
The above may more accurately express the reactions because several studies
have demonstrated that the presence of at least some (>2 improves the NOX
reduction (2, 26, 34, 72, 84).
There are primarily two basic classifications of catalyst used: metal
and C based. The metal catalyst may be subdivided into noble and nonnoble
metal catalysts. Much development has been completed on both the noble and
nonnoble catalysts, but the noble catalysts are generally considered
impractical for the selective reduction of NOX with flue gas containing SOX
(72, 88). The noble catalysts are easily poisoned by S compounds and are
more expensive. Nonnoble metal oxides are the catalysts used in most of the
processes reported herein. Although less efficient and requiring higher
temperatures (300~400°C) than noble catalysts, they are less expensive, more
resistant to SO^ poisoning, and do not promote l^O production as has been
reported for some noble catalysts (88) .
152
A
-------
TABLE 16.
COMPARISON OF DRY NOX REMOVAL PROCESSES
Ul
Process characteristics3
Simultaneous S02~NOX removal
Achieves moderate S02 removal
(>85%)
Achieves high NOX removal (>90%)
Operating conditions
Produces waste stream
Uses NH3
Forms NH4HS04
Operates with sensitivity to
particulates
Produces marketable byproduct
Current development status
Tested on coal-fired flue gas
Tested on pilot plant or larger
scale
Selective
catalytic
reduction
_b
-
X
-
X
X
X
_b
_b
X
Dry NOX
Nonselective
catalytic
reduction
X
X
X
-
-
-
-
X
-
_
removal process
Selective
noncatalytic
reduction
-
-
-
X
X
X
-
-
X
type
Adsorption
X
X
-
X
-
-
X
X
X
X
Radiation
X
X
X
X
-
-
-
-
-
X
a. An "X" indicates the process has
b. Depends on process.
this characteristic.
-------
The C-based catalysts are used for simultaneous NOX and SC>2 removal.
C is used commercially as an S0£ adsorbent and activated C's have been
developed with special structures or metallic constituents for aiding the
NOX and NH^ reduction reaction. Operating temperatures, averaging 2QO-250°C
with these C-based catalysts, are usually lower than the operating tempera-
tures with the base-metal catalysts. The lower temperatures provide in-
creased 862 adsorbing capability, though higher temperatures favor greater
NOX removal efficiency.
Several problems are associated with the use of SCR processes in com-
bustion flue gases. As already mentioned concerning noble catalysts, SOX
poisoning of the catalyst is a problem. SOX usually attacks the catalyst
carrier. For example, SOX, especially SOj, tends to react with A^O-j to
form an Al2 (80^)3, thus, decreasing the available surface area and catalyst
activity. Base-metal oxide catalysts also react with SOX to varying extents;
however, these base-metal sulfates are still reactive.
Other problems are the decrease In catalyst activity and the Increase
in pressure drop which are created from particulate pluggage of the reactor.
Methods used to minimize these problems are parallel passage reactors, moving-
bed reactors, regeneration steps, and specially shaped catalysts.
The formation of NH^HSO^ presents another concern with the use of SCR
processes, It is usually formed downstream from the reduction reactor as
the flue gases are cooled in heat exchange equipment according to the
following reaction:
NEL, , 4- SO,,, , + HLO, . -»• NH.HSO, n , (200)
3(g) 3(g) 2 (g) 4 4(1)
The conditions at which it is formed are shown in Figure 28. For example,
NH4HS04 would form at 210°C or below with 10 ppm each of 803 and NH3.
NlfyHSO^ is corrosive and also interferes with heat transfer in the heat
exchange equipment. At even lower temperatures it may solidify and form
NJfyHSO^. When these products do form, occasional steam blowing or washing
out with H20 are needed for their removal.
Another important factor which must be considered is the availability
and cost of NH3 for NOX reduction systems. It has been estimated, based
on a 100 ppra NOX regulation for a coal-fired boiler in 1985, that a 30-50%
increase in HHg production capacity in the U.S. may be necessary for utility
boilers alone |108) . This large increase in production demand and future
cost increases for NHj will significantly affect the NH3 availability and
economics for NOX removal technologies using NH3.
The companies which have investigated or are developing SCR processes
which are included in this report using metal or similar catalysts are the
Environlcs, Inc. ••- Eneron Corporation; Exxon Research and Engineering
Company; Hitachi, Ltd.; Hitachi Zosen; JGC Corporation; Kobe Steel, Ltd.;
Kurabo Industries, Ltd.; Mitsubishi Kakokl Kaisha, Ltd.; Mitsubishi Heavy
Industries, Ltd.; Mitsubishi Petrochemical Company, Ltd. (only catalysts
are offered); Mitsui Engineering and Shipbuilding Company; Mitsui Toatau
154
A
-------
1000
100
Q_
0.
#*
to
x
z
10
10 100
S03tPPM
1000
Figure 28. Temperatures Below Which NH4HS04 Forms (101)
-------
Chemicals, Inc.; Sumitomo Chemical Company, Inc.; Sumitomo Heavy Industries,
Ltd. (N0x-only removal process); Ube Industries, Ltd.; and Universal Oil
Products (both the simultaneous NOx-SC-2 removal and the N0x-only removal
processes). Other companies with SCR processes which are not described in
this report include Asahi Glass Company (2) (no information obtained) and
Nippon Kokan (13, 14). (Personal communications with Nippon Kokan indicated
development was just beginning with no information presently available for
release.) Companies developing SCR processes incorporating C-based
catalysts are Sumitomo Heavy Industries, Ltd.; Takeda Chemical Industries,
Ltd.; and Unitika, Ltd.
General advantages and disadvantages for SCR processes are indicated
in Table 17. A summary description of each of the SCI processes is listed
in Table 18.
TABLE 17. ADVANTAGES AND DISADVANTAGES OF
SCR PROCESSES
Advantage
Disadvantage
Achieves excellent NOX removal
(usually 90% or greater)
Creates no waste byproducts in most
processes
Demands lower capital investment
and revenue requirements than wet
processes
Uses gas-phase reaction chemistry
and, therefore, requires less com-
plex operation steps than wet
processes
Requires heating of flue gas to
attain or control reaction tempera-
ture (however , development underway
on several processes to locate
denitrification such that reheat
may not be required)
Emits NHj or forms
possibly
Operates with sensitivity to
particulates
Requires some difficulty to retrofit
Nonaelectlye CatalyticReduction
With nonselective catalytic reduction, the reducing gas reacts with the
oxldant components (primarily excess 0£) present in the flue gas. For a
typical coal-fired flue gas, sufficient reductant must be added for the
reduction of 02, SOX, and NOX. Hydrocarbons, (I.e., CH^, CO, and H2 are
the major reductants used. A typical overall reaction using CO as the re-
ductant is as follows:
m, ^ + aco
(g)
+ excess 0
„, ,
2(g)
3C0
0, v
2(g)
(201)
156
A
-------
TABLE 18. PROCESS CONDITIONS AND ECONOMICS
FOR DRY SCR NO* REMOVAL PROCESSES
Seve 1 opraent
utr.tun
COGioany (MM e^ulv)
Eneron
Bx)t(5nd
Hitachi ttd.
Hitachi Zo»«n
JCC Corp,
Kobe
Kurabo
KureVia
Mitsubishi
Heavy Ind,
Mitsubishi
KK
MitdublBhi
Petro. *
Mitsui 1 & S
Mlteul Toatsu
Sumitomo
Chemical
Sumitomo
Heavy Ind,
Sumitomo
Heavy Ind,
Take da
Ube
Unitika
Unltika
UOP
HOP
Pilot
Bench
Consa
e™e
Proto
Bench
Proto
Pilot
Pilot
Pilot
-
Co™
Proto
Cofsm
Bench
Pilot
Pilot
Pilot
Bench
Pilot
Proto™
ProtoP
"•'>
(0.003)
(170)
(275)
<23)
(0,3)
(10)
(1,6)
a. 3)
(4.7)
(67)
(30)
(100)
<0.5)
(3,3)
(3.3)
(3.3)
(0.07)
<1.5)
(40)
(40)
Removal
^jsjjyiciencyj^j^ React ion
NO^E S02 Catalyst tefflj>,,°C
65 (oil)
B5 (gat)
70-95
>90
>90
>«
90
>90
90
90
?90
95
90
>90
90
>90
85-90
90
90
>90
>90
>BO
>90
-b 0.03"/. ft 280-305
by wt • 255
90-95 - 315-370
-b - 300-400
-b - 300-400
-b - 380-420
-*> Base metal 350-400
-b CuOg 350-400
-b - . ' 150
-b - 350-400
.- Fe20J"H20 400-450
-b Titanate 300-600
-b Ba«e aetal 350
-b - 350-400
-b Baae metal 300-350
-b Metal oxido 270-370
95 Act. carbon 200-230
80 Act. carbon 210-250
-b - 350-400
-b - 320-410
>90 Act, carton 200-250
90 CuO/CuSO^ 400
-b CUS04 400
Space
velocity,
hr"1
35,000
50,000
2,000-5,000
10,000-20,000
4,000-6,000*
-
10,000
7,000-10,000
5,000
4,000-10,000h
3,000-5,000
10,000
5,000
3,000-10,000
10,000
5,000
1,000-1,700
1,000-3,000
10,000
10,000-15,000
1,000
7,000-10,000
4,000-8,000
Mol KHj
per mo! By-
.. _ VOX producti
2.0-2.5
0,7-1.0
0.9-1.1
0,8-1.2
1.1-1.3
1.0-1.1
0.9-1.0
1
1
1.0-1,3
1
1
1.0-1,2
1
I
2,0-2.5
0.7-1.2*
1.1-1.3
1,0-1,1
ffl
1.0-1.2
1.0-1,2
None
Nome
None
Hone
None
None
None
None
None.
None
None
None
None
None
None
H2S04
,1
None
-
H2S04
-
Hone
Reported
capital
inveatstent ,
S/WJ*
• 11'
-
45
16
27
12-21
35
-
-
36
J>
10
49
80J
30
46
90
-
24
60
131°
31°
Reported
mllls/kWhtt'
0.2C
-
-
1.5
-
1.3-1.7
1.4
-
-
1.8
_b
-
1,6
1.2-1
2.3
5.9
0.65
-
2,1
6.8
5.0°
1.4°
a. Unless otherwlBC noted, coat baaed on 19?6 dollars and Japanese location; see each detailed proceas description for additional ba&es.
b. Not applicable,
c. Based on 197$ cost and U.S. location.
d, Exxon has studied both slmultanaoul NO^-SOJ removal and N0x-r,n1y removal.
e. Being tested on 0,07-MB «d.ulv coal-£tr»d unit.
f, Area velocity is more appropriate terminology and ia 7-10 Km /hr/» ; vlth a catalyst surface area to voluise ratio of 550-660 m /m , s
S,V. of about 4,000-6,000/hr la obtained,
H. Kurabo nao newly devalcped iron-baaed eatalyat which is more suitable for coal-fired flue gas.
h. Ba**d on Intertaittent aoving catalyst bad on saffiidirty gaa: 5.V. is 1,600-2,300/br for parallel paeaage reactor witn dirty gas.
i. Offering catalyst only.
.], Capital Investment it based on treatment of dirty ga> uaing an ESP, The revenue requirement ii In 1973 dollars and based an treatment
of clean ga« at Che Migaahi Nlhon Methanol plant,
k. Vhit ta aol HHj/aol (NOy + S02),
1. CaS04 and Hll3 ia produced with waahlng mthod ol regeneration; H^SOi, is produced with heating method of regeneration,
ro, Moll JWj - (O.B) n»l« NO + (0.3-0.5) noli SO*.
n, UO? has (Htrfomed taita vith coal-{tred flue gas on 0.6-M9 equlv pilot plant in FGD node,
o. Those costs atra b***d OTL 1977 dollar* and U.S. location*.
p. H0x-only removal hai ba*n touted at the i«ne 40-f*J equiv treatment facility «hich noToally operate) vith simultaneous Sox and NOX
157
-------
Most of the development work In this area has been done on HNOj plant tall
gas and the exhaust from Internal combustion engines. Catalysts for these
cases are inappropriate for power plant use and nonnoble metal catalysts
which are efficient may require operating temperatures of at least 500°C
(930°F) (72). A number of nonselective reduction catalysts promote the
formation of H2S and carbonyl sulfide (COS) which are toxic. If CO is used
as the reductant, toxic metal carbonyls may be formed by the reaction of CO
with the active metal portion of the catalyst, walls of the reactor, or
trace metals in the flue gas. Of course, there is also the potential
problem of unreacted CO emissions.
The only nonselective catalytic reduction process contained within
this report is the Ralph M. Parsons Company process. Table 19 shows the
general advantages and disadvantages of nonselective catalytic reduction
systems. Process operating conditions and economics for The Ralph M. Parsons
process are included in Table 20.
TABLE 19. ADVANTAGES AND DISADVANTAGES OF
NONSELECTIVE CATALYTIC REDUCTION PROCESSES
Advantage
Disadvantage
Removes NOX and SOX
Produces S as a byproduct
Creates no waste byproducts
Demands lower capital investment
and revenue requirements than wet
processes
Uses gas-phase reaction chemistry
and, therefore, requires less com-
plex operation steps than wet
processes
Uses large amount of reductant to
reduce all oxidants in the flue gas
(02, SOX, NOX)
Requires greater development than
other methods
Increases potential for toxic gas
formation in flue gas
Increases potential for corrosion
Requires reheat
Operates with sensitivity to
particulates
Requires some difficulty to retrofit
Selective Noncatalytlc Reduction
The selective noncatalytic reduction method simply combines a reductant
(for example, NHj) with the NOX in a flue gas at the appropriate temperature
without any catalyst to convert the NO to N2 and H20. Though a complex-free
radical chain mechanism is involved, the overall reaction may be simply
expressed as follows:
158
A
-------
TABLE 20. PROCESS CONDITIONS AND ECONOMICS FOR DRY NO,, REMOVAL PROCESSES OTHER THAN SCR TYPE
Item
Development status
Size, MW equiv
NOX removal, %
S02 removal, %
Catalyst or adsorbent
Reaction temperature, °C
Space velocity, hr~^
NH3:NOX mol ratio
Byproducts
Reported capital
investment, $/kW
Reported revenue
requirement, mills/kWh
Nonselective
catalytic
reduction
(Ralph M. Parsons)
Not tested
-
-
-
_b
-
-
_a
S
,1
21. 7d
j
1.0d
Process
Selective
type (company)
noncatalytic
reduction
(Exxon)
Commercial
53
60-70
_a
None
705-760
_a
3.0-4.0
None
-
0.6-1.48
Adsorption
(Foster Wheeler)
Prototype
20
40-60
80-95
C
120-150
_a
_a
S
40-906
1.0-2.36
Radiation
(Ebara-JAERI)
Pilot plant
1.0
90
80
None
no
_a
_a
_c
f
l,000r
—
a. Not applicable.
b. Proprietary.
c. Multicomponent material, composition unknown.
d. Based on 1972 dollars
e. Based on 0.9-4.3% S in
f. Based on 1976 dollars
and U.S. location.
fuel and 1977 dollars with U.S.
location.
and Japanese location.
g. Based on $0.07-$0. 15/MBtu; assumed 9,000
with U.S. location.
Btu/kWh heat
rate for coal; 1976
costs
-------
4NH
-, ,
3(g)
+ 4NO
, .
(g)
+ 0
,,, ,
2(g)
0, .
2(g)
, ,
(g)
(202)
and 02 also combine with each other to form NO but, at the required
temperature, reaction (202) is favored and substantial NO reduction can. be
achieved. The optimum temperature needed when only NH3 is added is about
1000°C (~1850°F) but the optimum temperature is reduced to about 730°C
(1350°F) by injecting H2, either simultaneously with NH3 or Just downstream
of the NH3 introduction site.
One of the problems with this method is that the favorable temperature
range is very narrow for optimum NOX reduction. In power plant operations
with a varying load and subsequent flue gas temperature changes, the NHj:H2
ratio must be corrected to match the changing temperature. If an imbalance
between NH3:H2 ratio and flue gas temperature exists, either of two adverse
consequences may result. If the temperature is too high, NH3 will be con-
verted to form additional NO. If the temperature is too low, unreacted NHj
and NO may exit in the flue gas. The use of H2 may present problems for
power plant utilization either in the purchasing or generating of H^.
The safety aspect of applying H£ in a power plant must be considered, also.
The potential for NH4HS04 formation is relevant here as in SCR processes.
Also, the future NH3 availability and cost are very significant aspects
which must be considered as mentioned under the SCR section.
Exxon's Thermal Denox process is incorporated in the detailed section
of this report as a selective noncatalytic reduction method. The general
advantages.and disadvantages of selective noncatalytic reduction are given
in Table 21. Process operating conditions and economics for Exxon's Thermal
Denox system are shown in Table 20.
TABLE 21. ADVANTAGES AND DISADVANTAGES OF
SELECTIVE NONCATALYTIC REDUCTION PROCESSES
Advantage
Disadvantage
Requires no reheat
Requires no catalyst
Demands one of the least capital
investments of any NOX removal
method
Creates no waste byproducts
Uses homogeneous, gas-phase
reaction and, thus, requires least
complicated operation steps of
any process
Achieves only about 60-701 NOX
removal
Needs large amount of reductant
(NH3); NH3:NOX mol ratio >3:1
Operates with narrow optimum
temperature range with potential
for emission of NHj and NO during
boiler load variations
160
A
-------
Adsorption.
A system was developed using a special activated char with exceptionally
good S02 adsorption capability. In pilot-plant tests, it was detected that
this system also removed NOX. The NOX is adsorbed in the char, though the
mechanism is still unknown. The adsorbed NOX is changed to N2 as the char is
subjected to a temperature of about 650°C (1200°F) In a regeneration vessel.
The N2 exiting the regeneration section eventually reenters the adsorber and
passes through to the stack. The major problem with this system is the low
NOX removal capability (about 40-60%).
Bergbau-Forschung and the Foster Wheeler Energy Corporation developed
this system. The general advantages and disadvantages are listed in Table 22,
while Table 20 lists the process operating conditions and economics for the
Foster Wheeler process.
TABLE 22. ADVANTAGES AND DISADVANTAGES OF ADSORPTION PROCESS
Advantage . Disadvantage
Removes NOX and SOX Obtains only 40-60% NOX removal
Requires no reheat Requires waste disposal of ash
Produces S as a byproduct Uses more complex operation steps
than other dry processes
Demands lower capital investment
and revenue requirements than most Requires hot solids handling
wet processes
Electron Beam RadlatIon
A newly developed method of NO removal involves treating the flue gas
with electron beam radiation, NOX and SOX are both removed and a byproduct
comprised of S, N, 0, and H with unknown chemical composition is formed.
The major drawbacks at present are the high cost and waste product handling.
This method is included in this report under the Ebara Manufacturing
Company and Japan Atomic Energy Research Institute (Ebara-JAERI) process.
The advantages and disadvantages are displayed in Table 23. Table 20 in-
cludes a summary description of the Ebara-JAERI process.
161
-------
TABLE 23. ADVANTAGES AM) DISADVANTAGES OF RADIATION PROCESS
Advantage ; ; Disadvantage
Removes NOX and SOX (nay achieve Demands large capital investment
>90% NOX removal)
Needs development of waste product
Requires only electricity treatment
Uses less complex operation steps
than wet processes
162
A
-------
EBARA-JAERI PROCESS - DRY, ELECTRON BEAM RADIATION (NOX-SOX)
(2, 4, 5, 83, 105)
Ebara Manufacturing Company and Japan Atomic Energy Research Institute
(JAERI — a government organization) jointly developed an electron beam radia-
tion process for simultaneously removing S02 and NOX from flue gas. A flow
diagram of the process for a commercial power plant based on the process flow
used in bench-scale tests Is shown in Figure 29. The flue gas from a boiler,
after passing through the air heater, passes through a highly efficient
ESP for partlculate removal. The gas Is discharged from the ESP to the
reactor where it Is bombarded with an electron beam. Within the reactor,
a powder comprising S» N, 0, and H, and aa 12804 mist are generated. After
exiting the reactor, these byproducts are removed from the flue gas by
passage through another ESP and then the byproducts are treated. (This treat-
ment of byproduct is Included here as necessary for commercial operation;
however, it has not been tested at the pilot plant.) The gas continues to
the stack. During bench-scale tests a recycle stream is taken from the
reactor exit, passes through an ESP for removal of the byproduct formed in
the reactor, and fs feintroduced into the reactor. For a commercial plant, it
would be more practical to pass all the gas exiting the reactor through one
ESP and then take the recycle stream at the exit of this ESP and send it
back to the entrance of the reactor. This would eliminate one ESP.
As shown in Figure 30, at 110 C (230 F) and employing a flue gas con-
taining 80 ppm NOX and 600-900 ppm S02, an electron beam of 0.8 Mr ad (radi-
ation for 2 sec at an intensity of 0.431 Mrad/sec) removes 90% of the
NOX. An S02 removal of 80% is obtained with a 4 Mrad electron beam
(radiation for 10 sec at the same intensity). A lower temperature allowed
radiation to slightly increase the removal efficiency. The most probable
radiation rate is In the ICP-IO rad/sec range. The total rate is preferably
between 10-10 rad. The residence time of the gas in the reactor may vary
from 1-20 sec.
Various means of utilizing the byproduct formed in this NOX and 862
removal process have been studied on small laboratory— scale experiments. In
one method, the byproduct is heated to 90-120 C (194-248°F) to create an
anhydrous substance. This anhydrous material is decomposed to NOX and a
denitrogenated anhydrous material at 160-240 C (320-464 F) in the presence
of an inert gas (and/or ^2) and in the absence of Q^' The NOX is then con-
verted to N02 in the presence of air at less than 140 C (284 F) and the N02
is used for HN03 production. S02 is produced by heating the denitrogenated
anhydrous material to 240-270 C (464-518 F) in the presence of an inert gas.
This S02 may be used to form 12804.
A variation of the above process for handling the byproduct includes
heating the material to 165-240°C (329-464°F) in the presence of an inert
gas (excluding the use of N£) . N2> 02, and a denitrogenated material results
from the above operation. The N2 and 02 are released to the atmosphere.
The remaining substance is thermally decomposed to form S02 at 240-270 C
(464-518 F) in the presence of an inert gas and the S02 Is -used to produce
H2S04.
163
-------
ON
AIR
I I
ELECTRON
| BEAM |
ACCELERATOR
I I
^ STACK
Figure 29. Flow Diagram of Ebara-JAERI Electron Beam Radiation Process.
-------
100
0-4.3IX
Ln
b- 1.46
c- 8.61
d-4.31
I05 RAD/SEC
x I05 RAD / SEC
x 10s RAD/ SEC
x I05 RAD/SEC
3 4
TOTAL BEAM, I06 RAD
Figure 30. Effect of Beam Intensities on NOX and S02
Removal Efficiencies for Ebara-JAERI Process (5).
-------
A different byproduct treatment method combined the byproduct with CaO
in the form of a. paste. The paste is atearn-roasted at a pressure of 0,1-1.0
atmosphere and temperature of 250-500 C (482-932 F) to create a gas stream
of S02 and NQX and a deposit of CaSO^ and Ca(N03)2- The gas stream of S02
and NOX is recycled to the roasting section. By heating the residue of
CaS04 and Ca(N03)2 to 300-650°C (572-1202°P), NOX and HN03 are formed leaving
a CaS04 and CaO residue. The NOX Is oxidized at <140QC (284°F) to form N02
for HN03 production. By heating the CaSt>4 and CaO to >1300°C (2372 F) ,
S(>2 and CaO are manufactured. An alternative scheme separates the CaSOA and
CaO by heavy liquid separation means. In either case, the CaO is recycled
to the blending operation at the beginning of the process.
Another system which was investigated consisted of the injection of a
small quantity of H20 into the byproduct and creation of a paste by blending.
This paste was steam-roasted at 150-300°C (302-572°F) and 0.1-1.0 atmospheric
pressure with air inserted. This operation produced S02» NOX, H^jSO^j and
The S02 and NOX are recycled to the steam-roasting step. t^SO^ and
o are cooled with H20 to form a mixture of acids. Ebara states that this
mixture may either be separated into the components by concentration and
distillation or be employed for acid treatment of an ore or neutralization
of alkaline solution.
Status of Development
In 10 Nm-Vhr (0.003 MW equlv) capacity bench-scale experiments, oil-
fired flue gas was used with Inlet S02 levels ranging from 500-1500 ppm and
inlet NOX levels varying from 145-590 ppm. With various dosage rates, pres-
sures, and radiation sources, the S02 and NOX removal efficiencies ranged
from 36-68% and 63-100% respectively (83). The reactor was constructed of
stainless steel and the dimensions were 50 x 50 x 500 mm. The electron
beam accelerator was manufactured by Hitachi, Ltd., and was the Cockcroft-
Walton type. Ebara also operated a 1000 Nm^/hr (0.3 KW equiv) capacity,
oil-fired gas treatment unit. A pilot plant capable of treating 3000 Nm /hr
(1 MW equiv) of gas from an iron ore sintering plant was to be built at
Yawata Works, Nippon Steel Corporation in 1976. Future development should
include tests on a prototype-scale and establishment of a byproduct treat-
ment process.
Background of Proceas Developer
This process by Ebara-JAERI is still in the development stage and not
commercially available. Also, the accessibility of this process to the U.S.
market may not be as favorable as most processes since no contacts could
be found in the U.S.
Published Economic Data
•\
The estimated capital cost (5) for the 3000 Nm /hr (1 MW equiv)
capacity pilot plant to be constructed In 1976 is to exceed $1M. This
may not include any waste treatment facilities since none were tested in the
1000 Nm3/hr size unit. This cost exceeds $1000/kW which is quite high but
naturally will be reduced some with economy of scale. (Note; Proprietary
166
A
-------
information received recently from other sources indicates the capital invest-
ment could be an order of magnitude less than reported by Ando (5). This
lower capital investment may allow the process economics to be competitive
with other simultaneous NOX-S02 removal systems. The significant difference
in order of magnitude has not been resolved.) No operating costs are given
but the only major factor is the power consumption. However, it is stated
by Ebara-JAERI that the most economical method of operating this process may
be to remove the NOX from the flue gas by this electron beam process (where
lower intensity may be used) and then remove S02 by conventional FGD.
RawMaterial, Energy, and Operation Requirements
Mo raw materials are necessary for this electron beam radiation since
electricity is the only item claimed to be consumed. The power required for
only an electron beam accelerator for a 100,000 Nm-Vhr capacity unit (33 MW
equiv) Is estimated to be 1 MW. Therefore, the power usage for just the
accelerators, thereby excluding the ESP consumption, is 3.3% of the equiv-
alent total power output, The 1-MW size electron beam accelerator is possibly
as large as will be available. If so, a 1~MW accelerator will be required
for every 100,000 Nm3/hr of gas treated.
The manpower, technical support, and maintenance requirements are
unavailable.
Technical Considerations
The process appears to be simple and direct but process control capa-
bility may be complex involving control of the accelerator based on inlet
composition, temperature, and flow rate of flue gas. Figure 30 shows NOX
and S02 removal with a gas containing 80 ppm NOX and 600-900 ppm S02« The
sensitivity of NOX find SC>2 removal to inlet gas composition at two different
pressures is shown in Figure 31. It is stated that the flue gas must contain
at least 1% by vol of 02 for efficient removal of NOX and S02, and it is pre-
ferable for the gas to contain a quantity of water vapor greater than or
equal,to that of NOX. No estimate for comparative equipment size is available.
This process possibly could be applied to the retrofit of a plant with
or without an FGD unit. If an FGD unit is present, this process may be used
for only NOX removal and a less intense beam may be used as shown in Figure
30 than for simultaneous removal. However, the capital cost may be prohibi-
tive because the cost of the N0x-only system would be essentially the same
as for simultaneous NOX and S02 removal case. If there is no FGD and only
a cold ESP, the process would be used as described.
Since several acceleration and reactor units would be required for a
500-MW plant, for example, the major turndown capability would be the removal
of units from service as needed. The recycle streams around the reactor
would also provide some turndown control.
The only item known about materials of construction is that the bench-
scale reactor used in the initial tests was made of stainless steel, The
byproduct formed is corrosive to stainless steel when mixed with a small
amount of water.
167
-------
Oo
100
90
3? 80
£ 70
z
r^ 60
.
UJ
UJ
o:
50
40
2 30
20
10
0
0
2O
I
I
400
600
800
1000 1200
1400
1600
CONCENTRATION OF NOX OR S02 , PPM
Figure 31. Relationships Between Inlet Gas Concentrations, Pressure,
and NOX and SOX Removal Efficiencies for Ebara-JAERI Process (83).
-------
Environmental Considerations
The electron beam radiation process removes SC>2 in addition to NOX. As
noted in Figure 30, the beam intensity required for S02 removal is higher
than that for NOX only. In general, higher rates of irradiation are less
effective when used with gaseous materials. However, as shown in Figure 30,
more effective NOX and S02 removal are obtained with larger dosage rates in
this process. Within a certain range of beam intensities, an applicable
equation relating NOX removal to total dosage rate is as follows:
a ** 160b
where:
a = ppm NOX removed by electron beam radiation
b = total dose in id6 rad
Also, a higher pressure is more effective in removing the NOX and SC>2 than
a lower pressure as is pictured in Figure 31. Figure 32 displays the reactor
residence time difference required for only NOX removal compared to simul-
taneous NOX and S02 removal. Though the NOX can be almost completely removed
within 2 sec, 10% of the S02 still remains after 10 sec of residence time.
Waste disposal will be required for the byproducts exiting the reactor
and collected by the ESP. Whether any of the methods previously described
are applicable for power plants will have to be proven in larger scale tests,
The information obtained on the sensitivity of NOX and SC>2 removal efficiency
to various operating conditions is limited to that shown in Figures 30, 31,
and 32, It is unknown if there will be any interference from gas species or
the remaining particulates in coal-fired flue gas. The safety risks Involved
with this radiation process are also questionable.
Gritical_ JData.Gaps and Poorly JJnders^tood^ Phenomena
Included in the data gap is the maximum particulate loading allowable
in the flue gas entering the reactor. It also needs to be known if there
is any significant difference In removal efficiency for treating flue gas
at 110°C (230°F) as reported herein and treating the gas at 150 C (302 F),
the temperature which might be expected at the exit of the air heater on a
coal-fired power plant. The chemical principles of this process are not
given in detail.
The operating costs are not given but the consumption of electricity
will essentially compose the entire operating expense of the NOX and S02
removal section. The manpower, technical support, and maintenance require-
ments are not known, as is the materials of construction required except
for the reactor vessel. Information is also lacking on the size of the
reactor required for commercial-scale installations.
Handling of the byproducts formed will be necessary but the exact
method is not determined. The potential for work hazards has not been
ascertained.
169
-------
100
TEMPERATURE:
90~I20°C
INITIAL CONTENTS
SOz:600~900PPM
NOvsSOPPM
O
MEAN STAYING TIME OF GAS M THE REACTOR,
SEC.
Figure 32, The Relation Between the Staying Time of Gas in the
Reactor Chamber and the Removal of SG>2 or NOX for Ebara-JAERI process (83)
170
-------
Advantages andDigadvantages
The advantages and disadvantages are as follows :
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >90% NOX removal efficiency
3. Operates between 100 and 200 C, which may require only
negligible reheat
4. Does not require chemical raw materials
D is advant ages
1, Forms secondary source of pollution (powder composed of H, N, 0, and S)
2. Has not been tested on coal-fired flue gas
3. Requires somewhat clean (certain degree of particulate removal needed)
gas feed
4. Uses significant amounts of stainless steel or exotic materials for
process equipment
5. May require very high capital investment (see note on page 165)
171
-------
ENVIRONICS, INC. - ENERON CORPORATION - DRY, SCR OF NOX
Process Description and Principles ofOperation (67, 77, 78)
Environics» Inc., has been developing and has obtained a patent on an
NOX removal method which utilizes NH3 to reduce NOX to N£ and H£0 in the presence
of a catalyst. The work performed to date by Environics has been with gas-
and oil-fired flue gas. The process flow for a coal-fired flue gas is shown
in Figure 33. Since some problems were experienced with soot even with tests
on oil-fired flue gas, the flue gas exiting a coal-fired boiler first passes
through a "hot" ESP for particulate removal. NH3 is mixed with the flue gas
before entering the air heater-reactor vessel. The air heater contains the
catalyst and in this vessel the NOX is selectively converted to N£ and
by NH3 according to the following reactions.
6NO
, ,
(g)
a/ .
3(g)
0, v -f 6H-0, ,
2(g) 2 (g)
_. ,
2(g)
8NH
.., ,
3(g)
0/ >.
2(g)
12H,,0, ,
2 (g)
(203)
(204)
Following the reactor and air preheater the flue gas is sent to an FGD
unit and/or the stack.
The optimum operating conditions and NOX removal results based upon
pilot-plant operations with oil- and gas-fired flue gas are as follows '
Gas-f ired cpndltions
NO., inlet concentration
A,
NH3 inlet concentration
NHj exit concentration
Reaction temperature
Space velocity
Catalyst
NOX removal efficiency
_0il-fired conditions
NOX inlet concentration
NH3 inlet concentration
NH3 exit concentration
Reaction temperature
Space velocity
Catalyst
NOX removal efficiency
125 ppm
200 ppm
15 ppm
255 C (490°P)
50,000 hr-1 '
0.03% platinum by wt
85%
125 ppm
300 ppm
Not measured
282-304°C (540-580°F)
35,000 hr-1
0.03% platinum by wt
65%
Status of Development
Environics has performed tests for W)x removal on gas-fired flue gas
on a small scale with a furnace and on a laboratory pilot plant with a
boiler and a utility pilot plant. Studies on oil-fired flue gas have been
performed on the utility pilot plant and laboratory pilot-plant levels.
172
A
-------
FGD
AND/OR
STACK
AIR
ASSUMED NECESSARY BY AUTHOR FOR COAL-FIRED BOILER OPERATION
Figure 33. Flow Diagram of Environics, Inc. - Eneron Corporation Process.
-------
Initially, laboratory experiments were performed employing (1) a variety of
catalysts and (2) bottled gas mixtures of NO iti N£ with flow rates from 100-
300 sft-Vhr. As shown In Table 24 an NOX removal efficiency of 90-95% could
be obtained employing copper oxide (CuO), platinum (Pt), or vanadlmn^based
catalyst. Since these tests indicated a greater activity with Pt, Environics
concentrated on Pt catalyst during the remaining tests. Later a. gas-fired
furnace was used to study the effects on NOX removal by temperature, space
velocity, NH3:NOX mol ratio, and effects of S(>2 on activity of various
catalysts.
TABLE 24. SUMMARY OP CATALYST PERFORMANCE:
NO {ABOUT 300 PPM) IN N2 (67)
Space velocity, Temperature
Catalyst
CuO/Cr203
Ba/CuO/Cr^O-j
Pt
CuO/CoO
V205/MoO
Pt
hr"1
10,500a
10,500a
17,500b
10,500a
I0,500a
40,000C
range, °F
600-700
600-700
600-700
600
600-650
500
removal
efficiency
95-98%
95-98%
15-521
99% •
75-931
90%
a. Catalyst
b. Catalyst
c. Catalyst
bed dimensions =
bed dimensions =
bed dimensions =
1" dia x 2.0"
i" dia x 1.2"
3/4" dia x 2.0
deep (pellets) .
deep (pellets) .
" deep (honeycomb) .
A gas-fired boiler was next used In a laboratory pilot-plant study on
catalyst life -and effects of various parameters on NOX removal. The catalyst-
life teat extended about 4200 hr.
Environics was awarded an EPA contract in January 1973 for demonstrating
an NOX removal process employing reduction with NH3 in the presence of a Pt
catalyst on a utility pilot-plant scale. The original scheme for treating
flue gas from a gas-fired boiler was expanded to contain oil-fired testing.
A 170-MW boiler of the Los Angeles Department of Water and Power was utilized
for this study, A 150,000 sft^/hr (1.5 MW equlv) slipstream from this boiler
was used for the pilot-plant operation and the platinum catalyst was incor-
porated in the air heater after removal of some of the heat transfer elements.
Gas-fired operation extended over 2000 hr and an average NOX removal of 852
was achieved. The oil-fired operation slightly exceeded 400 hr because of
problems with maintaining a sufficiently high temperature, maintaining a
constant gas flow, NH4HSQ4 deposits, and soot deposits on the catalyst.
The maximum flue gas rate attained was 110,000 sft^/hr and NOX removal only
averaged 50%.
As a result of the short-lived tests with the oil-fired boiler at the
utility pilot plant, further work was accomplished on a 12,500 sft^/hr ca-
pacity laboratory pilot plant. Number 2 dlesel fuel oil with 5% S was burned
174
-------
with a space velocity of 50,000 hr"1 and temperature range of 26Q-338°C
(500~64Q°F) employed. The NOX removal still averaged far less than during
gas-fired operation.
Background of Process Developer^
Environics actually began work in October 1972 on the contract for
utility pilot-plant operation awarded by EPA. However, the literature search
and small-scale studies on various catalysts was performed much earlier. In
addition, in February 1975, a patent was granted to Environics for an
NQjj. removal method from combustion flue gases and in the presence of S02-
Eneron Corporation is a spinoff organization of Environics. Eneron is con-
tinuing work on NOX removal. Environics is in receivership and not active
at this time. This process has not been applied commercially. This system
should be accessible to the U.S. market.
Published Economic Data
The estimated capital investment for NOX removal facilities on a 480-MW
size gas-fired boiler is reported by Environics to be about $ll/kW, The
revenue requirements are stated as about 0,2 mill/kwh. These are 1975 costs
with a U.S. location. The above figures (67) are based on the following
Environics assumptions.
1. 480-MW plant with 50 Msft3/hr of flue gas
2. Inlet gas NOX concentration of 125 ppm
3. Catalyst cost of $l»000/ft3
4, NHj cost $0.18/!b including tankage
5. Space velocity of 50,000 hr"1
6. Catalyst installed' in existing air heater
7. Catalyst bed replaced at 5-yr 'intervals with 25% of replacement
cost of catalyst counteracted by precious metal reclamation
8. Cost of capital is 10%
It should be noted that in this case there is no ESP required since it
is a gas-fired boiler and the catalyst is installed in the air heater so that
a separate reactor is not included.
Raw Material^ Energy, and Operation Requirements
Based on the same case and assumptions as mentioned above in the
economic section, the amount of NH3 required is approximately 5,5 short
tons/day. No additional electricity or other energy requirements are needed
for this gas-fired case since no ESP is used and the catalyst is located in
the air heater; however, these energy requirements would be necessary for
the assumed process on a coal-fired boiler shown in the flow diagram. It
is estimated for the 480-MW gas-fired treatment case that less than one
man full time is needed for maintenance.
175
-------
Technxcal Considera t ions
The Environics NO removal method is similar to other simple dry pro-
cesses. As concluded by Environics the process control could simply include
the injection of NH3 adjusted based on the amount of NOX or NH3, or total of
both, in the gas effluent.
The sensitivity of NOX removal to inlet gas composition was demonstrated
in the small-scale and pilot-plant tests. Plugging of the catalyst was
experienced with soot from oil-fired flue gas in the utility pilot plant
though Environics states that larger honeycomb support for catalyst, inter-
mittent cleaning with air or water, and countercurrent operation of the air
heater may reduce this plugging. However, this indicates the need for par-
ticulate removal prior to the catalyst for coal-fired operations which possess
far greater particulate matter. Also, different results were obtained on
the NOX removal efficiency for the small-scale tests than on the operation
of pilot plants in oil—fired tests. The small—scale test, using gas-fired
flue gas which contained 300 ppm of injected bottled S025 displayed an NOX
removal of 80% at 232°C (575°F) after 450 hr. For the oil-fired pilot
plants, NOX removal only averaged 50% and the maximum achieved was 65% at
288-316°C (55Q-600°F). The NOX removal efficiency in one test decreased
from 60 to 20% at 304°C (580°F) in only 60 hr. Environics concluded it
was possible that the Pt catalyst was poisoned by some constituent of oil-
fired flue gas, other than SC>2. However, there is significant evidence
from other sources (5, 72, 88) that S02 does reduce the activity of Pt
catalysts,
The size of the reaction area required should be much less with the
Pt catalyst than for most other dry processes. The space velocity of
35,000-50,000 hr~l is much greater than any other process and the temperature
of 282-304°C (540-580°F) is lower than most other systems.
If the placement of the catalyst within the air heater is practical
for coal-fired operation, the process would be very suitable for retrofit
application. This is doubtful though considering the particulate loading
of coal-fired flue gas and problems of soot associated with oil-fired flue
gas. This system may be applicable to retrofit after an FGD unit since
little S02 would then be in the gas to be treated. However, reheat would
be required.
With the catalyst placed within the air heater and designed for full
load, the turndown capability may not be affected because as the load decreases,
NOX foundation decreases, though not necessarily in the same proportion. How-
ever, the turndown capability appears more limited than most other processes.
Several problems and loss in removal efficiency occurred during pilot-plant
operation with boiler load fluctuations resulting in lower than desirable
reaction temperature.
There is apparently no special material of construction required with
the catalyst located in the air heater. The preferred catalyst support is
a ceramic honeycomb structure.
176
-------
jEiwl_ronment:_al_ Cpnsideratiotis
The NOX removal sensitivity to various operating conditions may be seen
in Figures 34-41. Figures 34, 38, and 40 show data from small-scale tests
with gas-fired flue gas. Data from laboratory pilot-plant operations are
presented in Figures 35, 39, and 41 for gas-fired flue gas and in Figure 37
for oil-fired flue gas. Figure 36 reveals results from the utility pilot
plant with gas-fired flue gas. NOX removal sensitivity to the temperature,
shown in Figure 34, indicates a peak NOX removal in the 204-260°C (400-500 F)
temperature range. Figure 35 demonstrates there is little difference in NOX
removal between 243°C (470°F) and 260°C (500°F) while the NH3 In the exiting
flue gas decreases rapidly as the temperature increases from 470-500 F. The
maximum NOX reduction occurs at 254~260°C (490-500°F) in Figure 36. With
the oil-fired flue gas. Figure 37 exhibits that NOX removal was optimum in
the 282-304°C (540-580°) range but It was significantly less at 321°C (610°P).
The effects of space velocity on NOX removal are indicated in Figures 38 and
39. Both reveal the expected decrease in NOX removal as the space velocity
Increases. As the amount of NHj injected increases from 100-500 ppm in
Figure 34, the NQX removal is shown to increase at any one temperature.
Figure 40 denotes the increase in NOX removal and the constancy of the NH-j
concentration in the exiting gas at 500 F, as the NH3 concentration, in the
inlet gas increases from 300-600 ppm. However, Figure 41 Implies that If NH3
concentration surpasses 250 ppm, the NH3 in the effluent rises rather rapidly
and there is no Increase in NOX removal. As mentioned previously figure 35
suggests the amount of unreacted NH3 is decreased as the temperature increases;
and it also displays the Increase in NOX removal with the increase in NH3:NOX
mol ratio. NOX removal increase with additional NH3 in the inlet 'gas Is
also demonstrated in Figure 37.
There is no removal of pollutants other than NOX. There is no signifi-
cant amount of waste or byproduct produced as with wet processes. However,
the formation of NH4HS04 may present a. problem which would require removal
by cleaning with compressed air or water (as stated by Environics).
As noted earlier the NQX removal results In testing with oil-fired
flue gas were less than expected. Because of difference in results on
small-scale tests with bottled SC>2 injection (80% NOX removal) and laboratory
pilot-plant experiments with oil-fired flue gas (50% average NOX removal),
Environics felt some gas species other than S02 was poisoning the Pt catalyst.
There are no apparent work hazards.
Critical. Data Gaps and Poorly Understood ^Phenomena
The effects on air heater performance and subsequent boiler operation
would appear questionable in placing the catalyst in the air heater. Also,
if the catalyst were placed in a separate vessel ahead of the air heater for
coal-fired operation, the high temperature (e.g., 375 C or 700 F) may result
in lower NOX removal efficiency (Figures 34 and 36).
177
-------
PT CATALYST
INLET NO = 23OPPM
INLET O2 = 2%
SPACE VELOCITY^
50,000 MR"1
100 PPM NH3
350
400
450 500 550
TEMPERATURE, »F
60O
650
Figure 34. Effect of Temperature on NOX Removal Efficiency for Eneron Process (67).
-------
100
80
o
z
S 60
> 40
O
UJ
a:
X
O
2 20
NOx REMOVAL EFFICIENCY
N03 IN EXIT GAS
PT, CATALYST
SPACE VELOCITY' 50,000 HR-1
/
\
100
500* F
1-0 2.0
NH3/NOX MOL RATIO
8O
a.
a.
en
60
x
UJ
40
10
x
z
20
3.0
Figure 35. Effect of Temperature and NH3:NOX Ratio on
Removal Efficiency for Eneron Process (67).
-------
1001
80
o
z
UJ
y 60
u.
u.
UJ
40
00
o
X
O
Z
20
INLET N0xs 125 PPM
INLET NHs = 2000 PPM
SPACE VELOCITY= 45,000 HR
I
•s.
a.
a.
£
g
oc
80 ui
o
z
X
40 UJ
to
x
z
470 480 490 500 510
TEMPERATURE,°F
520
530
-JO
540
x
O
Figure 36. Optimum Test Conditions-Utility Pilot Plant for Eneron Process (67).
-------
100
80
O
w 60
o
uT
u.
UJ
o
2
X
O
40
20
OIL FIRED LABORATORY
PILOT PLANT TEST CONDITIONS
PT, CATALYST (NEW)
SPACE VELOCITY* 5O.OOOHR*"1
0 200 400 600 800
INLET AMMONIA CONCENTRATION, PPM
IOOO
figure 37, Effect of Inlet NH3 Concentration on NOX Removal
Efficiency at Vmrious Temperatures for Eneron Process (67).
181
-------
IOO
o
UJ
o
80
uj 60
I
UJ
oo
to
40
20
25
PT^CATALYST
fNLET NO =225 PPM
INLET Oa =2.5%
NH3/NO = 2
O TEMPERATURE = 450°F
A TEMPERATURE=490 °F
I
50
75 100 125 150
SPACE VELOCITY (xlO 3),HR-'
175
200
Figure 38. Effect of Space Velocity on NOX Removal Efficiency for Eneron Process (67)
-------
00
100
80
>
u
z
UJ
u.
u.
UJ
o
5
UJ
tr
x
O
60
4O
20
25
PT CATALYST
INLET N0xs 2IOPPM
TEMPERATURE* 500°F
50
I
75 100 125
SPACE VELOCITY (xlO 3HR~')
150
175
Figure 39. Effect of Space Velocity on Removal Efficiency for Eneron Process (67).
-------
100
80
o
z
UJ
o
UJ
60
o
25 40
UJ
(E
X
O
20
PT, CATALYST
INLET NOX=240PPM
INLET 02 = 2%
SPACE VELOCITY* 50,000 HR" '
TEMPERATURE=500°F
r
1
I
E
Q.
UJ
.
20 UJ
ro
X
100 200 300 400 500 600
INLET NH3 CONCENTRATION, PPM
700
0
800
Figure 40. Effect of Inlet NH3 Concentration on XOX
Removal Efficiency and NH3 in Effluent for Eneron Process (67).
-------
100
75
o
UJ
o
u.
u.
50
S
X
O
z
25
100
D (NOX + NH3) EXIT CONCENTRATION _
A NH3 EXIT CONCENTRATION
PT, CATALYST
INLET NOX= 250PPM
SPACE VELOCITY= 50,000 HR'
TEMPERATURE=490°F
200
150
100
E
o.
O
S
o
u
X
UJ
50
200
300 400 500 600
NH3 INLET CONCENTRATION, PPM
700
800
Figure 41. Effect of NH3 on NOX Removal Efficiency for Eneron Process (67).
-------
Though the exiting gas NHg concentration is displayed from several tests
on gas-fired operations, it is not shown for the oil-fired te.sts but is assumed
to be quite high since the NOX removal was much less than during tests with
gas-fired flue gas. Also, the maximum particulate loading allowable for the
process with the ceramic honeycomb structure for catalyst support is not stated,
Advantages and Pi sadvan t a ges
The advantages and disadvantages are listed below.
Advantages
There are no apparent advantages now.
Disadvantages
1. Can only achieve a maxmimum NOX removal efficiency of <70%
2. Forms secondary source of pollution (>10 ppm NH3 in outlet gas)
3. Has not been tested on coal-fired flue gas
4. Requires clean (S02~ and particulate-free) gas feed
5. Incorporates questionable design features (placing catalyst in air
heater)
186
A
-------
EXXON THERMAL DENOX PROCESS - DRY, SELECTIVE NONCATALYTIC REDUCTION (NOX)
ProcessDescription and Principles_ of Operation (6, 8, 79, 80, 81, 82)
Exxon has developed a thermal denitrification process (Thermal Denox)
which selectively reduces NOX by reaction with NH3 without the use of a
catalyst, NH3 is injected directly into the boiler in the proximity of the
steam superheaters and reheaters where the temperature is suitable for the
reaction to occur. In a typical case, the temperature ranges from 1900 F
(1028°C) prior to the secondary superheaters to 890 F (477 C) near the exit
of the steam superheating section. A simplified flow diagram of this pro-
cess is shown in Figure 42. This gas-phase homogeneous reaction involves
a complex free-radical chain mechanism. However, the overall reaction may
be written as follows:
4NH3, , + 4NO. , + 0,,, , -»• 4N9, , + 6H,0, (205)
J(g) (g) 2(g) t(g) 2 (g)
The following competitive oxidation reaction also occurs.
4NH., . + 500/ , -»• 4NOr . + 6H00, (206)
3(g) 2(g) (g) 2 (g)
A balance exists between reactions (205) and (206); though, in the applicable
temperature range, a sizeable reduction of NO can be attained since reaction
(205) is favored.
The optimum temperature range required with only the injection of NHo is
1700-1900°F (927-1028°C). By injecting a readily oxidizable gas, such as H2
along with NHo, the optimum temperature range for NO reduction can be lowered
to about 1300-1400°! (704-760°C). An H2:NH3 ratio of <3 is preferred
since excessive amounts of H2 can decrease the selectivity for reduction of
NO by NH3. The % may be injected either simultaneously with the NH3 or at
one or more intervals downstream from the point of insertion of NH^.
Figures 43 and 44 show data from laboratory studies for NOX removal without
and with the use of H2 respectively.
A maximum of 70% NOX removal has been obtained for commercial oil-fired
boilers. An NH^NOjj mol ratio range of 3.0 to 4.0 was employed.
S tat us of _Deye 1 ppment
Initial laboratory experiments with this Exxon denitrification process
used simulated mixtures of gases with helium (He) as a carrier gas. The gas
mixture was sent through a quartz tube reactor located within an electric
furnace. Operating conditions within the reactor included a positive pressure
of 0.2 atmosphere, a temperature of 982°C (1800 F) and a 0.075 sec reaction
time. The gas exiting the reactor was then analyzed. It was determined that
above 0.5% 02 concentration,the 02 level had little effect upon NOX removal,
However, the NOX removal became insignificant in the absence of 02- The NOX
187
-------
1
1
BOILER
FLUE
GAS
i
AIR
HEATER
TO F6D
AND/OR STACK
T
AIR
Figure 42. 'Flow Diagram of Exxon Thermal Denox Process,
188
A
-------
350
300
350
a.
a.
O 200
g
o
Z
O
O
150
100
50
I I
H2 = 0%
02 = 2%
1
I20O I3OO 1400 I5OO 1600 1700 1800 19OO 2000
TEMPERATURE,°F
Figure 43. Effect of Temperature on NO Removal and
Concentration Without % Addition for Exxon Thermal Denox Process (82)
189
-------
350
300
250
i 200
1
ty
o
8
150
100
50
0
I I
02: 2%
Hg: ADDITION RATE
UNKNOWN
NO PPM
1000 1200 1400 I6OO
TEMPERATURE, °F
1800
2000
Figure 44. Effect of Temperature on NOX Removal and NHj
Concentrations with H2 Addition for Exxon Thermal Denox Process (82)
190
A
-------
removal at 0%, 0.5%, and 2.2% 02 levels under laboratory operating conditions
can be seen in Figure 45. It was also discovered that 1-2 mols of $20 are
formed by the reaction of NO with NH-j and 02 for every 100 mols of NO reduced.
This process has been demonstrated commercially at the Kawasaki refinery
of Tonen Petrochemical KK, Exxon's Japanese affiliate. Tests were performed
on an oil-fired boiler functioning up to a maximum of 140 MBtu/hr and a gas-
fired furnace operating at 500 MBtu/hr. The boiler was operated as a swing
boiler, replacing an out-of-service boiler, while the furnace functioned
continuously. Up to 70% denitrlfication was obtained on these units with an
NOX concentration in the inlet gas of 150 ppm and an NH^:NOX mol ratio in the
range of 3:1 to 4:1 (see Figure 46). In addition to tests at the Kawasaki
refinery, demonstrations have been performed in Japan on a municipal in-
cinerator, 430 ton/hr oil-fired utility boiler, and a large pipestill furnace.
Background of Process Developer
This Thermal Denox process was developed by Exxon Research and
Engineering Company, a subsidiary of Exxon Corporation. U.S. patent
3,900,554 covering this technology was granted to Exxon on August 19, 1975.
This process for NQX removal is being licensed by Exxon to nonaffiliated
companies. With the aid of Tonen Technology KK, Mitsui Petrochemical
Industries, Ltd., acquired the first license in Japan. The technology was
employed at its Chlba plant on a 120 ton/hr oil-fired boiler with >50% NOX
reduction attained.
This technology is readily available to the U.S. market.
Published Economic Data
The reported cost of Thermal Denox ranges from $0.07 to $0.15/MBtu fired
on a retrofit basis and this Includes the capital cost (82), This cost is
based on a 60% NOX reduction with an NH3:NOX ratio of slightly greater than
2:1. Savings may be realized if the process is incorporated into the
design of new equipment.
Raw Material, Energy, and Operation Requirements
The only raw materials needed for this process are NH-j and %. The use
of H£ for reducing the required temperature or for temperature compensation
during a boiler fluctuation may not be practical for power plants from a
safety aspect. It is estimated that 60 short tons/day of NHj is required to
reduce 70% of the NOX from the flue gas of a 500-MW plant containing 600 ppm
NOX with an NH3:NOX ratio of 3:1 employed. The specific energy, operating
manpower, maintenance, and technical support requirements have not been
released by Exxon but should be minimal as compared to other NOX removal
processes.
191
-------
2.2% 02
TEMPERATURE=982°C
REACTION TIME = 0.075 SEC
(N0)0= 1020 PPM
0.5% 02
0.6 0.8
(NH3)0/(NO)0
Figure 45. Effect of NH3.-NO Mol Ratio and Various
Levels on NO Reduction for Exxon Thermal Der.ox Process (81) .
-------
0
1234
MOL RATIO OF NH3 TO NOX
Figure 46, Effect of NH3:NOX Mol Ratio on NOX Removal
Efficiency During Commercial Tests for Exxon Thermal Denox Process (82)
193
-------
Techni ca1 Cons iderations
Although the Exxon Thermal Denox process appears to be very simple in
comparison to other dry and wet NOX removal methods, its application may be
very difficult. For pure NHo and for NH3 + H2 mixtures of any given ratio
there is a narrow temperature 'range in which optimum NOx reduction occurs,
For a power plant boiler operating with a varying load, the temperature of
the flue gas at any location will vary with the changing load. Hence, it
would appear necessary to adjust the NHj:H2 ratio to match this changing
temperature. If a severe mismatch between NH^:!^ ratio and flue gas
temperature occurs, either additional NO may be produced from the oxidation
of NH-j to NO at high temperatures or the NH3 may fail to react with the NOX
at low temperatures. Either occurence would increase the pollutants emitted
to the environment (79)".
This use of H2 may pose problems for power plant applications. For
some locations it may not be practical to purchase large quantities of H2
from an external vendor. While H2 may be generated via NH3 cracking
(2NH3 -> N£ + 3H2) or other methods, this added processing step would represent
an additional cost. Also, the question of safety in handling H2 in a power
plant environment would require careful examination (79).
No data are available which indicate the relationship between NOX removal
and inlet gas composition. However, Exxon states that the concentration of
SOX or particulate matter in the flue gas should not interfere significantly
with the NOX removal capability since this NOX removal method only involves
a gas-phase homogeneous reaction. It is assumed that this has been proven
with respect to the oil-fired tests by Exxon. However, there is apparently
no proof that the particulate levels in coal—fired flue gas do not adversely
affect the NOX reduction efficiency.
This method of NOX removal is most applicable to retrofit since it
would only involve modifications to the existing boiler. At the commercial
demonstration unit at the Kawasaki refinery, it has been indicated that
installation of this process did not obstruct or impede normal operations,
Since this process injects a gaseous mixture into the boiler flue gas,
the only turndown capability required is the gas flow rates. However, when
the amount of turndown causes the temperature to drop, the.NOX removal
efficiency will also decrease.
The method of NHj injection is considered proprietary by Exxon. It is
known that no unusual materials of construction are required for the
injection.
EnyIronment al Cons iderat i one
As noted previously, should the temperature vary from the narrow
temperature range required, an excessive amount of NHg and/or NO could be
discharged out the system. In addition, it is possible for NH4HSQ4 to be
194
A
-------
formed from the combination of NH3» SO-j, and H20 in the cooler regions
downstream of the reaction zone. This compound is capable of corroding as
well as fouling equipment. Laboratory data indicated no difference in
corrosion between presence of 45 ppm NH^ and the absence of NH3, No additional
corroding or fouling was noted during a lengthy run with NH3 injection.
This process does suppress the oxidation of CO to C(>2 though it does
not generate CO. So, if any CO is present when the W3.^ reduces the NO, it
will remain and exit in the outlet gas. Exxon states that this is not a
problem for oil—,, and gas-.opeyated systems since the CO 'is completely oxidized
to C02 before the gases arrive at the NH3 injection point. However, it is
unknown whether incomplete CO oxidation will pose a problem with coal-fired
units.
Another species of NOX which will exit with the treated gas is ^0. For
every 100 mols of NO reduced, 1-2 mo Is of ^0 are formed.
A potential work hazard is produced if H/? is used to reduce the
temperature required or control the NOX removal during periods of variable
temperature. The f lananability limits are 4-7% H2 for H£ air mixture at
atmospheric pressure (89). If small amounts of H.2 are used (e.g., l^NHj mol
ratios <3:1) the risk may not be great. Also, Exxon indicates that the §2
may be substantially diluted with inert gas such as spent steam. This
dilution may further decrease the H2 flammability hazard.
Critical Data Gaps and 1 Poorly^ Understoocl jPhengmena
Figure 44 shows the curve with an initial NH3:NOX ratio of about 1.5:1.
The results are not available for an NH3:NOX ratio of about 3:1, which is
a typical ratio' for 60—70% NQX removal in commercial application. Also,
the H:NH ratio used for this curve is not known.
Figures 43 and 44 indicate the relationships among NOX removal, NH3
emitted, and reaction temperatures at an initial NH3:NOX ratio of about 1.5:1
It would be desirable to know the amount of NH3 emitted at commercial
conditions of 60-70% NOX removal and ranges of NH3:NOX ratios of 3:1 to 4:1.
The nature and technology of the NH3 injection is proprietary informa-
tion. The effects of particulates from coal-fired flue gas are uncertain
though Exxon expects none.
The exact energy requirements are unknown. Also, the estimated labor,
maintenance, and technical support requirements are unavailable.
Ad van t ages and Disadvan tages
The advantages and disadvantages of the Exxon Thermal "Denox are as
follows:
195
-------
Advantages
1. Has been applied to flue gas from commercial oil-fired boilers
2, Operates with full particulate loadings (>7 gr/sft^)
(this is hased on Exxon's expectation that particulate levels
have no influence on NOX removal efficiency but this has not been
proven completely for coal-fired flue gas)
3. Requires no additional post-boiler processing equipment
Disadvantages
1. Can only achieve a maximum NOX removal efficiency of <70%
2, Forms secondary source of pollution (large amounts of NHj or NO
may be emitted)
3. Has not been tested on coal-fired flue gas
4. Incorporates design features which may present significant process
control problems
5. Requires a relatively large NH-^NOx'mol ratio for equivalent NOX
removal (NH3:NOX mol ratio of >3:1)
196
-------
EXXON RESEARCH AND ENGINEERING COMPANY - DRY, SCR (NQX-SOX)
Process Description and Principlesof Operation (53, 100, 114, 115)
Exxon has acquired two British patents for NOX removal from gas streams.
British patent 1,438,119 is concerned-with the removal of NOX and S02 from
flue gas streams while patent 1,438,120 deals with removal of NOX only from
flue gas streams.
In the simultaneous NOX and S0£ removal process, a metal oxide catalyst
combines with the SOn in the flue gas to form a sulfate. In the presence of
this sulfate form of catalyst, NH^ reduces the NOX to $2 as follows:
S0= .
6NO(g) + 4NH3(g) - 5N2(g) + 6H20(g) (207)
S0=
3N02(g) + 4HH3(g) -, 7/2N2(g) -, 6H20(g) (208)
The harmless N2 then exits with the flue gas. As the catalyst becomes
converted to sulfates, the SOX in the incoming gas Is no longer removed, and
is found in the effluent gas. Regeneration of the catalyst la then required.
In regeneration, a reducing gas mixed with steam is contacted with the
catalyst to convert the sulfate to a free metal and to release SOn. The
free metal form of catalyst will be changed to the oxide form upon Introduc-
tion of the flue gas. Some means of handling and utilizing the released S02
is required. The regeneration gas may range from 10-501 H2 and 50-90% steam,
but is preferably 40% H2 and 60% steam.
The flue gas temperature required for this process ranges from 315-480°C
(600-900°F). Temperatures of 315-370°C (600-700°F) may be expected for a
coal-fired electric power generating facility. The moat preferable NH3:NOX
mol ratio utilized ranges from 0.7-1.0; Exxon states that increases in NH^
usage above stoichiometric quantities do not result in significantly greater
NOX reduction. The space velocity would range from 2000-5000 hr~^ and is
limited by S02 removal rather than NOX removal.
The NO -only removal method is very similar to the simultaneous NO -S02
removal system. The major difference is that continued regeneration of the
catalyst in the sulfate form to release SO^ is not required. The other
operating parameters (i.e., temperature, space velocity, etc.) are similar,
Status ofDevelopment
The development Information contained within the patents are limited
to bench-scale experiments. For the simultaneous NO -S02 removal system,
the first test involved the use of coal-fired flue gas containing 2500-3000
ppm S02, 400-600 ppm NOX, 6.8% Q£ and small amounts of flyash, A total flue
gas and NH-j rate of 6.35 sft-Vmin with a space velocity of 2000 hr was
197
-------
utilized. A 52-in,-long and 3—in.-diameter reactor with a fixed bed of CuO
on &.l2®3 was employed and an electrical heating jacket was used to cover the
reactor. The catalyst was about 1.45% Cu by weight and the bed was 48 in.
deep and occupied about 0.2 ft^. The operation of this system incorporated
the following cycle:
1. 9-min NOX-S02 removal period
2. 1/2 min, 1,25 sft /min steam purge on reactor
3. 1—min regeneration period with 40% by vol H2 and 60% steani at
0.4 g-mol/min of H2
4. 1/2-onin purge of the reactor with N2 at a rate of 0.25 sft^/min
Some of the results obtained from these tests are shown in Table 25.
TABLE 25. RESULTS OF COAL-FIRED BENCH-SCALE TESTS (114)
ppm NO
in inlet gas
542
542
553
553.
Average
NH3:NO
mol ratio
1.0
1.47
3.2
3.5
Average 1
removal over
9-min period3
78.4
84.2
92.4
86.8
a. Samples taken at 1-min intervals.
The SC>2 removal was maintained at 90-95%.
A second bench-scale experiment with the simultaneous NOX-S02 removal
process employed a synthetic flue gas comprised by volume of 2700 ppm S02»
1200 ppm NO, 5% 02 and 0.6-2.0% H20. A 1-in.-diameter glass reactor con-
taining a fixed bed of CuO supported on A1203 spheres with 8% by wt of Cu
was used. The cycles utilized in this test were:
1. 30 min of NOX-S02 removal
2. 1-^min N2 purge period
3. 2-min regeneration period with a gas mixture of 86% by vol H2
and 14% by vol H20 vapor
4. l~min N2 purge
The gas temperature and space velocity used were 343 C (650 F) and
5000 hr respectively. Both the relationship of NO removal to NH-jiNO mol
ratio and the SOX removal efficiency are demonstrated also by the results
shown in Table 26.
198
A
-------
TABLE 26. RESULTS OF BENCH-SCALE TESTS WITH SYNTHETIC FLUE GAS (114)
% NOV removal
in
ppm NO
inlet gas
1,200
1,200
1,200
ppm S02
in inlet gas
2,700
2,700
2,700
NH3:NO
mol ratio
0.5
0.76
1.0
Entire 30
min
55.0
69.9
82.5
Last 23
min
56.3
73.3
97.5
% S02
removal
91.5
93.7
91.5
The NO removal is shown for both the entire 30-min NOX and S02 removal cycle
and for just the last 23 min. The entire 30-mln NO removal average is
always lower because during the first minute of the cycle, most of the
catalyst is not in the sulfate form to aid the reduction of NO.
For the N0x-only removal method, the bench-scale experiment was made
using flue gas comprised by volume of 990-1010 ppm NOX, 2700 ppm S02, and
3—5% 02« The laboratory reactor was 7/8 in. in diameter and contained a
3-in.-deep catalyst bed of CuO supported on alumina. The space velocity
employed was 5000 hr"""*-. The relationship of NOX removal to the gas
temperature and NH3:NOX mol ratio is depicted by the results of the bench-
scale investigation shown in Table 27.
TABLE 27. BENCH-SCALE RESULTS ON NOX-ONLY REMOVAL (115)
NH3?NOX Gas inlet % NOX
molratio_ temperature, °F converted
0.25
0.33
0.33
0.38
0.4
0.5
0.67
0.67
0.73
0.8
0.8
650
650
750
650
750
750
650
650
650
650
750
35
45
50
52
55
75
95
95
100
100
100
These NOX reduction methods have been developed only in the laboratory
and British patents have been obtained by Exxon on these processes. In
personal communications with Exxon, it was discovered, however, that Exxon
has no commercially available catalyst. Also, licensing of these processes
are not being actively supported in the U.S. Thus, no greater detail is
given to these Exxon processes.
199
-------
FOSTLR WHEELER - BERGBAU-FORSCHUNG PROCESS - DRY, ADSORPTION (NOX-SOX)
Process Description and Principles of Operation (27, 28, 29, 30, 42)
The Foster Wheeler - Bergbau— Forschung (FW-BF) dry adsorption process
removes NOX as well as SOX from boiler flue gas by adsorption on a special
'char. The adsorbed NOX is- reduced to harmless N2 which is returned to the
main flue gas and exits the stack. The SOX is reduced to elemental S which
is condensed as a byproduct.
The FW-BF process consists of three sections: adsorption, regeneration,
and off-gas treatment (RESOX) . A flow diagram of the process is shown in
Figure 47. Adsorption begins with flue gas, having passed from the boiler
through the air heater and a highly efficient ESP, entering the adsorber.
The adsorber is a vertical column with parallel louver beds containing a
special activated char made from coal. The char is probably in pellet form
with a diameter of 3/8 in. and a length between 3/8 and 5/8 in. Regenerated
char is constantly being fed to a tank at the top of the adsorber and dis-
charged through tubes to the individual char beds by gravity. In addition
to carrying the moving char, the louvers direct the flue gas to flow cross-
ways through the adsorption beds and out the adsorber to the stack. The SOX
and NOX are adsorbed on the char which gradually but continuously progresses
downward through the adsorber. As water vapor, 02? and S02 are adsorbed on
the activated char, S02 Is converted to H2S04 by the following reaction.
H2°(g) +S°2(g) + 1/2°2(g) *H2SOA(i)' (209)
This H2S04 is then reported to be held within the char pellet pore structure.
S03= is similarly adsorbed and, though NO is also adsorbed, the exact mech-
anism is still being studied both in the U.S. and in Germany. Particulate
matter accumulates on the char pellet surface. The char flow rate is a
function of the SOX content of the flue gas but is slow enough to allow low
char abrasion. Feeders at the bottom exit of the char bed regulate the char
flow. The temperature in the adsorber is usually 121-149°C (250-300°F) and
the optimum is 135-149°C (275-300°F) , Therefore, if flue gas from a boiler
exits the air heater at a temperature >150 C (300 F) , tempering may be nec-
essary to attain the optimum temperature. The pressure drop across the
adsorber ranges from 4-10 in. of
The saturated char exits the adsorber bottom onto the discharge feeders
and proceeds to regeneration by a system of conveyors and bucket elevators.
Screens, at the end of the conveyor and before the bucket elevators feeding
char to the regenerator, sift the majority of the flyash deposit from the
char pellets.
In the regenerator, the saturated char is blended with hot sand and
the mixture slowly moves by gravity through the regeneration vessel. When
the activated char reaches a temperature at or above 649 C (1200 F) the
adsorbed H2S04 reacts with the C to produce S02 as follows :
2H-SO, ,' . + C, , +CO,,, , + 2H_0, + 2SOr., % (210)
2 4(1) (s) 2(g) 2 (g) 2(g)
200
A
-------
TO FLUE GAS
ENTERING AIR
HEATER
AIR
Figure 47. Flow Diagram of Foster Wheeler-Bergbau Forschung Dry Adsorption Process.
-------
The adsorbed NOX is converted to N2 as follows;
2ND, , + Cf N -»• CCU, , + N0, . (211)
(g) (s) 2(g) 2(g)
The gas stream produced in the regenerator, which is 25 to 40% by wt
S02 and the remainder C02, H20, and $2» passes from the top of the regener-
ator to the off-gas treatment section (RESOX).
The sand is inert to the reactions and is used only as a heat transfer
material. A feeder at the bottom discharge on the regenerator controls the
flow of the sand-char mixture. This feeder discharges the sand-char onto a
vibrating screen deck for physical segregation of the char from the sand.
The char Is collected from the top of the screen, spray cooled to 104 C
(220 F), and recycled to the top of the adsorber. The sand passes through
the screens, flows to a fluldized-bed heater, and is subsequently discharged
and returned to the regenerator. Direct combustion of No, 2 fuel oil or
coal is the heat source for the heater. The flue gas created in this heater
Is used to preheat the combustion and fluidization air entering the fluidized-
bed heater. The flue gas, containing S02, is then introduced into the main
flue gas before the air heater for further heat recovery and to allow the
S02 produced to be recovered in the adsorber.
The remaining particulate not screened from the char pellets between
the adsorber and regenerator is dislodged by the mixing action of the char
and sand within the regenerator. This flyash enters the hot sand loop.
The . S02~rich gas from the regenerator is sent to the RESOX reactor
mining crushed cc
led coal, at 649-
S and 02 as follows;
.containing crushed coal. As the gas passes countercurrent to the flow of
crushed coal, at 649-8l6°C (1200-1500 F) the SO? is converted to elemental
S°2(g) - S(g) + °2(g) <212>
The resulting 02 combines with C in the coal to form C02 by the following
reaction:
C(s)+°2(gr C°2(g) <213>
The other constituents of the regenerator off-gas, such as N2 and CG2S
are inert in the RESOX reactor. A discharge feeder regulates the coal flow
through the reactor. The unconsumed•coal and resulting ash exit the reactor
and enter a vessel for cooling and subsequent disposal.
The gas exiting the RESOX reactor is .directed to the S condenser where
S vapor forms molten S. The liquid S can be handled in either of two ways.
One method is to maintain the S in liquid form for tank-truck shipment in
an insulated tank outfitted with steam colls. An alternative is to solidify
the liquid S in cooling pits for subsequent shipment as solid S. The gas
202
A
-------
exiting the condenser contains C025 H20, N£» and S not in elemental form.
This gas is returned to the boiler, where the nonelemental S is oxidized to
S02, and becomes a portion of the flue gas entering the adsorber,
The average NOX removal efficiency with this FW-BF dry adsorption
process is 40-60% and the SOX removal efficiency averages 80-95%. The addi-
tional particulate removal efficiency on a gae having >99% of partictilates
removed by ESP is 90-95%.
The present information on conditions and flow rates is limited to data
from brief tests at a prototype plant at the Scholz Steam Plant of the Gulf
Power Company at Sneads, Florida, which had a 20-MW-equiv adsorber section and
40-MW-equiv regeneration and RESOX sections (see Status of Development).
This information is as follows:
Item Average or range
'Qasflow- to adsorber 174,000 Ib/hr
Charflow rate through adsorber 5,300 Ib/hr
Gas residence time in adsorber 13 sec
Char dwell time in adsorber 96 hr
Charflow rate to r-egener-atQj: 5,300 Ib/hr
Sandflow rate to regenerator 180,000 Ib/hr
Sand to char ratio (vol) 13.9:1
Coalflow to RESOX reactor 100-250 Ib/hr
Gasflow to RESOX reactor 750-2,000 Ib/hr
Status of Development
BF has been involved in developing a dry adsorption process for desul-
furization of flue gas since 1965. An activated char manufactured easily
and inexpensively from bituminous coal was made which possessed the following
attributes: (1) excellent S02 adsorption capability, (2) good physical
strength, (3) low pressure drop, and (4) high ignition point. Experience on
bench-scale tests and semlpilot-plant tests led to extensive work on a
105,000 sft3/hr capacity pilot plant at Welheim, Germany. The flue gas for
the pilot plant was a slipstream from a coal-fired steam generator. This
plant was operated over 2 yr and with one continuous period of 6000 hr. It
was discovered that this system also removed NOX and particulate matter as
well as SOX. NOX removal efficiency from small-scale pilot plants in
Germany averaged 40-60%.
FW received full licensing rights to supply this process under BF skill
and patents. Tests on small pilot plants in Livingston, New Jersey, and
in Japan also, averaged 40-60% NOX removal. The construction of.-a 20r-MW proto-
type unit of this adsorption process for treating coal-fired flue gas was
completed In May 1975 at the Scholz plant as part of a technology evaluation
program for Southern Company (holding company for several electric utilities),
At present, testing has been short-lived because of equipment problems. The
NOX removal averaged only 25-30% but problems were experienced with NOX
monitoring equipment. The removal of SOX averaged >96%.
203
-------
Background of Process Developer
BF is the research Institute for the German bituminous coal industry.
As noted earlier, BF has been performing research on this adsorption process
for many years. In addition to the pilot plants noted previously, BF is
operating a 150,000 Nm^/hr capacity (50 MW equiv) desulfurization demonstra-
tion unit 'at Kellermann Power Station owned by Steinkohlen-Elektrizitat AG
(STEAG) in Lunen, Federal Republic of Germany* The flue gas source is a
slipstream from a 350-MW coal-fired boiler.
FW manufactures steam generating equipment and is a contractor to
process plants. FW became very familiar with the BF adsorption process
during technical and economic assessments of pollution control systems by
the FW research department. Eventually, FW obtained licensing rights on
the BF dry adsorption system.
This simultaneous NOX:SOX removal process is still in the development
stages though it has been operated on coal-fired flue gas at several pilot
plants.. Arrangements are being made to continue the study at the Kellermann
plant by the addition of a RESOX unit to the existing prototype BF plant.
Since FW is a licensing agent this process should be readily available
to the U.S. market.
Published Economic Data
The estimated capital costs for an FW-BF adsorption system treating
flue gases from plants of >200-MW equlv are reported to range from $40/kW
for 0.9% S-containing fuel to $90/kW for fuel with 4.3% S (27, 28). Corres-
ponding operating costs fluctuate from 1.0-2.3 mllls/kWh with the same res-
pective fuel-S contents (27, 28). These values are 1977 costs for a U.S.
location. FW states that a graph of capital costs ($ /kW) vs MW rating and
% S in fuel indicates the most economical utilization of this process
in the 150-200-MW range. Beyond 200 MW there Is little change in the eco-
nomics. In cases of <50 MW, it is emphasized that several small boilers
could be connected to an individual treatment system and the economic
situation be improved by the overall rating of the system.
Raw Material, Energy, and Operation Requirements
The main raw material and utility requirements are water, makeup char
and sand, crushed coal, auxiliary fuel, and electricity. The consumption
and usage of the above are not yet available.
Maintenance requirements on this system are expected to be greater
than for other dry systems which use fixed beds or, at least, equipment
with fewer moving parts. Several mechanical problems caused trouble at the
Scholz pilot plant with adsorber—bed levels, char-sand separation, hot sand
conveying, char cooling, and char feed. However, several modifications
were made to correct the above problems. Labor and technical support
requirements are not known.
204
A
-------
Te chnical Consid eration
The FW-BF dry adsorption process is more complex than most other dry
processes as indicated by the large number of processing steps and generally
the flow of solids is more difficult to control as compared to the flow of
gases. The flow rate of the solids is controlled by the discharge feeders
on vessels, such as the adsorber, regenerator, and KESOX reactor. The flow
of char into the adsorber beds from the surge tank on top of the adsorber is
monitored with a level detector.
Little information has been obtained on the NOX and SOX removal sensi-
tivity to inlet gas composition. At the Scholz plant prototype unit an
average SOX removal of 961 was maintained as the inlet flue gas SOX compo-
sition varied from 900-2150 ppm. Figure 48 shows laboratory data on NOX
removal at a 720 ppm NOX level in the Inlet gas.
At the Gulf Power unit, the adsorber has two char-bed stages. Eight
parallel vertical beds (6 ft x 6 ft) form the first stage and four similar
beds (4 ft x 4 ft) are contained in the second stage. All beds are about
40 ft tall. The regeneration vessel used in this process has been in com-
mercial application, i.e., at a coke production facility by BF for over 5
yr; therefore, most of the design and operational difficulties have been
overcome. Also, the fluidized-bed heater has been used by many process
industries for several years.
This system would obviously not be suitable for retrofit on a plant
with an existing FGD since this process mainly removes.SOX from flue gas
and NOX removal is only secondary.
Turndown capability could be maintained to a limit by reducing flow
rates of char, 'sand, and crushed coal through the respective vessels.
However, a point could be reached where low gas and solid rates would inter-
fere especially with regeneration and RESOX operation. Several trains would
probably be required for a 5QO-MW size plant and trains could be removed
from service as needed for the necessary turndown capability.
Materials of construction are carbon steel for the most part with
some stainless steel being utilized in certain high-temperature areas.
Environmental Considerations
The only information obtained on removal efficiency sensitivity to
operating conditions is indicated in Figure 48 which shows NOX removal vs
space velocity on a bench-scale apparatus.
The FW-BF adsorption system removes SOX as well as NOX and some par-
ticulates. An average of 96% SOX removal has been obtained in pilot-plant
runs and an average of 40-60% NOX removal is expected. The maximum partic-
ulate removal is to be tested at the Kellermann plant unit by varying the
ESP efficiency which controls the particulate load of the gas entering the
adsorption system.
205
-------
50
u
z
QJ
O
t 30
UJ
O
a
x
O
Z
20
10
280 °F
4% Og
720 PPM NOX INLET
I
I
200 400 600 800
SF&CE VELOCITY, FT3 FLUE GAS /FT3 CHAR
1000
I20O
1400
Figure 48. Effect of Space Velocity on NOX Removal Efficiency for FW-BF Process (42).
-------
All the streams in this process except the main flue gas, S byproduct,
and crushed coal entering RESOX reactor are closed loop. So, there is no
waste disposal needed other than for the ash created in the process.
There is no information on the interference of any gas species and
contaminants.
In pilot-plant tests poor distribution of char to adsorber beds led to
problems of char stagnation, resulting in local hot spots within the adsorber
which not only reduce efficiency but could create a dangerous work hazard.
However, the new char distribution and surge tank above the adsorber with
feed lines to each bed is said to have corrected this problem. Additional
temperature monitoring of both gas and char throughout the adsorber is pro-
vided now. Handling of molten S could be a minor hazard not found in the
other processes which do not produce byproduct S.
Critical Data Gaps and PoorlyUnderstood Phenomena
Included in the information gap is the exact char cost, raw material
consumption, and the utility requirements. Also, the removal efficiencies
of both NOX and SOX need to be established for various inlet gas compositions
and operating conditions. The interference, if any, from any other gas spe-
cies and the maximum tolerable flue gas particulate loading need to be deter-
mined. The manpower requirements are also unknown.
Advantages, and Disadvantages
The advantages and disadvantages of the FW-BF adsorption process are
listed below.
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >95% SC>2 removal efficiency
3. Produces marketable byproduct (S)
4. Has been tested on coal-fired flue gas on pilot-plant or greater
scale
5, Operates between 100 and 200 C (In adsorber) which may require
negligible reheat
Pis advantages
1. Can only achieve a maximum NO removal efficiency of <70%
2. Forms secondary source of pollution (ash waste stream)
3. Requires significant amounts of energy for the regeneration step
4. Requires somewhat particulate-free gas feed
5. Requires hot solids handling
6. Uses moving-bed reactor which increases maintenance and catalyst
attrition
207
-------
HITACIIi, LTD., PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (2, 45, 50, 51)
Hitachi, Ltd., has developed a dry, SCR process for NOX removal. NHj
is the reduction agent; information describing the catalyst composition has
not been released. According to Hitachi, NOX in the flue gas is decomposed
into N£ and H20 by the following reactions at a temperature range of 25Q-40Q°C
(482-752°F) :
61)0 (g) + 4SH3(g) * 582(g) + 6H2°(g) «">
4NO(g) + 4SH3(g) + 02(g) + 4B2(g) + 6H20(g) (215)
Hitachi recommends two methods of applying their NOX removal process to
treatment of flue gases which have a very low particulate level. In the
direct method, the flue gas at about 300-400°C (572-752°F) from a hot ESP
is injected with OTU gas before it enters the reactor. This flow scheme is
shown in Figure 49. The gas is distributed in the fixed-bed reactor and
passes through the catalyst layers where the NOX reacts with the 1013. The
flue gas leaving the reactor enters the boiler air heater to heat incoming
air to the boiler. The flue gas then passes through a deaulfurization system
before being emitted at the stack. In the original concept by Hitachi, the
ESF was located after the air heater; however, its application was for oil-
fired boilers. If the same method was used for coal-fired boiler applica-
tion, the above scheme using a hot ESP would be necessary.
The second method uses a reheating scheme in which the flue gas is
treated after desulf urization. The flue gas is heated from about 60°C (140°F)
at the FGD outlet by passage through a heat exchanger and inline heater to
the reaction temperature (300-400°C). NH3 is injected into the gas as it
passes from the heater to the fixed-bed reactor. Upon exiting the reactor,
the treated gas passes through the heat exchanger to heat the Incoming,
untreated gas and then exits the stack.
The operating conditions which are applicable for the treatment of flue
gas from the combustion of heavy oil are as follows:
Reaction temperature 300-400°C (572-752°F)
Space velocity 10,000-20,000 hr"1
NH3:NOX mol ratio 0.9:1-1.1:1
Pressure drop across
the bed 100-150 mm H20
NOX removal efficiency >90%
Status of Development
Hitachi began working to find a catalyst for NOX removal as early as
1963 for the treatment of automobile exhausts. This work was later succeeded
by a project to develop an NOX removal system and the associated catalysts
208
-------
fO
O
NH9
AIR
> TO FGD
AND /OR STACK
Figure 49. Flow Diagram of Hitachi, Ltd., Process.
-------
for treatment of power plant flue gas. A 4000 "Sn^/hr capacity pilot plant
was constructed in 1974 using the direct method with crude heavy oil as the
combustion fuel. This pilot plant was used to test catalyst life and to
resolve potential scaleup problems and their solutions. A 200 Nm-Vhr capacity
bench-scale unit was operated in 1974 with Mitsubishi Petrochemicals to select
the optimum catalyst. The reheating method was employed during these tests
and the fuel source was heavy oil. For observation of the performance of the
process on a liquefied natural gas (LNG)-fired boiler, a 1000 Nm3/hr capacity
bench-scale unit started in 1975 using the direct method. Several other
pilot plants have also been operated.
NOX removal has been sustained at 90% efficiency for 7000 to 8000 hr with
negligible deterioration of the catalyst; therefore, a catalyst life of at
least 1 yr for commercial operation is expected. Hitachi states that tests
with coal-fired flue gas have been performed, but no details have been
released.
Backg roundof Developer
Hitachi constructs chemical plants, power plants, steel plants, etc.,
and manufactures equipment for environmental control, such aa ESP, desulfuri-
zation systems, NOX removal systems, and wastewater treatment and waste -
disposal facilities. Hitachi also produces the catalyst for the NOX removal
systems.
The accessibility of this process to the U.S. is favorable since there
are offices of Hitachi America, Ltd., in New York and San Francisco.
Hitachi has constructed a commercial plant to handle flue gas from a
170-MW equiv coke-oven plant with the reheating method. It was scheduled to
begin operation in October 1976. Two direct NOx removal facilities have been
ordered each for treating flue gas from 700-MW equiv-plants with LNG-fired
boilers.
Published_ Economic Data
The estimated capital cost (45) in 1976 for a 333-MW coal-fired plant
(assuming 3000 Nm^/hr ~ 1 MJJ) is $15 M, which is $45/kW. This is based on an
NOX removal efficiency of 90% with Inlet gas NOX content of 200 ppm. The
site is Japan and this cost includes a hot ESP.
RawMaterial, Energy, and Operation Requirements
NH3 is the only major raw material required. For treating flue gas from
a coal-fired boiler on a 333-JW plant (assuming 3000 Nm3/hr = 1 MW) with an
inlet flue gas NOX content of 200 ppm and 90% NOX removal efficiency, the
NH3 and utility requirements are as follows (includes hot ESP);
210
A
-------
Material Quantify
NH3 4.2 short tons/day
Electricity 6000 kW
Steam 52.8 tons/day
electrical requirement represents 1.8% of the total power output of the
plant.
No maintenance, labor, or technical requirements are yet available.
Technical Considerations
This process Is very simple as compared to wet processes. Process
control should Include automatic control of the NH3 flow rate. Hitachi
claims there are no effects on the catalyst with changes in NOX inlet con-
centration and that the catalyst Is free from adverse effects of S0~2.
With one of the largest space velocities of any dry process of 10,000 h
or greater and a typical reaction temperature for most dry NOx removal
facilities of 300-400°C (572-752°F), the reactor size should be about the
same or slightly smaller than for other dry systems.
For retrofit application on a. system with a cold ESP, the reheating
method for NOX removal would be required. Thus, an inline heater and heat
exchanger would be necessary. If an FGD unit is also present, some energy
saving in reheat could be realized by treating the gas between the ESP and
FGD unit.
Since several NOX removal trains are necessary for treating gas from a
5QO-MW size plant, turndown capability is accomplished by removal of trains
as necessary. The materials of construction required are not specified.
There are no byproducts with this process.
EnvironmentalConsiderations
The sensitivities of NOX removal on LNG and heavy oil combustion gases
to space velocity and temperature are shown in Figures 50 and 51. The NOX
removal efficiency increases as the temperature increases from 300°C to
400°C (572-752°F) and decreases as the space velocity Increases from 5,000
to 20,000 hr"1. Figure 52 shows a characteristic relationship of mol ratio
of NH3:NQX to NOX removal efficiency.
The NOX removal process of Hitachi does not remove any other pollutants.
There is no waste disposal necessary and apparently no secondary pollution
problem, since in some tests the NH3 concentration at the reactor outlet was
reported to be <5 ppm, even with a space velocity of 20,000 hr .
There are no apparent work hazards.
211
-------
100
80
55
u
0 60
ui
X
O
40
20
I I
LNG COMBUSTION
EXHAUST GAS
"HEAVY OIL
COMBUSTION
EXHAUST GAS
SVr 10,000 HR
_i
I
I
150
200
250 300 350 400
REACTION TEMPERATURE,°c
450
500
Figure 50. Relationship Between Reaction Temperatures
and NOX Removal Efficiency for Hitachi, Ltd., Process (50).
-------
ho
I—'
u>
3s 80
UJ
o
g
O
Ul
o:
60
40
20
LN6 COMBUSTION
EXHAUST GAS
HEAVY OIL COMBUSTION
EXHAUST GAS
TEMPERATURE* 350°C
NH3 / NO-1.1
I
5000 10,000 I5.0OO 20.OOO 25,000
SPACE VELOCITY, HR'1
Figure 51. Relationship Between Space Velocity and NOX
Removal Efficiency for Hitachi, Ltd., Process (50).
-------
NJ
100
90
o
z
UJ
u
b.
UJ
>
o
UJ
C
x
O
80
70
60
50
0.2
REACTION TEMPERATURE* 350°C
= 10,000
I
0.4
0.6 0.8 1.0 1.2
MOL RATIO OF NH3:NOX
1.4
1.6
Figure 52. Characteristic Curve of the Effect of Mol Ratio of
Ox on NOX Removal Efficiency for Hitachi, Ltd., Process (51)
-------
Critical Data Gapsand Poorly Understood Phenomena
Included in the data gap for the economics are the operating costs,
though the capital investment required is given. The catalyst probably has a
spherical or pellet structure* however, it has not been expressly stated.
Also, the exact nature of the catalyst and the catalyst cost have not been
released.
The maintenance and technical requirements are not known; however, they
should be similar to most other dry KOX removal processes. The major
materials of construction are not available.
Other than for the economic and the raw materials and energy information
reported herein, all the information has been from results on oil- or LNG-
fired boilers. Hitachi states tests have been performed on coal-fired
boilers, but details are not available,
Advantages and Disadvantages
Following are the advantages and disadvantages for the Hitachi process.
Mvantages
1. Achieves >90% NOX removal efficiency
2. Has been applied to flue gas from commercial coke-oven plant
3. Claims <10 ppm by vol NH3 in treated flue gas
Disadvantages
1. Requires somewhat particulate-free gas feed
215
-------
HITACF.x ZOSEN PROCESS - DRY, SCR (NOX)
Process Description and Principle's _of Operation (2, 4, 5, 10, 52, 58, 59)
Hitachi Zosen (Hitachi Shipbuilding and Engineering Company, Ltd.) has
developed an NOX removal process in which dry, SCR of NOX with NH3 occurs.
Hitachi Zosen is now developing a catalyst and reactor design which permits
treatment of flue gas with a high particulate loading. Therefore, the flue
gas from a coal-fired boiler may be fed directly to the reactor, upstream to
the air heater , without any particulate removal treatment as shown in Figure
53, Before entering the reactor, NH3 is injected into the flue gas. In the
reactor NOX is reduced to N£ by reaction with NH^ .in the presence of a catalyst
and a temperature range of 300-400 C' (572-752 F). The reactions occurring
are listed by Hitachi Zosen as follows:
6NO(g) 4- 4IH3(g) -, 5N2(g) + 6H20 (216)
6N02(g) + 4NH3(g) + 2N2(g) + 6NO(g) + 6H20(g) (217)
302(g) -> 2N2(g) + 6H20(g) (218)
Reaction (218) represents the reaction of excess NH3 with 02 at 400°C
to form N2» therefore, Hitachi Zosen reports that no problem exists with
excess NHj in the flue gas exiting the system. The treated flue gas passes
to the air heater for heating the air feeding the boiler, After the treated
flue gas is cooled in the heat exchanger, the gas is sent to the particulate
removal and desulfurization steps before exiting the stack.
The NH3:NOX mol ratio employed is about (0.8-1.2):!. The reaction
temperature range is 300-400°C (572-752°F) and Hitachi Zosen claims the*
pressure drop across the reactor is very low. The area velocity (flow rate
of gas/surface area of catalyst) is between 7-10 Nm3/hr/m2 and the relation-
ship of catalyst surface area to bulk volume ranges from 550-600 m /m . The
NOX removal efficiency is >90%. The process is to be able to function with
particulate loading up to 16 g/Nm3 (about 7 gr/sf t3) , making it possible to
use this system before an ESP or particulate scrubber on flue gas from a
coal-fired boiler.
The catalyst is manufactured into the shape of honeycomb units. These
units are welded together for the particular size required. The flue gas
passes parallel to the catalyst surface. The catalyst composition has not
been revealed for proprietary reasons; however, Hitachi Zosen does state
that it is constructed of common material. The expected catalyst life is
reported to be 1 yr.
Status of Development
Hitachi Zosen has been developing an NOX removal process and catalyst
since 1970. Five different catalyst aeries have been created and are known
as NOXNON 100, 200, 300, 400, and 500 series. The 100 series is nonselective
216
A
-------
PARTICULATE REMOVAL,
*» TO F6D
AND/OR STACK
AIR
Figure 53. Flow Diagram of Hitachi Zosen Process.
-------
and for use with CO, H2, and hydrocarbons as reductants. The other four
aeries are for use with NH3 as the reducing agent. Series 200 ±s for treating
"clean" flue gas, that is, gas which does not contain any significant amount
of SOX or participates. The 300 and 400 series are resistant to SOX. The
500 series is for use in gas with considerable particulate loading. There
have been over 21 different pilot-plant tests since 1973 in developing these
catalysts. The 500 series is still being tested and the results to date have
been promising even on gas containing 15 g/Ntn3 (6.44 gr/sft3) and 300 ppm SOX.
The following pilot-plant and bench-scale units are now testing the NOXNON
500 series catalyst.
Source of flue gas Test unit
(plant type) Fuel capacity, Nm-Vhr
Iron-ore sintering
Iron-ore sintering
Power
Glass-melting furnace
Heavy oil
Heavy oil
Coal
Heavy oil
5,000
200
200
200
Background of ProcessDeveloper
Hitachi Zosen is involved in shipbuilding, manufacturing of machinery,
steel structures, and plants for NH^, petroleum refining, and petrochemical
products. There is a liaison office in New York which should improve the
accessibility of the process to the U.S. market. The catalyst is manufactured
by a chemical company in Japan which is a subsidiary of Hitachi Zosen.
A commercial-scale, SCR NOX removal facility was constructed by Hitachi
Zosen in the fall of 1975 for a petroleum plant with an oil-fired boiler and
FGT capacity of 440,000 Nm3/hr. Another facility was completed in the fall
of 1975 for a petroleum operation with 350,000 Ntn-Vhr flue gas capacity. Two
more denitrification systems for steel manufacturing plants with 71,000 and
750,000 Nm3/hr flue gas rates have been completed and are reported to be
operating successfully.
Published Economic Data
The estimated capital cost for a denitrification unit to treat flue gas
containing 300 ppm NOX on a 250-MW-size plant is reported to be about $16/kW
(58). The corresponding operating cost would be approximately 1.5 mills/kWh
(59). These are assumed to be 1976 costs at a Japanese site.
RawMaterial, Energy, and Operation Requirements
The only major raw material required is NHj. For a 250-MW plant with
300 ppm NOX in the flue gas, an estimated 5 short tons/day of NH3 is required
for NOX removal.
On the same basis as above, the following energy consumption for
denitrification is required;
218
A
-------
Energy Quantity
Electricity 400 kW
Steam 4.1 tons/day
No reheat fuel is necessary. The electrical requirement represents 0.2%
of the total power output of the plant. The majority of the steam (2.9 tons/
day) is used for NHg vaporization, while the remaining 1.2 tons/day is for
soot blowing.
With no particulate scrubbing before the NOX removal system, soot
blowing of the air heater and the reactor are required. For pilot plants,
the reactor has required washing out of deposits once every 2 mo and Hitachi
Zosen states that this operation requires only a few min tp complete,
The technical support necessary for operation of this NOX removal system
should be similar to other dry systems.
Technical Considerations
Having only the reactor and NH3 storage and injection system, this
process would be a direct and simple dry process. Process control would
include automatic control of NHj flow rate. On a recently constructed plant,
the NH3 flow rate is automatically controlled based on the total inlet gas
flow rate and inlet NOX concentrations.
There is no information available yet on the sensitivity of the process
to inlet gas composition.
The equipment size for this system should be comparable to other dry
NOX removal systems. The reaction temperature is similar at 30Q-4QO°C
(572-752°F) and the area velocity range is comparable at 7-10 Ifa3/hr/m2.
In retrofit application, it would be most advantageous to treat the
flue gas for NOX removal prior to the air heater whether or not an FGD
system or only a cold ESP is present. More energy would be required to
reheat the flue gas to reaction temperature if the gas Is treated after the
air heater and even more energy is required if the FGD system is located
before the NOX removal system. The Hitachi Zosen process would have the
advantage over other processes of less area required since there are fewer
pieces of equipment. The turndown capability of this process is similar to
other dry systems in that an NOX removal train could be removed from service.
The major material of construction is reported to be carbon steel.
There Is no byproduct made in this process.
Should future results prove this direct FGT infeasible, Hitachi Zosen
has a more conventional NOX removal process, such as found on the commercial
plants constructed by Hitachi Zosen for oil-fired boilers. It employs the
heat exchanger, heater, and particulate removal discussed in the other dry
NOX removal processes and is comparable to other processes. The estimated
capital cost for this system is estimated to be $50-$60/kW based on 40-150
219
-------
ppm NOX, 1976 costs, and 300 yen/$. The NO removal efficiency is reported
to be >90% and operating conditions are similar to those previously reported.
The operating cost is estimated at 3.3 mills/kWh.
Envlronmental Conslderations
Little information has been reported on the sensitivity of NOX removal
efficiency to various operating conditions. The only data available indicate
the NOX removal efficiency approaching 90-100% at 350°C (662°F) and above
for most of the catalysts. The efficiency drops off to 60-70% at 300°C
(572°F).
There are no additional pollutants removed, other than NOX, by the
Hitachi Zosen process,
Excess NH-j at the outlet of the system is a potential problem; however,
Hitachi Zosen claims NHj exiting the system is very low and creates no
pollution problem. Some type of disposal for the reactor deposit washing
solution would be essential. Also, the catalyst is disposed of and not
reclaimed.
There are no apparent work hazards.
Critical Data Gaps and Poorly Understood Phenomena
There is no information available presently on the sensitivity of the
NOX removal efficiency to factors such as inlet gas composition and various
operating conditions. Specifically, there is no comment on any possible
adverse effects from Cl~ in the flue gas from a coal-fired boiler. These
effects might include interference of NOX removal efficiency or deterioration
of catalyst.
The composition of the catalyst is not known. The exact cost per pound
is not known though the estimated cost for the initial investment in catalyst
necessary for a 250-MW plant with flue gas containing 300 ppm NQX is given
by Hitachi Zosen as $1.7M (1976 cost and assuming 300 yen - $1).
There are few available facts on wash frequency and other washing
requirements on the reactor treating the flue gas directly from a coal-fired
boiler with no particulate removal. More information should be accessible
as the pilot-plant tests progress.
Maintenance, operating, and technical support requirements have not
been published.
Advantages and Disadvantages
Advantages
1. Achieves >90% NOX removal efficiency
2. Has been applied to flue gas from commercial oil-fired boilers
3. Operates with full particulate loadings (>7 gr/sft^)
4. Is probably <10 ppm by vol HHj in treated flue gas
Disadvantages
None is readily apparent now.
220
A
-------
4NH3(g) +
2NH3(g)
2N°2(g) -
+ N°2(g) '
h °2(g) -
f N°(g) -
* 2N2(g) -
1 £i"TJ f\
* OJul W v
h 3H2°(g)
JGC PARANOX PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation. (2, 4, 9, 62, 63, 64, 96)
JGC Corporation's "Paranox" process removes NOX from flue gas by SCR
employing NH^ as the reducing agent. The catalyst was developed by JGC and a
parallel-passage reactor is applied.
The flow diagram for the Paranox process is shown in Figure 54 and,
excluding the ESP assumed necessary for coal-fired flue gas, it is the same
as for the treatment plant at the Kashima Refinery of Kashima Oil. Flue gas
from a coal-fired boiler would pass through an air heater for heat recovery
to air feeding the boiler and then an ESP for particulate removal. The flue
gas then is heated to 380-420°C (716-788°F) by passage through a counter-
current heat exchanger and an auxiliary heater. NH3 is injected into the
flue gas before it enters the parallel-passage reactor. The main chemical
reactions in the reactor by which NOX is converted to N£ and ^Q by reduction
with NH3 are stated by JGC to be as follows:
4N°(g) + °2(g) ^ 4N2(g) + 6H2°(g) (219>
(220)
<221>
After exiting the reactor, the flue gas passes through the countercurrent
heat exchanger to heat the incoming, untreated gas passing to the reactor.
After exiting the heat exchanger, the treated flue gas passes to a desulfuri-
zation unit and/or the stack.
The operating conditions at the Kashima plant are as follows:
Reactor temperature 380-420°C (716-788°?)
NH3:NOX mol ratio (1.1-1.3);!
Pressure drop (see Figure 55) 180-200 mm H20
Particulate loading 0.05 g/Nm3 (0.02 gr/sft3)
The NHj injected into the flue gas has been diluted with steam at a
steam to NH3 mol ratio of 20:1. A stable NOX removal efficiency of >95% has
been demonstrated with an inlet gas NQX concentration of 60-250 ppm. The
concentration of NH3 exiting In the treated gas is reported to be <10 ppm,
The catalyst life is expected to be at least 1 yr.
Status of Development
The Paranox process consists of a denltrifIcation catalyst, developed
by JGC, and the parallel-passage-type reactor, developed by Shell and patent
rights possessed by Shell International Petroleum Maatichappij (SIPM). Nihan
Shell Gijutsu KK, a joint undertaking created by Shell Kosan Company and JGC,
is the licensing agent for the Paranox process in the Far East.
221
-------
• m
1 i
\ i
FLUE AIR
DOILCR GAS * HEATER
M A
I
AIR
i
i
I-COLD" 1 HEAT 1
""* ESP "" * EXCHANGER *"
1 1
1
TO F
AUXILIARY
HEATER
• i
t
i
60 FUEL NH3
AND/OR
STACK
Figure 54. Flow Diagram of JGC Paranox Process.
-------
Backgroundof Process Developer
JGG Corporation, formerly Japan Gasoline Company, Ltd., is an engineer-
ing and plant construction organization offering services in engineering,
consulting, procurement, construction, and operation management. The fields
of activity encompass petroleum, petrochemical, and diversified gas projects,
environmental conservation, nuclear projects, desalination, foods, and
Pharmaceuticals. JGC built the Shell FGD unit at Showa Yokkalchi Sekiyo and
has developed its own catalyst for SCR of NOX. It has constructed two
prototype units. One is at Kashima Refinery for Kashima Oil with a flue gas
treating capacity of 50,000 Nm3/hr and it has operated since November 1975.
The second is a 70,000 Nm3/hr capacity treatment plant for Fuji Oil. The
Paranox process may not be as accessible to the U.S. market as other processes
since no contact could be made with any JGC Corporation office in the U.S.
Published Economic Data
The capital cost for a 150,000 Nm3/hr gas treatment facility with 85-
90% NOX removal is estimated to be $1.33M, which is about $27/kW (96).
Values for operating costs have not been reported. The above estimate was
for oil-fired gas treatment and does not include an ESP and is assumed to be
1976 values for a Japanese location.
Raw Material, Energy, and Operation Requirements
The major raw material is NH3 and for an NHj:NOx ratio of 1.2:1, it is
estimated that 24 short tons of NH3/day are required for a 500-MW equiv
plant with an inlet flue gas containing 600 ppm NOX. Fuel and electricity
are the major energy sources required and no estimates for their usage are
available. The manpower and technical requirements are not stated, but are
expected to be the same as similar dry processes.
TechnicalConsiderationa
The Paranox process involves simple and straightforward chemistry. The
NH3 flow rate and reaction temperature should be capable of being automati-
cally controlled. The information obtained on NOX removal sensitivity to
inlet gas concentration is shown in Figures 55-58. In Figure 55 NOX in the
inlet gas ranges from 90-100 ppm while the outlet gas consists of 1-3 ppm
NOX. NOX removal results with varying inlet NO concentrations are shown in
Figure 56 while the effects of different 02 and 1^0 levels in the flue gas
on NOX removal efficiency are displayed in Figures 57 and 58 respectively.
The space velocity is not specified, but if in the 4000-7000 hr"~l range,
the equipment size necessary will be comparable to that of most other dry
processes with similar reaction temperatures.
The Paranox process Is suitable for retrofit on a plant with a cold ESP
and with or without an FGD unit. If placed after the FGD unit, more reheat
will be required. For coal~fired flue gas, the Paranox process may be as
223
-------
o
UJ
o
!*!
<#
o
W
cc
X
o
Z
§Tt
|_ CL Q.
£: K *<
z +.•£
U JB^
1 °*
v_) *- *.
o 5 ,2
™" *- rj
O
N
I
z
oT
UJ
or
K
o:
a,
1W
98
96
94
92
|{VJ
8.0
6.0
4JO
2.0
200
160
120
80
Q
C
1 l I
^•NHj/NOs I.3-J.4 NH^/NO* I.I
I ~- «^- , -5 1
~ ' —
M» ' ~^.
. _- ^..^r -*^^- • J^^_— - • •
^-REACTOR INLET
— —
^•REACTOR OUTLET
K -^ 4 ^..-.T., '" "
AV ^
J\r~*
I. ^ V
_
-—____ ___ -
. — r=. ^ ^ •»
_ _
anai . ^^^
... —
1 1 i
> 500 1000 1500 20C
PERIOD, HOURS
Figure 55. NOX Removal Efficiency, NOX Concentration, and
Pressure Loss Over 2,000 hr Test Period for JGC Paranox Process (2)
224
A
-------
N3
N3
too
.75
o
§
GC
X
O
50
HIGH TEMP. CAT
CAT=JPI02
TEMP«380°C
NH3/NO = I.I
SOg* 1500 ppm
H20SIO%
02«3%
N2-BALANCE
200 400
NO CONCENTRATION , ppm
Figure 56. Effect of NO Initial Concentration on NOX Removal Efficiency for JGC Paranox Process.
-------
NJ
NJ
100
6"
o 75
O
5
ttl
50
X
O
25
HIGH TEMP. CAT.
CAT = JPI02
TEMP =380°C
NO = 200 ppm
NHs= 2 20 ppm
S02=I500ppm
H20« 10%
N2 = BALANCE
I
0 I 23 4 5 6
02 CONCENTRATION, VOLUME %
Figure 57. Effect of G£ Concentration on NOX Removal Efficiency for JGC Paranox Process (64)
-------
N3
N3
IWW
o 75
UJ
o
u.
u.
Ul
< 50
o
UJ
QC
X
0
z 25
n
I 1 i 1
— —
HIGH TEMP. CAT.
~~ CAT. = JPI02 ~"
TEMP* 380° C
NO a 200 ppm
NH3 * 220ppm
~~ SO2 * I500ppm —
02 s 3 %
N2 = BALANCE
i 1 II
5 10 15 20
HjrO CONCENTRATION, VOLUME %
25
Figure 58. Effect of H20 Concentration on NO
Removal Efficiency for JGC Paranox Process (64; .
-------
suitable as some other dry processes mentioned for treating the flue gas
before the air heater. Savings in fuel for reheat would be significant and
elimination of the heat exchanger would also result. Of course, a hot ESP
with larger power consumption would be required.
Several NOX removal trains would be required for treating a 500-MW
equivwgas stream. Therefore, as for other dry processes, the major turndown
capability would be the removal of a train from the process. The auxiliary
heater would provide temperature control during boiler load fluctuations.
The materials of construction are not known.
Environmental Considerations
Information on the sensitivity of NOX removal efficiency to NHj:NQX
ratio over a 2000-hr test period indicates a decrease from 99 to 97.5% in
NOX removal with a decrease from 1.3:1 to 1.1:1 in NH3:NOX mol ratio (see
Figure 55). Figure 59 presents NOX removal in comparison to the NH3:NOX
mol ratio. The relationship of temperature to NQX reduction for two of
JGC's catalysts is given in Figure 60.
The Paranox process removes no pollutants other than NOX. There should
be no waste disposal required nor is there secondary pollution problems since
the NH3 is <10 ppm in the outlet gas. No information is known about
interference of gas species. There are no evident work hazards,
Critical Data Gaps and Poorly UnderstoodPhenomena
The operating costs as well as the energy requirements for the Paranox
process are not given, Also, the inlet gas NOX concentration basis is not
known for the reported capital cost of $27/kW for a 500-MW equiv--plant.
Included in the data gaps are the space velocity required in the reactor and
the possible adverse effects of other gas species expected with coal-fired
flue gas. In addition, the maximum particulate loading allowed is uncertain.
The exact manpower requirements (labor and technical) are not known
though they are expected to be very similar to other simple, dry processes.
It has not been ascertained whether there are any unusual materials of
construction required.
Advantages and Disadvantages
The advantages and disadvantages of the JGC Paranox process are as
follows:
Advantages
1. Achieves >901 NOX removal efficiency
2. Claims <10 ppm by vol NH3 in treated flue gas
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Requires, apparently, somewhat partlculate-free gas feed
3. Requires auxiliary heater to attain or control reaction temperatures
228
A
-------
100
N3
N3
75
UJ
o
it
I
UJ
oc
50
25
HIGH TEMP. CAT.
CAT. = JPI02
TEMP. = 380° C
NO* 200ppm
S02= I500ppm
02S 3%
H20S 10%
Ng* BALANCE
I I
1.0
NH3/NO MOLAR RATIO
20
Figure 59. Effect of NH3:NO Mol Ratio on NOX
Removal Efficiency for JGC Paranox Process (64) .
-------
100
§90
UJ
UJ 80
to >
JP50I LOW TEMP. CAT
70
x
O
60
JPI02 HIGH TEMP. CAT.
INLET CONDITIONS
NO = 200ppm
NH3 = 240ppm
02* 3%
H20= 10%
N2 = BALANCE
200
250
300
TEMPERATURE , °C
350
400
Figure 60. Effect of Temperature on NOX Removal Efficiency for JGC Paranox Process (64).
-------
KOBE STEEL PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (69, 70)
The Kobe Steel process Is a dry N0x-only removal system based on the
SCR of NOX using NHj, This dry NOX process is similar to the other dry
processes since it consists of three distinct sections: particulate removal,
NH3 injection, and catalytic reduction, as shown in the flowsheet in Figure
61. Kobe Steel has developed two different types of reduction reactors,
one for "clean" flue gas and one for particulate-laden gases, the latter of
which will be the only one considered in this description,
The flue gas from the boiler at 350-400°C (660-750°F) passes through a
hot ESP to reduce the particulate levels below 0,2-0.4 .g/Nm3 (0.09-0.17 gr/
sft^ at 32 F) and enters a mixing chamber where gaseous NH3, in an NH3:NOX
mol ratio of (1.0-1.1):!, is intimately mixed with the flue gas before entering
the KSL moving-bed reactor. As the mixture of flue gas and NHj pass over
and through the moving packed-bed sections containing the proprietary base-
metal catalyst, NOX is selectively reduced to N£ by the following reactions;
4NO(g) + 4HH3(g) + 02(g) -> 4N2(g) 4 6H20(g) (222)
6N02(g) + 8HH3(g) + 7N + 12H0 (223)
Since most of the NOX in the flue gas is present as NO, reaction (222)
is the primary reaction occurring in the reduction reactor. The clean flue
gas from the reactor passes through the air heater and then an FGD section
before exiting through the stack.
Status o f Dey e lo pjien t
Two bench-scale units using the Kobe Steel KSL moving-bed process to
treat 1000 Nm3/hr (0.3 MW equiv) of "dirty" flue gas was started up recently
and testing of the process on dirty flue gas (30-150 ppm SC>2, 300-400 ppm
NOX, 0.2-0.4 g/Nm3 of dust) is apparently continuing at the present time.
The longest period of continuous operation was 8000 hr (330 days) during a
test of the catalyst's life expectancy in particulate-laden flue gases.
Although this process has not been tested on coal-fired flue gas, the
Kobe Steel process should be able to adjust to the higher particulate and
NOX levels associated with coal-fired flue gas since it has been tested on
flue gas from a coke plant and a pelletizing plant containing approximately
0.2-0.4 g/Nm3 of dust and 300-400 ppm NOX.
The next development step for this process would be the continuation
of current testing in the 1000 Nm3/hr bench-scale units to generate data
on raw material, energy, and catalyst consumption. In addition, the ability
of the Kobe Steel process to handle the flue gas from a coal-fired boiler
needs to be demonstrated during long-term continuous pilot-plant operation.
231
-------
< 1
1 "HOT"
I
ASH
i
IV
Ha
1
r
REACTOR
\
SCREEN
— i
— *
AIR „ TO F6D
HEATER "" SYSTEM
AIR
DUST
Figure 61. Flow Diagram of Kobe Steel Process.
-------
Background of Process Developer
Kobe Steel, Ltd., is a large Japanese corporation heavily involved in
the iron and steel industry. With the coming of the strict air pollution
regulations in Japan in the early 1970's, Kobe Steel entered the air pollu-
tion control field to solve the SC>2 emissions problems from their plants.
Their CAL-SOX wet limestone FGD system (KSL FGD process) was developed and
has reached the commercial operation stage. In response to increasingly
strict NOX levels, Kobe Steel began the development of both a dry SCR pro-
cess using NHg and also a wet absorption-oxidation process. The development
of this wet absorption-oxidation has recently been suspended and the develop-
ment of a new wet process has been initiated. Their dry SCR process was
split into two types, one to treat clean gas and one to treat dirty flue
gas, with the only major difference being the design and type of reduction
reactor.
At the present time no American company has been licensed to market
this technology in the U.S. However, Kobe Steel does have a liaison office
in New York City through which inquiries about this process are handled.
Published Economic Data
Based on the premises of a 500-MW coal-fired boiler and 600 ppm NOX
in the inlet flue gas, the capital investment for the Kobe Steel process
has been estimated (43) as $4-7. Nm3/hr of flue gas. This value corresponds
to $12-21/kW of installed capacity (construction basis: Japan and 1976
dollars). The annual operating cost is expected to be $3-4 Nm /yr of flue
gas, which, assuming 7000 operating hr/yr, is equivalent to about 1.3-1.7
mills/kWh (43).
Raw Material,Energy, and Operation Requirements
The only major raw material requirement for this process would be NHj
with some makeup catalyst required to replace losses due to attrition. The
.only major energy requirement would be electricity. No information has been
published yet on the quantities of NH3,catalyst makeup, or electricity
required for this process.
In addition, the operating manpower and maintenance requirements have
also not been made available. Kobe Steel does state that the process is
fully automated and what little operating adjustments which are required
can be handled by personnel from other areas of the power plant.
Technical Considerations
The Kobe Steel flue gas denitrlfication process is a relatively simple
unit with only five pieces of equipment: a hot ESP, an NH3 storage tank,
an NH3 vaporizing unit, a mixing vessel, and a reduction reactor. The
only major process control problem is vaporizing and Injecting the exact
amount of'NH-j required to provide an NH3:NQX mol ratio of (1.0-1.1):! in
233
-------
the reduction reactor. Although the exact control mechanism has not been
specified by Kobe Steel, the NH3 is probably metered into the mixing tank
in direct proportion to the amount of flue gas passing through the system.
The'KSL moving-bed reactor will be the largest and most complex piece
of equipment in this system and a diagram of this reactor is shown in Figure
62, The catalyst pellets enter the reactor from the top and gradually drop
to the bottom of the reactor beds as they become clogged with particulates
from the flue gas. Although the residence time and"pellet size for the base-
metal catalyst have not been specified by Kobe Steel, from the numerous graphs
provided it is estimated for 90% NOX removal, the smaller catalyst size
_/3-7 mm (0.1-0,3 in.) in diameter/ will be required to obtain a reasonable
space velocity through the reactor. The use of the smaller catalyst pellets,
however, will sharply increase the- pressure drop through the reactor during
routine operation and increase the rate of catalyst movement through the bed.
The higher velocity of .the catalyst will, of course, result in larger catalyst
losses due to increased breakage and also to larger maintenance costs.
For the purpose of this study then the operating condition in the reduc-
tion reactor would be the following:
Parameter Value
NH3:NOX mo1 ratio (1.0-1.1):!
Temperature' 350°C (660°F)
Space velocity 10,000 hr"1
Catalyst size (shape) 3-7 mm dia (pellet)
Maximum tolerable 0.2-0.4 g/Nm^
dust level
These operating parameters for the reduction reactor in the Kobe Steel
process are very similar to those given for the other SCR processes.
The NOX removal efficiency is relatively independent of the flue gas
composition, excluding particulate levels, under normal circumstances. As
long as the temperature in the reactor remains near the recommended operating
temperature of 350°C the denitrification efficiency is insensitive to the
S02 concentration. The NOX removal efficiency is completely independent of
the H20, 02, and NO concentrations in the flue gas,providing that the 02
concentration is 1% or greater. The particulate levels must remain below
0,2-0.4 g/Nm^ to prevent excessive plugging in the reactor.
This process could not be easily retrofitted onto an existing 500-MW
plant equipped with cold ESP. The ductwork modifications required to remove
and return the flue gas between the economizer and the air heater would be
extremely difficult and expensive. Thus for retrofit applications the flue
gas would be removed after the existing cold ESP and sent to a countercurrent
heat exchanger where the raw flue gas extracts waste heat from the clean
flue gas exiting the denitrification section. The flue gas is further heated
in an oil-fired furnace to raise the temperature of 350°C (660°F) and then
mixed with gaseous NH3 before entering the reduction reactor. As the gas
234
-------
GAS
CLEAN GAS
Figure 62. Catalyst Reactor of KSL Moving Bed (70)
235
-------
mixture passes through the reactor, the NO is catalytically reduced to N£
and the effluent gas is sent to the heat exchanger to preheat the incoming
raw flue gas before entering the desulfurization section. Thus, the retro-
fitting of this system on existing power plants will increase both the
capital investment and operating cost since extra equipment and fuel will
be required. (For these retrofit applications Kobe Steel has estimated that
the capital and operating costs will be approximately double the cost for
installing the system on a new unit.)
Although Kobe Steel has not disclosed the number of separate trains of
denitrification equipment needed for a 500-MW boiler, from previous studies
It is assumed that two trains will be required. Thus by simply closing down
the trains as needed, the denitrification system will be able to partially
cycle with the boiler availability. The major consideration during turndown
will be to maintain a high temperature (350°C) in the reactor area to pre-
vent the formation and precipitation of NH4HS04 on the reactor and catalyst
surfaces,
Kobe Steel has not published any information on the materials of con-
struction required for this process,
Environmental Considerations
The Kobe Steel process has been shown in pilot-plant tests to be
capable of selectively removing 90% of the NOX by reaction with NHj. To
maintain this high removal efficiency the temperature must remain above
350°C (660°F) and the mol ratio -of NH3:NOX must remain at approximately
(1.0-1.1):!, When less NH3 is injected, the denitrification efficiency
will fall below 90%.
This process is for the selective removal of NOX and does not remove
any other pollutants from the flue gas.
Critical_JData Gaps andr Poorly Understood Phenomena
Many of the critical data gaps for this process involve the operating
requirements and costs for this system, Kobe Steel has not as yet published
any information on the NH3 and catalytic consumption and cost, and energy
requirements. In addition, the materials of construction, operating man-
power, and maintenance requirements have also not been made available. The
NHg level in the treated gas exiting the NOX removal system has not been
reported,
Advantages and Disadvantages
The advantages and disadvantages for the Kobe Steel process are listed
below;
Advantages
1. Achieves >90% NOX removal efficiency
236
A
-------
Disadvantages
1. Has not been tested on coal-fired flue gas
2, Has been tested only in a bench-scale unit
3. Requires .somewhat clean (particulate level must be below about
0.1-0.2 gr/sft3) gas feed
4, Uses moving-bed reactor which Increases maintenance and catalyst
attrition
237
-------
KUMBO KNORCA PROCESS - DRY, SCR (NOX)
Process Descrigtlon and grinciBles of Operation (4» 5, 7, 68, 71, 73, 98)
Kurabo's NOX removal process involves selective reduction of NOX with
MH3 using Kurabo's own developed catalyst. Two catalysts and support systems
have been developed by Kurabo. One is CuO on A1203 which was established
first. The second system, composed of an Iron (fe) based catalyst on
titanium dioxide (Ti02) carrier, was just recently developed and would be
more applicable to coal-fired FGT. The released information on this latter
catalyst and support material is limited and thus the CuO catalyst with
alumina support is used in this process description. The general removal
principles are the same for' either catalyst.
Based on the flow scheme of the 30,000 Nm^/hr capacity prototype unit
for oil-fired FGT, an assumed process flow diagram for a coal-fired FGT
operation is shown in Figure 63.
The gas from a coal-fired boiler, after having the particulates removed
(e.g., by ESP), is sent to an auxiliary heater to (1) maintain the reactor
temperature above 350°C (662°F), (2) prevent NH4HS04 formation and depositing
on the catalyst in the reactor, and (3) control reactor temperature with
fluctuating boiler load. Kurabo states that this may not be necessary if
the catalytic reactor can be located before the boiler-economizer or by-
passing a portion of the superheater flue gas. Initially the flue gas is at
350°C (662°F) prior to the auxiliary heater, and the gas is at the optimum
reaction temperature of 39Q-420°C (734-788°F) after the heater. NHj is
injected into the gas between the heater and reactor. Kurabo states that
803 reacts with the CuO catalyst to form copper sulfate (CuSO^). The CuS04
acts as a catalyst for the reduction of NO to N£ by reaction with MHj, The
reactor has a moving catalyst bed. The ephei-ical catalyst (3-5 mm dia) moves
slowly, downward continuously while the flue gas passes through the reactor
in crossflow. The catalyst removed from the reactor has the dust removed,
is reactivated (CuS04 •*• CuO) by heating to 800°C (1472°F), and then returned
to the reactor by a conveyor. The SOX and C02 from catalyst regeneration
are recycled to the gas exiting the reactor at 420°C (788°F), and the combined
stream passes through the boiler air heater to recover the heat before being
sent to desulfurization and/or the stack.
The chemical reactions for NOX removal in the reactor with the CuO
catalyst are given by Kurabo as follows:
SO,, . + CuO, , -»- CuSO, , , (224)
3(g) (s) 4(s)
In the presence of
4NO(g) + 4ffl3(g) + 02(g) + 4N2(g) 4 6H20(g) (225)
238
A
-------
AUXILIARY
HEATER
FUEL
1
i
N
i
H3
i
RE
i
I
REACTOR
i
i
3ENERATOI
|
I
?_Qff.
GAS
Al
R
TER
A
T
AIR
TO FGD
AND/OR STACK
Figure 63. Flow Diagram of Kurabo Knorca Process
-------
Otherwise,
6NO(g) + 4NH3(g) -, 5N2(g)
6N02(g) + 8NH3(g) - 7N2(g)
12H20(g)
In regeneration, the reactions at 800°C are:
CuSO. , N -*• CuO/ % + S0,x N
4(s) (s) 3(g)
A10(SO
X
,
x 3(a
A12°3(8)
nlS°3(g)
n2S02(g)
C(s)+02(g)"C°2(g)
(226)
(227)
(228)
(229)
(230)
The Al2(SOx)3 is created in the reactor as the catalyst (CuO) carrier,
A1203, reacts with the 803.
The "Knorea" system removes more than 90% NOX with an
mol ratio
of (0.9-1.2):! and a flue gas containing 1600 ppm S02, 280 ppm NOX, and
0.1 g/Nm3 of dust (0.04 gr/sft3). Greater than 80% of the particulate loading
is removed in the reactor but that is with inlet dust density of 0.1 g/Nm .
Pilot-plant operation has indicated that particulate loadings of 1-2 g/Nm3
(0.43-0.86 gr/sft3) to the reactor may be tolerated, although collection
efficiencies are not sufficient for particulate control. Therefore, an ESP
is necessary to remove dust to an acceptable degree prior to the gas entering
the reactor. The catalyst regeneration cycle takes 100 hr and the yearly loss
by crushing is 5% of the total amount of catalyst. Catalyst life is estimated
at 10,000 hr.
Typical operating conditions for
follows: •
Reactor temperature
Pressure drop across bed
Space velocity
NH3:NOX mol ratio
NOX removal efficiency
Inlet gas composition
S02
NOX
Dust
Duat removal
NH3 at outlet
Regeneration temperature
Catalyst life
the Knorca system are reported as
350-400°C (662-752°F)
100 mm H20
7,000-10,000 hr""1
(0.9-1.0):!
90%
1,600-2,000 ppm
300 ppm
0.1 g/Nm3 (0.043 gr/sft3)
801 at 0.1-0.2 g/Nm3 (0.043-
0.086 gr/sft3) particulate
loading at inlet
<10 ppm
800°C (1,472°F)
8,000-10,000 hr
240
A
-------
Status of Development
lurabo tested the Knorca system over a year on a 5,000 Nm^/hr flue gas
capacity pilot plant with an oil-fired boiler. This pilot plant had a fixed
catalyst bed in which pressure drop would increase after 30 hr of operation.
To solve this problem, Kurabo made use of the moving catalyst bed* Problems
were also experienced with NH4HS04 deposits in the air heater. This was
solved by reducing the NH3:NOX tnol ratio to about (0.9-1.0):!. A 30,000
Nm-*/hr capacity treatment plant was constructed by August 1975, and no
problem has been observed during operation of this plant with pressure drop
or NH4HS04 formation.
The reactor handling 30,000 Nm3/hr of flue gas from an oil-fired boiler
has three elements and each treats 10,000 NitP/hr of flue gas. The average
operating conditions are as follows;
NOX inlet concentration 280 ppm
NOX removal efficiency 93.6%
Particulate loading 94 rag/Km^ (0.04 gr/sft^)
Gas volume 24,000 Nm3/hr
Space velocity 8,000 hr~l
S02 inlet concentration 1,650 ppm
Gas temperature 400°C (752°F)
Pressure drop 65 mm H20
NH-j consumption 106.4 1/min
O mol ratio 1:1
Testing of effects of Cl"" on catalyst and aluminum (Al) support material
and a new support material have been performed. As a result, a Ti02 support
system with an Fe-based catalyst has been selected for use where Cl"" con-
centrations exceed 100 ppm. The Al support had a tendency to irreversibly
degrade at these Cl" levels. The Ti02 is inert to halogen poisoning and
does not require extensive regeneration. Though more expensive, the T102
is better when considering maintenance and operating costs for regeneration.
Background_of_ Process Developer
Kurabo is a producer of chemical and textile products. The Kurabo
Engineering Division was started in 1970 to develop pollution control systems.
Kurabo has developed an FGD system which is in use at over 140 industrial
sites. Other systems for wastewater treatment, E^O purification, etc., have
been created and are in wide spread use. T1W, Inc., a diversified worldwide
organization, has exclusive rights for supplying Knorca systems for boilers
in the U.S. and Canada. TRW has had over two decades of involvement with
Japanese business and has been involved since 1971 with Japanese technology
on pollution abatement. This process should be readily accessible to the
U.S. market. Kurabo states that the Knorca system is now commercially
available for oil-fired boilers.
241
-------
Published Economic Data
"\
The capital investment for a plant treating 500,000 Nirr/hr flue gas Is
about $5.83M (68). This was for a boiler using C-class heavy oil, 8,400 hr/yr
operating time, inlet gas NOX concentration of 280 ppm and reaction tempera-
ture of 400°C (752°F). The revenue requirement equaled $1.99M/yr ($237.00/hr)
or $7.59/kl of heavy oil (68), The capital investment is equivalent to about
$35/kW and revenue requirements are 1,4 mills/kWh. All these figures are
assumed to be 1976 costs with Japan as site location.
RawMaterial, Energy, and Operation Requirements
NH3 is the only major raw material required for the Knorca system. With
the NH3:NOX mol ratio of 1:1, the MS usage for a 500,000 Nm3/hr gas capacity
plant should be about 3.5 short tons/day at an NOX inlet concentration of
300 ppm.
The fuel oil required is 749 kl/day for the same conditions as above
in published economic data, which is equivalent to about 27.7 GBtu/day,
The following electricity and steam requirements are also on the same basis
as the above:
Material Quantity
Electricity 1,169 kW
Steam 15,3 tons/day
The electrical requirement represents 0.7% of the total equivalent electrical
output of the plant.
No details are available on operation and maintenance requirements,
but the operation is simpler and should require leas labor and maintenance
than wet NOX removal processes. The moving-bed and regeneration system
probably require more maintenance than dry systems with a fixed bed and no
regeneration. This system should not require any more technical support
than any other dry NOX removal process.
TechnicalConsiderations
This system is simple and can be fully automated. The NH3 flow rate
control is based upon the inlet gas NOX concentration and the gas volume.
As a double check, the NH3 usage rate is also compared with fuel consumption.
Experimental data from Kurabo show that 62 has an effect on NO removal.
As 02 present in flue gas Increases from 0-0.5%, the NOX removal increases
from about 50-90%. Above an D£ level of 0.5%, the effect on NOX removal
remains constant.
The size of the reactor vessels should be similar to other dry NOX
removal processes with reaction temperatures in the range of 350-400°C
(662-752°F) since the space velocity is about 7,000-10,000 hr"1. An
internal bucket conveyor is used for transferring the catalyst from the
242
A
-------
regenerator to the top of the reactor of the 30,000 Nm3/hr size prototype
unit. Since the catalyst flow rate Is small, the particular conveyor used
has a capacity over 100 times larger than necessary. Therefore, conveyors
would be available that possess the capacity required for a 50Q-MW-size
plant.
The regeneration unit for the 30,000 Nm3/hr capacity treatment facility
has electrical heating. This method of heating would be too expensive for
application to a large plant. The regeneration of the catalyst with hot air,
rather than electrical heating, would be the preferred regeneration mode for
large utility applications.
In retrofit application on a boiler with a cold ESP after the air
heater, treatment of gas from the cold ESP would require reheat for the gas
to reach reaction temperature. If a desulfurizatlon unit is also present,
treatment of gas from the FGD exit would require even more reheat. In the
above cases, the hot ESP shown in the process flowsheet would be eliminated
but a heat exchanger would be required for energy savings by allowing the
hot treated gas to preheat the incoming, untreated gas. The operating costs
for the above would naturally be greater since more reheat is required than
for the case shown In the process flow diagram.
For treating flue gas from a 500-KW plant, several NOX removal trains
will be necessary. The turndown capability would primarily be the removal
of a train from service. The auxiliary heaters provide reaction temperature
control during boiler operating fluctuations.
The majority of the material of construction is carbon steel.
Environmental ConsIderations
There are no byproducts with the Kurabo process. Kurabo states that
there were problems with NH4HS04 formation early in development tests but
that, since a lower NHjtNOx mol ratio and better control have been used,
no problem with the formation has been encountered. Kurabo representatives
state that if NH4HS04 did form, that steam and 1^0 would easily remove it;
however, this procedure would require shutting down the NOX removal unit and
also a method for disposal of the washing solution. The NHg in the outlet
is reported to be <10 ppm. The catalyst is not expensive enough to reclaim
entirely, but the heavy metal portion is reclaimed.
Greater than 80% removal of particulate is reported by Kurabo. However,
the inlet dust content was only about 0.1 g/Nm3 (0,043 gr/sft3) which is
much smaller than particulate loading typical of a coal-fired boiler (about
15 g/Mn3; 6.4 gr/sft3). The maximum particulate loading to the reactor
indicated from pilot tests is 1-2 g/Nm3 (0.43-0.86 gr/sft3). In some
instances, it may be possible to use multiclones prior to the reactor in
retrofit installation where secondary particulate control already exists
downstream of this NOX FGT system.
243
-------
Critical DataGaps andPoorly Understood Phenomena
The sensitivity of NOX removal efficiency to such factors as temperature,
space velocity, and SOX inlet concentration .are not yet reported.
Advantages and Disadvantages
The Kurabo Knorca process has the following advantages and disadvantages:
Advantages
1. Achieves >90% NOX removal efficiency
2. Operates with moderate particulate loadings (up to 1 gr/sft^)
3. Claims <10 ppm by vol NH3 in treated flue gas
Pi sadvantages
1. Has not been tested on coal-fired gas (has been tested on simulated
coal-fired flue gas)
2. Uses moving-bed reactor which increases maintenance and catalyst
attrition
3, May require auxiliary heater to attain or control reaction temperatures
244
A
-------
KUREHA PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (74, 75)
The Kureha process (Kureha Chemical Industry Company, Ltd.) Is a dry
NH3~based SCR system and is specifically designed to be installed after an
existing FGD system. Although not mentioned by Kureha, the catalyst is
apparently sensitive to SC-2 and the pretreatment step has been added to
remove most of the remaining SC-2 and dust. The process consists of three
sections: pretreatment, heat exchange, and the reduction reactor. A
block flow diagram is shown in Figure 64,
The flue gas from the boiler passes through a conventional FGD system
before entering the denitrification system. The flue gas enters the pre-
treatment section of the Kureha system where any remaining S02 and dust
are contacted with a very fine mist of alkali solution in a spray tower.
The resulting S02 and dust-contaminated mist is then reportedly removed in
a wet-type ESP.
The flue gas leaving the pretreatment section extracts waste heat from
the clean exhaust gas exiting the NH3 reduction reactor in a countercurrent
heat exchanger and is heated to 130 C (266 F). NH3 is injected and after
the temperature of the flue gas-NH3 mixture has been further increased to
150°C (30Q°F) by indirect contact with steam, it is sent to the catalytic
reduction reactor where NH3 selectively reduces NOX according to the following
reactions.
4NO, , -f 4NH,, . + 00, + 4N f , + 6H00 (231)
(g) 3(g) 2(g) 2(g) 2 (g)
6N00, s + 4NH,, , •* 6NO, , + 2N0, , -f 6H00, % (232)
2(g) 3(g) (g) 2(g) 2 (g)
Since NO represents 90-95% of the NOX in the flue gas, reaction (231) is
the primary reaction occurring in the reduction reactor,
The gas from the reactor passes through the heat exchanger to heat the
incoming flue gas before exiting through the stack.
Status of Development
The initial research and development work for the Kureha dry SCR process
was begun in 1975 with the goal of developing a system which would be rela-
tively inexpensive and could be installed after an existing wet FGD system.
The resulting process was tested in a 300 Nm-Vhr (0.1 MW equiv) bench-scale
unit during 1976. The type and composition of the flue gas have not been
disclosed. In late 1976 a 5000 Nm3/hr (1.6 MW equlv) pilot plant was built
based on the results from the earlier bench-scale unit and the operation
of this pilot plant was expected to begin in February 1977,
This process has not been tested on flue gas from a coal-fired boiler.
245
-------
STEAM
CLEAN GAS
TO STACK
\ I
a\
t
^ AIR
* HEATER
AIR
FGD
SYSTEM
^
i
ALKALI
MISTER
MIXING
TANK
^ WET
* ESP
1
WASTE WATER
PURGE
4 No OH
« WATER
t
^ HEAT
•tXCH/
i
kNtttt
>
CATALYTIC
REDUCTION
[REACTOR 1
AMMONIA
Figure 64. Flow Diagram of Kureha Process.
-------
The next development step for this process would be the continuation of
testing at their pilot plant to generate data for further scaleup to a proto-
type unit and also to demonstrate the long-term reliability of ' this system.
The ability of this unit to treat flue gas from a coal-fired boiler also
needs to be demonstrated.
Background of the Process Developer
Kureha is a medium-sized Japanese chemical company with major interests
in fertilizers and petrochemicals as well as basic organic and inorganic
chemicals. Kureha entered the air pollution control .field in 1967 with the
initial bench-scale testing of their Na£S03 FGD system. After additional
pilot-plant work during 1968, several commercial units based on this Na£SQ3
FGD technology were completed during the early 1970's. This original Na
process was modified in 1971 to incorporate acetic acid (CH3COOH) in the
scrubbing solution and this new system was renamed the sodium acetate
(NaC2H302)-CaS04'2H20 process. A bench-scale FGD unit using this NaC2H302~
CaS04'2H20 technology and treating 100 Nm3/hr (0.03 MW equlv) was' built in
1973 and continued to operate until a pilot-plant treating 5000 Nm^/hr (1.6
MW equlv) of oil-fired flue gas was completed in March 1975.
Initial investigation into the possibility of modifying their
«2H20 process to simultaneously remove S02 and NOX was begun in late
1971. A bench-scale treating 250 Nm3/hr (0.08 MW equlv) of oil-fired flue
gas was completed in 1973 and continued operation until mid-1974. The
Kureha NaC2H302-CaSOi'2H20 process for simultaneous S02~NOX removal was
scaled up to 5000 Nm3/hr (1.6 MW equiv) pilot plant In early 1976 and has
been operating since then.
Kureha has recently begun the development of this NH3~based dry SCR
process. The Initial research for this SCR process was completed in 1975
and a bench-scale unit (0.1 MW equlv) was operated during 1976. In late
1976 a pilot plant (1.6 MW equlv} was constructed and testing was scheduled
to begin in February 1977.
Kureha has offices in New York City which should make this process
more accessible.
Pub 1 ished Ecpnomi c Data
The estimated total capital investment and annual revenue requirement
for the Kureha NH3~based SCR system are considered proprietary information
by the process developer.
Raw Material , Energy , and Operation Requirements
The following raw material, energy, and operation requirements are
preliminary estimates by Kureha and are based on a unit treating 100,000
Nm3/hr (33 MW equlv) of flue gas containing 50 ppm S02» 200 ppm NOX, and
0.1 g/Nn>3 of dust. The S02 and NOX removal efficiencies were 99 and 90%
respectively. Under these assumptions the raw material requirements would
be as follows (all quantities in tons are assumed to be long tons) .
247
-------
Raw material Quantity
NaOH • 18 kg/hr
NH3 14 kg/hr
Catalyst 17.6 ton/yr
The utility requirements on the same basis are:
Utility Quantity
Electricity 200 kW
Process water 1.0 ton/hr
Steam 1.65 ton/hr
By using the FGD system operators to also monitor the denitrification
plant, no additional manpower will be required. The maintenance require-
ments have not been released.
Technical Considerations
The Kureha system is a. relatively simple system consisting of the
following three sections: a wet pretreatment section to remove particulates
and some SC>2, a heat exchange section to preheat the incoming flue gas, and
the reduction reactor. The only major process control problem will be the
injection of the exact amount of NHg required to provide an NH3:NOx mol
ratio of.(1.0-1.1):1 in the catalytic reduction reactor. , This NH3 feed
rate is automatically controlled by measuring the inlet flue gas rate and
injecting a proportionate amount of NHj.
The NOx removal efficiency is relatively independent of the inlet flue
gas composition providing that the particulates and S(>2 levels are kept
below 0,2 g/Nra^ and approximately 50 pptn respectively. These incoming par-
ticulate and S0.2 levels are reported to be further decreased to 1 mg/Nm^
of dust and 1 ppm S02 in the Kureha pretreatment section. At high dust
levels the catalyst pellets will become covered and the number of active
sites available for NOX reduction will decrease. In addition, this dust
coating will also plug some of the reactor internals and significantly
increase the pressure drop through the reactor. Failure to remove most of
the S02 in a conventional FGD system prior to entering the pretreatment
section of the Kureha process will result in the formation and precipitation
of corrosive NH4HS04 on the reactor internals and hence sharply decrease the
NOX removal efficiency.
The catalytic reduction reactor is a fixed-bed type. The catalyst
particles themselves are somewhat unusual in that they are made entirely
from the active base metal without any carrier such as Al2C>3 or activated
C. The nature of the base-metal catalyst Is proprietary but the catalyst
particles are shaped into pellets with typical dimensions of 5 mm in dia
and 10 mm in length. For 90% NOX removal the operating conditions in the
reactor ares
248
A
-------
Parameter Value
NH3;NOX mo1 ratio (1.0-1.1);! •
Temperature 150°C (302°F)
Space velocity 5,000 hr""^
Superficial velocity 0.75 m/sec
Catalyst size/shape 5 mm dia x 10 mm pellet
These operating parameters in the reduction reactor are very similar to
those specified by the other NH3~based dry SCR processes except for the
operating temperature. Since the Kureha system is added after a conventional
FGD system and the flue gas is further treated to reduce the inlet S02 con-
centration to about 1 ppm, the formation and precipitation of NH4HSQ4 should
not be a significant problem under normal operating conditions.
Unlike most of the other dry, NH3-based} SCR processes, a failure of the
conventional FGD unit will also require shutting down the NOX removal system.
By using a higher reaction temperature the other processes eliminate the
problem of NH^HSO^ formation and are 'not dependent on the FGD system per- '
formance.
Since this process is added after the FGD system, the Kureha system
should be relatively easy to retrofit to an existing power plant providing
sufficient land area is available nearby for siting the process equipment.
Although not mentioned by the process developers, it has been assumed
that two trains of denitrification equipment will be required for a 5QO-MW
coal-fired boiler. With each of these trains rated at approximately 55% of
the total capacity, this system should be able to cycle with the boiler.
With the exception of the pretreatment section, apparently only mild
carbon steel will be required for the remaining equipment. The equipment
and piping in the prescrubbing section will probably be constructed from
elastomer-lined carbon steel.
Environmental Considerations
The Kureha dry process has been able to average 90% NOX removal during
bench-scale testing. To maintain this high NOX removal efficiency the
particulate and S02 levels in the incoming flue gas must be kept to a mini-
mum. In addition, failure to maintain an NH3:NOX mol ratio of at least 1:1
in the reduction reactor will result in a decline in the NOx removal effi-
ciency.
This process, although designed for NOX removal, also removes residual
S02 and particulates.
Critical Data Gaps andPoorly Understood Phenomena
Because of the relatively new development status of the process,
critical data gaps or poorly understood phenomena are not given.
249
-------
Advantages and Disadvantages
Advantages and disadvantages of.the dry Kureha process are listed below,
Advantages
1. Achieves >90I NOX removal efficiency
2. Operates between 100 and 200 C, which may require only negligible
.reheat
Disadvantages
1. Forms secondary source of pollution (Na2S03~ash sludge waste stream)
2. Has not been tested on coal-fired flue gas
3, Requires clean (S02- and particulate-free) gas feed
250
A
-------
MITSUBISHI HEAVY INDUSTRIES PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (5, 20, 41)
Mitsubishi Heavy Industries, Ltd., (MHI) has developed a dry, SCR method
for NOX removal using NH3 as the reduction agent. A block flow diagram of the
MHI process for treating a coal-fired flue gas is shown in Figure 65 and is
based on the arrangement used in the pilot-plant treatment of dirty flue gas.
The flue gas exiting the boiler-economizer has the majority of partlculate
matter removed by passage through a hot ESP. The flue gas exiting the ESP
Is heated to reaction temperature by an auxiliary heater. The flue gas
having been injected with NHj enters the reactor where, at the required
temperature and in the presence of the catalyst used by MHI, the NOX is con-
verted to 1^2 by reaction with NH^ as follows :
4NH3(8) + 4N°(g) + °2(g) * 4N2(g) + 6H2°(g)
4NH3(g) + 2N02(g) + 02(g) -, 3N2(g) + 6H20 (234)
The treated flue gas leaves the reactor and passes through an air heater
to transfer the heat from the treated flue gas to the air feeding the boiler.
The flue gas passes from the air heater to a desulfurization unit and/or the
stack.
The reaction temperature ranges from 350-400°C (662-752°F) , the average
NH3:NOX mol ratio employed is 111 and the KOX removal efficiency averages
above 90%. The pressure drop and space velocity will vary based upon whether
a fixed- or moving-bed or parallel passage reactor Is used. The space
velocity for the fixed- and moving-bed reactor varies from 4,000-10,000 hr"-*-
on semidlrty gas while the range tested for parallel passage reactors is
1,000-2,300 hr"1 for dirty gas. MHI created a special "W"-ahaped vessel
enabling the catalyst to be intermittently moved downward through the reactor.
(Refer to critique topic Technical Considerations.) MHI is also testing a
parallel passage reactor.
Status of Development
MHI has tested the SCR methods at pilot plants on both clean and dirty
flue gas. A 10,000 Nm3/hr (3.3 Mtf equiv) capacity pilot plant treating flue
gas from an LNG-fired boiler at Minamyokohama Station of Tokyo Electric was
used for clean gas treatment studies. Several base-metal catalysts on A1203
carriers and combinations thereof were tested for NOx reduction capability
and the amount of NHj emitted in the outlet gas. The catalyst providing the
best results were composed of chromic oxide (0^03) with other oxides, e.g.,
Fe203_Cr2C>3 and vanadium pentoxide-chromic oxide ^05-0^03) . A catalyst
life test Indicated no reduction in activity after 5500 hr service.
Experiments on the dirty gas were performed on a 4000 Nm3/hr (1.3 MW
equiv) capacity pilot plant on gas from a low-S oil. The gas contained 160
ppm NOX> 80 ppm S02, 3 ppm SOj, and 20 mg/Nm3 (0.009 gr/sft3) of particulates
.251
-------
FUEL NH3
FGD
AND /OR STACK
r-o
m
t-O
Figure 65. Flow Diagram of Mitsubishi Heavy Industries Process.
-------
The metal catalysts evaluated Include nickel (Ni), Co, Fe, manganese (Mn),
chromium (Cr), V, and Co and combinations thereof. The most SOX-
reslstant catalyst carriers were T102 and Si02- There were two
unite in the pilot plant. One incorporated a fixed-bed reactor and the
other unit contained an intermittently moving-bed reactor. There was no loss
in catalyst activity after a 1500-hr test with the fixed-bed reactor. A list
of some of the pilot plants operated or planned for operation by MHI is given
in Table 28. It should be noted that both moving-bed reactors and parallel
passage reactors are being investigated. The NOx removal facilities for
treating clean or semidirty gas are considered commercially available. Tests
have been conducted on coal-fired boilers for at least 660 hr with the use of
a hot ESP. Future tests are planned for sending the flue gas containing
15 g/Nn\3 of particulate directly to the parallel passage reactor.
Background of Process Developer
MHI has been developing a SCR process since 1973 in laboratory and pilot-
plant tests. Two NQX removal facilities are scheduled to be constructed by
MHI for Kyushu Electric in 1978-1979 on two 6GO-MW LNG-fired boilers. Other
FGT units constructed by MHI include removal facilities for 382,000 (127 MW
equiv) and 200,000 Nm3/hr (67 MW equiv) heavy oil-fired boilers. These were
scheduled to start operation in September 1976 and December 1977 respectively.
Published Economic Data
No economic data have been published.
RawMaterial, Energy, and Operation Requirements
There is no information stated by MHI on raw material and energy require-
ments. However, it may be estimated that 20 short tons/day of NH3, the major
raw material required, is necessary to treat flue gas from a 500-MW plant
containing 600 ppm NQX if an NH3;NOX mol ratio of 1:1 is used. The major
energy sources required are fuel for the heater and electricity. No man-
power, maintenance, and technical support requirements are available.
Technical Considerations
The MHI process includes straightforward and simple steps and is much
less involved than wet NOX removal systems. The NH3 injection rate would
probably be controlled by the flue gas rate and NOX concentration. The fuel
rate to the auxiliary heater would be controlled by the reactor temperature.
During the pilot-plant work on clean gas, the effect on NOX removal of
flue gas Q£ content was studied and the results are shown in Figure 66. (The
particular catalyst used is not known.) As can be seen, with a space velocity
of 10,000 hr""l and ah NH3:NQX mol ratio of 1:1, the NOX removal efficiency
with 0% Q£ in the flue gas was very low as compared to NOX removal efficiency
with levels of 0.5% Q£ and greater. For example, at 375°C, the denitrifi-
catlon achieved was about 90% at a.1%.02 level while only about 50%
removal was achieved at an D level of 0%.
253
-------
TABLE 28. ACTUAL OR DESIGNED OPERATING CONDITIONS'
OF NOX REMOVAL PILOT PLANTS (41)
BY MITSUBISHI HEAVY INDUSTRIES
Ul
Inlet flue gas composition
Flue gas source
LNG
Low S oil
Low S oil
High S oil
High S oil
-
High S oil
Low S oil
Sintering mill
Heavy s oil
Heavy s oil
Heavy S oil
Low S oil
Coal
Capacity, SO*,
Nm /hr ppm
1,000
5,000
60
1,000
500
10,000
10,000
10
500
240
500
2,000
40,000
600
0
200
200
900
600
100
1,050
100
400
800
600
800-1,000
150
1,000-1,900
Particulates,
mg/Nm3
0
20-130
20-130
70
60
10
50-100
10-20
-
100-150
60
30-70
15-30
15,000b
Space
velocity, hr
10,000-20,000
5,000-10,000
6,000
10,000
7,000
10,000
8,000-10,000
10,000
8,000-10,000
2,300
-
-
1,000
1,000
Exit NH3
NOx minimum
concentration, removal efficiency,
Reactor type ppm by vol 7,
Fixed
Fixed and moving
Fixed
-
-
Fixed and moving
Fixed and moving
Fixed
-
Parallel passage
Parallel passage
Parallel passage
Parallel passage
Parallel passage
10
20
10
10
7
5
5
8
5
-
-
15
10
-
90 at 10-,000 hr"1 S.V.
90 at 5,000 hr'1 S.V.
90 at 6,000 hr"1 S.V.
90
95
95
95
95
96
-
-
85
90
90
Operating
period, hr
10,000
5,000
2,600
-
6,000
4,000
2,500
6,200
3,000
Design
Design
Design
Design
Design^
a. The data reported for the latter five units are design conditions; other data are from actual test operations.
b. Tests have been completed for 660 hr operation during which the particulate concentration was 15 g/Nm3. The design conditions assumed to have
been met.
-------
100
d* 80
S3
Ln
o
o
z
UJ
x
O
z
60
40
20
200
%02
I
I
250
300 350 400
TEMPERATURE, °C
450
500
Figure 66. Results from MHI's Pilot-Plant Tests Showing Effect of
Oxygen on NOX Removal Efficiency (20).
-------
In early bench-scale experiments on fixed beds with the dirty flue gas,
plugging problems were experienced on catalyst 3 mm in diameter. Catalyst of
8-mm diameter reduced the problem of plugging but the NOX removal efficiency
decreased. Therefore, later work used a catalyst of 4-6 mm in diameter.
The W-shaped vessel used in the interior of the moving-bed reactor as
shown in Figure 67 was developed by MHI. In tests at the pilot plant on the
dirty flue gas, the moving bed was used with a catalyst layer 100-200 mm deep
and in the absence of an ESP. The pressure drop increased from 7 to 13 mm
H20/day till it attained 160 mm H20; at this point 10-20% of the catalyst
would be removed from the reactor and supplanted by moving the bed. This
action lowered the pressure drop to about 60 mm t^O as shown in Figure 68.
The catalyst exiting the reactor is screened to remove dust and is circulated
back to the top of the reactor. In comparison to most dry HQX removal
processes, the size of equipment required will probably be similar. The
reaction temperature of 350-400°C is typical while the space velocity of
4,000-10,000 hr is in the average range of other dry processes.
If retrofitted to a plant after a cold ESP located after the air heater,
a large amount of reheat would be required. An even greater amount of reheat
is necessary if the NOX removal facility has to be located after an FGD unit.
A heat exchanger would be required for energy savings by transferring heat
from the hot treated gas to the untreated gas going to the reactor. Operating
costs for any of these retrofit cases would be greater than the process shown
in the flow diagram. If tests prove that it is feasible to send the flue gas
directly to a parallel-passage reactor without particulate removal, this
scheme will provide a significant saving in capital and operating costs.
However, this process would probably not be suitable for retrofit and lack
of space between the economizer and air heater in existing plants.
Since several NOX removal trains are needed for treating flue gas from a
500-MW plant, the main turndown control would be removal of individual trains.
The auxiliary heater would provide better reaction temperature control during
variations in boiler operations.
The materials of construction are not given.
Environmental Considerations
The sensitivity of NOX removal to reaction temperature and space velocity
is indicated in Figure 69 for tests performed on fixed-bed reactor with dirty
flue gas and using an ESP, With an NH3:NOX mol ratio of 1:1, a apace velocity
of 10,000 hr""1, and at 360°C, about 90% NOX removal was achieved. Figure 70
shows sensitivities of NOX removal efficiency and NH3 exiting in the outlet
gas to boiler load.
The amount of NH3 which may be expected to exit in the outlet gas should
average <10 ppm as shown In Figures 70 and 71 from tests on clean flue gas
and in Figure 72 from the work on dirty flue gas. The results from pilot-
plant teats with clean gas when NH-j emissions were lowered by a catalytic
256
-------
CATALYST
•Q.
Figure 67, Structure of Mitsubishi Heavy Industries'
Moving-Bed Reactor (20),
257
-------
200
oo
160
O
-------
100
o
t
90
UJ
cc
x
O
80
SV--4000JU
.1
NH3/NOX =1.0
I
I
280 300 320 340 360
TEMPERATURE,°C
380
400
420
Figure 69. Results of Tests with Fixed-Bed Reactor and Dirty Flue Gas After Use of an ESP
Showing Effects of Temperature and Space Velocity on NOX Removal Efficiency for MHI Process (20)
-------
100
90
I*.
tU
I 80
i
OC
x
O
z
70
INLET NOX-50-130 PPM
LJ
8
ou
20
10
0
2
1 1 1 1
^
1 1 1 i
40 270 300 330 360 390
i i i i
4000 6000 9000 12000 SPACE VELOCITY, HFT1
i i t i
8432
1 i i •
02,%
1 BOILER LOAD
NO Removal Efficiency and Outlet Gas
Figure 70.
Concentration at Different Boiler Loads for MHI Process (20).
260
A
-------
lit Q.
-I Q.
40
20
0
INLET NO* =
140-170 PPM
SV= 12,000 MR'1
340-360°C
I
J_
I
I
1000 2000 3OOO 40OO
OPERATING PEROID.HR
50OO
Figure 71. NOX Removal Efficiency and NHj Emission in
Relation to NH3:NOX Mol Ratio and Operating Period for MHI
Process (20).
-------
100
80
UJ
60
40
CL
Ou
20 —
UJ
4000
CATALYST DIAMETER =4-6mm —I
INLET NOX =!50ppm
380°C
NH3:NOX «l:l
6000 8000
SPACE VELOCITY VALUE, HR'1
IO.OOO
Figure 72. Effect of Space Velocity on NOX Removal Efficiency
and NH3 Emission for MHI Process (20).
-------
decomposer located after the reactor are shown in Figure 73. A denitrifl-
cation efficiency of 97% was obtained while the NH90% NOX removal efficiency
2. Has been applied to flue gas from commercial oil-fired boilers
3. Claims <10 ppm by vol NH3 in treated flue gas
Disadvantages
1. Requires somewhat clean, (indefinite level of particulate removal
required) gas feed
2. Uses moving-bed reactor which increases maintenance and catalyst
attrition
3, Requires auxiliary heater to attain or control reaction temperatures
263
-------
IOO
s
o
E
90
oc
X
o
30
Q.
£L
20
ID
0
REACTOR OUTLET
CONVERTER OUTLET
INLET NOX = I50PPM
NH3/ NOX - l.O
20O 250 300 350 400
CATALYST TEMPERATURE ,*C
450
Figure 73. Results of MHl's Tests on NOX Removal Efficiency and NH3
Emission Using an NH3 Converter to Reduce NH3 Emission (20).
264
A
-------
i.50
1.20
o
CO
o 0.90
10
Z
0,60
0.30
0-150
O
a.
UJ
o
o*
05
I
120
90
60
30
r i r r
1 r
i
295 280 265 250 235 220 205 I9O 175
GAS TEMPERATURE»*C
I6O 145
Figure 74. Deposit of 1^3-804 Compounds in Air Reheater from mil Tests (20),
265
-------
MITSUBISHI KAKOKI KAISHA PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (41)
The Mitsubishi Kakoki Kaisha, Ltd., (MKK, also Mitsubishi Chemical
Machinery) NOX removal process is based on dry, SCR of NOX with NH^. The
catalyst is manufactured by Santetsu Kogyo KK and is referred to as "SARC"
(Santetsu Ammonia Reduction Catalyst). It is a specially treated Fe20
and Is made without any carrier .
As shown in figure 75 for a coal-fired operation, MKK proposes to treat
the flue gas before the air heater by passing it through a hot ESP, injecting
NHj, and passing the mixture through a reactor containing the SARC catalyst.
In the presence of the catalyst and at the reaction temperature range of
400-450 C (752-842°F), NOX IB reduced to N£ by NH3 according to the following
reactions.
6N°(g) + 4NH3(g) * 5N2(g) + 6H2°(g) <235>
6N02{g) + 8NH3(g) + 7N2(g) + 12H20(g) (236)
4N°(g) + 4NH3(g) + °2+ 4N2(g) + 6H2°(g)
The treated flue gas exits the reactor and passes through the air
heater for heating the incoming air which feeds the boiler. The treated gas
is then sent to the desulfurization unit and finally to the stack.
The required reaction temperature varies according to the partlculate
loading and 803 level. MKK recommends 400-450 C for dirty gas with SOo as
for a coal-fired boiler and 250-300 C for a clean gas with a trace of §63.
The NH3:NOX rool ratio varies from 1:1 to 1.3:1. The pressure drop
across the reactor is normally about 80-160 mm E^O, but it can reach a maxi-
mum value of 400 mm H20. The space velocity varies as a function of the
dust content with a 3,000-10,000 hr range. The space velocity for the
coal-fired boiler case is usually about 3000-5000 hr~^. Also, the most
favorable gas linear velocity in the catalyst bed is 3-5 m/sec.
A fixed-bed reactor is applied in this case rather than a moving-bed
system which increases the wear on the catalyst and thus increases catalyst
consumption, A moving-bed system also results in larger maintenance cost.
Status of Development
MKK has been working on NOX removal since 1972. The initial effort was
in the study of wet simultaneous SOX and NOX removal processes. However,
now the emphasis is on the dry, SCR method. Several bench-scale tests have
been made. From March to. July 1975 a test on a 800 Nm3/hr (0.27 MW equiv)
capacity bench-scale unit was made using flue gas from a coke oven. A test
266
A
-------
NH3
-i*FGD AND /OR
STACK
AIR
Figure 75. Flow Diagram of Mitsubishi Kakoki Kaisha (MKK) Process.
-------
employing furnace flue gas at a rate of 100 Nm^/'hr (0.03 MW equiv) began in
June 1975, while in July another run began on a low-S, heavy oil flue gas
source with a gas capacity of 1400 Nm3/hr (0.47 MW equiv). The latter two
test units were still in operation as of April 1976. Also as of the same
date, a pilot plant treating 1500 Nm /hr (0.5 MW equiv) of gas from a high-S
heavy oil-fired boiler had been in operation since September 1975, A 2000
NnrYhr (0.67 MW equiv) capacity plant treating flue gas of a steel-ore sin-
tering machine began operation in January 1976.
Background ofProcessDeveloper
MKK has supplied at least 13 Wellman-Lord desulfurlzation units with
capacities of 30,000-1,200,000 Nm3/hr (10-400 MW equiv) since June 1971.
Four lime-CaSG4'2H20 FGD plants have been constructed since March 1973 with
35,000 to 250,000 Nm3/hr (12-83 MW equiv) FGT capacity. Two other FGD plants
treating 70,000 to 370,000 Nm3/hr (23-123 MW equiv) flue gas have been built
since June 1973.
There are two pilot*-plant size NQX removal plants constructed by MKK
which were to start up in 1976. One has a capacity of 5000 Nm3/hr (1.7 MW
equiv) with clean flue gas from a heating furnace. The other plant treats
14,000 Nm3/hr (4.7 MW equiv) of flue gas from a boiler fired with high-S
fuel oil at Nihon Yakin, Kawasaki Works.
The licensing rights for the SARC catalyst in the U.S. are held by
Pfizer, Inc. This arrangement should make the accessibility of the MKK
process more convenient to the U.S. The catalyst supply capacity of San-
tetsu is 300 tons/mo and MKK claims it holds top priority in having catalyst
orders supplied.
Published Economic Data
The capital investment for a 400-MW plant is reported to be $14.33M
which is equivalent to $36/kW (41). The revenue requirement is 1.8 milla/kWh
or $5.1 M/yr (41), These costs are on the basis of 150 ppm NOX, 7000 hr/yr,
space velocity = 5000 hr , catalyst entirely replaced every 2 yr, and four
equal NOX removal trains. The catalyst cost is listed as about $8,000/metric
ton. These figures are based on $1 = 300 yen, 1976 costs, and a Japanese
location.
Raw Material, Energy, and Operation Requirements
For the 400-MW basis, as above, the raw material and energy require-
ments are as follows:
Material Quantity
Electricity 2360 kW
Steam 2.9 metric tons/day
NH3 4.0 metric tons/day
Catalyst consumption 0.18 metric tons/day
268
A
-------
The electrical usage represents 0.6% of the total equivalent plant
power output. The estimated labor required for this plant is 12 man-hour/
day. No maintenance or technical support requirements are available but
they should be similar to other fixed-bed, dry KOX removal systems and less
than that required by wet processes.
Technical Considerations
This system is rather simple and automatic control of the NHo flow rate
is possible, though not mentioned by MKK. The size of the required reactor
should be slightly larger than some of the different processes since it has
a space velocity of 3000-5000 hr~* and reaction temperature of 40Q-450°C
(752-842°F).
For retrofit application on a system with a cold ESP and FGD unit it
would be necessary to use an Inline heater to heat the gas to the 400 C
reaction temperature and a heat exchanger to recover the heat from the treated
gas for the incoming, untreated gas. Less reheat would be needed if the gas
could be treated prior to the FGD unit. Even for application on a new coal-
fired plant, an Inline heater triay be required between the hot ESP and reactor
since the gas from the boiler-economizer is most often considered to be only
375°C (705 F).
The SARC catalyst is F^O-j recovered from a steel pickling waste liquor.
The catalyst can be made into any shape but presently is produced in five
standard sizes. SARC I-IV are all ring shaped and are progressively smaller
in size while SARC-P is a cylinder-shaped pellet. This cylinder-shaped
catalyst is recommended for flue gas with high particulate loading and where
stacking of catalyst would be required.
The turndown capability of this system is mainly by removal of an NOX
removal train from service. There would be several trains required for
treating gas from a 500-MW-size plant.
The materials of construction are not given by MKK,
Environmental Cpnslderations
MKK claims that there is no problem with the excess NH3 exiting the
system with a reaction temperature of 400 C (752°F) or greater while the
excess NH3 will exit at temperatures below 250 C (482 F). No mention of
NH^HSOj formation is made. There is no removal of pollutants other than NOX.
For a pilot plant with oil-fired boilers, >90% NOX removal was attained
with NH3:NOX mol ratio of 1.2:1, reaction temperatures >300°C (572°F),
space velocity of 5000 hr , and inlet gas concentration of 200 ppm NOX and
400 ppm SOX. An 80% NOX removal efficiency can still be reached with higher
space velocities, such as 5,000-10,000 hr"~*. Tests underway on a steel-ore
sintering plant have demonstrated that high level of Cl~ (exact level not
specified) do not adversely affect the SARC catalyst.
269
-------
Q
At the 14,000 Nin /hr pilot plant at Nihon Yakin, the operating conditions
employed are as follows;
Reactor temperature 350 C
NH3:NQX mol ratio 1.2:1
Space velocity 3,000 hr""1
Linear velocity 6,8-7.3 m/sec
Inlet gas SOX content 800-1,000 ppm
Inlet gas NOX content 100-150 ppm
The NOX removal which was achieved was about 95%. The catalyst arrangement
was then changed from a staggered to a stacked packing arrangement (decreasing
effective catalyst surface by 19%) to reduce accumulation of particulate
matter in the catalyst bed. The NOX removal subsequently declined from 95
to 85%. The man-hours required for stacking the catalyst on this 4.7-MW
equiv^plant were 160.
The waste catalyst is a potential raw material for pigment production
which may eliminate problems with disposing of the catalyst.
Critical Data Gaps and Poorly UnderstoodPhenomena.
The catalyst life is assumed to be 1 yr as reported by MKK in most
instances. However, in the economics, MKK used a catalyst replacement period
of 2 yr as a basis. There are also no statements by MKK on the effects of
particulate levels >0.1 g/Nm-* (0.04 gr/sft-*). There is no mention of mater-
ials of construction or maintenance requirements.
The maximum SOX level tolerable by this catalyst is not available. MKK
claims no adverse effects with a high SOX level; however, the maximum level
which has been tested apparently is 1000 ppm. The average level which can
be expected for a coal-fired power plant flue gas is about 2400 ppm. The
expected NH90% NOX removal efficiency
2. Uses catalyst which, when spent, has other potential applications
Pis advantages
1. Has not been tested on coal-fired flue gas
2. Requires clean (Indefinite level of particulate removal required)
gas feed
3. Incorporates questionable design features (stacking of catalyst in
reactor may be required)
270
A
-------
MITSUBISHI PETROCHEMICAL PROCESS - DRY, SCR (NO,.)
A
Process Description and Principles of Operation (54, 85)
Mitsubishi Petrochemical (MFC) is offering a highly efficient catalyst
for the SCR of NOX with NH3- Although the company is not offering a process,
the use of its catalyst would be the main difference among other dry SCR
processes; therefore, it has been treated as a process in this review. The
exact nature of the catalyst is not given by MFC but it is possibly a
titanate as reported by Dr. Jumpei Ando (Chuo University, Tokyo, Japan).
With respect to NO, MFC states that with their catalyst the reduction
of NO by NH3 requires the presence of 02 and that the reaction proceeds as
follows:
4NH,, . 4- 4NO, s + 0,, * + 4N + 6H00. (238)
3(g) (g) 2(g) 2(g) 2 (g)
With N0£ present, the following reaction takes dominance over the pre-
ceding reaction, when the NO:N02 mol ratio exceeds 1:1.
NO, . + N00/ ,+ 2MB,. . + 2N,, x + SH^O. . (239)
(g) 2(g) 3(g) 2(g) 2 (g)
The NH3JNOX ratio required is 1:1. The On concentration and
ratio versus NOX removal efficiency are shown In Figures 76 and 77,
There are three types of "PNX" catalysts manufactured by MFC for use at
different conditions. PNX-A is suitable at high temperature and has a
temperature range of 350-600°C (662-1112°F) . Type B has application at low
temperatures with a range of 300-450°C (572-842°F), while type C is used at
temperatures in a range of 300-600°G (572-1112°F). At a space velocity of
10,000 hr"1, each type is capable of NOX removal of >95%. The estimated life
is 1 yr for types B and C and 2 yr for type A. Types A and C are suitable at
high SOX concentrations (at least 2000 ppm) while B is suitable at 500 ppm
Status of Development
MFC screened several hundred catalysts for use in an NOx removal
facility treating flue gas containing SOX. A series of catalysts were
developed which are more resistant to SOX, Development is at the stage where
the company can supply catalyst samples.
Background of Process Developer
Little information has been obtained on the background of MFC. However,
MFC has a New York office making the catalyst more accessible to the U.S.
market.
271
-------
100
- 80
>
o
z
UJ
o
uT 60
u_
UJ
> 40
O
s
UJ
o:
c?20
CAT = PNX-A-30-42 MESH
CYLINDER
NO
NH3
H20
200 ppm
400ppm
9%
SV = 10000 hr "'
TEMP. * 350° C
0.3 0.5 I 2345 10
02 CONCENTRATION , %
20
Figure 76. Effect of 02 Concentration on NOX Removal Efficiency with MFC Catalyst (54)
-------
100
80
o
UJ
o 60
It
UJ
g 40
5
UJ
U>
X
O
20
NO = 300-800ppm
N02S !5-40ppm
TEMPERATURE =320°C
1.0
MOLS NH3 PER MOL NOX
2.0
100
80
60
40
o
o
10
I
UJ
IO
I
»-
UJ
20
0
Figure 77. Effect of NH3:NOX Mol Ratio on NO Removal
Efficiency with MFC Catalyst (85). X
-------
Published Economic Data
The cost of the catalyst is not reported by MFC.
Raw Materials, Energy,_and_0ge^tion ^Requirements
MFC is offering only a catalyst and not a process
Technical Considerations
As stated previously, PNX-A catalyst is for treatment of high-temperature
gas, PNX-B catalyst is for low-temperature gas, and PNX-C is for temperatures
between the two, Figure 78 shows the relationship between temperature and
NOX removal efficiency for each catalyst. Figure 79 gives relationships of
temperature and space velocity for each catalyst at actual size, 3 mm by
3 mm.
PNX-A and PNX-C catalysts are very resistant to S02 poisoning and capable
of being used with levels of up to 2000 ppm S02- PNX-B is recommended for up
to 500 ppm S02 level. Low SQ2 oxidation to SOj is also claimed for MFC
catalysts, which cause less corrosion and plugging problems of the equipment.
The catalysts are resistant somewhat to dust plugging. In the event of
heavy particulate loading however, the use of (a) a moving catalyst bed or,
(b) fixed bed with specially shaped catalyst may be required. The estimated
catalyst life is 2 yr for PNX-A and 1 yr for types B and C.
The catalysts may be manufactured into any shape but are primarily
designed into the following shapes and sizes:
Shape • Diameter (mm)
Cylinder 6, 8, 10, and 15
laschig ring 8x4, 10 x 5, and 15 x 7 (outside dia
x inside dim)
Pipe (under development) 30 x 22, 25 x 18, and 20 x 14 (outer
dia x Inside dia)
The cylindrlcally shaped catalyst is more suitable for a moving-bed
reactor with gas containing a high level of particulates because of its
higher physical strength. The Raschig ring-shaped catalyst has a higher
activity and lower pressure drop and is more suitable for fixed-bed reactors
with gas having low dust concentration. For a fixed-bed reactor with gas
containing relatively high concentrations of particulates, the pipe-shaped
catalyst is expected to be more suitable.
Envi ronmental Cons tderatione
The MFC catalysts allow the use of the equimolar mixtures of NH3 to NOX
to keep formation of ammonium (NH^) salts at a minimum level.
274
-------
100
80
o
UJ
fc 60
UJ
-------
100
80
60
38 100
o
z
£ 80
i so
UJ
x
o
60
40
20
350°C —
CAT.-*PNX-
300*350°C
250° C
CAT. « PNX-B
300° C
CAT. =PNX-A
N0«200-220ppm
NH3=220-250ppm
S02 B 300 - 400ppm
H20= 6.63%
02*II-I2%
C02B 4.9%
CO-0.023%
.CAT SIZE* 3mm0 x 3mml cylinder
I I I
5000 7500 10000 12500
SPACE VELOCITY, Hr"1
15000
Figure 79. Relationship of NOX Removal Efficiency to Temperature
and Space Velocity for Various MFC Catalysts (54).
276
A
-------
Critical Data Gaps and PoorlyUnderstood Phenomena
The exact composition of the catalyst is unknown though it is said to be
a titanate. There is also no mention of catalyst cost. Types A and C
catalysts are applicable with flue gas containing up to 2000 ppm S02 and 60
ppm 803 but it is not stated if higher levels, such as 2400 ppm S02 with coal-
fired boilers, are acceptable.
Advantages andDisadvantages
The advantages and disadvantages of the MFC catalysts are as follows:
Advantages .
1. Achieves >90% NOX removal efficiency
Disadvantages
1. Has probably not been tested on coal-fired flue gas
2. Requires clean (indefinite level of particulate removal required)
gas feed
277
-------
MITSUI ENGINEERING AND SHIPBUILDING PROCESS - DRY, SCR (NOX)
Proceas Description and Principles ofOperation (16, 86)
The Mitsui (Mitsui Engineering and Shipbuilding Company, Ltd.,) dry
process is an N0x-only removal system based on the SCR of NOX by NH3.
However, the Mitsui process can be tailored to fit the particular flue gas
conditions at a specific site by inserting any one of three different types
of catalysts (NXC-100, dust and SOX free; NXC-200, dust-free; NXG-3QO, dirty
gas) developed by Mitsui Petrochemical Industries, Ltd. Since only coal-
fired gas is being considered in this study, the NXC-300 catalyst designed
for heavy oil-fired flue gas will be considered. This dry SCR process is
very similar to the other dry processes in that it consists of three distinct
sections; particulate removal, NH3 injection, and catalytic reduction. The
basic outline for the Mitsui process can be seen from the block flow diagram
given in Figure 80.
The flue gas from the economizer at 400 C (750 F) passes through a hot
ESP to reduce the particulate levels below 0.2-0.4 g/NnH (0.09-0.17 gr/sft^
at 32 F) before entering the NH3 mixing chamber. Gaseous NH3 in an NH3:NOX
raol ratio of (1.0-1.1):! is injected and intimately mixed with the flue gas
before entering the reduction reactor containing the proprietary base-metal
catalyst. As the mixture passes through the packed-bed reactor at 350 C
(660 F) the NOX is selectively reduced to N£ by the following reactions:
6NO, , + 4NH,, . ->• 5N-, , + 6H,0, , (240)
(g) 3(g) 2(g) 2 (g)
6NO,,, x + 4NH,, v -*• 6NO, , + 2N0, x + 6H00, % (241)
2(g) 3(g) (g) 2(g) 2 (g)
4NO, , + 4NH,, , + 00, , -t-4N0, ,. + 6H00, , (242)
(g) 3(g) 2(g) 2(g) 2 (g)
Since NO represents 90-95% of the NOX in the flue gas, reactions (240)
and (242) are the primary reactions occurring in the reactor.
The clean flue gas from the reactor passes through the boiler air heater
and then the FGD section before exiting through the stack.
Status ofDevelopment
The Mitsui dry SCR process using the NXC-300 catalyst was initially
tested in a bench-scale unit treating 200 Nm-Vhr (0.06 MW equiv) of flue
gas having the following composition; 1000 ppm S02, 300 ppm NOX> 3% 02,
and 0.1-0.2 g/Nm-^ dust. During a recent 6000-hr (250 day) continuous opera-
tion, the bench-scale unit was able to maintain 90% NOX removal.
In October 1976 a pilot plant treating 5000 Nm3/hr (1.6 MW equiv) of
flue gas (SOX:1500 ppm, NOX:150 ppm, and 0.2 g/Nm3 dust) from a 500-MW power
plant burning heavy fuel oil was completed and began operating. During a
2300-hr (96 day) test, this pilot plant using the NXC-300 catalyst was able
278
A
-------
KJ
1
BOILER
ECONOMI;
:ER »
"HOT"
ESP
ACM
i
M
1
U
REACTOR
J
I
AIR
HEATER
AIR
- 4* TO FGD
SYSTEM
Figure 80. Flow Diagram of Mitsui Engineering and Shipbuilding Process.
-------
to maintain 90% NOX rejnoval, In addition methods of preventing NH4HS03 pre-
cipitation in the air heater downstream of the reactor are now being studied.
Recently the Mitsui process using the NXC-30Q catalyst was selected by
Mitsui Petrochemical Industries for a system to treat 200,000 Nm^/hr (67 MW
equiv) of flue gas from a heavy oil boiler. Preliminary plans indicate a
July 1977 startup for this commercial unit.
This process has not been tested on flue gas from a coal-fired boiler,
therefore, its ability to handle the flue gas from a coal-fired boiler needs
to be evaluated.
Background of. Proceaa Developer
Mitsui is a relatively large and diversified Japanese corporation which
entered the air pollution control field in the early 1970's. In addition to
the development of their lime slurry FGD system, their early involvement in
air pollution control also Included a licensing agreement to build the Dowa
Al2(S04)3'Al203-CaS04'2H20 FGD system.
In response to the adoption of stringent NOx regulation in Japan, Mitsui
began the development of their SCR process. In the initial testing of .the
Mitsui process, the NXC-100 catalyst was used and thus the testing was limited
to flue gas containing neither dust nor SOX (i.e., HNOj tail gas or natural
gas-fired boilers). The Mitsui process using the NXC-100 catalyst was
recently (October 1975) Installed at ' the Chiba Works of • the Mitsui Petro-
chemical and is treating 200,000 Nm3/hr (66.7 MW equiv) of flue gas.
Two new catalysts, the NXC-200 and NXC-300 series, were developed by
Mitsui Petrochemical. These two catalysts are more active than the NXC-100
series and, even more important, are resistant to SOX. The NXC-200 series
is specified for flue gas containing minimal amounts of dust and particulates
while the NXC-300 series was designed for duat-laden gases. The NXC-300
series was first tested in a 200 Nm-Vhr bench-scale unit and in October 1976
began operating in a 5000 Nm^/hr pilot plant.
At the present time no American company has been licensed by Mitsui to
market this process in the U.S. However, Mitsui does have a liaison office
in New York City, through which inquiries about this process are handled.
Published Economic Data
Since this process has only recently been tested in a pilot plant, the
following published economic data should be considered as a preliminary
estimate and subject to revision. Mitsui estimates (16) the total capital
investment for a system to treat 1,000,000 Nm /hr of flue gas (corresponding
to a 333-MW power plant) from a heavy oil-fired boiler as $3,600,000
(assumed construction basis: Japan and 1976 costs but excluding foundation
and civil work), or approximately $10/kW of generating capacity. The annual
revenue requirements for this process have not been published.
280
A
-------
Raw Material,Energy, and Operation Requirements
The following raw material requirements were estimated by Mitsui based
on a plant to remove 90% of the NOX from 1,000,000 Nm3/hr (333 MW equiv) of
flue gas. This inlet flue gas was from a heavy oil-fired boiler and was
assumed to contain 170 ppm NOX and 0.2 g/Nm-* of dust.
Material Quantity required
N% 129 kg/hr
Catalyst $1.2M/yr
This catalyst cost is based on a 1-yr lifetime but recent information indi-
cates that a 2-yr lifetime of the catalyst may be obtained and, hence, the
catalyst cost could be substantially reduced. The utility requirements
using the same basis were estimated as:
Utility Quantity required
Steam 80 kg/hr
Cooling water 9 tons/hr
Electric power 2900 kW
Land 2500 m2
This electrical requirement represents only 0.9% of the generating capacity
of the boiler.
Essentially no operating manpower is required for this system since it
is automated but occasional monitoring may be required. The maintenance
requirements have not been published but are expected to be a minor cost.
TechnicalConsiderations
The Mitsui process, as is typical for most dry, N0x-only removal pro-
cesses, is a relatively simple system with only five pieces of equipment:
a hot ESP, an NH3 storage tank, an NH3 vaporizing tank, a mixing device,
and the reduction reactor. The only major process control problem is vapor-
izing and injecting the exact amount of NH3 required to provide an NH3:NOX
mol ratio of 1:1 in the reduction reactor. This NH3 feed rate is controlled
automatically by measuring both the flue gas rate and the concentration of
HOX in the gas.
The NOX removal efficiency is relatively independent of the inlet flue
gas providing the particulate levels are below 0,2 g/Nm3. Significantly
higher particulate loadings and/or the presence of alkaline (i.e., Na, K,
Zn, etc.) 803* or Cl~ in the flue gas will decrease the NOX removal efficiency
since they form a layer over the catalyst particles. The denitriflcation
efficiency under normal operating conditions is completely Independent of
the H20, 02, and NOX concentrations in the flue gas. As long as the tempera-
tuee in the reactor remains near or above the recommended operating tempera
-------
S<>2 concentration. If the temperature falls below 350 C, significant amounts
of NHAHSO^ can form and plug the reactor and, thus, decrease the removal
efficiency.
The catalytic reduction reactor will be the largest piece of process
equipment and Is apparently a packed-bed reactor. The catalyst particles
are typically pipe shaped with diameters ranging from 5-50 ma and lengths of
300-1000 mm. These particle shapes are claimed to be highly resistant to
dust. Although these catalyst particles are claimed to be highly active,
the apace velocity through the catalyst bed appears to be limited to approx-
imately 5000 hr~^ to guarantee 90% NOX removal. However, even with dust
levels of 0.1-0.2 g/Nm^ of dust, there was no increase in pressure drop with
time during a 6000-hr test and it remained at about 100 mm of H20. Thus,
for the purposes of this study, the operating conditions in the reduction
reactor would be the following:
Parameter Value
NH3:NOX mol ratio 1:1
Temperature 350°C (660°F)
Space velocity 5000 hr"1
Catalyst size-shape 5-50-mm-dia pipe
Dust level (max tested) 0.2 g/Nm3
These operating parameters in the reduction reactor are very similar to
those specified by the other SCR processes.
This process could not be easily retrofitted onto an existing 500-MW
plant equipped with cold ESP. The ductwork modifications required to remove
and return the flue gas between the economizer and the air heater would be
extremely difficult and expensive. Thus, for retrofit applications, Mitsui
recommends the flue gas be rerouted after the existing cold ESP and sent
to a countercurrent heat exchanger where the raw flue gas extracts waste
heat from the clean flue gas exiting the denltrlflcation section. The flue
gas is further heated in an oil-fired furnace to raise the temperature to
350°C (660°F) and then mixed with gaseous NH3 before entering the reduction
reactor. As the gas mixture passes through the reactor, the NOX is cata-
lytically reduced to N2 and the effluent gas is sent to the heat exchanger
to preheat the incoming raw flue gas before the clean gas enters the desul-
furization section. Thus, the retrofitting of this system on existing power
plants will Increase both the capital investment and revenue requirements
since extra equipment and fuel will be required.
Although Mitsui has suggested only one train of denitrlfication equip-
ment needed for a 500-MW boiler, it is assumed from previous studies that
two trains will be used with each train rated at slightly more than 50% of
the boiler capacity. The use of two trains will allow the denltrification
system to be better able to cycle with the boiler availability. The major
consideration during turndown will be to maintain a high temperature (350 C)
in the reactor area to prevent the formation and precipitation of NH4HS04
on the reactor and catalyst surfaces.
282
A
-------
There are no exotic materials of construction required for this process
with all equipment made of carbon steel.
Envirgnmen tal Cons ider at ignis
The Mitsui process has been shown in bench-scale tests to be capable
of selectively removing 90% of the NOX by reaction with NHj. To maintain
this high removal efficiency the temperature must remain above 350 C (660 F)
and the mol ratio of NH3:NOX must remain at approximately 1:1, At lower
temperatures, NH4HS04 will form on the catalyst surface and the NOX removal
efficiency will quickly drop below 90%. When less than the specified amount
of NH3 is injected, the denitrification efficiency will also decline below
90%.
This process is designed specifically for the selective removal of NOX
and does not remove any other pollutants from the flue gas.
The only major source of secondary pollution from this system will be
NH3 emissions in the flue gas. These emissions are expected to be about
60 ppm and they may present a significant problem since, as the flue gas is
cooled, NH^HSO^ could be formed. However, Mitsui is currently conducting
pilot-plant tests to minimize NH3 emissions and associated problems.
Critical Data Gaps and Poorly Understood Phenomena
The critical data gaps for this process involve the operating require-
ments and costs for this system. Mitsui has not as yet published any infor-
mation on the annual operating costs and maintenance requirements.
Advantages and Disadvantages
Advantages and disadvantages for the Mitsui process are listed below:
Advantages
1. Achieves >90% NOX removal efficiency
Disadvantages
1. Forms secondary source of pollution (NH3 level in outlet gas may be
60 ppm)
2. Has not been tested on coal-fired flue gas
3. Requires somewhat clean (particulate level <0.1-0.2 gr/sft^ needed
before reactor) gas feed
283
-------
MITSUI TOATSU CHEMICALS PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (41, 87)
Mitsui Toatsu Chemicals, Inc., (MTC) is developing an SCR NOX removal
process. The NOX is converted to $2 and 1^0 by NHj in the presence of a
catalyst.
Based on a process flow scheme from MTC for dirty gas treatment , an
assumed flowsheet for NOX removal at a coal-fired power plant is given in
Figure 81. After the flue gas passes from the boiler and through the air
heater, NH3 is injected into the flue gas stream. The flue gas is then
heated to the required temperature range of 350°-400°C (662-752°F) by passage
through a heat exchanger and through the use of an auxiliary heater. Then the
flue gas has the particulates removed, for example, by a hot ESP, and then
enters the reactor. In the reactor in the presence of the catalyst, the
following reactions occur, as reported by MTC:
6N°(g) + 4NH3(g) * 5N2(g) + 6H2°(g)
6N02(g) + 8NH3(g) - 7N2(g) + 12H20 (244)
MTC also states that the excess JMj is decomposed to N2 and H20 by the
following reaction:
4M3(g) + 3°2(g) * 2N2(g) + 6H2°(g)
The treated flue gas is sent through the heat exchanger to transfer the heat
to the untreated flue gas flowing to the reactor. Upon exiting the heat
exchanger, the treated flue gas is sent to an FGD unit and/or the stack.
The NH3:NOX ool ratio used ranges from 0.9:1-1.2:1. The reaction
temperature is 350-40Q°C (662-752°F) and the applicable space velocity varies
from 3,000 hr~* to 10,000 hr"~l. The reactor would incorporate either an
intermittent moving-bed or a fixed-bed with a honeycombed-structured catalyst.
NO removal of >90% is reported.
Status of Development
MTC states that five different catalysts, divided Into groups composed
of Cu, Fe, or V, have been developed and are commercially available. MTC-102
catalyst (Cu catalyst group) is suitable for treating clean flue gas con-
taining <1 ppm SOx and 1 mg/Nm3 of dust at temperatures from 200-28Q°C. In
addition to this clean gas treatment, MTC- 104 and MTC-108 catalysts (Fe and
V catalyst groups respectively) may also be used in treating flue gas from
crude oil and naphtha combustion which contains 100 ppm SOX. The temperature
ranges required are 300-4SO°C for MTC-104 and 200-380°C for MTC-108. MTC-110
catalyst (V group) must be used with flue gas containing very much SOX and
284
A
-------
00
1
i
HEAT
EXCHANGE
K
aUXILIARI
HEATER
:b
"HOT"*
ESP
REACTOR
i
AIR NH3
TO
FGD
AND/OR
STACK
FUEL
ASSUMED NECESSARY FOR COAL-FIRED POWER PLANT
Figure 81. Flow Diagram of Mitsui Toatsu Chemicals Process,
-------
dust, such as fuel oil combustion gas. MTC-110 is used at temperatures of
270°C to 400°C and may be used at SOX concentrations of at least 1,000 ppm.
MTC has operated a 1,000 Nm3/hr (0.3 MW equlv) capacity bench-scale unit
at Its Osaka plant on clean gas from the combustion of liquefied propane
gas (LPG). The inlet flue gas contained <1 ppm S02i <1 mg/Nm3 particulates,
and about 200-300 ppm NOX. The MIC-102 catalyst was utilized to achieve
>90% NOX removal. No decrease in catalyst activity was found after 7,000 hr
of continuous operation.
A 1,000 Nm3/hr capacity bench-scale system was operated by MTC at its
Chiba plant and the flue gas source was the combustion of heavy oil. The
flue gas contained 100 ppm S02, 100-300 mg/Nm3, and 30-150 ppm NOX. The
NOX removal achieved was >90% with the MTC-104 catalyst. After 5000 hr of
continuous operation, no reduction in catalyst activity was observed. Parti-
culate removal before the reactor was required.
The MTC-104 catalyst was also used on a 3000 Nm3/hr (1 MW equlv) capacity
pilot plant at MTC's Osaka plant to treat gas being exhausted from an 1€D
unit. The flue gas was composed of 10-20 ppm S02 and 50-80 mg/Nni3 particu-
latee. NOX removal >90% was obtained.
In addition to the above tests, as of August 1976, MTC has been operating
4000 (1.3 MW equlv) and 8200 Nm3/hr (2.7 MW equlv) capacity NOX removal
facilities. These have been operating using a clean gas source since
September 1974 and June 5, 1975, respectively.
Background of Developer.
MTC manufactures fertilizers, industrial chemicals, synthetic resins
and plastics, pesticides, Pharmaceuticals, and dyestuffs. MTC has liaison
offices in New York which should promote the accessibility of this process to
the U.S. market.
NOX removal plants constructed and operating with clean flue gas as late
as 1976 include 84,000 (28 MW equlv) and 90,000 Nm3/hr (30 MW equiv) capacity
plants. Other clean gas treatment facilities in the planning stages (as of
1976) include 220,000 (73 MW equiv) and 300,000 Nm3/hr (100 MW equiv) size
units. Facilities capable of treating 230,000 (77 MW equiv) and 250,000
Nm3/hr (83 MW equiv) of dirty gas are also being planned.
Published Economie Data
For a 67-MW equiv size plant removing 90% of NOX with inlet flue gas
from LPG combustion containing 100 ppm NOX, 300 ppm SOX, and 200 mg/Nm3,of
dust, It is estimated (87) that the capital Investment required is $3.3M,
which is about $49/kW. The catalyst included in this cost is not stated but
may be either MTC-102 or MIC-104 since LPG combustion gas is the flue gas
source. It should be noted that the MTC-110 catalyst Is two to three times more
expensive than MTC-104. The revenue requirement is equivalent to 1.6 mills/
kWh (87). These figures are assumed to be baaed on 1976 costs and a Japanese
location.
286
-------
Raw Material, Energy, and Operation Requirements
The raw materials and energy requirements for denitriflcation have not
been obtained. It may be estimated, however, for the 67-MW equiv-NOx
removal unit treating flue gas with 100 ppm NOX, that 0.5 short tons/day of
NH3 are needed assuming an NH3:NQX mol ratio of 1.1:1 is applied. Data
covering operating manpower, maintenance, and technical support have not been
released.
Technical Considerations
The MIC method is similar to other dry NO removal processes in its
technical aspects. Though not stated by MTC, the NH-j flow will probably be
automatically controlled based on the NQx content in flue gas and the reactor
temperature will be maintained automatically by the auxiliary heater.
The reactor is the largest piece of equipment and should be the same or
slightly larger in size as compared to similar dry methods. Since the
reaction temperature is exactly the same as required by most SCR processes
(35Q-400°C), the comparative size of the reactor will depend on the maximum
allowable space velocity, which ranges from 3,000 to 10,000 hr~l.
MTC has operated NOX removal units with fixed-bed reactors containing
pellet-shaped catalyst and intermittent moving-bed reactors with spherical-
shaped catalyst. The fixed-bed reactor is most suitable for handling flue
gas with very low levels of particulate matter, such as LNG- or LPG-fired
flue gas. Intermittent moving-bed reactors have been employed with crude oil,
naphtha, and fuel oil combustion gases. The catalyst bed is moved as the
particulates accumulate on the catalyst to have the dust accumulation blown
away. MTC is currently studying a honeycomb-shaped catalyst structure for
use in a fixed-bed reactor and is searching for the optimum method for
treating coal-fired flue gas.
This process as described in the flow diagram can be retrofitted into an
existing power plant. For the process shown, large amounts of reheat are
required and thus heating and heat exchange equipment are needed. Naturally,
more reheat would be necessary if this system had to be retrofitted after an
FGD system than if placed only after a cold ESP or ahead of an FGD unit. As
true for other dry processes, this MTC system would prove much more economical
if it were possible to operate before the air heater.
A number of separate NOx reduction trains will be needed for a 500-MW
boiler. The major turndown capability will be the closing off • of any
combination of trains as required by boiler load variations. If an auxiliary
heater is used, this would provide some reactor temperature control capability
during boiler load fluctuations.
No information is presently available on materials of construction and
sensitivity of denitrificatlon to inlet gas composition.
287
-------
Environmental Considerations
The MTC process reduced more than 90% of the NOX in the flue gas during
pilot-plant tests. This efficient NOX removal was achieved with NH3JNQX mol
ratio between 1:1 and 1.2:1, reactor temperature of about 350-400°C, and a
space velocity between 3,000 and 10,000 hr~l. The effects of components such
as S02 and Cl~ at levels expected to be found in coal-fired flue gas have not
been made available. The highest known S02 level tested at present is about
1000 ppm S02. It is reported by MTC that the NH3 in the treated gas is
capable of being controlled at <5 ppm.
This system only removes NOX and does not eliminate any other pollutants
from the flue gas.
Critical Data Gaps and Poorly Understood Phenomena
Information not presently obtained from MTC for this process includes
the following items:
a. Specifics on reactor design and maximum expected pressure drop
b. Maximum tolerable SOX level in the inlet flue gas
c. Sensitivity of NOX removal to inlet gas composition and NH3:NOX
mol ratio
d. law material and catalyst consumption and utility requirements
e. Operating manpower and maintenance requirements
f. Materials of construction
g. Catalyst composition and availability
Advantages and Disadvantages
The advantages and disadvantages of the MTC process are as follows:
Advantages
1. Achieves >90% NOX removal efficiency
2. Claims <10 ppm by vol OT3 in treated flue gas
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Requires clean (degree of particulate removal required is uncertain)
gas feed
3. Uses moving-bed reactor which increases maintenance and catalyst
attrition
4. Requires auxiliary heater to attain or control reaction temperature
288
A
-------
THE RALPH M. PARSONS COMPANY PROCESS - DRY, NONSELECTIVE CATALYTIC
REDUCTION (NOX-SQX)
Process ^8cr±f_ti.on and Principles of Operation (31, 32)
This flue gas denitrlf icatlon system being developed by The Ralph M,
Parsons Company is a dry catalytic reduction process in which both the S02
and the NOX are nonselectively reduced in the presence of a catalyst in a
single step. However, this process is fundamentally different from any of
the other flue gas denitrification schemes discussed in this study. Instead
of consuming expensive NH^ to selectively reduce the NOX as in the other
dry NOX removal processes, the operating conditions in the boiler are
modified such that a neutral or slightly oxidizing atmosphere is maintained
in the boiler. This boiler atmosphere would be obtained by a combination
of holding the excess air in the boiler to a minimum and the injection of a
hydrocarbon fuel, such as natural gas or producer gas, into the radiant zone
of the boiler. This will minimize the amount of free oxygen remaining in
the flue gas and thus reduce the formation of SOX and NOX.
The flue gas from the boiler passes through a hot ESP in which 99% of the
particulates are removed and is then sent to a catalytic reduction reactor con-
taining an inexpensive, iionnoble metal catalyst. A simplified flow diagram
of the process is shown in Figure 82. The S02 is converted to hydrogen sul-
fide (H2S) and the NOX is reduced to N2 by the following hypothesized reactions:
3H2(g) + S°2(g) - H2S(g) +2H2° (246)
C°(B) + N°(g) * C°2(g) + 1/2N2(g)
2CO(g) * N02(g) -, 2C02(g) + l/2N2(g) (248)
In addition to these primary reactions, other secondary reactions are
catalyzed including the conversion of CO to C02 and also the conversion of
both COS and carbon disulfide (CS2) to K>2 and l^S.
The flue gas from the reduction reactor passes through the boiler air
heater and through a Stretford absorber to remove the H2S before the flue
gas is reheated and exits through the stack.
As the flue gas passes through the Stretford tower, the l^S is absorbed
by a Na2CO-j solution and forms sodium hydrosulfide (NaHS) by the following
reactions :
H2S(g) * H2S(aq)
Na2C03(aq) + H2S(aq) * NaHS(aq) + NaHC03(aq) (250)
289
-------
FUEL
GAS
1.* —
i T
1
"HOT" CATALYTIC AIR FroNO-
BOILER — »- ECONOMIZER -T-* ""I ^-REDUCTION •• UFATFP UITFR ~ S
i i tor REACTOR 1 HEATER MlZtn
K> 1 " 1
§ I t
FUEL ASH AIR
GAS
1
i
.^STRETFORD
ABSORBER
i
>
STRETFORD
UNIT
1
r
CLEAN
FLUE
GAS
— »• REHEAT
.^ELEMENTA
•*" SULFUR
Figure 82. Flow Diagram of The Ralph M. Parsons Company Process.
-------
Through a complex sequence of reactions involving soluble V compounds,
this hydrosulfide ion (HS~) is reduced to byproduct S. The overall reaction,
however, is written as simply:
2H2S(aq) + °2(g) - 2H2°(g) + 2S(s) <251>
Status of Development
The Ralph M. Parsons process is in a conceptual design stage and has
not been tested in a bench-scale unit. However, this process utilizes in
part, technology previously developed jointly by the Union Oil Company and The
Ralph M. Parsons Company during commercialization of the Beavon S removal
process.
This process has not been demonstrated on coal-fired flue gas.
An appropriate next stage in the development of this process would be
the demonstration of this system in a bench-scale unit treating actual power
plant stack gas (preferably coal-fired flue gas).
BackgroundofProcess Developer
The Ralph M. Parsons Company is a well-known international engineering-
construction firm based in Pasadena, California. In addition to an overall
involvement in the design of petrochemical and chemical, power generation,
and metallurgical facilities, Ralph M. Parsons is particularly well known
for their design of S recovery plants for treating acid gases.
Their Beavon process was originally developed jointly with Union Oil
Company to upgrade the SC>2 recovery of the Glaus process to meet the more
stringent environment regulations now being applied for SOX. The first
commercial Beavon unit (for S removal only) was completed in 1973 and 21
commercial units are currently operating,
This simultaneous S02~NOX removal process was originally conceived,
in the early 1970's, during the development of the Beavon process. Supple-
mental development work, other than the conceptlonal design data, has not
been completed. This simultaneous S02-NOX process has not been demonstrated
either in a bench-scale or a pilot-plant unit.
Published Economic Data
The Ralph M. Parsons Company has estimated costs (32) for three
different treatment systems depending on the method of generating the
reducing atmosphere. For the purposes of this study, case 1 will be for a
system with the simple injection of hydrocarbon fuel (natural gas) into the
radiant zone of a new 500-MW boiler and case 2 will be for a system using
an auxiliary boiler to generate the required amount of reducing gas for a
new 500-MW boiler. Also, a third case based on retrofitting the system on
an existing 225-MW boiler is Included. In each case the resulting flue gas
la passed through the same catalytic conversion section to convert S02 to
291
-------
and NOX to N£. The costs of each of the three cases given below are
approximate costs, based on a U.S. construction site and mid-1972 dollars,
Case 1 Case 2 Case 3
500-MW 50G-MW 225-MW
new \ new retrofit
Total capital investment
$ 10,380,000 10,843,000 7,972,000
$/kW 20.75 21.69 35.43
Annual revenue requirements
$/yr 2,932,000 3,501,000 2,170,000
Mills/kWh 0.84 1.00 1,38
Raw Material, Energy, and Operation Requirements
The following raw material requirements were estimated by Ralph M.
Parsons for a new 500-»MW coal-fired unit (coal analysis: 3.6% S, 9.0% ash,
12,600 Btu/lb heating value) baaed on case 1 where fuel is injected into the
radiant zone of the boiler.
Raw material Quantity
Natural gas 13,000 sft3/min
Catalyst and chemicals Not available
The utility requirements on the same basis given above are:
Utility Quantity
Electricity 3,940 kW
H20 115 gpm
Steam 2,500 Ib/hr
The only byproduct of this system is elemental S which is of marketable
quality. On the basis given above, about 4.73 long tons of elemental S per
hour will be produced.
The operating manpower needed for this system is about two men per shift,
Technical Considerations
The process equipment needed for the simple fuel Injection case is
minimal except for the Stretford unit. The major process control problem
for this case would be to inject the required amount of fuel into the boiler
since the addition of too much fuel would result in a reducing atmosphere In
the boiler. This reducing atmosphere could create major corrosion problems
in the boiler as well as increased operating costs for the process. On the
other hand, the injection of less than the optimum amount of fuel would
result in higher levels of S and NO In the flue gas and would lower both
the desulfurization and the denitriflcation efficiencies,
292
A
-------
The S(>2 and NOX removal efficiencies are dependent on the inlet flue gas
conditions since the flue gas must not contain any free €2 at the catalytic
reduction reactor and must contain an excess of reducing gases such as H£ or
CO. The presence of Q£ in the flue gas will consume the reducing gas pre-
ferentially with the result that neither the S02 or the NOX will be reduced.
The effects of the particulates in the coal-fired flue gas on the catalyst
and, hence, the removal efficiencies of both S and NOX, are unknown, but can
be expected to complicate the catalytic reduction reactor.
The conversion of SOX to I^S in the reduction reactor allows better
waste heat recovery from the flue gas since the problem of I^SO^ formation
is eliminated, A secondary economizer can be added after the boiler air
heater to cool and recover heat which is normally lost due to the presence
of 503 and H20 in the flue gas. As an additional benefit of this conversion
of S02 to H2S and its subsequent removal, the process developers believe
that the flue gas reheat step may not be required.
With impending shortages and higher prices of natural gas, the injection
of natural gas into the boiler to consume the free C>2 and generate a neutral
atmosphere is probably not a viable alternative. An alternative scheme
suggested by Ralph M. Parsons would be to use vaporized fuel oil or an
auxiliary boiler operating at substoichiometric air to generate a reducing
gas which would then be injected before the reduction reactor. Either of
these alternatives would have to be used in conjunction with both boiler
modifications to sharply reduce excess air to the primary boiler and with
the Injection of aome of the reducing gas into the primary boiler to remove
any remaining free 02. (The economics for this system are given as case 2
in the section, Published Economic Data.) A block flow diagram for this
system is given in Figure 83.
Although the sizes of the various pieces of process equipment have not
been made available, there is no exceedingly large or exotic equipment
needed for this process. The Stretford absorber and regeneration equipment
are expected to be the only major pieces of process equipment. However,
these Stretford units are used in commercial applications.
This process is expected to require two separate trains connected by
a common plenum from the boiler, each rated at 56% of the total boiler
capacity. By shutting down either of the trains, the system can provide a
reasonable turndown capability.
For retrofit applications, The Ralph M. Parsons Company recommends an
auxiliary boiler to generate the reducing gas to be injected before the
catalytic reduction reactor. Other modifications and technical considerations
have not been made available.
The process developers do not expect any change in the materials of
construction in the boiler. The catalytic reduction section and the Stretford
section will be constructed of carbon steel.
293
-------
1
•f
BOILER
1 »HnT" CATAL^
•b. PPHNnfJlTPD _^^h. ^__^^f7cniir>
| J ESP REAC
_ ASH
COAL SECONDARY
i '
t
rTIC AIR SECONDAR
TION »• HEATER ECONO-
'OR MIZER
f
AIR
1
r
1 ( "i
STRE1
1
r
rFORD
R8ER
r
STRETFOR3 ^E
UNIT
i
i
REHEAT —
LEMENTAL
SULFUR
Figure 83. Flow Diagram of Alternative Ralph M. Parsons Process.
-------
Environmental Considerations
Since the NCL. Is reduced to No and the SCL,. is removed as elemental S,
A fc A *
there are no apparent waste disposal problems associated with this process.
The generation of a reducing atmosphere in this system will result in
the formation of several toxic gases including CO, H^S, and COS which should
be considered as a work hazard. However, with a properly designed system
and the extensive experience of fhe Ralph M. Parsons Company in treating
such gas streams, the risks associated with these gases can be minimized.
CriticalData Gaps and Poorly Understood Phenomena
Although this process as applied to boilers is only in the conceptual
design stage and only a minimum amount of information is available, the
following critical data gaps have become apparent:
1. Operating conditions in the various parts of the system
2. Design of the catalyst reduction reactor and catalyst particles
3. Effect of particulatee and SOj on the catalyst
4. The effect of introducing a reducing atmosphere in the high
temperature section of the boiler.
5, Expected S02 and NOX removal efficiencies
Unfortunately at this stage of development many of these questions
cannot be answered. Further studies will be required to investigate many
of these critical data gaps.
Advantages and Disadvantages
Advantages and disadvantages of the Ralph M. Parsons process are listed
below. Additional advantages and disadvantages may become apparent as
development work progresses beyond the conceptual design stage,
Advantages
1. Removes NOX and S02 simultaneously
2. Produces marketable byproduct (S)
Disadvantageg
1. Has not been tested on coal-fired flue gas
2. Has not been developed beyond the conceptual design stage
3. Requires somewhat clean (degree of particulate removal required
is uncertain) gas feed
4. Incorporates design features which may present significant
process control problems
5. Requires flue gas reheat for plume buoyancy
295
-------
SUMITOMO CHEMICAL PROCESS - DRY, SCR (NOX)
Process Description and Principlea of Operation (4, 5, 90, 91, 101)
Sumitomo Chemical Company, Ltd., has developed a dry, SCR method for
removing NOX from flue gas. Treatment for a dirty gas would Include the
steps shown by the block flow diagram In Figure 84, The flue gas from the
holler passes through an air heater, an ESP for partlculate removal, and then
a heat exchanger and auxiliary heater to be heated to reaction temperature of
about 350°C (662°F). NH3 Is Introduced Into the flue gas after the auxiliary
heater and before the gas enters the reactor. In the reactor, the NOX is
reduced to N£ by NH3 In the presence of an S0x-resistant catalyst according to
Sumitomo by the following reactions ;
6N°(g) + 4NH3(g) + 5N2(g) + 6H2°(g) (252)
6N02(g) + 8NH3(g) * 7N2(g) + 12H20(g) (253)
The treated gas exits the reactor, passes through the heat exchanger to
heat the untreated flue gas going to the auxiliary heater, and then proceeds
to a desulfurization unit and/or the stack.
The average operating conditions and removal efficiency for dirty gas
are as follows:
Reaction temperature 350°C (662°F)
NH3!NOX mol ratio 1:1
Space velocity 10,000 hr"1
NOX removal efficiency 90%
S tatua of Development:
Sumitomo Chemical has conducted many investigations on NOx removal with
pilot plants. It is reported that testing more than a thousand catalysts in
clean and dirty flue gas led to the development of a base metal oxide carried
on A1203 as the most suitable catalyst. A prototype unit of 30,000 tfa3/hr
(10 MW equiv) gas capacity for treatment of an oil-fired flue gas has been
operated since July 1973. Large commercial units for both gas and oil-fired
flue gas are also in operation (see Background of Process Developer) .
Figure 85 indicates relationships of temperature and NH3:NOX mol ratio to NOX
removal efficiency for catalyst C-l on clean gas. (Results are discussed in
greater detail under Environmental Considerations.) A catalyst (A) for
decomposing excess NH3 was also developed and its effect with C-l catalyst at
350°C (662°F) is shown in Figure 86. As can be noted, there is <10 ppm NH3
in the outlet gas even with an NH^: NOX ratio of 1.5:1. Sumitomo stated that
the NOX removal efficiency increased about 5% with the use of catalyst A.
Figure 87 indicates that catalyst D which was produced to resist SOx is not
significantly affected by 503 while catalyst C-l is adversely affected. The
relationship of concentration of each of MS and S03 to the NH4HS04 formation
296
A
-------
NJ
BOILER
FLUE
GAS *
Al
HEA
I
R
TER
"COLD"
ESP
AIR
TO FGD
AND /OR STACK
FUEL
NHs
Figure 84. Flow Diagram of Sumitomo Chemical Process.
-------
ho
VO
00
100
80
o
z
UJ
2 60
40
UJ
§
O
2
UJ
o:
f350°C
0.5 IX) 1.5
MOL NH3PER MOL NOX INLET
120
a.
1000-
80 S
o
60 t
40
20
Figure 85. Relationship Among Inlet NH3:NOX Mol Ratio, NOX Removal Efficiency,
and Exiting NH3 Concentration Using the Sumitomo Chemical C-l Catalyst (91).
-------
NJ
VO
VO
100
8? 80
o
z
UJ
o 60
u.
u.
UJ
5 40
O
2
UJ
a:
O 20
350°C
C-l
CATALYS1
I
C-UA
CATALYST
0.5 1.0 1.5
MOL NH3 PER MOLNOx INLET
120
,00°-
80
o
60 t
X
40 U
20 •?
2.0
Figure 86. Relationship Among Inlet NH.,:NOX Mol Ratio, NO Removal Efficiency,
and Exiting NH3 Concentration Using an NH3 Decomposition Catalyst with the C-l Catalyst (101),
-------
100
UJ
O
O
58
5 80
UJ
o
u.
u.
uj 60
I
UJ 40
o:
X
O
z
20
(D)
CATALYST
AFTER S03
TREATMENT
C-l
CATALYST
* IO.OOOhr'1
NH3/NO=I
I
I
200
250 300 350
REACTION TEMPERATURE,°C
400
Figure 87. Comparison of NOX Removal Efficiency with
Sumitomo Chemical's S0x-Resistant D Catalyst and C-l Catalyst (91)
-------
temperature Is given In Figure 88. To avoid deposits of NH^HSO^ either the
concentrations of NHg and S03 must be maintained at lower than indicated
values or the temperature must be held above the values designated In
Figure 88.
Background of Process Developer
Sumitomo Chemical has constructed several commercial denitrification
facilities for oil- and gas-fired FGT. These are included in the following
table:
Capacity
Plant owner (Nm^/hr) Flue gas source Completion
Higashi Nihon Methanol 200,000
Nihon Ammonia 250,000
Sumitomo Chemical 200,000
Sumitomo Chemical 100,000
Sumitomo Chemical 200,000
Sumitomo Chemical 250,000
Sumitomo Chemical 300,000
Gas heating furnace
Gas heating furnace
Gas heating furnace
Gas-fired boiler
Gas-fired boiler
Oil-fired boiler
Oil-fired boiler
May 1974
March 1975
March 1975
February 1975
February 1975
March 1976
October 1976
Sumitomo has offices in New York, referred to as Sumitomo Chemical
America, Inc., and this should provide the accessibility of their NOX removal
process to the U.S. market.
Published Economic Data
„ The spring 1973 plant cost for the NOX removal facility treating 200,000
Nm /hr at the Higashi Nihon Cl^OH plant with flue gas containing 200 ppm NOX
and an NH3:KOX mol ratio of 1:1 was $830,000 (101). This cost is about
$12.6/kW. The corresponding revenue requirement is 1.2 mills/kWh (101). This
plant, however, operated with clean gas according to the flowsheet in Figure
89 and does not include an ESP, heat exchanger, and auxiliary heater. The
1976 cost for a plant to treat dirty gas has been reported to be about $80/kW
(5) which includes an ISP. The site location basis was assumed to be Japan.
Raw Material, Energy, and Operation Requirements
For the 200,000 Nm3/hr treatment facility with flue gas containing 200
ppm NOX and employing an NH3:NOX nol ratio of 1:1, the raw material and
energy requirements are as follows:
Material
J^uantity
NH3 0.73 short tons/day
Electricity 800 kW
The catalyst consumption is not reported In literature.
301
-------
1000
IOO
UJ
o
tsj
Q.
Q.
*»
10
I
z
10
10 100
S03,PPM
1000
Figure 88. Temperatures Below Which
Forms (101)
-------
REFORMER
FLUE
GAS
REACTOR
HEAT
EXCHANGER
NH3
STACK
AIR
Figure 89. Simplified Flow Diagram of the Flue Gas
Treatment Facility at Higashi Nihon Methanol Plant (101)
-------
The electricity needed represents 1.2% of the equiv total power output of
the plant. Fuel is not needed at the above plant and fuel requirements for
the case described in the flow diagram for dirty gas treatment (Figure 84)
are not available. There is no additional manpower required for the 200,000
Nm-Vhr capacity plant above.
Technical Cons IderatiQna
Plant operations using the Sumitomo Chemical dry, denitrification system
with oil- and gas-fired flue gas have reportedly functioned smoothly.
The D catalyst is claimed to be resistant to poisoning or deterioration
from 863. The NOX removal efficiency with D catalyst is slightly greater
with an inlet gas NOX concentration of 600 ppm than for 100 ppm.
With the reaction temperature of 350°C (662°F) and space velocity of
10,000 hr""1, the reactor vessel should be about the same size as most dry
denitrification processes where space velocities average 5,000-10,000 hr
and the reaction temperature is similar.
In retrofit application on a plant with only a cold ESP or FGD, the
required equipment would be just as shown in Figure 84. If an FGD unit is
present, more reheat will be needed if the gas is treated after desulfurizatlon
than If it could be treated before desulfurizatlon.
Since several NOx removal trains will be required to treat the flue gas
from a 500-MW plant, turndown capability would be maintained by removal of
trains from service as required. The auxiliary heater will aid in controlling
reaction temperature during boiler fluctuations.
The materials of construction are unknown,
Environmental pong ide rations
The increase in NOX removal efficiency with Increase in NHjrNOx mol ratio
can be seen in Figure 85 for C-l catalyst with clean gas at a space velocity
of 20,000 hr"1 at 300 and 350°C. Also from Figure 85, with the same NH3:NOX
ratio, increasing temperatures increase the NOX removal efficiency and decrease
the HH3 concentration in the outlet gas. For example, with NH3tNOx = 1:1,
85% NOX removal is attained at 300°C while 90% removal is obtained at 350°C.
At the same conditions, NH3 in the outlet gas decreases from about 40 ppa
to about 10 ppm as the temperature rises from 300 to 350°C. The reaction
temperature versus NO removal efficiency for D catalyst is displayed in
Figure 87. The NOX removal efficiency versus hours of operation for D
catalyst is given in Figure 90 at a space velocity of 10,000 hr"1, 350°C
temperature, and NH^lNO-j,. » 1:1.
The Sumitomo Chemical denitrification process removes no other pollutants
and creates no byproducts. There is no waste to dispose of and, though there
is potential for NH4HS04 formation, this occurrence can be avoided as long as
operating conditions, given by Figure 88, are satisfied,
304
-------
o
Ul
IUU
5?
>-"
O
z
g 90
E
UJ
rf
§ 80
UJ
a:
X
i 70
i
^^
\x^
— '
^ * -
I 1 1 I 1
D CATALYST WITH DIRTY GAS
NH3
~ SVs
350
1
:NOx«l:l
lO.OOOhr-'
°C
I 1 1 1 1
1,000 2,000
ON-STREAM TIME.HR
3,000
Figure 90. Result of Life Test for Sumitomo Chemical Catalyst (91)
-------
There are no apparent work hazards with this process.
Critical Data Gaps and Poorly Understood Phenomena
The cost, consumption, and life of the base—metal catalyst used by
Sumitomo Chemical are not known. The operating cost for denitrification of a
dirty gas is unavailable (the coat Included earlier was for treatment of a
clean gas). Also, fuel and electric requirements are not mentioned for treat-
ment of a gas containing SOX.
Although D catalyst is stated to be SOx resistant, the exact maximum,
tolerable level of SOX is not given. The average pressure drop across the
reactor system and the maximum allowable particulate level are not released
either.
Sumitomo Chemical reported that, with the use of the NHj decomposition
catalyst, the NOX removal efficiency Increased about 5%. It is uncertain
how this KOX reduction increase occurs.
The effects of gas species of small amounts, such as Cl~, is not known.
Also, it Is not known whether any unusual materials of construction are used.
Advantages and Disadvantages
The advantages and disadvantages of the Sumitomo Chemical process are
as fo Hows t
Advantages
1. Achieves >90% NOX removal efficiency
2. Has been applied to flue gas from commercial oil-fired boilers
3. Claims <10 ppm by vol NH^ in treated flue gas
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Requires somewhat clean (uncertain degree of particulate removal
required) gas feed
3. Requires auxiliary heater to attain or control reaction temperature
306
A
-------
SUMITOMO HEAVY INDUSTRIES PROCESS - DRY, SCR (NOX-SOX)
Process Description and Principles of ^ Operation (23, 102, 103, 104, 111)
Sumitomo Heavy Industries, Ltd., (SHI) uses an activated C catalyst in
their process for simultaneous NOX-SOX removal. The NOX is converted to N£
by catalytic reduction with NH3 and SOX is adsorbed on the catalyst in the form
of H2S04 and NH4HS04. This method is very similar to the Takeda Chemical Indus-
tries, Ltd. ,' process. The flow diagram for this process is given in Figure 91.
Flue gas, after leaving the air heater and having particulates removed,
is heated to reaction temperature by an inline heater. NHj is then injected
into the flue gas as it enters a moving-bed reactor operating at 200-230°C
(392-446°F) . In the reactor, the denltriflcation reactions are stated by SHI
to occur as follows;
6NO, , + 4NH,, N -»• 5N0, N + 6H«0 , , (254)
(g) 3(g) 2(g) 2 (g)
6N02(g) + 8NH3(g) ^ 7N2(g) + 12H20(g) (255)
6N°(g) + 302(g) + 8NH3(g) * 7N2(g) + 12H2°(g)
4N°(g) + °2(g) + 4NH3(g) *• 4N2(g) + 6H2°(g)
The NOX is reduced by the NH3 to harmless N£ and H£0 which pass through
the reactor to the stack with the main flue gas.
The removal of SOX from the flue gas proceeds by the following
equations:
2S°2(g) + °2(g) + 2H2°(g) * 2H2S°4(1) (258)
NH3(g) * HH4HB0 (259)
S°3(g) + ^(g) + H2°(g) * ^ (26°>
The H2S04 and NH4HS04 are adsorbed on the catalyst. Other minor reactions
occurring in the reactor include:
NH_ , . + HC1, % •»• NH.C1, . (261)
3(g) (g) 4 (g)
(262)
(263>
NH.Cl, %
4 (g)
c/ \ '
(g)
-*• NI
f °2(g) -
1* liOxJ / s •* \
fi f i f | |
" C°2(g)
+ HC1. v
(g)
307
-------
FUEL
U)
o
oo
STACK
LPG
FURNACE
REGENERATOR
t:
CATALYST
3
MFG
PLANT
Figure 91. Flow Diagram of Sumitomo Heavy Industries Process,
-------
A small amount of HC1 and NH,C1 formed In reactions (261) and (262) exits
in the treated flue gas. Reaction (261) represents a slight NH3 consumption
which must be compensated by an. additional amount of HHa being injected into
the flue gas. Reaction (263) represents a catalyst loss,
The reactor temperature ranges from 200-230°C (392-446°F), Other normal
operating conditions include a space velocity of 1000 hr"-*- and an NH3:NOX mol
ratio range of 2:1-2.5:1. The maximum particulate loading which has been
tested is 0.2 g/Nn3 (0,086 gr/sft3).
The treated flue gas passes through the reactor and is vented out the
stack.
The catalyst is continuously moving downward in the reactor with the flue
gas fed in a crossflow pattern to the catalyst bed. The catalyst exiting
the bottom of the reactor is screened in the presence of an inert gas, N2,
to remove deposits before entering the regenerator. A temperature of about
350°C (662°F) is maintained by the passage of hot gas from a furnace burning
LPG through a heat exchanger within the regenerator. The main reactions
taking place in the regenerator are as follows:
2H2S°4(1) + C(s) * 2S°2(g) + C°2(g) + 2H2°(g) (264>
C(a) - C02(g) + 2S02(g) + 2^0 (g) + 2NH3(g) (265)
3NH4HS04(g) * HH3(g) H- 3S02(g) + N2(g) + 6^0 (266)
The Inert gas exiting the regenerator contains 10-15% S02 and is suitable
for feed gas for manufacturing H/^SO^ The hot gas used to maintain the re-
generator temperature is recycled back to the LPG-burning furnace. Catalyst
is lost as a result of reactions (264) and (265) occurring and this loss
increases in direct proportion to the SOX content of the incoming flue gas.
Therefore, this process is not recommended for coal-fired FGT where S02 levels
exceed 600 ppm. (See Raw Material, Energy, and Operation Requirements.)
Additional reactions proceeding in the regenerator include;
2/3NH3(g) + C-0* * i/3N + H0 + C (267)
2/3NH.C1, , + C-0* -*• 1/3N,,, v + H>0, . + C, , + 2/3C1, , (268)
4 (g) 2(g) 2 (g) (s) (g)
(*C-0 represents 07 chemically bonded with or adsorbed on the
catalyst) .
This system may remove 85-90% NOX and 95% SOX from the flue gas.
309
-------
Status of Development
SHI began developing a process and testing of catalysts using laboratory
facilities in 1973. In 1974, bench-scale tests with a capacity of 150 Nm3/hr
(0.05 MW equiv) started and continued through mid-1975. This operation was
performed on flue gas from an oil-fired boiler and with a fixed-bed reactor.
SHI converted to a moving-bed reactor by drawing from 3 yr of experience at
the 175,000 Nm3/hr capacity oil-fired FGD plant of Kansal Electric Power
Company. SHI began tests in April 1976 with a 2000 Nm3/hr (0.7 MW equiv)
pilot plant treating flue gas from a steel mill sintering plant. Upon
completion of these tests in August 1976, the pilot plant was enlarged to
10,000 Nm3/hr (3.3 MW equiv) capacity and operated from October 1976 through
March 1977. Typical conditions for the operation of these plants are as
follows:
Inlet flue gaa composition
SOX = 150-200 ppm
S02 = 350-480 ppm
02 . 12-15%
H20 - 10-12% .
Particulates = 120-150 rag/Nm
Temperature = 200-230°C
Space velocity * 1000-1700 hr
NH3 Injected = 350-500 ppm
NOX removal - 85-90%
S02 removal - 95%
Background of Process Developer
Since 1965, SHI Investigated S0£ removal from flue gas from oil-fired
boilers. The FGD process of SHI has been in commercial use since 1972. The
SOx-NOjj simultaneous removal process is still under development for applica-
tions with flue gas containing <600 ppm S02 (see Raw Material, Energy, and
Operation Requirements). An N0x-only process Is being developed for coal-
fired FGT (see next process description). SHI has offices in New York which
would make the process more accessible in the U.S.
Published Economic Data
Assuming 300 yen equiv to $1 and 3000 Nm3/hr = 1 MW, SHI claims (23) a
1977 capital cost of $23M for a 500-MW equiv plant with flue gas containing
600 ppm NOX and 500 ppm S02, SOX removal of 95%, and IOX removal of 90%.
Included as part of the capital cost is $6.4M for the catalyst which cost
about $4170/metrie ton. This capital cost equals $46/kW. The operating
cost is equal to 5.9 mills/kWh (based on 233-MW-equlv plant with flue gas
containing 200 ppm NOX and 500 ppm S02 and 1976 costs) (102). These figures
are assumed to be based on Japanese location and to include both the removal
system and H2SO^ manufacturing plant though not stated specifically by SHI.
(1973 figures on FGD unit only Indicated the H2S04 manufacturing unit to be
10% of the Investment,)
310
A
-------
Raw Material, Energy, and Operation Requirements
The estimated NH3 usage rate is 55 short tons/day for a 50Q-MW-equiv
plant with 600 ppm NOX and 500 ppm SOX in the inlet gas, 95% SOx removal, and
90% NOX removal.
There is a substantial C loss, as noted in reactions (263), (264), and
(265). For the same 500-MW equiv-basis, as above, an initial catalyst charge
of 1716 short tons is needed. The catalyst makeup rate required is 3300
short tons/yr. The rate of C consumption increases in direct proportion to
the increase in SOX content of the flue gas, so that, C loss would be
correspondingly large for a 500-MW, coal-fired plant with flue gas containing
2400 ppm S02- SHI states that an S02 level of 600 ppm in the flue gas is the
maximum level allowed for economical operation.
The energy consumption values on the same basis as the above NH3 usage
rate are as follows:
Utility Quantity
Electricity 3100 kW
Steam 26 short tons/day
Cooling water 5 m^/hr
LPG 25 short tons/day
The electrical requirement represents 0.95% of the total equivalent
power of the plant.
The operating manpower requirement for a 500-MW equiv^FGT system is
estimated at two man/shift.
TechnicalConsiderations
This simultaneous SOX-NOX removal system is slightly more complex than
most dry NQX removal-only systems since it possesses a moving-catalyst bed
and a regeneration unit. It is similar to the Kurabo process in this respect.
SHl's process also includes an I^SO^ manufacturing unit for the treatment of
the S02~cohtaining regeneration gas.
Little information has been released on the sensitivity of the process to
inlet gas composition. It is known that the activated C loss is increased
with higher S02 concentrations in the flue gas.
The absorber vessel will be considerably larger for this system than most
competitive processes. Though the reaction temperature (200-300°C) is
considerably less than most SCR methods (4QQ°C), the space velocity (1000
hr~l) is significantly lower than the average of other processes (5000-10,000
hr ), thus, requiring a larger reaction vessel.
311
-------
This simultaneous SOX-NOX removal process may be applicable for retrofit
application for a coal-fired plant with an existing FGD facility. Since the
S02 level would be lowered to <6QO ppm in the gas after an FGD unit , the
activated C consumption loss would not be as great as previously mentioned.
Reheat would be required to obtain the required reaction temperature
(200-300°C) .
For a plant with a cold ESP and no existing FGD unit, an inline heater
for a small amount of reheat due to low reaction temperature is needed. This
is in comparison to the need of a hot ESP and possibly a heat exchanger for
other dry processes. The area required may be larger than other processes
because of the larger reaction vessels and I^SO^ manufacturing unit.
Since several FGT trains are required, adequate turndown capacity should
be maintained by simply removing a train from use. Also, the inline heater
should provide some reactor temperature control during boiler load variation.
The material of construction is primarily carbon steel but there are
some stainless steel portions.
The S02~rich gas exiting the regenerator is used to manufacture 98%
concentrated H2S04, which is considered to be of marketable quality.
Environmental Cons iderat ions
With flue gas containing 300 ppm NO and 600 ppm SOX, a space velocity of
1000 hr~l, and a 2:1 mol ratio of NH3:NOX> the NOX removal efficiency increases
from 70% to >95% as the reaction temperature increases from 200 to 250°C
(392-482°F). The SOX removal efficiency decreases from 90% to about 70% as
the temperature rises from 200 to 250°C. The NOX and SOX removal range
from 80-90% each at 220°C (428°F). These data are shown in Figure 92.
Sensitivity of NOX and S02 removal efficiencies to temperature, space velocity,
and NH3 injection is given in Figures 93 and 94. These results were obtained
from the pilot plant operating at 2000 Nni3/gr capacity at typical conditions
previously described in the Status of Development section. Again it is noted
that as the temperature increases, NOX removal increases while S02 removal
is reduced (Figure 93). As the space velocity rises from 1000 to 1700 hr"1 ,
EOX removal decreases as expected but still remains above 80% (Figure 93).
Figure 94 shows that the NOX removal efficiency increases as the inlet NH3
concentration increases. The Nlij exiting the system Is also displayed.
Along with NOX» this process simultaneously removes SOX from the flue
gas. A small amount of particulate matter is also extracted but the maximum
partlculate loading which has been tested is 0.2 g/Nm^ (0.085 gr/sf t^) ;
therefore, particulate removal Is considered necessary for applications of
this process to coal-fired boiler flue gas.
There should be no major waste- disposal problems; however, the dust
screened from the catalyst bed would require handling. There is potential
for secondary pollution with NH,C1 and excess NHj exiting the stack.
312
A
-------
100
85
»
o
UJ
o
fc
UJ
UJ
a:
CONDITIONS:
O NOXCONTENT: 300 PPM
• sox CONTENT: soo PPM
SPACE VELOCITY: 1000 HR-'
NH3/NOXRATIO : 2:1
200
230
250
280
TEMPERATURE , °C
Figure 92. Effect of Temperature on NOX and S02
Removal Efficiencies with SHI Carbon-Based Catalyst (102)
-------
100
*80
o
UJ
g 60
u.
u.
UJ
40
u>
O
ui 20
a:
I
I
180 190 200 210 220 230
REACTION TEMPERATURE,C°
100
*
80
>•
O
UI
5 60
u.
UJ
<
o
40
iu 20
a:
S02
1
500 1000 1500
SPACE VELOCITY , Hr
2000
'1
2500
Figure 93. Effect of Temperature and Space Velocity on
NOX and S02 Removal Efficiencies for SHI Process (103).
-------
100
80
w 60
Q
U.
y_
UJ
UJ
o:
40
20
1
E
Q.
CL
KX> fc
UJ
50 b
300 400 500 600
INJECTION CONCENTRATION OF NH3
x
z
ppm
Figure 94. Effect of NHj Injection Rate on
NOX and S02 Removal Efficiencies for SHI Process (103)
315
-------
There are no apparent work hazards,
Critical Data Gaps and PoorlyUnderstood Phenomena
A critical data gap is the time required for the C-based catalyst to
complete a cycle (reactor -»• regenerator -*• reactor). It will vary with the
S0£ concentration and particulate level of the flue gas.
Advantages and Disadvantages
A listing of the major advantages and disadvantages of the SHI process
is as follows:
Advantages
1. Removes NOx and 862 simultaneously
2. Achieves >95% S0£ removal efficiency
3. Produces marketable byproduct (98^
4. Is a slight modification of a commercially available FGD system
Disadvantages
1. Forms secondary source of pollution {NH3 in outlet gas probably
exceeds 10 ppm)
2. Has not been tested on coal-fired flue gas
3. Requires somewhat clean (uncertain degree of particulate removal
required) gas feed
4. Uses moving-bed reactor which increases maintenance and catalyst
attrition
5. Requires auxiliary heater to attain or control reaction temperatures
6. Has a low space velocity in the reactor (<5000 hr~l)
7. Uses catalyst which deteriorates by reaction with SO-
316
A
-------
SUMITOMO HEAVY INDUSTRIES PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (102, 111)
In addition to a simultaneous SOX-NOX removal process , Sumitomo Heavy
Industries, Ltd., (SHI) is developing an SCR method of NOX removal only.
The NOX is reduced to N£ by reaction with NHg in the presence of a metal-
oxide catalyst. Figure 95 shows a simplified flow diagram of the process.
Having passed through the air heater and through some type of particulate
removal (cold ESP for example) , the flue gas is heated by an inline heater
to the reaction temperature of 270~37Q°C (518-698°F). NH3 is injected into
the flue gas before it enters the reactor. In the reactor NOX is reduced
to harmless N£ by NH3 aided by a metal-oxide catalyst as follows:
6N°(8) + 4NH3(g) - 5N2(g) + 6H2°(g)
6N02(g) + 8NH3(gr 7N2(g) + 12H20(g) (270)
4NO, , + 4NH~, , + 0,, ,-*• 4N,, v + 6H00,, , (271)
(g) 3(g) 2(g) 2(g) 2 (g)
The reactor is a moving-bed type with the flue gas flowing in a cross-
flow pattern to the catalyst flow. The catalyst moves downward through the
reactor and after exiting the reactor, passes over screens for ash removal
before being conveyed back to the top of the reactor. The treated flue gas
exits the reactor and passes to the stack or an FGD unit*
The average NH3:NOX mol ratio is 1:1 and the space velocity is usually
>5000 hr~^ depending upon temperature. With the reaction temperature of
270-330°C (518-626°F) the NOX removal efficiency is <85%. The NOX removal
efficiency is >9Q% with reaction temperature above 350°C (662°F) .
Status of Development
Since late 1975, SHI has been testing an SCR process for the removal
of NOX using a metal-oxide catalyst. A bench unit of 500 Nm^/hr (0.2 MW
equlv) capacity with a moving-bed reactor was initially used for development
and has been enlarged to a capacity of 1500 Nm^/hr (0.5 MW equiv) for fur-
ther development. A steel mill sintering plant is the flue gas source. The
catalyst life has exceeded 5600 operating hr in bench unit tests.
Background of Process Developer
SHI has been involved in research and development of S02 removal from
oil-fired flue gas since 1965. The FGD process has been in commercial use
at a 175,000 Nm3/hr capacity plant since 1972. The NOX removal process
with the metal-oxide catalyst is still in the early stages of development.
317
-------
00
1 t
FLUE AIR
**""-" GAS "~ HEATER
i
AIR
"COLD"
ESP
AUXILIARY
HEATER
I.
FUEL
1
N
i
\
REACTOR
,
,
SCREEN
i
i
TO F6D
STACK
CATALYST
DUST
Figure 95. Flow Diagram of Sumitomo Heavy Industries Process.
-------
The SHI process should be readily available in the U.S. since there are
SHI offices in New York.
Published Economic Data
The estimated capital investment (111) for a 50Q-MJ equiv -plant is $14,8M,
which is about $30/kW. The revenue requirement is 2,3 milla/kWh with inlet
gas containing 600 ppm NOX (111), These are assumed to 'be based on 1976 costs
and a Japanese location,
Raw Material,, Energy ? and Operation Requirements
It is estimated that the required NH3 for a 5QQ-MW plant with flue gas
containing 600 ppm NOX is 20 tons/day.
The fuel and energy requirements (electricity and steam) are not given.
Operating, maintenance, and technical support requirements are not available
either. These however should be very similar to other dry, SCR methods for
removing NOX only,
Technical^ Cpnsiderat ions
This process is simple as is typical of other SCR processes. The major
control factors will be NH3 flow rate and reaction temperature, both of which
should be capable of automatic control.
In Figure 96 the- NOX removal is shown at various concentrations of
in the flue gas. For example, at 300°C (572°F) the % NOX removal was
80% at a Q ppm SQo concentration and 85% at a 600 ppm SC>2 concentration.
Since the space velocity of 5,000 hr~* is about average in comparison
to other processes (5,000-10,000 hr ) and the reaction temperature of 300 C
(572 F) is lower compared to the 400 C (752 F) of most processes, the reaction
vessel size that is required may be slightly smaller than for competitive
processes.
For retrofit application of this SHI process on a plant with an existing
FGD unit, more reheat is required to reach reaction temperature (270-370 C)
if the flue gas has to be treated after the FGD unit than if NOX removal is
applied before the FGD system. If a plant has only a cold ESP, the reheat
is the same as required above for treating gas before the FGD unit. In
either case, less reheat is needed for the SHI process if operated at about
270-330 C than most dry processes with reaction temperatures of 350-400 C.
Since several trains would be needed for treatment of flue gas from a
500-MW capacity plant, the main turndown capability is in the removal of
trains from service as necessary. An Inline heater would provide reactor
temperature control during boiler load variations,
There is no Information on materials of construction. Also, there are
no byproducts created by this process.
319
-------
u>
NJ
o
100
80
UJ
g 60
u.
O 40
UJ
cr
20
0
CONCENTRATION OF S02:
0-0 PPM
A - 300 PPM
X - 600 PPM
I
250
300
TEMPERATURE , °C
350
CONDITIONS:
NO CONTENT; 150 PPM
SPACE VELOCITY: 5,000 HR-I
NH3/NO RATIO: IM
Figure 96. Effect of Temperature and S02 on NO
Removal Efficiency with SHI Metal-Oxide Catalyst (102)
-------
Envirgnmental Cons iderat ions
The available data on sensitivity of NOX removal efficiency to various
parameters is shown in Figure 96. The NOX removal efficiency increased from
70 to >85% when the temperature of reaction was increased from 250 C (482 F)
to 300°C (572°F) with flue gas containing 150 ppm NOX and 300-600 ppm SOX.
With an increase in reaction temperature of 300 C to 350°C (662°F) the effi-
ciency increased only a few percent. No information is available on effects
of removal efficiency by varying NH3:NOX ratio, higher flue gas S02 concen-
trations, space velocity, etc.
The SHI process is claimed to remove some ash by use of a moving catalyst
bed which passes the catalyst from the reactor over a set of screens and then
returns the catalyst to the reactor. A possible disadvantage of the moving
bed and screens may be the early deterioration of the catalyst. No mention
is made by SHI of the physical strength of the metal-oxide catalyst. No
pollutants•other than NOX are removed to any extent from the flue gas by
this process,
There are no major wastes for disposal with this process; the dust
separated during the catalyst screening step will require handling. There
is the potential for excess NH3 to exit at the stack and though the formation
of NH^HSO^ is more likely at lower reaction temperatures, SHI states that
no reaction of NH3 with SOX occurs at 270-330°C (518-626 F). There are no
apparent work hazards.
Critical Data Gaps and Poorly Understood Phenomena
Included In the data gaps are the effects on the NOX removal efficiency
by factors such as NH-jsNO^. mol ratios and S02 concentrations in the inlet
gas >600 ppm. The NOX removal as a function of space velocity and the
effects of. Cl~ in flue gas from a coal-fired boiler are unknown.
The pressure drop across the reactor, the maximum tolerable particulate
level, the expected catalyst life, and catalyst cost are not available. It
is not known whether there is a problem with excess NH^ exiting at the stack
nor is formation of NH4HS04 in the system mentioned.
In considering requirements for this process, the fuel, electricity,
and steam consumption have not been given yet by SHI, The exact NH^ usage
has not been stated by SHI though it can be estimated at 20 short tons/day
for a 500-MW capacity-plant with flue gas containing 600 ppm NOX and based
on an NH3:NOX ratio of 1:1.
There is also a lack of information on maintenance and technical support
requirements, and the required materials of construction.
Ad van t ag e s_ and Pisadvan t ages
The advantages and disadvantages of the SHI dry, SCR method of removing
NOX are as follows.
321
-------
Advantages
1. Achieves >90% NOX removal efficiency
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Requires somewhat clean (degree of particulate removal required
is uncertain) gas feed
3. Uses moving-bed reactor which increases maintenance and catalyst
attrition
4. Requires auxiliary heater to attain or control reaction temperature
322
A
-------
TAKIDA PROCESS - DRY, SCR (NOX-SOX)
Process Description and Principles of Operation (5, 48)
Takeda Chemical Industries, Ltd., is offering a simultaneous NOX-S02
removal process. NOX removal is accomplished by a SCR of NOX with NH-i in
the presence of an activated C catalyst. S(>2 is adsorbed by the catalyst
in the form of H2S04 and
The optimum temperature range for 90% NOX removal and greater than or
equal to 80% S02 removal in the NOX-S02 removal vessel is 210-250°C (410-
482°F). It is therefore assumed that for a coal-fired power plant, the flue
gas would pass from the air preheater through .a cold ESP and then be heated
by heat exchanger and heater to reaction temperature (see Figure 97 or 98).
Between the heater and reactor, NH3 is injected into the flue gas stream.
In the reactor NOX is reduced to N£ according to the following equations ;
(g) - 5N2(g) + 6H2°(g)
6N02(g) + 8NH3(g) -+ 7N2(g) + 12H20 (273)
While the NOX is reduced to N2, the S0? is being adsorbed on the C catalyst
in the form of 1^804 and NH^HSO^. These are formed as follows:
2S°2(g) + °2(g) + 2H2°(g) * 2H2S04(1)
+NH3(g) -NH4HS04(s) (275)
Typical operating conditions include the following:
Reaction temperature 200-250°C (392-482°F)
Space velocity 1,000-3 ,.000 hr"1
NH3:(NOX + S02) mol ratio (0.7-1.2);!
The catalyst life is guaranteed for 1 yr. The pressure drop in the
catalyst layer over a bed height of 50 cm is 30-70 mm H20. The NOX removal
efficiency averages at least 90% and SOX removal efficiency is a minimum of
80%.
The treated flue gas exits the reactor, passes through the heat exchanger
to transfer heat to the untreated flue gas, and then exits the stack.
As the adsorption of H2S04 and NH4HS04 continues, the NOX and S02
removal efficiency is reduced. Therefore, the H2S04 and NH^HSO^ must be
removed periodically to renew the activity of the C catalyst. This removal
may be accomplished by either washing with H20 or heating in the presence of
an inert gas. With the washing method, a 10-20% H2S04 and NH4HS04 solution
is obtained from forcing H20 through the reactor. After CaO or Ca(OH)2 is
added to this solution, the mixture is sent to a steam stripper where gaseous
323
-------
1
BOILER
FLUE
GAS
/
AIR
HEATER
"COLD"*
ESP
AIR
N>
t±r:
NH3
I STEAM]
STRIPPER*
HEAT I
EXCHANGER-
ADJUSTING-*-
CELL
CoS04
ASSUMED NECESSARY FOR COAL-FIRED POWER PLANT OPERATION
Figure 97. Flow Diagram of Takeda Chemical Industries, Ltd., Process (Regeneration by Washing).
-------
STACK
1
BOILER
FLUE
GAS *
AIR
HEATER
"COLD"*
ESP
U)
NO
Ul
AIR
H2S04
0=3
•-FUEL
•AIR
ASSUMED NECESSARY FOR COAL-FIRED POWER PLANT OPERATION
Figure 98. Flow Diagram of Takeda Chemical Industries, Ltd., Process (Regeneration by Heating).
-------
NH3 i-3 taken off the overhead and CaSO^ is recovered as the stripper bottoms
(see Figure 97). The NH3 is recycled to the injection step. With the heating
method of regenerating, which is shown in Figure 98, an inert gas is passed
through the reactor at 350-450°C (662-842°F). The I^SO^ and NHpSCH on the
C catalyst are decomposed Into 862, H20, N2» and C02- The gas, composed of
5-15% S02, is converted to 98% H2S04 in an H2S04 manufacturing plant. This
heating method Is the most preferable since wastewater treatment would be
required with the washing method.
Status of Development
Takeda has been doing research since 1968 on NOX and SOX removal.
Experimental studies were performed on NOX removal versus % C^, %
H20 in flue gas, space velocity, and temperature and also pressure drop
versus linear velocity. (Refer to the section, Technical Considerations.)
Also, a bench-scale plant with 100 Nm3/hr capacity from a thermal power
station has been operated. The flue gas composition included 60-130 ppm NOX,
50-150 ppm SOX, and 6-7% 02- There are two pilot-plant systems in operation
with the heating method of regenerating the catalyst. One has flue gas from
a sintering furnace with a 10,000 Nm3/hr (33 MW equiv) gaa capacity. The
inlet flue gas NOX and SOX levels are 200 and 300 ppm respectively. The
particulate level is 100-200 mg/Nm-* and the average space velocity, 1100 hr~l.
The temperature ranges from 210-230°C (410-446°F). The minimum NOX and SOx
removal levels are 80 and 90% respectively. The other plant with a capacity
of 4500 Nm-Vhr (1.5 MW equiv) treats flue gas from a glass-melting furnace.
The NOX and SOX inlet levels are about 300 ppm each. The dust level is
1-5 mg/Nm^ and the average space velocity, 1300 hr~l. The temperature ranges
from 200-250°C (392-482°F). The NOX and S02 removals are 90% each.
Background of Proce s s Deyeloper
Takeda is the process owner and catalyst supplier for this simultaneous
NOX-SC>2 removal process and retains the licensing rights. There is an office
of Takeda in New York City which should make the process accessible in the
U.S.
PublishedEconomicData
The economic data available are the estimated capital cost and operating
cost of utility and raw material consumption for a plant with 100,000 Nm3/hr
(33.3 MW equiv) gas treatment capacity (48). The estimated capital cost is
about $3M which is equivalent to $90/kW. The operating cost equals $21.80/hr
and is equivalent to 0.65 tnllls/kWh, The basis for the above costs includes
the following:
Inlet NOX concentration 200 ppm
Outlet NOX concentration <20 ppm
Exhaust temperature 180°C (356°F)
Space velocity 2,000 hr-1
SOX concentration 50 ppm
326
A
-------
These costs do not include the system for catalyst regeneration. The
cost of the activated C is approximately $4170/metrlc ton and about 60 tons
of C catalyst is initially required. It is also assumed that 1976 costs and
a site in Japan are basis for the above economics.
Raw Material^, Energy» and Operation Requirements
As in the economic data section, the following raw materials and energy
requirements are for a 100,000 Nm^/hr gas capacity treatment facility. The
basis is the same as mentioned above and only includes the operations
involved in NOX removal and do not include the catalyst regeneration facili-
ties. The requirements are as follows:
Electricity 230 kW
Fuel oil 3.09 short tons/day
NH3 0.42 short ton/day
The electricity required for the NOX removal portion represents 0.7% of the
total power equivalent of the plant.
Technical Considerations
Takeda's system is fairly simple compared to wet NOX removal systems.
The regeneration section would make this slightly more complex than some of
the other dry N0x-only removal processes.
The information on removal sensitivity to inlet gas composition is pro-
vided by experimental data as shown in Figures 99-102. The gas used contained
300 ppm NOX and 50 ppm S02 and 300 ppra NH3. In Figure 99 the NOX removal
increased from 45 to 95% as the 63 increased from 0 to 5%. The NOX removal
decreased from 95 to 90% as the H20 Increased from 5 to 10%. Also, the
frequency of regeneration required is dependent upon the concentration of
SOX in the flue gas. It is stated that with an inlet concentration of 150
ppm SOX, the regeneration is required every few days. For a flue gas from
a coal-fired boiler with 2400 ppm SOX, regeneration would be required approxi-
mately every 15 hr. Though the heating method is the most preferable regene-
ration method. If the washing method were used, H20 equivalent to five times
the catalyst in weight would be required for each period of regeneration.
In Figure 100 the pressure drop data indicate an increase in pressure
loss from 35 to 200 mm 1^0 when linear gas velocity increases from 0.15 to
0.5 m/sec. This may be a greater pressure loss than most of the other dry
systems. Since the space velocity is rather low in comparison to most of
the dry processes, the reactor vessels would be about the same to slightly
larger. Also, more vessels would be required if the washing step is required.
However, with the low reaction temperature, a cold ESP and smaller heaters
and heat exchanger are satisfactory.
This process would be suitable for retrofit applicability on a plant
with or without FGD equipment. If an FGD process is already present, this
system could be placed after the FGD unit and the frequency of regeneration
in the Takeda process would be greatly reduced. More reheat would be required
327
-------
100
TEMP.* 200°C
S:V; » 1100 Hr.'1
40
30
CONCENTRATIONS, % BY VOL
Figure 99. Relationship of 02 and H2<) Concentration In
Flue Gas to NOX Removal Efficiency for Takeda Process (48)
328
-------
200
E
E
*t
I
yj
o:
3
CO
01
UJ
cc
CL
too
1
o
o:
V-
^ m*
S 50
o:
UJ
a.
to
co
O
25
I
J_
J_
10
25 50 100
LINEAR VELOCITY, cm/SEC
Figure 100. Relationship of Linear Velocity and
Pressure Drop for Takeda Process (48).
329
-------
100
90
a?
o" 80
z
UJ
u
u.
u.
Ul
_i 7O
O
UJ
x 60
O
5O
4O
30
0
I
1
1000 2000 3000 4000
SPftCE VELOCITY, HR-1
5000
Figure 101. Relationship of Space Velocity and
NOX Removal Efficiency for Takeda Process (48) .
330
-------
CO
CO
100
s?
u.
UJ
o
UJ
o:
ox
80
60
40
20
0
50
SPACE VELOCITY = 1100 HfH
I
100 150 20O
TEMPERATURE, °C
250
300
350
Figure 102. Relationship of Temperature and NOX Removal Efficiency for Takeda Process (48),
-------
to reach reaction temperature, however. If there was no FGD, the only con-
sideration would be if there is enough space to place the Takeda system on
the present site. It could be Installed past the air heater and very little
reheat would be: required to reach reaction temperature.
The turndown capability should be acceptable since for a 500-MW plant,
several trains will be required for NOX and S02 removal from the flue gas,
and reactors would be taken off stream as needed.
The information obtained on materials of construction has applied only
to the regeneration systems. The heating regeneration process uses carbon
steel while stainless steel is required for the washing method of regeneration.
Environmental Considerations
Data on NOX removal efficiency versus operating conditions were given
for experimental data on gas containing 300 ppm NOX, 50 ppm S02» 300 ppm NH3,
4-5% 02, 9-11% H20, and 10-11% C02. The NOX removal decreased from 98 to 85%
as the space velocity Increased from 1000 to 4000 hr"1 at 250°C (482°F). The
NOX removal efficiency decreased from 95 to 55% as the space velocity
increased from 1000 to 4000 hr""1 at 200°C (392°F). The NOX removal Increased
with an Increase in reaction temperature. For example, NOX removal increased
from 40 to 80% as the temperature Increased from 150-200°C (301-392°F) as a
space velocity of 1100 hr"~l. Removal efficiency went from 80 to 95% as
temperature rose from 200 to 250°C (392-482°F) at 1100 hr"1 space velocity.
Takeda's process also removes S02- The S02 removal, however, decreased
with an increase in temperature. With a 1100 hr"-*- space velocity the SOX
removal was 95% at 200°C (392°F), and 80% at 250°C (482°F) with an inlet gas
SOX concentration of 50-150 ppm.
The effect of NH^:(NOX + SOX) ratio on NOX and SOX removal may be seen
in the following data.
NHji (NOX _+ SO^.) mol ratio
0.8 1.0 1.2
Removal efficiency, %
NOX
SOV 76 84 90
NOX 87 93 96
The conditions under which the above data were taken are as follows:
Temperature 220°C
Space velocity 1,000 hr
Inlet NOX 200 ppm
Inlet SOX 500 ppm
Inlet 02 15% (vol)
H20 10% (vol)
332
A
-------
Waste disposal problems would primarily depend on the method of regenera-
tion used. In the washing method, a solution of CaSC>4 would have to be
handled. In the heating method, any waste streams from an H2S04 plant would
have to be treated.
Critical Data Gaps and Poorly Understood Phenomena
Further details required about the Takeda process include the capital
and operating costs for the regeneration system, and the operating conditions
for the regeneration system. The catalyst consumption rate and NH3 level in
the treated gas are also not reported.
Advantages and Disadvantages
The advantages and disadvantages of the Takeda system are as follows;
Advantages
1. Removes NOX and SC>2 simultaneously
2. Achieves >90% NOX removal efficiency
3. Produces marketable byproduct (981 H2864)
Disadvantages
1. Requires significant amounts of energy for the regeneration step
(may require regeneration every 15 hr for treating coal-fired flue gas)
2. Requires somewhat clean (degree of particulate removal required is
uncertain) gas feed
3. Requires auxiliary heater to attain or control reaction temperature
4. Has a low space velocity In the reactor (<5000 hr~l)
333
-------
USE PROCESS - DRY, SCR (NOX)
Process Description and PrinciplesofOperation (109)
The Ube (Ube Industries, Ltd.) process is a dry N0x-only removal system
based on the SCR of NOX using NH3. By inserting any one of the three different
types of catalyst (B-2, S0x-free; KV, low SOX; or K-2, moderate SOX levels)
developed by Ube, this process can be specifically designed for the particular
flue gas conditions. The K-2 catalyst, which is now considered the best Ube
catalyst for treating coal-fired flue gas, was developed for treating flue
gas containing a maximum of approximately 1000 ppm SOX. Thus, although Ube
apparently does not have a catalyst available for treating coal-fired flue
gas containing 2400 ppm SOX, the Ube process will be described assuming that
the K-2 catalyst is being used.
This dry SCR process is very similar to the other dry processes in that
it consists of three distinct sections: partlculate removal, NHj injection,
and catalytic reduction. The basic outline of the Ube process can be seen
in the block flow diagram given in Figure 103.
The flue gas from the economizer at 350-400°C (660-750°F) passes through
a hot ESP to remove 99% of the incoming particulates before entering the NH3
mixing chamber. Gaseous NHj is injected into this mixing chamber at an
NH3:NOX mol ratio of (1.0-1.1);! and immediately mixed with the flue gas
before entering the catalytic reduction reactor. As this gas mixture passes
through the packed-bed reactor at approximately 350 C (660 F), the NOX is
selectively reduced to N2 by the following reactions.
6NO, . + 4NH-, , + 5N0/ \ + 6H00, , (276)
(g) 3(g) 2(g) 2 (g)
6NO,,, . + 8NH,, ,-»• 7®»f . + 12H00, . (277)
2(g) 3(g) 2(g) 2 (g)
Since NO'represents 90-95% of the NOX in the flue gas, reaction (276)
is the primary reaction occurring in the reduction reactor.
The clean flue gas from the reactor passes through the boiler air
heater and then the FGD section before exiting through the stack.
Status of Development
Very little information has been released concerning the development
of the Ube process. From preliminary Information it appears that this
process, using the K-2 catalyst, has been tested in a pilot plant treating
10,000 Nm3/hr of tail gas from an HN03 plant. During the first 8 mo of
operation treating flue gas containing 3000 ppm NOX and no 862» the Ube
process was able to maintain 95% removal of NOX with no decrease in catalyst
activity.
No information has been published concerning any test results from
treating either coal-fired flue gas or any other type of flue gas containing
334
A
-------
U)
u;
ESP
REACTOR
i
1
AIR
HEATER
ASH
NH
AIR
TO FGD
SYSTEM
Figure 103. Flow Diagram of Ube Industries Process.
-------
more than 700 ppm S02- The size of the test facility treating flue gas con-
taining 700 ppm SC>2 and 300 ppm NOX has not been released.
The next stage of development would be to construct a pilot plant to
demonstrate the ability of this process using the K.-2 catalyst to treat flue
gases containing more than 1000 ppm S02-
BackgroundofProcess Developer
Ube is a relatively large, diversified Japanese chemical company with
interest In chemicals, shipbuilding, construction materials, and transporta-
tion. Ube entered the air pollution control field in the early 1970*s with
the development of particulate control devices, both wet and dry, and a wet
FGD system. In 1973 Ube began testing the applicability of base-metal cat-
alysts for removing NOX from exhaust gases. This catalyst research led to
the Ube Industries SCR process which has undergone pilot-plant testing for
the past year.
At the present time no American company has been licensed by Ube to
market this process in the U.S. However, Ube does have a liaison office in
New York City through which inquiries about this process are handled.
Pub11shed Economic Data
No information has been released concerning the total capital invest-
ment or the annual revenue requirements for this system.
Raw Material, Energy, and Operation Requirements
The major raw materials required for the Ube process are NH3 and the
base-metal catalyst. Utility requirements would include electricity, steam,
and cooling water. The quantities of these raw materials and utilities and
also the other operating requirements such as the operating personnel have
not been published.
Technical Consideratlons
The Ube process is a relatively simple system with only five pieces of
process equipment: a hot ESP, an NH^ storage tank, an NH/j vaporizing tank,
a mixing vessel, and the reduction reactor. As Is common for the dry SCR
processes, the only major process control problem is vaporizing and injecting
the exact amount of NH3 required to provide an NH3:NOX mol ratio of
(1.0-1.1):! in the reduction reactor. The NH3 to be injected into the system
is metered Into the mixing tank In direct proportion to the amount of flue
gas passing through the system.
The NOX removal efficiency is relatively Independent of the inlet flue
gas providing that the incoming particulate levels in the gas are very low.
Although the maximum allowable particulate loadings have not been published,
the use of small diameter pellets and a packed-bed arrangement In the reduc-
tion reactor seriously undermine the usefulness of this process for treating
336
A
-------
coal-fired flue gas. Its ability to treat coal-fired flue gas, even with a
very efficient hot ESP, without plugging seems highly questionable. At the
present time Ube is modifying their process before testing begins on oil-
fired boiler flue gas containing higher dust levels.
The catalytic reduction reactor will be the largest piece of process
equipment in the Ube system and is a packed-bed type reactor. The catalyst
particles are small pellets with typical dimensions of 5 mm in dia and 5 mm
in height. Thus, for 90% NOX removal the operating conditions in the reduc-
tion reactor would be the following;
Parameter Value
NH3:NOX mo1 ratio (1.0-1.1):!
Temperature 400°C (750 F)
Space velocity 5,000 hr"1
Catalyst size-shape 5 mm dia x 5 mm high pellet
SOX level 2,400 ppm
Dust level Small dust content
Under these conditions the lifetime of the catalyst is expected to be >8000-
hr (1 yr).
This process would be difficult to retrofit onto an existing power plant
equipped with cold ESP. The ductwork modifications required to remove and
return the flue gas between the economizer and the air heater would be expen-
sive. Thus, for retrofit applications, the flue gas will be removed after
the existing cold ESP and passed through a countercurrent heat exchanger.
The incoming flue gas is heated by extracting waste heat from the clean flue
gas exiting the denitrlficatlon section. The inlet flue gas is further
heated In an oil-fired furnace to raise the temperature to 400 C (750 F) and
then mixed with gaseous NH3 before the mixture enters the reduction reactor.
From the reactor the flue gas passes to the heat exchanger to preheat the
incoming flue gas before the gas enters the desulfurization section. Thus,
the retrofitting of this system on existing power plants will increase both
the capital and operating cost since extra equipment and fuel will be required,
Although the number of separate denitrification trains needed for a 500-
MW boiler has not been disclosed, it is assumed that several trains will be
required and the flue gas will be sent to either train through a common
plenum. This type of arrangement will allow the denitrification system to
partially cycle with the boiler availability by simply closing off any com-
bination of trains. The major consideration during turndown is to maintain
a temperature of 350 C in the reactor area to prevent the formation and
precipitation of NH^HSO^ on the reactor internals.
Ube has not published any information on the materials of construction
required for this system.
337
-------
Environmental Considerations
The Ube process, during pilot-plant testing on HN03 tail gas (no SOX
or partieulates) was capable of selectively reducing 90-95% of the NOX.
This high NOX removal efficiency is dependent on both the reactor tempera-
ture remaining above 370 C (698 F) and the mol ratio of NH3:NOX remaining
in the range of 1.0:1 to 1.1:1.
This process is designed specifically for the selective removal of NOX
and does not remove any other pollutants from the flue gas.
Critical Data Gaps and Poorly Understood Phenomena
Many of the critical data gaps for this process involve the operating
requirements and costs for this system. Ube has not released information
on the capital or operating coats, catalyst consumption, utility require-
ments, pilot-plant data, materials of construction, or operation require-
ments and only scant data to determine status of development.
Advantages and Disadvantages
Advantages and disadvantages for the Ube process are listed below.
Advantages
1. Achieves >90% NOX removal efficiency
Disadvantages
1. Has not been tested on coal-fired flue gas
2, Requires clean (S02~ and particulate-free) gas feed
338
-------
UNITIKA PROCESS - DRY, SCR (NOX-SQX)
Process Description and Principles of Operation (2, 4, 46, 112)
Unltika has been developing two types of NQX removal systems , a simul-
taneous NQX-SOX and an N0x-only removal unit. The simultaneous NQX-SQ2
removal process is considered almost fully developed by Unitika and is herein
described.
Unitika* s system passes the flue gas, having already passed through a
particulate removing device and heated to the relatively low reaction tem-
perature of 200-250°C (392-482°F) , through a packed tower containing fixed
C beds. The S02 in the flue gas reacts to form several products which are
removed by adsorption in the C catalyst while the NOx is reduced by NH^ to
N£ which is removed with flue gas exiting the reactor. An. assumed flow
scheme for coal-fired FGT, based on the pilot-plant process flow, is depicted
in Figure 104. The NOX is converted to N2 by NH3 as follows:
6H°(g) + 4™3(8) * 5K2(8) + 6H2°(g)
6N02(g) + 8KH3(g) * 7S + 12H0 (279)
There are several S0£ removal reactions and all the products from the
reactors are diffused and adsorbed into the catalyst. The reactions are as
follows :
2S°2(g) + °2(g) + 2H2°(g) * 2H2S°4(1) (280)
2NH3(g) + (NH4)HS04(1) (282)
S°2(g) + 2NH3(g) + H2°(g) * (NH4>2S03(s) (283)
(NH4)2S°3(s) + S°2(g) + H2°(g) •" 2(M4)HS03(s) (284)
At 200°C (392°F) or above, reaction (280) is predominant while at below
180°C (356°F) reactions (281-284) take priority.
It should be noted that NH3 is also involved in the S02 removal as well
as NOX removal. Unitika applies the following equation for required NH3
consumption :
NHL - 0.8NO + (0.3-0.5) SO
«j X
Greater than 90% SOX and NOX removal is claimed.
339
-------
STACK
1
I
BOILER
FLUE ^
GAS
|
AIR
HEATER
*
"COLD"
ESP
.
•I
i
[CHANGER
4N
O
AIR
FUEL
ASSUMED NECESSARY FOR COAL-FIRED
POWER PLANT
REACTOR TOWER
ABSORBED
BED
SO* CONTAINING GAS
TO
H8S04 PLANT
NH
Figure 104. Flow Diagram of Unitika, Ltd. Process.
-------
When the C bed becomes saturated, it is removed from the process and a
regenerated bed Is placed onstream. The C is regenerated in the presence of
a low-02 gas at 300-350°C (572-662°F). The S02 is released for I^SO^
production as the S0^= and 803" are decomposed to S02 and N2 as follows :
, 2S02(g) + 2H20(g) + 02(g) (285)
2H2S°4(1) + C(s) * 2S°2(g) + C°2(g) + 2H2°(g) <286>
02(g) + S02(g) + 4H20(g) + H2(g) (287)
02(g) - 4S02(g) + 10H20(g) + 2N (288)
302(g) -, 2S02(g) + 8H20 + 2N (289)
302(g) - 4S02(g) + 10H20 4- 2N (290)
At a 4500 Nm^/hr (1.5 MW equiv) capacity pilot plant, a packed tower con-
sisting of four compartments with fixed C beds is utilized in treating gas
containing 400 ppm S02 and 500 ppro NOX. Three compartments are fed with flue
gas while the fourth is being regenerated. The total tower height is 12 m
and the C bed height is 0.6 m. The pressure drop is <100 mm t^O. The
space velocity is 1000 hr and NH3 Is injected to give about 600 ppm in
the flue gas. The adsorption time is 24 hr and regeneration time is 8 hr.
The S02 concentration in the gas from regeneration is 5-10%.
The C consumption is given as about 50% of a. charge per year. The C is
said to hardly deteriorate and is resistant to damage by SOX and partlculate
matter. The catalyst life exceeds 1 yr.
Status of Development
Unitika began developing an SOX removal process in 1971. This removal
technique with activated C was evaluated with a 150 Nm^/hr and a 600 Nm-Vhr
capacity models. When NOX removal became necessary, Unitika Incorporated NOX
removal into the SOX removal process. A pilot plant of 4500 Nm3/hr flue gas
capacity with an inlet composition of 500 ppm NOX and 400 ppm S02 is currently
in operation at Union Glass Industry Company, Ltd. , at Hirakata, Japan.
Background of Process Developer
Unitika, Ltd., has a liaison office in New York. This should make
Unitika 'a process accessible to the U.S. market. Based on the successful
results of the pilot plant at Union Glass, a larger NOX-S02 removal plant
will be erected.
341
-------
Published Economic Data
The capital investment for a simultaneous NOX-S02 removal system and
resulting acid recovery section treating 500,000 NnrVhr (167 MW equiv) of flue
gas is reported to be $10M or about $60/kW (46). The revenue requirement,
with an inlet gas containing 1000 ppm SOX and 300 ppm NOX, is equivalent to
$27.3/kl of heavy oil burned in the boiler. Using 12,000 Nm3 = 1 kl of oil,
the revenue requirement equals $1138/hr which is 6,8 mills/kWh (46). The
cost for the catalyst is about $3800/short ton. All the above values are
assumed to be based on a Japanese location and 1976 costs.
Raw Material^ Energy, and Operation Requirements
The estimated requirements for an NOX-S02 removal system treating a
500,000 Nm3/hr stream of flue gas containing 1000 ppm SOX and 300 ppm NOX
are as follows;
Material Quantity _
NH3 13.3 short tons/day
Fuel, LPG 55.5 short tons/day
Electricity 2,708 kW
The electrical demand Is equal to 1.62% of the equivalent total output of the
plant,
No information on requirements for the I^SQ^ production are available,
along with operating, maintenance, and technical support needs. However, the
H2SO^ manufacturing needs and costs should be the same as for the production
of industrial-grade l^SQ^ made by the contact process method.
Technical ConsideratlonB
Compared to wet systems, this system is simple with a packed tower
reactor and regeneration gas producer. It is slightly more complex than some
other dry processes since a system for handling S02~rich gas from regeneration
for H2S04 production is included.
The NOX and S02 removal, and consequently, absorption and regeneration
cycles will be affected by inlet gas composition. On a pilot plant with
400 ppm S02 and 500 ppm NOX in the flue gas, absorption time is 24 hr with
an 8-hr regeneration time for 90% NQX~SOX removal. Data with inlet NOX of
200 ppm and SOX of 1000 ppm, showed UO^ removal still about 90% after 4 days,
but SOX removal had fallen below 90% after only 22 hr. With coal-fired flue
gas which would contain about 2400 ppm 80%, the regeneration frequency should
be significantly greater than that Just mentioned due to the richer SOX
content in the flue gas.
Although the reaction temperature is lower than most dry NOX removal
systems, the space velocity is only 1000 hr~l and the reactors would probably
342
-------
be considerably larger than for other processes. Also, more reactor area is
required since part will be undergoing regeneration.
Unitika's process should be suitable for retrofit in the case of a plant
without FGD. A small amount of heat is required but not as much as other
processes whose reaction temperature is 300-400°C (572-752°F), The minimum
land required for a 500,000 Nm^/hr (167 MW equiv) gas treatment facility is
4000 m2.
This simultaneous NQX~S02 process should provide adequate turndown
capacity. This could be handled by simply removing absorption beds from use.
Carbon steel is probably the major material of construction.
Environmental Considerations
S02 removal is much more sensitive to space velocity than NOX removal.
At 240°C (464°F) and a space velocity of 1000 hr""1, the S02 removal was 93%
and it decreased to 86% with an increase in space velocity to 1400 hr~^ at a
constant temperature. The NOX removal varied from 95 to 80% as the space
velocity increased from 2000 hr~l to 2700 hr~^. An increase in temperature
decreased the S02 removal. An S02 removal of 98% at 200°C (392°F) with space
velocity of 1000 hr"-*- was decreased to 93% as the temperature increased to
240°C (464°F) at the same space velocity. The relationship of S02 removal
and NOX removal to temperature and space velocity are given In Figures 105
and 106.
S02 is removed along with NOX in Unitlka's system. The S02 is released
after regeneration and the S02~rich gas is used in producing a high-purity
H2S04.
There should be no waste^disposal problems except for possible exhaust
gases during the making of byproduct from the S02~rich regeneration gas.
No data are given on effects of Cl~ upon this process. There are no
apparent work hazards.
Critical Data Gaps and Poorly Understood Phenomena
The Information lacking on this process includes the operating manpower,
maintenance, and technical support requirements. The expected NH^ emissions
in the treated gas and the maximum particulate level allowed in the inlet flue
gas have not been reported.
Advantages and Disadvantages
The advantages and disadvantages of the Unitika simultaneous NOX-S02
removal system are as follows:
343
-------
100
90
UJ
o
C
fc 80
§
i
CM
O
CO
70
60
CONDITIONS
INLET
NOX 180-240 ppm
SOX 200-280ppm
H2° 9.0%
DUST 5-9mg/Nm3
02 13.0%
OUTLET
10-15 ppm
!8-25ppm
9.0 %
5-9mg/Nm3
13.0%
1
I
I
200 210 220 230
REACTION TEMPERATURE,°C
240
Figure 105. Relationship of S02 Removal Efficiency to
Reaction Temperature and Space Velocity for Unitika Process (46)
-------
U)
-p-
100
90
UJ
a so
u.
UJ
O
X
O
70
60
220
CONDITIONS
INLET
NOX I80-240ppm
SOx 200-280 ppm
H20 9.0%
PARTICULATES 5-9mg/Nm3
Oa 13.0%
I I I
OUTLET
!5-20ppm
200-280ppm
9.0%
5-9mg/Nm3
13.0%
230
240
250
260
REACTION TEMPERATURE, °C
Figure 106. Relationship of Reaction Temperature and Space Velocity
on NOV Removal Efficiency for Unitika Process (46).
-------
Advan tages
1, Removes NOX and S02 simultaneously
2. Achieves >90% NOX removal efficiency
3. Produces marketable byproduct (H^SO,)
Pis advan tages
1. Requires significant amounts of energy for the regeneration step
2. Has not been tested on coal-fired flue gas
3. Requires somewhat clean (degree of particulate removal required
is uncertain) gas feed
4. Requires auxiliary heater to attain or control reaction temperatures
5. Has a low space velocity in the reactor (<5QOO hr~ )
346
A
-------
UNITIKA PROCESS - DRY, SCR (NOX)
Process Description, and PrinciplesofOperation (46)
In addition to the simultaneous NOX~SC>2 removal system, Unitika, Ltd.,
has developed an NOX—only removal process. Though details about the process
are unavailable, it is known to be an SCR process. More information should
be reported in the near future. The Nl^iNOjj tnol ratio is (1,0-1.1):! and
reaction temperatures range from 320-410°C (590-770°F). Information concerning
the catalyst composition is not available. NOX removal efficiencies of >90%
have been achieved on bench-scale tests.
Statusof Development
Unitika has applied this NCy-only method on 150 Nm3/hr (0.05 MW equlv)
capacity tests for 45 days and 200 Nm3/hr (0.07 MW equiv) flue gas capacity
tests for 300 days. This technology is to be tested during 1977 with equip-
ment capable of treating 70,000 NtirVhr (23 MW equlv) of flue gas.
Background of Process Developer
Unitika has a liaison office in New York which should make this process
available to the U.S. market.
Published Economic Data
The estimated capital investment for the NOX removal system capable of
treating 500,000 Nm3/hr (167 MW equiv) of flue gas is $4M or about $24/kW
(46). The revenue requirement for this facility with 300 ppm NOX in the
inlet gas is $8.60/kl of oil. Assuming a kl of oil = 12,000 Nm3 of gas, the
revenue requirement is about 2.1 mills/kWh (46). All of the above values
are assumed to be based on a Japanese location and 1976 costs,
Raw Materials, Energy, and Operation Requirements
The estimated consumption of these for an NOX removal system handling
500,000 Nm3/hr of flue gas containing 1,000 ppm SOX and 300 ppm NOX is as
follows:
Material Quantity
3 short tons/day
Fuel 16.7 short tons/day
Electricity 917 kW
Hence, the electrical requirement represents 0.55% of the total equivalent
plant output, assuming 3000 Nm3/hr - 1 MW. Operating manpower, maintenance,
and technical support needs are not available.
347
-------
Technical Cons ider at_lpn_%
Data from NOX removal bench-scale tests, indicate the NH^rNO mol ratio
ranges from 1:1 to 1.1:1 and the reaction temperature varies from 320-410 C
(590 to 770°F). The space velocity tested varied from 10,000 to 15,000 hr"1.
The NOX removal essentially remained constant at 90-92% during 45-day opera-
tions at inlet gas SOX concentration of 100-1000 ppm and also for 360 ppm
SOX levels for a 300-day-long test. The conditions and results of three
bench-scale tests are shown in Table 29.
TABLE 29. BENCH-SCALE TEST CONDITIONS FOR UNITIKA (NOX) PROCESS (46)
Catalyst name
Test duration, days
Gas flow rate, Ntn^/hr
Space velocity, hr~l
Inlet NOX, ppm
Inlet SOX, ppm
Reaction temperature, C
NH3:NOX mol ratio
NOX removal efficiency, %
OL-3
300
200
15,000
180
360
370-410
i:,l:l
90-92
UL-2
45
150
10,000
180
1,000
320-380
1:1
90-92
UL-2
45
150
10,000
300
100
340-360
1:1
90-92
With the high space velocities employed in this process, the size of
the reactor should be slightly smaller than roost SCR processes using 300-400 C
reaction temperatures. The majority of the construction material for equip-
ment will probably be carbon steel.
Environmental Considerations
The only information obtained concerning environmental considerations
is on sensitivity of NOX removal to operating conditions as given in Table
29.
Critical Data Gaps andPoorlyUnderstood Phenomena
The data gaps Include the following areas.
1. Process flow scheme and process control plan
2. Development; history and present status
3. Operating manpower, maintenance, and technical support requirements
4. Retrofit applicability and turndown capability
5. Catalyst composition and shape
6. Waste disposal requirements
7. 113 emissions in outlet gas
8. Maximum particulate level allowed
348
A
-------
Advantages and Pisadvantages
The advantages and disadvantages of the Unitika process for removal
of NOX only are as follows:
Advantages
1. Achieves >90% NOX removal efficiency
Disadvantages
1. Has not been tested on coal-fired flue gas
2. Has been tested only in a bench-scale unit
3. Requires somewhat clean (degree of particulate removal required
is uncertain) gas feed
349
-------
UNIVERSAL OIL PRODUCTS PROCESS - SHELL CuO PROCESS - DRY, SCR (NOX-SQX)
Process Description and Principles of Operation (92, 93)
Universal Oil Products, Inc., (UOP) offers a dry simultaneous HOX-SOX
removal process using the Shell CuO process. The S02 in the flue gas reacts
with CuO to form CuS04 and the S02 is released later during a regeneration
step to be used in alternative ways. In the presence of the CuO and CuS04
catalysts, NOX is reduced to harmless N2 and 1^0 by reaction with NH3- Two or
more reactors are needed because while NOX and SOX removal (the acceptance
stage) is underway in one reactor, regeneration of the catalyst in the swing
reactor is being performed.
A proposed flow scheme for treating flue gas from a commercial-scale
coal-fired boiler is shown in Figure 107. The flue gas leaving the boiler-
economizer at about 400°C (750°F) has NHg injected into the gas stream and
proceeds to a reactor in the acceptance stage. The reactor contains a fixed
bed of the parallel-passage design. The flue gas passes through the parallel-
passage reactor by flowing across the face of the acceptor material and not
through it. The acceptor material is Cu on a special A^Qj support. The Cu
is rapidly oxidized to CuO upon introduction of the flue gas to the reactor
as follows:
Cu(s) + l/202(g) * CuO(s) (291)
The S02 in the flue gas reacts with this CuO to form CuS04 as shown below:
S02(g) -4- l/202(g) 4- CuO(8) - CuS04(s) (292)
The CuS04 is a catalyst for the reduction of the NOX in the flue gas by
combining with OT^j to create N2 and H20 as follows:
6™(g) + 4NH3(g) * 5fl2(g) + 6H2°
The treated flue gas then leaves the reactor and passes through the air heater
and particulate removal equipment, such as an ESP, before going to the stack.
When the acceptor material becomes saturated with the S02 so that the
removal efficiency attains the limiting value, the flue gas flow is diverted
to a reactor containing regenerated acceptor material. The spent acceptor is
regenerated with a steam-diluted, l^-containing regeneration gas which may be
produced, for example, for steam-naphtha reforming or coal gasification. This
regeneration occurs at the same temperature of 400°C as acceptance. Therefore,
heating or cooling of the acceptor is not required. The CuSO^ is converted to
elemental Cu, S02, and H20, while any unreacted CuO is reduced to Cu and H20.
The following equations express the above reactions.
350
-------
RflRTICULATE REMOVAL
AND STACK
BOILER
FLUE GAS
LO
NH3
OFF
GAS
PRODUCT
(S,S02 (t),OR H2S04)
Figure 107. Flow Diagram of UOP Process.
-------
CUS°4(S) + 2H2(g) * Cu(s) + S°2(g) + 2H2°(g)
CU°(S> + H2(g) -^Cs) + H2°(g)
The regenerated reactor Is now ready to be returned to the acceptance opera-
tion mode. The reactor is purged before and after changing from modes of
acceptance to regeneration and back to acceptance.
The regeneration off -gas contains SC^, %() vapor, and traces of unreacted
reducing gas and may be handled by various means depending on the desired
product and the inerts in the reducing gas. Since this FGT process is cyclic,
the off-gas production is also cyclic and before further processing of the
regeneration off-gas, flow variations must be minimized. The bulk of the H£0
vapor also needs to be condensed. Therefore, the off -gas passes through a
waste heat boiler to recover sensible heat and a direct-contact cooler for
heat removal below the dew-point temperature. A gas-compressor /gas-holder is
considered the simplest system for dampening the flow. S02~concentrated gas
is released from this equipment on flow control to either of the following
production purposes: (1) elemental S production by the modified Glaus method,
(2) liquid S02 by liquefaction, and (3) 12804 production by oxidation. The
choice of the above methods for regeneration off-gas treatment is primarily
economic. Figure 108 exhibits the option for the workup sections for the SC^-
rich regeneration off-gas.
Normal operating conditions which may be expected for this process are as
follows :
Maximum particulate Full loading (>10
Pressure drop across the reactor 5-6 in. H20
NH3:NOX niol ratio (1.0-1.2):!
Average space velocity 5,000-8,000 hr
Temperature 4QO°C (750°F)
UOP states that NOX and SOX removal efficiencies of 90% for each may be
achieved .
There are several other important characteristics of this process. With
this FGT system tied into the boiler system between the economizer and air
heater, the most practical flue gas reaction temperature control method,
though not the only way, involves regulating either the flow of flue gas or
boiler feedwater through or around the exchanger. This controls the heat
transfer rate. Also, the FGT system and boiler may operate independently of
each other with an open bypass system. The FGT system pulls the flue gas
from the flue gas duct into its fan and expels the gas Into the header to
which the reactor inlets are connected. The treated gas then returns to the
flue gas duct downstream of the FGT inlet. With no flow restriction in the
flue gas duct between the FGT inlet and return, a bypass around the FGT unit
is created. A continuous stream of gas is fed to the reactors by the FGT fan
regardless of the flue gas generation rate. Any excess flue gas from the
352
A
-------
i REDUCING GAS
REACTOR SECTION
FLOW SMOOTING
WORK UP SECTION
SUPPLY SECTION
SECTION
U)
Ul
U)
Nl
STEAM
NAPHTHA
REFORMING
COAL
GASIFICATION
FLUE
NOTE: GAS
DASHED LINE REPRESENTS
SECTIONS INCLUDED IN
ECONOMIC EVALUATION.
GASHOLDER
COMPRESSOR
ABSORBER
STRIPPER
MODIFIED
CLAUS
t
SULFUR
V
LIQUEFACTION
\
S02
OXIDATION
H2S04
Figure 108. Diagram of Alternative Methods Used With UOP Process for the Following Sections:
(1) Regeneration Gas Supply, (2) Flow Smoothing, and (3) Workup Section (92).
-------
boiler above the FGT fan rate may bypass the FGT system while gas may be
recycled from the FGT discharge to the FGT fan inlet if the flue gas produced
in the boiler is less than the FGT fan rate.
Also, all the reactions occurring in the acceptance cycle are exothermic,
and by having the FGT system located upstream of the air heater, this heat of
reaction added to the flue gas is recoverable and reduces the fuel required
for the boiler. The heat of the regeneration off-gas is recovered in the
waste heat boiler and contact cooler. UOP states that 863 is also removed
by the Cu-cn-A^O^ acceptor with the result that the dew point is lowered
allowing stack temperatures of the treated gas to be equivalent to flue gas
from low S fuels.
Status of Development
Shell began laboratory bench-scale testing on a catalyst search and
desulfurlzation process in the early 1960's. In 1967 a unit of 0.2-0.3-MW
capacity was built near Rotterdam (at Pernis) which used a parallel—passage
reactor and unit cell acceptor for desulfurization. The fuel source was
high-S fuel oil. This unit was operated over a 4-yr period for more than
20,000 cycles. In August 1973, a 40-MW equiv FGD unit was placed on stream
at Showa Yokkaichi Sekiyu (SYS), Japan. This unit was designed for an oil-
fired boiler flue gas containing 2500 ppm SO^. About 90% S02 removal was
obtained. This unit at SYS has functioned with simultaneous NOX-SOX removal
since mid-1975 when NH/j injection equipment was permanently installed. Also,
a 0.6-MW equiv-FGD pilot plant of Tampa Electric Company (TECO) in Tampa,
Florida, has been operated for 3 yr on coal-fired flue gas from 3.5% S coal.
The slip stream of flue gas was taken either upstream or downstream of the
ESP. Some of the demonstrated results from the above operations include the
following:
Pernis
• The preferred regeneration gas is diluted ^-containing gas.
* The acceptor possessed excellent physical and chemical stability and
will have a life in excess of 8000 cycles.
* Corrosion rates with FGD conditions were established for a number
of metals.
* Pressure drop across the reactor remained low and constant,
SYS
• Unit cells were handled and loaded into reactors with no difficulty.
* Actual performance closely matched results from computer programs
simulating commercial performance indicating the usefulness of
computer programs for reactor design and design optimization.
354
A
-------
* Engineering scaleup consists mainly of paralleling standardized unit
cells.
• The unit is operated by an automatic sequence controller requiring
practically no operator attention after startup periods.
• Special valves functioned properly to isolate reactors during
regeneration,
* Open bypass performs as intended to make FGT and boiler operations
independent of each other.
TECO (In addition to some items already stated for SYS)
« Acceptor is chemically and physically stable after 13,000 cycles (equiv
to a 3-yr life) with 90% S02 removal in 4-m bed and pressure drop of
5-6 in,, H20,
* Process performed well with particulate loading equal to 10 gr/sft^
(about 23 gm/Nnr*) (greater during soot blowing); in situ cleaning of
reactor internals was required and developed,
Background of_Process Developer
Shell International Petroleum Company developed this process and has been
involved in the program since the early 1960*9. The TOP Process Division holds
the worldwide licensing rights to the Shell FGD process (with the exception of
the Far East). The Shell FGD system is commercially available for oil-fired
boilers. The simultaneous NOx and S02 removal has been used at SYS unit since
1975 on an oil-fired boiler while the TECO pilot plant has operated on a coal-
fired flue gas for 3 yr. This process is readily accessible to the U.S.
market.
Published Economic Data
UOP estimates the total capital Investment for a N0x-S0x removal system
treating flue gas from a 500-MW plant to be about $131/ktt (92). The revenue
requirement for this same 500-MW case is equiv to 5 mills/kWh (92). These
values are based on mid-1977 costs with a mldwestern U.S. location. The
factors describing the above FGT system and providing the basis for the
economics Include the following:
Flue gas to reactors, Nm^/hr 1,657,000
Temperature of flue gas, °C 400
S02 content, ppm 2,580
NOX content, ppm 634
S02 removal, % 90.5
NOX removal, % 90.0
355
-------
Number of reactors, accepting/regenerating 6/2
Bed length, m 6.0
Reactor diameter, m 7.38
Acceptance time, mln 64.3
Acceptor quantity, kg/reactor 54,000
System pressure drop, cm 1^0 40
As Indicated in Figure 108 there are a variety of combinations possible for
the sections supporting SOX removal. The dotted lines in this figure reveal
the integrated units chosen for the above economic evaluation. The gas-
compressor/gas-holder was chosen because of economics. The modified Claus
system and steam-naphtha reforming unit were chosen because more information
was available on these sections.
Raw Materials, Energy, and Operation Requirements^
The amounts of raw materials and energy required for the same 5QQ-MW case
stated in Published Economic Data follow*.
Material Quantity
NH3 21 short tons/day
Electricity 7,712 kW
Steam, (net) 187 short tons/day
Naphtha 57.23 bbl/hr
The above steam requirements are based on 867 short tons/day steam usage for
FGD and NH3 injection and a steam credit of 680 short tons/day for the workup,
modified Claus, and steam-naphtha units. The total electric usage represents
1.5% of the total equivalent power output of a power plant.
Also, a heat credit totaling 131 MBtu/hr is claimed by UOP. Sources of
feeat credit in the FGT section include fan compression, reaction heat from
both acceptance and regeneration, and dew-point suppression.
For the 500-MW case, the labor and supervision needs were estimated to be
21,000-hr annually. At an average cost of $8/hr, this equals $168,000/yr. The
maintenance costs were estimated to be about $1,046,000, which is about 6% of
the total annual revenue requirement. UOP states that for the 4Q-MW equiv
unit at SYS, no additional manpower was needed. Excluding startup and initial
operation period, the required operator attention was almost nil.
Technical Considerations
The UOP system is composed of several more operation steps than most dry
SCR NOX removal methods, since S02 processing equipment is needed for treating
the regeneration off-gas. However, the units used in the UOP process involve
proven technology and straightforward chemistry. Flue gas valves of large
diameter are needed to provide tight seals at high temperatures and over
356
A
-------
several thousand opening and closing cycles. These valves must separate the
reactors undergoing acceptance or regeneration from the regeneration or main
flue gas respectively. UOP states that the valves selected from this purpose
have performed satisfactorily at SYS and TECO.
The NQX and SOX removal unit is controlled and protected with the use of
an automatic sequence controller by setting the acceptance and regeneration
cycle lengths, monitoring all valve positions, temperature, and flow rates.
At SYS, the NOX at the outlet of the FGT system is measured continuously by
the chemlluminesence method and the NH3:NQX ratio is controlled manually.
It should be noted that at the end of the regeneration period the acceptor
Is In the form Cu. With all the acceptor In this form and several minutes
required for oxidation of the Cu to CuQ, if NQX removal were attempted at the
beginning of the acceptance stage, the NOX level in the treated gas would
actually increase because of NH3 conversion to NOX. At SYS, NH3 injection is
automatically controlled to start a few minutes after the acceptance cycle is
begun. More efficient methods recommended for a new design are as follows:
(1) preoxidize the acceptor before introducing the flue gas and Injecting NtTj
or (2) allow a portion of the acceptor bed to remain unregenerated as CuSO^.
This second method may be the best solution.
the NQX and S02 removal efficiencies are considered independent of inlet
NOx and S02 concentrations respectively. The NOX reaction is first order and
the NOX reaction rate is directly proportional to the NOX concentration. The
NOX removal rate will Increase as bed length or cycle length increases.
The reactor size for this NOX and S0£ removal method should be about the
same as other SCR processes with similar operating conditions. The number of
reactors required for this Shell CuO process is greater than other SCR
processes since the same vessel is used for both acceptance and regeneration.
The ratio of reactors in the acceptance stage to those in the regeneration
cycle for the 500-MW case is 6:2. However, this is a design variable, for
example, if lower-S western coal had been the basis, the ratio would probably
be 5:1.
In -examining retrofit applicability, the UOP Shell CuO process with
simultaneous NOX and SOX removal would require reheat equipment such as a heat
exchanger and heater if installed after the air heater. The simultaneous
NOX-S02 removal process would not be suitable for retrofit on a plant with an
existing FGD system. However, the UOP Shell Cuo process can be operated in
an N0x-only removal manner as reported under a separate detailed process
description.
With the open bypass system, the turndown ratio Is unlimited, capable of
operating from full boiler load to complete recycle. Naturally, operating
with total recycle produces a decrease in reactor regeneration frequency.
357
-------
The materials of construction are considered proprietary information;
however, UOP states that no unusual materials are used. UOP has derived
corrosion rates for various metals under FGD reactor conditions. Service life
of at least 15 yr for the reactor internals has been estimated with the proper
construction materials.
A marketable quality byproduct is manufactured by this process. Depending
upon economics, elemental S, liquid S(>2» or a concentrated S02 gas feed to an
H2S04 plant may be produced.
Enyir onmental Cons id era t ions
An advantage of the UOP Shell CuO process is that it is capable of
simultaneous NOX and S02 removal. Data from tests at SYS with a 4—m bed are
presented in Figure 109 showing the NOX and SOX concentration varying with
time at different NH3:NOX mol ratios at the reactor outlet. The SOX con-
centration decreases rapidly from above 1200 ppta to below 100 ppm as Cu is
converted to CuO which then reacts with S02 to form CuS04- However, am CuO
is changed to CuS04, the S02 level in the treated flue gas rises to about
400 ppm after 120 min. The NOX removal increases with cycle length as more
Cu is converted to CuO and eventually CuS04, the catalyst for the NO^
reduction reaction. The effects of NH3:NOX ratio and the fraction of acceptor
converted to CuS04 upon NOX removal are indicated in Figure 110. These data
are from similar tests at the SYS unit. A 90% removal of both SOX and NOX is
expected with the use of a 5-m bed and by allowing 1 m to remain unregenerated;
that is, the acceptor is in the CuS04 form for NOX removal even at the first
introduction of flue gas.
There is no significant waste disposal needed for this system other than
that associated with the production of byproducts from the S02- An environ-
mental concern would be the leakage of excess NHj to the atmosphere; however,
UOP claims NHj in the treated gas averages 1 ppm.
The parallel-passage reactor has been shown to be capable of operating at
normal conditions with full particulate loading (>10 gr/sft3). Initial
testing with full flyash loading to the reactor at the Pernis unit showed no
deterioration in performance. However, at SYS the high V- arid Na-containing
flyash gradually fouled the reactor Internals forcing runs to be limited to
operating periods of 1-2 mo. The higher particulate loading with coal-fired
flue gas at TECO of 10 gr/sft-* or greater caused a decline in performance
after only days of operation. As a result of this problem> a procedure was
developed and implemented at the Tampa pilot plant which provided in situ
cleaning of reactor Internals during normal operation. This technique allowed
stable performance with high particulate loadings; pressure drop across the
reactor and desulfurization were not affected. In situ cleaning of reactor
internals will be implemented at SYS also.
Testing at TECO also demonstrated that Cl£ and Cl~ in flue gas had no
adverse effect upon the acceptor performance. The loas of Cu was negligible.
358
A
-------
450
FLOW = 137,000 Nm3/h
SO 2 - I260ppmv
NOx ' 293 ppmv
REACTOR BED LENGTH * 4 METER
NOX AT NH, /N0«0,0
20
40 60 80
ACCEPTANCE TIME, MIN
100
120
Figure 109. Performance of Shell FGD Reactor at SYS,
Instantaneous S02 and NOX Slip (92),
359
-------
0.
en
CD*
z
IUU
80
60
50
4O
30
20
10
8
6
5
4
3
2
1
3d
-fa
>fta
X
\^
IS^
Wv
^c^
vO
^
»*~
IXW~-
~^P«N
Sn-L
|.?X
^J
^O^
-jS
>
"•-
F-*^—
"Yl
^N
\L
X
V
N
"HTni
XQ
^^
s.
\
k
\
\
CONDITIONS:
BED LENGTH ^
^
\
\
TEMPERATURE 400°C
REYNOLDS NUMBER 2000
i- NORMAL SYS OPERATION
NH3
MO'L
V
u
o
\
\
\
-
/NO
AR RJ
0.3
0.6
0-8
0.9
1.0
1, 1-
\
X
V
M10"
—
1.2 ~
X
\
O.I 0.2 0.3 0.4 0.5 0.6 0.7
LOADING (Of )
0.8 0.9 1.0
Figure 110.
Mol Ratio
NOX Slip, Percent of Intake vs Loading^ CuSQ,/(CuO + CuSOA)
5;r«Ae«!«iil3or/eaealyst wifch l^g process '-(92),
360
A
-------
Since the regeneration gas contains B^, which may detonate with 02
mixtures, the regeneration gas must be prevented from coming into contact with
the main flue gas. The flammability limits for H2~air mixtures at atmospheric
pressure are 4-74% H2 (89). The reactor is purged before and after each
regeneration cycle. Also, UOP states that since the ^-containing regeneration
gas is diluted with steam, no detonation would result even in the case of
failure of both the purge system and sequence control systems. With this
extreme example, combustion may occur within the closed reactor which would
raise the pressure from atmospheric to about 1.5 atmosphere, the design
pressure for the reactor. Also, UOP feels the safety of this process is
proven by the thousands of hours of safe operation.
Critical Data Gaps and Poorly Understood Phenomena
Although there are no major poorly understood phenomena, the use of an
A^Oj base for the CuSO^ catalyst is questionable since there is a trend of
competing processes to be changing to alternate, supposedly more stable
supports. It is recognized that UOP refers to their catalyst as being on a
"special" A1203 support.
Advantages and Disadvantages
The following are the advantages and disadvantages of the UOP-Shell CuO
simultaneous NOx-SOj^ removal process'
Advantages
1. Removes NOX and S02 simultaneously
2. Achieves >90% NOX removal efficiency
3. Produces marketable byproduct (S, or H2S04, or S02(j%)
4. Has been applied to flue gas from commercial oil-fired boilers
5. Is a slight modification of a commercially available FGD system
6. Operates with full particulate loadings (>7 gr/sft^)
7. Claims <10 ppm by vol NH3 in treated flue gas
8. Claims full turndown capability
Disadvantag^es
1. Has not been tested on coal-fired flue gas
361
-------
UNIVERSAL OIL PRODUCTS PROCESS - SHELL CuO PROCESS - DRY, SCR (NOX)
Process Description and Principles of Operation (92, 93)
In addition to simultaneous NOX-SOX removal, the Shell CuO process
offered by Universal Oil Products (UOP) may be used for NOX removal only.
N0x-only removal may be carried out by eliminating the regeneration step
and subsequent S02 removal facilities from the Shell CuO simultaneous NOX— SOX
removal process,
This process is depicted by the flow diagram in Figure 111. The flue
gas leaves the boiler economizer and NH3 is injected into the flue gas. The
entire gas stream enters the fixed-bed, parallel-passage reactor. The reactor
bed contains Cu (on special Al2<33) which, upon introduction of the flue gas
to the reactor, is oxidized to CuO as follows:
Cu(s) + 1/2°2(gr CU°(s)
The CuO then reacts with the SC>2 to form CuS04 according to the following:
S02(g) 4- l/202(g) + CuO(s) -CuS0 (297)
The CuS04 acts as catalyst for the reduction of NOX by NHj to form N2 and
H20 as expressed below.
6NO, , + 4NH,, % -»- 5N0, , + 6H~0, , (298)
(g) 3(g) 2(g) 2 (g)
The treated flue gas exits the reactor and flows through the air heater,
particulate removal, and FGD equipment before leaving the stack.
Once the CuO is converted to CuSO^, the catalyst remains in this form
for NOX removal and, thus, SOX in the flue gas passes through the reactor
without being removed.
The reactor operating conditions are similar to those mentioned for the
UOP simultaneous NOX-SOX removal process and are as follows:
Maximum particulate loading Full loading (> 10 gr/sft^)
Pressure drop across the reactor 5-6 in. H20
NH3:NOX mol ratio (1.0-1.2):!
Average space velocity 4,000-8,000 hr~^
Temperature 400°C (750°F)
NOX removal efficiency of at least 90% is expected with 4-m-long beds and
removal efficiencies of 95—97% and 99% have been achieved during prototype-
scale testing on 6-m- and 7-m~long beds respectively.
362
A
-------
AIR
PARTICULATE REMOVAL,
FGD, AND /OR
STACK
Figure 111. Flow Diagram of HOP Process.
-------
There are several characteristics of this UOP NOX~only removal process
which are the same as for the UOP simultaneous NOX~SOX removal process. With
this NOX treatment system connected to the boiler system between the economizer
and air heater, the most practical flue gas reaction temperature control method,
though not the only way, involves regulating either the flow of flue gas or
boiler feed water through or around the exchanger to control the heat trans-
fer rate. Also, the FGT system and boiler may operate independently of each
other with an open bypass system. The FGT system pulls the flue gas from
the flue gas duct into its fan and expels the gas into the header to which
the reactor inlets are connected. The treated gas then returns to the flue
gas duct downstream of the FGT inlet. With no flow restriction in the flue
gaa duct between the FGT inlet and return, a bypass around the FGT unit is
created. A continuous stream of gas is furnished to the reactors by the FGT
fan regardless of the flue gas generation rate. Any excess flue gas from
the boiler above the FGT fan rate may bypass the FGT system while gas may
be recycled from the FGT discharge to the FGT fan inlet if the flue gas pro-
duced in the boiler is less than the FGT fan rate.
All the reactions occurring in the reactor are exothermic and by having
the FGT system located upstream of the air heater, this heat of reaction is
recoverable from the flue gas. Fan compression heat is recoverable too.
Status of Development
Shell began laboratory bench-scale testing and a catalyst search and
desulfurization process in the early 1960's. In 1967 a unit of 0.2-Q.3-MW
capacity was built near Rotterdam (at Pernis) which used a parallel-passage
reactor and unit cell acceptor for desulfurization. The fuel source was
high-S fuel oil. This unit was operated over a 4-yr period for more than
20,000 cycles. In August 1973 a 40-MW equiv-FGD unit was placed on stream
at Showa Yokkaichi Sekiyu (SYS), Japan. This unit was designed for flue gas
containing 2500 ppm SC>2 and processes gas from an oil-fired boiler. About
90% S02 removal was obtained. This unit at SYS has functioned with simul-
taneous NOX-SOX removal since mid—1975 when NH3 injection equipment was
permanently installed. Also, a 0,6-MW equiv-FGD pilot plant of Tampa Elec-
tric Company (TECO) in Tampa, Florida, has been operated for 3 yr on coal-
fired flue gas from 3.5% S coal. The slipstream of flue gas was taken
either upstream or downstream of the ESP. Some of the demonstrated results
from the above operations includes the following.
Pernis
* The preferred regeneration, gas is a diluted H2-containing gas
* The acceptor possessed excellent physical and chemical stability
and will have a life in excess of 8000 cycles
* Corrosion rates with FGD conditions were established for a number
of metals
* Pressure drop across the reactor remained low and constant
364
A
-------
SYS
• Unit cells were handled and loaded Into reactors with no difficulty
* Actual performance closely matched results from computer program
simulating commercial performance indicating the usefulness of com-
puter programs for reactor design and design optimization
« Engineering scaleup consists mainly of paralleling standardized
unit cells
* The unit is operated by an automatic sequence controller requiring
practically no operator attention after startup periods
» Special valves functioned properly to isolate reactors during regene-
ration
* Open bypass performs as intended to make FGT and boiler operations
independent of each other
TECO (In addition to some items already stated for SYS)
* Acceptor is chemically and physically stable after 13,000 cycles
(equivalent to a 3-yr life) with 90% SC>2 removal in 4-m bed and
pressure drop of 5—6 in. H20
* Process performed well with particulate loading equal to 10 gr/sft^
(about 23 g/Nm ); in situ cleaning of reactor internals was found
to be needed and a method was developed
This N0x-only removal process has been tested on pilot plants and the
4Q-MW equiv prototype-scale SYS unit with the acceptor material completely
in the form of CuS04.
Backgroundof Process Developer
Shell International Petroleum Company developed this process and has
been involved in this program since the early 1960's. The UOP Process
Division holds the worldwide licensing rights to the Shell FGD process
(with the exception of the Far East). The Shell FGD system is commercially
available for oil-fired boilers. The simultaneous NOX and S<>2 removal has
been used at SYS unit since 1975 on oil-fired boiler while the TECO pilot
plant has operated on a coal-fired flue gas for 3 yr. This process is
readily accessible to the U.S. market.
Published Economic Data
The reported capital investment for an N0x-only removal system treating
flue gas from a 500-MW plant is $31/kW (92). For the same facility, the
reported revenue requirement is 1.4 mills/kWh (92). A Midwestern U.S. loca-
tion and mid-1977 erection charges provide the basis for the above economic
365
-------
values. The parameters characterizing this NOX removal system are as follows:
Flue gas to reactors, Nm^/hr 1,582,000
Temperature of flue gas, C 400
S02 content, pptn 2,580
NOX content, ppm 634
NOX removal, I 90
Number of reactors 4
Bed length, m 4
Reactor diameter, m 8,70
Catalyst quantity, kg/reactor 51,480
Pressure drop across system, cm H20 23
Raw Material, Energy, and Operation Requirements
The raw material and energy requirements for the above 500-MW case
are as follows:
Material Quantity
NH3 21 short tons/day
Electricity 3,242 kW
Steam 13.2 short tons/day
The total electrical usage equals about 0.6% of the equivalent power output
of the 500-MW plant. Also, a heat credit constituting 26.5 MBtu/hr is
reported for fan compression and reaction heat recoverable at the air heater.
The labor and supervisory requirements for the same 500-MW example are
reported as 1750 hr/yr at an average cost of $8/hr. This represents an
annual cost of $14,000. The maintenance needs are reported to cost $194,000
annually, UOP reported that no additional manpower was needed for the 40-MW
equiv-unit at SYS and that the operator control needed was practically nil
after startup and the initial operating period.
Technical Considerations
The UOP Shell CuO process for N0x-only removal is simple requiring only
a reactor, NH3 storage, and injection equipment. The N0x-only removal system
represents significant reduction in complexity, number of operation steps,
and capital and operating costs in comparison to the UOP simultaneous NOX-S02
removal system.
The NOX removal efficiency is reported to be independent of inlet gas
NOX concentration. The NOX reaction is first order and the NOX reaction
rate is directly proportional to the NOX concentration. The NOX removal
rate will increase as bed length increases.
The reactor size may be about average for dry SCR systems since the
reaction temperature is similar and assuming the space velocity ranges from
4000 hr"! to 8000 hr"^. fhe number of reactors reportedly needed for the
500-MW example is four.
366
A
-------
The UOP NQx-only removal system is suitable for retrofit on a plant
with or without an FGD system. Of course, reheat will be needed for the
system located after the air heater. Also, more reheat is required if the
NOX removal system is placed after rather than before a low-temperature FGD
unit .
There are no byproducts made with this process.
With the open bypass system the turndown ratio is unlimited, capable
of operating from full boiler load to complete recycle.
The materials of construction are considered proprietary information;
however, UOP states that no unusual materials are used. UOP has derived
corrosion rates for various metals under FGD reactor conditions. Service
life of at least 15 yr for the reactor internals has been estimated with
the proper construction materials.
Environmental Considerations
NOX removal data from tests at pilot plants and the prototype-scale
(40 MW equiv) unit at SYS are shown in Figure 112. The acceptor used was
completely in the sulfated form. The results are plotted as a function of
bed length since NOX conversion fluctuated with time on stream, reactor
internals design, and operating conditions. NOX removal efficiencies of
95-97% for 6-m-length beds and 99% for 7-m-length beds have been obtained
during tests.
No waste disposal is needed for this system. The NH3 exiting with the
treated gas reportedly averages 1 ppm.
The parallel-passage reactor has been shown capable of operating at
normal conditions with full particulate loading ( >10 gr/sft-*) . Initial
testing with full flyash loading to the reactor at the Pernis unit showed
no deterioration in performance. However, at SYS the high V- and Na- contain ing
flyash gradually fouled the reactor internals forcing runs to be limited to
operating periods of 1-2 mo. The higher particulate loading with coal— fired
flue gas at TECO of 10 gr/sft-* or greater, caused a decline in performance
after only days of operation. As a result of this problem, a procedure was
developed and implemented at this pilot plant which provided in situ cleaning
of reactor internals during normal operation. This technique allowed stable
performance with high particulate loadings; pressure drop across the reactor
and desulfurlzation were not affected. Cleaning of reactor internals will
be implemented at SYS also.
Testing at TECO also demonstrated that Cl£ and Cl~ in flue gas had no
adverse effect upon the acceptor performance. The loss of Cu was negligible.
C r it ic a 1 Da t a Gaps and Poor ly. Und er s t o o d Ph enomena
Although there are no major poorly understood phenomena, the use of an
A1203 base for the CuS04 catalyst is questionable since there is a trend of
367
-------
100
en
o" 1°
*v
%
N\\
\
1
V
v\
\ X
^ \
^\
™ \ |
^
I \
^ \
\ ^
\
\
\
\
\ '
%,
"*
\
\
\ c
\
\
\
\
CONDITIONS:
Cu AS CuSO
NH3/NO I.1H.5
PERF
IAL EXPEC"
ORMANCE
FED
• PERFORMANCE
AFFECTED BY
QUTSmr ranrnnd
J
^%
\
\
\
s
\
\
\
'.
\
V
^
\
\
\
^
\
\
s.
\
\
\
\
\
0
2 3 • 4
iED LENGTH, METERS
Figure 112. Unconverted NOX as a Function of
Catalyst Bed Length for UOP Process (92).
368
A
-------
competing processes to be changing to alternate, supposedly more stable
supports. It is recognized that OOP refers to their catalyst as being on a
"special" A^Orj support,
Advantages and Disadvantages
The advantages and disadvantages of the UOP-Shell CuO N0x-only removal
process are as follows:
Advantages
1. Achieves >90% NOX removal efficiency
2. Has been applied to flue gas from commercial oil-fired boilers
3. Is a slight modification of a commercially available FGD system
4. Operates with full particulate loading (>7gr/sft^)
5. Claims <10 ppm by vol NHn in treated flue gas
6. Claims full turndown capability
Disadvantages
1. Has not been tested on coal-fired flue gas
369
-------
LEAVE THIS PAGE BLANK
370
-------
CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
Although much development work has been performed for flue gas denitri-
fication, most processes have been developed only through bench-scale or
pilot-plant operation using flue gas from gas- or oil-fired sources. The most
tested and advanced type of FGT process is now SCR. Several commercial, oil-
fired SCR denitrificatlon facilities have been operated; about half of the
NOX removal processes included in this report are the SCR type. The majority
of the SCR processes are reportedly capable of excellent NOX removal (>90%).
Looking at other dry methods of NOX removal, the selective noncatalytic example
(Exxon Thermal) has been tested on a commercial oil-fired boiler but the NOX
removal efficiency achieved may only be 60-70%, The dry, adsorption process
(Foster Wheeler) has been tested on a coal-fired flue gas at a prototype unit.
However, NOX removal efficiency was only 25-30%. The Ebara-JAERI radiation
process has reached the pilot-plant status and may remove about 90% of the
NOX. The major drawbacks with the dry flue gas denitrification schemes which
will require attention in future testing on coal-fired sources are the sensi-
tivity to particulates In all dry processes and the formation of NH4HS04 for
NH^-reductant processes.
Though not as well developed as SCR methods, the oxidation-absorption-
reduction method Is the most advanced wet NOX removal method. Two of the
oxidation-absorption-reduction processes have been tested on prototype units,
with the remainder having been tested on bench-scale or pilot-plant units.
Reportedly, these processes obtain moderate NOX removal (80-90% efficiency).
The absorption-reduction processes have only reached bench—scale or pilot-
plant status with 60-85% NQX removal efficiency. Of these processes, only
PENSYS has been tested on coal-fired flue gas. Of the oxidation-absorption
processes which have been tested, one has operated on a prototype unit and
the other has operated on a pilot plant with coal-fired flue gas with no
published results to date. Of the absorption-oxidation processes, only pilot-
plant studies or smaller are presently active. The major adverse factors to
be considered with wet processes in future testing include waste stream treat-
ment, expensive oxidants, and the complicated series of processing steps which
are usually required.
The reported economics vary greatly for these NOX removal processes,
even when the processes are within the same classification. The economics
are given as reported from the process developer and represent various bases
which impairs any meaningful comparison. Revisions in the economics may be
expected as further testing is accomplished on larger facilities.
371
Preceding page blank
-------
It is concluded that several adverse process characteristics should be
considered in future development regarding both dry and wet NOX removal pro-
cesses. In addition, further testing must be done on pilot-plant and proto-
type units with coal-fired flue gas before any FGT processes are considered
feasible by the utility industry for commercial, coal-fired power plant
application.
One purpose of this first phase of the NOX removal process study has
been to perform a state-of-the-art review of all the flue gas denitrification
systems currently undergoing development. The primary evaluation criteria
were the technical and commercial feasibilities of each of these denitrifi-
cation processes. Once this initial technical survey of all of the processes
had been completed, the most viable processes were selected for preliminary
economic analysis in the second phase of the study. These selected processes,
which are listed in Table 30, have been chosen based on three criteria.
• Technical considerations
* Development status
* Representative sample
TABLE 30. PROCESSES SELECTED FOR FURTHER STUDY IN PHASE II
Process
Type of process (classification)
UOP Shell Copper Oxide
UOP Shell Copper Oxide
Hitachi Zosen
Kurabo Knorca
Moretana Calcium
Dry simultaneous S02~NOX
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Dry NOX only
(Selective catalytic reduction)
Wet simultaneous S02~NOX
(Oxidation-absorption-reduction)
Ishikawajima-Harima Heavy Industries Wet simultaneous SO2~NOX
Asahi Chemical
MON Alkali Permanganate
(Oxidation-absorption-reduction)
Wet simultaneous S02~NOX
(Absorption-reduction)
Wet NOX only
(Absorption-oxidation)
372
A
-------
Itach of these selection criteria Is probably obvious except for the
third, representative sample. There are five different types of dry pro-
cesses and four types of wet processes in various stages of development and
an attempt was made to select one process from each type. Thus, in addition
to selecting the best processes from the standpoint of both technical feasi-
bility and development status, an attempt was made to select only the best
process, based on these two criteria, from each type of flue gas denitrifica-
tion process. A preliminary economic analysis of differing types of processes
in the second phase would allow a fair comparison between the various tech-
nologies available for flue gas denitrification. This broad coverage would
also prevent possible later questions on why certain types of flue gas deni-
trification processes were dropped from consideration simply because they
had only been tested in a bench-scale unit.
TECHNICAL CONSIDERATIONS
The primary consideration involved in selecting processes for further
study was whether the process was technologically feasible, i.e., whether
the basic reactions and operations involved in the process were either well-
known or had been demonstrated in actual operation. With the exception of
three processes, all of the denitrification systems included in this study
have been tested on oil-fired flue gas. Thus most of these processes have
demonstrated their capability in treating clean flue gas.
Since coal is the primary utility boiler fuel in the U.S. and is expected
to be used almost exclusively in all new fossil-fueled utility boilers, a
second technical consideration is the ability of the process to handle the
increased particulates, SOX, NQX, and Cl~ levels associated with coal com-
bustion. Although essentially all of the development work for these deni-
trif ication processes have been done on oil-fired flue gas, most of these
processes are expected, with minor modifications, to be able to handle the
flue gas from coal-fired boilers. With the addition of either an efficient
hot ESP (primarily for the dry processes) or a closed-loop prescrubblng
section (primarily for the wet processes), the particulate problem is
expected to be controlled. In addition, most of the dry processes have
developed catalysts or designed reactors which minimize the effects of dust,
i.e., the catalysts in packed-bed reactors are either honeycomb, ring, or
pipe shaped or if the catalyst particles are present in pellet form they
are packed in either a moving-bed reactor or a reactor with a specially
designed internal structure.
For the dry N0x-only processes increasing the NOX concentration in con-
verting from oil-firing to coal-firing will increase the NHj consumption
rate and, hence, substantially affect the economics of the process. However,
from a technical point of view the increased NOX concentration should not
cause any problems. The increased levels of the SOX are also not expected
to have a significant impact on,the dry N0x-only processes. The Cl~, however,
depending on its actual chemical makeup, may present problems by reacting
with the catalyst particles.
373
-------
For the wet simultaneous SOo~NOx processes with a prescrubber, the
increased particulates, SOX, NOX, and Cl~ in the flue gas are expected to
only have a minimal impact from a technical standpoint.
Although the projected economics of each of these processes would be
a valuable addition to the selection criteria, the economic data included
for most of these processes are estimates made by the process developer
and as such are based on widely ranging premises. Any attempt to convert
these estimates to a single standard set of premises without investigating
the estimate details to determine the items covered would only add additional
uncertainties, such that all the resulting calculated values would he essen-
tially equivalent within the errors of uncertainty. For this reason, the
processes were not selected based on their projected economics.
DEVELOPMENT STATUS
The second major selection criteria was the development status of the
process. The development status was chosen since it reflects a degree of
achieving full— scale commercial operation and also Indirectly the confidence
level which the process developers have in the technical and economic feasi-
bility of their system. Ideally, the develo-pment of the selected process
would have reached the commercial stage (>50 MW) and be treating flue gas
from a coal-fired boiler. However, since only one or two of the processes
have been extensively tested on coal-fired flue gas, this selection criteria
was reduced to simply the size of the unit in which the process was currently
being tested. For the purposes of this study, the following size ranges
have been defined for each stage of development.
Stage of development Size of unit
Commercial MW ^
Prototype 5< MW <50
Pilot plant 0.55 MW < 5
Laboratory MW
-------
TABLE 31, STATUS OF DEVELOPMENT
OF THE FLUE GAS DENITRIPICATION PROCESSES
Process . Size, MH equivalent
Commercial scale
Hitachi Zosen 275
Hitachi, Ltd. 170
Sumitomo Chemical 100
Mitsui Engineering and Shipbuilding (Dry) 67a
Exxon Thermal Denox 53
Prototype-scale
Moretana Sodium 40
UOP (NOX-SG2) 40
UOP (NO^-only) 40
Tokyo Electric-Mitsubishi Heavy Industries 33
Mitsui Toateu 30
JGC Paranox 23
Foster Wheeler-Bergbau Forschung 20
Kobe Steel (Wet) 17b
Kurabo Knorca 10
Horetana Calcium 8.3
Pilot-plant scale
Mitsubishi Kakokl Kalaha 4.7
Takeda • 3.3
Sumitomo Heavy Industries (ROX-S02) 3.3
Ishlkawajlma-Harlma Heavy Industries 1.6
Kawasaki Heavy Industries 1.6
Kureha (Wet) 1.6
Kureha (Dry) 1.6
Pittsburgh Environmental and Energy Systems 1.6a
Eneron 1.5
Unitlka (KC^-SOa) 1.5
MON Alkali Permanganate 1.3
Mitsubishi Heavy Industries (Dry) 1.1
Ebara-Jaerl 1.0
Mitsubishi Heavy Industries (Wet) 0.6
Bench-acale
Sumitomo Heavy Industries (NOx-only) 0,5
Chlyoda Thoroughbred 102 0.3
Kobe Steel (Dry) 0.3
Chisso Engineering 0.10
Unitlka (N0x-only) 0.07
Mitsui Engineering and Shipbuilding (Wet) 0,05
Asahl Chemical 0.02
Exxon (SCR) 0.003
Nottested on flue gas
Nissan Engineering
Ralph M. Parsons
Ube (Dry and Wet)
Mitsubishi Petrochemical Uncertain
n. Under construction
b. Terminated (new process under development).
375
-------
reached the pilot-plant stage. Thus by. only considering the status of
development, many important and perhaps fundamental comparisons would be
eliminated. Therefore, the third selection criterion given in the following
section was incorporated into the evaluation section,
REPRESENTATIVE SAMPLE
During the selection processes, the above-mentioned criteria, i.e., the
technical considerations and the status of development, were modified in
order to obtain a representative sample of all the types of flue gas denitri-
fication processes currently undergoing development. In other words, attempts
were made to select the best process from each type of treatment scheme if
that process was deemed technologically feasible. For example, although the
Asahi Chemical process has only been tested in a laboratory-scale unit, it
represents substantially different technology than the other selected pro-
cesses. In addition, it is fairly representative of an entire class of NOX
removal processes that, based on the other selection criteria, would not be
acceptable for further study in Phase II. Although an example of each of
the various technologies could not be selected, those categories which are
not represented among the selected processes were categories which only had
one or two example processes. Thus those categories did not represent a
sizable portion of the current NOx FGT technology,
The selection of two wet, oxidation-absorption-reduction processes, as
shown in Table 30, appears to be consistent with the specific purposes of
having a representative sample. However, these two oxidation-absorption-
reduction processes, the Moretana Calcium and the Ishikawajima-Harlma, have
a major difference In that in the former C102 is used as the gas-phase
oxidant where as Ishikawajima-Harima has chosen 63 as the gas-phase oxidant.
These two oxidants, although serving the same purpose, offer substantially
different complications In both their generation equipment and the resulting
processing scheme and, therefore, different costs and technologies. Thus,
the selection of these two processes is consistent with the avowed purpose
of obtaining a representative sample.
Similar differences are noted in the dry NOX removal processes. The
UOP Shell CuO system represents a unique process for two reasons. This
process can operate not only as a dry, N0x-only removal system, but also as
a dry simultaneous S02~NOX system which allows a comparison of both types
of dry process with a somewhat simplified calculation. The second reason
for selecting the UOP process Is that it has a parallel passage reactor
design which minimizes the adverse .effects of the heavy particulate loading
In the coal-fired flue gas. This reactor system has also been tested on
coal-fired flue gas, although only in the FGD mode.
Both the Hitachl-Zosen and the Kurabo processes are based on the SCR
of NOX using NH3, but they have one fundamental, and therefore important,
difference. The Hitachi-Zosen process uses a typical fixed-bed reactor
containing a honeycomb-shaped catalyst to prevent excessive plugging by
the dust in the flue gas, whereas the Kurabo process is based on a continuous
moving-bed reactor containing spherical-shaped catalyst particles. Thus, by
376
A
-------
Including both the HItachi-Zosen and the Kurabo processes in the preliminary '
economic analysis phase, a general comparison of the moving-bed reactor with
a fixed-bed reactor for NOX removal will be possible.
If sufficient funding to select three additional processes were available,
the following dry processes could be included for further study: the Ralph
M, Parsons, the Exxon Thermal, and the Takeda. The Ralph M. Parsons process
Is a nonselective catalytic reduction process and, hence, would result in a
direct comparison of nonselective with selective catalytic reduction. The
evaluation of the Exxon Thermal process would be a good example for comparison
in that the NH3 is injected directly into the boiler to selectively reduce
the NOX and thus eliminate the need for an add-on catalytic denitrification
system. The Takeda process, since it is based on both SCR of NOX and activated
C adsorption of 862, would provide a comparison of base-metal catalysts with
the activated-C-type reduction process for SCR as well as for comparison of
differing dry simultaneous S02~NOX processes.
It should be pointed out, that in some cases there are several other
processes equivalent to those chosen for further study in Phase II. Their
elimination should not be construed as adverse comments on these processes,
but only as representing similar technologies that do not possess significant
differences from those processes selected. For example, the Ishikawajima-
Harima process was chosen as the representative for the 03~based oxidation-
absorption—reduction processes but several others would be equally well
representative of this group, including Mitsubishi Heavy Industries process
and possibly the Chiyoda Thoroughbred 102 process. The Mitsui Engineering
and Shipbuilding dry process is similar to the Hitachi-Zosen process and the
Kobe Steel process is similar to the Kurabo process, but, because of the
scope of the second phase, only one of each type could be selected. Within
the errors of calculation, the results obtained in the preliminary economic
analysis for the selected processes would be applicable to these processes
which were not selected.
RECOMMENDATIONS
From the preceding conclusions a more detailed analysis, Including
process design and preliminary economics, for the eight selected NOX removal
processes is recommended to be Included in the second phase of this study.
Premises will be chosen such that a uniform comparison of each selected NOX
removal system will be obtained for a single case. The general premises for
this process design and preliminary economic study will be established on a
new, 500-MW coal-fired power plant burning coal with 3.5% S (dry basis). The
NOX and SOX emissions in the untreated flue gas will be about 600 ppm and
2400 ppm by vol respectively.
During Phase II the process design for each selected method will include
the following: process flow diagram material and energy balances, fuel
stream requirements and waste treatment needs. The preliminary economics
will Include both capital investments and revenue requirements. The capital
Investment based on mid-1979 dollars will encompass costs for major equipment
site preparation and installation labor and material for electrical, ductwork,
377
-------
piping, etc. Titie revenue requirements will be based on mid-1980 dollars.
This section will contain annual revenue requirements and also projections
for Che availability and cost of process reagents as a function of time.
With these recommendations, Phase II studies will provide a direct
comparison of the various flue gas denitrification technologies. The results
should establish which of the following alternatives are more economically
promising.
1. NO absorption vs gas-phase oxidation
2, Wet simultaneous process vs dry simultaneous process
3. Wet simultaneous process vs wet FGD + dry, M>x-only process
4. Wet NOX only vs dry NOX only
5, Dry simultaneous vs wet FGD + dry, N0x-only process
378
-------
REFERENCES
1. Aerotherin Division of Acurex Corporation, NOX Control Review, _!_, No. 1,
March 1976. ~"
2. Ando, J. (Chuo University, Tokyo, Japan). Private communication,
August 1976.
3. Ando, J. SOX and NOX Remo_val _Technology in Japan-1976. Environmental
Technical Information Center> Japan Management Association, Tokyo,
Japan, February 1976.
4. Ando, J., R. D. Stern, and J. D. Mobley. "Status of Flue Gas Treatment
Technology for Control of NOX and Simultaneous Control of SOX and NOX
in the United States and Japan." Paper presented at the 69th Annual
Meeting of the American Institute of Chemical Engineers, Chicago,
Illinois, November 28-Deeember 2, 1976.
5. Ando, J., H. Tohata, and G. Isaacs. NO^ Abatement for Stationary
Sources inJapan. EPA-600/2-76-0136 (NTIS PB 250 586), January 1976.
6. Anonymous. "A Way to Lower NOX in Utility Boilers." Environ.Sci.
Technol., 2, No. 3, 226-228, March 1977.
7. Anonymous. "Export of NOX Removal Technology." Chem. Econ. Eng. Rev. ,
_8, No. 11, 49, November 1976.
8. Anonymous. "Exxon is Given Award for Emissions Advance." Chem. Mark.
Rep., 210, No. 14, 12, October 4, 1976.
9. Anonymous. "50,000 Nm^/hr Flue Gas Denitrification Plant." Chem. Econ.
Eng. Rev., _8, No. 3, 39-40, March 1976.
10. Anonymous. "Hitachi Zosen's 'DeNOx Plant' - A Solution to Contamination
Problems." Hitachi Zosen News, 19, No. 1, 10-12, Spring 1976.
11, Anonymous. "IHI Bares Flue Gas SOX, NOX Removal Plant in Test Run."
Jpn. Chem. Week, 17, No. 862, 5, November 4, 1976.
12. Anonymous. "MHI Develops Simultaneous NOX, SOX Removal Method." Jpn.
Chem. Week, JJ, No. 831, 7, May 13, 1976.
13. Anonymous. "Nippon Kokan Iron Ore Catalyst Denitrates Flue Gas." Jpn.
Chem. Week, 17, No. 862, 2, November 4, 1976.
379
-------
14, Anonymous. "NOX Removal for Sintering Furnace." Chem. Econ. Eng. Rev.,
_8, No. 10, 51, October 1976.
15. Anonymous. "Regenerative S02 Removal Process." Environ. Scl. Technol.,
jj., No. 1, 22-23, January 1977.
16. Aokl, M. (Mitsui Engineering and Shipbuilding Company, Ltd. , Tokyo,
Japan). Private communication, March 1977,
17. Araki, T. (Sumitomo Metal America, Inc., New York, New York). Private
communication, April 1977.
18. Armento, W. J. Effects of Design and Operating Variables on NOX from
Coal-Fired Boilers -Phase I. EPA-650/2-74-OQ2a (NTIS PB 229 986),
January 1974.
19. Atkinson, E. S. (Hooker Chemicals and Plastics Corporation, Niagara
Falls, New York). Private communication, May 1977.
20. Atsukawa, M., et al. "Development of NOX Removal Processes with
Catalyst for Stationary Combustion Facilities" (11 pp). Mitsubishi
Heavy Industries, Tokyo, Japan, 1976 (in Japanese).
21. Atsukawa, M., Y. Nlshlmoto, and N. Takahashi. "Study on the Removal
of Nitrogen Oxides from Effluent Stack Gases." Tech. Rev., 53-59,
May 1968.
22. Atsukawa, M. and N. Shinoda. "Status of MHI-Lime/Limestone-Gypsum
Desulfurlzation Process in Japan." Paper presented at American Chemical
Society Centennial Meeting Symposium, New York, New York, April 4-9, 1976,
23. Atsumi, T. (Sumitomo Heavy Industries, Ltd., Tokyo, Japan). Private
communication, April 1977.
24. Axeworthy, A. E., G. R. Schneider, and V. H. Dayan. Chemistry of Fuel
Nitrogen Conversion toNitrogenOxides inCombustion. EPA 600/2-76-039
(NTIS PB 250 373), February 1976.
25. Bartok, W., A. R. Crawford, and A. Skopp. "Control of NOX Emissions
from Stationary Sources." Chem.Eng. Prog., 67, No. 2, 64, February
1971.
26. Bauerle, G. L., S. C. Wu, and K. Nobe. "Catalytic Reduction of Nitric
Oxide with Ammonia on Vanadium Oxide and Iron - Chromium Oxide." Ind.
Eng. Ghem. Prod. Res. Dev. . _14_, No. 4, 268-273, 1975.
27. Bischoff, W. F. "FW-BF Dry Adsorption System for Flue Gas Clean Up."
In Proceedings: Flue Gas Desulfurization Symposium - 1973, pp 1081-1100,
New Orleans, Louisiana, May 14-17, 1973".EPA-650/2-73-038• (NTIS PB 230'
901), December 1973.
380
A
-------
28. BIschoff, W. F. (Foster Wheeler Energy Corporation, Livingston, New
Jersey), Private communication, March 1977,
29, Bischoff, W, F, and P. Steiner. "Toward a Cleaner Environment...The
Currently Available Options.,.Dry Adsorption." Heat Eng., pp. 150-152,
July-September 1972.
30. Bischoff, W. F., J. J. Strum, and R. E. Rush. "BF Dry Adsorption System:
Part 1, FW-BF Gulf Power Demonstration Unit Interim Results." In Pro-
ceedings : SymposiumonFlue Gas Desulfurization, New Orleans, Louisiana,
March 1976, Vol. II, pp. 877-898. EPA-60G/2-76-136a (NTIS PB 255 317),
May 1976.
31, Brocoff, J. C. (The Ralph M, Parsons Company, Pasadena, California).
Private communication, April 1977.
32. Brocoff, J. C. (The Ralph M. Parsons Company). Process for the Treat-
ment of Sulfur and Nitrogen Oxides Formed During Power Generation, U.S.
Patent 4,012,488, March 15, 1977.
33. Carberry, J. J. "Some Remarks on Chemical Equilibrium and Kinetics
in the Nitrogen Oxides - Water System." Chem. Eng. Sci. , 9_, No. 4,
189-194, 1959.
34. Chakrabarti, G. and C. Chu. "Reduction of Nitric Oxide with Ammonia on
Copper Chroinite and Nickel - Copper Oxide Catalysts." Atmos. Environ.,
j6, 297-307, 1972.
35. Chlyoda Chemical Engineering and Construction Company, Ltd. The
Chiyoda Thoroughbred 102 Simultaneous S02 and N0y Removal Process (15 pp),
Yokohama, Japan, June 1974.
36. Chiyoda Chemical Engineering and Construction'Company, Ltd. The
Chiyoda Thoroughbred 102 Simultaneous S02 and NOX Removal Process (11 pp),
Tokyo, Japan, January 31, 1977.
37. Crowe, J. L. jtudy pf_SludgeDisposal from Sulfur Dioxide and Partic-
ulate Removal Processes. Tennessee Valley Authority, PRS-1, May 1974.
38, Driscoll, J, N. Flue Gas MonitoringTechniques and Manual Determination
of Gaseous Pollutants, pp 219-263. Ann Arbor Science Publishers, Ann
Arbor, Michigan, 1974.
39. Environmental Protection Agency. Monitoringand Air Quality Trends
Report,1974. EPA-450/1-76-001 (NTIS PB 252 269), February 1976.
40. Environmental Protection Agency. "Standards of Performance for New
Stationary Sources." Fed. Regist., 36, No. 247, p. 24,879, December 23,
1971.
381
-------
41. Faucett, H. L. Japan Trip Report (unpublished). Tennessee Valley
Authority, Muscle Shoals, Alabama, April 1977.
42. Friedrich, J. L. and R. H. Pai. "Nitrogen Oxide Reduction." In The
Proceedings of the NOy Control Technology Seminar, pp 9-14. EPRI SR-39,
San Francisco, February 5-6, 1976.
43. Fukita, S. (Kobe Steel, Ltd., New York, New York). Private communica-
tion, March 1977.
44. Fukui, S., et al. (Chiyoda Chemical Engineering and Construction
Company). Method for Simultaneous Removal of SOX and NOX, U.S. Patent
4,011,298, March 8, 1977.
45. Fukui, T. S. (Hitachi, Ltd., New York, New York). Private communication,
October 1976.
46. Fukunaga, F. (Unitlka America Corporation, New York, New York). Private
communication, April 1977.
47. Habelt, W. W. and B. M. Howell. "Control of NOX Formation in Tangen-
ttally Coal-Fired Steam Generators." In The ProceedingsoftheN0y
Control Technology Semin_ar, p. 26. EPRI SR-39, San Francisco, February
5-6, 1976.
48, Hara, Y. (Takeda Chemical Industries, Ltd., Tokyo, Japan). Private
communication, April 1977.
49. Harrison, J. W. Technology and Economics ofFlue Gas NOXOxidation by
Ozone (66 pp). Research Triangle Institute, Research Triangle Park,
North Carolina 27709, November 1976.
50. Hitachi, Ltd. HitachiDry-Type NOX Removal System Utilizing Selective
Catalytic Reduction (13 pp), Tokyo Japan, July 1975.
51. Hitachi, Ltd. Hitachi Stack Gas NOX Removal Plant (8 pp), Tokyo, Japan.
52. Hitachi Zosen. Flue Gas Denitrification Plant (8 pp), Osakl, Japan.
53. Hoover, W. (Exxon Research and Engineering Company, Linden, New Jersey).
Private communication, February 1977.
54. Hosoda, K. (Mitsubishi Petrochemical Company, Ltd., New York, New York).
Private communication, April 1977.
55. Hradesky, E. L., A. W. Leavitt, and E. Shapiro. Initial Field Test,
Systems Engineering and Econometic AnalysisofanAndyanced FlueGas
DesulfurizationProcess: SULF-X, Final Report. Prepared for the
Appalachian Regional Commission, December 1976.
382
A
-------
56. Ibaraki, H. (Sumitomo Metal Industries, Ltd., Osaka, Japan). Private
communication, April 1977.
57. Idemura, H. "Simultaneous S02 and NOX Removal Process for Flue Gas."
Ch em. EC on. Eng. Rev., _6, No. 8, 22-26, August 1974.
58. Ikawa, K. (Hitachi Zosen International, S.A., New York, New York).
Private communication, March 1977.
59. Inaba, N. (Hitachi Zosen, Tokyo, Japan). Private communication,
October 1976.
60. Ishikawajima-Harima Heavy Industries Company, Ltd. Outline of Wet
Type Simultaneous SOX-NOX Removal Pilot Plant - Ozon^l^idizing - Wet
Lime and Gypsum Process (5 pp), Tokyo, Japan.
61. Iwasaki, K. (Kureha Chemical Industry Company, Ltd., Tokyo, Japan).
Private communication, April 1977.
62. JGC Corporation. All About JGC - Air Pollution Control Techniques '
(10 pp), Tokyo, Japan.
63. JGC Corporation. Information from JGC (7 pp), 19, Tokyo, Japan,
February 1977.
64. JGC Corporation. Technical Information on_Paranox Process (13 pp),
Tokyo , Japan.
65. Kawasaki Heavy Industries, Ltd. Flue Gas_ DesulfurizationL and rjeNOx
Process Using Magnesium Compound (19 pp), Tokyo, Japan.
66. Kawasaki Heavy Industries, Ltd.. Technical Informationon theKawasaki
Flue Gas Desulfurization Process (28 pp) , Tokyo, Japan, October 1976.
67. Kline, J. M., P. H. Owens, and Y. C. Lee. Catalytic Reduction of
NitrogenOxideswithAmmonia; Utility Pilot-Plant Operation. EPA-
600/7-76-031 (NTIS PB 261 265), October 1976.
68. Kobayashl, H. and K.. Muraki (Kurabo Industries, Ltd., Osaka, Japan).
Private communication, October 1976.
69. Kobe Steel, Ltd. Characteristics of KSL H-Type NOX Removal Catalyst
(9 pp), Kobe, Japan.
70. Kobe Steel, Ltd. NOX Removal Facility Using KSL Moving Bed (7 pp),
Kobe, Japan. ^
c*
71. Koppang, R. R. (TRW, Redondo Beach, California). Private communication,
March 1977.
383
-------
72, Koutsoukas, E. P., et al. Assessment of Catalysts for---Control of NOX
from Stationary Power Plants, Phase I, Volume 1. EPA-650/2-75-001b
(NTIS PB 239 745), January 1975. /
73. Kurabo Industries, Ltd. Knorca NOxRemoval System for FlueGas (5 pp),
Osaka, Japan.
74. Kureha Chemical Industry Company, Ltd. Kureha KPN Process - DryProcess
. for MOX Removal (12 pp), Tokyo, Japan, March 1977.
75. Kureha Chemical Industry Company, Ltd. Kureha K.DSNProcess -_ Simultaneous
Removal of NOX and SO? (33 pp) , Tokyo, Japan, March1977.
76. Leavitt, A. W. (Pittsburgh Environmental and Energy Systems, Inc.,
Pittsburgh, Pennsylvania). Private communication, April 1977.
77. Lee, Y. C. (Eneron Corporation, Newport Beach, California). Private
communication, March 1977.
78. Lewis, W. H. Nitrogen Oxides Removal, p. 166. Noyes Data Corporation,
Park Ridge, New Jersey, 1975.
79. Lyon, R. K. (Exxon Research and Engineering Company, Linden, New
Jersey). Private communication, March 1977.
80. Kyon, R, K. (Exxon Research and Engineering Company). Method for
the Reduction of the Concentration of NO in Combustion Effluents Using
Ammonia. U.S. Patent 3,900,554, August 19, 1975.
81. Lyon, R. K. "The NH3-NQ-02 Reaction." Int. J. Chem.Kinet. , _§_, 315-
318, 1976.
82. Lyon, R. K. and J. P. Longwell. "Selective Noncatalytic Reduction of
NOX by NH3." In The Proceedings ofthe NOX Control Technology Seminar,
1-20. EPRI SR-39, San Francisco, California, February 5-6, 1976.
83. Machl, S., et al. (Ebara Manufacturing Company and Japan Atomic Energy
Research Institute). Process for Removing Nitrogen Oxides and Sulfur
Dioxide from Effluent Gases. U.S. Patent 4,004,995, January 25, 1977.
84. Markvart, M. and V. Pour. "Selective Reduction of Nitrogen Oxides with
Ammonia." Chcm. Prum., _1_9_, No. 1, 8-12, 1969.
85. Mitsubishi Petrochemical Company, Ltd. Outline of theCharacteristics
of NOX Removal Catalyst 'PNX' (37 pp), 1976.
86. Mitsui Engineering and Shipbuilding Company, Ltd, NOX Removal Process -
Selective Catalytic ^Reduction Using Ammonia (19 pp) , Tokyo, Japan.
87. Mitsui Toatsu Chemicals, Inc. Technical Information on Mitsui Toatsu
(MTC) Process for NOK Removal (16 pp), Tokyo, Japan, August 1976.
384
A
-------
88. Kobe, K., G. L. Bauerle, and S. C. Wu. Parametric Studies p_f_Catalysts
for NOX Control from Stationary Power Plants. EPA-600/7-76-026 (NTIS
PB 261 289), October 1976.
• 89. North American Manufacturing Company. North American Combustion Hand-
book (10 pp), Cleveland, Ohio, 1952.
90. Oka, H., E. Ichiki, and T. Shiraishi. "Process Removes NOX Efficiently."
Hydrocarbon Process., 53, No, 10, 113-114, October 1974.
91. Okada, T. (Sumitomo Chemical America, Inc., New York, New York).
Private communication, May 1977,.
92. Pohlenz, J. B. (Universal Oil Products, Des Plaines, Illinois). Pri-
vate communication, May 1977.
93. Pohlenz, J. B. "The Shell Flue Gas Desulfurization Process," Paper
presented at the Third Annual International Conference on Coal Gasifi-
cation and Liquefaction, Pittsburgh, Pennsylvania, August 3-5, 1976.
94. Polek, J. R. (Catalytic, Inc., Philadelphia, Pennsylvania). Private
communication, April 1977.
95. Ricci, L. J. "EPA Sets Its Sights on Nixing GPS *s NOX Emissions."
Chem. Eng., 84, No. 4, 32-36, February 14, 1977.
96. Ricci, L. J. "Nixing NOX Emissions." Chem. Eng., 84., No. 8, 84-89,
April 11, 1977.
97. Sawai, K. and T. Gorai. "Simultaneous Desulfurization and Denitrifica-
tion of Waste Flue Gas by CEC Process." Sulfuric Acid Ind. , 29, No. 9,
1-5, 1976 (in Japanese),
98. Shima, T. Knorca Exhaust Gas N0y Removal Process (20 pp). Kurashiki
Basek Company, Ltd., (in Japanese).
99. Siddigi, A. A., J. W. Tenini, and L. D. Killion. "Control NOX Emissions
from Fixed Fireboxes." Hydrocarbon Process., _55_, No. 10, 94-97,
October 1976.
100. Stahl, S. (Exxon Research and Engineering Company, Linden, New Jersey).
Private communication, March 1977.
101. Sumitomo Chemical Company, Ltd. Sumitomo's Denitration Process (20 pp),
Osaka, Japan, February 10, 1975.
102. Sumitomo Heavy Industries, Ltd. jPrpcess for Treatment of Combustion
Flue Gases for Removal of N0y (18 pp), Tokyo, Japan, May 1976.
103. Sumitomo Heavy Industries, Ltd. Sumitomo's NOx/SOg Removal System,
(18 pp), Tokyo, Japan, April 1977.
385
-------
104. Sumitomo Shipbuilding and Machinery Company, Ltd. Economic Evaluation.
Dataon Sumitomo Shipbuilding and Machinery Company's Flue Gas j3esulfur-
ization Process Using Activated Carbon_ as Absorb_ent (3 pp) , Tokyo, Japan,
September .1973.
105. Taniguchi, M., et al. (Ebara Manufacturing Company, Ltd.)- Process
for Utilizing Collected Material Removed from Effluent Gas, U.S. Patent
3,981,815, September 21, 1976.
106. Tennessee Valley Authority. Pilot-Plant Studyof an Ammonia Absorption-
Ammonium Bisulfate Regeneration Process, Topical ReportPhases Iand II.
EPA-65Q/2-74-Q49a, (NTIS PB 237 171), June 1974.
107. Terasaki, I. (Asahi Chemical Industry Company, Ltd., Tokyo, Japan).
Private communication, April 1977.
108. Teixeira, D. P. "Status of Utility Application of Homogeneous NOX
Reduction. " In The Pro ceedings j>J_ J:_he NOx. Con j:ro1_ TechnoIpgy geminar,
p. 3. EPRI SR-39, San Francisco, California, February 5-6, 1976.
109. Ube Industries, Ltd. Catalytic Removal of Nitrogen Oxide with JP3
from Exhaust Gasesof Industrial Processes and StationaryFacilities
(5 pp), Tokyo, Japan.
110. Uchiyama, H., et al. (Fuji Kasui Engineering Company, Ltd., and
Sumitomo Metal Industries, Ltd.). Process for Removing Sulfur Oxides
from Gases. U.S. Patent 3,892,837, July I, 1975.
111. Uede, T. (Sumitomo Heavy Industries, Ltd., New York, New York). Pri-
vate communication, May 1977.
112. Unitika, Ltd. "Simultaneous NOX-SOX Removal Equipment." Jpn. Text.
News, pp. 84-87, February 1976.
113. Ushio, S. "Japan's NOX Cleanup Routes." Cheni.^ Eng. , 82, No. 15,
70-71, July 21, 1975.
114. Weltz, A. B. (Exxon Research and Engineering Company, Linden, New
Jersey). Removal of Nitrogen Oxides and Sulfur Oxides from Gases.
British Patent 1,438,119, June 3, 1976.
115. Weltz, A. B. (Exxon Research and Engineering Company, Linden, New
Jersey). Removal of Nitrogen Oxides and Sulfur Oxides from Gases.
British Patent 1,438,119, June 3, 1976.
116. Yatnada, S., T. Watanabe, and H. Uchiyama. Bench-Scale Tests_on
Simultaneous_Removalof SOy and NQ^ by Wet Lime^ and Gypsum Process
(17 pp). Ishikawajima-Harima Heavy Industries, Tokyo, Japan.
117. Yoshimaga, M, (Ishikawajima-Harima Heavy Industries Company, Ltd.,
New York, New York). Private communication, April 1977.
118. Zeldovich, J. "The Oxidation of Nitrogen in Combustion and Explosions."
Acta Physicochim URSS (Moscow), 21, p. 577, 1946.
386
A
-------
APPENDIX A
GAS-PHASE OXIDATION OF NO
387
-------
APPENDIX A
OXIDATION OF NO
The easiest and cheapest method of converting insoluble NO to N02 would
be to react the NO with the excess 62 already present in the flue gas.
Although the oxidation of NO by free 02 In the flue gas is theoretically
possible, in actual boiler operations there is insufficient reaction time
during the passage from the boiler to the stack. For the 95-100% conversion
of the NO in the flue gas by free 02 as required in the oxidation processes,
the residence time necessary in the system would be on the order of 1 hr.
Thus, for practical operations a method to promote the oxidation rate is
required.
The primary method of increasing the oxidation rate of NO currently under-
going development is through the use of gas-phase oxidant. These gas—phase
oxidants are capable of rapidly and selectively oxidizing the NO with reaction
times on the order of half a second. The two gas-phase oxidants currently
being considered are Oj and C102- The use of 03 results in the following
reactions:
N°(8) + °3(8) ^°2(g) + °2(g) (1>
2NO(8)+303
With good distribution of the 03 into the flue gas duct, reaction (2) can be
minimized and better 03 utilization can be achieved. 0^ is favored by many
process developers since its use generates no undesirable byproducts in the
flue gas and the equipment required for the generation of 03 is relatively
simple and easy to control. The primary disadvantage of using 03 as the gas-
phase oxidant is the expense Involved in generating the 03.
C102 also rapidly oxidizes NO at a reaction rate similar to that of 03.
The use of C102 results in the following gas-phase reaction:
2NO. + CLO_, . + H_0 -»• NO,, v + HNO,, . + HCl, N (3)
(g) 2(g) 2 2(g) 3(g) (g)
The use of C102 is favored by some since, as compared to ©3, it is cheaper to
produce and also because 1 mol of C102 oxidizes 2 mols of NO. In addition,
C102 converts 1 mol of NO into HNOg which is much more soluble than N0£. The
disadvantages associated with the use of C102 include the high cost of C102?
the formation of undesirable reaction products such as HCl and HN03 in flue gas,
and the extensive chemical processing system involved in generating C102- This
generation of C102 by chemical reaction also results in the slow response to
changing levels of NOX in the flue gas.
388
•j
-------
Brief introductions to the technical and economic considerations for
both 03 and C102 generating systems follow,
03 GENERATION
Since most of the oxidation-absorption-reduction processes under current
development favor Og as the gas-phase oxidant to convert relatively insoluble
NO to M>2, the following discussion is included to briefly describe 03 gener-
ation technology and costs. Other sources of information on 03 are available
for a more detailed explanation of the technology. (J. W. Harrison, Technology
and Economics of FlueGas NOX Oxidationby Ozone Research Triangle Institute,
Research Triangle Park, North Carolina 27709, November 1976.)
03 can be generated by any of five relatively exotic methods including
electrolysis of H20, photochemical reaction, radioehetnical reaction, thermal
reaction, and corona discharge in either air or 02- The least expensive and
the most common method for generating 03 on the scale required for commercial-
sized FGT systems would be corona discharge in air or 02- In this method
either air or 02 is fed to a bank of Oj generators (see Figure A-l for a
diagram of a typical Oj generator) where, under a pressure of 1.2-2.0 atmosphere,
an electrical discharge is passed through the feed gas. If the feed gas is air,
the resulting exit gas stream will contain approximately 1% (by wt) 03 and, if
Q£ Is used, the exit stream will contain 1.5-2.0% (by wt) 03. The yield of 03
per unit of energy also depends on the feed gas: for air feed approximately
0.05-0.06 kg (0.11-0.13 lb) of 03 are produced per kWh, while for 02 feed about
0.11-0.13 kg (0.24-0.29 lb) of 03 are produced per kWh of electricity. Although
at first glance it would appear that generating 03 from 02 would be more eco-
nomical; if the cost of the 02 plant is included, the air-feed system is more
economical. This is readily apparent from the costs given in Table A-l. Thus
for a coal-fired boiler generating flue gas containing 600 ppm NOX, the air-
feed 03 generator system itself adds $53/kW to the capital cost of the plant
and 4.7 mills/kWh to the annual revenue requirements for the S02—NQx removal
system, (Cost basis: assumed U.S. and 1976 dollars). The use of 02 as the
feed stream would approximately triple both the capital cost and annual
revenue requirements for the 03 generating system.
In addition Table A-l also Includes a comparison of the 03 generating
costs for treating flue gas from a coal-fired boiler containing 600 ppm NOX
and for treating flue gas from an oil-fired boiler containing 200 ppm NOX. In
Japan where most of these processes were developed, most of the boilers are
fired with oil and hence capital costs and annual revenue requirements are
only one-third those required for treating coal-fired flue gas.
The 03 generating system using air feed consists of two stages as shown
in Figure A-2: a preparation stage including filters, coolers, and driers to
remove the particulates and H20 from the air and the actual 03 generation
section. The air must be dried to a dew point of at least -60°C (-77°F) since
the presence of H^O in the air stream can result in both sparking and also the
decomposition of the 03. If H20 Is present in the air stream it can also react
with the small amount of N205 produced in the 03 generator to form HNO^. If
the more expensive 02~fed system is used, however, these problems are elimin-
ated since the 02 plant produced a clean, dry 02 stream.
389
-------
TRANSRWliER
AIR OR
OXYGEN
IN
LEGEND;
A - COOLING WATER
8 - STAINLESS STEEL TUiE
C * DISCHARGE GAP
D - SLASS TUBE
Figure A-l. Welsbach Ozonator (49)
390
A
-------
COMPRESSOR
AIR
n^
_
> r
—1
1
I T
,„„.,;, O m 4- f
AIR 3
TO
H"""1 1TT~I 07nPJATOR ™>-^c
ILTER COOLERS DRIERS
OZONE FROM AIR
OXYSEN
PLANT
02 ^
OZONATOR
03 4- 02 TO
* PROCESS
OZONE FROM OXYGEN
Figure A-2. Ozone Generation Systems (49)
391
-------
TABLE A-l. SUMMARY OF CAPITAL INVESTMENT, REVENUE REQUIREMENTS, AND
ENERGY REQUIREMENTS FOR THE OXIDATION OF FLUE GAS NO BY
03a IN 500-MW OIL-FIRED AND COAL-FIRED PLANTSb
Oil fired (200 ppm NO) Coal fired (600 ppm NO)
Item Air feed 02 feed Air feed Q£ feed
Capital investment,
$/kWC 17.60 56,20 52.60 152.40
Revenue requirements,
mills/kWhc 2.0 3.8 4.7 9.1
Energy requirements,
% of boiler capacity 3.1 3.6 9.4 10.7
a. Based on atoiehioraetrle 0., requirement.
b, J. W. Harrison. Technology and Economic^
Ozone, Research Triangle Institute, Research Triangle Park, North
Carolina 27709, November 1976.
c. Assumed U.S. location and 1976 dollars.
For a 500—MW coal-fired power plant generating flue gas containing 600
ppm NOX, the requirement for 90% NOX removal will result in an estimated
electrical consumption of 47,500 kW or approximately 9.4% of the generating
capacity of the boiler if the 03 is generated by corona discharge in air.
This electrical consumption is for the 03 generating system only; the energy
requirements for the other process equipment in the simultaneous S02~NOX
removal system would be in addition to this 47,500 kW. This, of course,
represents a substantial loss of output for the boiler and hence higher costs
for the remaining electrical energy available for sale.
The 63 generating system, in addition to its high electrical consumption,
also requires significant quantities of cooling H2<3 and process air. Since
about 90% of the electrical energy consumed to produce the 03 is rejected as
waste heat, the cooling water requirement for the 500-MW boiler system is
substantial (approx 550 gpm). Failure to adequately remove this waste heat
increases the rate of 03 decomposition and hence would increase the already
high energy cost for each unit of 03. The air feed rate to the 03 generators
for this 500-MW coal-fired boiler would be approximately 100,000 sft3/ntln,
Not only will this high air feed rate require large and expensive pieces of
process-handling equipment but the injection of this much air into the flue
gas ducts represents a 6.5% increase in the amount of gas to be scrubbed and
of course reheated. The adverse ramifications of this increase in flow rate
on the process economics have not been studied as yet.
One other major technical consideration concerning the 03 generation
section is the relatively small capacities of the 03 generators currently
available. The largest commercial 03 generators available at the present
time have a rated capacity of only 560 kg (1232 Ib) of Og/day. Since a
392
A
-------
500-MW coal-fired boiler based on the premises of this study (600 ppm NOX in
the flue gas and 90% NOX removal) would require approximately 49,000 kg
(107,700 Ib) of 03/day, each 500-MW boiler would require a battery of at least
88 individual 03 units rated at 560 kg of 03/day. Assuming a 10% excess
capacity factor, a minimum of about 100 03 generators would be required in the
battery for a 500-MW coal-fired boiler.
C102 GENERATION
Although most of the oxidation-absorption-reduction processes use 03 as
the gas-phase oxidant, the processes offered by Sumitomo Metals (Moretana
Calcium and Sodium processes) opt to use C102 to oxidize NO to N02. Even
though the use of C102 results in the introduction of high levels of Cl~ in
the circulating solution and increased amounts of effluent wastewater, it is
much cheaper to use since each mol of C102 oxidizes twice as many mols of NO
as each mol of 03 and also because the cost/mol of oxidant, although still
expensive, is less for C102«
Since C102 is a commonly used industrial chemical, particularly in the
paper industry, numerous methods of producing it have been developed over the
years. However, most of these processes developed by the paper industry
result in the production of Na2S04 which can be reused in the pulping process
but which cannot be used in this S02~NQX removal process. The Resting process
generates byproduct NaCl which can be used to regenerate NaC103, but un-
fortunately this process also generates relatively large amounts of undesir-
able Cl2 gas. Thus, Sumitomo Metals and DAISO Engineering developed a
modified version of the Resting (HC1) process for the simultaneous S02~NOX
removal system in which an additive is used to minimize the formation of free
•Cl2 in the outlet gas. NaCl, HC1, and the proprietary additive are reacted
to generate a weak (3% by vol) C102 gas stream and a recycle Na salt solution.
This salt solution is recycled and undergoes electrolysis to regenerate the
required NaClOg to complete the cycle. The only raw materials required appear
to be NaOH, HC1, and the additive.
Under the premises of this study (90% NOX removal and 600 ppm NOX) a 500-
MW coal-fired power plant would consume approximately 46-51 tons/day of C102
depending on the ratio of C102:NOX used (46 tons/day corresponds to a C102:
NOX = 0.55 and 51 tons/day to C102:NOX = 0.60). The C102 required for an oil-
fired boiler flue gas containing 200 ppm NOX would be proportionately less,
i.e., 15-17 tons/day depending on the C10*2:NOX ratio. Although the capital
investment and the annual revenue requirements for the Sumitomo-DAISO C102
generating system have not been released, a similar capacity (50 tons/day)
C102 generating system using the Hooker R-2 method has been'estimated ( E. S.
Atkinson, Hooker Chemicals & Plastics Corporation, Niagara Falls, N. Y.,
Private communication, May 1977) as $2.0M for the capital investment ($4/kW)
and an estimated $8.8M for the annual revenue requirements (2.51 mills/kWh),
(Cost basis: assumed U.S. and 1976 dollars).
One of the main drawbacks associated with the use of C102 as a gas-phase
oxidant is the very slow response of the system to changes in the NOX con-
centration of the inlet flue gas. Since the C102 is generated by chemical
393
-------
reaction and cannot be stored, an abrupt increase in the NOX concentration in
the inlet flue gas cannot be treated effectively for a period of time as the
C102 generator gradually increases production. Although the average NOX
removal efficiency can be high, whenever the inlet NOX concentration increases
sharply, the removal efficiency can drop and remain at lower levels during the
time required for the C102 generating system to increase production.
Although the amount of inert material injected with the C102 results in
a large gas stream (13,600 aft-Vmin at 127°F), it is only one-tenth the size
of the gas stream used for the 63 injection system and represents an increase
of <1% in the total amount of gas to be scrubbed and consequently reheated.
394
A
-------
APPENDIX B
SOLUBILITY OF VARIOUS GASES IN
395
-------
TABLE B-l. SOLUBILITY OF VARIOUS GASES IN H20a
Gas
N2
°2
NO
N02
C02
H2S
S02
Solubility (in g/£)
0.010
0.044
0.063
1.26
1.74
3.98
117.0
a. J. N. Driscoll, Flue Gas Monitoring Techniques
andManualDeterminationofGaseous Pollutants
Ann Arbor Science Publishers, Ann Arbor,
Michigan, 1974, pp. 219-263-
396
GPO 929.454
------- |