Tyco Laboratories, Inc.
Bear Hill
Waltham, Massachusetts 02154
DEVELOPMENT OF THE CATALYTIC
CHAMBER PROCESS
FOR THE
MANUFACTURE OF SULFURIC AND NITRIC
ACIDS FROM WASTE FLUE GASES
by
B. Keilin
A. L. Walitt
Final Report, Contract No. PH 86-68-75
Covering period
29 October 1967 to 30 September 1969
Prepared for
Environmental Protection Agency
Office of Air Programs
Research Triangle Park, North Carolina 27711
-------
DEVELOPMENT OF THE CATALYTIC
CHAMBER PROCESS
FOR THE
MANUFACTURE OF SULFURIC AND NITRIC
ACIDS FROM WASTE FLUE GASES
by
B. Keilin
A. L. Walitt
Final Report, Contract No. PH 86-68-75
Covering period
29 October 1967 to 30 September 1969
Prepared for
Environmental Protection Agency
Office of Air Programs
Research Triangle Park, North Carolina 27711
-------
1.
II.
III.
IV.
V.
Table of Contents
SUMMARY. . . . .
. . . . . . . . . . . . . . . . . .
RECOMMENDA TIONS. .
. . . .
. . . . . . . .
. . . .
A.
B.
Catalyst Evaluation.
Pilot Plant Conversion
and Operation
. . . .
. . . . .
. . . . . . .
. . . .
. . . . . . . . . . . .
C. Demonstration Plant Design
. . . . . . . . . . . . .
INTRODUCTION. .
..................
CURRENT PROCESS
. . . . . . .
. . . . . . . . . . .
A.
B.
General. . . . . .
. . . . . . . . . .
. . . . . .
The Catalytic Chamber Process. . . .
Flue Gas Condition. . .
. . . . . . . .
C.
D.
. . . .
. . . . . . . . . .
Oxidation Reaction Vessel.
. . . . .
. . . . . . . .
E. High Temperature Scrubber. .
. . . . . . . . . . . .
F.
G.
Catalytic Stripper. .
Nitric Acid Absorber
................
. . . . .
. . . . . . . . . . .
H. Process Economics. . . . . . . .
. . . . . . . . .
I.
Effect on the Sulfuric Acid. . . . . . . . .
Industry . . . . .
THE EVOLUTION OF THE CA TAL YTIC. . .
CHAMBER PROCESS: CONCEPTUALIZATION
AND EXPERIMENTATION
. . . . . . .
A.
B.
Genera 1 . . . . . . . . . . . . . . . . .
. . . . .
The Originally Proposed Concept. . . . . . . . . . .
ii
Page No.
1
3
3
3
3
5
7
7
7
9
13
14
18
18
20
27
31
3.1
31
-------
Table of Contents (Cont.)
C.
D.
Feasibility of Original Concept.
Batch Kinetic Study. , . . .
.......
. . . .
.......
. .. .. . ..
E. Nitric Oxide Oxidation. . .
. .. . ..
. , .. . . . .. .. .
F.
Collection of Nitrogen Oxides
and Acid Mist
. . .. . .. .
.. . .. . .. .
G. Solubility of No, N02 and Fly
Ash in H2S04
H. The Modified Chamber Process
.........~..
.. . . .. ..
. .. . . .. ..
I.
Laboratory Evaluation of the
Modified Chamber Process
........
.. . . .
J.
The 10 SCFM Pilot Plant:. . , . . .
Design, Construction, and Operation
.........
K. Process Improvements
...............
APPENDIX 1 . . . .
APPENDIX 2. . . .
. . . .. . .. .. .. . .
""""".
.....................
APPENDIX 3 . .
.. .. .. .. "
""""""
.. .. .. .. " .
VI.
REFERENCES. .
8..............
. . .. .
Page No.
33
35
43
49
55
59
75
80
III
147
153
167
173
-------
List of Illustrations
Figure No.
7.
8.
9.
10.
11.
12.
13.
14.
15.
1.
Catalytic Chamber Process
. . . .
. . . . . . . . .
2.
3.
High Temperature Scrubber. .
. . . . .
. . . .
Vapor Pressure of Water Over
80% H2S04
. . . . . . .
. . . .
. . . .
4.
5.
Catalytic Stripper
. . . .
. . . . .
. . . . .
The Lead Chamber Process
Flow Diagram
. . . .
. . . . .
. . . .
6.
Originally Proposed NOfOxidation
Process for Recovery of S02 from
Power Plant Stack Gas
.............
Apparatus for Measuring Rate
of Oxidation of S02 by N02,
and of NO by 02
Oxidation of NO by 02 - 420 mt.'
Effect of N02 on the Oxidation
Rate of NO
...............
. . . . .
. . . . . .
. . . . . . . .
. . . .
Effect of Temperature on the
NO Oxidation Rate
...........,...
Oxidation Rate of NO in the Vapor. . . . . . . . . . . . .
Phase Catalytic Oxidation of S02
Equilibrium Constant in the Reaction. . . . . . . . . . . .
2 NO + 02 ~ 2 N02 vs. Temperature
Conversion Rate of NO to N02 as . .
a Function of Temperature
. . . . . . . . . . . .
Apparatus for Continuous Oxidation. . . . . . . . . . . . .
of NO to N02
Rate of Oxidation of NO to N02' . . . . . . . . . . . . . .
U sing Charcoal Catalysis
iv
Page No.
8
15
17
19
29
32
36
38
40
41
42
45
46
47
48
-------
List of Illustrations (Cont.)
Figure No.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Oxidation Rate of NO at High NO
Concentrations Without Catalyst
Oxidation Rate of NO at High NO . . . . . . . . . . . . . .
Concentrations Without Catalyst, 60 0 C
. . . . . . . . . . . . . III
Mist Collection Study Apparatus
Flow Diagram
Collection Efficiency for N 203 and
Pressure Drop Across Fiber Mist
Collector Unit
.....,..",...
..e-,.,......
Solubility of NO in 74~ H2S04
Containing N02, 40 0 c
Solubility of Nitrosyl Sulfuric.
Acid in H2S04
......,..,.....
,..............
Modified Chamber Process
"""'"
. . , . . . .
Vapor Pressure of N20g and H20 over 80%. . . . . . . . . .
H2S04 Solutions
Nitrose Vapor Pressures Over 80% wt 82S0 4 Solutions. . . . .
S02 and N02 Recovery Efficiency. . . . . . . . . . . . . .
in a Packed Bed Scrubber
N02 Removal Efficiency as a Function
of NO Concentration and Contact Time
. . . . . . . . . . .
F low Sheet for 10 SCFM Pilot Plant. .
Gas Supply System - Pilot Plant.
. . . .
.......
. . . . ,
. . ., . .
. . . .
Pilot Plant Reactor
............
"~"III"
Flooding Velocities in Packed Tower
Scrubber
. . . . . " .
. . . . .
. . . . . .
.......
....,......
Acid Concentrator
........"..,.......
Enthalpy-Concentration Diagram. . . . . . . . . . . . . .
for Aqueous Sulfuric Acid at 1 A tm
v
Page No.
50
51
52
54
56
57
60
63
64
77
78
82
88
89
92
93
96
97
-------
List of Illustrations (Cont.)
Figure No.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
Pilot Plant
. . . . .. .. .. ..
.. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. ..
Experimental Apparatus for. . .
H20 Condensation Experiments
a vs t for Reaction S02 + N02 --+ S03 + NO .
.. .. .. ..
.. .. .. .. ..
.. .. .. ..
. . .. . .
.. .. .. ..
Reaction Rate Constant as a Function. . . .
of Temperature for Reaction
Adsorption Column. . . . . . . . . . . .
.. .. .. ..
.. .. .. ..
.. .. .. .. .. .. .. ..
McCabe-Thiele Plot for NOx Scrubber. . . .
in Catalytic Chamber Process
.. .. .. .. .. .. . ..
High Temperature Stripper. . . . .
with NO Oxidizer
............
Modified High Temperature Process. . . . . .
NO Oxidation Times for Charcoal. . . .
Catalyzed and Uncatalyzed Reactions
.......
" .. .. .. ..
.. .. .. ..
Water-Jacketed UV Absorption Cell. . . . . . . . . . . . .
Absorption of S02 at 2860 A . . . . . . . . . . . . . . . .
Absorption of N02 at 5100 A . . . . . . . . . . . . . . . .
Nitrogen Dioxid1 Calibration. . . . . . . . . . . . . . . .
Curve 2950 cm-
Sulfur Dioxide Calibration
Curve 1150 cm-1
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
Nitric Oxide Calibration. . . .
Curve 1900 cm-1
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
Nitrous Oxide (~20) Calibration
Curve 2230 cm-
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
Carbon Monoxide Calibration
Curve 2160 cm-1
.. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
Carbon Dioxide Calibration
Curve 3640 cm-l
................
Layout and Operating Conditions. . .
of U. S. Steel Agri -Chemical Chamber
Process Sulfuric Acid Plant
...........
vi
Page No.
107
113
121
122
124
125
132
133
136
148
149
150
156
158
159
161
162
163
169
-------
Table No.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
IX.
X.
XI.
XII.
XIII.
XIV.
XV.
XVI.
XVII.
XVIII.
List of Tables
Material Balance for Catalytic Chamber.
Process
. . . . .
Energy Balance. . . . . . . . . . .
. . . . . .
Estimated Major Equipment Cost. .
......8
Estimated Capital Cost Summary. . .
A nnual Operating Costs.
. . . . . .
. . . . . .
. . . . . .
Return on Investment. .
. . . . .
.......
RetUrn on Investment. . .
. . . . .
. . . . . .
Comparative Characteristics of Power Plant
Flue Gas and Sulfur Burner Gas
. . . .
Rate of Oxidation of S02 by N02 . . . , . . . . .
Solubility of NO in 74% H2S04 Containing. . . . .
S05NH
Contamination of Sulfuric Acid by Fly Ash.
. . . .
Material Balance-Modified Chamber Process
Heat Balance-Modified Chamber Process. . . . .
Heat Balance: Modified Chamber Process for. . .
800 Megawatt Power Plant Installation
Assumptions for Economic Analysis
.......r.
Estimated Equipment Cost for Modified. . . . . .
Chamber Process
A nnual Operating Cost Estimate: Modified
Chamber Process
. . . .
Return on Investment: Modified Chamber.
Process
. . . .
vii
Page No.
10
11
21
22
23
24
25
34
39
55
58
61
68
69
70
71
73
74
-------
List of Tables (Cont.)
Table No.
XIX.
XX.
XXI.
XXII.
XXIII.
XXIV.
XXV.
XXVI.
XXVII.
Cost of 10 SCFM Pilot Plant. . . .
. . .. . . . .
. . . . .
Comparison of Alternative Flue. . . .
Gas Supply Systems
. . . . . . . . . .
Vapor Pressure of Water Over
Sulfuric Acid Solutions
. . . . . . . . . .
. . . .
Experimentally Measured Gas System Pressure Drop
as a Function of Gas Flow Rate
. . . .
Experimental Results: . . . .
Pilot Plant
. . . . . . . . . . . . . .
Results of Condensation
Experiments
Results of Kinetic Experiments
. . . . . . . . . . . . . . . . .
.............
Calculated Reaction Rate Constant for S02 Oxidation by N02 .
Results of High Temperature. . . . . . . . . . . . . . .
Absorption Experiments
XXVIII. Contact Time for Oxidation
of Nitric Oxide
XXIX.
XXX.
XXXI.
................
Results of Catalytic HNS05 . . . . . . . . .
Oxidation Experiments
Results of Catalytic Oxidation of HNS05 Using
Water-Glass Treated Charcoal
. . . . . . .
. . . . . . .
Infrared Spectral Band Frequencies
for Several Gases of Interest
. . . . . . . . . . . .
viii
Page No.
83
86
101
106
110
116
118
120
130
135
140
144
154
-------
1.
SUMMARY
A process has been developed for the simultaneous removal
of sulfur dioxide and nitrogen oxides from pow~r plant flue gas at nor-
mal stack gas effluent temperatures. The process, based on the Lead
Chamber Process for sulfuric acid manufactUre, incorporates two tech-
nological breakthroughs developed on this contract to produce sulfuric
acid and nitric acid without the need for additional raw materials or heat
beyond that in the flue gas. Sulfur dioxide removal is accomplished by
oxidation with nitrogen dioxide followed by absorption of the resulting
sulfuric acid and the oxides of nitrogen in a recycle stream of sulfuric
acid. The acid is stripped of its nitrogen oxide content and is recircu-
lated to the absorber after the product acid is withdrawn. The nitrogen
oxides are reoxidized with the excess nitrogen dioxide being converted
to nitric acid and the bulk of the oxide recycled for further sulfur
dioxide oxidation. Thus the nitrogen dioxide acts as a reactive catalyst.
The two major modifications to the original Chamber Process
which make this new method technologically and economically feasible
are: a high temperatUre isothermal scrubber which permits the
scrubber to recover the sulfuric acid and the oxides of nitrogen without
absorbing any diluent water (which would require heat to remove), and
a catalytic stripper by which the oxides of nitrogen in the scrubber effluent
acid are rapidly oxidized in the liquid phase, thus avoiding the vapor
pressure limitations of recovering a dissolved gas from solution and
thereby eliminating the need for extra heat.
Experimental confirmation on a limited laboratory scale was ob-
tained for the two vital parts of the new Catalytic Chamber Process: high
temperature absorption of sulfuric acid and nitrogen oxides in sulfuric
acid, and catalytic oxidation of nitrogen oxide solutions in sulfuric acid
- 1 -
-------
solution to N02 using activated charcoal as the catalyst. In addition,
kinetic studies of the sulfur dioxide oxidation reaction with nitrogen
dioxide were conducted which established the practicality of performing
this reaction at the low concentration levels present in power plant flue
gas. A 10 SCFM pilot plant was constructed for use in evaluating the
various process stages.
A preliminary economic analysis based on an 800 megawatt
power plant, producing 1.44 million standard cubic feet per minute of
flue gas, showed that an add-on Catalytic Chamber Process clean":up
plant would cost about 11.3 million dollars in capital costs and about
3.7 million dollars per year in operating costs before taking a credit
for acid by-products. If 80% sulfuric acid and 52% nitric acid can be
sold for $20 and $40 per ton (100% basis) respectively, the plant can
make a gross pretax profit of 2.5 million dollars per year.
If the sulfuric acid could be sold for only $12.50 per ton (100%
basis) and nitric acid sold for $40 per ton, the plant would still make
a gross, pretax profit of about $0.58 million per year. Even if HN03
recovery were not desirable and therefore not provided for at a given
location, sale of sulfuric acid at a mere $14 per ton (100% basis) would
recover the total operating cost.
In addition to its application to the removal of pollutants from
power plant stack gases, the Catalytic Chamber Process may provide
a possible alternative to currently used methods of manufacturing sul-
furic acid from conventional sources of sulfur dioxide. Since the nitro-
gen oxides are oxidized separately from the sulfur dioxide oxidation
reaction, and consequently there is no longer a difficulty in removing
the nitrogen oxides from concentrated sulfuric acid solution, it would
now be possible to operate the Lead Chamber Process at high acid
strengths, thus removing the major drawback to this currently obsolete
process. The Catalytic Chamber Process appears to be the first signi-
ficant, new process concept for sulfuric acid manufacture to be suggest-
ed in the past 50 years.
- 2 -
-------
II. RECOMMENDATIONS
The Catalytic Chamber Process which is presented in detail
in Section III has been conceptualized on the basis of somewhat
incomplete supporting experimental work. The process stages of
major interest: high temperature sulfuric acid and nitrogen oxide
absorption, and catalytic oxidation of nitrogen oxides have been
individually confirmed on a laboratory scale, but no work has been
performed on an integrated, continuous basis. In addition there are
areas which are virtually impossible to evaluate on a small scale
(such as the need for mist eliminators) thus making it imperative
that the scale of work be expanded. The following. recommendations
for future work are made with those considerations in mind:
A. Catalyst Evaluation
The activated charcoal catalyst used in the small scale
experimentation appears to have stability problems which must be
evaluated. In addition, it is important that the fjeld of catalysts be
properly screened to see which is the best catalyst to use. There
are many candidate materials and only one has been tested.
B. Pilot Plant Conversion and Operation
The pilot plant is currently set up to evaluate an early version
of the Catalytic Chamber Process and it is necessary to convert to
equipment to conform with the latest system including the high temperature
scrubber and the catalytic stripper. This should be the first item of
business in a continued research program.
C. Demonstration Plant Design
As soon as the general feasibility of the Catalytic Chamber
Process is established work should commence on a scale-up beyond the
10 SCFM capability of the current pilot plant. It is strongly recommended
that design work begin as soon as possible on a 1000 SCFM demonstration
plant which would put all the process problems on a realistic basts. Such
things as the need for mist eliminators, the ability to control the composition
- 3 -
-------
of gases in the process and the realistic limit on absorption efficienty
can only be fully evaluated at the 1000 SCFM scale. In conjunction with
this an economic evaluation of the capital and operating costs should be
undertaken.
- 4 -
-------
III.
INTRODUCTION
During the past several years the problem of air pollution from
power plant stack gases has received much attention from government
and civic agencies with the result that a great deal of research and develop-
ment effort has been devoted to the clean-up of these contaminated gases.
The primary offender in stack gases is sulfur dioxide (3000-6000 ppm) ,
although the nitrogen oxides present are also considered detrimental to
human, plant and animal life. 1
HistOrically, contaminated gaseous effluents from plant stacks have
been treated in wet scrubbers to remove undesirable components, but these
wet processes have the disadvantages of releasing a gas of low temperature
and high water content which prevents its ready dissipation in the atmos-
phere. More recently several dry processes have been proposed which
avoid these problems.
Oxidation of the 502 to 503 and recovery of the sulfur values as
sulfuric acid results in a low operating cost because no raw materials
are needed. A route to effect the oxidation of 502 to sulfuric acid and
remove NOx as well as 502 from power plant stack gases is through the
use of nitrogen dioxide as a reactive, regenerable catalyst in a manner
similar to the old Lead Chamber Process technique. The work on this
contr~ct (NAPCA PH86-68-75) was undertaken to determine the ability
of N02 to oxidize the dilute 502 present in flue gas in a recycling scheme
in which the NO produced from the 502 oxidation reaction is reoxidized
to N02 and thus is available for reuse. This report describes the con-
ceptual progress made during the 21 months of the contract as well as
the experimental work performed to support the theory.
- 5 -
-------
IV.
CURRENT PROCESS
A.
General
This section of the report will discuss the process concept in its most
advanced form. The fundamental aspects of the process have been established
experimentally although it has not been operated in an integrated manner. The
process will be described, experimental verification given and future plans
presented. A process flow sheet, heat and material balances and a preliminary
economic evaluation will be presented. Succeeding sections will discuss the
original process concept and will follow its evolution to the current system
including the experimental support work done throughout the contract period.
The discussion will include the breakthroughs that make this process feasible
and the potential impact of these developments on the sulfuric acid industry
in general.
B. The Catalytic Chamber Process
The rr~ost recent conceptUalization of the Tyco/ NAPCA process for the
removal of S02 and NOx from power plant stack gas has been named the Cataly-
tic Chamber Process because of the novel manner in which the oxides of nitro-
gen are recovered and oxidized for reuse as the S02 oxidant. In this process,
(see Flow Sheet, Figure 1) the raw flue gas, having had most of the fly ash
removed, is contacted with a stream of recycle N02 which oxidizes the S02 to
S03' which hydrolyzes to sulfuric acid through contact with water vapor in
the gas stream. The reacted gases pass through a sulfuric acid scrubber which
removes the newly formed acid and the oxides of nitrogen and the cleaned, wet
stack gas is vented to the atmosphere at relatively high temperature, insuring
easy dissipation. The nitrogen oxide-bearing sulfuric acid (nitrose) is
passed on to a stripper column where it is contacted with a catalyst (activated
charcoal, for example) counter-current to a small side stream of the raw
flue gas. Here the oxides of nitrogen are catalytically oxidized to N02 in
the liquid phase with the flue gas providing the necessary oxygen. The gas
stream from the stripper, now containing all the oxides of nitrogen as N02
and some of the flue gas is passed through a water scrubber which recovers
the excess N02 (equivalent to the net NOx recovered) as nitric acid before
being mixed with the bulk of the flue gas stream for further 802 oxidation.
- 7 -
-------
Cooler H2S04
S03
p, S02
Flue
Gas
7% H20
Cooler N 203
Stack Gas -<
250 ° F
7% H20
60 ppm NOx
Reactor
N02 in
Jue gas I I'" ~
HN03 Gas
Absorber 1 100 of
Product
HN0352%
t
10%--->,-
I~l
Cooler
Flue Gas
300 ° F
0.3% S02
7% H20
600 ppm NOx
Gas
N02
250 0 F
High
Temperature
Scrubber
250° F
H2S0480%
O. 11 M HNS05
250 of
I
Ca ta lytic
Stripper
Filter
H2S04 80%
O. OOlM HNS05
Fig. 1. Catalytic Chamber Process
- 8 -
Product
H2S04
-------
An amount of sulfuric acid equivalent to the 502 recovered is removed as
product from the denitrated acid leaving the stripper before the acid is re-
turned to the scrubber for recycle.
Table I shows a complete material balance for the process shown in Figure
1 along with the temperatures of the various process streams. Table II presents
an energy balance for the system on the same basis as the material balance.
Examination of these tables will show that the process requires no additional
heat (in fact, it is exothermic overall) and a minimum of cooling water; both
facts contribute markedly to the low operating cost which will be itemized
later.
c. Flue Gas Condition
It is obvious that the composition and temperature of the power plant flue
gas will vary with the plant being treated, depending on such things as ambient
temperature and humidity, type of fuel, size of the plant, efficiency of hear
recovery and anti -pollution control devices already in use. Rather than try to
cover a whole range of conditions, the following "average" gas composition was
assumed, based on OAP's recommendation, and was used throughout the study:
. . . . . . . . . . . . . . . 14.6% V
C02
H20
5°2
°2
N2
NO
x
...............
........
.......
7.3% V
0.3% V (6400 ppm by weight)
...........
. . . .
2.8% V
. . . . .
. . . . . . . . . . 74.8% V
. . . . .
..........
0.06% V (600 ppm by weight)
Ash
........
.......
0.2% wt
Temperature. . . . . . . . .. 300°F
The effect of the fly ash on the operation of the process is of some
interest although by the very nature of the process, its influence is minimized.
Most power plants have some type of dust collection system to clean up the
stack gas before it is vented to the atmosphere, thus partially solving the prob-
lem. If an appreciable amount of ash is introduced to the proposed 502 clean-
up system, there are three ways to approach the problem:
- 9 -
-------
Table I. Material Balance for Catalytic Chamber Process*
Temp.,
Stage C02 H20 S02 02 N2 H2S04 Ash N02 NO Total of
Gas from stack 215 44 6.4 30 702 2.0 0.06 0.54 1()00 3()()
Gas to stripper 21.5 4.4 0.64 3.0 70.2 0.2 0.006 0.054 ]()O 2S()
Gas from stripper 21.5 3.97 1. 18 70.2 10.48 1f17.33 2:=;()
Gas to HN03 absorber 21.5 3.97 1. 18 70.2 10.48 107.33 11)1)
Gas from HN03 absorber 21.5 3.97 1. 18 70.2 9.42 0.23 106.50 1()()
Water to HN03 absorber 1. 03 1. 03 7()
Product HN03 (52%) 0.89 0.97 1. 86 1()O
(HN03)
Gas to reactor 215 43.57 5.76 28.18 702 1.8 9.43 O. 75 1(1)6.49 3()1)
Gas to scrubber 215 41.95 28.18 702 8.82 1.8 5.28 3.45 1 W)/1. 48 2S'1
Gas to stack 215 39.75 28.18 702 0.091 0.059 98S.()8 2S()
.... Acid to scrubber 716 2864 0.053 0.034 3S8().1)87 2SI)
o
Acid from scrubber 718.21 2872.82 1.8 5.24 3.43 36()1. S() 2S0
Acid from stripper 718.5 2873.8 2.0 0.053 0.034 3S94.387 2,')1)
Acid from filter 718.23 2872.74 0.053 0.034 3,')9]. OS7 2,')1)
Cake from filter 0.27 1.06 2.0 '" '" 3.33 2;:;0
Product H2S04 (80%) 2.23 8.74 '" '" 10.97 251)
* Assumptions:
1. Basis: 1000 lb of entering flue gas
2. 10% of flue gas passed through the stripper
3. Stripper is 99% efficient in denitrating the scrubber acid
4. Scrubber is 99% efficient in NOx recovery
5. All S02 entering stripper oxidized by N02 and scrubbed out
6. Acid is filtered after the stripper and filter cake is 60% solids.
-------
He(~~ ~~ 1~)(Yfn2 (H 2CC~17S2) (1. 8)
Water absorbed in scrubber (H20 + S03 --+ 80% H2S04)
(0.09 mol/O. 8), (18 lb/mol) (967 Btu/lb) 1,950
w(6~r22~5~~f)d(i/~)t(Tif~/~~~05 + H2~~67 Btu/lb)
Table II.
Energy Balance (Basis:
FlueGas at 32°P)
1000 lb
Heat Inputs
Entering flue gas
(1000 lb) (0. 25 Btu lb/lb/o P) (300 - 32)
Water into HNO absorber
(1. 3 lb) (1 Bt~ lb/ 0p), ' (70 - 32)
Heat of reaction (NC2 + S02)
(9. 8 lb) (1350 Btu/ lb)
Heat of absorption (N203 + H2S04)
(8. 73 lb) (271 cal/g ) (1. 8)
Total
Heat Outputs
Stack gas effluent
(985 lb) (0. 25 Btu/lbr F) (250 - 32)
Product H 2S0 4
(12. 3 10 ') (95 Btu/lb)
Heat of reaction (HNSO + 0,)
(0.2298 mol) (8280 c~/m01) (1,8) .'
Product HN03
(1. 86 Ib) (0.66 Btu/lb/oP) (100 - 32)
- 11 -
Bru
67,000
39
13,200
4,260
30
2,000
88,479
53,600
1, 170
3,430
83
-------
Table II - Continued
Heat Outputs
Cooling reacted flue gas (342 -+ 250 'F)
(1007 lb) (0. 25 Btu/lb) (342 - 250)
Cooling stripper gas for HNq~ absorber to 100 'F
(108 lb) (0.075) (0. 25 Btu/lt1/o F) (250 - 10;)
Btu
23,200
300
Cooling scrubber (heat of absorption - N 203 + H 2S0 4)
(8. 73 lb) (271 cal/g) (1. 8) 4, 260
Vaporization of water in scrubber (from HNS05 genera-
tion)
(0.2295 mol) (1/2) (18 Ib/mol) (967 BW/lb) 2,000
Total 87, 943
Error = 536/ 88,479 = 0.6%
- 12 -
-------
1) Depending on the use of the acid and the amount of fly ash in the raw
flue gas, the ash can be ignored and the acid sold with a small amount of sus-
pended solids. This might well be possible for acid sold to the fertilizer industry
where acid purity is not critical.
2) Precipitation equipment could be added to the flue gas stream prior to
entering the S02 removal plant to remove the bulk of the ash.
3) The ash could be allowed to enter the system with the gas which would
be scrubbed clean of ash in the scrubber section. The acid could then be filtered
either downstream of the scrubber or downstream of the stripper, although the
soluble ash compounds would still be present in the acid.
One of the controlling factors in this problem is the effect of the ash on the
catalyst in the stripper. If the ash acts as a poison for the catalyst, then it would
have to be removed before it entered the stripper column. Since the catalyst has
not been chosen and the effect of the ash is not known, this area has not received
a great deal of emphasis. The solution is straightforward: filtration of the acid,
with the effect of the ash on the overall process not being of concern to us at
this time. (The one point of some interest; the solubility of the ash in sulfuric
acid has been examined and the experimental results showing very low ash
solubility can be found in Section V -G.) Since the soluble ash compounds could
not be filtered out, the catalyst would have to be resistant to the compounds.
D. Oxidation Reaction Vessel
The purpose of the reactor is to provide sufficient volume to allow the
oxidation of the S02 to take the place according to the following reaction:
S02+N02 -) S03 +NO
The S03 is quickly hydrolyzed by the large excess of water vapor present in the
flue gas forming sulfuric acid mist which remains in the gas stream. There is
the possibility that some of this acid will condense on the walls of the reactor
which should be considered when choosing materials of construction, but the
greatest part of the acid will continue through the reactor to the scrubber. Sul-
furic acid mist is a difficult material to remove from gases in general and
there is no reason why an appreciable amount should condense in the reactor.
The actual size of the reactor is controlled by the residence time required
- 13 -
-------
to effect complete oxidation of the S02 to S03. Under the gas conditions shown above,
this reaction time is under 5 seconds as is shown by experimental verification in
Section V-D. The total heat of reaction (see energy balance in Table II) is quite small
in this dilute gas stream and the gas temperature is only raised from 300 ° F to about
340oP. This low temperature avoids the requirement for the large heat dissipating
chambers needed in the standard Chamber Process and it is quite likely that in actual
practice the reactor chamber will vanish altogether with the S02 oxidation occurring
in the connecting ductwork. (The other reason for large chambers in the standard process,
that of residence volume for the reoxidation of the nitric oxide formed in the S02 oxi-
dation, is also avoided in the Catalytic Chamber Process by performing this oxidation
catalytically in a separate stage. This will be discussed in detail later.)
E. High Temperature Scrubber
The reacted flue gases enter the scrubber as shown in Figure 2 and consist of
inert gases, sulfuric acid mist, unreacted S02 and S03' water vapor and equimolar
quantities of NO and N02. The reason the oxides of nitrogen must be in equimolar ratio
is because individually NO and N02 are virtually insoluble in sulfuric acid while together
they combine with sulfuric acid to form a crystalline material known as nitrosylsulfuric
acid which is soluble in sulfuric acid up to 60% by weight:
N02 + NO + 2H2S04 -> 2HNS05 + H20
Thus, in the scrubber the reacted gases are contacted with sulfuric acid which
scrubs out the sulfuric acid mist, the nitrogen oxides and, depending on the acid tem-
perature, the water vapor.
Here we are faced with one of the major problems attempting to apply Chamber
Process technology to flue gas clean-up: the large excess of water in the stack gas above
that which is needed to make sulfuric acid. In the standard Chamber Process the water
to S02 mole ratio in the entering gas is 1: 15 and more water must be added to make the
acid. In the Tyco Catalytic Chamber Process this ratio in the flue gas is 25: 1 and
scrubbing this stream with cold sulfuric acid will cause dilution of the acid and create
the necessity of reconcentrating the acid for recycle, an expensive, heat consuming
process. The way around this is to scrub the reacted gases with hot sulfuric acid as
is explained below.
If the gas is scrubbed with sulfuric acid at 80 ° F as in the standard
Chamber Process, all the water will be absorbed, thus diluting the acid. How-
- 14 -
-------
Stack Gas
250 0p
7% H20
H2S0~ 80%
250 ° F
o 001M HNS05
P lue G
250 0p
7% H2
87~0 ppm N203
H2S04 (g)
S03
S02
15'0 ppm 1
N203
"
as
2
° 0
H~g9~ 80%
. 11M HNS05
Pig. 2. High temperature scrubber
- 15 -
-------
eve r, if the co 1 u m n is run a t a temperature w her e the vapor
pressure of water over the incoming acid is equal to the partial pressure of
water in the incoming gas, no water will be absorbed in the column and the
acid will leave the scrubber at the same concentration as it entered. For
80% acid and a gas stream containing 7.3% water (52 mm Hg partial pressure),
this temperature is 250°F (see Figure 3).
The question arises as to whether or not the oxides of nitrogen can be
absorbed at this temperature. Using a small laboratory column in continuous
counter-current flow it was found that about 99% of the oxides of nitrogen were
removed by scrubbing with sulfuric acid at 250 of. The details of this work are
given in Section V -K3 . It is clear that a scrubber operating at the appropriate
temperature can remove the oxides of nitrogen as well as the sulfuric acid
while permitting all the incoming water to pass into the atmosphere.
Of major importance is the ability of the scrubber to remove all sulfuric
acid mist from the reacted flue gas. It has been determined experimentally
(see Section V -J 1~ that on a pilot plant scale (10 SCFM) no measurable amounts
of acid mist are able to pass through the scrubber when operated at the 100° F
level. It is likely that no real engineering data can be determined on the acid
mist problem until a pilot plant of 1000 to 100,000 SCFM capacity is in operation.
More details on the scrubber effluent gas can be found in the section discussing
the performance of the laboratory pilot plant (V -} 13).
Since the solubility of nitrosylsulfuric acid increases with acid concentration,
it should be possible to operate the scrubber at acid strengths above 80%. The
80% level was chosen because this is the optimum concentration for the standard
Lead Chamber Process upon which the original work was based. In operating
the isothermal scrubber in the Catalytic Chamber Process, the only limitation
on the acid strength is that imposed by the sulfuric acid content of the exit gas
at the operating temperature of the scrubber. As the acid concentration is
increased, the operating temperature of the scrubber must be increased in order
to maintain the vapor pressure of the water over the acid equal to the partial
pressure of water in the entering gas. At some acid strength, the temperature
reaches a level where the vapor pressure of sulfuric acid becomes high enough
such that the effluent gas would contain an undesir able amount of sulfuric acid
vapor. This condition is reached at an acid concentration of about 92% where
the operating temperature of the scrubber would be about 4000 F and the H SO
.24
- 16 -
-------
b.O
::r:
8
8
~
Q)
H
::!
UJ
UJ
Q)
H
p...
S
p..
C\3
:>
o
N
::r:
100
------------
10
I
I
I
I
I
I
~
I
I
I
I
I
I
I
J
1
140
160
200
240
180
220
Temperature, of
260
280
Fig. 3. Vapor pressure of water over 80% H2S04. Data
of Berl and Saenger 4
- 17 -
-------
vapor pressure is 1.9 mm Hg. A standard Chamber Plant could not operate at 93%
acid because of the difficulty in recovering the oxides of nitrogen from a nitrose
of greater than 80% sulfuric acid content. This limitation has been avoided by
using a catalytic stripper as will be discussed in the next section.
Catalytic Stripper
F.
The most critical part of the new process is the recovery and reoxidation
of the oxides of nitrogen. In the st&.ndard Chamber Plant the nitrogen oxides
are sufficiently concentrated that the reoxidation can occur simultaneously with
the sulfur dioxide oxidation even though the reaction rate is comparatively show
a t the 0 per a tin g temperature use. Nitrogen oxide recovery is effected
by the dilution and heating of the nitrose which causes an increase of the vapor
pressure of the oxides which are then swept out of the acid by the high temperature
gases from the sulfur burner. Both heating and diluting are impractical when treat-
ing dilute, wet flue gases in the Catalytic Chamber Plant because they involve a
requirement for heat beyond that which enters the system with the flue gas.
Heat balance calculations on such a system show that the heat penalty would be
enormous; as much as 40% of the total heat generated by the power plant would
be necessary.
The breakthrough that was achieved during this contract t hat a v 0 ids
this heat penalty was the discovery that by contacting the nitrose with a catalyst
such as activated charcoal in the presence of oxygen, the oxides of nitrogen
are oxidized to nitrogen dioxide in the liquid phase and the NO is recovered quantita-
x
tively in a relatively small volume of carrier gas. Figure 4 shows a schematic
diagram of the catalytic stripper that would be used to accomplish the nitrogen
oxide recovery.
The exact mechanism of the oxidation reaction has not yet been determined,
but the net effect is that the total nitrogen oxide content of the nitrose is con-
verted to the relatively insoluble nitrogen dioxide which is spontaneously evolved
from the sulfuric acid. No additional heat is required to drive the reaction to
completion; as a matter of fact the overall process is somewhat exothermic.
Details of the experimental verification of this oxidation can be found in Section V -K4c,
G.
Nitric Acid Absorber
One of the major advantages of the Catalytic Chamber Process is that it
provides for the recovery of nitrogen oxides that are formed during the com-
bust ion of coal in the power plant, thus eliminating another source of air
- 18 -
-------
~~9p 80%
O. 11M HNSO
5
Flue Gas I
250 0 F
N02
Charcoal
Packing
> ~~9~ 80%
O. 001M HNSO
5
Flue Gas
250 0 F
Fig. 4. Catalytic Stripper
- 19 -
-------
pollution. The oxides, which are primarily in the form of nitric oxide, mix
with the recycle nitrogen dioxide and are scrubbed from the flue gas in the
isothermal scrubber. Sufficient nitrogen dioxide must be provided to ensure
that the ratio of nitric oxide to nitrogen dioxide is unity by the time the reacted
gases reach the scrubber so that the total nitrogen oxide content can be absorbed
in the sulfuric acid.
Thus, when the nitrogen oxides are recovered in the catalytic stripper,
there is an excess in the recycle stream equal to the amount of nitrogen oxides
recovered by the process. This excess N02 can be converted to 52% nitric acid by
scrub~ing a sidestream of the stripper gas effluent with water according to the
reaction:
3N02 + H20 -+ 2HN03 + NO
Since the absorption of the nitrogen dioxide does not occur efficiently at high
temperatures, the sidestream must be cooled to 1000 F before it enters the
scrubber. The nitric oxide in the carrier gas is returned to the incoming flue
gas. Although some heat is lost in this process, the total volume of the side-
stream is so low that this loss has little effect on the overall heat balance.
It is interesting to note that in general, power plants make a distinct
effort to minimize the amount of nitrogen oxides being generated in the boiler
in order to avoid pollution. As will be shown in more detail in a later section,
the Catalytic Chamber Process, being designed to recover the nitrogen oxides
in the flue gas as nitric acid, can be operated more profitably as the amount
of nitrogen oxides in the stack gas increases. Discussions with two large
power generating companies have shown that the oxide level can be increased
substantially without any loss in power generating efficiency.
H. Process Economics
A preliminary economic analysis of the Catalytic Chamber Process
has been performed based on guidelines established by the National Air Pollution
Control Administration. This analysis is shown on the following pages in
Tables III through VII and refers to a plant operated in conj unction with an
800 MW power plant producing 1. 44 million SC FM of flue gas.
- 20 -
-------
Table III. Estimated Malor Equipment Cost*
Item Function No. Size Cost
Cyclone Separator Remove coarse 3 487,000 cfm 245,000
fly ash
Scrubber Remove N 203' 3 35 ft. dia. $1,500,000
H 2S0 4 70 ft. high
and fly ash 8 ft/sec
Stripper Recover N203 2 32 ft. dia. 250,000
60 ft. high
2 ft/ sec
HN03 Absorber Recover NO at 44 ft. dia. 450,000
x
52% HN03 60 ft. high
Filter Remove fly ash from 3 38,000 gal/min 250,000
acid acid flow
Product Storage Store H2S04 3 100,000 gal 100,000
Tank
Exhaust fan Provide differential 3 12" H20 60,000
pressure 87 x 106
SCFH
Gas Cooler Cool &tripper gas 1 2.5 x 106 20,000
BTU jhr
2100 feet square
area
Gas Cooler Cool reactor gas 6 124 x 106 525,000
BTU/hr
105,000 feet square
area
Pumps Move acid from 8 38, 000 gall min 50,000
scrubber to 60-ft head
stripper
Pumps Move acid from 8 38,000 gal/min 50,000
stripper to 70-ft head
scrubber
$ 3, 500,000
* 800 megawatt (e) power plant producing 1. 44 x 106 SCFM flue gas.
- 21 -
-------
Table IV. Estimated Capital Cost Summary*
Item
1. Major Equipment (see Table III)
Factor
2. Erection labor and foundation
1. 00
.38
4. Instruments
. S5
.05
3. Piping
6. Electrical
.03
.05
5. Insulation
7. Building
8. Plant facilities
.20
.05
9. Plant utilities
.08
.05
10. Receiving, shipping, etc.
11. Physical Plant costs
12. Engineering and Construction
2.44
.60
13. Direct plant cost
14. Contractor t s fee
3.04
.10
15. Contingency
. 15
16. Fixed capital costs
3.29
17. Total fixed capital costs = $3,500,000 x 3.29:: $11,515,000:
* 800 megawatt (e) power plant producing 1. 44 x 106 SCFM flue g~s.
- 22 -.
-------
Table V. Annual Operating Costs*
Raw materials and chemicals
Catalyst replacement
Packing replacement
Water (3500 GPM at $0. 101M gal)
Direct labor
Supervision
Maintenance at O. 05 of fixed capital
Supplies at 15% of maintenance
Utilities
Power
Heat
$500,000
100,000
168,000
108,000
27,600
653,000
98,000
145,70g t
Total direct cost
$1,800,300
Payroll burden (20% of direct labor and
. supervision)
Plant overhead (50% of direct labor, supervision,
maintenance, and supplies)
Packing and shipping
Waste disposal
Other
Total indirect cost
16,600
416,600
425,000
25,000
873,200
$
Depreciation (10% of fixed capital)**
Taxes (2% of fixed capital)
Insurance (1% of fixed capital)
Other
Total fixed cost
1, 151,500
230,300
115,150
$1,496,950
Total operating cost
$4, 1 70, 450
(or $1. 64/ton of coal)
* BOO-megawatt (e) power plant p~oducing 1. 44 x 106 SCFM
flue gas. (Coal consumption = 2.22 x 10 tons/yr)
t Heat is needed only for start -up and process adjustment
situations.
**10 year straight-line depreciation.
- 23 -
-------
Table VI. Return on Investment
(Basis: 80% H2S04,600 ppm NOx in Flue Gas)
Credits *
H2S04 at $20/ ton
HN03 at $401 ton
$ 5, 030,000
1,100,000
Total Operating Cost
6,130,000
4,170,450
Net Operating Profit
$/ ton coal t
$1,959,550
$0.88
Return on Investment (before tax)
17%
*
Production figures based on the material balance in Table I are
27, 600 tons /yr HNO (100%,)
314,000 tons/yr H2Sd4 (80%)
252,000tons/yr H2S04 (100%)
t Coal consumption = 2.22 x 106 tons/yr
- 24 -
-------
Table VII.
Return on Investment
(Basis: 93% H2S04,2000 ppm NOx in Flue Gas)
Credits *
H2S04 at $30/ton
HN03 at $40/ton
$7,530,000
4,470,000
12,000,000
Total Operating Cost (including an
additional 430, 000 for heating the
flue gas to 4000 F to maintain 93%
H2S04 in scrubbing system)
Net Operating Profit
$/ ton coal
4,600,000
$ 7,400,000
$3.34
Return on Inves tment (before taxes)
64%
*Based on:
252,000 tons/yr H2S04 (100%)
112,000 tons/yr HN03 (100%)
- 25 -
-------
~
Table III lists the Major Equipment Items and their cost. The estimated
total is $3.5 million. An item that was not included, but may be necessary is
mist eliminators for the gas effluent from the scrubber. At this point it is not
known if they will be necessary, so they will not be considered until more engi-
neering data is available. Table IV shows the development of the equipment multi-
plier used to obtain the total fixed capital costs of $11.5 million. These figures
are also based on recommendations of NAPCA.
Table V is a breakdown of the estimated annual operating costs for the
Catalytic Chamber Process. It is interesting to note the low cooling water cost
and the complete absence of additional heat costs. These two items, which were
major expenses in earlier versions of the process, have been completely eliminated
due to the technological breakthroughs achieved in the Catalytic Chamber Process
Concept.
Table VI presents the very attractive profitability picture for the process
in Figure 1 having the feedstream described in Section IV B (0.3% S02; 600 ppm
vol. NO ). The assumption is made that 80% sulfuric acid can be sold at $20
x
per ton (100% basis) and 52% nitric acid can be sold at $40 per ton (100% basis).
These numbers are reasonable in today's marketplace although they might not
be if every coal burning power plant started producing acid. More will be said
about this matter later.
If the plant is run at the highest possible acid concentration, 93%, a con~
siderable price advantage should be realized. At current market prices 93% sulfurh
acid should be worth $30 per ton (100% basis), thus giving an increase in the revenu
at very little cost. There would be a slight increase in operating cost because the
gas would have to be heated to about 4000 F to maintain the proper temperature in
the isothermal acid loop. This would be partially offset by the fact that the cooling
stages would no longer be necessary. In any case the fact that the acid is kept at
one temperature keeps the additional heating cost at a very low level. The
maximum that this heat would cost would be $430,000 per year and this may be
reduced by both the credit in savings on cooling water and any heat that could be
recovered from the cleaned stack gases.
- 26 -
-------
In addition, it has been found through discussions with two large power
producers that it would take very little effort and would cost virtually nothing
to supply a raw flue gas containing 2000 ppm of nitrogen oxides instead of the
600 ppm assumed for the economic analysis in Table VI, Therefore, at no
additional cost more than three times as much nitric acid could be recovered
for sale. An economic analysis was made based on both the sale of 93% sulfuric
acid and the production of additional nitric acid, the return on investment
would be 64% as shown in Table VII.
This number seems to be high, but it was calculated using the limits of
the OAP guidelines for capital and operating costs. The only variable whose value
is difficult to determine is the selling price of the acids. The factOrs that go into
the evaluation of the market value of sulfuric and nitric acids produced from
power plant stack gases are very complex and are 'beyond the scope of this
contract. In addition, marketing factors should be taken into account: will
power generating companies want to go into the chemical business? Should acid
be manufactured if it has to be shipped fairly long distances to custOmer, etc?
Some of these questions are discussed in the next section and make the picture
a little more clear. The major point is that the Catalytic Chamber Process can
be operated at a profit while purifying power plant stack gases to a highly accept-
able level, so long as acid markets exist and remain intact over the life of this
air pollution control equipment.
1. Effect on the Sulfuric Acid Industry
The Catalytic Chamber Process has the potential for pro-
found effect on the sulfuric acid industry5, 6 first through the increase
in the overall production of acid if coal burning power plants use the
process for pollution clean-up, and secondly because of the possibility
of a new process for acid production from conventional sulfur sources.
The problem is quite complex and a full treatment is beyond the scope
of this report, but a few points should be made at this time.
From a technological viewpoint, the Catalytic Chamber Process
could revive the concept of the Lead Chamber Process from its present
obsolescence. There are two basic, interrelated reasons for this obso-
lescence: the conventional Chamber Process cannot economically pro-
duce acid stronger than about 77%, and conventional Chamber Plants can-
not be built, as a practical matter, much larger than 100 tons per day.
- .:?7 -
-------
'''!III
Examination of the conventional process (see Figure 5) shows why
these problems exist. The acid concentration is determined by the ease
of recovering the nitrogen oxides from the scrubber acid in the Glover Tower
(stripper). As the sulfuric acid concentration is increased, the nitrosylsulfuric
acid (and thus the nitrogen oxides) becomes more soluble and the N203 vapor
pressure at a given HNS05 concentration decreases. Above 77%, the heat require
to denitrate the acid is greater than can be supplied by the hot gases from the
sulfur burner, thus limiting the sulfuric acid concentration in the Glover Tower tc
this level.
The size limitation of the conventional process is based on both the
large amount of heat generated during the sulfur dioxide oxidation and the
slow reaction rate of nitric oxide oxidation. The gases in the chambers
react in a relatively concentrated mixture so that a great deal of heat is
given off due to the oxidation of S02' This heat must be dissipated in order
to facilitate the reoxidation of the nitric oxide produced. The NO oxidation
reaction has an inverse rate vs. temperature relationship; i.e., the higher the
temperature, the lower the reaction rate. Since the rate of NO reoxidation is
slow in any case (thus requiring long residence times), care must be ta~(en to
keep the temperature down to effect reoxidation as fast as possible. To remove
the necessary amount of heat the chambers must have very large surface areas
and to fully reoxidize the NO the chambers must provide large residence volumes
thus creating the requirement for huge box-like chambers. The necessary size
of the chambers for a small plant is obtained from Fig. 52 in Appendix 3. The
size of a conventional Chamber Plant producing 1000 tons/day (ten times the
capacity of the standard Chamber Plant) staggers the imagination.
Most of the acid strength and size problems are avoided in the Catalytic
Chamber Process. The nitrogen oxides are no longer recovered by heating
the nitrose; instead the denitration is accomplished by catalytic oxidation of
the nitrosylsulfuric acid at moderate temperatures. This means that no
additional heat is needed for the stripping step and it will work as well with
highly concentrated acid as with 77% acid. In addition, the 93% concentra-
tion limit placed on the flue gas clean -up process is not applicable to a pro-
cess which treats a dry gas such as the off-gas from a sulfur burner be-
/
- 28 -
-------
Ammonia' 2.41b
77~ H SO
Ai
o '2" '4
64% H2S04 H20, 840 lb
r: 700 lb
"- Glover Lead Gay-
tower - chambers Lussac
Sulfur tower
r: 7350 lb burner N2' 02
H20,S02
NO
x 2.57 mm Btu
2620 lb
8360 lb
78% HC)S0.4' 8240 lb
Product: 78% H2S04' 2565 lb
(10(>% H SO = 20001b)
2
4
Sulfu
N
co
Fig. 5. The Lead Chamber Process Flow Diagram*
7
*From Duecker and West, 3rd ed., p 119.
-------
cause there is no need to run the scrubber at high temperature conditions to
avoid absorbing H20. The process would be water deficient so there would be
no excess water to force through the scrubber and the whole process could be
run at low temperatures. At low temperatures, there would thus be no problem
with sulfuric acid vapors to keep the maximum concentration down.
The size problem is considerably reduced in the Catalytic Chamber
Process because there is no need to minimize the temperature in the chambers
to accelerate the NO reoxidation reaction. In the new process the nitrogen
oxide oxidation would take place separately in the catalytic stripper, so the
emphasis on cooling the chambers is reduced. The S02 oxidation reaction
rate is increased with increasing chamber temperatures making it even more
attractive. There would still be the necessity to cool the gas at some point
because the system would have excess heat, but this could be turned into an
economic advantage. The gas from the sulfur burner enters the system at
about 1600°F, thus being about 1300°F hotter than necessary. This is a
very high grade heat source and could be easily adapted to electrical power
generation to supply power to the plant complex. ... q- ..
These process consid"erations are in the conceptual stage
of development, but the advantages to be derived from them cannot be over-
looked. Future work should be devoted to examining the engineering aspects
of the process to determine if the Lead Chamber Process can indeed make
a come -back.
-" 3'0 -
-------
V.
THE EVOLUTION OF THE CATALYTIC CHAMBER PROCESS:
CONCEPTUALIZATION AND EXPERIMENTATION
A.
General
This section of the report will follow the step -by -step develop-
ment of the Catalytic Chamber Process from the concept presented in the
original proposal through the experimentation, concept reformulation, pilot
plant design and construction, heat and material balances and economic anal-
ysis which led to the current process as outlined in Section IV. The approach
will be chronologically oriented rather than divided into process stages. This
is intended to show the evolution of the process as a whole rather than the de-
velopment of individual unit operations.
B.
The Originally Proposed Concept
As originally proposed, the Tyco process for flue gas clean -up
was essentially the Lead Chamber Process with only minor modification,
as can be seen from Figure 6. The distillation column is equivalent to the
Chamber Process' Glover Tower and the packed bed collector is the Gay-
Lussac Tower, with only the acid filtration step and the acid collection
technique being new to the process. At that time, there was no provision
for recovering nitric acid from the process.
In the originally proposed Chamber Process for stack gas clean-up, the
flue gas at 9000 F enters the distillation column where it contacts a recycle stream
of sulfuric acid containing dissolved oxides of nitrogen and excess water. The
hot gas concentrates the acid and recovers th~ oxides of nitrogen for sulfur
dioxide oxidation. The gas mixture continues on to the retention chamber
which provides the residence volume to enable the S02 oxidation to occur:
.
502 + N02
--+
S03
+ NO
-" 31 -
-------
~
MAKEUP
WATER
STACK 9000F
GAS
NO + N02
DISTILLA -
TION
COLUMN
AkUL
COOLER
70%
H 2 SO 4
Fig. 6.
200°F
RETE NTION
CHAMBER
NITROSE
ACID
300°F
( 140°F)
(GOOC)
DRY STACK
GAS
IRRIGATED
FIBERGLASS
PACKED
BED
COLLECTOR
(30°C)
ACID- ASH SLURRY
WASH
WATER
ASH
WATER
RECYCLE
25°C
ACID
Originally Proposed N02 -Oxidation Process for
Recovery of S02 from Power Plant Stack Gas
- 32 -
-------
The S03 interacts with the water in the gas stream to form sulfuric acid
mist which remains entrained in the gas. In addition the retention chamber
provides the volume for the reoxidation of NO to N02:
2 NO
+ °2
-
2 N02
The NO which is formed during the S02 oxidation must be reoxidized in
this system before it enters the scrubber jcollector in order to allow recovery
of the oxides. Both nitric oxide and nitrogen dioxide are relatively insoluble
in sulfuric acid separately, but when present in equimOlar amounts they are
soluble up to 60% by weight through the formation of nitrosylsulfuric acid:
NO
+ N02
+ 2 H2S04
-
2 HNSOS
+ H20
Therefore, the mole ratio of NO to N02 must be accurately maintained at
unity in order to efficiently scrub the nitrogen oxides from the gas stream.
The sulfuric acid mist, the oxides of nitrogen and the entrained fly
ash in the f1 u e gas are scr';lbbed out of the gas stream with clean, cool
sulfuric acid in the fiberglass packed -bed collector. The clean, dry gas is
then vented to the atmosphere. The collector has to be packed with fiber-
glass in a fashion similar to a commercial mist collector in order to achieve
maximum recovery of the sulfuric acid mist, a notably difficult material to
remove .from gas streams.
The nitrose from the collector is filtered to remove the fly ash and
is then passed through the distillation column for recovery and recycle of
the nitrogen oxides. In addition, the hot gases vaporize the excess water
that was picked up in the scrubber so that the acid leaving the distillation
column is both free of the oxides of nitrogen and V; of the desiTed concentration.
Product acid is drawn off from this stream after being cooled and the remain-
der is recycled to the collector.
C.
Feasibility of Original Concept
It was well known that the nitrogen dioxide oxidation of sulfur di-
oxide was a feasible technique as evidenced by the long. standing success of
- 33 -
-------
the Lead Chamber Process for sulfuric acid manufactUre. What remained
to be determined was whether the technique would work with a wet, dilute
flue gas as well as with the dry, concentrated product of a sulfur burner.
Table VIII summarizes the differences between the two gases.
TABLE VIII.
Comparative Characteristics of Power Plant
Flue Gas and Sulfur Burner Gas (Lead Chamber Process)
Flue Gas
Sulfur Burner Gas
Temperature (OF)
Sulfur Dioxide Concentration (%)
Water: S02 Ratio
300
0.3
24: 1
1600
10 -13
1:15
1) Temperature. The Lead Chamber Process requires a hot (16000P)
gas in order to concentrate the acid and vaporize the oxides of nitrogen in the
Glover Tower. With flue gas being available at only 300°F, could the proposed
process operate at some reasonably economical level considering the heat that
would have to be put into the system? This could be determined only after the
process had been worked out in some detail and complete material and heat
balances were developed, after which an economic analysis could be made.
2) Gas Concentration. The reaction kinetics of the Lead Chamber
Process have been worked out in some detail dJJring the century or so that the
7 ~ .
process has been used for acid manufactUre,' and it was well known that the
limiting reaction was the slow reoxidation of the nitric oxide in the lead
chambers. However, virtually nothing had been done at the low sulfur dioxide
levels present in flue gases. What remained to be done was to determine if
the sulfur dioxide could be oxidized at relatively high rates at 0.3% concen-
tration and if the resulting nitric oxide could be oxidized back to nitrogen di-
oxide within a reasonably sized reactor.
3) Water Content. The lead Chamber Process requires that water
be added to it because the feed gas contains only about 1/15 of the water needed
to make sulfuric acid from the sulfur dioxide. The power plant flue gas, on
the other han~, contains more than 20 times too much water for acid production so
some technique must be employed to remove water from the system. The
question was: can this be done economically"?
- 34 -
-------
These three points were the basic questions that had to be answered
before any scale-up could be attempted. There were other problems that also
had to be analyzed before the feasibility of the process could be determined.
Among these were:
- Maximum S02 recovery efficiency
- Maximum nitrogen oxide recovery efficiency
- Operating temperatures of the scrubber, reaction chamber
and stripper
- Optimum product acid concentration
- Effect of fly ash on product acid purity
- Liquid and gas phase analytical techniques
D. Batch Kinetic Study
In order to evaluate the effect of low sulfur dioxide concentrations in
the flue gas a series of kinetic studies were initiated using the batch reactor
shown in Fig. 7. Concentrations of reactants (N02 and S02) were measured
spectroscopically in the ultraviolet range as a function of time, the amount of
o
N02 being determ~ed by the fractional extinction at 5100 A and the S02 obtained
from that at 2560 A. Later in the program techniques were developed using
the infrared region of the absorption spectra which permitted more complete
gas analysis at no loss in accuracy. The details of the UV analytical procedure
can be found in Appendix 1 and the procedures for the IR technique in Appendix 2.
As shown in Fig. 7 gases were expanded from three known volumes
simultaneously into the water-jacketed UV absorption cell via a 4-way stopcock,
all bores of which were of equal diameter and volume. One line contained
S02 + N2' a second NO + N2' and the third N02 + 02 + N2. The N2 was used
as an inert additive to adjust the pressures of all three gas mixtures to 760 torr
prior to expansion into the evacuated cell. Each of the gas mixtures was pre-
pared by expansion of pure gases into a calilirated 500 cc bulb, adding N2 to
atmospheric pressure and allowing several hours for gas mixing. The volumes
of burettes were about 25 cc each, and all three sections were of the same
diameter tubing. However, expansion into the cell was carried out with the
storage bulbs open to their respective burettes so that pressure drops due
to expansion would be negligible. New gas mixtures were prepared for each run.
- 35 -
-------
CA)
0')
4-way mlxlnQ valve
e
u
u/V Absorption Cell
~ water jacketed
o
m
S02+
N2
~ - STOPCOCK
Mechanical
Vacuum
Pump
Constant
Temp. water
Bath
Circulating Pump
Fig. :
Apparatus for measurin~ rate of oxidation of S02 by N02 . and of NO by 02 0
-------
To evaluate the response and mixing times in the apparatus, a test
run was made with two of the lines filled with air and the third with a mix-
ture of 33% N02 ' balance 02. Simultaneous expansion into the evacuated
cell resulted in a state wherein equilibration occurred within fi ve seconds
'see Fig. 8). The pressure of N02 in the cell using the spectroscopic
measurement agreed with that predicted from the known system volumes.
In the experiments reported, the cell was first filled with 24 torr of
water vapor (no liquid phase was initially present in the cell), prior to simul-
taneous expansion of the three gas mixtu~es into it. A few runs made with
150 tOrr H20 vapor present in the cell indicated that the oxidation rate differ-
ence attributable to high water vapor concentrations is negligible.
Following these calibration tests, reactions between N02 and S02 were
performed in the batch reactor apparatus. The conditions and results of these
runs are lis!ed in Table IX.
~ . . - .
As can be seen from the results (Table IX and Figs. 9 and 10), the
rate of S02 oxidation appears very much faster than that of the reoxidation
of NO. Figure 11 clearly shows a dependence of rate on the square of the
NO concentration, as would be expected. The effect of temperature on re-
action rate for the NO reoxidation step is clearly shown in Fig. 10. The
rate can be seen to decrease by about a factor of two for a 70°C temperature
rise under test conditions. This observation agrees well with data previously
published. 9, 10 An inspection of Fig. 9 also shows that the initial concentra-
tion of N02 present seems to have little if any effect on the rate. The
overall S02 oxidation rate, within the limits of pressures employed in this
study, is controlled by the NO reoxidation time.
- -
To insure that the product formed was actually H2S04 and not some
other species, an experiment was performed using a large bulb (1100 ml volume),
with initial pressures of 2. 1 mm S02' 22 mm NO , 22 mm 02 ' 24 mm H20,
and the balance nitrogen. After contacting the mixture, the bulb was rinsed
several times with water and analyzed for sulfate by the barium nitrate gravi-
metric method. The ashed barium sulfate precipitate weighed 27.50 mg; theory,
for the pressure of S02 used, is 29.30 mg. A quantity of sulfuric acid equal to
94% of theory was thus found. A duplicate run performed later using the same
- 37 -
-------
60
z
o
CJ)
CJ)
~
CJ)
~ 70
a:
.-
lLI
>
~
..J 80
lLI
a:
50
90
,
Baseline
o
15
Fig. 8.
\
NO +02 Added
30
45
TIME, Seconds
p~o = 120.8 mm
pO = 127.3 mm
02
60
90
75
Oxidation of NO by 02 - 420 mt-i.
- 38 -
-------
TABLE IX
Rate of Oxidation of 802 by N02 (Batch tests)
Initial Partial Pressures, mm Hg(3) Reaction Time (Sees)
Run 02 S02 NO N02 H20 Temp. SO (1) NO (2)
No. °C 2 2
- -
1 21. 6 6.4 16. 7 10. 8 24 56 84 282
2 56 4.2 62 28 24 56 22 26
3 20.0 2.3 21. 5 21. 5 24 56 "'5 23
4 20.0 2.3 33.5 10. 8 24 56 <5 <5
5 20.5 20.5 33.5 10.8 24 56 >300
6 20.0 2.3 33.5 10. 8 24 56 <5 <5
7 21. 0 2.6 30.4 32.2 24 56 <5 <5
8 20.0 2.3 11. 0 33.9 24 57 <5 38
9 20.0 2.3 12. 0 34.0 24 57 <5 43
10 20 2. 1 10.5 10.5 24 56 <5 40
11 20 2.1 10.5 10.5 24 56 <5 4-1
12 20 2.1 10.5 10.5 24 56 <5 40
13 20 2.2 7.15 7.15 24 56 "'5 50
14 20 2.2 7.05 7.05 24 56 'Us 48
15 20.5 2.2 23.3 10. 7 24 56 <5 17
16 20.5 2.2 23.3 10. 7 24 56 <5 14
17 20 2.2 23.0 10. 7 24 127 <5 30
18 20 2.2 23.0 10. 7 24 127 <5 31
19 20 2.2 23 10. 7 24 90 <5 27
20 20 2.2 23 10. 7 24 90 <5 21. 3
21 20 2.2 23 10. 7 24 90 <5 27
22 20.4 2.2 21. 1 21. 4 24 56 <5 14
23 20.4 2.2 21. 1 21. 4 24 56 <5 16
(1) Time to reduce S02 to 10% of initial concentration.
(2) Time to return N02 to initial concentration.
(3) Balance of mixture N2' total pressure 1 atm.
- 39.-
-------
20
o
z
~
0
Z 10
0
.....
c::x:
0
-
X
0
~
0
0
LO 5
a::
0
~
-
0
Q>
II)
-
w
:2E
.....
z 2
0
-
.....
U
c::x: 1500C 1000C
w I I
a::
~ -With initial 100 mm N02
o -No initial N02
Initial NO pressure 150mmHg
60°C
I
2.4
2.7
103 IT
3.0
, oK
Fig. 9.
Effect of N02 on the oxidation rate of NO.
- 40 -
-------
130
120
110
u
o
Q)
~ 100
;...J
CiS
H
Q)
0..
S
Q)
~
c:: 90
o
......
;...J
u
CiS
Q)
p::;
80
70
60
50
TIME
080
o
Time for complete re -oxidation
of NO formed to N02
NO + 1/2 02 -> N02
40
(SEC. )
Fig. 10. Effect of temperature on the NO oxidation rate,
- 41 -
-------
o 100 . N02 = 3502
Q)
If) 6. N02 = 5502
~
- 0 N02 = 10502
Q)
> 0 N02 = 15502
Q)
N
0
Z
0
c
.-
C7I
~
0
0 8.
-
"0 ~
Q)
E
~
0
-
0 10
z
-
0
c
0
-
0
"0
IC
0
I
Q)
~
~
0
-
"0
Q)
~
::::J
CT
Q)
~
Q)
E
.- 1
~ 10 100
I
NO/502 Mole Ratio
Fig. 11. Oxidation rate of NO in the vapor phase
catalytic oxidation of S02.
- 42 -
-------
procedure yielded 31. 4 mg (again the theoretical yield was 29.3 mg). The
averaged results of the two determinations show essentially 100% conver-
sion of 502 to sulfuric acid, and it can be concluded that the reactions
which occur are as have been postulated~ .
It can be seen from this experimentation that there is no problem
in obtaining rapid oxidation of 502 with N02' but the simultaneous re-
oxidation of NO back to N02 is far too slow to be performed in this manner.
The next section shows further experimentation designed to examine ways to.
improve this situation.
E.
Nitric Oxide Oxidation
It can be shown that only two variables affect the oxidation of
nitric oxide to nitrogen dioxide: temperature and nitric oxide concentration.
It was determined experimentally that N02 concentration has no substantial
effect (see Figs. 9 and 11) and it was shown previously that the concentration
of 502 has a similar ineffectiveness. .
From the reaction equation ( 2 NO + 02 -... 2 N02 ) one would
expect that the reaction rate would be proportional to the square of the nitric
oxide concentration and directly :Qroportional to the oxygen concentration.
From this assumption, Burdick II has developed the following reaction rate
equation which has been found to agree quite closely with experimental evi
dence.
Kt =
1
2 Co - C 1
r ~ - ~J
2.3
(2 Co - C1)2
log (2 Co - Cl + C)(Cl)
2 CoC 1
where:
K = Reaction rate constant
C = Concentration of NO at time t
Co;:: Original concentration of 02
C 1;:: Original concentration of NO
:~ 43 -
---
-------
From this it would appear that the nitric oxide reaction rate could be in-
creased by increasing the oxygen concentration, but as a practical matter
this cannot be done. The oxygen content of the gas mixtures is determined
by the source of the flue gas and would be very difficult to change. In addi-
tion, increasing the oxygen content would slightly decrease the nitric oxide
concentration which has been shown to have far greater importance.
Burdick also found that the water content of the gas has an effect on
the reaction rate. Very dry gases react about ten times as fast as mixtures
with small amounts of water vapor, but beyond this small amount increasing
water content has no effect. It is clear, therefore, that this factor, too, can
be ignored since it would be impossible to completely dry the gas before the
reaction took place. The variables that should be examined, therefore, are
reaction temperature and NO concentration.
Figure 12 is a plot of the equilibrium constant, Kp' in the nitric
oxide oxidation reaction as a function of temperature, calculated from hand-
book free energy data. It points out the somewhat unusual inverse tempera-
ture relation where the reaction rate decreases as the temperature increases.
It is to be noted that below 300°F (425°K), the temperature of the flue gas
entering the process, K exceeds 107. This indicates that at this condition
p
the equilibrium is highly favorable to oxidation of NO to N02' This was
verified experimentally as shown in Fig. 13 (see below for experimental
technique). There is obviously a decided advantage to keeping the NO oxi-
dation temperature well below 300°F.
A study of the external reoxidation of NO to N02 was made using
the small-scale continuous kinetic apparatus shown schematically in Fig. 14.
The same spectroscopic method of determining N02 concentrations as used
before was employed, but with a cell of much shorter path length. Calibra,..
tion curves at 5100 R showed that at this path length the UV absorption method
was reproducible and sensitive up to ca 300 torr. The effect of initial NO
concentration was evaluated with the dependent variable being the rate of con-
version of NO to N02' The same procedures were used to determine the
effect of N02 concentration and temperature as given in FigB. 12 and 13.
As shown in Fig. 15, the rate of oxidation of NO to N02 can be
markedly increased when the reaction is catalyzed with coconut charcoal.
- 44 -
-------
a.
~ 8
aa
o
10
6
4
2
300
400
500
TOK
600
Fig. 12. Equilibrium constant in the reaction 2 NO + 02 .= 2 N02
vs. temperature. (Stoichiometric reactant concentrations,
Tatm. total pressure. ) .
.. 45 -
-------
20
10
REACTION
TIME,
Seconds
5
3
10
INITIAL COMPOSITION OF GAS:
02 -100 mm Hg
NO-150 mm Hg
N02- NONE
N2 - BALANCE
20
%CONVERSION
50
NO- N02
JOO
Fig. 13.
Conversion rate of NO to N02 as a function
of temperature.
- 46 -
-------
FLOWMETERS
NO
CYLINDER
COMPRESSED
AIR
CYLINDER
WATER BUBBLER
SATURATOR
(GO.C)
N2
CYLINDER
3-WAY
MIXING STOPCOCK
FIXED BED REACTOR
(empty volume 25 mL)
TO EXHAUST
N02
CYLINDER
Pig...14: Apparatus for Continuous Oxidation of NO to N02
- 47 -
-------
1.0
FRACTIONAL NO
CONVERSION TO N 02
BLANK (no catalyst)
(II seconds' retention time)
t 0.1
0.01
10 100
TIME (min) FROM START OF GAS FLOW
Fig. 15. Rate of Oxidation of NO to N02 using Charcoal Catalysis
CHARCOAL BED IN SAME
REACTOR AT SAME GAS
FLOW RATE.
(II seconds' retention based
on empty reactor.)
8.4 mm Hg NO'V(l%)
20 mm Hg 02 'V (21f2Cfo)
BALANCE N2
COCONUT CHARCOAL
CATALYST
1000
-------
As presented in Figs. 18 aDd 17, at an initial concentration of 105 mm 02'
J35'
-t5& mm NO , 355 mm N2 ' and 150 mm H20, 85% oxidation of the NO to
N02 was accomplished in 9 seconds' retention.
The results of the batch reactor study previously described indicate
that with contact time of 1 - 2 seconds at least 90% of the S02 in flue gas can
be converted to H2S04. The longer times needed to reconvert low concentra-
tions of NO to N02 can be avoided by carrying out this step separately at
higher concentrations and then feeding this gas (as a 3: 1 molar N02/NO
mixtUre) into the flue gas stream. After S02 oxidation, a 1: 1 NO/N02
ratio will exist. This mixtUre (N203) can then be scrubbed, tOgether with
the sulfuric acid mist, from the flue gas stream in a packed bed collector unit.
The N203 would then be recovered from the acid, thus creating a concentrated
NO mixture which could be easily oxidized to N02 before being recycled.
This scheme for separate NO oxidation will be shown in a later sectiondes-
cribing the second generation Modified Chamber Process for S02 Removal
from Power Plant Stack Gases.
F.
Collection of Nitrogen Oxides and Acid Mist
The vapor phase catalytic process for S02 oxidation with N02
depends for its effectiveness on efficient (> 99.9%) collection of particulate
matter and nitrogen oxide vapors. This was to be done using a stream of
recycled sulfuric acid as the collection medium in a packed bed collection
apparatus surmounted by a fiberglas mist collector of commercial design. To
evaluate this aspect of the process, the apparatus shown in Fig. 18 was con-
structed. A laboratory-scale high efficiency flat bed mist collection unit and
its associated feed tank were procured from Monsanto Co., manufactUrers
of commercial-scale fiberglas bed mist collection equipment. The unit, which
has a gas tlow capacity of about 1 cfm, was fed as shown, and its collection
efficiency for both nitrogen oxides and particulate matter determined. The
variables studied included the effects of gas velocity, pressure drop,. gas
composition and acid recycle rate.
- 49 -
-------
120
INITIAL 02=105 mmHg
INITIAL N2 = BALANCE
110 T = 25°C.
A RETENTION TIME =9 sees.
% N02 INITIALLY
100 SEE FIG. 18 FOR
APPARATUS
90
rr>
80 I
o
CI 15 ><
I
E 70 CI
E I
o (\I
a..Z 60~ 0 E
E
....J ~
-------
RETENTION TIME =9 SECONDS
220 % N02 INITIALLY.
PRESATURATOR = 60GC
REACTOR = 60GC
200 INITIAL P 02=105mm.Hg 20D
INITIAL N2 = BALANCE
(355 mmHg)
180 INITIAL H20=150mmHg
SEE FIG. 18 FOR
APPARATUS
160 rt)
A b
15.0 )(
140 01
J:
01 NE
I E
el20
E NO
o Q..Z
z
0.100 ....J
10-0 <[
..J I-
-------
0.8 ml./min.
WATER
RESERVOI R
400ml. BEAKER
11z. DIA. HOSE OR PIPE
1/2" DIA. INSIDE
FLOWMETER
0.3-3cc/min
WATER
f\.DWMETtR
0.2-2 ct.
WRAP WITH HEATING
TAPE -100 WATTS MINIMUM.
SLANT DOWNWARD
SLIGHTLY FOR WATER
TO RUN THROUGH.
REACTION
CHAMBER
81 FLASK
OOW "[!SED
AIR TANK
(300 FT.')
92.1.1... N20,
880 880 1.1 cfm ~os
..Lilli in. me.llnin
FLOWMETER
"-0--, U'I o.3-3Jlmin.
(!IO~
eel...
NEEDLE
V-.vES 25-250cclmin
MDUCER FLOWMETER
SHUTOff FIBERGLASS
MIST ELIMINATOR
UNIT
SOt NO N02 MANOMETER
cn. cn. CYL. ACID + N203
DRAINAGE
(1Ib.) (Stb.) (~Ib.)
4..
BEAKER
Fig. 18. Mist collection study apparatus flow diagram
- 52 -
I
-------
After calibration of the flowmeter and UV cell, using a 6-inch path
length cell, several experiments were made using this apparatus. In two
initial runs, 20 cfh each of NO and N02 ' 1. 5 cfm of air, 90 ccfmin of
S02 and 0.3 ccfmin of liquid water were introduced into the flow reactor.
The mist collector had H2S04 flowing through the bed at a rate of 60 ccfmin.
For a pressure drop of 35 inches of water, an N02 removal efficiency
of 91. 3% was found. Reduction of the pressure drop to 20 inches of water re-
duced the efficiency of N02 collection to 76%. Removal of the S02 was
found to be essentially 100% in both cases.
Within the reactor, formation of a small amount of white crystalline
material was noted. From its properties this material appears to be nitro-
sylsulfuric acid (S05NH).
A second series of runs was conducted. For these, the following flow
rates were used in all cases:
Air
N02
NO
S02
H20
21,600 ccfmin of gas (0.75 cfm)
278 ccfmin of gas (1. 3 volume %)
62 ccfmin of gas (0.3 volume %)
130 ccfmin of gas (0.6 volume %)
0.3 ccfmin of liquid (1. 7 volume %)
The H2S04 flow rate through the mist collector was then varied, with values
of 10 ccfmin to 180 ccfmin being studied. The results for N02 collection
efficiency and pressure drop as a function of acid (70% H2S04) flow rate are
shown in Fig. 19~ Fresh acid was used for each determination. For the
same gas flow rates, the pressure drop across the mist collector bed tended to
increase as the acid flow rate increased.
From the above results, it can be seen that at higher acid flow rates,
flooding of the pores of the packing apparently reduces the amount of gas
liquid interfacial area present at which absorption of the N02 can occur.
At lower flow rates, however, insufficient acid is probably moving through
the bed to take up all of the N02'
These preliminary results show that even with only a single absorp-
tion stage, N 203 is scrubbed from the gas stream at high efficiency. While
a countercurrent column scrubber is evidently required to attain the necessary
99.9+% N203 collection efficiency, the number of stages needed should not
be excessive.
- 53 -
-------
100 .....--.~
90 / "".
80
en 70
c::x:
t!:>
.......
~6
z
~
050
a::
IJ...
rt)
o
~40
IJ...
o
~30
>
o
~
w
a:: 20 0 0
~
o
10
o
/
."" //
~<
o 0
100
200
H2S04 FLOW IN FIBERGLASS BED, mIl min
Fig. 19. Collection efficiency for N203 and pressure drop across
fiber mist collector unit.
- 54 -
4.0
3.0
CD
I'T1
o
1)
::0
I'T1
CJ)
CJ)
2.0 c
::0
I'T1
o
::0
o
1)
:::J
J:
N
1.0 0
o
8
-------
G.
Solubility of NO , N02 and Fly Ash in H 2S0 4
Data were collected on the solubility of NO and N02 in sul-
furic acid as a function of acid concentration and temperature. Since the
process depends for its effectiveness on efficient removal of oxides of nitro-
gen by scrubbing the effluent gas stream with recycled sulfuric acid, this
solubility information is essential.
The solubility of NO in 74% sulfuric acid containing dissolved N203
is given in Table X below.
TABLE X So lubi lity of NO in 74% H 2S0 4 Containing SO NH(2)
5
NO Absorbed
No
Solvent mljml Solvent Solvent, N02' Total NO, -C0/N02 ratio,
g/l.. g/l', ,g/l, mol/mol,
% N 203 40°C 70°C 40°(; 40°C 40°C 40°C
1. 18 1. 47 1. 61 1. 72 11. 7 9.4 1. 23
4.3 6.40 6.20 7.48 42.6 35.4 1. 27
1.11 2.85 2.93 3.33 11.0 10.5 1. 46
3.86 10.10 10.20 11. 8' 38.2 36.8 1. 48
These data are shown graphically in Fig. 20 and are compared with
a theoretical line for pure nitrosylsulfuric acid. The curve from "the data
in Table X shows that the nitric oxide is not sufficiently soluble in the nitrose
to allow the use of a large excess of NO relative to N02 in the reaction
chamber. As shown in Fig. 21, the solubility of N203 in sulfuric acid is
much higher than the solubility of NO. This increased solubility occurs because
an acid'-soluble compound, nitrosylsulfuric acid, is forqIed which is obviously
essential for efficient nitrogen oxide recovery.
.
- 55-
-------
50
45
" a
~
~
40 ~
&-
.f a
-5
o;t 0'"
0 35 +'"
(/)
N
::t:
~ /
0
~ 30
I'-
CIa(
......
CI
0 25
z
20
15
10
10
15
..0
Fig. 20. Solubility of NO in 74% H2S04 containing N02' 40°C.
- 56 -
-------
50
v
o
en
(\J
::I:
C'
o
o 20
......
an
o
en
z
::I:
C' 10
an
o
en
z
::I:
1L. 5
o
>-
r-
...J
CD
:::>
...J
o 2
en
100
/'
o
20
30 40 50 60 70 80
ACt 0 STRENGTH, W % H2S04
10
90
100
Fig. 21. Solubility of nitrosyl sulfuric acid in H2S04 (6)
,- 57 -
-------
Since this process envisions oxidation of S02 to sulfuric acid in
the presence of fly ash, it was necessary to determine the extent of con-
tamination of the sulfuric acid produced by contacting with fly ash. Samples
of fly ash from a coal-burning power plant (New England Electric Co., Salem,
Mass.) were obtained and contacted with sulfuric acid as shown in Table XI
below. The amount of ash dissolved may be controlled by reducing the
contact time before' ash solids filtration.
*
TABLE XI Contamination of Sulfuric Acid by Fly Ash
Contact Acid 10f Acid Analysis, %t
Time Temp. Concentration sh
Min. DC % H2S04 Dissolved Fe Al Si Ca
120 25 30 4.95 .010 .018 .005 nil
30 25 30 5.97
120 25 70 8.86 .010 .023 .012 nil
30 25 70 7.51
120 75 30 20.6 .051 .015 .016 trace
30 75 30 9.3
120 75 70 16.1 .018 .008 .010 nil
30 75 70 10.2
* All.e~~eriments were performed using reagent grade sulfuric acid, with
an InItIal ash concentration of 0.5% wt equivalent to that expected in an
actual process.
t ASTM analytical methods used.
- 58 -
-------
H. The Modified Chamber Process
1) Conceptual Change. The experimentation described in Sections
D-G of Part V of this report indicates that the original Chamber Process
concept for flue gas clean-up as presented in Section V -B cannot be used
without modification. The primary drawback to the basic Lead Chamber Process
is the difficulty in reoxidizing the NO formed during the S02 oxidation back
to N02 for recovery and recycle. The long residence times required to
reoxidize the very dilute NO would necessitate the use of impossibly enormous
reaction chambers, making the process impractical.
As discussed in Section V - E, the way to avoid this extended reaction
time is to collect the nitrogen oxides in sulfuric acid, concentrate them in a
separate gas stream, and then oxidize the nitric oxide content before recycling
the nitrogen dioxide to the reactor. This higWy significant conceptual ch,ange
makes it technically possible to use Chamber Process technology for removing
sulfur dioxide from power plant flue gas and is incorporated in the process
scheme shown in Fig. 22 and now called the Modified Chamber Process.
Complete heat and material balances were necessary before an economic
analysis could be made to determine if the modified process is economically
feasible as well.
2) S02 Reaction Chamber. In the Modified Chamber Process the
flue gas at 300°F is fed to the reactor chamber where it is mixed with nitro-
gEm oxides so that the net mole r.at~o before S02 oxidation is N02: S02: NO =
3:1:1. (See the total material balance, Table XII for details.) The purpOse
of using this mole ratio is to provide an excess of N02 in the reaction mixture
to make sure that a maximum amount of sulfur dioxide is oxidized and still
have an equimolar mixture of NO and N02 when the reacted gas enters the
scrubber:
3 N02 + 1 NO + 1 S02 ~ 2 N02 + 2 NO + S03
- 59 -
-------
TO STACK
flue gas
3000 F
0.3% S02
600 ppm NO
x
(j)
o
3300 F
REACTOR
H20
COOLER
HNO
ABS6RBER
NO
OXIDIZER
Product
HN03 (52%)
stripper
gas
21000 F
Fig. 22. Modified chamber process
H2S04 80% wt 800 F
1500 F
SCRUBBER
H2S04 76%
HNS05
275 0 F
N203
STR IPPER
H2 SO 4
78%
clean
gas
21000 F
3400
ACID
COOLER
Product
H2S04
FILTER
ACID
CONCENTRATOR
-------
Table XI1. Material Balance - Modified Chamber Process
(Basis ~ 1000 lb Flue Gas)
Stage C02 Jh2- ~02 ~2 ~2 J:!2S04 Ash NOx N02 NO Total T, of
Flue gas to process 215 44 6.4 30 702 2.0 0.6 1000. 0 300
Gas from cyclone 215 44 6.4 30 702 O. 8 .03 .57 998. 8 300
Gas from oxidizer 45. I 24.5 0 3.9 147 0 13. 77 2.43 236. 7 300
Gas from N203 mixer 260 66. 7 0 32.3 849 9. 8 O. 8 9.20 6.00 1233. 8 340
Recycle acid to 300 1200 0 0 0 1500 80
collector (80%H2S04)
Acid from collector 362.6 1209. 8 O. 8 9.20 6.00 1588.4 285
(76% H2S04)
Gas to stack 260 3.4 0 32.3 849 0 0 .003 .002 1144.7 150
Acid from stripper 347.0 0 1211.8 1.2 0 0 0 1560.0 375
(78% H2S04)
0) Gas to stripper 45. 1 9.2 1.3 6.3 147 0.4 O. 13 - - 210 2100
~ 45. 1 24.5 0 6.0 147 0 0 9.28 6.05 237.9 300
Vapor from s tri pper
A cid from 303 0 0 0 1211. 6 0 0 0 0 1515 395
concentrator (80% H2S04)
Fly ash cake 0.3 0 0 0 0.9 1.2 0 0 0 2.4 80
Product acid (80%H2S0 4) 2.7 10. 7 13.4 80
Gas to concentrator 39 8.0 0 5.4 128 O. 4 O. 1 181 2100
Exhaust from 39 52 0 5.4 128 0.4 O. 1 225 400
cuncentrator
Water to HN03 1.7 1.7 70
absorber
Vapor from HN03 45.1 24.5 0 6.0 147 0 0 0 7.7 6.4 222.6 300
absorber
Product HN03 (52%) 0 1. 45 - - 1. 45 2. 9 100
HN03
Water to acid 2240 2240 70
cooler (no waste
hea t boiler)
-------
The large heat of conversion of S02 to sulfuric acid does not
result in a high temperature rise in the flue gas stream because the
502 concentration is at a low level. The reactor hea t dissipa tion pro-
visions necessary in the original chamber sulfuric acid process are
unnecessary in this process, and the heat of reaction is removed by
recirculated acid in the s c rub be r . The heat of reaction of 502
and dilution to 80% H2S04 is about 1350 BTU/ lb 100% H2S04' Porma-
tion of 9.8 Ibs. H2S04/ 1000 lb. flue gas, for example, as shown in
Table I, corresponds to a gas tempera ture rise of
9.8 (1350) BTU
0.241 BTU x 1234 lb.
Ibop
, or 44 ° F ,
and part of this energy will be lost by radiation and convection to the
surroundings. Also, since reoxidation of the NO formed in the reaction
to N02 is performed externally to the main flue gas stream, higher
temperatures in the flue gas do not adversely affect the rate, as they do
in the conventional chamber process. (The rate of oxidation of NO to
N02 decreases as the temperature is increased, while that of S02 to
H2S04 increases as the temperature is increased. )
3) Scrubber. In the packed bed scrubber the reacted gases are
contacted with a countercurrent stream of cool, clean sulfuric acid which
scrubs out the sulfuric acid mist formed in the reaction chamber along
with the equimolar mixture of nitrogen oxides, the fly ash, and the water
vapor in the gas. The clean, dry stack gas can now be vented to the at-
mosphere having been cooled to about 150oP. This will be heated for
better dispersion as will be shown later.
Effective removal of both sulfuric acid and oxides of nitrogen from
the gas stream is an essential part of this process. Vapor pressure data
indicate that with an efficient gas scrubbing system, negligibly small amount~
of H2S0 ~ or N203 will escape to the atmosphere. Both Ber14 and Yush-
manov 1 have measured the vapor pressure of N203 and water over sulfuric
acid of various nitrose concentrations and temperatures. Pig. 23 and 24
illustrate these data for 80% wt H2S04 solutions.
- 62 -
-------
1000
100
CI
J:
E
E
~
lJJ
a::
::>
en
en
lJJ
a::
a..
a::
o
a..
~
10
Fig. 23.
20
.
H20 over :80% H2S04
N203 concentration = O.213g/1 H2S04
Temperature, 0 C
140 120 100 90
I 1 I 1
1.00
0.10
0.01
30
1
0.001
3.4
Vapor pressure of N203 and H20 over 80% H2S04 solutions,
Yushmanov.13
- 63 -
-------
01
:J:
E
E
100
..
CD
~
~
fn
fn
CD
~
~
..
o
Q.
o
>
0\
0\
o
'\
o
\
o
\0
'\0
- total vapor pressure (H20 + N203)
over 80% H2S 04
(0.10 moles HNS05/ I H2S04 or
3.6 N203/1 H2S04)
o
10
\
o
\0
\0
~o-- 0
o 0-... \
°A
Nz03 Vapor Pressure Only / 0", 0
0\0
Temperature ° F
300
250
150
200
o
1.8
1.9
0.1
1.3
1.4
1.6
103/ T, oR-I
1.1
1.5
Fig. 24. Nitrose vapor pressures over 80% wt H2S04 solutions
(Ber 1 and Saenger) 4
- 64 -
-------
After our experimental experience and an additional literature search,
it was seen that the fiberglas packed mist eliminator would not be a practical
NOx scrubber because of the excessively high pressure drops through this kind
of device. Laboratory experimentation with the small fiberglas packed absorber
yielded pressure drops up to 36 inches of water, obviously too high to be useful.
To overcome this resistance huge draft fans would be necessary in a 800 MW
power plant handling almost 1.5 million cubic feet per minute of flue gas. There
is clearly an advantage to minimizing the pressure drop through the scrubber.
To attain lower resistance through this column it was decided to use more con-
ventional packings: both rings and saddles offer lower pressure drop than the
fiberglas packing. The primary advantage of the fiberglas, that of removing
acid mists, is lost with conventional packings, but it is not clear at this stage
of development how serious the mist problem will be. It was decided to go with
standard ceramic packing and evaluate the mist problem at a later stage of
development.
4) Nitrogen Oxide Recovery. The nitrose leaving the scrubber enters
a packed bed stripper column where it is treated by a countercurrent stream
of hot (21000 F) flue gas. The nitrogen oxides are vaporized and pass out of
the column with the carrier gas, having been concentrated by a factor of five
compared to the original flue gas stream. This is sufficient to allow separate
reoxidation in a reasonable volume before being recycled to the reaction chamber.
If this absorption and desorption of nitrogen oxides is performed efficiently
enough, it is possible to reap a considerable economic benefit at this stage in
the process. If the oxides are recovered at an efficiency of 96% or higher there
will be a surplus of N203 above recycle needs because of the recovery of
nitrogen oxides present in the raw flue gas. If part of this recycle stream is
cooled and scrubbed with water, this surplus can be converted to marketable
52% nitric acid:
3 N02 + H2 ° - 2 HN03 + NO
The higher the nitrogen oxide content of the flue gas, the more nitric acid can
be made and the greater the economic advantage.
- 65 -
-------
5) Acid Concentrator. The acid which has been denitrated in
the stripper is recycled to the scrubber, but first it must be concentrated
to remove the water that was absorbed from the wet flue gas. This excess
water is sufficient to dilute the acid from 80% in the incoming scrubber
stream to 76% in the effluent. Some water is removed in the stripper
during the vaporization of the nitrogen oxides, but it is still necessary to
strengthen the acid from 78% back to 80%.
In the conventional Chamber Process the nitrogen oxide vaporization
and the concentration are performed in the same stage: the Glover Tower.
This can be done because the process is water deficient and the water that
is stripped from the nitrose can be used in the chambers in the production
of sulfuric acid. In the Modified Chamber Process for flue gas treatment,
however, there is a large excess of water and if all the absorbed water is
vaporized in the stripper and recycled to the reactor, there will be a steady
buildup of water in the system. This is avoided by recycling only part of the wate
by vaporization in the stripper and to remove the remainder from the system
completely in a separate acid concentrating step. Under steady state conditions,
all the excess water that is absorbed by the H2S04 is removed by boiling the
stripper effluent in an acid concentratOr.
The acid concentration step is accomplished by passing enough hot,
clean (21000P) flue gas countercurrent to the denitrated stripper acid to boil
the acid and venting the wet gas to the atmosphere. This effluent gas stream is
mixed with the stack gas from the scrubber tower, thus heating the scrubber
gas sufficiently to provide adequate buoyancy for proper atmospheric dis-
persion. Since the gas is vented directly to the atmosphere, it must be
clean enough not to be an air pollutant itself. This means that the source
of the concentrator gas must be a furnace burning sulfur -free fuel, such
as natural gas or desulfurized coal. The requirement for a "clean" fuel
places an additional economic burden on the process, which will be evaluated
in a later section.
6) Ply Ash Piltration. The concentrated acid is cooled from 395 of
(the boiling point of 80% sulfuric acid) to 800p and is recycled to the scrubber.
If the system is scrubbing fly ash out of the flue gas, then the acid must be
- 66 -
-------
filtered before recycling; the produce acid is removed from the stream
after filtration. In order to avoid filtering the entire recycle acid stream,
an alternative approach would be to allow the fly ash concentration in the acid
to build up so that the ash removed with the product acid per unit time would
be equal to the ash introduced with the flue gas in the same amount of time,
and only the product acid need be filtered.
7) Heat Balance. Using the total material balance shown in Table XII,
a complete heat balance can now be calculated for the Modified Chamber Process
as described in this section. The heat balance on the same basis as the material
balance is shown in Table XIII, while Table XIV presents the balance for an
800 megawatt power plant installation (1.44 x 106 SCFM of flue gas).
8) Economic Analysis. At this stage in the discussion it is now possible
to perform a preliminary economic analysis for the Modified Chamber Process
based on conceptual changes described in this section and the heat and material
balances given in Tables XII, XIII, and XIV. The assumptions made in this
analysis are listed in Table XV.
Table XVI shows a breakdown of the equipment type and size required
for the Modified Chamber Process together with its estimated delivered cost.
Using the same 3.29 factor as in Table IV, this data is then used to develop
an estimate of the total capital cost as shown at the bottom of Table XVI. Based
on the total capital cost and general guidelines for plant operation, the annual
operating cost was developed in Table XVII. Taking credits of $20/ton (100%
basis) for 80% sulfuric acid and $78/ton (100% basis) for 52% nitric acid, an
income statement was calculated in Table XVIII showing a gross loss of $151,000
which corresponds to a loss $0.07/ton of coal. These figures show that the
Modified Chamber Process will break even.
9) Overall Evaluation. The Modified Chamber Process appeared to
to be economically feasible, but closer examination of the process revealed
that there was some question as to the practicality of one significant operational
aspect: the extra heat requirements in the stripper and the concentrator.
- 67 -
-------
Table XIII. Heat Balance
Modified Chamber Process
Heat Inputs (Basis: 1000 lb. entering flue gas):
Cooling stac~ gas, 300 -> 150° F: 1190 lb. x
0.241 BTU x 150°F =
lb. of
Cooling stripper gas, 2100 -> 300° F:
224 lb. x . 241 x 1800° =
Condensing water in recycle acid:
(63 lb. - 15. 3 lb.) x 1357 BTU
lb.
=
Heat of reaction, S02 -> 80% H2S04:
11. 6 lb. H2S04 x 1350 BTU =
lb.
Cooling concentrator gas, 2100 -> 400°F:
203 lb. x .241 x 1700° =
Total heat input per 1000 lb. flue gas =
Heat Outputs:
Cooling 80% H 2S0 4 :
0.47 BTU x (3950
lb. of
1515 lb. x
80° F) =
Water vaporized in concentrator:
44 lb. x 1357 BTU
lb.
=
Unaccounted for (6.6%) =
-~-
43,000 BTU
97,200 BTU
64, 700 BTU
15, 650 BTU
83, 000 BTU
303, 650 BTU
224,000 BTU
59, 000 BTU
20,050 BTU
303, 650 BTU
-------
Table XIV.
Heat Balance: Modified Chamber Process for
800 Megawatt Power Plant Installation (7.25
MM lbj hr entering flue gas) (38. 8 tons per
hour 100% H2S04 produced)
Input
Cooling stack gas:
312 MM BTUjHr
Output
Cooling stripper gas:
704 MM
Cooling water: 1717 MM
Condensing water in
recycle acid:
469 MM
Water vaporized in concentrator:
484 MM
Cooling concentrator gas:
603 MM
Total: 2201 MM BTUj Hr
2201 MM
- 69 -
-------
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
( 16)
Table xv.
Assumptions for Economic Analysis
280 tons of coal burned an hour, flue gas composition
as given in Table 1.
Flue gas rate 1.44 X 106 SCFM from an 800-megawatt
power plant.
Two percent of the sulfur in the coal burns to S03' the
remainder to S02.
Ash content of the flue gas 2 lb per 1000 lb.
Yearly maintenance cost (labor, supervision, and
material) equal to three percent of depreciation base.
Plant facilities are charged at five percent of the cost of
units, and plant utilities-seven percent of cost of units
plus plant facilities.
A total charge on investment of 14 percent annually is
made to include taxes, insurance, and return on
investment.
S02 is quantitatively converted to H2S04.
Gas residence time in S02-oxidation chamber is two seconds.
Gas residence time in reoxidation chamber is ten seconds.
Recovered N203 from stripper contains lIb H20 per lb
N203.
Pressure drop: twelve inches of water over entire system.
802 reactor temperature: 3000 F.
Sulfuric acid (80%) is sold for $16 per ton
Scrubber efficiency for N 203 and H 2S0 4 :99. 9% .
Operating life of packing: two years.
- 70 -
-------
Table XVI. Estimated Equipment Cost for Modified Chamber Process
(800 Megawatt Power Plant Installation)
Estimated
Item No. Required Duty Size Purchase Cost, $
Cyclone separator 3 Remove coarse fly ash 487,000 cfm 245,000
from flue gas
Re-oxidation 1 Provide retention time 183,000 cfm, 70,000
chamber for re -oxida tion of 10 sec. gas retention
NO to N02 30 ft. qia. x 43 ft.
height
Scrubber unit 3 A~sorb N~3' acid 35 ft. dia x 70 ft. 1,500,000
mIst and as height. 8 ft.j sec
gas velocity
Stripping unit 2 Remove N203 for 32 ft. dia x 60 ft. 250,000
recycle from acid height. 2 ftj sec
-:J and water vapor velocity
.....
Coal furnace for 1 Provide hot gas for 704 MM BTU! hr 80,000
stripper stripper heat re-
quirement
HN03 absorber 1 Convert part of N02 44 ft. dia x 60 ft. 450,000
to HN03 and remove height 183,000
as 52% HN03 cfm 2 ft/ sec gas
veloci ty
Acid concentrator 3 Concentrate HfiS04 360, 000 lbj hr 300,000
to 80% strengt H20 evap
-------
Table XVI(Cont.)
E s tim a ted
Item No. Required Duty Size Purchase Cost, $
Furnace for 3 Provide hot gas for 603 MM BTUjhr 150,000
concentrator concentrator
Filter 3 11, 600 lbjl1..Y cake Rotary drum unit 250,000
Acid cool~r 3 Cool c'oncentrated Shell and tube ex - 320,000
acid for recycle changers
21, 000 ftZ area
each, 400 MM
BTU! hr. each
Product storage 3 Store cold 80% 500, 000 gal. total 100,000
tank H2S04 volume
Exhaust fan 3 Provide differential 12 in. water suction 60,000
pressure for gas 87 MM SCFH total
-;:J flows flow
t.:)
Total Equipment Cost (Purchased) $3,855,000
Total Capital Cost (Based on Factor of 3.29 Developed in Table IV) $12,683,000
-------
Table XVI1. Annual Operating Cost Estimate: Modified
Chamber Process (800 Megawatt Power
Plant Installation)
Packing replacement
Water, 23,600 GPM at $00 10/M gal
Direct labor
Supervision
Maintenance (0.05 of fixed capital)
Supplies at 15% of maintenance
Power, 24.2 MM Kw hrs at $0.006/ Kwhr
Heat, 1307 MM BTU/ hr at $0.50/ MM BTU
$ 75,000
1,133,000
72,300
10,800
634,200
95,000
145,000
5,220,000
Total direct cost
$ 7, 385, 400
Payroll burden (20% of direct labor and
supervision)
Plant overhead (50% of direct labor and
supervision, maintenance and supplies)
Packing and shipping
Waste disposal
16,600
406,200
400,000
25,000
Total indirect cost
$ 847, 800
Deprecia tion (10% of fixed capi tal) *
Taxes (2% of fixed capital)
Insurance (1% of fixed capital)
$ 1,268,300
253,700
126,800
Total fixed cost
$ 1, 648, 800
Total Operating Cost
$ 9,882,000
* Assumes 10 yr straight-line depreciation.
- 73 -
-------
Table XVIII. Return on Investment: Modified Chamber
Process (800 Megawatt Power Plant
Installation)
Credits
80% Sulfuric Acid, 38.8 tons/ hr (100%) at
$20!ton
52% Nitric Acid, 5.2 tons! hr (100%) at
$~4! ton
$ 6, 201, 000
3,530,000
Total Credits
Annual Operating Cost (Table XVI)
$ 9, 731, 000
9,882,000
Gross Profit
(151,000)
Loss on investment before taxes
151,000
12,683,000
;::: 1%
151,000 = $0 07
Cost $! ton coal 280 x 8000 .
- 74 -
-------
Calculations show that these two stages require about 37% more heat
than the power plant generates, implying that an additional 37% more
coal would have to be burned, about half of which must be sulfur-free.
It is true that the economic analysis showed that this could be done and
still break even, but it was felt that there would be considerable resis-
tance in the power industry to putting in enough furnace capacity to con-
sume 37% more fuel than was currently being burned. Efforts would have
to be made to further modify the process to limit the extra heat require-
ments.
It was therefore decided to continue work on the experimental
verification of the Modified Chamber Process while concurrently working
on new schemes to reduce heat requirements. The remainder of this
section of the report will discuss in turn: 1) the laboratory work aimed
at establishing the feasibility of the Modified Chamber Process and 2) the
conceptual and laboratory work intended to reduce the heat requirements
of the process thereby improving both the economics and commercial
practicability .
I. Laboratory Evaluation of the Modified Chamber Process
Initially it was decided to divide the laboratory feasibility studies
for the evaluation of the Modified Chamber Process into two phases: a
1 SCFM bench scale evaluation of individual process stages, and a 10
SCFM pilot plant analysis of an integrated system. Initial work with a
bench scale scrubber system demonstrated that this scale of study would not
be practical because the efforts necessary to optimize the individual process
stages would take far more time than they merited and would unreasonably delay
the scale-up to the pilot plant. It was felt that the optimization of the
individual stages of the process could be accomplished just as rapidly
and far more meaningfully in the pilot plant stage than on a bench scale.
The bench scale scrubber experiments did have appreciable value
although they were not carried through to complete optimization. Using a
3 foot column, 2 inches in diameter, packed with 1/4 inch Bed saddles,
- 75 -
-------
absorption studies were carried out by countercurrent scrubbing of an
artificial flue gas (N02:NO:S02::3:1:1 in a nitrogen carrier gas) with
80%, 80 of sulfuric acid at an L/ G ratio of 1. 03. Absorption efficiencies
of about 90% for both S02 and N02 were determined by ultraviolet spectro-
photometric analysis of the effluent gas stream although maximum absorp-
tion was not obtained for both gases at the same time. Typical data for
these experiments are shown in Figure 25 with the packed column described
above being used as a scrubber. The NO and N02 flow rates were held
constant at a 3:1 ratio while the S02 constant of the gas mixture was
varied from a N02/ S02 ratio of about 3. 0 to 1. 5.
The data in Figure 25 show high S02 removal efficiencies at high
N02 excesses, while the N02 removal goes toward a maximum at low N02
excesses. This suggests that the N02 oxidation of S02 is not going to
completion in the reactor and for the most part the N02! NO ratio is not
near enough to equality to allow complete absorption in the scrubber.
These experiments, which show the difficulty in obtaining N02
removal efficiencies approaching 100%, have prompted a closer look into
the mechanism of the adsorption of NO and N02 into sulfuric acid. An
extensive literature search has resulted in the compilation of a large
amount of information on the physical and chemical properties of nitrosyl
sulfuric acid, but more data was needed in order to determine what was
occurring in our system.
The experiment was set up such that a mixture of NO and N02 in
n~trogen was passed through the reaction chamber and into the scrubber,
with the scrubber effluent gas being monitored for N02 by means of the UV
spectrophotometer. The two conditions that were varied were the ratio of
NO to N02 and the gas flow rate. The sulfuric acid flow rate was held
constant, as was the N02 concentration in the nitrogen carrier gas (0.9%).
Figure 26 shows the N02 removal efficiency as a function of NO! N02 ratio
at different contact times.
Two features of the data are obvious: that N02 removal efficiency
increases with both NO!N02 ratio and with contact time.
- 76 -
-------
90
-
?fl 80
-:J
-:J
>-
u
z
w
U
IJ....
IJ....
W
...J
~ 70
w
a::
60
50
o
o
o
o
o
o
N02/S02 RATIO
3.0
2.0
75
100
125
S02 GAS FLOW RATE cc / min
o
N02 FLOW = 210 cc Imin
N02/NO = 3/1
80% H2S04 FLOW =30cc/min
SCRUBBER:
2 in. DIAMETER
40 in. PACKING HEIGHT
1/4 in. INTALOX SADDLES
PACKING
o
1.5
150
Fig. 25. S02 and N02 recovery efficiency in a packed bed scrubber
-------
>.
0
c 90
CD
0 0
-
-
UJ 0
c 80
>
0
E
CD
a:: 0
(\j
0
z 70
0
0
0
60
Reactor:
1.5 in. diameter
18 in. length
50
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.C
Mole Ratio NO / N 02
o ~ 0-0-0-
-yo
.,.......0
o
o
0/
o
Condit ions:
Constant H2S04 flow (80% acid)
(21cc/min)
Constant NO 2 con cent rat ion
(O.9%)
Balance N2
100 -0
o
o
Contact time:
.
o 6.6 sec
o 3.3 sec
o LOsec
Scr ubber:
2 In. diameter
40 In. pa ckl n g height
1/4 in.lntalox sad dies
packing
Fig. 26. N02 removal efficiency as a function of NO concentration
and contact time
- 78 -
-------
The fact that the N02 removal efficiency improves with increasing
NO content (in the lower two curves) indicates that NO is being removed
2
as HNS05 (nitrosylsulfuric acid), at least in part. The fact that the gas
is not completely' removed at an NO/N02 ratio of 1.0 (the stoichiometric
ratio for HNS05 formation) suggests that the absorption reaction did not go
to completion, either because the gas-liquid contact time was insufficient
for complete reaction, or the equilibrium at the timperatUre used did not
allow it.
The increase in N02 removal with increased gas-liquid contact
time may be due to two effects: additional formation of HNS05 or dissolu-
tion of N02 in sulfuric acid. The fact that 100% of the N02 was removed
at NO/N02 ratios below 1.0 (top curve of Fig. 26) suggests that N02 is
soluble eno\,lgh to be absorbed alone in sulfuric acid. Experimentation
has shown that 25 - 55 % removal of N02 at 0.0 NO/N02 can be achieved
with the bench-scale scrubber, with the exact level being dependent on gas-
liquid contact time. Comparison of this data with published literature shows
agreement and points out the possibility of removing all oxides of nitrogen
as long as the NO/N02 ratio is maintained at:::; 1.0.
The work described above was very useful in determining opera-
ting information which c.:>uld be of value in pilot plant operation: It
established the importance of controlling the NO/N02 ratio of the reacted
gas stream entering the scrubber, and showed that consideration must
be given to such matters as accurate flow meter calibration, gas and
liquid stream temperature control, and accurate, complete analysis of
all streams to permit the calculation of material balances, and thus verify
absorption efficiencies. These matters were kept in mind during the
design and construction of the 10 SCFM pilot plant which is discussed in
the next section.
- 79 -
-------
J. The 10 SCFM Pilot Plant: Design, Construction, and Operation
1) Choosing a Basic Process Construction Approach. Three
alternative schemes were considered for pilot plant construction: Pfaudler's
"Glassed Steel" equipment, Corning's "Glassplant" system, or a beefed-up
laboratory type system such as might be available from any of several
laboratory supply houses. The features that were considered in evaluating
each of these systems included durability (the ability of the completed sys-
tem to withstand a wide variety of possible operating conditions and some
physical punishment while being run on a round-the clock basis), versatility
(the capability for easy change of components) and cost.
The Pfaudler "Glassed Steel" system utilizes steel equipment that
is internally protected from corrosive process materials with a layer of
fusion-bonded glass. It is quite rugged and is widely used in full scale
plants that handle corrosive materials or food products. It can be obtained
in a wide variety of individual components, which would allow considerable
versatility, but, as would be expected, it is the most costly of the three
systems. One other disadvantage of the Pfaud1er system was the fact
that we would be working at the very bottom of their size range, which
might make it difficult to obtain certain components.
The Corning "Glassplant" system utilizes heavy duty glass com-
ponents for use in small plants in the food and corrosive product industries.
The system is surprisingly rugged, and they manufacture some rather
large process units (including such items as 40 foot absorption towers
- 80 -
-------
and 50 gallon reactors). The equipment was available in
"kit-form" and is very versatile with the cost being lower than the Pfaudler
system. A g a i n we w 0 ul d have been working at the low end of the size
range, but it appeared that a complete range of components was available
in the size desired.
The laboratory type equipment available from custom suppliers
such as Ace Glass and Scientific Glass could have been designed and built
to meet our specifications, but here we would have been working at the
very top of their size range. Components such as columns for the absorber
and stripper would have had to have been custom blown to our design and
would have lacked a good deal in ruggedness and versatility. The cost was
undoubtedly the lowest of the three systems, but not as low as expected.
Based on this overall system analysis the decision was made to use
the Corning Glassplant Flanged Conical Pipe System in fabricating maj or
components as well as for the connecting piping. Although somewhat
higher in cost than the custom glass blown system, the advantages of using
interchangeable stock items of industrial quality more than compensated
for the increased cost.
Figure 27 presents a schematic diagram of the pilot plant system
with Table XIX giving the cost of the various pieces of equipment. The
design considera tions for each stage will be discussed separately below.
2) Flue Gas Supply. In supplying a flue gas to the system, there
are two obvious alternatives: using a coal burner to make "real" flue
gas, or putting together an artificial flue gas from pure gases. Upon close
examination it became clear that it would be far too expensive to obtain a
- 81 -
-------
Gas
Supply
System
Clean Gas to Vent
Sampling Port
Recycle
Acid
Reactor
N02
Manometer
Scrubber
Sampling Port
Sampling
Port
NO
Reoxl d izer
and Condenser
Reacted
d Gases
Oxi es
of Nitrogen
Dilute
HN03
Flowmeter
Sampling
Port
Diluted Acid
Holding Tank
and Pump
Sampling Port
Water to Vent
Strippe r
Acid
Concentrato r
Flow meter
Flue Gas
Holding Tank
and Pump
Sampl ing Port
Sampling Port
Holding Ta n k
and Pump
Fig. 27. Flow sheet for lOSe FM pilot plant
- 82 -
-------
Table XIX.
Cost of 10 SCFM Pilot Plant
l.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Flue gas supply system
Reactor gas supply
Reactor
Scrubber
Stripper
Acid concentrator
NO oxidizer
Heating jackets
Sulfuric acid pumping system
Piping and fitting
Sampling port assemblies
Packing (3/8" lntalox Saddles)
- 83 -
$ 11 00
880
370
1460
1100
960
475
700
1000
900
1100
250
$ 10,295
-------
coal burner (about $10,000), aside from the fact that small coal burners do not
really supply a flue gas representative of a power plant.
An artificial flue gas can be made up to any' composition desired by mixing
pure gases. The advantages are fairly obvious: the gas composition can be
made to order, thus excluding oxygen if desired, and the capital outlay is rela-
tively small. The disadvantages include high gas cost, a dependence on suppliers
to maintain fairly high volumes of gas, and the fact that the gas would be quite
different from real flue gas.
An alternative compromise scheme would utilize a natural gas burner
to provide a stream of hot moist flue gas which would approximate the flue gas
from a coal burning power plant if it were doped with sulfur dioxide and fly ash.
The gas would be too hot to use directly, but it might be possible to use the vent
gas from a commercial hot water heater where the gas temperature would be
relatively close to the 3000 F needed for the experimentation.
Further analysis of this last flue gas supply system (the natural gas
burner) shows that the raw flue gas from a natural gas burner must be diluted
because of its high water content but still offers the superior system. The natura]
gas, which is 95% methane, produces combustion gases which contain about 16%
water, as compared to the combustion gas of a coal burner which contains about
5 - 7% water. If the flue gas from natural gas were to be used, the water con-
centration would have to be reduced, either by removing some of the water (which
is impractical at elevated temperatures) or by diluting the flue gas.
The flue gas from a standard natural gas burner is produced at about
14000 F with 16% water. If the water content were to be diluted back to 5% it
would mean mixing the flue gas with 700 F air at a 2/1 ratio. The resulting gas
would be at a temperature of about 5000 F, within the desired range. This diluted
gas would have an oxygen concentration of 15% which could have some effect on
the concentration of NO produced during the oxidation of 50 by the NO . Althoug]
2 2
the oxygen would be present in large excess of the stoichiometric amount needed
to reoxidize the NO back to N02' the low concentration of NO in the flue gas and
the short contact time of the mixture before it reaches the scrubber would probab
prevent NO from being oxidized.
- 84 -
-------
According to Burdick, llit would take 15 minutes for a gas
containing 2% NO and 3% 02 to attain 90% reaction at 77°P. The
system currently under consideration would contain about 0.3%
NO and 15% 02' Considering that the reaction rate is directly
proportional to the square of the NO concentration and to the first
power of the oxygen concentration, it would take about 135 minutes
to attain the same degree of reaction in the new system at the same
temperature. Since the reaction rate decreases with temperature
increase, very little NO will react during the few seconds it is in
contact with oxygen at 3000p before it enters the scrubber.
Thus it appears that a diluted combustion gas from a natural
gas burner would give us a flue gas of the proper water composition
in the desired temperature range. The equipment needed would simply
be a small gas burner, rated at 0.25 cfm of natural gas to deliver 10 cfm
of hot flue gas. Natural gas is mixed with combustion air at a ratio
of 10/1, giving about 2.75 SCPM of 14000P flue gas which is then diluted
3/1 with room temperature air to give 11.0 SCPM at 400° P. This product
can then be run through a simple heat exchanger to drop the temperature
to any desired level. The same heat exchanger can be used to raise the
temperature if any future testing requires hotter gases. (Hotter gases
can also be obtained by lower dilution ratios, but this would of course
change the wa ter content of the product. )
Based on this analysis, the natural gas burner system was selected
for the pilot plant. Any disadv:antage in gas handling was more than offset
by a distinct advantage in operating cost. An artificial flue gas system
using nitrogen from a cryogenic supply system as the carrier gas would
have cost about $400 per month to operate while the monthly natural gas
cost is essentially negligible. (See Table XX for cost analysis.)
3) Reactor Gases. During the bench scale experimentation,
oxides of nitrogen were supplied individually from cylinders of pure
gases. This has worked out quite well for all gases except N02 which
is a liquid at room temperature.
- 85 -
-------
Table XX.
Comparison of Alternative Flue Gas
Supply Systems
Liquid Nitrogen System
Natural Gas Burner (Reactant Gases Added)
Flue Gas 02 15.6 02 0.0*
Composition N2 77.1 .N2 94.7
(% by volume) C02 2.0 C02 0.0*
H20 5.0 H20 5.0
S02 0.3 S02 0.3
100.0 100.3
Capital
Equipment Costs
Gas Burner
Fan
$100
75
Heat Exchanger
Heat Exchanger
200
$375
Monthly
Operating Costs t
Natural gas
$ 2.50
Nitrogen
Equipment
R en tal
Total Annual Operating Cost and
Equipment cost
$405
* Can be added if desired.
t Based on a 22
day month and a 5 - hour day.
- 86 -
$300
-
$300
$450
40
$490
$ 6180
-------
The flow of liquid N02 from the cylinder is therefore difficult to meter
and control. It was therefore decided to supply N02 by mixing NO with air
and providing enough residence time in a small reactor to allow complete
oxidation to N02 in a relatively dilute form. As long as the partial pressure
of N02 is below 0.75 atmospheres it will not condense at room temperature
and will be easy to handle. S02 will be fed to the system directly from a
cylinder as shown in the complete gas system schematic in Fig. 28. As
discussed in the previous section, water will be supplied by the combus-
tion gas from the gas burner and the H20 concentration will be measured
by a hygrometer in a small bypass stream.
4) Reactor. Continuous gas phase reactions are usually carried
out in pipe-like reactors, with the volume of the pipe depending on the
residence time necessary for complete reaction. Studies performed
earlier have shown that 1.5 seconds is sufficient for complete oxidationofS02
under the conditions existing in the proposed system. With a flow of
10 SCFM and a reactor temperature of 300°F, the reactor' volume would
have to be 0.40 ft3. Figure 29 shows the pipe -type reactor designed
around the Corning Conical Pipe System which will allow changes in
reactor volume by addition of lengths of standard 4" pipe.
Provisions will be made to insulate the reactor to simulate actual
process conditions, although some he a t 10 s s can be to 1 era ted.
As is shown in the diagram, the temperature will be monitored at several
points in the system. The pressure drop in the gas line (i. e., through
the reactor and the scrubber) will be measured by a manometer placed
on the upstream side of the reactor. From preliminary studies, it was
anticipated that most of the pressure drop will occur through the scrubber
since that is the only place where there is an obstruction in the line (packing).
The reactive gases (NO, 502' N02) will be introduced into the line just
before the reactor. No insulation is to be used on the piping since the 1"
glass pipe (5/32" wall thickness) should not lose much heat at the high gas
velocities being used.
5) Scrubber. The scrubber is to be a countercurrent gas absorption
column packed with glass Raschig rings. The column will be constructed of
Corning glass sections to allow us to vary the packing height from 1 - 10 ft.
- 87 - '
-------
To Reactor
Flo
Mete
-
-,.
) 11\
N02
-I--
w S02
r Supply
~ '---- NO
Oxidizer
HYlrrometer
- r 1 To Vent
I I
Diluent
.Air
~
NatUral .... Air
Gas
Burner
NO
J Supply
Tank
Natural Gas
& Air
Fig. 28. Gas Supply System
Pilot Plant
- 88 -
-------
Therm~meter
00
(0
To Manometer
..
4" diameter
, Pipe Sections
Thermometer
50"
3
ReactOr Volume... 0.40 ft
Fig. 29. Pilot Plant Reactor
T:J:r
~
To
Scrubber
-------
The absorber column diameter required for 10 SCFM of flue gas was
calculated using the empirical method described by McCabe and Smith.14 The
general approach is to determine the gas flooding velocity of the column con-
taining a certain packing and pick a column diameter which will result in an
optimum gas velocity equal to one-half the flooding velocity. (Justification
for the optimum choice of one-half the flooding velocity can be found in Sherwooc
and Pigford, 15 "Absorption and Extraction").
A logarithmic plot of the group Gy 2 av (/1 xt)O. 2j gc E3 PxPy versus
the group (G j G )./p j p was determined empirically for a great many
x y y x
systems. The parameters are:
G = mass velocity of liquid, Ib/ft2-hr
G x = mass velocity of gas at flooding, lbj ft2 - hr
P: = density of iiquid, lbj ft3
Py = density of gas, Ibjft3
av = surface area of dry packing per unit packed volume, ft2 j ft3
/1~ = viscosity of liquid, cp
gc = ~Newton' s law conversion factor, 4.17 x 108 ft/lb/lb force-hr2
E = porosity, or fraction voids, in packed section
In our system the gas flow rate is set at 10 ft3/ min and G G calculated
x y
to be 1. 5. The values of the other parameters are given below:
p = 1. 64 lb/ ft3 at 2250 F
x
Py = 0.0582 lbj ft3 at 225°p
a = 122 ft2 j ft31
~ = O. 64 for 1/2" Raschig rings
/1~ = 1. 4 cps (77 v sulfuric acid at 2500 F)
Some of these values are averages or approximations, but the errors'
introduced are considered to be small, probably under 10%, and it turns
out that the value of Gy is relatively insensitive to even larger variations
in these parameters.
- 90 -
-------
The value of the group G jG J p j p - was calculated and for
x y y x
the chosen conditions is 0.0357. Using the empirical curve from McCabe
and Smith (Fig. 30) the value of the group G 2 a (/1' )0. 2jg E 3 p p wac<
. . y v x c xy Co
determll1ed to be 0. 17. Solvll1g for G gives a gas flooding velocity of
2 y
920 lbj ft -hr. The mass gas flow is
10 ft3j min x 60 mini hr x 0.0582 lbj ft3 = 34.9 lbj hr
Assuming an optimum gas velocity of one-half the flooding velocity, the
cross -sectional area of the column is
34.9 lbj hr
920j2 Ibjft2-hr
= 0.0758 ft2
This is the cross-sectional area of a column with a diameter of 3.75
inches. As' a result, the column diameter was chosen to be 4 inches.
Since proposed process modification calculations have shown
that it might be possible to run the scrubber at considerably lower liquid
to gas ratios, the optimum column diameter calculation was made again.
Based on the revised calculations discussed in an earlier report, a value
of 0.45 was used for the parameter GIG instead of 1. 5. The optimum
x y
column diameter for this new system was calculated to be 3.58 inches.
Obviously the 4 inch column will be appropriate for a wide variety of
conditions as long as the gas flow rate is kept at 10 SCFM.
An effort was made to calculate the optimum height of the column,
but the mass transfer coefficients necessary to obtain reasonably accurate
values are not available. Therefore, a variable height column was
designed.
Figure 31 shows the schematic diagram for the design of the scrubber
using the Corning system. This column sections are 1 foot long, making
the apparatus quite versatile. The feed sections in the column are for the
purpose of introducing the gas feed stream at different heights without com-
pletely disassembling the tower. These sections will also be packed and
are provided with thermometer ports for monitoring the column temperature
at different points.
-"91 -
-------
0.\
'"
0
.."""
.. ~..
~
",'" Pr}~~
'" '" ..
~ b,\J
00 I
0.001
001
Fig. 30.
00 <
o 0 <
C\Jp<
rf"~~
~<':-:;""'''--I- ~
T ').~ltf >,It
II '3 "0 .
>- . R'(f«'c,'i'...c ..J Q
V I.---i -=.c I,*~,' ~ r U Q
> < "'~. .~~.o
"L:\:.I.,o
'-~x
CI /;,A,t
ce,;..... D
o .f'~;?\~.
o x-)...
8" .s:;;~ 0
J::~\:: x
< -0 -6>
o .~
1.0
<:
->
P.' < '0
~["X 0 >
~...- "\~. .,. 0 ~,
A V A - r:::.-s.<.! 0
D p.~
o
......
Symbol System
o H20.oir
C1 H20- Hz
P H20-COZ
Lt Glycerol.oir
~ Butync ocid - oir
* CH30H-orr
v Turbine oil-oir
1\ Transformer oil'olr
> B'IOOoil,olr
< 10-Coil-olr
D Oil No 1- Olr
p Oil Nol-COz
b.. Oil Nol-HZ
A Oil No2-Olr
V Oil No.2'C02
+ Oil No.3-oir
,
>
.
""
. ...
(
0.1
Gx ..M
Gy r-;;
<
1\
t\
1.0
10
Flooding velocities in packed tower (McCabe
and Smith). 14
,.. 92 -
-------
Thermometer
Thermometer
Thermometer
-=--=-~:~.dl--J Gas Exit
...
~4" Diameter
Pipe Sections
Alternate Gas
Feed Points
...
u
Thermrmeter
~~ Gas Feed
Acid
Exit
. 31. Scrubber
FIg. .
- 93 -
-------
6) Stripper. The stripper is a packed column, also of 4 inch
diameter and variable height. The 4 inch diameter is somewhat over-
sized for this equipment but will not cause any loss in efficiency and is
being used to achieve plant simplicity by means of interchangeable parts.
All piping will be 1 inch in diameter, while all process components will
be 4 inches in diameter. This size is based on the optimization of the
scrubber detailed in the previous section since it was considered that the
scrubber was the critical component in the process. (If the scrubber does
not work to maximum efficiency, pollutant gases will escape to the atmos-
phere. )
The stripper will be operated by passing hot gases countercurrent
to the nitrose stream coming from the scrubber. The equilibrium data4
indicate that in order to remove the NZ03 a large amount of water will
also be stripped from the nitrose. This wet gas constitutes a recycle
stream which adds a blower operating cost burden to the process. One
of the tasks to be undertaken during the operation of the pilot plant is the
development of methods either to remove this water from the recovery
stream or to avoid its pickup in the first place.
The stripper and the NO reoxidizer will be set up on an experimental
basis so that process modifications can be easily incorpora ted during the
continuous operation of the pilot plant. The plant will be initially constructed
so that the system can be opera ted continuously regardless of the efficiency
of the stripper and the oxidizer. This means that steady state will be
achieved in the reactor, the scrubber and the acid concentrator even if
the nitrogen oxides are not being recovered and recycled efficiently.
7) Acid Concentrator. In order to recycle the acid used in the
scrubber, it must be stripped of its oxides of nitrogen and concentrated
back to 80%. Since the concentration of sulfuric acid is a well established
process which does not need additional development at this time, we do
have the alternative of dumping the acid, which amounts to about 7 gal/hr.
Since this would be inconvenient as well as costly, a distillation column
will be put in the system to concentrate the acid to the desired level. The
.-:.94 -
-------
goal here is to make the pilot plant as complete and self-contained as
possible. This will help in later evaluation of heat requirements, as
well as enable us to introduce process modifications in a realistic manner.
The concentrator will be constructed of 4 in. Corning pipe, with a
reflux head and condenser on top of a short packed distillation column.
Figure 32 shows the setup of this apparatus.
The acid concentrator has been designed as a distillation column
heated with electric heating mantles. The mantles will heat the cool feed
(about 150°F) from the stripper (initially directly from the scrubber) and
boil off sufficient water to increase the concentration to 80%. Based on a
water input of 5% in the flue gas stream, a reactor inlet N02/S02 ratio of
2.0 and a scrubber L/G of 1.0 moles of acid per mole of flue gas, the
liquid leaving the scrubber should have the following composition:
H20 23. 7%
H2S04 75.8%
N203 0.5%
On a basis of 10 SCFM of flue gas entering the system, the flow rate of
liquid scrubber effluent would be 1. 19 lbj min of acid wi th this composi-
tion. To increase the concentration to 80% acid, 0.0595 lbj min of water
ha ve to be rem oved.
In order to determine the amount of heat needed to remove this
water, the enthalpy-concentration diagram shown in Fig. 33 was used.
The technique used for determining the heat input is to subtract the enthalpy
of the original solution from that of the total of the hot liquid and the
vaporized water:
.6.H = n2H2 nlH 1
HI = H06% H2S04 at 150°F)
H2 = H(80% H2S04 at 395 OF) + H(steam at 395 OF)
n = lb of material
- 195 -
-------
DilUTE H2 S04 FEED
150°F
CONDENSER
SECTION
REFLUX HEAD
H20 DISTillATE
PACKED
DISTI llATION
COLUMN
THERMOM ETER
H EATER SECTION
350°F
CONCENTRATED 80% WT
H2 S04 PRODUCT
Fig. 32. Acid concentration
- 96 -
-------
,.Q
...-.j
'-
::J H2
~
:>.
P..
........
C'IJ
..c:::
.j...J
c:
~
HI
Percentage H2S04
Fig. 33, Enthalpy-concentration diagram for aque.ous sulfuric
acid at 1 atm.16Reference states enrhalples of pure
liquid components at 32°F.
- 97 -
-------
The enthalpies of the sulfuric acid solutions are shown in Pig. 29 as
HI and H2° The enthalpy of the superheated steam is determined from
standard steam tables.17The values for each are:
HI = -75 BTU!lb
H2 (acid only) = 47.5 BTU/lb
H2 (steam) = 1137.5 BTU/lb
nl = 1. 19 lbj min (79% acid)
n2 = 0.0595 lb H20! min + 1. 13 lbj min (80% acid)
Calculating from these values, .0.H is 209 BTUj min or 3680 watts in
equivalent electrical power. Using heating mantles along 28 inches
of the 4 inch pipe that makes up the distillation column res ults in a
heating requirement of 9.3 wattsj in2, well within the design capability
of the heating mantles.
One aspect of the design of the concentrator system that must
be considered is the heat transfer capability of the glass pipe. In other
words, can the glass transfer the required amount of heat fast enough
at a reasonable thermal gradient across the glass. To get an idea of
this, the required gradient was calculated from the following equation:
Q = U A .0. T
(1)
where
Q = heat transferred BTU/ min
U = over-all heat transfer coefficient BTU/ hr/ ft2j 0p
A = surface area for heat transfer
,6. T = temperature gradient across the pipe
This equation is an oversimplification of the situation since the tempera-
ture gradient varies over the length of the heated section. It is more
accurate to use the log mean temperature difference which is defined: 17
.0. tL =
.0. t2 - .0. tl
2.303 log (.0. tl )
.0. t2
(2)
- "98 -
-------
Corning suggests using a value of 55 BTUjhrjft2jop for the
over-all heat transfer coefficient for glass heat exchangers with a
cool liquid on one side and a condensing vapor on the hot side. That
situation is more efficient than the conditions existing in the proposed
acid concentrator where there is a liquid film on the cold side and the
heating mantle (with an air space) on the hot side. From an examination
of over-all heat transfer coefficients of other systems, an estimate of
2 ° ----
35 BTUj hrj ft j F is probably reasonable. The value of .6. tL was cal-
culated from equation (1) using the following values.
Q = 209 BTUjmin
U = 35 BTUj hrj ft2j of
A = 395 in2 (28 inch mantle for 4 inch pipe)
- 0
Th~ value of .6. tL is then 132 F.
The outside wall temperature can now be calculated from equation
(2) if one assumes that the heating mantle temperature is constant along
the length of the pipe, using 150°F for the inlet acid temperature and 395°p
for the outlet temperature. Therefore, if T is the outside wall temperature
.6. t 1 ;:;: T - 150
A t2 = T 395
Rearranging equation (2) and solving for T:
T = 150 - 395 C
1 - C
where
C = 10(245/2.303.6. tL)
Substituting 132°F for .6. tL:
C = 6. 42
Substituting in equation (3):
T = 4410 F
- 99 -
-------
This means that in order to boil 1. 12 lbl min of 75. 8% sulfuric
acid and concentrate it to 80% in a 28 inch long, 4 inch glass pipe, the
outside of the glass pipe has to be maintained at 441 DF. (This, of
course, depends on the actual operating heat transfer coefficient. If
in practice it is lower than the value used above, the pipe temperature
will have to be higher.)
The recommended maximum operating temperature for the
Corning pipe system is 450DF which is very close to the wall
temperatures needed for evaporation of the water from the boiling acid.
Upon consultation with Corning it appears that this maximum temperature
is set for two reasons: (1) the maximum use temperature of the Teflon
gaskets used at the flanged joints is 450°F and (2) the glass pipe is
tempered at the ends to increase its strength where the stress on the
flange is applied and also to maintain the roundness of the pipe under
stress, avoiding high local stress concentrations in the pipe. Corning
admits to being somewhat conservative in their temperature rating of
this equipment and says that with care to avoid thermal shock, wall
temperatures of 500°F can be tolerated. Thus it would appear that
our system can be operated without serious problems.
There is another approach to this acid concentration problem
which would avoid the use of the high pipe temperatures - run the
distillation under vacuum,: As the total pressure is reduced, the boiling
point is also reduced, thus allowing us to concentrate the acid at lower
tempera tures.
Table XXI show vapor pressure data for water over sulfuric acid
solutions at various concentrations and temperatures in the range of interest.
From the data in the table below, it can be seen that to lower the
boiling point of the 80% sulfuric acid about 110°F, the pressure has to be
reduced to 108.7 mm Hg. If heat is put in at the necessary rate, the
vaporizati,on process should take place fast enough to concentrate the acid to
the desired level.
- 100 -
-------
Table XXI. Vapor Pressure of Water Over Sulfuric
Acid Solutions (mm Hg)4
Tem~erature Acid Concentra tion
75. 7% wt 78.0% wt 80.0%wt
( F)
140 4.9 2.9 1.7
15~ 8. 7 5.2 3.4
176 14.7 9.3 6.5
194 24.3 15.6 11. 5
212 39.1 25.6 18.8
230 60.3 40.9 30.9
248 91. 4 63.7 48.0
266 134.6 95.9 72.6
284 144.2 141. 3 108. 7
370 760
385 760
395 760
- 101-
-------
The equipment that has been ordered can be easily adapted to opera-
tion under moderate vacuum and the heaters have been sized to provide mOre
than enough heat to boil the acid at atmospheric pressure. This
will allow a wide range of operating temperatures and pressures, giving
us substantial versatility for acid concentrator performance. If there
are difficulties with either of these two approaches to the concentrator,
there are two other alternatives: preheating the acid feed, or using
special Corning heat exchanger pipe which has an over-all heat transfer
coefficient three times that of the standard Corning pipe (although it is
not primarily intended for column use).
8) NO Oxidizer
The NO oxidizer will be a straight rube pipe reactor
virtually identical to the S02 oxidation reactor shown in Fig. 25. The
length will depend on the concentration of NO in the stripper gas and
the required residence time for reoxidation to N02.
In the evolution of this process it has been found that there is
essentially no advantage to using a N02/NO/S02 ratio of 3/1/1. The unit
will work equally well with a mole ratio of N02"s02 of 2/1. Of course
there must be a slight excess of N02 in order to assure the complete
oxidation of S02' but the 200% excess discussed previously is probably
unnecessary. Some adjustment in mole ratio will be required since it
is impractical to assume that all the NO from the stripper will be
oxidized to N02 and some reasonable level of reoxidation will have to be
planned, e. g. 90%. There will then have to be an excess of N02 to
permit the mole ratio of N02/NO to be 1/L as the gas enters the scrubber.
These two requirements for slightly higher N02 levels than would be
assumed theoretically would probably be satisfied by a 10-20% ex~ess
of N02 in the recycle stream.
9) Pumps
The two important criteria in selecting pumps for sulfuric
acid service are flow capacity and materials of construction. Considera-
tions such as suction head, discharge head, horsepower and uniformity
of flow (pulsations) are only of secondary importance.
- 102 -
-------
The capacity of the pump has to be in the 0.1 -0.2 GPM range,
thus establishing the requirement for a metering pump with a
capability for carefully controlled flow. This feature is quite important
since the efficient operation of the system depends on the ability to
balance the material flows in each stage of processing.
The materials of construction depend on the temperature of
the acid being pumped, which of course depends on the process stage
in question. An examination of Fig. 27 reveals that three pumps will
be used in the system. The acid leaving the scrubber will be relatively
hot. If the scrubber inlet acid is at 80° F, the diluted acid which is to be
pumped to the stripper will pick up sensible heat from the reacted flue
gas and heat of solution from reaction with the water, thus leaving it
at about 275 of. The acid leaving the stripper will be somewhat hotter,
de pen din g on the tern peratUre of the stripping gas and the flow
ratio of gas to acid. Temperatures will probably be in the 285-300 of
range. It is the acid leaving the concentrator that will be the hottest.
This acid will be at its boiling point which is 395 of for 80% acid.
Examination of corrosion tables indicates that the only materials
which have adequate service lives for 80% sulfuric acid above 200 ° F
are high silica iron (known by the trade name of Duriron), glass, platinum,
gold and tantalum. The last three are out of the question because of
price, which leaves glass and Duriron as the only possible materials.
An extensive search of the industrial literature indicates that there are
no metering pumps made of Duriron and the glass metering pumps that
are available have thermal shock problems that restrict their use to
temperatures below 150 of. Since Duriron is not machinable, it will be
difficult if not impossible to obtain a custom -manufactured pump out of
this metal, which leaves only one solution - the acid which leaves all
three stages will have to be cooled before it is pumped. (It should be noted
here that for the higher flow rates that will be present in larger scale
plants, Duriron pumps of the proper size are readily available. )
- 103 -
-------
For acid temperatures below 2000 F, pumps are readily available
in the desired flow range. Carpenter 20 stainless steel is recommended
for use with 80% sulfuric acid up to about 2000 F, and the pumps that will
be used have the following specifications:
Mark II Chemical Metering Pump (Clark-Cooper Co. )
Liquid end
Pump housing
Plunger
Check balls
Seals
Packing
Speed
Flow rate
Carpenter 20
Cast iron
Carpenter 20
Ceramic
Teflon
Teflon/asbestos
88 strokes/minutes
o - 0.175 GPM
10) Flow Meters
There are essentially three different types of flow
meters in the system: gas service in the 10 SCFM range, gas service
in the O. 1 SCFM range and liquid in the O. 1 GPM range. The meters
were chosen on the basis of ease of operation and compatibility with'
the rest of the system.
Several companies make flowmeters of the glass rotameter
variety, and they are all about the same in price and operational
features. Only one company, however (Schutte and Koerting Co. )
makes their flowmeters compatible with Corning Conical Pipe. It is
possible, of course, to use standard threaded connections which are
cheaper, but the connection to the glass pipe has to be made somewhere
and the adapters necessary make the total cost relatively close. The
convenience of using the same type of connectors compensates for the
additional cost.
- 104 -
-------
Concerning the sulfuric acid service, there was an alternative
of either calibrating the metering pumps or using flowmeters. While
it is likely that the pumps can be calibrated, it is felt that this would
involve a somewhat indirect control method and slow changes in pump
performance could not be determined. A situation which would require
occasional pump recalibration would be impractical. There was also
a choice of using one acid flowmeter and valving it so that all three
process streams could be monitored individually. It turns out that the
pipe and valves would cost as much as the additional meters (aside
from the contamination problem while switching from one stream to
another). The decision was made to buy three meters, one for each
acid stream.
11) Heating Mantles
There will be several places in the system that will
require heating because of process requirements or because the mini-
plant will have higher heat losses than an actual
plant and will need reheating. The acid concentrator, which is a dis-
tillation column, will need a heated section at the bottom to act as a
reboiler. The scrubber will probably need some heated sections to
compensate for heat losses, as will the stripper. There may also be a
need to reheat the sulfuric acid after it is pumped.
There are two ways to perform these reheating operations:
run the process streams through heat exchangers or use external
heaters. In the case of the scrubber and the stripper there is really
no choice, and heating mantles will be used. Although this technique
is somewhat more expensive than using heating tapes, the mantles pro-
vide more uniform heat and are much easier to use since they will not
require additional insulation after they are installed.
Since the entire system is designed in a modular fashion using
short sections of 1 and 4 inch pipe, the heating mantles will be obtained
in the same way. By this technique the mantles can be purchased without
designing special shapes or sizes and can be used where needed. The
mantles will be obtained from Glas-Col Company.
- 105 -
-------
12) Construction and Start-Up
The pilot plant was constructed in the general manner shown
in Fig. 27 using the Unistrut framework system for support. No
problems were incurred in the construction aside from the usual equip-
ment delivery delays and the final system is shown in the photograph
in Fig. 34.
A significant problem arose when the plant was started up. There
was a considerable pressure drop across the reactor and scrubber. The
pressure drop was too high to be overcome by the natural convection gas
burner being used. Table XXII presents the back pressure of the system
at various gas flow rates with the scrubber dry. (The "system" for these
tests included the 10 cfm flowmeter, the S02/N02 reaction chamber, the
scrubber-packed with 8.5 ft. of 3/8 in. Intalox saddles, nine elbows,
four thermometer well constrictions and thirteen feet of 1 inch glass pipe.)
Table XXII. Experimentally Measured Gas Sys tern
Pressure Drop as a Function of Gas Flow Rate
Gas Flow Rate
(dm)
Pressure Drop
(Inches of Water)
2. 8
3. 7
4. 7
5. 7
6.8
7. 8
8. 9
10.0
2.0
3.0
4.5
6.5
8.5
10.8
13. 5
16.0
Data supplied by U. S. Stoneware indicates that the column would have
a pressure drop of O. 3 inches of water per foot of packing when operated
dry and about 0.4 inches of water when operated at an L/G ratio of 2.0
and a gas mass velocity of 600 lb /ft2 /hr (10 cfm in a 4 inch column). The
- 106 -
-------
,.
Fig. 34. Pilot Plant
- 107 -
-------
difference of O. 1 inches of water per foot of packing results in a
correction of about 1 inch of water additional pressure drop at
10 dm when the acid is flowing through the column. Actual measure-
ments of pressure drop in the dry column show a drop of less than
2 inches of water instead of the 2.4 predicted which suggests that
1 inch additional pressure drop should be more than enough to correct
for liquid flow.
This total pressure drop of about 17 inches of water is less
than the drop experienced with the small 1. 0 cfm bench scale system
used in earlier stUdies, which is probably due to proper sizing of
pipes and fittings (the old set-up used some small-bore fittings which
undoubtedly contributed a large portion of the pressure drop). However,
this pressure drop is still too high to allow a natUral draft to be formed
in the gas burner.
There were several ways to approach the pressure drop prob-
lem and the following were considered:
1. pressurized gas burner. . . There are commercially
available gas burners which mix the air and the gas and force-feed the
mixture to the burner under pressure. In general these devices have
relatively low operating pressures and are fairly expensive.
2. vacuum exhaust. . . The vent from the scrubber could
be fed into a vacuum pump which would pull the gases through the whole
system. This would be relatively simple to install and only moderately
expensive.
3. in -line fan. . . The exhaust gas from the gas burner
could be fed to an in -line fan which would force the gases into the pro-
cessing system. The gases would have to be cooled before they reached
the fan in order to simplify the fan design. This approach is straight-
forward, but would involve modifications in the fan and the piping since
these devices are not usually designed for operation with small bore
piping systems.
- 108 -
-------
The system decided on what was a moderately high pressure in-line
blower which takes the output of the gas burner at about 250 ° F and forces it
into the system. The blower has a maximum exit temperature of 350 ° F and
since the blower itself heats the gas it is necessary to introduce a relatively
cool gas. The blower was installed and appears to have solved the back pressure
problem.
13) Initial Experimentation
Initial runs on the pilot plant were aimed at evaluating the scrubber
efficiency for removing both sulfur dioxide and nitrogen oxides. The reactor
was heated to 3000 F and an artificial flue gas at 10 SCFM containing about 0.3%
502 and 1% water was used as the feed stream. No effort was made to use
the proper amount of water in the gas because the whole system was not going
to be tested and it was therefore not essential to have any more water than the
minimum necessary to make the acid. The flue gas was mixed with N02 in
various N02/S02 ratios and the reacted gas stream fed to the scrubber. The
scrubber acid was fed at an L/G of about 1.0 mole/mole and a temperature of
90°F. ,The gas was sampled after the scrubber and was analyzed by infrared
absorption for oxides of nitrogen and sulfur dioxide. The effluent acid was
analyzed for its nitrosylsulfuric acid content by KMn04 titration.
Table XXIII shows the results of a number of experiments where the
N02/S02 ratios were varied from about 4 to 2. In all cases the S02 was com-
pletely absorbed, but as would be expected, the N02 absorption depended on
how much excess N02 was present above that required to form an equimolar
NO: N02 mixture after the S02 oxidation reaction.
Examination of the data shows that the amount of nitrogen oxides absorbed
in the scrubber was usually about twice the amount of sulfur dioxide in the feed
stream. Since no S02 was apparent in the exit gas, it appears that the N02
oxidized all of the SO forming enough NO to allow the absorption of an equal
2
amount of N02' both being equal to the original amount of S02 present in the gas.
- 109 -
-------
Table XXIII. Experimental Res ults: Pilot Plant
Scrubber Scrubbing Materia]
Entrance Gas Exit Gas % HNS05 in Efficiency Balance
Partial Pres - Partial Pres - Scrubber Outlet % NOx Errort.
sure (torr) sure (torr)* Acid Removalt % "
S02 NO ** S02 N02
2
2.28 9.45 <0.01 5.0 1.28 47 10.0
2.80 9.45 <0.01 4.1 1.63 57 9.1
2.80 7.85 <0.01 2.4 1.95 70 4.2
2.80 7.48 <0.01 1.9 2.00 75 4.0
2.80 6.63 <0.01 1. 52 2.18 77 -10.2
2.80 5.74 <0.01 0.07 1. 78 98+ 16.6
2.80 5.74 <0.01 0.08 2.01 98+ 6.2
2.80 5.74 <0.01 0.08 1.98 98+ 7.6
*NO could not be measured in the effluent gas, because by the time the gas was
analyzed (infrared) all the NO was oxidized to N02. Therefore, tOtal NO only
was measured. x
t Based on inlet and outlet NO concentrations in the gas.
x
t Conditions: NOx (in) - NOx (out)
x 100
NO (in)
x
** No NO was added to reactor inlet gas.
L/G = 1.0 mole/mole
Scrubber inlet acid temperature = 900 F
Gas flow rate = 10 SCFM (1% H20)
Acid flow rate = 0.085 GPM (80% H2S04'
0% HNS05 )
Reactor inlet gas temperature = 3000 F
- 110 -
-------
Good material balance closure was obtained in all the runs with
the error being less than 10% in all cases. The experimental conditions are
very close to those in the standard Lead Chamber Process, which has been
used for more than a century, and the results agree with Lead Chamber
Process commercial operating data. It is clear that control of the NO /NO
2'
ratio is vital to achieve high enough NOx absorption to make the process
profitable.
K. ?rocess Improvements (Bench Scale)
1) General Approach to Process Modification
Examination of the schematic of the Modified Chamber
Process in Fig. 20 and the heat and material balances for the process
in Tables XI and XII reveals that the required 37% excess heat discussed
in Section V - H9 is divided approximately in half between the stripper
and the acid concentratOr. This means that half the heat is needed to re-
move the excess water abs~rbed by the acid in the scrubber and the other half
needed to concentrate the N203. The basic approach to the heat requirement
problem then, was to 1) try to prevent tbe water from being absorbed in
scrubber, and 2) find another way to concentrate the N203.
2) Reactor Gas Condensation
Initial work in trying to prevent the water from being
absorbed in the scrubber acid was to remove the water from the flue
gas stream before it got to the scrubber. It was thought that by cooling
the gas between the reactor and the scrubber, enough water could be
condensed so that the scrubber acid would not be excessively diluted.
For instance, if the partial pressure of water in the flue gas stream
were reduced from 50 torr (about 7%) to 27 torr (3.5%) by cooling the gas
to 90 of, thus condensing out the water as dilute sulfuric acid, then
the acid in the scrubber would be diluted to only about 78% instead of 76%
as previously. Then when the acid was heated in the stripper, the excess
- 111 -
-------
water would be removed along with the N203 and recycled to the reactor
with the concentrated N203' Then the water would continue back to the
condenser where all the water in excess of 3.5% would drop out. By this
method the flue gas entering the reactor is enriched with water, but loses
everything down to 3.5% in the condenser. At steady state, flue gas con-
taining 3.5% H a enters the scrubber and a constant amount of water is
2
recycled from the stripper. The diluted acid from the condenser would
be concentrated and sold and since this is a relatively small amount of
acid compared to the large amount of recycle acid, the heat requirement.
should be cons iderably reduced.
. -
This scheme was tried experimentally by running the reacted
flue gas into a condenser and analyzing the effluent liquid and gas as
shown in Fig. 35. The nitrogen (at room temperature) was bubbled
through a water pre-saturator, with the water temperature being
113 0 F. This would provide a water content of about 9% in the gas
if it came to equilibrium, but there was sufficient cooling of the gas
before it entered the reactor that it was highly probable that the actual
concentration was much lower. Other methods were used to determine
the water content of the reactor inlet gas accurately and these will be
mentioned later.
The reactor tube was wrapped with electrical heating tape, with
the heat input being that which would provide an exit gas temperature of
300 of at the thermocouple. Since the gases entered at a much lower
temperature, there was a temperature gradient along the reaction tube.
The hot reacted gases pas sed into the coils of a condenser.
The shell of the condenser contained circulating cooling water that was kept
at constant temperature in an external thermo stated bath. The water
flowed through the condenser at such a rate that the exit and
entrance temperature of the water was about the same. This created
a constant temperature region along the length of the condenser.
- 112 -
-------
Reactor Tube
Thermocouple
Electric Heating Tape
9% H20
S02
1130 F
Flowmeters
Out
Water
Pre satu rator
Water From
Constant
Temperature
Bath .
in
Condensate
Collection
Tube
N2
NO
300 0 F
To Vent
Condenser
IR
Analysis
Cell
98% wt
H2S04
Absorber
250 cc
Fig. 35. Experimental apparatus for H20 condensation experiments
- 113 -
-------
The gases and the condensate passed into a collection tube, with the
gases being piped out into an absorber containing 98% sulfuric acid. After
passing through an infrared analytical cell, the gas was vented to the atmos-
phere. The purpose of the sulfuric acid absorber was to remove all the
sulfuric acid mist and water from the gas before it went through the IR cell.
This cell contains sodium chloride windows which would have been affected
by the aqueous vapors in the gas stream. The acid would absorb water, acid
mist and equimolar quantities of nitric oxide and nitrogen dioxide but would
not collect sulfur dioxide and excess nitrogen dioxide in significant quantities,
Thus a complete nitrogen oxide balance could be made by analyzing the
absorber acid for nitrosylsulfuric acid and the gas for both NO and N02' The
amount of water in the gas was determined from the increase in volume of
the absorber acid during the experimental run (assuming 0% H20 in the vent
gas). A complete sulfur balance could be made if it were possible to analyze
the absorber acid for sulfuric acid content, but the large volume of acid
required to provide a constant rate of absorption of the oxides of nitrogen over
a period of time made it impossible to detect the pickup of very small quantities
of sulfuric acid. It was necessary to absorb the nitrogen oxides over a
significant period of time in order to develop a high enough concentration of
nitrosylsulfuric acid to allow accurate analysis.
The only parameter that was varied during the experimentation
was the temperature of the circulating water in the condenser. The other con-
ditions were maintained as follows:
Nitrogen flow rate
Sulfur dioxide flow rate
Nitrogen dioxide flow rate
Nitric oxide flow
28, 300 cc/min (1 cfm)
85 cc/min
230 cc/min
85 cc/min
113 0 F
300 0 F
Water presaturator temperature
Reactor gas temperature
Reactor retention time
0.46 sec (assuming an average reactO
temperature of 1850 F)
- 114 -
-------
In addition, the residence time in the 98% H2S04 absorber was main-
tained at abdut 30 minutes with the sulfuric acid volume in the absorber
being about 250 cc (both were determined exactly for each experiment).
The results of the initial condensation experiments are shown in
Table XXIV.
At 140°F there was virtually no condensation, although this is
below the sublimation temperature of nitrosylsulfuric acid and we
might have expected to see crystals forming on the condenser walls.
There was some fogging of the condenser coil, but nothing could be
collected. These results verify earlier experiments which demonstrated
the futility of trying to condense solid nitrosylsulfuric acid. (After the
experimentation was completed, it was observed that crystalline ma-
terial had formed in some liquid puddles in the reactor. This was un-
doubtedly nitrosylsulfuric acid, but in small quantities and difficult to
collect. )
At 122°F there was somewhat more condensation, bUt it still
could not be collected. It was not until the condenser temperature was
'dropped to 1040 F that appreciable condensate was formed. This ma-
terial,was collected and analyzed for both acid content and the presence
of nitrogen oxides. As the table shows, the solution was 49% H2S04
. in this case, while the nitrosylsulfuric acid content was negligible.
When condensation was performed at 860 F, the',amount of condensate
increased, as might have been expected, and the acid content decreased
to 45%. Again there was negligible nitrosylsulfuric acid present.
Material balance concentrations showed that the acid collected
in the two low temperatUre condensation experiments was only about
25% of the acid which had been formed in the reaction. This points out
the well-known difficulty of removing sulfuric acid mists from gas
streams. (It is assumed that the remaining 75% was absorbed in the
H2S04 absorber.)
It is interesting to note that the amount of N203 retained by the
sulfuric acid in the absorber decreased as the gas' temperatUre increased,
although nitrosylsulfuric acid is more soluble in sulfuric acid at higher
temperatures. This is partly due to the higher N203 vapor pressure at
- 115 -
-------
Condenser
Water
Inlet Temp (0 F)
86
104
122
140
....
....
a:.
Duration
of Run
( min)
30
30
30
31
* Conditions - - See page 113
*NO (in)-NO (out)
x x
NOx (in)
Table XXIV. Results of Condensation Experiments*
98% H2S04 Absorber
Final Acid Water Total N2.03
Volume Pickup in Acia
(cc)*** (ce) (g moles)
260
266
275
273
10
16
13
0.103
0.087
0.081
0.071
x 100 = Error
***Initial H2S04 volume = 250 cc
Exi t Gas
Concentration
(mm Hg)
NO N02 S02
4.0 1.2
4.2 1.1
4.2 1.2
3.0 1.4
0.7
0.7
0.7
1.8
NOx*
Material
Balance
Error (%)
-1.4
0.0
3.1
7.2
Final
Volume of
Condensa te (ec)
2.1
1.8
<0.5
o
H2S04
Coneen tra tion
(% wt)
45
49
-------
higher temperatUres and partly due to the shorter contact time of the gas
in the acid since the gas volumetric flow rate was higher at the higher
temperature.
One of the most significant results of the experimentation was the
excellent reproducibility and reliability of the analytical procedures. Since
the reactor conditions were the same in all four experiments, one would
expect that the exit gas would have the same composition in all cases. For
the most part this is true, with the exception being the 1400 F condensation
which was the first run made. This run is hard to explain, but even here
the total material balance on the nitrogen oxides was closed to within 7.2%.
It is interesting to note that the NOx content of the exit gas was essentially
constant even in this case although the NO/N02 ratio varied.
From the quantity of water picked up in the acid absorber, it is
possible to determine the actUal water content of the original gas stream.
All the water in the gas stream was trapped in the acid scrubber since the
IR analysis of the exit gas did not show any traces of water in any of the
experiments. From the data in Table XXIV it can be calculated that the
gas contained 2% water on the average. The reasons that the gas did not
contain the anticipated 9.5% are that (1) the residence time of the gas in
the presaturator wa"s too short to allow the gas and liquid to come to
equilibrium and (2) the gas cooled down after it left the presaturator and
before it entered the reactor, thus allowing condensation of some of the
water. In any case there was more than enough water in the gas to provide
for the formation of sulfuric acid, and the low water content had no effect
on the results. Since it was found that a portion of the acid condensed with
the water making a dilute acid (about 45%) it was possible that the product
could be concentrated to a usable level and thus partially solve the water
problem. However, the difficulties in cooling the gas to these low temper-
atures (86 -1 040 F) and recovering the acid mis t made the technique im-
practical and the work was discontinued.
-117-
-------
The experimental set-up used for the condensation experiments
p!ovided an opportunity to perform a valuable kinetic stUdy on the sulfur
dioxide oxidation reaction. Previous experimentation had shown that better
than 90% of the S02 could be oxidized in less than 5 seconds contact time
with the N02' but it was never clearly defined as to how much less could
be tolerated. The gas phase material balance on the condensation experi-
ments revealed that only about 40% of the S02 reacted during 0.46 seconds;
obviously an insufficient amount of time being allowed for the reaction. It
was decided to evaluate the kinetics of the S02/N02 oxidation reaction:
N02 + S02 ~ NO + S03
H20
Two additional experi.ments were performed using the condensation
apparatus where the residence time in the reactor was first doubled (0.92 S~
and then doubled again (1.84 sec), with all exit streams being analyzed as
before. The results are shown in Table XXV and compared with the data
from ~e 1040 F experiment shown in Tabl e XXIV.
Table XXV.
s 00 ~ ;C
Results of Kinetic Experiments ifBi 8Ft
N02 + S02
H 20( g)
-+ S03 + NO
Reactor
Residence Duration S02 Concentration
Time of Run in Exit Gas % S02
(sec) (min) (mm Hg) ** Reacted
0.46* 30 1. 15 43.6
0.92 60 O. 80 60.8
1. 84 90 0.40 80.4
* This run from Table XXIV
** S02 inlet concentration = 2.2 mm Hg.
- 118 -
-------
These two additional experiments were carried out under the
same conditions as the previous condensation one, with the gas composition
being the same and the condenser again being operated at 104 of.
(The condenser was not really necessary for these experiments, but
it was decided not to change the conditions.) The retention time was
changed, but the total volume of gas passed through the absorber
was kept constant in order to be able to analyze solutions of similar
concentrations.
The data in Table XXV show that the S02 concentration in the
exit gas decreased as the retention time increased, a not unexpected
result. From this data we can calculate the reaction rate constant
for these conditions. For the general reaction:
A + B ~ Products
where a and b are the initial concentrations of A and B respectively,
and the reaction rate is proportional to the concentration of both
reactants, we can write the following general equation for the rate
of reaction: 11
dx = k (a - x) (b - x)
dt
with x being the decrease in each concentration after time t. This
can be integrated using the boundary condition that x = 0 when t = 0:
t = 2. 303 log b (a -x)
k(a-b) a (b-x)
(4)
k, the reaction rate constant, can be determined by plotting t versus
2. 303 log b (a -x)
a-b a (b-x)
-
a
- 119 -
-------
for several values of t or k can be calculated directly for each case.
,
This was done for the three sets of data shown in Table XXV. The
calculations and results are summarized in Table XXVI.
t
(sec)
0.46
0.92
1. 84
Table XXVI. Calculated Reaction Rate Constant for S02
Oxidation by N02
a (S02) b (N02) x
(torr) (torr) (torr)
2.04 5.5 0.89
2.04 5.3 1.24
2.04 5.3 1.64
a*
k
O. 250
0.224
0.210
0.115
O. 206
0.386
*
- 2. 303 I b(a -x)
a - og
(a - b) a(b-x)
Figure 36 shows a plot of t versus a. The slope of the line
gives the value for k of O. 196. (Since the graph was plotted as
a = kt the slope gives k directly rather than Ilk. )
Although the data falls on a perfect straight line, there is a
slight error since the line does not pass through the origin. This
accounts for the differe11ce between the slope and the calculated
values of k. The results are good enough to allow comparison with
published values of k for this reaction. The most complete stUdy of
this reaction was done by I. N. Kusminych, E. J. Turchan and
M. S. Archipowa who determined values for k at several concen-
. d f 18
tratlons an a range 0 temperatures. Their temperatUre corre-
lation of k is shown in Fig. 37.
From the Kusminych data it would appear that k should be
0.372 at 300 of rather than the O. 196 found in our stUdy. An examina-
tion of the reaction chamber in Fig. 35 will reveal the source of this
discrepancy. Our apparatus was set up with the gas entering the heated
- 120 -
-------
a
.4
.3
.2
. 1
o
.3
.6
1.2
1.5
1.8
Fig. 36.
.9
t (8 ec)
a ~ t for Reaction 802 +N02 ~ 803 +NO
- 1 21 -
-------
~
.....,
c=
CtI
.....,
00
c=
o
u
3 .2-
CtI
H
c=
o
......
.....,
u
CtI
Q)
p::;
. 4
.3
o
. 1
150
300
Fig. 378.
200 225
Temperature (OF)
Reaction Rate Constant as a Function of 1emperature (Data
of Kusminych, Turchan, and ArchipowE\-) for Reaction:
802 +N02 -> 803 +NO .
- - 122 -
175
250
275
-------
reactor at about room temperature. Since the reactor heater was set so
that the exit temperature was 300 ° F, there had to be a temperature gradient
along the length of the tube. With the heat being introduced by a heating tape,
it is possible to approximate the temperature profile of the reactor by a
straight line relationship. This means that the resultant k should be repre-
sentative of an over-all reaction at a linearly averaged temperature. From
Fig. 37 it can be seen that a k of 0.196 is representative of this reaction
occurring at a temperature of 210°F. The linear average of the temperature
gradient in the reactor would be 185 ° F. We can interpret this difference as
either a nonlinear temperature profile along the length of the reactor or a
nonlinear relationship between temperature and reaction rate. In either case
the agreement between the two is good.
Using the data of Kusminych and equation (4), it is possible to cal-
culate the theoretical retention time necessary for 90% reaction at a constant
temperature of 300 ° F. With a k value of 0.37 and initial concentration of
2 mm Hg of S02 and 6 mm Hg of N02' it would take 1.32 sec to achieve 90%
reaction. If the N02/S02 ratio were reduced to 2/1 instead of 3/1, it would
take 2.3 sec to achieve the same goal. These residence times are within
the capability of the pilot plant reactor system.
3) High Temperature Scrubber
The concept of the high temperature scrubber as described in
Section IV-D is based on the idea of maintaining the scrubber at a temperature
such that the vapor pressure of water over the entering acid was equal to the
partial pressure of water entering with the flue gas. (See Table XXI for water
vapor pressure data.) By this technique all the water entering with the flue
gas would leave with the clean stack gas and the acid would not be diluted. The
primary question was whether or not the hot acid would absorb the oxides of
nitrogen at these temperatures and thus keep the N203 losses to acceptable level.
In order to develop reasonable design criteria for a scrubber of this
type, an analysis of the absorption capabilities of the system was conducted
using the McCabe-Thiele method.14 The following is a brief development of
the method and the results for the system under consideration.
- 123 -
-------
G
Y
i'
a I
b x
L
xa
G
Y
L
b
Fig. 38. Absorption column (NO Scrubber)
x
In considering the absorption of a gas in a liquid in a counter-
current scheme as shown in Fig. 38, it is possible to predict absorber
performance using the McCabe-Thiele method. An equilibrium line is
plotted as shown in Fig. 39, and an operating line is determined by the
terminal conditions of the column. The lower end of the operating line
represents the conditions at the tOp of the column, and the upper end, the
bottOm of the column.
A material balance can be taken around the column using the
f 11' . 14
o OWing notatlon:
xa = solute concentration in the inlet liquid
xb = solUte concentration in the effluent liquid
Ya = solute concentration in the effluent gas
Yb = solUte concentration in the inlet gas
L = liquid mass flow rate
G = gas mass flow rate
- 124 -
-------
McCabe-Thiele Plot
Catalytic Chamber Process - NO Scrubber
x
.IYO = 2. 6~
~= .13
2.8
---~~---..--,..--
- -Y~=2.-62-I/A
xb;': .11
Entrance gal:> concentration
2. 62 mm Hg
2.4
:r
E
E 2.0
(l)
~
:;!
C(J
(lJ
(l)
H
c..
H 1.6
o
0..
rn
>
CVJ
0
ZN
~
Proposed operating line
L/ G :::: 2.1 mole/mole
>,........
Operating line
.-- from Material
Balance (Equili-
brium Conditions)
LjG :::: 1.7 mole/
mole
1.2
Equilibrium line (250 OF)
Data of Berl and Saenger 4
Minimum L/ G :::: 1.6 mole/mol
.8
.4
o
o
.025
. 050 . 075 . 100
x, moles HNSO,c::/ L HZS04 (80% wt)
.125
. 150
Fig. 39. McCl;lbe-Thiele plot for NOx scrubber in Catalytic Chamber
process
- 125 -
-------
An assumption is made that Land G are constant throughout the
length of the column without a significant error. The material balance
for the solute is:
L xa + G Y b = L xb + G Ya
This can be rearranged to give:
Yb - Ya - L
xb - xa G
From examination of Figure 39; it is clear that L/G defines the slope
of the operating line.
The equilibrium line plotted in Figure 39 is taken from the data
of Berl and Saenger4 for an operating condition of 250 of as determined
earlier. (The temperature is based on a scrubber that is treating a gas
containing 6. 7% excess water with 80% sulfuric acid and must therefore
maintain a water vapor pressure of 51 mm Hg. The original flue gas con-
tained 7.3% water, with the difference being used to form 80% sulfuric
acid with the oxidized 502. )
Operating terminal conditions are based on data obtained from the
material balance in Table I. The one condition that cannot be varied is
the concentration of N203 in the gas entering the scrubber. From the
material balance:
8. 73 lb N203
Yb = 1000 lb gas
mol N203
76 lb N203
30 lb gas
mol gas
760 mm
Yb = 2.62 mm (as N203' or 6900 ppm NOx)
Again, using the material balance:
76 lb N203 mol N203
Ya =
106 lb gas 76 lb N203 mol gas
Ya = 0.023 mm Hg (as N203' or 60 ppm NOx)
30 lb gas
760 mm
- 126 -
-------
8. 65 Ib N203 mol N203 2 mol HNS05 454 g - mol HNS05
xb ;:;:
3021 Ib H2S04 76 lb N203 mol N203 lb-mol HNS05
x lb H2S04
4S4 g H2S04
1727 g H2S04
i H2S04 (80%)
Xb ;r::; O. 13 mol HNSOS/ 1 H2S04
0.066 lb N203
x =
a 3000 lb H2S04
xa = 0.001 mol HNSOS/l H2S04
(or 0.96% Wt HNS05)
2 11727
76
(or 0.007% wt HNS05)
The terminal points for this operating line are therefore:
Xa = 0.001 mol HNS05/ f. H2S04
Ya :; 0.023 mm Hg
Xb= 0.13 mol HNS05/f. H2S04
Yb = 2.62 mm Hg
This line is als9 plotted on Figure 39.
Since the data shown in this operating line were obtained from
equilibrium condition assumptions, one end of the line actually touches
the equilibrium line. This means that it would take an infinite number
of theoretical stages to perform the desired operation and it is, there-
fore, impos sible.
Given the equation for the slope of the operating line, it is
possible to obtain the L/ G ratio for this operating line and the minimum
L! Ehwhich iEl the slope of the equilibrium line. The minimum L! G is 1.7
~nd the L/ G for the material balance-based operating line is~. Os,
as given earlier.
In order to achieve the absorption desired in this process, it
is necessary to determine more realistic terminal points. If we assume
that the acid entering the column must contain some N203' then we have
to increase the aUowable effluent gas content of N203. At the same time,
a fa~ter separation can be achieved if the liquid flow rate is increased,
thus diluting the effluent acid. The following terminal conditions were
chosen for trial:
- 127 -
-------
Yb = 2.62 mm Hg (6900 ppm NOx)
xb = 0.11 mol HNS05/ £ H2S04
y = 0.046 mm Hg (120 ppm NO )
a x
xa = 0.001 mol HNS05/£ H2S04
This new operating line is also plotted on Figure 39, and the usual step"
wise McCabe-Thiele procedure is used to determine the theoretical number
of equivalent plates. The number of plates is 10, certainly a reasonable
number.
By increasing the liquid flow, the L/G ratio is increased. Using
the same material balance equation as before, the L/G for the new line
is 2.1 mole/mole. Obviously, by increasing this ratio (and thereby diluting
the effluent add), the value for xb will be decreased and the number of
theoretical plates will decrease. Alternatively, if more N203 is allowed
to escape to the atmosphere, fewer theoretical plates: are needed. The
figures given here for this operating line are based on 150 ppm of N203
escaping with the stack gas which represents 25% of the NO which came
x
in with the flue gas.
In order to evaluate the concept of high temperature scrubbing,
a 1 1/2-in. glass column was packed with 2 ft of Bed saddles and was
fed with countercurrent streams of N203 in N2 and 80% H2S04 at an L/G
ratio of about 1. 7 mole/mole. All entering streams and the column were
heated so that the absorption would take place at about 2500 F. All exit
- 128 -
-------
streams were analyzed for NOx content, and a complete material
balance was made. Table XXVII shows the column conditions and
the concentrations of the exit and entrance streams for two runs. It is
significant to note that without optimizing column height or operating
condi tions, recoveries in exces s of 98% were achieved wi th very low
*
material balance errors.
These data are very encouraging and show that there will be no
problem in scrubbing the reacted flue gas with 80% H2S04 at 250 of
and removing 99% of the NOx content. It will also be possible to reduce
the exit gas concentration to about 150 ppm NO. Obviously, use of
x
high te~perature isothermal. scrubber is a practical way of preventing
the absorption of water by the scrubber acid and thus eliminate the need
for the acid concentration.
4) Concentration of N20S in the Stripper
a) Ga.s Phase Catalytic Oxidation of NO
As mentioned above, about one-half of the
heat requiremept of the Modified Chamber Process can be eliminated if
a technique can be found to concentrate the N20S in the stripper gas to a
*Note: The actUal recoveries were probably higher than those shown
because of the difficulty in reading the NO peak in the IR spectrum used
to analyze the exit gases. At the high temperature used, there is a good
deal of water removed (about 50 mm Hg) and one of the water bands
partially dve.rlaps the NO band, thus making it hard to properly read the
exact NO concentration. The figures reported are the worst case with a
strong likelihood that more than 75% of the NO reported is really water.
Adjusting "the recovery for the water band shows recoveries of about
99.5% N02 and effluent NOx concentrations of less than 40 ppm. For
future experimentation, the NO can be recalibrated in the presence of
water to eliminate any error, but for the time being these results are
satisfactory.
- 129 -
-------
Table XXVII. Results of High Temperature
Absorption Experiments
Run 1
Run 2
Entering acid temperature, of
Column temperature, of
265
265
Top
240 255
255 255
300 300
24 30
Bottom
Entering gas temperature, of
H2S04 flow rate, ccl min
Entering gas flow rate, ccl min-STP
N203
N2
Entering N203 concentration, ppm,
Exit acid concentration, HNS05 -wt %
35 35
12000 12000
2900 2900
0.86 0.86
Exit gas concentration, ppm
NOx removal, %
<100 <65
10 10
110 75
10 13.8
> 98 > 98. 7
NO
N02
Total NO
x
Material balance error, %*
*Material balance error = NOx (in-gas) -
[NOx (out-gas) + NO (out-acid) J
x ,
;
NO (in-gas)
x
- 130 -
-------
level which will allow the oxidation of the NO content of the gas in a
reasonable volume without heating the nitrose. In the originally proposed
chamber process, the limitation is the vapor pressure of N203 above the
hot nitrose entering the stripper from the scrubber. Therefore, in order
to obtain a concentration of N203 in the stripper gas that is appreciably
higher than the concentration in the reacted flue gas, the nitrose has to
be heated considerably.
If a high temperature stripper is operated as in Fig. 40, it is
obvious that the volume of gas necessary to recover the N203 is exactly
equal to the original amount of flue gas (if we assume equilibrium con-
ditions). This is clearly impractical since the concentration of N203
would be so low as to require enormous contact time with the 2-3% oxygen
present in the flue gas to effect 100% oxidation to N020
The only alternative is to remove the dissolved NO by thermal
x
stripping, which brings us to the problem of handling recycle water. If
the stripper is operated hotter than the scrubber so as to raise the vapor
pressure of N203' less gas is needed to carry off the NOx then was present
in the original flue gas. The N203 would then be more concentrated, thus
speeding up the reoxidation process.
Figure 41 shows such a scheme where the scrubber is operated
so that the gas exiting to the atmosphere contains as much water as enters
with the reacted flue gas. At the same time, the stripper conditions are
such that only 31% of the original flue gas at 12500 F is required to heat the
acid to a temperature such that the concentration of the N203 in the exit
gas is raised to about 1 %. This amount of required heat is equivalent to
burning an additional 18% coal, which might well be within the capabilities
of the average power plant.
The heat requirement is so low because the acid no longer has to be
heated to boiling ( 3950 F for 80% acid). This is because reconcentration
of the acid is not required since no water has been removed from the
original flue gas stream. The only transfer of water in the system is
- 131 -
-------
5800 ppm N02
y
NO Oxidizer
Flue Gas
5% H20
2.8% 02
2400 F
80% H2S04
0.140 moles HNSOS
per liter ( 1 % Wt)
240 of
12~ooppmN2031
5% H-20
2400 F
80% H2S04
240 of
Fig. 40. High Temperature Stripper with NO Oxidizer
- 132 -
-------
------1
Reactor
NO Oxidizer
o. 69 of Total Flue Gas
300 of
5% H20
240 of
5% H20
80% H2S04
240 ° F
~
Co_oling
Water
Cooling
Water
7.9% H20
2900ppm N2 3
2400 P
240 of I
79% ~2S0 4
1% HNS05
~-85:-;-.
8700ppmN203
14% H20
Stripper
80% Ii?S04
285 ° 1-<-
0.31 of Total Plue Gas
1250 0p
5% H20
Fig. 41. Modified High Temperature Process
- 133 -
Cooling
Water
-------
the vaporization of water in the stripper and the reabsorption of
the same amount of water in the scrubber.
One matter that must be determined is the concentration of
HNS05 in the acid leaving the stripper. (Since the stripper will not
operate under equilibrium conditions, it is hard to predict the actual
operating conditions and HNS05 concentrations. These are probably
best determined experimentally.) The amount of N203 in the recycle
acid will determine the amount of NO lost to the atmosphere since
x
the stack gas is in equilibrium with this acid stream. It is important
to keep this to a minimum in order to take advantage of the opportunity
to manufacture and sell nitric acid using the NO in the original flue
x
gas as the raw material source.
In order to make this system operate properly, three heat
exchangers would have to be added, as shown in Fig. 41. The flue
gas will have to be cooled so that it enters the scrubber at 250 of.
rather than the 330 of of the combined reacted flue gases. Heat
will have to be removed from the scrubber in order to maintain the
isothermal condition. This excess heat is developed by the absorption
of N203 and the solution of water in the scrubber. The third heat
exchanger is used to cool the acid leaving the stripper from 285 of
back to 250 of for recycle. This last quantity of heat is the largest
heat loss in the system and may have some value, although it is not
a high quality heat source.
After leaving the stripper the gas contains N 203 which must be
oxidized to N02 before it is mixed with the remaining flue gas and fed
to the reactor. The reaction between NO and 02 occurs faster at
lower temperatures which would require the gas to be cooled at this
point. However, even if the gas were to be cooled, the contact time
required for the uncatalyzed reaction would be quite long, assuming
that the flue gas contains 2. 8% oxygen. Table XXVIII shows the contact
times necessary to achieve 80 or 90% reaction with this level of oxygen
with and without a charcoal catalyst. The data are plotted in Fig. 42.
- 134 -
-------
Table XXVIII. Contact Time for Oxidation of Nitric Oxide (2. 8% 02)
NO
Concentration
(%)
0.3
0.3
1.0
1.0
5.0
5.0
Degree of
Reaction
(%)
90
80
90
80
90
80
Reaction Time (see)
Car bon
Catalyzed Uncatalyzed
22.3
9.9
7.5
3.2
2.5
1.4
4667
2062
1550
668
525
284
The data in this table are based on the reaction rate equation
developed by Burdick and given in Section V-E.
Burdick found that below 50 0 C, k had a value of about 12
and dropped as the temperature of the uncatalyzed reaction was
increased. He found the same temperature effect with catalyzed
reactions of dry gases, while the rate constant rose to about
10, 000. If moist gases were used, the value of k dropped to about
1000 but now increased as the temperature increased. He accounted for
this by assuming that the moisture in the gas tended to adsorb on
the charcoal, using up active catalytic sites, but heating desorbed
the water, thus increasing the activity of the catalyst.
Thus, it appears to be advantageous to catalytically oxidize the
NO at as high a temperature as possible. The gas would not have to be
cooled and relatively low concentrations of NO could be tOlerated ("'1%).
Examination of Table XXVIII and Fig. 42 reveals that very fast reaction
times can be achieved with 2.8% oxygen and NO concentrations above 1 %.
The catalyst so markedly increases the reaction rate that consideration
might be given to using all the flue gas in the stripper at low temperature
and thus avoid the water recycle and the sensible heat loss incurred by
.the thermal cycling of the acid.
- 135 -
-------
4000
u
iJ)
CI)
3000
iJ)
E
~
s::
o
.,....
.....
u
C\3
iJ)
0:::
'1:J
iJ)
N
~ 2000
.....
cO
U
s::
:J
1000
b
Each curve can be
read for uncatalyzed
times (left scale) or
for catalyzed times
(right scale).
80%R
eqCt,
1011
o
o
1
2
20
15 u
iJ)
CI)
..
iJ)
E
~
s::
o
'.....
.....
u
cO
iJ)
0:::
'1:J
iJ)
10 ~
.-4
5
cO
u
5
'8,
o
o
3 4
NO Concentration. %
NO Oxidation Times for Charcoal Catalyzed and Uncatalyzed Reactions (2.8% 02,
25-500 C) to Achieve 80% and 90% Yields From the Reaction:
NO + 1/2 02 -> N02
Fig. 42.
- 136 -
5
-------
Use of this catalytic NO oxidation technique with the process
shown in Fig. 41 would require the use of 18% additional coal to
provide the heat needed to concentrate the N203 to about 1%.
This would permit 90% oxidation in 7.5 seconds (or 80% oxidation
in 3. 2 sec), thus reducing the size of the required oxidation chamber
to a more reasonable volume. In a system designed for an 800 mega-
watt power plant (7. 25 million lb/hr of flue gas) this would mean using
an oxidizing chamber 25 ft in diameter and about 50 ft high to achieve
80% oxidation.
These schemes depend on maintaining the proper balance of
N02 and NO in the various stages of the system. If the NO oxidizer
were set up to convert only 80% of the NO to N02' then additional N02
must be kept in the system so that after the reaction with S02 the N02
to NO ratio would be 1/1. This would also change the concentration of
HNS05 in the scrubber effluent acid and would affect the operation of
the stripper.
b) Liquid Phase Reduction of HNS05
A lthough the catalytic oxidation of NO as des-
cribed in the previous section has a great deal of merit and does result
in the reduction of the extra process heat needed, it still requires 18%
additional fuel. If the catalytic oxidation of NO could be accomplished after
a stripping operation that did not require this additional fuel; a further
saving could be realized.
Basically. what must be done is to reduce the solubility of the
oxides of nitrogen in the sulfuric acid. As was discussed in an earlier
section, NO and N02 alone are relatively insoluble in sulfuric acid, bur
in equimolar quantities they are very soluble. Since the gases are
absorbed in the scrubber in equimolar quantities they will be present
in the nitrose in that mole ratio. In order to accelerate their removal
from solution it would be necessary to unbalance the mole ratio, either
by oxidizing the NO to N02 or reducing the N02 to NO. '
- 137 -
-------
A methcxl was suggested for the liquid phase reduction of
N 203 to NO and thus speed up the vaporization of the oxides. In
choosing the technique we kept in mind that it was important to
avoid using a reagent that would leave an unwanted component
in the acid after the denitration was accomplished. The choice
for the reducing agent was ammonium sulfate, assuming the
following reaction:
10 HNSOS + (NH4)2 S04 + 2 H20 = 12 NO + 11 H2S04
The experiments consisted of a flask containing the HNS05 and
(NH4) 2S04 solution. The air above the solution was purged with nitrogen
to avoid oxidation of any resultant NO before it could be analyzed. The
analysis was performed with the Perkin-Elmer one-meter folded-path gas
cell in the Model 457 infrared spectrophotOmeter.
No visible reaction occurred at room temperature, but a
vigorous bubbling was observed above 60 0 C with the effluent gas
being only slightly brown in color. Analysis of this gas showed
very low concentrations of NO and N02' considerably lower than
their equilibrium vapor pressures at this temperature. The experi-
ment was rerun using C02 as the inert gas in order to determine
whether or not the reduction had gone all the way to nitrogen. The
effluent gas was virtUally colorless. The analysis showed no oxides
of nitrogen and about S mm Hg of C02' indicating that all the N203
in solution had indeed been reduced to nitrogen (which cannot be
analyzed by infrared techniques). Analysis of the remaining solution
revealed almost complete removal of the HNSOS' An experiment was
run where an attempt was made to catalyze the reaction with charcoal
at lower temperatUres, but the results were not encouraging. The con-
cept still appears iiiteresting, but work was discontinued until a better
reducing agent could be found.
- 138 -
-------
c) Liquid Phase Catalytic Oxidation of HNSOS
A clue to a possible answer to the problem
was obtained by examining the reactions occurring during the nitrogen
oxide recovery and reoxidation:
2 HNSOS + H20 ~ N203 + 2 H2S04
N203 -+ NO + N02
NO + i 02 -+ N02
2 HNSOS + H20 + i 02 -+ 2 N02 + 2 H2S04
The net reaction equation suggests that the process is in effect,
the oxidation of nitrosylsulfuric acid, implying that if a way could
be found to oxidize the HNSOS in the liquid phase, the process might
be accelerated.
This idea was evaluated experimentally by passing a nitrose solu-
tion countercurrent to air in a packed column and measuring the N203
concentration of the effluent gas. As might be expected from equilibrium
considerations, the N203 content of the gas was very low. In order to
accelerate the oxidation of the nitrose it was decided to push the reaction
by packing the column with some material that might reasonably be ex-
pected to act as a catalyst. Activated charcoal was decided on because it
was known to catalyze the air oxidation of nitric oxide. The previous
experiment was rerun using an activated charcoal packing in the column
with the visual result that the acid foamed when it contacted the charcoal
giving off brown fumes which were apparently N02' Analysis of the
effluent gas str.eam showed an enormous increase in the N02 content,
well in excess of the equilibrium vapor pressure above the nitrose under
these conditions. The results are shown in Table XXIX.
Clearly, the presence of charcoal had a startling effect on the
solubility of the nitrogen oxides in sulfuric acid. Analysis of the effluent
acid showed that effecti vely all the nitrogen oxide content of the nitrose
had been removed and transferred to the gas stream without the limitation
of equilibrium vapor pressures to restrict the oxide concentration of the gas.
.., 139 -
-------
Table XXIX.
Results of Catalytic HNS05 Oxidation Experiments*
Exit Gas Concentration (mm Hg)t
-
--""--
~
Carrier Gas
Packing
NO
N02
S3ddles
<0.01
< 0.01
< 0.01
0.024
N2
Air
Saddles
Air
Charcoal
<0.01
0.75
* Test conditions:
Temperature = 25°C
Asid flow rate = 10 eel min
Acid concentration = 0.7% HNS05 in 80% H2S04
Gas flow = 0.25 SOFM.
Packing height = 20"
Column diameter = 1.5"
t All concentrations were taken 5 minutes after the gas started
to pass through the liquid and packing. Equilibrium vapor press of N203
under these conditions is 0.15 mm Hg. Exit acid showed no residual
NO content.
x
- 140 -
-------
Although all the NOx content of the acid had been removed (as
determined by analysis of the effluent acid), a material balance around
the column disclosed that only one-third of the NO that had apparently
, x
been fed to the column had been recovered as N02 in the effluent gas.
There are two contributing factors to this anomaly: (1) experiI11ental
error in measuring acid flows and gas concentrations, and (2) the partial
reduction of the N02 to N20. The largest source of N02 disappearance
was undoubtedly the formation of N20 which was not measured during
these experiments, but was detected in subsequent experimentation.
This problem is discussed below. The important conclusion to be derived
from these tests is that this treatment completely removed the NOx con-
I
tent of the sulfuric acid.
This experimentation was the basis for the catalytic stripper
discussed in some detail in Section IV - F . This breakthrough makes
possible the reoxidation of the oxides of nitrogen without the additional
heat requirement of the Modified Chamber Process, thereby saving a
considerable heat expense.
One problem that has appeared in these experiments was that
the N02 seemed to be reacting with the carbon forming C02 and N20,
which caused the charcoal bed to break up and form fines which plugged
the column. A series of experiments were performed to try to determine
the cause of these changes and ways to avoid them.
(1) Degassing the charcoal. On the suggestion of applications
chemists at Witco Chemical Co. who have an extensive background in
!
uses of activated carbon, we attempted to slow down the oxidation of
the carbon by removing the adsorbed oxygen from the surface of the
carbon particles. Although it is known that the adsorbed oxygen plays a
significant role in the catalytic behavior of the carbon, it was felt that
there was excess activity available on the carbon and some could be
sacrificed in order to increase stability.
Samples of activated carbon were kept in boiling water for pro-
longed periocl"s of time to degas the particles before using them to
pack a column. The column passed water very easily, but when the
- 141 -
-------
air was passed countercurrent to 0.7% HNS05 solution in 80% H2S04'
the effluent acid rapidly became discolored with finely divided carbon
particles and the column soon plugged up and no liquid could pass,
although the air was able to go through the bed. It was apparent, from
the evolution of brown fumes from the top of the column, that HNS05 was
being oxidized to N02. The gas analysis showed the presence of N02'
N20, and C02' although quantitative measurements were not possible.
Other experiments were made with this column using concurrent flow
of gas and liquid, but there was only slight improvement. It was obvious
that the degassing did not prevent the breakdown of the carbon.
(2) Other forms of carbon. Again working on the hypothesis that
some of the catalytic sites on the carbon were too active and were causing
the oxidation and breakdown of the particles, we tried using coal (both
hard and soft types) and coke to catalyze the reaction. Batch experiments
at room temperature showed no reaction occurring and the experiments
were discontinued.
(3) Silica gel. Batch tests were run with particles of silica gel
used to attempt the catalytic oxidation of the HNS05. Using 80% H2S04
solutions of the nitrose, it was found that no N02 was evolved even at
temperatures above 2000 F. In addition, the gel particles appeared to
be absorbing water and shattering in the process. It is possible that the
active sites on the silica gel were involved in absorbing water and were
not available for HNS05 oxidation, but the fact that this process will
probably not be feasible for acid concentrations above 92% tends to miti-
gate against further experimentation with silica gel.
(4) Water glass treatment. There is evidence in the literatUre
that the oxidation of activated carbon can be inhibited by coating the
carbon particles with a layer of silica formed by precipitation from water
glass solutions. The carbon is first washed with a dilute (5%) solution of
- 142 -
-------
sodium silicate and then washed with a 20% HCI solution to precipitate
the silica.
This treatment was performed on a batch of activated carbon from
Witco (type 256, particle size 4 x 10 mesh which was then packed in a
column. Countercurrent contact of a nitrose solution with air produced
N02 in the gas stream but also produced an effluent acid stream that was
discolored although considerably lighter than that with untreated carbon.
A second batch of the same carbon was treated three times with
the water glass and another column was prepared. A number of experi -
ments were performed with this column; the results and experimental
conditions are shown in Table xxx.
It is significant that the water glass treatment did prevent the
oxidation of the carbon while maintaining the catalytic activity of the
activated charcoal, and that the remedial effect was only temporary, with
catalyst oxidation and particle breakdown occurring after several hours of
operation. The overall results suggest that the washing with 80% H2S04
has removed the protective silica coating, implying that it would be
worthwhile to try to find a coating material that is not soluble in 80 to
92% H2S04.
An unexpected result was the heating of the column. As shown in
the table, the column temperature rose as soon as the acid contacted the
charcoal. Prior to the first run, the acid was passed through the column causing
a temperature rise from 70 to 160 ° F. When nitrogen was passed through the column
bed, the temperature dropped to 135°F and remained at this steady state
value throughout the first four experiments, which took about 4 1/2 hr of
continuous running to complete.
It is also surprising that the oxidation of the N203 took place in
the first four experiments without the presence of oxygen. It would seem
that some reaction took place between the nitrose and the charcoal, but it
is not clear what was happening. For the N203 to be oxidized, another
- 143 -
-------
Table XXX. Results of Catalytic Oxidation of HNS05 Using Water-Glass Treated Charcoal
Experi - Column Effluent Gas Composition, mm Hg N203 Material
ment Carrier Temperature, NO N02 N20 C02 Remo.+ed Balance *
No. Gas °F* % Error, %..
1 N2 135 1.8 1.9 <0.01 0
2 N2 135 12.0 O. 65 <0.01 0
3 N2 135 2.9 0.30 <0.01 0 85 -20
4 N2 135 4. 1 1.7 <0.01 0 88 +22
..... 5 Air 95 0.03 1.9 <0.01 0 78 -4
~ i
~
6 Air 85 <0.01 0.30 <0.01 0.08 45 -34
.
7 Air 80 <0.01 O. 17 <0.01 0.20 37 -31
8 Air 105 <0.01 1. 95 <0.01 1. 05 88 -53
*Sready state temperature.
+Based on analysis of the effluent acid.
:f:Minus error indicates less N203 recovered than was put in the columr::
NO (in) - NO (out)
Material balance error = x x x 100
NO ( in)
v
-------
material had to be reduced, and there is no evidence that this occurred
(very little N20 was formed and the material balance deviation was
relatively small).
The next three experiments (performed after the column had been
washed with water for 30 min and allowed to sit wet (overnight) show a
sharp decrease in the amount of NO being formed (which is expected,
since the oxygen present would oxidize the NO to N02) and an erratic
quantity of N02 in the exit gas. At the same time, the amount of C02
increased, suggesting that the silica was wearing off, allowing the car-
bon to be oxidized. At this time, the effluent acid started becoming
black with suspended carbon fines. However, it is hard to understand
why the NO and N02 production decreased when the silica started to
wear off. It mi~ht be said that the newly formed N02 was reacting with
\
the charcoal and was reduced, but then there would have been more N20
present in the exit gas (and the IR analysis spectrum showed no sign of
a reduced species).
It is clear that a good deal more work is needed to understand
the behavior of the silica treated charcoal. A decision will have to
be made as to whether it is worthwhile to continue to work with the
charcoal or to devote our experimental effort toward finding other
suitable catalysts.
(5) Results of Process Modification Efforts. The goal in this
work was the elimination of the additional heat requirements demanded
by the Modified Chamber Process. The resulting process, dubbed the
Catalytic Chamber Process as described in Section IV, accomplishes
this goal to the extent that the process is now slightly exothermic.
Additional work is required to determine operating conditions, catalyst
systems and general engineering data, nut the concept appears soundly
supported by reliable dAta and gives promise of being a workable solution
to the power plant flue gas pollution problem.
- 145 -
-------
APPENDIX 1.
Ultraviolet Spectrophotometric Analytical Techniques
Sulfur dioxide and nitrogen dioxide were analyzed in dilute gas mix-
tUres using a modified Spex Model 1700 single beam infrared spectrophoto-
meter and an external hydrogen arc light source. The sample was placed
in an optical cell (with quartz windows), similar to the one in Fig. 43 and
positioned in the light beam between the source and the entrance slit of the
photometer. Changes in transmission due to absorption by the gases in
the sample cell were monitored on a strip chart recorder. The full scale
recorder response time was about 5 seconds.
Figure 44 shows the ultraviolet absorption of S02 as a function of
partial pressure as determined experimentally by Tyco. The reproduci-
o
bility of the data at 2860 A was quite good in the 0.1 - 1% range. Similar data
o
for N02 absorption at 5100 A are shown graphically in Fig. 45. Again the
reproducibility of measurement was good, and the sensitivity more than
adequate in the concentration range of 1 - 10 volume %.
Since N02 absorbs to a limited extent at 2860
-------
-+
,t>o
OJ)
Quartz
Window
Gas Inlet
~
H20
l~
Path Length
H20
H20
tOUT
Jacket
Quartz
Window
Fig. 48. Water-jacketed UV absorption cell (15 ml volume).
-------
w
>
i= 0.4
«
...J
w
a:: 0.3
0.2
1.0
10
PS02 mm. Hg
Fig. 44. Absorption of SO? at 2860 J.. (Tyeo
Experimental data.)
-149 -
-------
1.0
~I~
~
>-
I-
en
z
w
I-
Z
w 0.10
>
I-
-------
balance was assessed due to S02' The accuracy of this procedure is
verified by the observation that, using a linear plotting method, extrapolation
of the reaction rate data to zero time yields a value for initial S02 concen-
tration that corresponds closely to its actual level.
A major drawback to the UV method is its inability to detect nitric
oxide in the concentration range of interest. As a result many of the gas
mixtures analyzed during the experimental work have been incompletely
evaluated, making it impossible to do complete material balances. As a
result, although the UV method is quick and dependable, a new method had
to be developed which would permit the analysis of all gaseous species in
multicomponent mixtures. The infrared detection method that was adopted is
presented in Appendix 2 along with wet chemical methods for absolute quanti-
tative measurements.
- 151 -
-------
APPENDIX 2. Analyti€al Techniques
A.
Introduction
Experimental work performed during the early part of the contract
clearly pointed out the need for accurate, reproducible techniques for analyz-
ing all of the various process streams in our pilot plant system. Without this
capability for complete analysis, material and heat balances cannot be worked
out, making it difficult to justify conclusions based on incomplete data. Work
has been completed on a series of analytical procedures which will make possible
the complete analysis of all process streams. The following discussion de-
scribes the spectroscopic and classical wet chemical techniques that are used
to analyze the various system streams and some of the work that has been done
to further develop these methods. A list of relevant band frequencies for the
spectroscopic determinations is given in Table XXXI. All possible gas phase
components (NO, N02 ' 502 ' CO , C02 and N20) are determined by the
infrared method.
B.
General Gas Handling Methods
The flow rates of all gases used were measured with Brook Instrument
Nos. 601, 602 and 604 flowmeters which were calibrated against a standard
bubble flowmeter for each of the gases in question. This technique was followed
for each of the gases used except N02 which required, due to its reactivity and
physical properties, the use of special analytical techniques for flowmeter cali-
brations. A discussion of these techniques will be given in the section discuss-
ing nitrogen dioxide.
The general calibration procedure employed for each of the gases con-
sisted of mixing measured streams of the desired gas and nitrogen and passing
this mixture through the optical cell.
The spectroscopic work was accomplished through the use of a one
meter folded path Perkin -Elmer gas cell with sodium chloride windows and
a Perkin -Elmer double beam infrared spectrophotometer (Model No. 457). As
the flow rates of each of the components was known, the partial pressure of
the gas could then be calculated. Then by obtaining a log-log plot of partial
pressure vs. relative intensity for a particular band, a calibration curve
was made for each of the desired gases.
- 15~ -
-------
TABLE XXXI.
Infrared Spectral Band Frequencie s
for Several Gases Of Interest
-1
Band Frequencies (cm )
Gas
NO
N02
S02
N20
1900 *
2950 *, 1750, 1600, 1260, 750
2500, 1360, 1150 *
3490, 2800, 2570, 2470, 2440, 2230 *,
2210, 1300, 1270
2160 *, 2110
3640 *, 2260, 700
CO
C02
*
Denotes band most useful for quantitative analysis work.
- 154 -
-------
c. Nitrogen Dioxide
Two methods, one chemical and one spectroscopic, have been used to
analyze N02 in the gas phase. The chemical technique involved absorption of
the N02 in 3% hydrogen peroxide, followed by titration of the resultant HN03
with standard NaOH in the presence of methyl red. This method is only specific
for N02 if there are no acidic species present in the gas, thus making it unusable
for analysis of the effluent scrubber gas which may contain sulfuric acid mist, NO
and 502. The major utility of this method is for the analysis of pure gas streams,
for instance, in the calibration of flowmeters and infrared spectra.
Since this is an absolute method, a few precautions must be taken to insure
accuracy. At least two absorbers must be used in series to make sure that all
the N02 that enters the system is absorbed for analysis. In addition, care must
be taken to avoid a significant amount of void gas volume in the tubing leading to
the absorbers. This can be eliminated by using capillary tubing and passing through
a volume of gas several hundred times the void volume of the tubing. Data ob-
tained in the laboratory indicates an accuracy of :t: 3% in the calibration of flowmeter So
In the handling of N02 in the laboratory some unusual procedures had to
be followed: As the gas condenses at 21 °c at atmospheric pressure, the portion of
the calibrating system through which pure undiluted N02 was passing had to be kept
in a constant temperature box to avoid condensation problems. The temperature of
this box, during all calibrations and operations was maintained at 35 ° c.
In order to use the infrared analysis method, a series of calibration spectra
were obtained using known concentrations of N02 (using a flowmeter calibrated with
the above chemical method). The resulting relative intensity vs. concentration curve
was then used to evaluate unknown gases. The bands of interest were 2950, 1750,
1600, and 750 cm -1, with the 2950 cm -1 band appearing to be the most useful.
Figure 46 shows a typical calibration curve for 2950 cm -1.
Some problems have arisen through the use of the cell which require
careful analysis of the spectral data obtained with it. First of all, the
window material appears to be reacting with the N02 to form a permanent
layer of sodium nitrite. This does not appreciably affect the trans-
- 155 -
-------
1.0
o
o
c
c
.....
CJl
(j)
>.
.....
.....
en
c:
(J)
.....
.5 . 1
(J)
:::-
.....
.....
CC!
,.....;
(J)
p::;
o
o
o
o
o
Fig. 46.
I I
.1 1.0
N02 Partial Pressure, mm Hg
Nitrogen Dioxide Calibration Curve 2950 em -1.
L
10.0
.01.
.01
-------
parency of the window, although it does create a sharp peak at 1355 cm -1
which has some influence on the 1260 cm 1 N02 peak.
Another more significant problem is the apparent decrease in N02
concentration with time when the gas sample is left in the cell and spectra
taken over an extended period of time. As the N02 decreases, three new
peaks appear at 880, 1330 and 3550 cm -1 which suggest the presence of
organic nitrates or nitrites. These may result from the reaction of N02
with adhesives or seals in the cell. This phenomenon indicates that the
spectrum must be obtained as quickly as possible to avoid the effect of this
reaction. Data obtained over a long period of time indicate that errors in
concentration will be less than a few percent if the spectrum is taken
within 20 minutes, a condition well within the capabilities of the equipment.
This work has demonstrated that the N02 can be analyzed in the
100 10,000 ppm range with an accuracy of 10% if the proper precautions
are taken. The sodium nitrite peak has been shown to have no effect on the
analysis, and the time dependent N02 concentration drop can be neglected
if the sample is analyzed fast enough.
D.
Su Uur Dioxide
Gas phase S02 can be analyzed by spectroscopic means similar to
the technique used above for N02' The peaks of interest are at 2500, 1360
and 1150 cm -1, with the 1150 cm -1 band appearing to be the best. (This "best"
peak has been chosen for all gases on the basis of reproducibility and usable
concentration range.) No problems have been encountered with this gas and
reproducible results in the 100 - 10,000 ppm range have been obtained with
an experimental error under 10%. Figure 47. shows a calibration curve for
-1
502 at 1150 cm .
E. Nitric Oxide
NO can be analyzed in gas streams using the same spectroscopic
technique as described above, with the high frequency peak of the NO doublet
band at 1900 cm -1 being utilized. (See Fig. 48 for calibration curve.) Care
must be taken in analyzing gas mixtures containing NO and oxygen since the
NO slowly oxidizes to N02' Since the NO oxidation is very slow at low
concentrations it is possible to obtain very good data by measuring the spectral
- 157 -
-------
1.0
....
t11
00
c
......
en
C
Q)
~
..s.1
Q)
:>
......
~
Ctl
..-(
Q)
p:::
.01
.01
.. 1 S02 Partial Pressure (mm Hg) 1. 0
Fig.. 47. Sulfur Dioxide Calibration Curve 1150 em -1
10..0
-------
1.0
>.
~
......
rn
s:::::
Q)
~
s::::: 1-
- .
I Q)
...... .::;
c:J1 ~
co qj
.......
I Q)
p::;
.01
.01
. 1 NO Partial Pressure (mm Hg)
Fig. 48. Nitric Oxide Calibration Curve 1900 cm -1.
1.0
10.1
-------
intensity drop over a period of time and extrapolating back to zero time
for original concentration. Experimentally it was determined that at
the low NO concentration levels expected in the operation of the pilot
plant it would be possible to ignore this oxidation if the oxygen content
were very low and the sample evaluated immediately. Errors of less than
10% would result if care were taken in analyzing the sample.
F. Nitrous Oxide (N 2<21-
In a few instances N20 has been found in trace quantities as an
unwanted reduction product. Gas phase N20 can be determined spectro-
scopically by the same techniques used for NO and S02. The observed
peaks in its spectrum are at 3490 cm-l, 2800 cm-l, 2570 cm-l, 2440
cm-l, 2210 cm-l, 2230 cm-l, 1300 cm-l, and 1270 cm-l, with the
2230 cm-l band appearing to be the most usable. (The other bands dis-
appear too rapidly with decreasing partial pressure to be of use.) No
problems have been obtained in the 100 - 10,000 ppm range. Figure 49
shows a calibration curve for 2230 cm -1. The flatness of this curve makes it
difficult to use above .01 mm Hg, but it has been of value to determine qualita-
tive N20 levels.
G. Carbon Monoxide
Carbon monoxide may appear as an unwanted product from
reactions between nitrogen oxides and be charcoal catalyst used in the
modified stripper. Gas phase CO can be determined spectroscopically
by the same methods used for the other gases. The observed peaks for
this material occur at 2160 and 2110 cm-l. The 2160 cm-l peak is the
most intense and has been used for calibration purposes. Results are
shown in Fig. 50.
H. Carbon Dioxide
Gas phase concentrations of carbon dioxide can be determined
by the infrared method. The principal bands occur at 3640 2260 and
-1 -1 '
700 em . Of these, the 3640 cm band appears the most satisfactory
for use in calibrations. Figure 51 shows the calibration curve for this
band.
- 160 -
-------
.-
0)
.-
>.
.....
.....
rn
c=
(I)
.....
.5 . 1
(I)
>
.....
.....
-------
1.0
>.
~
......
(I.J
c::
Q)
~
.5 . 1
.... Q)
0) >-
N ......
~
I Ct!
~
Q)
p:::;
.01
001
,
Fig. 50.
I
. 1 CO Partial Pressure (mm Hg) 1. 0
Carbon Monoxide Calibration Curve 2160 em -1.
10.0
-------
1.
>.
4...J
-.-I
00
~
I Q)
4...J
J-£ ~ 1
0) _.
Co\) (])
>
'.-1
4...J
as
r-4
Q)
p::;
.01
.. 01
o
Fig. 51.
I
. 1 C02 Partial Pressure (mm Hg) 1. 0
Carbon Dioxide Calibration Curve 3640 em -1
10.0
-------
I. Nitrosylsulfuric Acid
A classical wet chemical analysis method is being used to
analyze the liquid effluents from the scrubber and the stripper to determine
the nitrogen oxide content of the streams. Aliquots of the acid solutions
are titrated with potassium permanganate (0. 1M) which acts as its own
indicator, to a faint pink end point which remains for at least 5 min. It
may be noted that, near the end of the titration, the reactions involved
proceed more slowly due to the very low concentration of NHS05 remaining.
This procedure depends on the hydrolysis of the HNS05 to HN02 which is
then oxidized by the permanganate to HN03:
NHS05 + H20 -+ H2S04 + HN02
5 HN02 + 2 HMn04 -+ HN03 + 3 H20 + 2 Mn (N03) 2
One problem with this technique is that it will not accurately
determine any excess dissolved N02' which is hydrolyzed to an equi-
molar mixture of HN02 and HN03' Since the nitric acid cannot be
oxidized any further, the permanganate will not react with half of the
N02 dissolved in the acid. (It should be emphasized that all the N02
associated with N203 will be determined because no HN03 is obtained
when the N 203 is hydrolyzed.) This will probably not be a significant
problem since very little N02 should be dissolved in the acid.
The technique has been tried several times in the laboratory and has
shown good reproducibility (about 1%).
One precaution that will be necessary is to avoid very high
concentrations of sulfuric acid in the analysis. Titration of 50% nitro-
sylsulfuric acid in 98% H 2S0 4 with 0. 1 molar KMn04 resulted in the
liberation of a great deal of heat which caused the nitrosylsulfuric acid
to break down with the resulting evolution of brown NO fumes which were
2
lost to the analysis. When diluted with 80% sulfuric acid so that the H2S04
- 164 -
-------
and HNS05 concentrations were lower (about 2% HNS05)' the temperature
did not rise as much during the titration and no N02 evolution was evident.
Care must be taken to add the permangate solution sufficiently slowly so
as to avoid significant heating of the nitrosyl solutions. The use of a
surrounding ice bath may be employed.
J. Sulfuric Acid
Sulfate will be analyzed using the standard barium sulfate
gravimetric technique whereby the solution is treated with an excess
of a barium chloride solution and the resulting BaS04 precipitate is washed
and weighed. This procedure was tried several times in the laboratory
and it appears that it is the type of procedure that takes experience to
perfect. The precipitate has to be properly washed and dried before
being weighed, but once this technique is mastered, good results can be
obtained.
K. Total Acidity
Acid solutions will be titrated with standard sodium hydroxide
to obtain the total H+ content.
L. Water
The water content of gas streams will be measured through the
use of a hygrometer. The primary application for this technique will be
in the incoming flue gas line where the hot flue gas from the natura} gas
burner (with a 20% water content) is diluted to a cooler gas (about 400 0 F)
with a 5% water content. The hygrometer will be put in a cooled side
stream (about 100 OF) which will then be vented. The water content of
the incoming flue gas can be adjusted by changing the dilution ratio. The
side stream will also allow variation in the flue gas flow without changing
the burner or dilution settings.
- 165 -
-------
APPENDIX 3.
Visit to an Operating Chamber Plant
On August 9, 1969, Dr. Keilin and Mr. Walitt visited a
Chamber Process Sulfuric Acid Plant operated by U. S. Steel Agri-
chemicals, a subsidiary of U. S. Steel engaged in fertilizer manufacture.
The plant was located in Nashville, Tennessee, and produced sulfuric
acid exclusively for use in an adjacent fertilizer plant.
There was interest in visiting an operating plant for two reasons.
First, the Tyco/EPA Modified Chamber Process for removal of sulfur
dioxide from power plant flue gas is very similar to the original Chamber
Process and we wanted to get a good idea of actual plant operating charac-
teristics. Second, the newly modified Catalytic Chamber Process could
well improve the process economics of the Chamber Process to a degree
where it could again compete with the Contact Process, and it was there-
fore important to determine the real limitations of the old process. U. S.
Steel Agri-chemicals (formerly Armour Agri -chemicals) was quite coopera
tive and invited us to see their acid-making facilities.
Discussions with engineering and managerial people at the Agri-
chemical headquarters in Atlanta revealed two significant facts: (1) they
are phasing out all Chamber Process plants because it has become cheaper
to buy acid than to make it this way and (2) no one has bothered to study
the Chamber Process closely enough to understand its basic chemistry and
thus determine the fundamental problems that have caused the obsolescence
of the process.
All new acid-making capacity being installed by Agri-chemicals
utilizes the contact process, although the product of the chamber process
is ideal for use in fertilizer manufacture. Contact acid (either purchased
or manufactured by Agri -chemicals) must be diluted back to 60° Be (78%)
before being used. Obviously, the low strength (also 78%) of the chamber
process acid is not the limiting feature.
A tour of the Nashville plant and a detailed discussion of the
process with Mr. Clifford Roquet, the" acid maker" in charge of the
operation of the plant, gave an insight into the problems that have
- 167 -
-------
contributed to the obsolescence of the Chamber Process. The plant
that was visited was in its last days of operation, but we were assured
that operating conditions were typical.
The appearance of the plant was a revelation in itself. Every-
thing was very old and in a poor state of repair. It had been built in
1895 and rebuilt in 1922 after a disastrous fire. There was little
evidence that any design modifications had been made during rebuilding,
and few changes have been made since then. (As a matter of fact, some
of the drawings of Chamber plant equipment in Lunge's book" Sulfuric
Acid and Alkali, " printed in 1891, looked very much like equipment in
this plant. )
Figure 52 shows the layout of the Nashville plant and illustrates
some of the known operating conditions. Because of the almost complete
lack of instrumentation, no other significant data could be obtained con-
cerning plant operation. Process control was crude and limited in scope.
The primary control technique was to observe the color of the exhaust
plume from the second Gay- Lussac tower and vary the S02-air balance
in the combustion chamber accordingly. This ratio could be increased
by putting bricks in the draft inlets in the furnace, thus cutting down the
amount of influent air. The color of the plume varied from reddish
brown (when too much NO entered the towers and was lost to the atmosphere
where it was oxidized to N02)' to white (when too much N02 entered the
tower and was lost as nitric acid mist) .
Despite the crudity of the operation, the plant consistently produced
776 tons per week of 60° Be sulfuric acid for use in the adjacent fertilizer
plant. There was little or no unscheduled down-time, and only two men
per shift were needed to operate the entire plant. In addition, examination
of production economics showed that the cost of the sulfur raw material
contributed up to 70% of the total manufactUring cost of the acid. Why then
are chamber plants obsolete?
The answer to this question involves an examination of the sulfuric
acid industry as a whole and Agri -chemicals' plans for the future. Although
the Nashville plant was operating relatively smoothly, it was obviously due
- 168 -
-------
Pig. 52.
Layout and Operating Conditions of U. S. Steel Agri -Chemical
Chamber Process Sulfuric Acid Plant, Nashville, Tennessee
A. Layout
4
3
2
Chamb~rs
,
5 l~ 6
Gay-Lussac Towers
, Gay- Lussac Towers
B. Operating Conditions
1. Production: 600 tons (100%) per week as 60° Be (77. 7%).
2. H2S04 recycle flow: 3.5 times acid production.
3. NOx recovery: 32 oz (as NaN03) per ft3 of recycle acid
in Gay- Lussac tower
4. N02/S02 mole ratio in first chamber: 0.05
C. Chamber Conditions
Chamber No.
Dimensions
Temperature, 0p
Product Acid
1
2
3
4
5
6
50 x 28 x 20
75 x 28 x 20
110 x 28 x 20
110 x 28 x 20
75 x 28 x 20
50 x 28 x 20
286
272
252
230
212
166
None
None
52° Be
52° Be
52° Be
52° Be
- 169 -
-------
for replacement before too long. Agri -chemicals was faced with the
choice of either building or buying. They found that they could buy
66° B~ acid from a manufacturer who used a smelter off-gas as his
raw material in a contact plant and could deliver it to the Nashville
fertilizer plant (by barge) cheaper than the old chamber plant could
make it. This points up the fact that when the sulfur raw material
is available to the acid manufacturer at no cost, he can produce acid
at a very attractive price compared to acid manufactured by conven-
tional means. This would be true for acid made from power plant
stack gas as well as smelter off-gas.
Obviously, if Agri -chemicals wanted to make the acid and com-
pete with this supplier's price, they would have to change their manu-
facturing approach and find a way to cut costs. One way they have tried
is to make larger quantities at a single location and ship the acid to the
various fertilizer plants. This is where the basic problems with Chamber
Plants become apparent.
There are essentially two reasons why you cannot build large
Chamber Plants and ship the acid to where it is needed: (1) it costs too
much to ship 78% chamber acid, and (2) equipment size makes it
impractical to build a Chamber Plant of greater than 100-ton-per-day
capacity. Both problems were examined in some detail to see if they
were unavoidable for fundamental reasons.
Why can't high strength acid be made in a Chamber Plant? After
much discussion with Agri -Chemical people, we concluded that our original
reason must be correct: the extreme difficulty of recovering oxides of
nitrogen from high acid strength nitrose. Normally, in the Glover tower,
the 60° B~acidfrom the Gay-Lussac tower is diluted with the 52° Be acid
from the chambers and contacted with hot combustion gas ( 1600 OF) from
the sulfur burner which heats the dilute acid and vaporizes the dissolved
N203. Since the solubility of N203 increases with acid concentration,
excessively high temperatures would be required to raise the N203 vapor
pressure to effect nitrogen oxide recovery from nitrose of higher acid
strength. To achieve these elevated temperatUres without creating the
- 1 70 -
-------
requirement for additional heat in the system, the L/G ratio would have
to be reduced. This in tUrn means that the nitrose would have to be more
concentrated in N203 to supply the required nitrogen oxide catalyst which
would therefore require more heat for N 203 removal, etc. It appears
that trial-and-error optimization have resulted in a process that uses an
acid recycle of 3. 5 times the acid production rate and a product concen-
tration of 600 Be.
An understanding of the second problem, excessively large
equipment, requires an examination of what is going on in the lead
chambers. These large structures serve two purposes: (1) to provide
an adequate volume for reoxidation of the NO formed when N02 oxidizes
502' and (2) to provide a large surface area for dissipation of the heat
generated by the 502 oxidation and the absorption of S03 in the water
spray. Re-examination of the plant operating conditions in Fig.46 shows
that the N02/S02 ratio is O. OS, which means that the chamber volume
has to be big enough to allow the NO to be reoxidized 20 times before all
the 502 is oxidized. At first glance, this would seem to be very inefficient
considering the slow rate of NO oxidation, but if the NOz/S02 ratio were
increased substantially, the heat generated in the first chamber would be
too much to be dissipated by this inefficient box-like structure. The
solution to the problem was to keep the N02 dilute and add more boxes.
The rule of thumb used in the industry is 5 ft3 per lb of H 2S0 4 produced
per day. Even the more efficient Mills-Packard towers require about
2.5 ft3 per lb of H2S04 produced per day. After considering these problems,
it is clear that any attempt to build a 500 to 1000 ton per day chamber sulfuric
acid plant in the conventional manner is doomed.
How then does all this affect the prospects for the development of
the Tyco/EPA Catalytic Chamber Process for the removal of S02 from
power plant flue gas? The problem of manufactUring a low strength acid
does not exist, since the catalytic stripper enables us to recover the
oxides of nitrogen from high acid strength nitrose (up to 92%) without
having to worry aboUt vapor pressure limitations. The catalytic process
operates at low temperature and permits N 203 recovery from cool gases,
such as the 300 0 F power plant stack gas.
- 1 71 -
-------
The problem of equipment size is somewhat complex, but does
not depend on the same limitation as the conventional Chamber Plant. In
the Catalytic Chamber Process, the S02 from the flue gas is mixed with
N02 in the ratio of N02/S02 = 2/1, thus eliminating the need for the NO
to be reoxidized in the gas phase. In addition, the heat dissipation
requirement is eliminated because the gases are very dilute (0.3% as
compared to 10 to 12% in chamber plants), and the S03 formed is not
dissolved in water in the reactor. These factors tend to reduce the
required chamber volume problems to the point where they can probably
be eliminated completely. With no excess heat to dissipate, no NO
reoxidation, and no sulfuric acid absorption, it is likely that the last
remaining function of the chambers, S02 oxidation, can be carried out
in the connecting duct work.
Although multiple towers will be required for nitrogen oxide
absorption and recovery, the catalytic nature of the Glover tower
equivalent (stripper), which requires very little stripping gas. to effect
nitrogen oxide recovery, should help reduce the size of the equipment.
All things taken into account, we found no fundamental properties
of Chamber Plants that would create potential problems for the Tyco/EPA
Catalytic Chamber Process for S02/NOx pollution control. In addition,
it seems likely that the new process can be adapted to the more concentrated
gases of a sulfur burner to permit the production of 66° Be' acid and perhaps
high strength oleums.
- 1 72 -
-------
VI. REFERENCES
12.
13.
14.
15.
16.
1.
Stern, A. C. "Air Pollution, "Vols. 1 and 2, Academic Press,
New York (1962).
2.
3.
Anon., Chemical and Engineering News, July 8, 1968.
Bienstock, D., Field, J. H., Katell, S. and Plants, K. 0.,
Journal of Air Pollution Control Association, 15, 459-464 (1965).
4.
Berl, E. and Saenger, H. H., Z. Anorg. Allg. Chern., 202,
113-134, (1931).
5.
Connor, J. M., The Economics of Sulfuric Acid Manufacture,
presented at the Sixty-First Annual Meeting, A. I. Ch. E., Los
Angeles, California, December 1-5, 1968.
Manderson, M. C., Sulfur Outlook into the Early 1970' s, pre-
sented at the Sixty-First Annual Meeting, A. I. Ch. E., Los
Angeles, California, December 1-5, 1968.
6.
7.
Duecker, W. W. and West. J. R., "The Manufacture of Sulfuric
Acid," Reinhold Press, New York (1959).
Berl, E., Transactions of the A.I.Ch.E., 31, 193 (1935).
8.
9.
Bodenstein, M., Z. Electrochem., 24, 183, 381 (1918).
Bodenstein, M., Z. Phys. Chern., 100, 68 (1922).
Burdick, C. L., J. Amer. Chern.., Soc., 44, 244 (1922).
10.
11.
Seidell, A. and Linke, W. F. ,"Solubilities of Inorganic and
Metallic Organic Compounds, "Vol. 2, Am. Chern. Soc., New
York (1965). '
Yushmanov, E. V., Journal of Chemical Ind. (USSR), 17, no. 8,
48 (1940). -
McCabe and Smith, "Unit Operations of Chemical Engineering, "
McGraw-Hill, New York (1956), p. 633. "
Sherwood and Pigford, "Absorption and Extraction, " McGraw- Hill,
New YQrk (1952), p. 245.
Hougen and Watson, "Chemical Process Principles, " part I,
Wiley, New York (1943).
- 1 73 -
-------
17.
18.
Perry, J. H., "Chemical Engineers Handbook, " McGraw-Hill,
New York (1950).
Kusminych, 1. N., Turchan, E. J., and Archipowa, M. S.,
z. Amorg. Allg, Chern., 226, 31~20 (1936).
- 174 -
-------
ACKNOWLEDGMENT
A grateful acknowledgment is hereby made to Mr. E. Saunders
and Mr. E. Rissmann for their efforts on the work described in this
report.
- 175 -
------- |