Tyco Laboratories, Inc.
Bear Hill
Waltham, Massachusetts 02154
DEVELOPMENT OF THE CATALYTIC CHAMBER PROCESS
Final Report
Contract No. CPA 70-59
by
A. Gruber
A. Walitt
23 April 1970 - 22 October 1970
Prepared for
Office of Air Programs Environmental Protection Agency
411 W. Chapel Hill Street
Durham, North Carolina 27701
-------
Tyco Laboratories, Inc.
Bear Hill
Waltham, Massachusetts 02154
DEVELOPMENT OF THE CATALYTIC CHAMBER PROCESS
Final Report
Contract No. CPA 70-59
by
A. Gruber
A. Walitt
23 Apri11970 - 22 October 1970
Prepared for
Office of Air Programs Environmental Protection Agency
411 W. Chapel Hill Street
D.1rham, North Carolina 27701
-------
Section
L
IL
III.
IV.
V.
Table of Contents
SUMMARY. . . . . .
.. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. ..
.. .. .. .. .. ..
.. .. .. .. .. .. .. .. ..
INTRODUCTION. . . . . . . . . .
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. ..
A.
B.
C.
D.
Background. . . . . . . . . . . . . . .
The Tyco Catalytic Chamber Process. . . . .
Previous Work. . . . . . . . . . . . .
.. .. .. .. .. .. ..
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. ..
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. ..
Goals of the Contract. . . . . . . .
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. ..
.. .. .. .. ..
CATALYTIC STRIPPER - CATALYST SELECTION. . . .
.. .. .. .. ..
.. .. .. .. .. ..
A.
B.
C.
Prior Work. . . . .
.. .. .. .. .. .. .. ..
.. .. .. .. .. ..
.. .. .. .. .. ..
.. .. .. .. ..
Candidate Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Qualitative Screening. . . . . . . . . . . . . . . . . . . .
CATALYTIC STRIPPER
ACTIVATED CHARCOAL EVALUATION.
.. .. .. .. .. ..
A.
B.
C.
D.
Batch Tests. . . . .
.. .. .. .. ..
.. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
Continuous Column Experiments (1- In. Column) . . . . . . . . . . . . . .
Continuous Column Experiments (2 - In. Column) . . . . . . . . . .
Summary of Results of Charcoal Evaluation. . . . . . . . . . . . .
HIGH TEMPERATURE SCRUBBER - LABORATORY EVALUATION.
.. .. .. .. .. ..
A.
B.
C.
D.
E.
F.
Background. . . . . . . . . . . . . . . . . . . . . . . . . .
Equipment Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Operating Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analytical Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Experimental Results. . . . . . . . . . . . . . . . . . . . . . . .
Summary of Results of High Temperature Scrubber Evaluation. . . . . .
VAPOR PRESSURE OF NITROGEN OXIDES OVER
HNS05/H2S0 4 SOLUTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . .
A. Literature Search. . . . . . . . . . . . . . . . . . . . . . . . . . .
B. Laboratory Work at Tyco . . . . . . . . . . . . . . . . . . . . . . . . . .
C. Interpretation of Tyco Results. . . . . . . . . . . . . . . . . . . .
iii
Page
1
3
3
3
9
10
11
13
13
14
19
19
21
27
53
55
55
55
55
58
58
64
65
65
69
79
-------
Section
VI.
VII.
VIII.
Table of Contents (continued)
LABORATORY SCALE MINIPLANT - MODIFICATION
AND EXPERIMENTATION. . . . . . . . . . . . . .
. . . . . .
. . . . . . . .
A. Goals of Miniplant Operation. . . . . . . . . . . . . . . . . . . . . . . .
B. Miniplant Modification. . . . . . . . . . . . . . . . . . . . . . . . . . . .
C. Analytical Instrumentation and Sampling Procedures. . . . . . . . . . . .
D. Experimental Results. . . . . . . . . . . . . . . . . . . . . . . . . . . .
ACKNOWLEDGEMENTS
. .' 131
. . . . . .
" . . . . . . .
.. . . . . .
. . . . . .
REFERENCES. . . . .
. . ., 133
APPENDIX I: PROCESS ECONOMICS
. . . . .
. . . . . . . . . .
. . . . . .
APPENDIX II: MINIPLANT DATA SHEET
APPENDIX III: MINIPLANT EXPERIMENTAL PROCEDURES
APPENDIX IV: MINIPLANT DATA SHEET
iv
Page
81
81
82
87
88
-------
Figure No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
List of Illustrations
Catalytic chamber process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
High temperarure scrubber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Catalytic stripper.......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pr-ogram schedule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Apparat'.ls for qualitative catalyst evaluation. . . . . . . . . . . . . . . . . . . . . . . . .
Appararus for continuous, quantitative testing of catalysts. . . . . . . . . . . .
Continuous stripper experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2-inch continuous column for catalyst evaluation. . . . . . . . . . . . . . . . . . . . .
Laboratory apparatus for continuous evaluation of
sttipper catalyst. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . .. . .. .. . .. .. .. .. .. . . . . .. .
Continuous laboratory stripper test: Experiment I-A. . . .. .... .. ... ..
Continuous laboratory stripper test: Experiment 1-8 .. .. . . . .... .. ...
Continuous stripper experiment: Experiment l-C. . . . .. . . .. .. . . .. . . ..
Continuous stripper test: Experiment I-D..........................
Denitration of recycled acid:
Experiment 1- E. , . . . . . . . . . . . . . . . . . . . . .
Denitration of recycled acid:
Experiment 1- F.. . .. .. .. .. .. . . .. .. .. .. . .. . .. .. . .. . . ..
Column optimization; Experiment I-G, air flow
rate variation.. .. . . .. . . .. . . .. . .. .. .. . .. . .. . .. .. .. . . .. .. .. .. .. .. .. . . .. .. . . . . . . . .. .. .. .. .. .. . . ..
Column optimization: Experiment 1- H, air flow
rate variation. .. . .. .. . . . . . .. .. .. .. .. .. .. .. .. .. .. .. . .. .. .. .. .. .. .. .. . .. .. . .. . .. .. .. .. .. .. .. .. . .. .. .. .. .. ..
Column optimization: Experiment 1- I, air flow
rate variation.. . .. .. . .. . .. . . .. .. .. . .. .. .. . .. . .. .. .. .. .. .. .. .. .. .. . . . . .. . . .. .. . .. .. . . .. .. .. .. .. .. . ..
Co-current flow: Experiment I-J. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Countercurrent column operation: Experiment 1- K . . . . . . . . . . . . . . . . . .
Denitration using reagent grade acid in nitrose:
Experiment l-L . .. . .. .. . . . .. . . .. .. .. . .. .. .. .. .. .. . .. . . .. .. .. .. .. . . . .. .. .. .. . .. .. . . .. .. . . . . . .
Stripper blank - saddle packing: Experiment 2-A.. . .. . . . . .. .. . .. .. . .
Laboratory scale high temperature absorption column. . . . . . . . . . . . . . . .
v
Page No.
4
6
8
11
15
22
28
29
31
33
34
36
37
39
40
42
43
44
45
47
49
53
56
-------
List of Illustrations (Cont)
Figure No.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
Laboratory apparatus for continuous evaluation of high
temperature scrubber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nitrogen oxide partial pressure, nitrosylsulfuric
acid concentration dependence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nitrogen oxide partial pressure, temperature dependence. . . . . . . . . . .
Nitrogen oxide partial pressure, sulfuric acid concentration
dependence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
N203 vapor pressure analysis apparatus. ... . ... . . . . . . . . . . . . . . ., .. .
Total vapor pressure nitrose solutions containing
O.1M HNS05 . . . . . . . ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nitrogen oxide partial pressures over nitrose solutions
containing O.lM :HNS05 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nitrogen oxide partial pressure over nitrose solutions;
concentration dependence - experimental values. . . . . . . . . . . . . . . . . . . .
Miniplant flowsheet .,...........................................
Miniplant showing 502 reactor and NO oxidizer. . . . . . . . . . . . . . . . . . . . .
Miniplant showing stripper and scrubber columns. ... . .. . . . . ... . .. . .
Miniplant analytical instrumentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gas sampling and analysis system.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Miniplant scrubber evaluation wet feed: NOx only. . . . . . . . . . . . . . . . . .
McCabe-Thiele analysis of high temperature scrubber... . ...... .....
Scrubbing efficiency of packed column; experimental and
extrapolated values used to determine HTU .. . . . . . . . . . . . . . . . . . . . . . .
Experimental operating lines - scrubber design. . . . . . . . . . . . . . . . . . . .
Scrubber / stripper operating in miniplant -
Data from Table XVII. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Scrubber / stripper operation in miniplant -
data from Table XVIII. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Scrubber / stripper operation in miniplant -
data from Table XIX ...........................................
Scrubber / stripper operation in miniplant -
data from Table XXIII. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Modified catalytic stripper
recycle concept. . . . . . . . . . . . . . . . . . . . . . .
vi
Page No.
57
67
68
70
71
75
76
77
83
84
86
89
90
102
108
113
116
119
120
121
122
128
-------
Table
I
IL
III
IV
V
VI
VII
VIII
IX
x
Xl
XII
XIII
XIV
XV
XVI
XVII
XVIII
XIX
XX
XXI
XXII
XXIII
XXIV
List of Tables
Results of Qualitative Screening Tests. . . . . . . . . . . . . .
" " " " " " "
Batch Experiments - Characterization of Charcoal Catalyst.
" " " " " " " " "
Continuous Column Tests - 1- In. Column. . . . . .
" " " " "
" " " " "
Batch Tests With Activated Charcoal. .
" " " " "
" " " " " " " " " "
" " " " "
Comparative Analyses of Acid Samples
" " " " " "
" " " " " " " " "
" " " " "
High Temperature Scrubber Experimental Results 2500 F Steady
State Opera tion 2-in. x 40-in. Column Packed with 1/4-in.
Procelain Intalox Saddles. . . . . . . . . . . . . . . . . . . . .
" " " " " " "
High Temperature Scrubber Experimental Results 2500 F Steady
State Operation 2 -in. x 40-in. Column Packed with 1/4-in.
Porcelain Intalox Saddles. . . . . . . . . . . . . . . . . . . . .
" " " " " " "
Operation of Scrubber Using Pure NO as Feed Steady
State Operation. . . . . . . . . . . . . . . . . . . . . . . .
" " " " "
Tyco Sample Compositions. . . . . . . . . . . .
" " " " " " "
" " " " " " " "
Tyco Vapor Pressure Data for 80.08 wt % H2S04 . . . . . . .
Tyco Vapor Pressure Data for 84.04 wt % H2S04' . . . . . . . . . . . . . .
Tyco Vapor Pressure Data for 90.16 wt % H2S04' . . . . . . . . . .
Tyco Vapor Pressure Data for 95.0 wt % H2S04' . . . . . . . . . . . . . . .
Miniplant Experimentation - Reactor Evaluation.
" " " " " "
" " " " "
Miniplant Reactor Experiments. . .
" " " " " " " " "
" " " " "
Miniplant Evaluation of Scrubber. . . . .
" " " " " " "
" " " " " " " "
Miniplant Scrubber Evaluation. . . . .
" " " " " " " "
" " " " " " " "
Miniplant Scrubber Evaluation. . . . .
" " " " " "
" " " " "
" " " " "
Miniplant Scrubber Evaluation. .
" " " " "
" " " " " " " " "
Residence Volumes and Times in Miniplant Reactor and
Analytical System. . . . . . . . . . . . . . . . . . . . . . . .
. . .. 104
Miniplant Reactor Experiments.
" " " " " " " " "
" " " " " " " " " " "
Equilibrium Data. . . . . . . .
. . . . .. 109
" " " " " " "
" " " " "
" " " " " "
Miniplant Scrubber Evaluation. . .
" " " " "
" " " " "
" " " " , " "
Extrapolated Scrubbing Efficiencies Based on Experimental
Column Tests. . . . . . . . . . . . . . . . . . . . . . . . .
" " " " " " " " "
vii
Page
17
20
24
26
51
59
60
63
72
73
73
74
74
94
96
97
99
100
. 101
106
. 111
112
-------
Tables
xxv
XXVI
XXVII
XXVIII
XXIX
XXX
XXXI
XXXII
List of Tables (continued)
Calculation of Mass Transfer Coefficients (K a) and
Height of Transfer Unit (HTU) . , . , , , . ,g, . , .
. ,. ,. . ,. ,. ,. ,. . ,. ,. ,.
Design and Conditions for Operation of LaboratOry and
Miniplant Catalytic Stripper. . . , . . . . . , . . , . .
. . ,. . ,. ,.
Miniplant Stripper Verification Experiment No.1,
,. ,. ,. ,. ,. ,. . ,. ,.
Miniplant Stripper Verification Experiment No.2. .
. ,. ,. ,. ,. ,. ,. ,. ,. ,. ,. .
Estimated Major Equipment Cost.
. ,. . ,. ,. ,.
. . ,. ,. ,. ,. ,. ,. ,.
,. ,. ,. ,. ,. ,. .
Estimated Capital Cost Summary.
,. ,. ,. ,. ,. . .
. ,. ,. ,. . ,. ,. ,. . ,. ,.
Annual Operating Costs. . ,
,. ,. ,. ,. ,. ,. ,. ,. ,. ,. ,. ,. ,. ,.
,. ,. ,. ,. ,. ,. ,.
Miniplant Startup Procedure. . . . . . . , . . . . . . . . . . .
,. ,. ,. ,. ,. ,. ,.
XXXIII-XXXVI Miniplant Data Sheets. . .
,. ,. ,. ,. ,. ,. ,. ,.
,. ,. ,. ,. ,.
,. ,. ,. ,. ,. ,. ,. ,. . ,.
viii
Page
117
123
126
127
1-1
1-2
1-3
III-I
IV-I-IV-4
-------
SUMMARY
The objective of this contract was to: (1) confirm the technical feasibility of the Cata-
lytic Chamber Process, and (2) obtain preliminary engineering design data necessary for
process scale-up. The contract work consisted of a laboratOry-scale investigation of the
separate process stages and of an integrated 10 SCFM unit called the "miniplant."
The Catalytic Chamber Process, which is based on the Lead Chamber Process for
sulfur acid manufacture, accomplishes the removal of sulfur dioxide by oxidation with nitrogen
dioxide followed by adsorption of the resulting sulfuric acid and the oxides of nitrogen in a
recycle stream of sulfuric acid. The nitrogen oxides are recovered by catalytic stripping of the
sulfuric acid; the net nitrogen dioxide recovered by the process is converted to nitric acid; and
the bulk of the nitrogen dioxide is recycled for further sulfur dioxide oxidation. Three major
modifications of the original Lead Chamber Process were investigated during this contract:
(1) the sulfur dioxide oxidation and the nitric oxide oxidation are performed in separate stages,
thus avoiding the need for large reactor volumes to complete the very slow NO oxidation; (2) a
high temperature scrubber is used which permits the absorption of the sulfuric acid and oxides
of nitrogen but not water, which would dilute the acid; and (3) a catalytic stripper is used to
simultaneously remove the oxides of nitrogen from the scrubbing acid and oxidize them to
nitrogen dioxide.
The following results were obtained:
1. The S02 reactor achieved 84-94% conversion of S02 to S03 at a residence
time of 20-22 see, and inlet S02 concentration of 3200 -3950 ppm, 290 -324 0 F, and atmospheric
pressure.
2. The high temperature scrubber achieved 65% and 45% removal of the NO in
x
the inlet gas using 8.5 ft and 5 ft of 3/8-in. Intalox saddles, respectively. * The column diameter
was 4 in. and the inlet NO concentration was "'7000 ppm.
x
*These data were extrapolated to show that 97% of the NOx could be removed in 30 ft
of packed bed.
1
-------
3. The catalytic stripper achieved 95% denitration of the inlet acid. The
stripper exit gas contained 0.39% NO and 0.75% N02' The inlet acid concentration was
0.38% wt nitrosylsulfuric acid in 80% sulfuric acid and the stripping temperatUre was 280-
310°F.
4, New NO vapor pressure data were obtained for nitrose solutions containing
x
up (095% H2S04 and at a temperature of 350°F. These new data agree well with the earlier
data of Berl and Saenger. 5
The results of the contract work indicate that the following additional work should be
performed:
a. A parametric study of process variables using the 10 SCFM miniplant,
b. Investigation and demonstration of a process control system,
c. Preliminary design of a 2000 SCFM pilot plant.
2
-------
L INTRODUCTION
A. Background
The Catalytic Chamber Process is aimed at solving one of our most pressing air pollu-
tion problems: the emission of enormous quantities of sulfur dioxide and nitrogen oxides into
the atmosphere from the stacks of power plants burning fossil fuels. This problem has become
so severe that at the present time the amount of sulfur dioxide emitted into the atmosphere in
the United States is equivalent to more sulfuric acid than is produced for industrial purposes. *
In addition to the serious sulfur dioxide problem, there has been a growing awareness of the
detrimental effect of nitrogen oxides on man and his surroundings. It has also been established
that N02 is a key participant in the photochemical oxidation of unburned hydrocarbons from
automotive exhaust to form smog.
The Catalytic Chamber Process features the oxidation of S02 to S03 through the use of
nitrogen dioxide as a homogenous, gas-phase, reactive catalyst in a manner similar to the old
Lead Chamber Process. The initial work performed under contract to NAPCA was undertaken
to determine the ability of N02 to oxidize the dilute sulfur dioxide present in flue gas. This
work has led to the present Catalytic Chamber Process concept.
B. The Tyco Catalytic Chamber Process
The most recent conceptualization of the Tyco process for the removal of S02 and NOx
from the power plant stack gas has been named the Catalytic Chamber Process because of the
novel manner in which the oxides of nitrogen are recovered for reuse as the S02 oxidant. In
this process, the oxides of nitrogen are removed from the scrubbing solution by contacting
the nitrose with air in a fixed bed of activated charcoal. The charcoal catalytically oxidizes
the NO values in the nitrose to N02 which, being only slightly soluble in sulfuric acid, is
spontaneously evolved from solution into the gas phase. The Catalytic Chamber Process is
composed of three major process units: (1) the S02 reactor, (2) the high temperature
scrubber, and (3) the catalytic stripper (see Fig. 1). The raw flue gas is cleaned of most of
*In 1966, for example, U.S. sulfuric acid production was about 22 million tons as compared
to 44 million tons of equivalent acid from all stationary sources and about 20 million tons
from power plants alone.!' 2
3
-------
Gas to stack
2500 F
7% H20
150 ppm NO
x
~
80% H2S04
@)
HIGH
TEMPERA1URE
SCR UBBER
REACTOR
G)
Gas Gas
o NO N203
N02 H2S04
S02 @ Air N02
'2' ~03ABSORBERtW
~ ~ ~H20
Air 9
NO 'e)
CD
80% H2S04
HNS05
@
(f)
Product
HN03
Air
@
CATALYTIC
STR IPPER
Flue gas
3000 F
G)
Product
H2S04
G)
Fig. 1. Catalytic chamber process
4
-------
the fly ash and is contacted with a stream of recycle N02' The N02 oxidizes the S02 to S03'
The S03 and H20 in the flue gas react to form H2S04(g). The reacted gases are passed through
a sulfuric acid scrubber to remove the newly-formed acid and the oxides of nitrogen. The clean,
wet stack gas is vented to the atmosphere at about 2500 F and is easily dissipated. The nitrogen
oxide-bearing sulfuric acid (nitrose) from the scrubber is passed countercurrent to air across
activated charcoal catalyst. The nitrose is catalytically oxidized to N02 with air, and the N02
exits the stripper with the air stream. Part of the N02 rich gas stream from the stripper,
equivalent to the net NO recovered by the process, is used to produce nitric acid in a conven-
x
tional nitric acid plant. The bulk of the N02 rich gas stream is recycled to the S02 reactor for
further S02 oxidation. Sulfuric acid, equivalent to the S02 recovered by the process, is
removed as product from the denitrated acid leaving the stripper.
1. Reactor
The reactor accomplishes the oxidation of S02 to S03 and the hydrolysis of
S03 to H2S04:
S02 + N02 -- S03 + NO
503 + H20(g) -- H2S04(g)
Important features of this stage include: (1) maintaining a sufficiently high
temperature to avoid condensation of the acid mist in the reactor, and (2) providing twice
the stoichiometric amount of N02 so that the full reacted gas will contain an equimolar amount
of NO and N02' thus enabling recovery of the oxides of nitrogen. The reactor residence time
is long enough to permit high yield oxidation of the S02' but not long enough to permit appreciable
oxidation of the nitric oxide. The NO is oxidized in the catalytic stripper (see below), and the
N02 is recycled to the reactor to oxidize more 502'
2. Scrubber
The high temperature scrubber removes the sulfuric acid mist and the oxides of
nitrogen present in the reactor exit gas (see Fig. 2). The oxides of nitrogen must be present
in equimolar quantities. NO or N02 alone are insoluble in sulfuric acid; but together, as
N203' they react with sulfuric acid to form nitrosylsulfuric acid, which is soluble in sulfuric
acid:
N02 + NO + 2H2S04 -- 2HNS05 + H20
The scrubber operating temperature is that temperature for which the H20 vapor
pressure of the incoming acid is equal to the partial pressure of water in the incoming gas.
5
-------
St k
2
7
1
N
ac gas 0.0
50°F
% H20
50 ppm
2°3
e Gas ,~
H2S04 80%
250 ° F
01M HNS05
Flu
350 ° F
7% H20
8730 ppm N203
H2S0480%
260 ° F
O.IIM HNS05
Fig. 2. High temperature scrubber
6
-------
Thus, no water is absorbed in the column and the acid leaves the scrubber at the same concen-
tration as it entered. For 80% acid and a gas stream containing 7.3% water, this temperature
is 2500 F. It should be noted that the Lead Chamber Process uses cold sulfuric acid for scrubbing.
Since the solubility of nitrosylsulfuric acid increases with increasing H2S04 concentra-
tion, it may be desirable to operate the scrubber at acid strengths above 80% H2S04' the
concentration utilized by the Lead Chamber Process. For the Catalytic Chamber Process, a
limitation on the acid concentration is imposed by the sulfuric acid content of the exit gas. As
the acid concentration is increased, the operating temperature of the scrubber must be increased
in order to maintain the water vapor pressure over the entering acid equal to the partial pressure
of water in the entering gas. As the scrubber temperature is increased, the sulfuric acid
content of the exit gas increases. An acid concentration of about 92%, for example, requires a
4000 F scrubber temperature and results in about 2 mm H2S04 in the exit gas. The limitation of
the standard Lead Chamber Process in the increasing degree of difficulty of denitrifying higher
strength sulfuric acid is overcome in the Catalytic Chamber Process through the use of the
catalytic stripper, as discussed in detail below.
3. Catalytic stripper
In the standard chamber plant, the nitrogen oxides are sufficiently concentrated
so that reoxidation can occur simultaneously with sulfur dioxide oxidation even though the
reaction rate is quite slow in comparison at the operating temperature used. Nitrogen oxide
recovery is effected by the dilution and heating of the nitrose, which causes an increase of
the vapor pressure of the oxides that are then swept out of the acid by the high temperature
gases from the sulfur burner.
Both heating and diluting are impractical when treating 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 20% of the tOtal heat generated by the power plant
would be necessary.
The breakthrough achieved during this research which avoids this heat penalty was the
discovery that, by contacting the nitrose with a catalyst such as acitivated charcoal in the
presence of oxygen, the oxides of nitrogen are oxidized to nitrogen dioxide in the liquid phase
and are recovered quantitatively in a relatively small volume of carrier gas. Fig. 3 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 converted to the relatively insolu-
ble 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
negligibly exothermic.
7
-------
Air
N02
Air
Fig. 3. Catalytic stripper
8
H2S04 80%
0.11M HNSOS
Charcoal
packing
H2S04 80%
Oo001M HNSOS
-------
4. Process economics
Investmentt and operating costs have been estimated by Tyco Laboratories, Inc.,
and by the Tennessee Valley Authority, as shown below:
Tyco
Size 800 M W
Flue Gas Flowrate (MM SCFM) 1.44
NO concentration, ppm 600
x
S02 concentration, ppm 3000
N02 concentration 30
S in coal % wt 3.5
Investment Cost $MM 11.5
Annual Operating Cost 3.7
Refer to Appendix I for a breakdown of the above costs.
TVA
500 MW
1.0
400, 1200
3000
20, 60
3.5
10.7
2.4
The net operating costs are highly dependent on the market price of the product acids.
If the nitric acid and sulfuric acid prices are $40/ton and $9.50/ton, respectively, the net
operating costs are zero. * If the prices are higher, the Catalytic Chamber Process will pro-
duce a return on the investment.
C. Previous Work
The work performed under a prior contract (PH 86-68-75) had several intermediate
goals which are worth discussing before going into detail on the current work. This background
will clarify the orientation and goals of the current contract.
1. Initial evaluation of the Lead Chamber Process
The earliest work at Tyco concerning this technology was the evaluation of the
Lead Chamber Process for direct application as a flue gas cleanup technique. These experi-
ments involved kinetic studies of the S02 oxidation and NO reoxidation reactions, as well as
preliminary analysis of the NO scrubbing process.
x
It was concluded from this work that the Lead Chamber Process could not be used for
flue gas cleanup without modification. The oxidation of S02 by N02 could be accomplished
under the dilute conditions present in the flue gas but the simultaneous NO oxidation was im-
possible. In addition, scrubbing with cold sulfuric acid removed all the water from the flue
gas and diluted the acid. Heat was needed for the recovery of the oxides of nitrogen from the
scrubbing acid and the reconcentration of the acid - almost 40% more heat than was provided
by the flue gas. Subsequent work concentrated on finding ways of avoiding the disadvantages of
the basic Lead Chamber Process.
*Prices are based on 100% acid strengths.
9
-------
2. Modification of the Lead Chamber Process
Three major modifications of the original process were conceived: (1) NO
reoxidation separate from the S02 oxidation, (2) high temperatUre NOx scrubbing to avoid
dilution of the recycle acid, and (3) catalytic oxidation of the oxides of nitrogen in the liquid
phase to enable N02 recovery and recycle without additional heat (see Part B of this section).
Each of the three process stages (reactor, scrubber, and stripper) was evaluated by labora-
tory-scale experimentation. * Then, a 10 SCFM laboratory-scale pilot plant (miniplant) was
designed, built, and operated in order to evaluate the new process stages on an integrated
basis.
D. Goals of the Contract
The goals of Contract CPA 70-59 were: (1) to continue the individual bench-scale
evaluation of the reactor, the scrubber, and the stripper, and (2) to operate the miniplant.
More specifically, the work was to include the following tasks:
1. Selection of a catalyst to be used in the catalytic stripper,
2. Operation of a laboratory-scale catalytic stripper to determine operating
characteris tics,
3. Operation of a laboratory-scale, high-temperature scrubber to verify the
high-temperature removal of oxides of nitrogen from the reacted flue gas
without the dilution of the acid,
4. Determination of N203 vapor pressure data for nitrose solutions at higher
concentrations than were currently available in the literature,
5. Modification of the miniplant according to the results of 1-4 above,
6. Operation of the miniplant on a continuous basis to obtain process data.
A program schedule of the time spend on each of the six tasks is shown in Fig. 4.
* .
Refer to the fmal report of Contract No. PH 86-68-75 for more details.
10
-------
Months
May June July Aug Sep Ocr
1. Catalyst Screening
2. Catalyst Optimization
Continuous Studies
3. High Temperature I
Scrubbing Test
(N203 Absorption
4. Vapor Pressure Studies
(N203 Above HNS05
Solutions)
5. Miniplant Modification I
6. Miniplant Demonstration
and Data Accumulation
Fig. 4. Program schedule
11
-------
II. CATALYTIC STRIPPER - CATALYST SELECTION
A. Prior Work
It was demonstrated that nitrosylsulfuric acid in sulfuric acid (nitrose) could be deni-
trated quantitatively by contacting the nitrose with air in the presence of activated charcoal
catalyst. The proposed mechanism for the denitration/oxidation reaction is:
4 HNS05 + 2 H2 ° + 02
activated charcoal
catalyst > 4 N02 + 4 H2S04
(1)
The NOx in the effluent gas from the stripper was almost completely in the form of N02;
the difficult gas phase NO oxidation was thus avoided.
Activated charcoal was an effective catalyst for the above reaction but appeared to
have stability problems. There was evidence that the N02 had reacted with the carbon to form
small quantities of carbon dioxide and nitrous oxide (N20). In addition, the carbon bed even-
tually became plugged with very fine pieces of carbon, which were assumed to result from
the physical breakdown of the activated carbon particles. Thus, it was decided to look at
other potential catalysts during Contract CPA 70-59 to see if there was a material better
suited to the task.
B. Candidate Catalysts
A screening program was set up to determine if materials other than activated char-
coal could be used to catalyze Reaction (1). The screening program consisted of two phases:
(1) a literature study in which candidate materials were chosen on the basis of physical pro-
perties, and (2) an experimental study of the ability of the candidate materials to catalyze the
oxidation of nitrosy lsulfuric acid.
The candidate catalysts had to be insoluble in hot (250 to 4000 F) sulfuric acid solutions
(75 to 90%) containing about 1% nitrosylsulfuric acid and resistant to oxidation at temperatures
between 70 and 4000 F. Thus, the oxides of nickel, cobalt, and iron and many of the pure metal
catalysts were eliminated from consideration. Refractory metals and their carbides, nitri
-------
1. Carbides
The carbides were considered prime candidates because of their general stability
and the similarity between the carbide and activated carbon structures. Boron carbide, silicon
carbide, tungsten carbide, nickel carbide, titanium carbide, and molybdenum carbide were
considered. Boron carbide was considered the best candidate because it is commercially
available and is relatively inexpensive, even though the surface area is approximately 1 m2/g
as compared to 500-1000 m2/g for the activated carbon. Molybdenum carbide was considered
the worst candidate because molybdenum compounds are apparently somewhat soluble in sul-
furic acid.
2. Silicides, nitrides, and selenides
This candidate list included the silicides and nitrides of boron, silicon, and
tungsten. The selenides were eliminated because of their toxic nature.
3. Tungsten bronzes
These materials were chosen because of their acid stability. They have been
placed low on the priority list because they are somewhat hard to obtain (they are not widely
commercially available).
4. Other materials
Consideration was given to other materials, including pure metals such as
tungsten, titanium, platinum, rhenium, and tantalum. They were not considered prime can-
didates because most of them are extremely expensive and are known to be easily poisoned.
There was also the possibility of using some of the refractory compounds with a trace of
metal (platinum, rhenium, etc.) as a promoter. Such mixed systems are of considerable
interest but will receive secondary consideration because of the greater difficulty in obtaining
them.
C. Qualitative Screening
After a list of candidate materials had been determined, several were tested quali-
tatively by a simple experimental procedure designed to evaluate their gross ability to cata-
lyze the liquid-phase oxidation of nitrosylsulfuric acid at high temperatures. The apparatus
used (see Fig. 5) consisted of a large test tube fitted with a fritted glass sparging tube and a
gas outlet which led to an infrared analytic cell. A quantity of catalyst material was placed in
the tube and 25 ml of nitrosylsulfuric acid (0.7%), in 80% sulfuric acid at 2100 F, was added to
the catalyst. The air bleed was inserted, and the test t'.lbe was placed in an oil bath, also at
2100 F. The air was passed through the system for 10 to 15 min, after which the acid was
cooled and analyzed for nitrosylsulfuric acid content using potassium permanganate to deter-
mine oxidizable nitrogen oxide species. Details of this analytical procedure can be found in
14
-------
Ga s to
infrared cell
..
Air
,{
Fritted
sparger
Oil bath,
2100 F
25 ml
H2S04
(HNS05)
Catalyst
Fig. 5. Apparatus for qualitative catalyst evaluation
15
-------
Appendix II. An infrared spectrum of the effluent gas was taken during the run. The spectrum
and acid content were compared with control runs on crushed saddles and activated carbon.
The results of a series of these tests are shown in Table 1.
As was expected, the saddles had no catalytic effect on the acid and no significant
denitration was observed. The activated carbon reduced the HNS05 level of the acid, but not
as much as was expected. A sample of the charcoal was first treated with the HNS05 solution,
washed in water, dried at 2500 F for 2 hr, and then run in the standard test. The acid was
reduced to a concentration of only 0.075% HNS05' which was more in keeping with previous
experience. Apparently, the original charcoal was wet and little air was adsorbed on the
surface, and the air sparger was not efficient enough to allow good contact in the three-phase
system: air, catalyst, and acid. When the charcoal was washed and dried, more air was
adsorbed on the charcoal surface, thus providing oxygen for the denitration reaction.
It was interesting to note that, when the charcoal was used directly as-received, the
resulting acid contained many fines (as had been observed in the earlier contract work). How-
ever, when the charcoal was washed in acid, then washed in water and dried, subsequent acid
treatment showed no evidence of the fine charcoal particles. Obviously, the initial contact with
the acid causes some reaction, breaking off the fine particles; but additional treatment with
acid does not affect the charcoal and it will be physically stable. This is very significant in
light of previous observations that the fines tend to plug the charcoal columns after initial
treatment with acid.
Materials such as silicon carbide, tUngsten silicide, titanium carbide, and platinum
black showed no evidence of catalytic activity and can probably be eliminated from further
consideration. Others such as nickel, nickel boride, nickel carbide, nickel with platinum,
tUngsten boride, and titanium boride can be eliminated, since they showed a tendency to react
with acid or dissolve in it.
There remains a group of materials which shows some activity, although none appear
to be as active as the activated carbon. Boron carbide, silicon nitride, boron nitride, tUngsten
carbide, and vanadium carbide are in this latter category. Boron carbide is of doubtful use-
fulness, since it may likely contain a relatively large amount of free carbon, and it is this
carbon that might well be causing the denitration oxidation reaction. This is reinforced by the
observation that the purer carbide obtained from Cerac Company had less tendency to catalyze
the reaction (see tests 4 and 4).
Although there was no material that was as good as activated charcoal, this did not rule
out replacement of the carbon with one of the other materials such as tungsten carbide or sili-
con nitride. The next step in ascertaining the utility of these potential replacements was to
test them in a more quantitative fashion in a continuous column containing material balance
data. A column was arranged so that acid and gas could be fed countercurrently at known rates
and the effluent materials collected and analyzed.
16
-------
Table 1.
Results of Qualitative Screening Tests (Temperantre:
200 to 210oP; Original Acid Concentration: 0.720% HNS05)
Catalyst
Time of
Test, min
Final Acid
Concentration,
% HNS05
Other
Interactions
With Acid*
1. Crushed saddles
5. Boron carbide~
(Cerac)
6. Nickel
7. Nickel boride
8. Nickel carbide
9. Nickel with platinum
10. Silicon nitride
10
20
15
10
20
15
15
lOt
15
10
15
15t
10
10~
0.718
0.720
0.714
0.450
0.401
0.412
0.485
0.076
0.704
0.688
0.550
0.413
0.569
0.637
None
2. Activated carbon
(Witco grade 256)
Evidence of N20
3. Silicon carbide
(Carborundum no. 6-RA)
4. Boron carbide
(Norton)
None
Some N20
Some N20
12. Boron nitride
10
10
10
10
15
15
10
15
10
15
10
15
15
10
10
10
15
15
10
10
0.681
0.427
0.463
0.686
0.590
0.497
0.586
0.428
0.354
0.566
0.653
0.669
0.603
0.576
0.505
0.542
0.670
0.594
0.292
0.647
Some reaction
Some reaction
Reaction
Slight reaction
None
11. Tungsten boride
Reaction
None
13. Tungsten silicide
14. Tungsten carbide
None
None
15. Titanium carbide
16. Vanadium carbide
17. Titanium boride
18. Platinum black
None
None
Some reaction
None
*This refers to either solubility of the catalyst in the acid or a tendency for the system
to oxidize the catalyst.
tThis carbon sample was prewashed in HNS05 solution, washed in H20, and dried.
:j;Additional B4C was added after 5 min and after 10 min
~This B4C was of higher purity than the Norton material.
17
-------
There would be little difficulty in column testing the granular charcoal or boron
carbide, but the other promising materials mentioned were far more difficult to test in the
continuous apparatus since they were available only in extremely fine powders (325-mesh or
finer). In order to test them in a column, it would have been necessary to find a way to make
porous pellets from the powders. A company that specialized in powder metallurgy techniques
was located (Advanced Metallurgy, Inc., of McKeesport, Pennsylvania), and they felt that they
could make macroporous pellets from the fine powders; but their price was quite high and it
was decided to more fully evaluate the charcoal before pellets were made.
As later results will show, the carbon performance was good enough that it was not
considered necessary to go back and more fully evaluate the other potential catalysts. There
are many advantages for using activated carbon, including availability and price, and it was
felt that carbon should be used if possible. The work on the other catalyst materials was dis-
continued at this stage with the conclusion that there are other materials that can be used in the
catalytic stripper and if there are any major problems in the future we could return to this
\IDrk.
18
-------
IlL CATALYTIC STRIPPER - ACTIVATED CHARCOAL EVALUATION
A. Batch Tests
1. Prewashing with HNS05
The tests in Table I show that the activated charcoal provided about 40% denitra-
tion of a 0.72% HNS05 solution if the carbon was used in the as-received form. If the carbon
was first washed in HNS05 solution and then washed in water and dried, about 90% of the
HNS05 in the solution was removed.
To further evaluate this effect as well as the long term stability and reusability of the
carbon, an experiment was set up where the carbon was tested for catalytic activity on a batch
basis, then washed, dried, and retested. The carbon used was Witco type 256, 4 x 10 mesh.
After each test, the carbon sample was weighed and screened to test for particle breakdown.
The batch test procedure was the one described in Section II of this report. As shown in Part A
of Table II, the results of these repeated tests indicate that the carbon activity is maintained
over a period of six repetitions of the test with no physical or catalytic activity degradation.
It was interesting to note that, although the activated carbon lost a small amount of
material in the form of fines the first time it was contacted with nitrosylsulfuric acid, subse-
quent denitration runs and washings did not produce any further fines. The carbon appears to
have stabilized after this first washing.
2. Teflon impregnation of the charcoal
One idea concerning the maintenance of the physical stability of the carbon was
to impregnate the carbon particles with TFE Teflon. The charcoal particles were allowed to
absorb the Teflon from a liquid dispersion of the polymer after which the excess liquid was
drained off and the charcoal particles sintered at 280°C. It was felt that not only would the
Teflon improve the physical stability of the charcoal, but it would improve the diffusion of the
oxygen into the pores of the carbon by making the surfaces more hydrophobic.
A sample of carbon was treated in this fashion and tested in the same way as above
(water washed and dried between tests). The analysis of the final acid concentration shows
that the carbon had lower activity after the Teflon treatment, and after several tests the car-
bon appeared to maintain a constant level of activity although at a considerably lower level than
without the Teflon impregnation (see Part. B of Table II). Since the untreated carbon appeared
19
-------
Table IL Batch Experiments - Characterization
of Charcoal Catalyst
Final Acid
Test Concentration *
A. Witco carbon type 256 1 .104
4 x 10 mesh 2 .054
1250 m2j g (total surface area) 3
carbon tested with standard 4
test procedure then washed to 5 .061
effluent pH ~ 4.0; dried 2 hr 6 .064
at 120 DC (tests repeated six
times)
B. Same carbon as A with Teflon 1 .375
sintered into structure; test 2 .542
procedure the same as A 3 .541
4 .550
5 .568
C. Witco carbon type 517 1 .366
20 x 40 mesh 2 .055
1050 m2j g (total surface area); 3 .145
standard test - two reruns
with no wash or dry treatment
between runs
D. Nuchar carbon type C-190-N, 1 .176
very fine « 325 mesh)
E. Same carbon as C; .247
standard test using N2 instead 1
of air
* Inlet acid for all experiments had a concentration of 0.670% HNS05 by weight.
20
-------
to be quite stable physically, there seems to be no advantage at this time to continuing this
line of experimentation.
3. Evaluation of smaller particle carbon
There was some thought that there was not really enough contact time to allow
the oxygen and the acid to diffuse into the fine pore structure of the carbon particles and react
there, and that most of the reaction was occurring on the outer surface of the carbon particles.
If that were true, an equal volume of a small sized carbon would have a higher outer surface
area and the reaction should be accelerated. This concept was evaluated by testing Witco type
517 carbon which had a particle size of 20 X 40 mesh. A sample of this carbon was tested
three times without washing between tests.
The results are shown in Part C of Table II with the first test on as -received carbon
exhibiting lower activity than the larger size particles (which had a higher total surface area,
although a lower outer surface area). Additional tests show improved activity although the
third test seems to be going to lower activity. The increase in activity after an initial test
has been demonstrated several times for different carbons and suggested that the as-received
carbon does not have maximum activity. A washing with acid is necessary to remove adsorbed
material from the carbon surface and free it for catalytic activity.
Another test was run with an even finer carbon: Nuchar type C-190-N with a particle
size of less than 325 mesh. As shown in Part 0 of Table II, the results are good but not sig-
nificantly better than any other carbon.
4. Using nitrogen as the treatment gas
A sample of the Witco 517 carbon was run in the standard batch test and the final
acid concentration was considerably lower than the original acid (see Part E of Table II). This
result suggests that either the acid is reacting with oxygen adsorbed on the carbon, or the acid
is only being thermally stripped. The test results using saddles would tend to deny this latter
possibility since considerable stripping did occur. The question then arises as to whether or not
the original adsorbed oxygen that is used is ever replaced during the reaction. The test results
in Part A of Table II are not completely applicable here as the acid was washed out of the carbon
and the particles dried, thus allowing the readsorption of more oxygen. In order to evaluate
this concept, the continuous column apparatus was used to get long term activity data on a single
bed of carbon.
B. Continuous Column Experiments (1- In. Column)
1. Apparatus and procedure
To obtain long term stability and activity characteristics of theWitco type 256
activated carbon, a continuous column apparatus was constructed using Corning 1-in. conical
pipe. As shown in Fig. 6, the apparatus contained a removable section so that the catalyst
could be examined or replaced without dismantling the entire apparatus.
The acid was heated in a container above the column and then dripped into the column;
the flow being controlled by a stopcock. (This control was difficult to achieve and the flow
21
-------
Valve
-
Inlet acid
heater
Gas outlet to
infrared cell
Acid inlet from
acid reservoir
(gravity feed)
Thermometer
Outlet gas
mist trap
Thermometer
Removable packed
catalyst sectioQ
(heated)
Flowmeter
Acid collection
reservoir
t Air inlet
t Acid outlet
Fig. 6. Apparatus for continuous. quantitative testing of catalysts
22
-------
rates were somewhat erratic.) The acid passed through the charcoal bed and was collected in
a receiver below the carbon bed. Periodically this receiver was drained and the acid analyzed
for HNS05 using the KMn04 method (see Appendix II).
An attempt was made to preheat the entering air stream, but the procedure was very
inefficient and it was finally decided to feed the air at room temperature. (Since the total heat
capacity of the air is small compared to the acid, this had little or no effect on the operating
temperature of the column). The exit gas was fed to the infrared batch cell for analysis.
2. Continuous tests
Several tests were run with this column with the results being given in Table
III. The column was somewhat difficult to operate in a completely continuous manner since
flooding frequently occurred in the charcoal beds due to packing support plates which did not
have enough open spaces to allow free drainage. There was no apparent plugging due to cat a -
lyst breakdown. The flooding problem was solved in later experimentation as will be described below.
Test A in Table III shows that, after running 89 ml of 0.639% HNS05 through the column
(which contained about 80 g of fresh, washed, and dried carbon, Witco type 256), the acid was
denitrated to 0.384% HNS05' a 40% reduction. A second run, test B, showed that the rate of
denitration was dependent upon the aCid-gas flow ratios, with denitration being increased when
the acid flow was reduced. During this stage, denitration of about 75% was achieved.
The material from test B was run again and the activity of the carbon appeared to go
rapidly down to zero with the final effluent acid being almost as concentrated as the input acid.
To try to duplicate the wash and dry batch experiments (Table II-A), the column was washed
with water followed by a 2-hr period where air was passed through the column at room tem-
perature. Acid was then passed through the column again using countercurrent air feed with
the result shown in test D in Table IlL The impression that is gained from these results is
that the wash water was never really dried out of the column and the low initial acid concen-
trations merely reflected some dilution of the acid.
Test E was made with the finer Witco type 517 carbon in three beds (in series), but
again the carbon appeared to drop in activity. This is shown more clearly in test F (which
used a single layer of Witco type 517 carbon) where the activity of the carbon dropped off
steadily during the test. Both of these tests were run at room temperature to see if activity
was improved at lower operating temperatures. (It should be recalled that the gas phase
oxidation of NO to N02 takes place at higher rates as the temperature is reduced. These tests
were aimed at evaluating the possibility of a gas phase oxidation mechanism. The results did
not support such a hypothesis.)
An extended test was run at high temperatures with the Witco type 517 carbon with the
results shown as test G in Table IlL The temperature was controlled fairly close to 2000 F
with the acid flow rate varying between 0.5 and 1.0 ml/min. During the 530 min of actual
operation, 412 ml of acid was passed through the column with the effluent acid showing a
23
-------
Tabk J/1. C')llIinuolis Colullln Tests - I-In. Columl1
Output Acid
Air flow. T ellllx'r,llu I'L'. .\dd Flow. L/G CUIll dlative Cumulative ConL'l:ntration
CDtalyst cc;min of cc. min Illolcm )Ie Till1~'. min Volullle>. Ill! % IINS05
A. Witeo carbon type 256 115 200 3.01 18.5 25 75
4 x 10 m ?sh 30 89 0.384
fresh carbon (80 g)
B. Type 256 (dried) 115 140 4.41 27.1 25 0.563
100 220 0.81 5.7 100 0.172
C. Type 256 (dried) 115 175-220 31 18.5 10 30 0.356
Material from B 45 115 0.621
D. Material from C: 200 200 3.21 11-3 10 30 0.0182
water passed through 200 175 2.51 8.8 300 0.500
column followed by 2
hr of air drying at R. T.
in column
E. Witco type 517 200 R.T.. 11 3.5 80 0.176
three layers. 200 R.T. 11 3.5 105 0.523
1 1;2-2 in. each,
~ 20 x 40 mesh (70 g)
,j>.
F. Type 517 single layer 200 R. T. 31 10.6
-4 in., dired (25 g)
G. Type 517
three layers
dried 2 hr (120°C);
1
2
200
70-175
185
13
3.5
25 0.238
55 0.420
85 0.666
115 0.640
90 304 0.296
120 80 0.209
175 155 0.023
215 180 0.367
240 205 0.175
270 230 0.191
280 255 0.310
320 280 0.362
360 305 0.497
405 350 0.527
445 375 0.677
480 385 0.566
510 395
530 412 0.659
r hr hold at 1250 F
3
4
5
6
7
8
9
10
11
12a
12b
12c
185
185-210
205-220
210
220
175-220
100-195
175-205
175-205
190-205
190-205
--205
53
17.6
Notes: 1. Input acid = 0.£39% w HNS05
2. Probably includes sume wash water that was not dried out
3. This is in addition to acid needed to soak dry column initially
4. 'Input: acid = 0-661% 'VV' HNSOs
-------
steadily decreasing catalyst activity. At the end of this period, the effluent acid was at the
same concentration as the incoming acid although the operating conditions were no different
from the point at which the column was achieving better than 70% denitration. This type of
behavior suggested that the denitration oxidation reaction was dependent on the oxygen ad-
sorbed on the fresh carbon and little or no additional oxygen was being adsorbed during the
operation of the column.
3. Theoretical considerations
It would appear from test G above that the denitration reaction continued until
a given amount of oxygen present on the carbon had been used up and then the carbon showed
no further activity. This type of behavior called for a closer look at what might be occurring
during the reaction.
A possible mechanism might include the diffusion of the oxygen through the acid to an
adsorption site followed by the adsorption of the oxygen onto the carbon. Simultaneously, the
nitrosylsulfuric acid has to diffuse through the acid to the oxygenated active site where the
oxidation reaction would take place. Thus it would be one of these three processes that might
be the rate limiting step in the overall reaction: (1) oxygen diffusion, (2) oxygen adsorption,
and (3) HNSOS diffusion. It is also possible that it could be the actual oxidation reaction itself
that would be the limiting step, but then there should be no variation in denitration rate with
time. The behavior pattern observed is also inconsistent with the concept that the diffusion
of the HNS05 to the active site is the rate limiting step. If the mechanism suggested above is
correct, then it would appear that it is either the diffusion of the oxygen to the active site or the
adsorption of the oxygen at that site that is the limiting step in the reaction.
If it is the adsorption at the active site that limits the overall reaction rate, this could
be ascertained by intentionally removing all the oxygen from the charcoal by prolonged expo-
sure to boiling water, drying in nitrogen at temperatures above the proposed operating tem-
perature, cooling in nitrogen to prevent oxygen adsorption, and then attempting to use the so-
treated charcoal in a standard batch test. If indeed it is the adsorption rate that limits the
reaction, this treatment should render the charcoal virtually inert and a batch test should show
no activity.
A series of batch tests was run which sheds some light on the problem. The results
on these tests are shown in Table IV. What is strikingly obvious in the results shown in the
carbon tests (tests C and D in Table IV) is that within a moderate limit of error all samples
were denitrated to the same degree, regardless of prior history. That this should be true even
of the sample that was never exposed to air throughout the entire test is quite unusual, but
explainable: Exposing the carbon to boiling water limits the degassing to 212 of, and it is
well known that this type of treatment is not effective in removing all the oxygen adsorbed on
activated carbons. It would seem that sufficient oxygen remained on the carbon (even after
the high temperature drying) to achieve a high level of denitration during the batch test.
25
-------
Table IV. Batcn Tests With Activated Charcoal*
Test
Temperature,
op
Final Acid Concentration,
% HNS05
0.698
A.
B.
Blank (saddles/air)
Blank (saddles/N2)
0.726
250
250
C.
Carbon (as-received)
1. air
2. air
3. air
275
265
265
0.192
0.083
0.201
D.
Degassed Carbon
Contacted with boiling water
for 1 hr under nitrogen then
dried (260-300 OF) and cooled
under nitrogen
1. Exposed to air 1 hr before
test: test run with air
2. Kept under N2 until test:
test run with air
250 0.153
250 0.096
250 0.163
250 0.114
3. Kept under N2 until test:
test run with N2
* All carbon tests on Witeo type 256 activated charcoal using stock HNS05 solution
containing 0.740% HNS05 by weight.
26
-------
4. Continuous column test - temperature variation
One test was run in the I-in. column to see the effect of varying the temperature
on the denitration capability of the activated charcoal. Three beds of Wit co type 256 carbon
were set up in series with provision made for determining the acid temperature in the second
bed. The acid was again collected in the reservoir at the bottom of the column and sampled at
regular intervals. Both the acid and the air were fed to the column at 75 to 80 of and were
heated in the column by heating tapes wrapped around the pipe sections. The column tempera-
ture was varied during the experiment.
Fig. 7 is a graphic representation of the history of this experiment. The graph shows
the concentration of the effluent acid as a function of the total amount of acid that was passed
through the column. The circles indicate the data points taken and are numbered in order
while the column temperatures are indicated next to the data points. The air flow was also
varied during the experiment, and the flow rates are shown at the bottom of the graph.
At the beginning of this run, about 70% denitration was achieved at a relatively low tem-
perature (point 1 at 145 OF). This activity dropped off sharply and was probably due to loosely
adsorbed oxygen and air trapped in the pores of the carbon. Then, after point 3 at 165 of, the
activity increased as the temperature rose to 220 of at point 6, but dropped off again at points
7 and 8. Apparently the additional oxygen available at the higher temperatures was used up
and activity decreased. At this point, the air flow was increased by an order of magnitude
and the activity increased again and remained relatively constant from point 9 to point 13 when
the air flow was increased again. Again the activity increased although the temperature
dropped to 210 of. From there to the end of the experiment at point 17, the activity of the
carbon dropped greatly as the temperature was reduced to 185 of.
It would appear from this experiment that both temperature and air flow play important
parts in the denitration oxidation reaction. When the air flow was increased at constant tem-
perature the activity increased,and when the temperature was increased at constant air flow
the activity again increased. These phenomena were evaluated more completely in subsequent
experiments.
C. Continuous Column Experiments (2-In. Column)
1. Column redesign
The main problems with the I-in. column were consistent acid holdup while air
was flowing and the difficulty in preheating the acid and feeding it at a controlled rate. These
problems were solved by utilizing a 2-in. column constructed in a modular fashion from 2 -in.
Corning flanged pipe. The packing support plates were redesigned to permit easy flowthrough
of the acid, and a simple acid heater was installed which allowed the close control of incoming
acid temperature. Acid flow rates were controlled by the use of a variable speed peristaltic
pump (Randolph pump, model 250). As shown in Fig. 8, the pump delivered the acid to a
heated I-in. tube, and the acid was fed to the column by overflow through aU-tube arrange-
27
-------
Lt')
o
CIJ
z
.......
......
cf2
~
d
o
.....
~
cd
~
Il)
g
8
-:J
.....
U
-<
Inlet acid concentration
----------------0---.-- - - - ---- ---
3 165°F
I-in. column
Witco type 256 activated <:arbon
4.7 in.:J (in 3 beds)
Acid flow: 2 cc/min (average)
.8
.6
11 1850F
195°F
185°F
4
Effluent acid
concentra tion
.2
I 145°F
50-1001
Air = 285 cc/min
~
II 10 11,00 1285 I
110
1- Air flow,
c / min
10
100
200
300
400
500
Cumulative Acid Throughput Volume, ml
Fig. 7. Continuous stripper experiment
28
600
-------
Acid
pump
Gas exit
Condensate trap
ACid
heater
Thermometer
Gas
entrance
Acid
reservoir
Acid
exit
Fig. 8. 2 -inch continuous column for catalyst evaluation
29
-------
ment. The U-tube system provided a gas seal at the top of the column. A photo of the
apparatus is shown in Fig. 9.
The column was composed of three beds in series and thermometer wells were installed
so that the temperature of each bed could be monitored. The packing support plates were made
of Teflon and were slotted so that more than 60% of the area was open. Small glass rods were
placed vertically in some of the slots to act as downcomers for the acid. A large empty sec-
tion was placed at the bottom of the column to act as an acid reservoir which could be period-
ically drained for sampling. The entire packed section could be removed from the column
without disturbing the basic setup so that the catalyst could be easily changed or examined.
An initial attempt was made to heat the incoming air, but again this proved inefficient
and the air was fed at room temperature. The gas outlet was fitted with a small amount of
glass wool for stopping water vapor and a trap was provided for collection of the water. The
gas outlet was fed to the infrared cell for analysis.
2. Continuous column testing (2 -in. column)
During a period of several weeks, many experiments were performed in the
2-in. column. Since the column contained a removable section which housed the catalyst bed,
it was possible to run several experiments with one bed, remove that bed section and replace
it with another, and then return to the first bed without disturbing the packing in either bed.
Three beds were evaluated during this phase of the work and were set up as follows:
Column 1: Three beds of Witco type 256 activated charcoal in series con-
taining a total depth of 10 in. of catalyst and about 30 cu in. of material
Column 2: Identical in size and configuration to column 1 except that the
beds were packed with 1/4 -in. ceramic saddles to act as a blank for the
charcoal test
Column 3: One bed of Witco type 256 activated carbon packed to a depth of
22 in.
The results of the testing in each column are given graphically by showing the HNS05
concentration of the effluent acid as a function of the amount of acid passed through the column
and comparing this concentration with that of the input acid. The plots show cumulative acid
throughput as the abscissa with the total figures giving the amount of acid passed through each
column since the column was first packed. Thus the graphs show not only the results of each
specific experiment, but also the long term stability and activity of the catalyst bed.
3. T es ting with column 1
Throughout the testing with this column, the acid flow rate was maintained at
about 10 ml/min while the air flow rate was varied depending on the goals of the individual
experiment. The temperatures of the three beds were continuously monitored, and it was
found to be difficult to maintain the same temperature in all three catalyst beds. The tempera-
30
-------
~b
'<. ..;'
~, ('
. I J*;
,
+l
\.
Fig. 9. Laboratory apparatus for continuous evaluation of stripper catalyst
31
-------
tUre that will be used to describe the operating condition of the column will be the highest
temperatUre which consistently occurred in the middle section. The bottom bed was usually
within 10 of while the top section was often 20 OF cooler. Acid samples were taken and
analyzed for HNS05 content every 15 min.
Experiment 1- A: denitration at 200 of - In this experiment, the tempera-
tUre was maintained fairly constant at 190 to 205 OF as can be seen from the data in Fig. 10.
Although the catalyst showed good activity for the first 1500 ml of acid throughput, the level
of denitration dropped off sharply after that, as shown by the jump in output acid concentration
after point 12 on the graph.
The first 22 points suggest a behavior pattern that was described in Section B-3 above.
The catalyst was providing oxygen that had been adsorbed prior to the test and when that oxy-
gen was used up, the reaction virtually stopped. At point 23, the acid was stopped and air was
forced through the column (at a rate in excess of 1000 cc/min). After this, the low flow of
300 cc/min of air was resumed and the activity appeared to have increased. It would seem
that the period of air flushing restored some of the carbon's activity after which the activity
again decreases toward the plateau level of points 15 through 22. At point 26, the air flow was
doubled to 600 cc/min, and the catalyst activity again increased noticeably. After point 27,
the air was pulsed through the column: 1 min at 300 cc/min and 1 min at such a rate that the
acid would not flow through the carbon beds. This was maintained for about 30 min through
data points 28 and 29.
This experiment apparently shows that after the carbon's initial quantity of adsorbed
oxygen is used up, it becomes largely inactive. However, if the air supply is greatly increased,
some of the original activity can be restored.
Experiment 1-B: denitration at 250 of - After point 29, the column was shut
down and cooled overnight with experimentation being resumed the next day at the same operat-
ing conditions (low air flow) as during most of test 1- A. As can be seen in Fig. 11, the ef-
fluent acid showed the same level of activity as when the air had been pulsed but quickly
resumed the low plateau level of the previous day (points 15 through 22). After point 33, the
air flow was increased to 800 cc/min and again the denitration increased. It was clear that a
higher air flow rate was needed to maintain acceptable denitration rates at the 200 of tempera-
ture level.
At this point (data point 35), the column temperature was increased and immediately
the denitration rate increased. Although there was an inexplicable jump at points 40, 41, and
42, the activity of the catalyst appeared to level off at about 70% denitration when the column
temperature leveled off near 250 of.
This experiment suggests that higher temperatures are necessary to achieve an ac-
ceptable level of denitration, but whether this is due to increased diffusion rates or because
of the capability to provide larger required activation energies for adsorption was not certain
at this time. In any case, it appears that 70% or better denitration is possible at this tempera-
ture and at this acid/ catalyst contact time.
32
-------
2-in. column
Witco type 256 activated carbon
30 in.3 (in 3 beds)
Feed acid
",,-----~---
.8 ,,-
'"
------D"""
2000F
---0,--- - - - -- - - - -0---
It')
0
CI)
~ .6
~
~
..
s::
0
.....
C,,) ....
C,,) C'IS ~
~
Q)
g
c
0
"0
......
C) .2
<:
185°F
205°F
Air flow = 300 cc/min
from points 1-22
High air flOJ R~sumed
No acid 300 cc/min
air flow
Effluent acid
Started
pulsing air
8
195°F
200°F
12
500
1500 2000 2500 3000
Cumulative Acid Feed Volume, ml
3500
1000
Fig. 10. Continuous laboratory stripper test: Experiment 1- A
4000
-------
.8
It:>
0
C/)
~
cf2 .6
~
a
0
.....
....
co
H
....
!:: .4
C)
u
!::
0
U
'0
.....
U
--<
.2
Air flow raised
to 800 cc/min
-----, I
't,_____Fee~ci~__c-_----
200°F Started heating column
33 f 2150F
35
31 2000
175°F
1
Column not operated for
18 1/2 hr after sample 29
(Fig. 10); Cooled, then
reheated
250°F
245°F ~ 45 49
2~0
Effluent acid 248 F
4500
5000
5500
6000
6500
7000
Cumulative Acid Throughput Volume, ml
Fig. 11. Continuous laboratory stripper test: Experiment 1-B
34
-------
Experiment 1- C: extended period denitration - Test 1- B was terminated
before it could be ascertained that we were not just utilizing additional oxygen that might be
available at higher temperatures, but were really in a steady state situation. Therefore, test
1- C was run after a 3 -day shutdown and is shown by data points 50 through 102 in Fig. 12.
This run which covered 13 hr and almost 7 t of acid makes it clear that steady state can be
achieved at 50 to 60% denitration at 250 to 260 of and the existing contact time. In test l-C,
the entire operation was started at room temperature and data points taken as the column was
slowly heated to its steady state maximum of about 260 of. Again the catalyst showed poor
activity at low temperatures but increased significantly as 260 OP was approached. It is
interesting to note the difference in the slopes in test I-B (points 36 through 39) and l-C
(points 52 through 62) which appear to be directly proportional to the heating rate in each
experiment. (One additional point should be noted before the discussion continues: the low
concentration of the input acid in the 11,000 to 12,500 ml total acid period.) The practice was
to mix the acid in 2 t quantities and take a sample from the hot feed acid midway during the
2-t feed period. These analysis points are shown by the boxed data points on the inlet acid
concentration line. The sample taken during the period in question exhibited bubbles which
were not seen during any of the other sample analyses and might have been due to a wet or
dirty sample flask. It is felt that this inlet level appears lower than it really was and the fact
that the outlet acid showed no corresponding change in concentration tends to verify this feeling.
Experiment 1-0: denitration at 290 of - Having achieved greatly improved
continuous denitration levels at 260 of, it was decided to evaluate still higher temperatures.
Test 1-0 (Fig. 13) was therefore run at a steady state operating temperature of 290 of with
the resulting improved denitration level shown between data points 108 and 119. Under these
conditions, the stripper caused about 75% denitration of the nitrose.
These higher temperatures were chosen in order to determine the effect of temperature
variation on the denitration capability of the charcoal and to try to obtain sufficient informa-
tion to understand the mechanism whereby the charcoal effects denitration of the nitrose. The
290 of used in experiment 1-0 was higher than originally recommended for the recycle tem-
perature of the 80% acid loop, but this temperature range is certainly worth investigation
since variations in input flue gas conditions and changes in acid concentration might well
require higher operating temperatures. It is clear from these first few experiments that
denitration of nitros
-------
c:..:>
0)
.8____- -
Feed acid
-- ----1-- ----
---- -- ---,,---------
U")
o
V)
~
102
c:f2 .6
~
Effluent acid
d'
o
......
.....
(1j
H
.....
!::
Q)
U
!::
8
245°F
.4
After 3 day
shutdown, test
was started from
room temperatUre
acid - 10 cc/min
air - 600 cc/min
L/G == 11.8 moles/mole
Steady state operations from
points 71 to 102.
TemperatUre constant at 2600 F
Effluent acid concentration = "'.35%
2600F
"0
......
u
-< .2
7000
7500
8000
9000
14,000
9500
10,000
8500
13,500
Cumulative Acid Feed Volume, ml
Fig. 12. Continuous stripper experiment: Experiment 1-C
-------
.8
It:)
0
U1
z
:r:
cf2 .6
~
d'
0
......
....
Cd
~
....
c:: .4
Q)
u
c::
8
'0
......
U
-------
Saenger, the vapor pressure of NP3 above a 0.100 mole HNS051 £ solution in 80% H2S04
would be 5.8mm Hg. So if the column were operating under equilibrium conditions with no
catalytic activity, the partial pressure of nitrogen trioxide would be about half of the partial
pressure that must exist in order for the experimentally determined degree of denitration to
occur. The denitration is undoubtedly far more effective than this since it is highly likely that
we are operating with such short contact times that we are nowhere near equilibrium. This
was verified when additional experiments were run under the same conditions with inert pack-
ing (see below: experiments with column 2 with ceramic saddle packing).
Experiments l-E, I-F: multiple stage denitration - The effluent acid from
experiment I-D was saved and used as the feed stream for another denitration experiment run
at 290 of (experiment I-E). This acid had a composite concentration of 0.320% HNS05' As
shown in Fig. 14 at steady state at 290 OP, the column caused denitration to about 0.08% HNS05'
again about 75% denitration. The effluent from this experiment, with a composite concentration
of 0.110% HNS05' was passed through the column again (experiment 1- F, Fig. 15). The ef-
fluent from this third stage of stripping had a concentration of 0.03% HNS05' having had about
73% of its HNS05 removed.
The overall result of these three experiments (I-D, l-E, I-F) was that a nitrose of
0.800% HNS05 was denitrated to 0.030% HNS05 in three passes through the column. Thus,
more than 96% of the HNS05 in the nitrose solution was removed by an effective bed depth of
about 30 in.
Examination of the experimental conditions shows that the column was more effective
than that. Each pass through the column removed about 75% of the HNS05' but the feed for
each of the last two experiments was not the steady state effluent from the previous run. This
was due to the fact that the feed streams were composites of the total run which included the
higher strength effluent from the early parts of each run where the column was heating up. If
the 0.800% HNS05 nitrose had been passed through 30 in. of carbon and each 10 in. accomplished
the removal of 75% of the input nitrose (which seems to be a reasonable assumption in light
of the experimental evidence), the final effluent would have been 0.013% HNS05' The overall
effect is a denitration of more than 98%, closely approaching the desired 99% denitration level.
For the purpose of engineering design calculations, it is important to have some m2a-
sure of the contact tim2 between the acid, the air, and the charcoal catalyst. The residence
time of the acid in the column is somewhat difficult to determine exactly because the difference
in path length for the acid would create a distribution of residence times that would be dependent
on the path configuration. This could be measured by use of a tracer material whose concentra-
tion in the acid effluent could be monitored, but this degree of sophistication is probably not
necessary at this stage. A good preliminary design can be achieved by optimizing criteria
which are easily available: L/G, acid space velocity, and gas space velocity.
38
-------
~
o
~
~ .6
~
~
d'
.9 4
.j...J .
Ctl
H
.j...J
!::
Q)
~
8
'0 .2
.....
u
<
.8
Air flow:
Acid flow:
L/G:
835 cc/min
10 ml/min
8.5 mole/mole
Acid feed
----------0-------------- -..,
250°F \
121 \
,
o
17,000
290°F
126
290°F.
130
Effluent acid
290°F
18,000
135
Cumulative Acid Feed Volume, ml
19,000
Fig. 14. Denitration of recycled acid: Experiment 1- E
39
290°F
140
2900F
143
20,000
-------
g- .4
.....
....,
C(j
~ \
~ \
8 2 \
] . 290 L - - - - - - - -0- - - - Acid~eed - - - - - - - -D-
of ° Effluent acid
144 295 F 290°F 2950F 295°F 295°F
15() I
21,000
Ln
0 .6
r.rJ
Z
:r:
cf<
~
.8
Air flow:
Acid flow:
L/C:
835 cc/min
10 ml/min
8.5 mole/mole
o
20,000
22,000
Cumulative Acid Feed Volume, ml
Fig. 15. Denitration of recycled acid: Experiment 1- F
40
-------
The I.)G for each run is indicated on the denitration graphs in this section of the report
and varies between 3 and 30 moles of acid per mole of air, with values around 10 giving the
best results. For an acid flow rate of 10 cc/min, the acid space velocity is 1.7 hr -1 (using the
350 cc void volume in the 2 -in. x 10-in. charcoal bed and assuming 70% voids in the bed). On
the same basis, a gas flow rate of 835 cc/min yields a gas space velocity of 140 hr -1 (STP).
Experiments 1-G, 1- H: air flow variation - Experiments I-A through 1- F
were made without any effort to optimize the flow conditions of the column. Emphasis was on
operating temperature and degree of denitration. A series of experiments was run where the
degree of denitration was evaluated as a function of gas and liquid flow rates to try to develop
preliminary estimates of flow parameters for future engineering design studies, as well as for
the Tyco miniplant.
The first experiment, no. 1-G (Fig. 16), evaluated the effect of air flow rate on the
degree of denitration. The column was started up and brought to steady state at about 260° F
at point 167 using an air flow of 835 cc/min and an acid flow of 10 ml/min (UG = 8.5 mole/mole) .
(Due to a change in the electrical hookup before this experiment, there was some difficulty in
reaching 290°F. This was corrected after this experiment.) After data point 174 was taken,
the air flow was decreased from 835 cc/min to 320 cc/min (UG = 21.1 mole/mole). The tempera-
ture rose to about 270°F, and the degree of denitration was apparently reduced. After point 181,
the air flow was further reduced to 96 cc/min (L/G = 74.3 mole/mole) giving a similar reaction
in temperature and denitration.
Since the column was not operated at 290° F throughout experiment 1-G, it was decided
to run the test over again at that temperature. Experiment 1-H shown in Fig. 17 was a repeat
of the previous experiment except that the air flow variation was reversed: low flow at the
start and high flow at the finish. The results of this experiment again show a general trend
towards reduction in denitration as the air flow is reduced.
Experiment 1-1: acid flow rate variation - The results of the previous ex-
perimentation indicated that for the 2-in. column being used it was necessary to pass 835 cc/min
of air in order to effect 75% denitration of 10 cc/min of acid. Another experiment, no.1-I, was
run to see if it were possible to denitrate more acid with that same gas flow rate.
In this experiment, the L/G was doubled (to 17.0 mOle/mole) by increasing the acid flow
to 20 ml/min. As is shown in Fig. 18, the nitrosylsulfuric acid concentration of the nitrose was
reduced from 0.79 to about 0.5%, a 37% reduction. It should be noted that, since the column
activity had apparently dropped off somewhat during the last few experiments, this degree of
denitration should be compared to the previous run no. 1- H which showed a 50% denitration
level. In any case, it would seem that denitration is definitely a function of L/G under the condi-
tions present in this series of experiments.
Experiment 1-J: cocurrent column operation - In order to evaluate the
catalytic stripper concept as completely as possible, the column was run under cocurrent gas/
liquid contact conditions. The acid was fed to the top of the column at 10 ml/min, and the air
was also fed to the top of the column at 835 cc/min. This experiment, no. 1-J, was performed
at about 290°F and the results are shown in Fig. 19.
41
-------
.8
I!':>
o
(f)
z
::r:
cf2
....
~ .6
Acid feed
-- ------0-- -- - - ---- - ---0--- - ----
.........
I-- 95 cc/min ~--
2800F 295°F
320 cc/min ~ 185 295°F
187
L/e = 74.3 mole/mole
~
260~
170
180
~
~
c
.S
....
Ctl
H
....
C
(j)
g .4
o
U
'0
'u
<
I"
Air flow
835 cc/min
~I
L/e = 8.5 mole/mole
Effluent acid
f74
o
23,000
24,000 25,000
Cumulative Acid Feed Volume, ml
26,000
Fig. 16. Column optimization: Experiment l-e, air flow rate variation
-------
l{')
o
C/)
z
::r::
cf2
.....
~
s::
o
.....
....
Ct!
H
....
s::
Q)
u
s::
o
u
:s
~ u
~ -<
Acid flow: 10 ml/min
,.-----;r---- ----
/
Acid feed /
/'-----------~---_/
/
/
.8
.6
280
of
275 190
Of='
4 189
290°F
Effluent acid
213
A ir flow
\.-105 cc/min -+-
.2
210 cc/min
-t- 420 cc/min --\.-
~
835 cc/min
L/G = 67.3 mole/mole
L/G = 33.7 mole/mole
L/G = 16,8 mole/mole L/G = 8.5 mole/mo
o
27,000
28,000
30,000
29,000
Cumulative Acid Feed Volume, ml
Fig. 17. Column optimization: Experiment 1- H, air flow rate variation
-------
It")
o .8
C/)
z
:r::
cf2
~
~
c:~ .6
.s
~
Cd
....
~
c:
Q)
()
c:
8 4
:9
()
~
Acid flow: 20 ml/min
Air flow: 835 cc/min
L/C = 17.0 mole/mole
......
" Acid feed
~--- -- -- - -0--- -- - ----- - -0---
230°F
215
285°F
220
290°F
222
290°F
224
Effluent acid
.2
o
31,000
32,000 33,000
Cumulative Acid Feed Volume, ml
Fig. 18. Column optimization: Experiment 1- I, acid flow rate variation
44
-------
10 .8
o
[J)
z
:r:
cP
.....
a: 6
~ .
s::
o
....
.....
CtS
H
.....
s::
OJ
g ~
o
U
"0
.u
-<
Acid feed
-- 0---- -- ---- ----- ---
''''''-0-----
o
33,000
285°F
295°F
290°F
235
226 Effluent acid
Air flow: 835 cc/min
Acid flow: 10 ml/min
L/G: 8.5 mole/mole
34,000
35,000
Cumulative Acid Feed Volume, ml
Fig. 19. Co-current flow: Experiment I-J
45
295°F
238
-------
Although the acid was denitrated only by about 40%, some interesting properties of this
type of operation were noted. First of all the N02 in the exit gas stream appeared visually to
be more concentrated than that in the countercurrent experiments. Infrared analysis of this
exit gas clearly showed this level of N02 (about 5.4mm Hg partial pressure or about 0.7%) and
permitted the calculation of a material balance within 13%, a precision that had not been easily
obtainable in the countercurrent experiments.
At the same time, there was apparently far less water vapor in the exit gas stream
which may account for the improved material balance. In the countercurrent experiments,
there was a great deal of condensation at the point where the gas comes off the column because
the exit line is not insulated and thus the gas is rapidly cooled. In the cocurrent experiment,
the condensation must be taking place in the bottom of the column (which is relatively cool)
and is removed with the exit acid. It appears, then, that the condensing water in the counter-
current experiments is scrubbing the exit gas, thus removing some of the N02 before it enters
the infrared cell. To analyze this hypothesis, the countercurrent operation was carefully
observed, and it could be seen that the water condensed all through the exit gas line, and
at some times the exit gas actually bubbled through pools of condensed water. Some of this
condensed liquid was removed from the system and analyzed and was found to be very acidic
(about pH 0.5). Thus, the N02 was reacting with the water to form nitric acid. These points do
not make a case for the cocurrent operation of the stripper but the experimentation does pro-
vide some insight into the operation of the countercurrent operation of the column.
Experiments 1- K, 1- L: evaluation of apparent catalyst deactivation - Ex-
periments 1-H and I-I showed an apparent reduction in activity of the activated carbon in
column 1. At the time when no. 1-H was started, more then 26 £ of nitrose solution at high
temperature had passed through the column, but from all external evidence very little change
could be seen in the carbon. There had been some carbon fines removed during the first ex-
periment, but the effluent acid had been free of particulates ever since. Had the carbon really
lost its activity and if so, why?
Experiment l-K was run to see if the catalyst had lost its activity. Experimental con-
ditions were virtually identical with those of run no. 1-0 which showed about 75% denitration.
The results (see Fig. 20) showed that the 0.760% HNS05 feed solution was denitrated to 0.320%
for a denitration of 58%. This is a significantly lower level of catalyst activity similar to the
results of experiment 1- H.
In evaluating this apparent deactivation, two mechanisms were considered: (1) physical
deterioration, and (2) chemical poisoning. As was mentioned above, there was no evidence of
physical breakdown of the carbon as might be measured by the production of fines during opera-
tion, and in addition, a visual examination of the carbon through the glass walls of the column
showed no gross change in the size or shape of the particles. The conclusion was made that
the deactivation was chemical in nature.
Since the only materials that entered the column were air and nitrosylsulfuric acid
solution, it was concluded that the poison must be entering with the acid. Physical examination
46
-------
It) .8
o
V)
z
::c:
I:fi
~
-J
....
~ .6
a
,9
....
(1j
I-<
....
c::
(l)
g 4
o
U
'0
.....
u
-------
of the acid plus observations of earlier experiments gave a clue to the possible source of the poi-
son. The acid that had been used throughout the stripper evaluation was 80% H2S04 that had been
purchased in bulk about a year prior to use. About 2000 lb of this commercial grade acid had
been purchased from a local supplier in order to avoid the problem in diluting large quantities
of acid. The acid had been stored in special acid resistant black polyethylene carboys. When
fresh, the acid had been almost water white in color, but more recently it had
been noticed that it was now brownish in color although still clear. A check with the supplier,
Nyanza O1emical, revealed that the color was due to some carbonization of the polyethylene
as well as some leaching of carbon filler from the black polyethylene carboy. Titration of the
fresh acid with KMn04 showed the requirement for a negligible amount of the oxidizing agent
to form the characteristic purple end point. It was therefore initially assumed that the dis-
colored acid was not interfering with the analysis of the nitrose solution.
However, it was observed in early experiments that the effluent acid from the stripper
was noticeably lighter in color than the feed acid. Thus, the charcoal was decolorizing the
acid. Since the deactivation was observed, a comparison was made among several column
effluent solutions that had been saved from runs made over a period of time, and it was noted
that the early effluents were obviously decolorized but the latest runs were not. The ability
of the carbon to adsorb the color-producing material from the acid had definitely been sharply
reduced. If the carbon's adsorption sites for decolorizing were
the same as the catalytic sites for nitrose oxidation, it would be possible to explain the reason
for the stripper deactivation.
In order to properly analyze the solutions for evidence to support this hypothesis, it
was necessary to develop another technique for measuring the nitrogen oxide content of the
acid solutions. The method chosen was the nitrometer, an apparatus widely mentioned in the
literature as appropriate for this purpose. The technique involves the reaction of fixed nitro-
gen in solution with mercury resulting in the evolution of nitric oxide. The volume of the NO
is measured, thus permitting the calculation of the NO content of the original solution. The
x
procedure is somewhat more time consuming than the KMn04 titration which is why it was not
used originally.
A series of experiments was performed to determine if the 40 £. of contaminated com-
mercial acid had deposited sufficient material on the carbon to reduce its catalytic activity.
First, the column was used to denitrate a nitrose solution under conditions almost identical
to run no. 1-D, except that reagent grade acid was used to make up the nitrose. This test,
experiment 1- L, again appeared to demonstrate that the catalyst activity had been reduced
with the carbon achieving only about 40% denitration of the nitrose (see Fig. 21). What was
more significant was that the effluent acid was strongly discolored although the feed acid
was water white. It appeared that the more pure nitrose was causing the desorption of the
contaminant that had been loading the catalyst.
48
-------
It'>
o
V) .8
z
:r:
I:f2
....,
~
5.6
.....
....,
Cd
H
....,
c:
Q)
u
c:
o
U~
:9
~ u
co -<
.2
Acid feed
--- - --- ------ - --[]- - ---
------
295°F
250°F
266
Effluent acid
288
Air flow: 835 cc/min
Acid flow: 10 ml/min
L/G: 8.5 mole/mole
o
40,000 41,000
Cumulative Acid Feed Volume, ml
42,000
Fig. 21. Oenitration using reagent grade acid in nitrose: Experiment 1- L
-------
To continue this cleaning, it was decided to wash the column with hot reagent grade
acid rather than nitrose made from this acid. The first liter of acid (which passed through
the column at about 175 OF) was made up of 96% reagent grade acid and resulted in a very dark
effluent. The second liter of wash solution was 80% reagent grade acid and came out of the
column almost as clean as it went in. It was hard to tell if the concentrated acid had attacked
the carbon or if it had just accelerated the cleaning of the column. To test this, a small
amount of fresh carbon was heated to 175 of in 96% reagent grade acid. The acid turned inky
black apparently with carbon fines. The acid was drained off and the test repeated with more
acid. No more fines were visible although the acid turned greenish in color completely dif-
ferent from the effluent from the column. It was therefore assumed that the 96% acid had
accelerated the washing of the column and was not causing the breakdown of the carbon. As
further support for this conclusion, the effluent from the column was cenrrifuged for 30 min
without obtaining any sediment. Previous attempts to cenrrifuge effluent solutions which ap-
peared to show carbon fines (as when fresh carbon was treated with nitrose) had resulted in
partial clarification and the collection of sediment.
In order to put all these data in perspective and obtain a quantitative approach to the
problem, several solutions were analyzed both by titration and with the nitrometer. The re-
sults of these analyses are shown in Table V. The first test on the fresh acid indi
cates that there is some oxidizable material in the acid that contains no fixed nitrogen. It is
present in very low concentrations which justifies the original assumption that it would not
interfere with individual acid analyses. Test 2 shows no reactive species detectable in the
reagent grade acid by either test.
The third set of analyses shows a reproducible difference in HNS05 concentration when
performed by KMn04 titration as compared to nitrometer testing. The fact that the titration
is consistently low is attributed to a small loss in NO during titration because the sample
x
heats up due to the exothermic oxidation, thus vaporizing some dissolved nitrogen oxides.
The difference between the two techniques appears to be fairly constant at about 5%.
The fourth test shows that the HNS05 analysis when using barrel acid exhibits a smal-
ler difference between the KMn04 titration and the nitrometer than when using reagent grade
acid, possibly because of the contamination in the barrel acid.
The fifth test shows about 10% more" HNS05" by titration than by nitrometer. Since
the titration had been running lower than the nitrometer on standardization sample as dis-
cussed above, this 10% increase is quite significant. It appears to be due to the removal of
some oxidizable material from the column.
Test 6 performed on the effluent from the attempt to denitrate reagent grade acid shows
a 20% higher HNS05 concentration by titration than by nitrometer. Here the clean acid ap-
pears to be rapidly washing out the contaminant from the carbon.
The results of test 7 clearly show that there are oxidizable species in the column that
contain no oxides of nitrogen. This result in conjunction with the others above appears to give
50
-------
Table V. Comparative Analyses of Acid Samples
HNS05 Concentration, %*
Sample KMn04 Titration Nitrometer
1. Fresh barrel acidt 0.0037 nil
2. Reagent grade acid (80%) nil nil
3. Reagent grade acid plus 0.628 0.664
HNS05 0.630 0.670
4. Barrel acid plus HNS05 0.655 0.667
5. Effluent from stripper run 0.368 0.333
no. 1-G
6. Effluent from stripper run 0.480 0.385
no. 1- L 0.485 0.415
7. Effluent from reagent grade 0.197 0.030
acid wash:/:
* All results are reported as apparent HNS05 concentrations even where the KMn04
titration is clearly measuring species other than NOx'
t Barrel acid refers to 80% H2S04 that had been stored in black plastic carboys for
about 1 yr.
:/:This is a mixture of 96% and 80% acids from the 175 of columns wash described in
the text.
5'1
-------
strong support to the original hypothesis that the column adsorbed some contaminant, probably
organic, which partially poisoned the column of activated carbon. In actual use, this type of
poisoning should not be a factor since flue gas does not contain similar contaminants.
4. Testing with column 2 - experiment 2
In order to determine the effect of vapor pressure denitration (thermal stripping) on
the operation of the stripper column, an experiment was run using ceramic saddles in the column in-
stead of activated charcoal. Column no. 2 was set up so that it was physically the same as column 1
(except for the difference in surface area between the ceramic saddles and activated charcoal):
three beds in series containing about 30 cu in. of saddles to a depth of about 10 in. Experimenr
2 was run at about 290 of with flow rates of 10 ml/min of acid and 83S cc/min of air. These
conditions were identical with those in experiment 1-D which showed about 7S% denitration.
Fig. 22 shows the results of experiment 2 with the 0.670% HNSOS feed acid being
denitrated to about 0.S70% HNSOS' This is a denitration level of 1S%, obviously far lower
than that achieved with the activated charcoal.
At equilibrium at 290 of with the acid and gas flow rates and concentrations as present
in this experiment, the acid should have been denitrated to about 0.31% HNSOS' Clearly, the
column conditions were far from equilibrium.
A somewhat unusual aspect of the experiment was the fact that at the beginning of the
experiment the effluent acid had a considerably higher HNSOS concentration than the feed acid.
Since this is impossible, it is necessary to examine the conditions and the analytical technique
to explain the situation.
The analysis for HNSOS in sulfuric acid is accomplished by titrating the acid with
potassium permanganate, a strong oxidizing reagent. The KMn04 oxidizes the NOx content
of the acid to HN03 and thus allows the calculation of the amount of fixed nitrogen in the acid
solution. However, the KMn04 titration is not specific for NOx but will react with any oxidiza-
ble material in solution. What must have occurred in this experiment is that there was some
contaminant on the fresh saddles, perhaps an organic of some sort, and the sulfuric acid
washed the material off the saddles. When the effluent was titrated, the KMn04 reacted with
both the NOx in solution and the organic contaminant, thus showing a high apparent HNSOS
concentration. This contaminant was eventually washed off the saddles and the steady state
effluent concentration shown after point 6 was apparently due only to HNSOS'
S. Testing with column 3
A total of five runs were made on column 3 which consisted of one 22-in. bed
of activated charcoal in a 2-in. column. Nitrose and air were passed in countercurrent streams
and the gas and acid streams analyzed as in previous tests. All five test runs were made
at steady state temperatures of 290 OP with fluid flows being held constant at 10 ml/min for
the acid and 835 cc/min for the air. The nitrose concentration was in the O.S to 0.6% HNSOS
range.
52
-------
1.0
220°F
Lt)
o
C/)
~.8
~
....
~
~
r::
o
.,'j .6
ctt
H
....
r::
Q)
()
r::
o
U
en :94
c..:I
()
<
Acid feed
--------- -- --- ----- --- ---
300°F
Effluent acid
Acid flow: 10 ml/min
Air flow: 835 cc/min
L/G: 8.5 mole/mole
.2
o
o
200
400
600 800 1000
Cumulative Acid Feed Volume, ml
285°F
10
1200
1400
Fig. 22. Stripper blank - saddle packing: Experiment 2 - A
1600
295°F
13
-------
A total of 14 £ of nitrose were passed through the column during five runs (totaling
about 2S hr of operation) and after an initial period where impurities were washed out of the
carbon, the effluent acid maintained a concentration of less than 0.007% HNSOS' which repre-
sented a denitration of about 99%. Analysis of these extremely low concentrations was quite
difficult, and a technique was used that gave the maximum possible HNSOS content rather than
an exact figure. Instead of titrating the acid solution to an end point (which was very difficult
to see at these low concentrations), a given amount of KMn04 was introduced into an aliquot
of the acid. If this amount of reagent showed that the end point had been exceeded (which it
invariably did), it could be safely said that the HNSOS concentration was less than an amount
equivalent to the KMn04 added. The amount used was equivalent to 99% denitration, so it is a
fair statement that, under the steady state conditions indicated above, denitration was greater
than 99% complete.
D. Summary of Results of Charcoal Evaluation
The tests with the Witco activated charcoal in batch and continuous equipment has pro-
duced positive results which verify the concept of the catalytic stripper and indicate that the
charcoal can be used for denitration of nitrosylsulfuric acid solutions in the Tyco Catalytic
Chamber Process:
1. The apparent physical instability of the carbon particles observed in earlier
testing has been shown to be temporary in nature. Initial contact of the carbon with hot nitrose
results in the production of fines which apparently result from initial rapid reaction on the char-!
coal surface. After a short period of time, the production of fines stops and there is no further
physical breakdown of the particles. At the same time, there is no apparent reduction in
catalytic activity when the carbon is stabilized.
2. The reaction of the carbon with oxides of nitrogen which resulted in the pro-
duction of nitrous oxide (N20) is apparently a minor problem. The amount of N20 produced is
a negligible fraction of the total NO content of the stripper effluent gas and should have no effect
x
on either the long term stability of the carbon or the efficiency of NO recovery in the process.
x
(N20 concentrations in stripper off-gas samples were consistently less than 10 ppm.)
3. Denitration of nitrose solutions in the 1% HNS05 concentration range is in
excess of 99% in columns containing 22 in. of activated charcoal, with acid space velocities on
-1 -1
the order of 0.8 hr , gas space velocities of 65 hr , and l/G values of 8.5 mole/mole. These
figures are based on the experimental conditions used in the testing and do not indicate optimum
values, although they do suggest starting points for engineering scaleup.
54
-------
IV. HIGH TEMPERATURE SCRUBBER - LABORATORY EVALUATION
A. Background
Small scale laboratory experiments during the previous contract showed that oxides
of nitrogen could be absorbed in sulfuric acid from a nitrogen carrier gas stream at 250 ° F
with high recovery efficiencies. Replicated runs in a I 1/2-in.-diameter column packed with
Berl saddles to a depth of 24 in. demonstrated greater than 98% scrubbing efficiency at 255°F
at an L/G of about 1. 7 mole/mole. Preliminary evaluation of the high temperature scrubbing concept in
the 10 SCFM miniplant verified this absorption capability. Because of the need for scaleup
data which were not available from previous test runs, it was decided to construct a more soph-
isticated laboratory scrubber and investigate the characteristics of the high temperature
scrubber on a broader basis.
B. Equipment Design
A laboratory scale scrubber was designed based on the Corning flanged pipe system to
permit modular construction. Fig. 23 shows a schematic diagram of the scrubber which is
made of 2-in. Corning flanged pipe with feed lines consisting of I-in. pipe and transfer lines
made of 1/4-in. pipe. A photo of the apparatus is shown in Fig.24.
Acid was pumped through a I-in. -diameter section which was wrapped with heating
tapes capable of heating the acid to temperatures in excess of 250°F. The acid was pumped
with a peristaltic pump capable of delivering flows up to about 300 ml/min. The acid outlet
had a reservoir section to act as a gas trap with acid samples being taken periodically to
drain the system and to permit analysis of the acid. The column was heated with external
heating tapes and was packed with 40 in. of 1/4-in. Intalox saddles. There was one redistrib-
uter half-way up the column, and observation of fluid flow in the column showed good distrib-
ution of flow acrOss the column.
C. Operating Conditions
The column was operated by contacting a NOx-bearinggas with sulfuric acid in a counter-
current mode. The gas was prepared by mixing a nitrogen carrier gas with N203' the latter
being supplied as the trioxide in commercially available gas cylinders. The acid was the
same material used in the continuous column tests for activated charcoal analysis: 80%
sulfuric acid, commercial grade.
The flow rates used in the column were calculated based on standard theoretical assump-
tions. Given the L/G, the type and size of the packing, and the column diameter, it was possi.
ble to calculate the flooding velocity of the column. Making the assumption that the optimum
gas velocity is one-half the flooding velocity, the optimum gas and liquid flow rates could be
55
-------
Acid
heater
Acid
pump
Thermometers
404 in. of packing
1/4-in. intalox
saddles
Acid reservoir
Acid exit
Fig. 23. Laboratory scale high temperature absorption column
56
-------
.---- - -
~,. " I
.Jf'i .~
ir'~ ,
~ 1 p....,
r [a
,
-_. - - ---
"
r-
m
M.
:,3
"
f ~
.
.
~
".-:-;.
\ .
.~.~
Fig. 24.
Laboratory apparatus for continuous evaluation of high temperature
scrubber
57
-------
obtained. These flow rates were:
Nitrogen carrier gas 66,201 cc/min (2.35 SCFM) at STP
N203 232 cc/min (3500 ppm by volume) at STP
Sulfuric acid (80% H2S04) 205 cc/min (at 700 F)
These figures are based on L/G of 2.1 mole/mole with a column operating temperature of 2500 F.
It was intended that the L/G be varied as well as the basic gas flow rates. As will be seen in the
discussion of the results, the carrier gas flow was varied between 11,500 and 54,000 cc/min
and the L/G between 2.0 and 10.0 mole/mole.
D. Analytical Techniques
The inlet gas was analyzed by taking a sidestream from the gas feed and running it
through the infrared analysis cell. This is a batch technique and requires about 30 min to run,
10 min to pass the sample through the cell, 13 min to run the infrared spectrum and 5 to 10
min handling time. The analysis is based on calibration curves obtained previously and is con-
sidered to be quite accurate and reproducible. Error is less than 5%.
The outlet gas was analyzed by first freezing out the water vapor and then running the
gas through the infrared cell. The water was trapped by passing it through a saddle-filled U-
tube immersed in alcohol at - 45 0 C. Considering that trapping of the water might well remove
N02 from the gas stream, provisions were made for analyzing the NOx content of the trapped
solution. After each experiment, a quantity of 3% H202 was added to the trap which was then
sealed off to prevent gas leakage. The trap was warmed up to tha w out the frozen trapped
water and the U -tube shaken for several minutes to absorb the gaseous NO . The liquid was
x
decanted and the trap rinsed twice to collect the total dissolved NO . The peroxide solution
x
plus the washings were titrated with standard NaOH to measure the NO content. The results
x
invariably showed a negligible amount of NO absorbed in the water.
x
The outlet acid was analyzed by titration with KMnO 4 as previously discussed in connec-
tion with the stripper experiments. The nitrometer was not used.
The inlet acid was fresh material taken from the barrel and was considered free of NO
x
(see Table V for analysis of the NO content in fresh barrel acid). The acid was not recycled.
x
E. Experimental Results
A total of 20 test runs were made with the apparatus with the experimental conditions
and test results given in Tables VI and VII. The first significant point to be noted in the data is
the wide variation in the N02/NO ratio in the inlet gas, despite the fact that gases obtained
from both Matheson and Air Products were claimed to have a mole ratio of 1: 1. This erratic
gas concentration variation seemed independent of the technique used to withdraw the gas from
the cylinder. Nitrogen trioxide is made available as a liquid under pressure, and it is possible
to withdraw the material as either a gas or a liquid from the cylinder. In the first three runs
(Table VI), the N203 was removed as a liquid and completely volatized in the line, while in the
remaining tests the gas above the liquid was withdrawn.
Despite the problems in providing the proper gas mixture, there is much to be gained
from these experiments. The most obvious point is that the gas side efficiency is consistently
58
-------
Table VI. High Temperature Scrubber Experimental Results 2500 F Steady State Operation
2-in. x 40-in. Column Packed With 1/4-in. Porcelain Intalox Saddles
Experiment No. 1 2 3 4 5 6 7 8 9 10
Inlet N02' ppm 5000 4600 2000 4000 1500 3700 4200 1600 1000 1200
Concentration,
mm Hg* NO 2900 2900 1000 3800 2400 3300 2800 2100 2000 2700
NOx content, % by 7900 7500 3000 7800 3900 7000 7000 3700 3000 3900
volume
Total gas flow, i/ min 11.5 11.5 11.5 11.5 11.5 15.0 23.0 17.0 34.5 27.5
(STP)
Acid flow, 52 58 50 100 150 66 100 50 100 100
cc/min (70 OF) t
UG mole/mole 3.6 4.0 3.5 6.9 10.4 3.5 3.5 2.4 2.3 2.9
c.n Material balance 20 15.8 13.3 ~ 2.8 73.5 26.5 29.5 11.6 10.6
to error, %:j:
Gas side abso~tion 77.6 96.9 72.7 82.8 93.1 97.5 98.5 98.4 95.8 95.7
efficiency, % *
Acid side adsor~tion 57 81.1 61.3 ~ 65.0 24.0 72.0 67.0 84.2 84.0
efficiency, % **
Notes:
* As measured by infrared analysis.
t As measured by outlet acid flow at about 200 "F.
Inlet NO (gas) - Outlet NO (acid + gas)
:j: % Material balance error = x x = x 100.
Inlet NO (gas)
x
~ Material balance error negative, greater than 100% acid side absorption efficiency.
** Efficiency based on effluent gas and feed gas.
*** Efficiency based on effluent acid and feed gas.
-------
Table VII. High TemperatUre Scrubber Experimental Results 2500 F Steady State Operation
2-in. X 40-in. Column Packed With 1/4-in. Porcelain lntelox Saddles
Experiment No. 11 12 13 14 16 17 19 20
Inlet N02' ppm 13,200 2500 4200 5500 3300 1700 2000 2100
Inlet NO, ppm 5000 5600 3200 3000 4300 6600 6100 4100
Total Inlet 18,200 8100 7400 8500 7600 8300 8100 6200
NO , ppm
x
Total gas flow, i/ min 54.5 35.0 52.0 52.0 20.0 29.5 20.0 20.0
(STP)
Acid flow, 150 100 150 150 100 150 100 100
cc/ min (70 OF) ~
LlG mole/mole 2.2 2.3 2.3 2.3 4.0 4.0 4.0 4.0
Material balance 4.8 8.7 8.8 7.1 1.6 3.1 10.4 -6.1
0':> error, % ~
0
Gas side absorption 84.8 78.5 97.7 95.4 88.1 96.4 80.0 83.7
efficiency, % *
Acid side absorption 80.0 69.8 88.9 88.8 86.3 90.4 69.6 89.8
efficiency, % **
Notes:
t As measured by infrared analysis.
~ As measured by outlet acid flow at about 200 OP.
Inlet NO (gas) - Outlet NO (acid + gas)
~ % Material balance error == x x
Inlet NO (gas)
x
* Efficiency based on effluent gas and feed gas (infrared analysis).
** Efficiency based on effluent acid and feed gas.
x 100.
-------
high, indicating that the scrubber is doing an excellent job in removing the oxides of nitrogen
from the gas stream. Since both the inlet and outlet gas analyses are based on the same
calibration curves and are thus independent, these high removal efficiencies can be consid-
ered quite reliable. (It should be noted that these results closely reproduce the original exper-
iments described in the final report for Contract PH-86-68-75.)
The problem is, of course, that the acid side efficiencies do not agree with the gas
based figures and thus the material balances do not close. There are some general trends
which might help to explain the discrepancies. First of all one might expect these large mat-
erial balance errors because of the variations in the inlet gas NO/N02 ratio. Some of these
runs were made successively on the same day, yet there was still considerable variation in the
inlet gas composition. It is therefore reasonable to assume that there might have been some
change in the inlet gas composition from the time the inlet gas was analyzed to when the outlet
gas was analyzed.
If the N02 is reacting with the water in the 80% H2S04 to form HN03 in the outlet acid,
it would not have been detected by titration with KMnO 4 which only determines oxidizable spec-
ies. Nitric acid can only be determined by measuring total nitrogen oxides with the nitrometer
and subtracting the oxidizable nitrogen oxides obtained by KMnO 4 titration. The difference
must be fully oxidized NO;. This type of analysis was performed on a portion of a large com-
posite sample obtained from several of the first eight runs. It was found that the nitrometer
analysis of NOx was about 8% higher than the KMn04 titration value, which is a major portion
of the material balance error in most of the first eight runs. It is interesting to note that in
the remainder of the runs there was usually an excess of NO and the material balances closed
better. NO of course does not dissolve in or react with sulfuric acid and it might be expect-
ed that the material balance would be better in these cases.
The fact that there was an excess of NO in many of the runs leads to an interesting an-
alysis of the data. It is significant to note that, despite the imbalance in the N02/NO ratio,
the absorption efficiency in the scrubber remained consistently high. If the absorption was
assumed to take place through the mechanism of equimolar combination of NO and N02 with
H2S04' then a maximum absorption efficiency could be calculated for each run. For example,
in run no. 17, the entering gas contained 1.3 mm Hg of N02 and 5.0 mm Hg of NO so the max-
imum absorption should have been 1.3 mm Hg of N02 and 1.3 mm Hg of NO, or 2.6 mm Hg NOx
out of a total of 6.3 mm Hg of NO for an overall efficiency of 36%. In this case, the measured
x
overall scrubbing efficiency was 96.4%, far more than is theoretically possible.
This phenomenon is hard to explain until one examines the outlet gas concentration for
several of the experimental runs and observes that there is a considerable amount of N02
present even though it was NO that was present in excess in the inlet gas. If one considers that
there is a considerable delay (approximately 10 min) between gas sampling and the time when the
infrared spectrum is obtained, it seems quite likely that some oxygen has been introduced into
the gas between the delivery of the NOx from the N203 tank and the time when the gas is fed to the
infrared cell.
61
-------
If oxygen were present in the gas as it was fed to the scrubber, some of the NO would
have been oxidized to NO thus forming a more equimolar mixture of the two nitrogen oxides
2
and allowing the absorption of more NO than would have been possible based on the original
x
feed gas analysis. If rehtively high absorptions were assumed in the scrubber based on the
combination of equimolar amounts of NO and N02' then there should be low concentrations of
N02 in the outlet gas. The fact that there was consistently large amounts of N02 compared to
the NO in the outlet gas suggests that the NO continued to oxidize due to the presence of oxygen
and by the time the gas was analyzed much of the NO had been oxidized.
To check this hypothesis, several runs were made with a feed gas consisting of nitrogen
and NO with no N02 be intentionally fed to the system. NO was obtained from a high pressure
cylinder (Matheson gas) with the contents analyzing by I.R. about 2 mm Hg «0.3%) of N02
and the remainder NO (> 99%). Any appreciable N02 detected in the operation of the scrubber
must necessarily have been produced by NO oxidation.
The scrubber was run under conditions similar to those in previous experiments except
that both high and low temperatures were used. The resulting data are shown in Table VIII.
The fact that the feed gas showed NO to N02 mole ratios similar to those experiments where
N203 was used as the feed gas lends strong support to the idea that oxygen is being introduced
to the system and is causing NO oxidation. The fact that the scrubber absorption efficiencies
were quite low suggests that most of the oxidation is occurring in the infrared analysis cell
after the gas passed through the scrubber. but there was still enough oxidation in the scrubber
to cause up to 20% of the NO to be absorbed. The significant point is that a gas which analyzes
x
more than 99.5% NO appears to have 30-50% N02 when it is fed to the scrubber system. Clear-
ly some oxygen is entering the system and causing oxidation of the NO.
There is no apparent source of oxygen in the system, but only very small quantities of
the gas would be necessary to explain the phenomena observed. It is possible that the gas
moving at 40 ft/sec through the transfer lines in the scrubber feed system could be aspirating
some air into the lines. Unfortunately, at the present time, it is impossible to analyze for
low concentrations of oxygen, but a careful study is being made to determine possible sources
of unwanted oxygen.
One might be tempted to conclude that the problem in the experiments discussed above
is due to large errors in gas analysis. This would be reasonable except for the fact that the
low material balance errors (5-10%) provide strong evidence that the gas analysis on the basis
of total NO is correct. The material balance is obtained by comparing the amount of NO
x x
coming into the system with the feed gas (infrared gas analysis) with the total amount of NO
x
leaving the system with the outlet gas (infrared gas analysis) and the outlet acid (analysis by
titration with potassium permanganate). Thus, the low material balance errors are the result
of two completely independent analytical techniques which suggests a high degree of precision
in the measurements.
62
-------
Table VIII. Operation of Scrubber Using Pure NO as
Feed Steady State Operation*
Run 27 Run 28
Inlet N02' ppm 2500 4000
Inlet NO, ppm 6100 4600
Outlet N02' ppm 3400 900
Outlet NO, ppm 3400 6300
Operating temperature, Of' 75 250
Material balance error, % t 6.9 3.3
Gas side adsorption efficiency, % t 19.1 15.7
Notes
* Acid flow = 100 cc/min, total gas flow = 20.0 ~ /min, I)G = 4.0 mole/mole
Inlet NO (gas) - Outlet NO (gas + acid)
t % Material Balance Error = x x x 100
Inlet NO (gas)
x
t Efficiency based on effluent gas and feed gas (infrared analysis).
63
-------
The conclusion that can be dra wn is that the gas analysis is correct in giving the total
amount of NO entering and leaving the scrubber system, but it may very well be incorrect in
x
giving the instantaneous proportions of NO and N02' To get this type of analysis, it is necessary
to have rapid continuous monitoring equipment analyzing the gas streams. Continuous analysis
is impossible with the infrared batch cell, but the new infrared (Beckman) and ultraviolet
(DuPont) equipment make this type of inline analysis a routine task.
This discussion of the possible reasons for the variety of phenomena observed in the
high temperature scrubber experiments should not be allowed to confuse the basic results of
the testing. The high temperature scrubber does work at high efficiencies and the analytical
data that verifies this conclusion are consistent and reproducible with material balance closure
within 5-10%.
F. Summary of Results of High Temperature Scrubber Evaluation
Despite some experimental anomalies, it was shown reproducibly that mixtures of NO
and N02 can be scrubbed with 80% sulfuric acid at high temperat'Jres with high efficiency. The
data obtained tend to reproduce and verify data obtained in earlier tests during the prior con-
tract.
There is some evidence that it may not be necessary to have a precise 1: 1 ratio of the
two gases in order to obtain high absorption efficiencies at high temperatures. Continued work
in this area would be necessary to determine any possible unforeseen absorption mechanism.
This would be demonstrated in miniplant operation using gases more closely approximating flue
gases.
64
-------
V. VAPOR PRESSURE OF NITROGEN OXIDES OVER HNSO /H SO SOLUTIONS
524
Although the vapor pressure of N203 over nitrose solutions has been investigated in
the literature, considerable spread in data had been noted prior to this effort and no specific
data were available in the high temperature regions critical to the Tyco process. The purpose
of this subtask was to:
1. Review and evaluate the literature data for accuracy and reliability of
method
2. Experimentally verify, if necessary, the available vapor pressure data of
nitrogen oxides above HNS05/H2S04 solutions in the area of H2S04 strengths
from 80 to 90% and temperatures of 250 to 350 0 F.
A. Literature Search
M . 1 5-20 1 h . . f. 1 . f d.
any artlc es re evant to t e cOmpOsitIOn 0 mtrose so utlons were oun 10 our
literature search. Most of the articles were not key source papers for actual vapor pressure
data, but were merely supportive of various mechanisms outlined here and are footnoted at
the appropriate locations in the text.
Key articles for vapor pressure data were those of A. N. Tseitlin and V. T. Yavorskii20
5
and E. Berland H. H. Saenger:
1. Tseitlin and Yavorskii
The subject reference20 describes a series of experiments in which the authors
attempted to determine the partial pressure of nitrogen oxides over nitrose solutions in the
original range of 56 to 84 % H2S04 with up to 33.6 wt % HNS05 added at temperatures of 122
to 194 of by a dynamic gas flow method. Three thermostated sample vessels containing
identical nitrose solutions prepared from known equimolar quantities of NO and N02 were
equilibrated with a flowing stream of nitrogen. The flowing stream of nitrogen was scrubbed
with concentrated sulfuric acid scrubbers which were assumed to remove the nitrogen oxides
from the nitrogen gas stream. The scrubbers were then analyzed for nitrose content by a
method not mentioned in the paper. The partial pressure of nitrogen oxides over the nitrose
solutions were calculated from the calculated nitrose concentration of the scrubbers and k""
nitrogen flow rates.
65
-------
Several conceptual and experimental flaws are evident in this work. In the first place,
the authors neglect to mention the method of analysis of the scrubbers. Thus, one does not
know whether a strong oxidant was used to titrate the scrubber contents or if a strong reducing
agent such as mercury was used to convert the supposed HNS05 to NO. Secondly, the authors
incorrectly assumed that concentrated sulfuric acid may be used to remove nitrogen oxides
from a gas stream and dissolve them as HNS05' As shown later in this paper, at high sul-
furic acid concentrations, HNS05 disproportionates to give N02 + and NO. Thirdly, the authors
prepared their solutions at constant (NO + N02) weight fractions and then attempted to derive
an empirical equation relating all of their vapor pressure data. Since NO + NOZ dissolves in
H2S04 to form HNS05 and H20, they would have achieved a better empirical correlation if
they had used constant HNS05 concentrations at various H2S04 concentrations. This is a re-
sult of the fact that the water content of nitrose solutions does not change if the solutions are
made up with HNS05' but the water content of nitrose solutions increases for solutions made
up with NO + NOZ' Fourthly, that the authors technique was inadequate to determine partial
pressure of gases in equilibrium with nitrose solutions is confirmed by their incorrect values
for the partial pressure of water. 19 Especially noteworthy is their value of 2.22 torr for the
partial pressure of water above 81.25 wt % HzS04 at 50 °C which they obtained, especially since
other authors' data 1,2, 3 lead one to expect a value less than 0.6 torr,
2. Bed and Saenger
These authors5 used a batch method for obtaining vapor pressure data over
nitrose solutions, similar to that which will be described in subsequent Tyco experimentation.
Within their range of study, their data appear to be precise (0.1 mm Hg) internally consistent
and consistent with other literatures6, 7, 11, 12,21 which have tended to be less precise. The
vapor pressure data of Bed and Saenger can be treated to obtain the partial pressure of nitro-
gen oxides over nitrose solutions with acid concentrations in the range 64 to 80 wt % HzS04'
The total pressure above nitrose solutions containing 0 to 0.5M HNSO was determined for the
5
temperature range 30 to 150 °C for eight concentrations of HzS04 in the range 64 to 80 wt %
H2S04'
Making the reasonable assumption that the partial pressure of nitrogen oxides is equal
to the total observed pressure at a particular HNS05 concentration minus the total observed
pressure at 0 HNS05 concentration and the same HzS04 concentration and temperature, one
may obtain the partial pressure of nitrogen oxides over nitrose solutions from their data. In
addition, one may conclude that:
1. Henry's law may be used to describe the data. Thus, in Fig. 25, a Henry's
law constant of 410 is obtained and the data correlate well with the calculated line.
2. One may extrapolate to higher temperatures by applying the relation log
P = A + BIT. Thus, in Fig. 26, a plot of log P versus reciprocal temperature is seen to be
linear with A = 5.286 and B = -1835.
66
-------
5
Berl and Saenger
60
70.0 wt % H2S04
1300 C
40
PNO = H . CHNSO
x 5
H = 410
b()
::r::
E
E
:<
o
Z
0...
20
.05
.10
.15
C
HNS05' mole HNS05/ £
Fig: 25. Nitrogen oxide partial pressure, nitrosylsulfuric acid
concentration dependence
67
-------
5
Ber! and Saenger
40
70.0 wt % H2S04
O.lM HNS05
log P = 5.286 1835 /T
bG 20
:r:
6
6
><:
0
Z
0..
10
8
6
2.4 2.5 2.6 2.7 2.8 2.9
1000/T 0 K
Fig. 26. Nitrogen oxide partial pressure, temperature
dependence
68
-------
3. At any given temperature. and HNS05 concentration, the logarithm of the
partial pressure of nitrogen oxides above nitrose solutions is a linear function of the ~S04
wt %. For example, Fig. 27 verifies this for 0.1M HNS05 at 130 °e. Note that the partial
pressure of nitrogen oxides above nitrose decreases as the sulfuric acid concentration in-
creases.
B. Laboratory Work at Tyco
Work on the determination of the vapor pressure of nitrogen oxides over nitrosylsul-
furic acid solutions entered the experimental stage after reexamination of the literature. The
literature work supported the conclusion that there was no existing data available for nitro-
sylsulfuric acid solutions for acid concentrations above 80 wt % sulfuric acid or temperatures
in the 250 to 350 of range, although reasonable extrapolations might be made from Berl and
Saenger at conditions close to their area of study.
1. Procedure
The experimental apparatus used at Tyco was similar to that of Berl and
Saenger and is shown in Fig. 28. The solution to be examined was put into flask A, which was
then cooled to about -20°c' The entire system was then evacuated to a pressure of Imm Hg.
The flask was then tipped so that the acid manometer was partly filled with the acid solution
thus forming a liquid seal between the flask A and the mercury manorneter. The flasks and
acid manometer were then put into an oven and connected (through a hole in the wall of the
oven) to the rest of the system.
To measure the total vapor pressure of the system, the oven was slowly heated to a
given temperature and held there. As the acid warmed and a pressure was created on the flask
side of the acid manometer, the acid manometer became unbalanced. To offset this imbalance,
argon was bled into the system on the other side of the acid manometer. The pressure of the
mixed system outside the oven was measured with the mercury manometer. Thus, the pres-
sure reading on the mercury manometer was equal to the total vapor pressure of the system
in flask A. A series of readings could be obtained by increasing the temperature of the oven,
holding it at each desired level until equilibrium was obtained. Equilibrium was assumed to
have been attained when, for any given temperature, the mercury manometer pressure reading
was constant for at least 15 min. One complete run gave the temperature dependence of the
total vapor pressure of a solution containing a certain concentration of nitrosylsulfuric acid
in sulfuric acid of a specific concentration.
2. Experimental results
The experimental data obtained at Tyco and some derived quantities are given
in Tables IX through XIII and in Figs. 29 through 31. Table IX is a listing of data regarding
the composition of the four nitrose solutions which we investigated. The strength of the sul-
furic acid is expressed as the weight percent of ~S04 in the solution prior to the addition of
69
-------
100
bO
:I:
E
E 10
~
)<
0
Z
0..
5
5
Berl and Saenger
0.1 mole HNSOS/£
T = 1300 C = 2660 F
50
Acid Concentration,
Fig. 27. Nitrogen oxide partial pressure, sulfuric acid
concentration dependence
70
-------
I
I
I
Oven ambient
Argon
-
Acid
manometer
Mercury
manometer
Fig. 28. N203 vapor pressure analysis apparatus
71
- Vacuum
-------
Table IX. Tyco Sample Compositions
Experimental Data Derived Quantities
-"- -- ~ ~ -----....
HNS05 Molarity Calculated +
Specific O.IN KMnOil0 ml NO/ml sample, Assuming Oxida- Oxidation Calculated HNS05' Calculated N02 '
-.1 Gravity sample, ml ml tion State," 3 State mol/l mol/l
t\)
1. 74 18.8 0.093
1.77 19.2 2.4 @ 250C 0.096 3.04 :t 0.1 0.096 0.002
1.815 18.5 0.096
1.84 18.8 2.7 @ 240C 0.094 3.31 :t 0.08 0.094 0.017
-------
Table X. Tyco Vapor Pressure Data for 80.08 wt % H2S04 *
Temperature,
°C
Observed Pressure,
torr
Calculated Pressure (torr),
Corrected to 0.1 mol/£ HNSO
5
61.7
76.7
98.9
115.6
132.2
144.4
165.6
3
7
16
35
69
113
184
3
7
16
36
71
116
190
* 0.097 mole HNS051 £ .
Table XL Tyco Vapor Pressure Data for 84.04 wt % H2S04 *
Temperature,
°C
Observed Pressure,
torr
Calculated Pressure (torr),
0.1 moll!
70
85
115.5
133
155
163
2
6
20
38
93
145
2
6
20
39
95
148
* 0.098 mole HNS05/!.
73
-------
Table XIL Tyco Vapor Pressure Data for 90.16 wt % ~S04 *
Temperature,
°C
Observed Pressure,
torr
Calculated Pressure (torr),
O. 1 mol/ I
57
82
102
127
144
162
1
3
6
20
35
60
1
3
6
21
36
63
* 0.096 mole HNS05/ I .
Table XIIL Tyco Vapor Pressure Data for 95.0 wt % H2S04 *
Temperature,
°C
Observed Pressure,
torr
Calculated Pressure (torr),
0.1 mol/I
70
86
120
131
141
158
166
1
2
8
13
18
25
39
1
2
9
14
19
27
41
* 0.094 mole HNS05/ 1.
74
-------
1000
Temperature, 0 C
150 130
,
302 ,266
, '" ,
'...... '" "
....... " ,
" , "
........ "" 64%
" ......... "67%'
, "~%"',
'Z5.7% , " "
" ......... " ,
......... ....... .......
, .......
, .........
, .........
, ,
"'-
.......
.......
.......
100
b/)
::r:
E
E
212
OF
,
"
"
"
"-
........
,
,
,
"
,
........
,
,
"
~
......
C'iI
~
B
0..
Q)
~
:::I
UJ
UJ
Q)
ct
......
5
~
10
---
5
Ber! and Saenger data
Tyco experimental data
2.4
2.5
2.6
1000/T 0 K
2.7
2.8
Fig. 29. Total vapor pressure over nitrose solutions
containing O. 1M HNS05
75
-------
~ ><
o
(l.,Z
"
"-
....
of "
"-
......
,
,
',67%
,
........
Temperawre, 0 C
100
130
100
212
bO
:r:
S
S
,
,
"
.......
"- ,95%
"-
,
"-
.......
.......
,
,76%
,
,
,
,
"-
"
"-
"
,
2.8
150
302
,
"-
......
,
......
......
......
......
......
,
,
........
........
,
"
266
10
,
,
,
"
"
,
" 80%
"
"
"-
"
"-
--- Berl and Saenger5 ""
data '"
Tyco experimental data'" "-
"
"
, 70%
,
.......
,
........
,
2.4
2.5
Fig. 30. Nitrogen oxide partial pressures over nitrose
solutions containing O.lM HNS05
76
-------
~
o
Z
0...
80
Concentration, wt % H2S04
10
0.1M ,HNS05
1300 C = 266 0 F
5
o Her 1 and Saenger data
. Tyco data
0.2
70
.
~
\
\
\
\
\ .
ExtrapolatIOn
\
\
\
\
\
\
90
Fig. 31. Nitrogen oxide partial pressure over nitrose solutions.
Concentration dependence - experimental values
77
.
-------
the HNSOS. The original analytical data for HNSOS analysis are given in columns 3 and 4 so
that future investigators may have access to experimental quantities rather than derived
functions. The later columns refer to derived results.
If a sample is known to contain all dissolved nitrogen as HNSOS' the concentration
of the HNSOS may be determined by permanganate titration or with a nitrometer. In the case
of the permanganate titration, the HNSOS is oxidized to HN03 which is an oxidation state
change of two. Thus, for each two equivalents of KMn04 required to titrate the sample, there
is 1 mol of HNSOS present. In the case of the nitrometer, all of the HNSOS is reduced to NO
and one measures the volume of NO formed to determine the total number of moles of HNSOS
present. The permanganate titration is more precise and faster than the nitrometer determina-
tion and is thus normally used to the exclusion of the nitrometer.
If one has reason to suspect that the nitrogen may be present as other than HNS05, it is
. 8 9 10 11 12 14 15 18 21
necessary to perform both analyses. We know from the lIterature" , , , , , ,
that, at acid concentrations less than 65 wt % ~S04' the HNS05 disproportionates to form
HN03 and NO thereby causing the mean oxidation state of the dissolved nitrogen species to be
greater than three and less than five. The nitronium ion, N02 ~ is a well characterized
oxidizing agent (nitrogen oxidation state of 5) which is stable in concentrated sulfuric acid9, 10
exceeding 84 wt % ~S04. We thus had reason to expect that at sufficiently high concentrations
of sulfuric acid, the HNS05 might disproportionate to N02 + and NO.
The oxidation state of the dissolved nitrogen species is calculated as follows: (1) the
permanganate titration is performed to oxidize all of the nitrogen species to the +5 oxidation
state thereby giving the number of equivalents required to do this, and (2) the nitrometer is
used to reduce all of the nitrogen species to NO, thereby giving the total number of moles of
nitrogen species. Dividing the number of equivalents by the number of moles gives the oxida-
tion state change experienced during the permanganate oxidation. Subtracting this from five
gives the mean oxidation state of the dissolved species. As an example, consider the 95.0 wt
% H2S04 sample. In this case, 18.8 ml of 0.100N KMn04 were required to titrate 10 ml of
sample. This is equivalent to 0.188 equivalents of KMn04 per liter of sample. In the nitro-
meter run, 2.7 ml of NO were produced from a 1 ml sample at 24 °C at which temperature the
molar volume of an ideal gas is 24.37 £ / mol. This corresponds to 0.1108 mol of NO per liter
of sample. Dividing 0.187 equivalents/ £ by 0.1108 moll £, one obtains an oxidation state change
of 1.69 :f: 0.08 equivalents/mol. Subtracting this from 5.00 one obtains an oxidation state of
3.31 :f: 0.08. This, in turn, leads to calculated concentrations of 0.094M HNSOS and 0.017M
N02 + at 24 °C. Since we are postulating that the N02 + is formed by disproportionation of
HNSOS' viz:
+ -
3 HNS05 -. 2 NO + N02 + HS04 + ~S207'
+
we know that 2 mol of NO are formed per mole of N02 ' and thus in this case 0.034 mol of NO
78
-------
went into the gas phase for each liter of solution at 24°C. We will come back to this conclu-
sion in our subsequent interpretation of data (see Section V. C).
Tables X through XIII list the experimental vapor pressure data. The total observed
pressure data, corrected to 0.100M HNS05 using Henry's law, are plotted in Fig. 29 both for
the Berl and Saenger and the Tyco data. The Tyco 80.08 wt % ~S04 data fall close to but
slightly below the Berl and Saenger 80.02 wt % H2S04 data indicating that we are in substantial
agreement. Due probably to the large volume (500 m!) of our solution and the fact that we did
not stir this solution, we believe that our observed pressure values may be slightly less than
the equilibrium values. Nevertheless our values are close enough to those of Berl and Saenger5
to indicate agreement. We have thus confirmed the reliability of the Berl and Saenger5 method.
C. Interpretation of Tyco Results
Fig. 29 demonstrates that the curves for our higher acid concentrations lie progres-
sively lower than the 80% curve, as expected. For our 90.16 wt % ~S04 and 95.0 wt % H2S04
data, the observed total pressures are substantially higher than the approximate values of
water vapor pressure which may be obtained from the data of Greenewalt.7 Thus, since the
correction term for water vapor pressure is small, it is possible to calculate the partial
pressure of nitrogen oxides with a low percentage error. Note that in Fig. 30 the partial pres-
sure of nitrogen oxides for 90.16 and 95.0 wt % ~S04 are higher than for 80.02 wt % ~S04'
This is in contrast to the trend noted by Berl and Saenger at lower concentrations of ~S04 for
which the partial pressure of nitrogen oxides was observed to decrease steadily as the sul-
furic acid concentrations increased. We contend that this is due principally to the formation
of NO by disproportionation of HNS05 as outlined earlier. Since the samples were degassed
prior to vapor pressure measurements by cycle of pumping with a mechanical vacuum pump
while frozen at dry ice-acetone temperatures followed by thawing, all of the NO formed at
room temperature was lost and the high observed pressures at higher temperatures must have
been due to a positive temperature coefficient for the disproportion at ion reaction.
Fig. 31 is a plot of partial pressure of nitrogen oxides above nitrose solutions (O.IM
HNS05) at 130°C versus H2S04 concentration. Note that data for the 64 to 80 wt % ~S04
if extrapolated to higher acid concentrations indicate that the partial pressure of nitrogen
oxides would be quite low at all acid concentrations above 80%. However, our data for 90 and
95% acid indicate that the disproportionation of HNS05 occurs at these concentrations. It is
important to determine just where in the 80 to 90% range this disproportionation reaction be-
comes important. At acid concentrations below 84%, the ~O/~S04 mol ratio is greater than
one and the activity of water (which could act to destabilize N02 +) is much higher than at acid
concentrations higher than 84%.6 We thus have reason to believe that, since HNS05 dispro-
portionation depends in part on stabilization of N02 +, this disproportionation reaction will be-
come important when the sulfuric acid concentration exceeds 84%.
79
-------
VI. LABORATORY SCALE MINIPLANT - MODIFICATION AND EXPERIMENTATION
A. Goals of Miniplant Operation
1 . Conceptual goals
The research work during the 2 1/2 years preceding the operation of the mini-
plant had been successfully devoted to the conceptual and experimental development of a pro-
cess which could remove the oxides of sulfur and nitrogen from power plant flue gas. The
resultant process includes major modifications of the century-old process on which it is
based and these new process stages require extensive experimental characterization and ver-
ification before large scale operations can be attempted.
Experimental work on the separate process stages during the latter phases of the pre-
vious contract and the early stages of this contract gave ample evidence that the process was
very promising, but there was the definite need to demonstrate that the individual stages
could be put together and made to work as an integrated unit. Thus, one of the major goals
of the miniplant operation was to provide a more comprehensive experimental verification of
the viability of the Tyco Catalytic Chamber Process.
2. Scale-up goals
After experimental verification of the process concept, the long range goal is
to design, construct and operate a full scale plant in conjunction with an industrial power
plant. In working toward this goal, it will be necessary to scale up in stages, with each stage
being used to develop engineering data to permit design of the next size plant.
Toward this end, the miniplant is to be used to determine a range of operating charac-
teristics of the maj or process stages to permit scale-up to pilot plant level. Thus, by the
time the miniplant evaluation period is completed, it should be possible to design a pilot plant
that will permit the optimization of the process so that full scale plants can be built.
3. Evaluation of the major process stages
Although there are a great many aspects of the process that must be optimized,
the work can be broken down into the characterization of the three main process units: (1)
the reactor, (2) the scrubber, and (3) the stripper. Proper optimization of these stages will
require a complete parametric study of the process operation, and this will have to be done
before the industrial scale pilot plant is designed and built. The goals of this contract were
limited to the verification of the feasibility of the important process stages and the determin-
ation of the order of magnitude of some of the key operating variables. Some of the signif-
icant variables examined in this manner were:
81
-------
ReactOr:
Operating temperature range
Residence time
Scrubbing efficiency
Number of equivalent scrubbing plates
Height of a theoretical plate
Operating temperature range
Stripping efficiency
Residence time
Bed depth
Scrubber
Stripper
B. Miniplant Modification
1. Background
At the start of this contract, the miniplant was in operational condition for
evaluating a process concept that preceeded the development of the catalytic stripper. The
plant consisted of a reactOr, a low temperature scrubber, a hot air stripper (packed with
ceramic saddles) and an acid concentratOr, With the change in process concept, the acid
concentratOr was no longer needed and the scrubber and stripper had to be extensively modified.
Much of the equipment needed for the required modifications was made available by salvaging
parts of the old equipment, but some new items were needed. Details of this are given in the
following sections of this report,
2. Flowsheet
The flowsheet for the modified miniplant is shown in Fig. 32, The plant is set
up so that each stage could be isolated and the effluent stream either analyzed or dumped.
There is a bypass to the stack between the reactor and the scrubber so that the flue gas could
be contacted with oxides of nitrogen and brought to steady state before being fed to the scrubber.
This avoids the problem of diluting or concentrating the acid while the scrubber is being heated
to its operating temperature. Most of the bypass lines were equipped with condensers to avoid
venting wet gas streams, which would cause acid condensation in the vent fan. The flowsheet
also shows the location of the sample pOrts leading to the analytical system which will be dis-
cussed in detail below.
3. Reactor
The S02 oxidation reactOr was not altered from its earlier design since its
function had not changed. It consisted of about 57 in. of 4-in. -diameter pipe (volume: 0.414
ft 3) with a thermometer well mounted in the middle. (See Fig. 33>. Originally the reactor
had two heating mantles on it, each 12 in. long. To these were added heating tapes so that the
entire length of the reactOr was heated. The heating tapes and exposed portions of the tube
were insulated with aluminum -foil backed fiberglass insulation. At various times, different
baffles were placed between the flanges at the joints to create turbulence to enhance mixing
and the chamber was so designed that these could be replaced or removed very quickly without
shutting down the operation of the plant. The residence time of the gas in the reactor at a
flow rate of 10 SCFM (14.4 ACFM at 250°F) is 1.7 sec.
82
-------
To stae k
By-pass
to stack
CONDENSER
To stack
CONDENSER
Gas sample line
Gas sample line
Gas sample line
Water drain
CONTROL VALVE
FLOWMETER
MANOMETER
By-pass
to stack
S02
Manometer
HIGH
TEMPERA TIJR
N203
SCRUBBER
Control
valve
By-pass
to tack
BLOWER FLOWMET
(XI
~ FLOW TER
Water FLOWMETER
drain
Air NO
GAS Gas sample line
BURNER
Compressed air
Air
atural gas For sa mple 0
drain
Fig. 32. Miniplam flowsheet
-------
-.I
Fig. 33. Miniplant showing S02 reactor (above) and NO oxidizer. (This photo
was taken before the heating tapes and insulation were put on.)
84
-------
4. High temperature scrubber
The high temperature scrubber and the original low temperature scrubber had
essentially the same purpose, i. e., to remove the sulfuric acid mist from the gas stream and
to recover the oxides of nitrogen. No modifications were made to the 4-in. -diameter packed
scrubber (Fig. 34) other than to equip the column with heaters so it could be operated at the
desired temperatures.
For heating, the column was provided with four heating mantles and several heating
tapes with the entire length being insulated. The heating elements were ganged into two units
so that temperature control could be accomplished by two-zone heating. A thermometer well
was located half way up the column. In addition, there were thermometer wells located so
that the temperature of all inlet and outlet streams could be monitored.
The packing remained as 3j8-in. ceramic lntalox saddles packed to a height of 8.5 ft.
Several redistributors were placed in the column to prevent channeling through the column.
A manometer was installed so that the pressure drop across the scrubber could be determined
and isolated from the rest of the gas loop (see Fig. 32).
A reservoir was provided at the base of the column so that there would be a constant
source of liquid for the stripper feed pump. An acid sample port was provided before the
reservoir.
5. Catalytic stripper
The stripper column (4-in. diameter by 6 ft. high) consisted of four beds, each
containing 2 in. of saddles supporting 8 in. of Witco type 256 activated carbon. (See Fig, 34).
The saddles w~re used to prevent the relatively fine carbon particles from falling through the
holes in the glass support plate. One redistributor was installed half way down the column.
There was an acid reservoir and sampling port at the bottom of the column. The column was
also equipped with a manometer to determine the pressure drop.
The column was equipped with heating mantles and heating tapes so that the stripper
could be heated to temperatures up to about 300°F. Three thermometer wells were installed
so that the temperature could be measured at various heights in the column. Again, there
were provisions for measuring the temperature of all inlet and outlet streams.
The stripper was fed with acid from the bottom of the scrubber and air from a com-
pressed air line. Both lines had flowmeters so that flow rates could be monitored. The eff-
luent acid was pumped back to the scrubber while the effluent gas was either recycled to the
reactor or vented to the stack. Samples could be obtained of all input and output streams.
6. Nitrogen oxide feed
At startup, the N02 needed for feed to the S02 oxidation reactor is provided by
air oxidation of bottled nitric oxide. As shown in Fig. 32, the NO and air (from a compressed
air source) are fed to a pipe reactor which allows about 45 sec residence time at a flow rate
of about 0.3 SCFM. This should be sufficient to provide 90% oxidation of NO to N02 under condi-
tions necessary to provide 7000 ppm of N02 to a 10 SCFM gas stream.
If the miniplant is running at steady state and the stripper is providing sufficient NO
x
for recycle, the bottled NO is turned off and the stripper effluent gas is piped into the S02
reactor through the NO oxidizer. The NOoxidation reactor was wrapped with heating
85
-------
Fig. 34.
Miniplam showing stripper (left) and scrubber columns (this photo
was taken before the heating m'tmles, heating tapes and insulation
were put on)
86
-------
tapes so that the recycled gases from the stripper could be kept hot to avoid water conden-
sation, When used as a chamber for NO oxidation, the heaters are turned off since this
oxidation reactions Occurs with higher rates at low temperatures.
7. Fluid flows - pumps and blower
The miniplant was originally designed for a flue gas input flow of 10 SCFM.
All other flow rates were calculated based on the gas flow rate, using theoretical consid-
erations and laboratory data to size fluid handling equipment,
The acid flow rate in the scrubber/stripper loop was based on an L/G ratio of 2.5 mole/mole
in the scrubber as determined in the previous contract. Recommendations by TV A and Tyco Lab
experiments revealed a scrubber L/G ratio of about 4.6 mole/mole to be more feasible. The
original pumps had a maximum flow rate of 0.2 gpm, with 0.39 gpm required for an L/G of
4.6 mole/mole. In view of delivery time limitations on the high temperature corrosion resist-
ant positive displacement pumps, it was decided to perform the miniplant scrubber experi-
ments at maximum pumping rates and modify gas flow rates accordingly for scrubber operation
("'5 SCFM). Three pumps were available for the scrubber/ stripper loop with two for continuous
operation, and one as a spare.
The laboratory scale stripper utilized an L/G of 16 lb/lb in most of the experimen-
tation, and it was decided to use this order of magnitude in the miniplant stripper. With 80%
acid being fed at 0.14 gpm, compressed air feed of Z SCFM was used giving an L/G of approx-
imately 13.
C. Analytical Instrumentation and Sampling Procedures
1. Background
During the previous contract (No. PH 86 -68-75), gas analysis had been performed
using two spectrophotometric methods: (1) infrared and (2) ultraviolet photometric analysis.
The ultraviolet method was quite satisfactory for analysis of N02 and S02' butthe particular set-
up was not very sensitive and it could not be used to analyze for NO. The infrared system was more
sensitive and could be used for NO; N02 and S02 analysis, but it could only be operated on a
batch sampling basis which made it impossible to do NO and N02 balances since by the time
the batch test was run, the NO would have oxidized to N02' Water interference in the IR,
while overcome by scrubbing or cold traps, limits the flexibility of this method.
Z. Inline continuous analytical monitoring
For the continuous miniplant operation, it was necessary to obtain suitable
instrumentation that could be used on an inline, continuous basis. After extensive discussions
and examination of industrially available instrumentation, Tyco decided to purchase two in-
struments which appeared to be well suited to the task at hand. These instruments are:
1. DuPont Model 400 Ultraviolet Photometer: This instrument has been
used by TV A for analysis of SOZ in laboratOry and pilot plant applications. The model pur-
chased by Tyco is capable of monitOring SOZ and NOZ sequentially using the attenuation of
ultraviolet radiation at two specific wave lengths. The maj or advantage of this instrument
over other commercially available analyzers is that the sample can be maintained at very
87
-------
high temperatures which is very important when working with a gas stream that contains very
high concentrations of water which would condense out at the operating range of many of the
other available instruments (usually around 125 OF maximum).
2. Beckman Model 315A Infrared Analyzer: This device is used to measure
the concentration of NO in the sample gas streams. As with all infrared photometers, it is
very sensitive to the water content of the gas stream since the infrared absorption band for
water overlaps the absorption band for NO. For this reason, the gas stream was first passed
through the DuPont analyzer and then through a condenser trap so that the water could be re-
moved before sending the gas into the lR analyzer chamber.
Both instruments were equipped with dual sensitivity ranges so that both high and low
gas concentrations could be measured. The concentration ranges of the two instruments are:
o 10, 000 ppm (by volume)
o 1,000
o 5,000
o 500
o - 5,000
0-1,000
A photograph of the instruments is shown in Fig. 35.
3. Miniplant gas sampling
N02
S02
NO
It was necessary to analyze the gas streams at four points in the system:
1. Inlet gas after the introduction of SOZ and recycle NOZ
2. Reacted gas just after the reaction chamber
3, Stack gas just after the scrubber
4. Recycle N02 stream just after the stripper.
Gas samples were removed from the system at these four points and conducted to the
analytical instrumentation through 1j4-in. glass pipe which was heated to prevent condensation.
Fig. 36 shows a schematic diagram of the gas sampling and instrumental analysis system.
The procedure for gas analysis is to flow the gas from the desired sample port through
the sample lines into the DuPont analyzer and obtain sequential N02 and SOZ concentration
values, while simultaneously reading the NO concentration obtained on the Beckman analyzer.
Then a sample would be taken from another sample port and the procedure repeated. This
procedure obviously depends on being able to equilibrate the instruments for a particular gas
sample in a reasonably short time. In practice, rapid scanning of the various sample ports
was very difficult to achieve, as will be discussed later in the section on experimental results.
D. Experimental Results
1. Plant start-up
The start-up of the miniplant actually was accomplished in two phases: (1) in-
strument calibration and evaluation, and (2) miniplant operation. (See Appendix III for startup
procedures.) The first phase, which was carried out at the same time as the overall performance
evaluation, lasted about 1 week and permitted operating personnel to become acquainted with the
characteristics of both the instruments and the plant equipment. Operation of the miniplant con-
tinued for the next 2 -week period.
88
-------
.. i1!J¥'~ .;.,.,<
-
-.1.:
...
, ;
..
-1 A."
Fig. 35. Miniplant analytical instrumentation
89
I
~~
'f,...
(,
~
i - J
,...
d.-
-------
~------1
AmP:ifier I
Recorder ~ ----- -----.
I A mPl:fier
To stack
u. V. Photometer
N02 and S02
Refrigerator
condenser
I.R. Photome~er
(NO)
Flowmeter
To stack
CJ:>
o
High Low High Low High Low
cone cone cone cone cone cone N2 From From From From
S02 S02 N02 N02 NO NO sample sample sample sample
point point point point
1 2 3 4
Fig. 36. Gas sampling and analysis system
-------
2. Instrument evaluation
Electronic calibration of the DuPont and Beckman spectrophotometers was
quite straightforward and was accomplished in a short period of time. However, due to the
variety of gas sample sources, calibration and use of the instruments to analyze gas streams
was more involved and took more time to get used to,
A major consideration in operating the instruments is that the measurements must be
made at constant temperature and pressure in the sample chamber or else corrections must
be made. This is because the instruments do not really measure concentration but the absolute
number of molecules in the radiation beam. Thus the machine would give the same reading
for two gases, one of which had twice the actual concentration (of S02' for example) as the
other, but was read at one half the absolute pressure. In order to avoid this problem, a
manometer was installed in the sample line leading to the DuPont instrument, and flows from
the different gas sources (either standard gas or sample port) were adjusted for constant pres-
sure and not necessarily constant flow rate.
The discussion above and other problems concerning the instruments were worked out
with the assistance of Mr. Robert Saltzman of DuPont who had major responsibility in the develop-
ment of the Model 400 instrument. He spent a day at Tyco examining the sample system and
the application of the photometer and concluded that accurate analysis of the various process
streams would be very much dependent on sampling technique. He indicated that most of the
problems he had experienced in the past were not due to machine malfunction
but to problems in either sampling technique or sample stream contamination.
The problem of contamination is a very real one in the system under discussion because
of the chemical species involved. The Model 400 Ultraviolet Photometer measures S02 con-
centration by attenuation of the energy at 302 I?1J.. N02' which is monitored at 436 m1J., also
absorbs at 302 m1J. but to a lesser extent. Great care must therefore be taken to properly
~ ~
balance the machine which has the capability to zero out the N02 influence when measuring
S02. In addition, the nitrosyl group absorbs very strongly at 302 m1J. and any deposition of
nitrosylsulfuric acid on the windows of the UV cell would cause large errors in the measure-
ment of S02 in the gas in the cell. Contamination on the cell windows can be zeroed out of the
measurement, but if there is a continuous buildup of the material there would be a continually
changing interference which would make the analytical results of marginal value.
The way to avoid interference of this sort is to maintain high temperatures in the cell
and in the sample lines. If nothing is permitted to condense prior to the cold trap, there should
be no problem with interference. It was felt that the gas phase constituents of HNSOS would have
no effect on the readings since the nitrosyl group apparently breaks down to nitrogen oxides
which can be accounted for. There is the possibility that sulfuric acid mist might absorb
nitrogen oxides to form nitrosylsulfuric acid, but if the temperatures are kept high enough,
this effect would be minimal. In actual operation, it was found that at temperatures below
250 0 F there was a slow, continuous deposition of HNS05 - bearing acid on the quartz windows
of the sample cell, but as the temperature was raised this problem disappeared.
91
-------
The overall evaluation of the analytical instrumentation was that they are very dependable
machines and give good results as long as the sampling system was very carefully controlled.
It is important to know at all times the exact nature of the sample entering the cell: its
residence time in various stages of the miniplant and sampling system, its temperature, how
much material might have deposited out of the gas between the miniplant and the analytical cell
and any other change in condition.
3. Equipment performance
The first week of miniplant operation was on an 8-hr day basis in order to
evaluate the instrumentation and the plant equipment. Without introducing any reactive com-
ponents to the gas or acid streams, the various stages of the plant were run to examine their
operating characteristics under process conditions.
The acid loop was started up with acid circulating at about 0.1 gpm at room temperature.
Pumps, flowmeters and valves were operated and watched for a period of time, as were the
scrubber and stripper columns. All phases of the acid system worked satisfactorily at low
temperatures. Acid distribution throughout the columns was adequate and the pumps operated
properly. However, when the system was heated up and run for a while, some intermittent
problems became apparent.
The major problem in the acid system which caused several days delay in the preliminary
and operating phases of the miniplant evaluation period was pump malfunction as a result of
seizure of the ball check valves. By observation of the acid flowmeters it was possible to know
when the pumps stopped working properly. With piston pumps such as were being used, the
function of the check valves is to permit acid to be pulled into the pump from the source during the
suction stroke and not allow the acid to be pulled back in from the discharge side of the pump.
Conversely, on the discharge stroke, the ball check valves should shut the suction side of the
pump so that the acid is pumped out of the discharge side of the pump. It was found that cor-
rosion had caused the balls to wedge in the valves so that the check valve on the discharge side
of the pump would stay open during the suction stroke with the result that the acid was pulled
back into the pump from the delivery pipe. To correct this problem, all the check valves were
disassembled and modified so that the balls would not stick in the valves. There were no further
problems with the pumps after this was done, but unfortunately several days were lost before
the problem was corrected and eliminated satisfactorily.
A less serious problem arose after the acid loop had been operated for several days.
Several of the valves in the system were large stop cocks which required a lubricant to permit
easy operation. After being used several times, it was found that the acid species in the gas and
liquid loops were dissolving the lubricant causing the stop cock to stick. In some cases, the
valve became frozen and had to be heated up to free it. All types of lubricants were used to
remedy the problem including both silicone and fluorocarbon greases, but the problem continued.
The only way that was found to avoid the problem was to service the valves regularly before they
92
-------
froze. In the future, this type of valve will be avoided and either Teflon or glass valves will be
used. Saunders-Grinnell glass lined valves are used in the acid pump bypass lines as control
valves and also as shutoff valves to isolate the pumps, and they are working quite satisfactorily.
The blower performed satisfactorily throughout the miniplant operating period although
for 2 days flow decreased intermittently. It was found that the oil level of the blower had dropped
very low. This was regularly checked thereafter, with no further problems noted.
Overall, the physical performance of the miniplant was very satisfactory and by the end
of the 3-week operating period it was operating smoothly. Some modifications will be necessary
for future work because of required expansion of operating characteristics, and increased
flexibility, but the operating capabilities of the plant are sufficient for further optimization
studies.
4. Process evaluation - initial sttldies
Using the start"1.lpprocedure shown in Appendix III, the 2 -week period of round-the-
clock operations was begun with a 12-hr interval during which the acid and gas loops were
slowly brought up to operating temperatures and fluid flow rates and kept there for several
hours. This was done to make sure the equipment was continuing to function properly before the
reactive gases were introduced into the system.
It was during these early experiments that most of the pump problems occurred and it
wasn't until late in the first week that useful operating data could be obtained. Typical results
during initial reactor evaluation tests are shown in Table XiV. The S02 oxidation yield of
89.2% is quite satisfactory and although the NO balance was not as accurate as had been
x
desired, it was still in reasonable agreement.
Continued experimentation was performed in the miniplant which evaluated both the
scrubber and the stripper in conjunction with the reactor. After several days of integrated
operation, a pattern seemed to develop which demonstrated the following performance levels:
S02 removal efficiency
Scrubber efficiency
Stripper efficiency
Although the S02 oxidation reaction appeared to be satisfactory (although somewhat
erratic), the scrubber and stripper were both performing below the capacity observed during
the extensive laboratory testing. (It should be noted here again that all performance data
obtained during the miniplant experimentation were based on the DuPont and Beckman spectro-
photometers unless otherwise noted.) It was therefore decided to modify the experimental
approach and concentrate on one process stage at a time, in order to determine its operating
characteristics and find out the reason for the unusually poor performance of the scrubber and
75-95%
10-30%
40%
the stripper.
93
-------
Table XIV. Miniplant Experimentation - Reactor Evaluation*
Input gas concentration (into reactor) t
s~
N~
NO
3700 ppm t
6000 ppm
3000 ppm
Output gas concentration (from reactor)
S02
~~
400 ppm
3200 ppm
5000 ppm
Reactor residence time
S02 oxidation efficiency
NO material balance loss
x
2.2 sec
[ (3700
[(9000
400) /3700] x 100 = 89.2%
8200) /9000] x 100 = 8.9%
Notes:
* Gas flow rate: 8 SCFM
Reactor temperature: 300 -320° F
Reactor volume = 0.5 ft3
t All gas concentrations measured as volume ppm on DuPont and Beckman
spectrophotometers.
t Input gases measured as fed individually to the reactor.
5. Miniplant evaluation - reactor testing
A series of runs were made in the miniplant to evaluate the performance of the
S02 oxidation reactor. The gas was fed to the reactor from the blower using the gas burner to
add water to the stream. S02 was fed from a cylinder and N02 was introduced by oxidizing
NO with air in the NO oxidizer chamber. The effluent from the reactor was sent through the
94
-------
bypass to the stack after a sidestream was removed for analytical purposes. In other words,
the reactor was run identically to the way the integrated plant was to run except that the reactor
effluent was vented rather than being fed to the scrubber.
Results of several of the runs are shown in Table XV. Again the 802 removal efficiency
is quite high, in the 75 to 95% range although the NOx material balance was somewhat erratic.
One general observation that can be made was that the change in 802 was not exactly equal to
the reduction in N02 and the increase in NO. One of the difficulties was that the NO was not
completely oxidized in the NO oxidizer, although it had sufficient residence time in this chamber
according to the data of Burdick.22 Another problem is that in the low temperature runs the
NO probably had an opportunity to oxidize in the reactor and possibly in the sample lines. In
general, it can be noted that the data show less NO and more N02 in the effluent than would be
expected. This observation is in keeping with the idea that some of the NO was oxidized before
it was analyzed.
In connection with the analytical equipment, it should be remembered that for a period
of time the gas passes through a cold trap that has a fairly large residence volume. Thus the NO
could be oxidized before it reaches the Beckman spectrophotometer and would therefore not be
detected at all. For example, if oxidation occurred in the cold trap, the NO in the DuPont analyzer
would not be detected and would be N02 in the Beckman and again would not be measured. This
would show up as an overall shortage of NO , which is what has been observed.
x
The significant point of these data is that the experimental setup consistently gave high
802 removal rates under a variety of conditions. The apparentN02/NO ratios in the effluent gas
from the reactor were in the range of 0.60 to 1.93 (see Table XV) and thus deviate from the
desired ratio of 1.0. This was due to the way in which the N02 was prepared. There does not
seem to be any reason why it cannot be controlled under more representative operating conditions.
During this work, no effort was made to maintain this control except in terms of scrubber opera-
tion' which will be discussed later. In the experiments concerning reactor evaluation, control
of the nitrogen oxide ratio was not emphasized.
6. Miniplant evaluation - high temperature scrubber tests
The experimental results discussed above indicated that the 802 oxidation reactor
was apparently operating to give satisfactory removal efficiencies. If this were true and the
N02/NO ratios in the gas entering the scrubber were as given in Table XV, then the scrubber
should have been removing large percentages of the NO in the reacted flue gas. The scrubber
x
efficiencies observed during some of these tests were in the 10 to 30% range, much lower than
expected. Table XVI shows typical data during one of this scrubber evaluation runs. The 802
removal efficiency was 69 to 87% and the NOx feed to the scrubber showed an N02/NO ratio in
the range of 1.2 to 2.2. Perfect scrubbing of such gas mixtures on a 1: 1 basis of N02/NO
would have shown absorption efficiencies of 60 to 90%. The fact that the observed efficiencies
were 6 to 18% indicates that either the scrubber was not working very efficiency (and
thereby contradicting a large amount of laboratory scrubbing data) or the input gas to the
scrubber was not accurately analyzed.
95
-------
Table XV. Miniplant Reactor Experiments
1 2 3 4 5 6 7 8
* 9
Reactor gas flow rate, SCFM 7 7 7 5 5 8.2 9.0
Reactor residence time, sec 2.8 2.5 2.7 2.7 4.5 3.5 2.1 1.9
Reactor temperature, of 100 324 180 182 120 320 300 290
S02 input concentration, ppm 2930 3200 3500 3740 4150 3350 3950 3350
N02 input concentration, ppm 6000 5000 6000 6000 5900 4500 7750 5100
NO input concentration, ppm 3000 3000 3000 3000 2150 2100 1500 2950
~ S02 output concentration, ppm 100 300 150 150 550 500 250 550
C7)
N02 output concentration, ppm 1800 t t 1700 5600 4400 4950 3900
NO output concentration, ppm 2750 t t 2850 2900 3175 2400 1400
S02 removed, % 96.7 90.6 95.7 96.1 90.6 85.0 93.6 83.6
NO material balance, % t -50 t t -50 -5.6 -14.8 -20.5 -34.2
x
Effluent N02/NO ratio 0.66 0.60 1.93 1.39 2.06 2.79
* Air plus burner gas.
t Output NO data not available.
x
t Minus sign (-) indicates a loss in the NO output, a plus sign (+) indicates a gain.
x
-------
Table XVL Miniplant Evaluation of Scrubber*
(Wet Feed: NOx + S02)
Reactor** Effluent, Scrubber Effluent,
ppm ppm
Time, N~/NO NOx Scrubbing,
min NO N02 S02 ratio NO N02 S02 %t
o 3850 4800 950 1.25
10 3225 4300 550 1.33
15 3000 4100 325 5.7
70 2500 4000 400 1.6
115 2325 5100 500 2.2
135 1900 4400 525 15.2
150 2000 4100. 550 17.8
Notes:
* I Scrubber conditions
Gas flow rate = 5 SCFM
Acid flow rate = .12 gpm
Scrubber temperature = 250°F
3/8-in. lntalox saddles
Packing height
Column diameter
= 8. 5 .it
= 4 in
S~ - 3050
NO - 3500
N02 - 6000
t 07 . ppm NOx (reactor effluent) - ppm NOx (scrubber effluent) 100
70 NO scrubbmg = x
x ppm NO (reactor effluent)
. x
** Reactor feed:
97
-------
Additional runs were made to ascertain the low scrubbing rate and the results of one
test is shown in Table XXI. The scrubbing efficiency was 8 to 9% although the S02
removal efficiency was about 85% and the N02/NO ratios was 2.2. A scrubbing efficiency
of about 60% was possible on a 1: 1 absorption basis. It should be noted that an NOx material
balance around the scrubber closed within 15% indicating that the analysis was not greatly in
error.
The results of tests like those; shown in Tables XVI and XVII suggested that there was a
significant difference between the laboratory scrubbing tests and those performed in the mini-
plant. Examination of the operating conditions in both sets of tests show that, in addition to
some differences in apparatus and temperature control, a major change that was initiated in the
miniplant was the use of the 502 oxidation reaction products rather than an artificially prepared
mixture of nitrogen oxides in a nitrogen carrier gas. Thus the gas that was fed to the laboratory
scrubber was inert in the sense that it could not undergo further internal reaction. On the other
hand, the scrubber feed in the miniplant was undergoing a continuing reaction between S02 and
N02' The significance of this was not apparent until some tests were run in the miniplant that
approximated the laboratory runs.
A miniplant experiment was conducted where the scrubber feed consisted only of moist
gas containing partly oxidized NO. This was the same gas as was used in the reactor tests
except that no 502 was introduced into the gas stream. Analysis was performed in the usual
manner by sampling the effluent gases from the reactor and the scrubber and feeding them to
the DuPont and Beckman spectrophotometers. Acid samples were taken from both the scrubber
and the stripper so that a complete NO material balance could be made around the scrubber.
x
The results of this first test are shown in Table XVIII and indicate that the scrubber was
now working at 60-65% efficiency with the feed gas containing the oxides of nitrogen in a nearly
1: 1 ratio. The NO material balance around the scrubber closed within 4% indicating a good
x
agreement of the acid and gas analytical techniques. (It can be noted here that the stripper is
not completely denitrating the scrubber effluent acid. This will be discussed in detail in a later
section of the report.) The significance of this test is that the scrubber continued to operate at
relatively high efficiency (compared to the earlier runs) as long as the feed acid was low in
HNS05' which tended to verify the concept that a scrubber could remove equimolar NO and
N02 with high temperature sulfuric acid.
Another run was made in the miniplant with conditions essentially the same as in the run
shown in Table XVIIL The results of this second run are shown in Table XIX and again demon-
strate a high degree of scrubbing as long as the acid feed to the scrubber is low in oxides of
nitrogen. As shown graphically in Fig. 37 the scrubber efficiency went down as the NO content
x
in the acid feed increased. The reason for this is the theoretical limitation whereby the acid can-
not reduce the NOx concentration of the gas below that equal to the equilibrium vapor pressure
over the incoming acid. Since this is true, as long as the feed acid contains a significant amount
of NO there must be some NO in the scrubber effluent gas regardless of the height of the col-
x x
umn. Because of this, the true efficiency of the column is obtained during the early stages of the
98
-------
Table XVIL Miniplant Scrubber Evaluation (Wet Peed: NOx + S02) *
Scrubber Stripper, Scrubber Effluent Gas
Acid Concentration Acid Concentration Concentration, ppm NO
Time, ~ ~ x
min % H2S04 % HNS05 % H2 SO 4 % HNS05 NO N02 S02 NO Scrubbing
x
Peed: 2100 4500 500 6600
0 78.20 0.086
5 78.02 0.205 2400 3700 650 6100 7.6
10 78.02 0.194 2400 3700 6100 7.6
15 78.02 0.194 78.55 0.074 2400 3600 650 6000 9.1
30 78.02 0.197 78.02 0.108 2400 3600 700 6000t 9.1
45t 78.02 0.124
60t 78.20 0.088
90t 78.20 0.051
Notes
* Gas flow
Acid flow
Scrubber temperature
3/8 -in. Intalox saddles,
4-in. diameter column
Reactor input S02 = 3.300 ppm S02
t Material balance error around scrubber was 14.9%
t Stripping only: no NO fed to scrubber.
x
= 5 SCFM
= .12 gpm
= 250 to 265°P
8.5 ft high
99
-------
Table XVIIL Miniplant Scrubber Evaluation
(Wet Feed: NO only)
x
Scrubber Acid Stripper Acid Scrubber Effluent Gas
Concentr ation Concentr a tion Concentration, ppm
Time, ~ --""- - --- ---"-- - ........... ---
min % ~SO 4 % HNS05 % ~S04 % HNS05 NO N02 NO
x
Feed: 3700 3300 7000
o 80.74 0.019
5 80.36 0.268 1800 700 2500
10 80.36 0.316 1775 650 2425
15 80.36 0.324 80.36 0.026 1750 650 2400*
20 1750 700 2450
25 1750 675 2425
..... 30 80.36 0.379 80.36 0.179 1725 750 2475
o
0
40 1750 825 2575
45 80.00 0.490 80.00 0.253
50 1750 950 2700
60 80.00 0.673 80.00 0.390 1775 1100 2875t
75t 80.00 0.435
90 80.00 0.409
Notes:
*NO material balance error around scrubber was 3.7%.
x
t NO material balance error around scrubber was 3.3%.
t Stripping only: no NO fed to scrubber.
x
Gas flow rate = 5 SCFM
Acid flow rate = .12 gpm
Scrubber temperature = 230 - 2500 F
3/8-in. Intalox saddles, 8.5 ft high
NO Scrubbing,
x %
64.3
65.3
65.7
65.0
65.3
64.6
63.2
61.4
58.8
-------
Table XIX. Miniplant Scrubber Evaluation
(Wet Feed: NOx only)
Scrubber Acid Stripper Acid Scrubber Effluent Gas
Concentration Concentration Concentration, ppm
Time, --- - -------- -""- ~ Scrubber Tern - NOx Scrub-
min % ~S04 % HNS05 % ~S04 % HNS05 NO N02 NO perature, 0 F bing, %
x
Feed: 3425 2500 5925 248
0 79.8 0.085
5 78.9 0.330 1350 750 2160 252 64.6
10 78.7 0.323 1350 800 2150 255 63.8
15 78.9 0.330 1350 900 2250* 254 62.8
20 79.4 0.148
30 79.4 0.403 78.9 0.211 1400 1100 2500 248 57.8
35 1400 950 2350 246 60.4
40 1425 900 2325 60.7
45 79.1 0.453 78.9 0.281 1450 900 2350 247 60.4
50 1500 850 2350 60.4
60 78.5 0.498 78.7 0.352 1550 800 2350 247 60.4
75 78.7 0.527 1675 850 2525 57.3
90 78.2 0.543 78.6 0.419 1775 1000 2775 253 53.3
105 78.4 0.573 78.6 0.442 1850 1100 2950t 50.4
120 78.0 0.557 78.4 0.457 2000 1200 3200 46.2
135 77.9 0.557 78.2 0.473 2100 1450 3550 242 40.3
150 77.7 0.581 77.9 0.476 278
180~ 77.3 0.555 77.7 0.483 3425 1400 4825 18.7
195 3500 1475 4975 289 16.3
210 77.0 0.553 77.3 0.456 2650 1900 4550 23.3
225 2850 2050 4900 17.4
240 76.8 0.520 77.2 0.448 2950 2050 5000 ~ 289 15.7
Notes:
*NO material balance error around the scrubber was 9.2%.
x
t NO material balance error around the scrubber was 5.6%.
x
~ S02 introduced at the 165 min mark and kept flowing at about 750 ppm concentration until the 195 min point.
~NO material balance error around the scrubber was 0.4%.
x
Stripper temperature 220 270' F.
101
-------
1:f2
~
»
~
Q)
.....
~ 40
4-<
~
bO
~
:z
,.q
;:J
I-;
rJ5 30
~
o
z
60
50
20
10
S02 introduced
50
100
150
200
Time, min
Fig. 37. Miniplant scrubber evaluation
wet feed: NO only
(data from Ta6le XIX)
102
-
5
o
(J
o
~
()
(1)
~
r-t
.4 ;;!
r-t
,.....
o
~
o
....,
C/1
r-t
'i
.3 -5'
'1:j
(1)
"i
tT1
....,
....,
-
~
(1)
.2 g
>-
()
,.....
0.
~
~
.1
('sQ
:r:
z
C/1
o
c.n
-------
run when the feed acid is relatively pure. This will not be a problem when the stripper is prop-
erly operating so as to completely denitrate the acid (see parr 10 of this section for a discussion
of stripper operation) .
The last two tests indicate that there is a difference between scrubbing the products of
the reactor and scrubbing pure NO. Since the analysis shows the same range of NO /NO feed
x 2
ratios as in the scrubbing tests exhibiting low efficiency, it was hard to understand the reason
for the large difference in scrubbing efficiency. These tests led to a complete evaluation of
what is actually occurring in the gas stream and the relationship between the miniplant system
and the gas analysis system.
7. System evaluation
It is tempting to dismiss some of the anomalous results quoted above as being due
to either experimental or analytical error, but the consistancy and reproducibility of the results
coupled with good material balances around the scrubber verify the data. The question was:
what was the difference between the scrubber experiments NO alone was fed to the scrubber
x
and the tests where a mixture of NOx and S02 was passed into the system?
Careful examination of the data led to the conclusion that the real difference between the
two sets of tests was that in one case an inert gas was being fed to the system (inert in the sense
that little or no reaction was occurring in the gas phase as it passed through the system), and
in the other case a continually reacting gas was in the lines. In addition, it was concluded that
the sample lines and the ultraviolet analysis cell had become part of the reaction system and
the analyses made did not truly reflect the concentration of the gases in the miniplant. Instead
the analyses showed the makeup of a gas stream that had undergone further reaction since it
had been in the miniplant and thus the effluent and feed streams were not as had been recorded.
If these two conclusions are correct, then the scrubber data can be explained. The
S02 was mixed with NOx and fed to the reactor chamber where the oxidation of the S02 began.
The gas stream was then split, with the bulk of the gas going to the scrubber and the rest, a
very tiny fraction, was sent to the sample lines and photometers for analysis. As it passed
through the hot sample line and into the UV cell, the gases continued to react. When the
instruments finally reported the concentrations of the oxides of nitrogen and sulfur dioxide,
the composition of the gas was considerably different from that entering the scrubber. The S02
reported was lower than that entering the scrubber as was the N02' while the NO observed on
the Beckman instrument was probably much more concentrated than in the scrubber feed gas.
The residence times of the gas in the various parts of the system were calculated.
Volumes of different parts of the system and the residence times at specific gas flow rates are
shown in Table XX. It is clear that the gas passed through the sample system much more
slowly than it should have. It turned out that this was due to an improperly calibrated rotameter
which gave actual flows of 3.75 £/min (0.13 cfm) when 1.0 cfm was expected. The bulk of the
samples were taken at about 1.0 £/min which gave the residence times as shown in the table.
103
-------
-
Table XX. Residence Volumes and Times in Miniplant Reactor and
Analytical System
System Phase
Residence
3
Volume ft
Gas Flow
Rated at STP
Residence
*
Time, sec
S02 oxidation reactor
0.4143
11> CFM
5 CFM
2.5 CFM
1.7
3.4
6.8
Sample lines between
reactor gas effluent
and UV cell
0.0041
1.0 CFM
1.0 £/min
3.75 £/min
0.2
4.9
1.3
0.9t
26.1
7.0
UV cell
0.0153
1.0 CFM
1.0 £/min
3.75 £/min
Notes
*
Calculated at 300 0 F
t See text for discussion of effective cell residence times.
In the determination of the total residence time, it should be realized that if there is a reaction
occurring in the UV cell, the concentration of any reacting species can be roughly approximated
as the linear average of the incoming concentration and the effluent concentration. The residence
time of this average gas sample is therefore estimated to be one-half the total residence time of
the gas in the cell. Thus, for the majority of the experiments, the effective gas residence times
were 3.4 sec in the miniplant reactor, 4.9 sec in the sample lines and 13.1 sec in the UV cell.
Using these residence times, it is clear that the reason for the poor scrubbing effi-
ciencies when the reactive mixture was fed to the scrubber was that the S02 had not yet been
fully oxidized and there was a very large excess of N02 entering the scrubber. Although the
laboratory data indicated that high scrubbing efficiencies could be achieved without a perfect
1: 1 ratio of nitrogen oxides, the feed stream to the scrubber in the miniplant experimentation
was probably excessively strong in N02 and thus hindered absorption. (It should also be noted
that the laboratory experiments that showed good adsorption of NO despite an absence of the
x
1: 1 ratio of nitrogen oxides usually had an excess of NO, not N02) .
104
-------
A significant point in the experimentation is that the high temperature scrubber worked
relatively well in absorbing oxides of nitrogen when the scrubber feed gas was known to contain
close to a 1: 1 ratio of N02 and NO when S02 was not present.
The analysis that has just been discussed has a significant effect on the earlier analysis
of the reactor. It was assumed from the reactor evaluation experimentation that the S02 had
been almost completely oxidized in the 2 to 3 sec of residence time in the miniplant reactor.
Some experimental evidence of the oxidation of S02 in the sample lines was obtained by
withdrawing the sample of the reactor effluent gas at varying sample line flow rates. At a flow
rate of 1 f/min (17 sec total residence time), about 90 to 95% of the S02 was removed from the
gas stream. At a sample line flow of 4 f/min, and a total residence time of 7.9 sec, about 30
to 40% of the S02 was oxidized.
A resummation of the miniplant reactor data of Table XX is presented in Table XXI
showing actual residence time and reaction rate constants. Clearly, it has taken residence
times in order of 20-23 sec to achieve 84-97% S02 oxidation. This is about four times the
residence times estimated based on the experimentation performed during the previous con-
tract (No. PH-86-68-75).
As in the previous report, the rate constants are based on the overall reaction:
N02 + S02 He) S03 + NO
2
using the general equation
dx
dt = K (a - x) (b - x)
where
x = decrease in each concentration in mm Hg after time t, sec
a = initial S02 concentration in mm Hg
b = initial N02 concentration in mm Hg
K = rate constant
This can be integrated using the boundary condition that x = 0 when t = 0, and solved:
K - 2.303 b(a - x)
- t(a - b) log a(b - x)
105
-------
Table XXL Miniplant Reactor Experiments
1 2 3 4 5 6 7 8
Reactor gas flow rate, SCFM 9 7 7 7 5 5 8.2 9.0
Reactor residence time, sec 2.8 2.5 2.7 2.7 4.5 3.5 2.1 1.9
Sample gas flow rate, £ /min 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
Sample line residence time, sec 18 18 18 18 18 18 18 18
Total residence time, sec 20.8 20.5 20.7 20.7 22.5 21.5 20.1 19.9
Reactor temperature, of 100 324 180 182 120 320 300 290
....
0 Input concentration, ppm
C1> 2430 3200 3740
a. S02 3500 4150 3350 3950 3350
b. N02 6000 5000 6000 6000 5900 4500 7750 5100
c. NO 3000 3000 3000 3000 2150 2100 1500 2950
d. N02/S02 2.47 1.56 1. 71 1.60 1.42 1.34 1.46 1.52
Output concentration S02' ppm 100 300 150 150 550 500 250 550
S02 removed, % 96.7 90.6 95.7 96.1 90.6 85.0 93.6 83.6
Reaction rate constant, K* 0.071 0.066 0.073 0.081 0.053 0.058 0.045 0.048
* Units of mm, sec, of. Note: K based on given input gas concentrations and output S02 removal efficiency.
-------
Rate constants (K) from 0.045 to 0.081 have been obtaineEl in the present effort. The calculated
rate constant (K) for 90% reaction with 3000 ppm S02 and 6000 ppm N02 lin see -1 is 0.150. Rate
constants of 0.200 were obtained in the kinetic studies of the previous work and covered a range
from 0.004 in early batch studies to 0.250 under a variety of conditions. The primary difference
between the present effort and the past effort is in the method of analysis and sampling, with
continuous methods being used only in this study. It is anticipated that present residence times
can be improved through further investigation of the reaction mechanism and optimization.
However, no indication is available at this writing whether the reaction times and rates can be
improved by a factor of 3-4. Thus longer residence times may have to be provided in the
reactor section.
8. Scrubber design calculations
Operation of the miniplant scrubber was intended to both verify the concept of the
high temperature scrubber and provide preliminary design data for process scaleup. Experi-
mentation was designed so that calculations could be made at the conclusion of the current
contract for the overall design of the scrubber. Engineering design data to be developed was
to include: (1) mass transfer coefficients (K a), (2) height of a transfer unit (HTU), (3) number
g
of transfer units (N ), and (4) overall height of the scrubber column (Z) .
t
The initial approach to the engineering design of the scrubber was to develop the
theoretical number of transfer units needed by the use of a McCabe-Thiele analysis. Fig. 38
shows the graphical determination of 4.31 theoretical transfer units. It should be noted that
this analysis is based on an assumed L/G of 4.6 moles of 80% H2S04 per mole of dry gas, which
was developed by a optimization cost study of the process by TV A under contract to OAP. In
performing the graphical analysis, a 50% equivalent tray efficiency was assumed. The equili-
brium data was obtained from the extensive work of Ber! and Saenger5 as discussed in Section V
of this report (see Table XXII for a tabulation of the values for 80% H2S04 nitrose solutions at
2500 F) .
This approach to the development of the number of transfer units was verified using a
more basic approach. Treybal23 presents an equation for the calculation of this value:
£n ( y 1 - ~ \ (1 - 1 ) +! ]
Y2 - ffiX2 J A A
1 -
(1)
NtOG =
107
-------
40
I I
I I
I I
I I
I I
- - - - - (J- - - - - - - - - - - - - - - -
Y1 = 2.62 mm Hg
5
Operating line
x, y
35
(L/G = 4.6
moles/mole)
NtOG = 4.31 (graphically)
30
-.t'
o
.....
x
~
H
.....
(1j 25
;>,
H
"0
(j)
]
"--
C'?
o 20
N
Z
rJJ
(j)
......
o
S
>-<
15
Equilibrium line 5
(Ber I and Saenger)
10
.8
5
10 15 20 25
X, moles N203/mole 80% H2S04' x 104
30
Fig. 38. McCable-Thiele analysis of high temperature scrubber
108
-------
Table XXIL Equilibrium Data*
Total Vapor N20S Vapor** y, moles ~Os/ x, moles N20S/
Pressure, Pressure, mole dr g-moles HNS05 g-mole 80%
mm Hg mm Hg air X lOt! per1. H2S04 ~S04 X 104
50.73 0 0 0 0
51.16 0.22 3.1 0.025 3.98
51.60 0.44 6.2 0.050 7.95
52.03 0.65 9.2 0.075 11.93
52.47 0.85 12.3 0.100 15.90
52.99 1.13 16.0 0.125 19.88
53.52 1.40 19.8 0.150 23.86
54.00 1.67 23.6 0.175 27.83
54.67 1.97 27.8 0.200 31. 80
55.93 2.60 36.8 0.250 39.76
Notes:
*Experimental data of Ber! and Saenger5 at 250°F, 80% H2S04
**Total vapor pressure less the vapor pressure of H20 (50.73 mm Hg).
109
-------
where
NtOG == overall number of transfer units, gas
A == absorption factor == L/mG (dimensionless)
x
== concentration of solute in liquid, mole fraction
y
== concentration of solute in dry gas, mole fraction
m
== slope of equilibrium line (dimensionless)
.Qn
== natural logarithm
2
== total liquid rate, lb moles/hr -ft
== total dry gas rate, lb moles/hr-ft2
L
G
The value of NtOG was calculated using the following values:
L/G == 4.6 mole/mole
m
== 0.77
A
== 6.0
Yl
Y2
== 37.2 moles N20/mole dry gas
== 0.8 moles N20lmole dry gas
== 0 moles N20lmole dry gas
x2
The feed and effluent gas concentrations, y 1 and Y2' were obtained from a material balance
around the scrubber based on an 502 input of 3000 ppm. The L/G value came from the TVA
optimization, and the assumption was made that the inlet acid concentration (x2) was negligibly
low. (If the stripper denitrates the nitrose at a 99% efficiency, this assumption is easily justi-
fied.) The calculation from equation (1) gave a value of NtOG of 4.42 which is in very good
agreement with the 4.31 obtained graphically from the McCabe-Thiele analysis.
With these values as background, experimentation was performed to determine working
values for the engineering design data needed for process scaleup. The experimental runs shown
in Table XVIII was used in calculation design data, but it was necessary to evaluate the scrubber
at various packing depths to verify the scrubbing efficiency obtained by one height of packing.
The miniplant scrubber was modified so that only 5 ft of the total 8.5 ft of packing was used
for countercurrent contact between the gas and the acid. The column was fed with a mixture of
nitrogen oxides as shown in Table XXIII and analyses of the gas and liquid effluents were taken
as in previous tests.
110
-------
Table XXIIL Miniplant Scrubber Evaluation (Wet Feed:
NOx only in flue gas; 5-ft bed height)
Effluent Effluent
Scrubber Acid Stripper Ac id
Concentra tion Concentra tion
Time, min %H2S04 % HNS05 %H2S04 %HNS05
Scrubber
Feed:
o 79.08 0.007
5 79.08 0.221
10 79.08 0.229
15 78.73 0.239 78.73 0.026
30 78.73 0.261 78.73 0.147
35
45 78.73 0.320 78.73 0.173
60** 78.73 0.202
Notes:
Scrubber Effluent
Gas Concentration
~
NO N02 NOx
NO *
x
Scrubbing, %
3325
2950
6175
2150 1200 3350 45.6
2250 1250 3500 43.4
2150 1250 3400 44.8
2200 1200 3400 44.8
2250 1350 3600 41.6
2325 1550 3875 37.3
*L/G = 2.4 mole/male (5 SCFM gas, .12 gpm acid)
**Stripping only: No NO fed to scrubber.
x
The analysis of scrubber operation showed that 5 ft of packed columns was capable of
removing 37 -47% of the NO in the feed gas as long as the input acid was relatively free of
x
nitrogen oxides. The tests shown in Table XVIII show that an 8.5 ft of packed column can
remove 59-66% of the NO under the same operating conditions.
x
These two data points can be used to analyze the performance of taller columns. If
5 ft of packing can remove 45% of the NO then the next 5 ft can remove 45% of the remainder
x
for a total of 70%. This can be expanded for greater packing heights as shown in Table XXIV.
The same treatment can be performed on the other set of data and this is also shown in
Table XXIV. Both sets of data are plotted in Fig. 39 and are mutually consistent since they
both fall on the same efficiency curve. Examination of this curve shows that 97% scrubbing
efficiency can be achieved with about 30 ft of packed column, certainly a reasonable packing
height. (99% scrubbing could be achieved with 35 to 40 ft of packing) .
111
-------
Column
Height, ft
Extrapolated Scrubbing Efficiencies Based on Experimental Column Tests
%00 %00
x ~fu~ x
Removed Height, ft Removed
Table XXIV.
5
10
15
20
25
30
35
45* 8.5 65*
70 17 88
83 25.5 96
91 34 99
95
97
98
*Experimental data obtained in the absence of S02 is the scrubber gas.
The height of a transfer unit can be obtained from the curve in Fig. 39 by making a
simplifying assumption. This assumption is developed by Sherwood and Pigford24 by examina-
tion of the following definite integral which is used for the estimation of tower parameters:
SY1
NtOG =
Y2
dy
Y - Ye
(2)
where
NtOG = overall number of transfer units, gas
Y
= gas phase concentration of solute
Ye
= gas phase concentration when at equilibrium with the liquid
phase.
112
-------
35
30
25
1
o
o
.....
o
~
t1 15
.....
Q)
::r:
Feed NO = 6200 ppm
x
-- Hill = 8.3 ft
'I Reduction to lie
of original gas
I concentration (one
I transfer unit)
I
I
~
8~ 20
:::s
8
CD, m Experimental data points
m 8.5 ft of column
CD 5.0 ft of column
o Extrapolated values based on 5 ft
column experim~nt
o E.xtt'ap:>lated values based on
8.5 ft column experiment
10
5
20
'40 60
% NO Absorbed
x
100
Fig. 39. Scrubbing efficiency of packed column; experimental
and extrapolated values used to determine HTU
113
-------
The simplifying assumption is that if the vapor pressure of the solute over the liquid is very
low compared to the concentration of the solute in the gas being scrubbed, then y is small
e
enough to be ignored. The equation can then be easily integrated:
Yl
NtOG =: 1'n Y2
If one transfer unit is being considered (NtOG = 1), then:
Yl
1'n- =: 1
Y2
(3)
and
Yl
- =: e
Y2
(4)
This means that, if the assumed conditions pertain to the system in question, one transfer unit
will reduced the outlet gas concentration to lie (or 37%) of the feed composition.
In examining this approach with respect to the N203/H2S04 system existing in the
scrubber, it can be shown that the vapor pressure of N203 over the dilute nitrose solutions
existing in the system is relatively low compared to the gas composition. Typically the vapor
pressure would be about 15-20% of the gas phase concentration (see Fig. 38 for a comparison of
the operating line with the equilibrium line). This is not a negligible value, but it is low enough
to allow the use of the technique for a first approximation.
To determine the HTU by this method, the curve in Fig. 39 is examined and the column
height is read at 63% scrubbing efficiency. As shown on the graph, one HTU is equal to 8.3 ft
of packed column. The actual number of transfer units could then be determined by dividing
the height needed to obtain 99% scrubbing (35 -40 ft) by the HTU (8.3 ft). This gives a value of
NtOO of 4.2 -4. 8 which is fairly close to the predicted values of 4.3 to 4.4.
The determination of HTU can be done more rigorously using the experimental data in
Treybal's simplified equation for the determination of number of transfer units in a dilute gas
system in conjunction with equations for mass transfer coefficient, K a, and height of a
23 y
transfer unit, HTU
Yl
NtOO =:
(y
Y2
*
y )M
(5)
114
-------
where
* *
( - * - (y - y ) 1 - (y - y ) 2
Y Y)M - * *
in (y - y }i(Y - Y ) 2
(6)
Y 1 = feed gas concentration, mole fraction
Y 2 = effluent gas concentration, mole fraction
*
Y = equilibrium vapor pressure of solute over acid in contact with
scrubber gas of concentration y
*
Thus (y - Y ) is the vertical distance between the operating line and the equilibrium line.
The other equations are:
*
G(Yl-Y2) =KgaZPt(Y-Y)M
(7)
or, from equation (5) :
GN
K a = -.!OG
g Z Pt
(8)
where
K a = overall mass transfer coefficient, gas,
g
lb-moles/hr ft3 - atm
G
= gas flow rate, moles/hr
ft2
Pt
Z
= total pressure, atm.
= hei ght of packing, ft.
and
Z
HID =-
Nt
(9)
where N here refers to the number of transfer units in a column with Z ft of packing.
t
The equilibrium data from Table XXlI is plotted again in Fig. 40 with the operating line
for an I)G of 4.6 plotted for reference. In addition, the experimentally determined operating
lines for the tests shown in Tables XVIII and XXIII are plotted on the same graph. The values
115
-------
40
Y1 = 2.62 mm Hg
- --------------------
"'"
o
-
35
~Theor~tical
operatmg
line, x, Y
L/G = 4.6 mole/mole
30
Experimental operating
lines, from following
Tahles:
-----4~..:min XXIII
15 min XVIII
10 min XXIII
x
H
.,....
Ct!
t 25
"0
Q)
.......
o
6
'----
C")
o
ZN 20
rJJ
Q)
.......
~
><
10
Equilibrium line 5
(Bed and Saenger)
x*. y*
5
10 15 20 25
X, moles N203/mole 80% H2S04 X 104
Fig. 40. Experimental operating lines - scrubber design
116
30
-------
*
of(y - y )are taken as the vertical distances between these experimental operating lines and the
equilibrium line and are used in equation (5) to determine experimental values of N. It should
t
be remembered that in this case Nt is the number of transfer units in the given column to effect
the absorption shown by the experimental operating line and is not the total number of transfer
units needed to achieve the desired 99% absorption of NO . This was done for the three cases
x
chosen for analysis.
The values of Nt were then used to calculate the mass transfer coefficients using
equation (8) . HTU values were determined using the values of N in equation (9). All calculated
t
results are shown in Table XXV and are compared to the values obtained by the simplified
analysis of the data using the curve in Fig. 39.
What is signifjcant about these data is the close agreement of the 'calculated values with
each other and with the simplified analysis. Although the values of HTU are somewhat higher
than usually encountered in industrial practice, the number of these units needed to effect the
desired absorption is relatively low as shown by the calculated values of NtOG in Table XXIX.
It is also interesting to note that these values of NtOG are quite close to the predicted values
of 4.3 and 4.4 determined by McCabe-Thiele graphical analysis and mathematical computation.
The values of mass transfer coefficient, height of transfer unit, and number of overall
transfer units obtained experimentally by operation of the mini plant will be invaluable in fur-ure
work. Both plant scale-up and future miniplant operation will be aided by the development of
these figures. The values are not the final ones since considerably more experimentation will
be performed to refine figures, but they certainly present the order of magnitude of these
vital parameters. It should be noted here again that there was no 502 present during these
NO absorption experiments.
x
Table XXV. Calculation of Mass Transfer Coefficients (K a) and
Height of Transfer Unit (HTU) * g
z, Packed Nt K a~
Reference Height, ft HTU, ft
t g
Table XVIII 8.5 1.16 1.31 7.4
Table XXIII 5 0.61 1.17 8.3
(10 min)
Table XXIII 5 0.55 1.06 9.1
(45 min)
Fig. 39 1 1.17 8.3
Averages 1.18 8.3
*In these experiments the following conditions pertained: G = 9.59 moles gas;hr-ft2;
Z is given for each test above; Pt = 1 arm total pressure.
tN refers to the number of transfer units in the packed height shown in the table.
t
~moles/hr-ft~ -atm.
117
-------
9. Water balance in scrubber
One of the key features of the high temperature scrubber was its capability of
scrubbing NO from the reacted flue gas while maintaining the concentration of the recycle acid.
x
Examination of the miniplant operating data shows that this capability was demonstrated in
several runs. The tests described in Tables XVII and XVIII show good water balance maintenance
in both the scrubber and the stripper loops. In the test shown in Table XIX, there was a slow
decrease in acid concentration indicating that water was being absorbed, causing dilution. This
test shows the need for proper temperature control in the system, not only in the scrubber itself,
but also in the acid feed. Although the scrubber temperature was maintained near 250 0 F
throughout the test, the input acid was cooler and thus caused absorption of water at the top of
the column. In general, the data clearly show that the concept of scrubbing at high temperatures
to avoid water absorption is feasible.
10. Stripper evaluation
A tabulation of stripper results during simultaneous operation of the scrubber,
as presented in Tables XVII, XVIII, XIX, and XXIII are graphically represented in Figs. 41, 42,
43, and 44. Fig. 43 is the most meaningful as the stripper-scrubber operation was performed over
a prolonged time period.
In these figures, starting time zero indicates the time scrubber operation is begun.
Approximately 10 min are required at a flow rate of 0.16 gal/min to obtain a sample at the
bottom of the stripper, due primarily to residence in the feed line. The accumulated stripper
data with time is thus indicative of conditions occurring in the scrubber ten minutes earlier.
Furthermore, the stripper bottoms are collected in a reservoir of approximately 5 gal hold
capacity while being continuously refed to the scrubber. Thus, where partial stripping is
achieved (as in the case of these data), the rate of nitrosylsulfuric acid buildup in the scrubber
acid feed is retarded as a result of dilution in the reservoir. Both of these factors are evident
in the initial shapes of the curves presented.
Two pertinent conclusions can be drawn from the data presented:
1. The present miniplant stripper design cannot denitrate at a rate fast enough to
maintain integrated operation without modification.
2. Despite the buildup of nitrosyl sulfuric acid, some denitration was obtained.
As can be noted from the relatively parallel stripper-scrubber curve in Fig. 43, denitration
at steady state appears to occur in the stripper at an essentially constant rate.
In view of the extremely encouraging earlier laboratory data on the stripper, a reassess-
ment of thescale-updesign conditions in the miniplant stripper was made. With limited available
laboratory design and optimization data, the miniplant stripper was sized for a similar L/G ratio
and roughly comparable residence times in the packed bed. No change was made in the type or
size of charcoal catalyst used. It was hoped that satisfactory stripping could be achieved at con-
siderably higher space velocities than obtained in laboratory equipment. A comparison of the
118
-------
It)
o
Cf.)
z
:r:
f:f:!
.....
.....
co
~
~
a
o
.,...
.....
cIS
~
Q)
u
=
o
U
"0
.,...
U
~
~
Q)
='
E
~
.3
J
Scrubber
Gas feed to scrubber
roppoo
Stripper conditions
L/G: 6.0 mole/mole
Air feed: 2 SCFM
Acid feed: 0.12 gpm
Temperature: 240-2800 F
~--..o..................
--0-- - - - Stripper
-- -
- - --0.-
cr- --
--
-.0
60
75
15
30
45
Fig. 41. Scrubber/stripper operation in miniplant - data from Table XVII
Time, min
-------
.I
.6
10
0
U)
z
::r: .5
Cfi
~
a
0
.....
.....
cd .4
~
Q)
(j
~
0
U
"0
..... .3
(j
-------
It) .5
o
U')
z
::r:
~
~ .4
~
0
'P
CIS
~
Q)
~ .3
8
'0
....
u
.... <:
to.:) =
.... .2
Q)
E
~
.1
Scrubber
,0- - -<>-
-c"'-
,g""-
/", Stripper
/"
/'
",
?
I
I
I
/
/
I
/
J
Oenitration
--~
--
- "0-
---""---0
Stripper conditions
L/G: 6.0 mole/mole
Air feed: 2 SCFM
Acid feed: 0.12 gpm
Temperature: 230-2700 F
30
60
150
90
120
Time, min
180
21Q
Fig. 43. Scrubber/ stripper operation in miniplant - data from Table XIX
-------
.5
In
0
VJ
Z
:r::
c? 4
~
d'
0
..-.
~
C1I
b .3
,::
Q)
u
,::
0
U
"0
..-.
u .2
.... ~
t.:I ~
t.:I ,::
Q)
;j
!E
~
.1
Stripper conditions
L/C: 6.0 mole/mole
Air feed: 2 SCFM
Acid feed: 0.12 gpm
Temperature: 240-2700 F
Scrubber
--
/;y- Stripper
/
/
/
/
/
if'
--0
--
~-
-
15
30
Time, min
45
60
Fig. 44. Scrubber/stripper operation in miniplant - data from Table XXIII
-------
conditions in the laboratory and miniplant catalytic strippers are presented in Table XXVL
The conditions used in later experiments are also noted in the table.
Table XXVL Design and Conditions for Operation of Laboratory and Miniplant
Ca talytic Stripper
Mi~ant --
Initial
Lab Design Verification Experiments
~-
1 2
I Column
Diameter of column, cm 5.08 10.16 10.16 10.16
Height of catalyst bed, cm 25.4 81.2 81.2 81.2
Volume of catalyst bed, m.£ 510 6580 6580 6580
II Flow
(L) Acid flow rate, g/min 17.3 1,310 65 104
(G) Gas flow rate, g/min 1.08 97 6.15 6.15
L/G mOle/mole 9.3 7.8 6.2 9.8
III Space Velocity{Liquid),
Volume of acid, .£ per hr 1.18 12.0 0.34 0.54
per volume of catalyst
As noted by Johnstone and Thring,25 the overall rates of many heterogeneous chemical
reactions are controlled by the rates at which the reacting substances are brought together or
products removed at the interface. Apart from such dynamic factors, the rate of a heterogeneous
chemical reaction is influenced by the interfacial area between the phases, the extent to which
reactants and products are absorbed at the interface, and the specific catalytic activity the inter-
face may possess. The reaction velocity is normally represented by a generalized equation of
the form:
da
U 7' - -d = tp [K F (a, a2 . . . a ) as]
t n n
where:
123
-------
U = reaction rate, expressed as moles of reactant A which react per unit
volume and time
a = concentration of reactant A per unit volume
t = time
-------
available and that the conditions used in the laboratory stripper were not optimized. For these
exerpiments, quantities of unstripped nitrose solution were accumulated during scrubber opera-
tion and retained without going through the stripper. The nitrose solutions were subsequently fed
to the stripper (scrubber and scrubber lines operating only as a means of heating the retained
solution to operating temperatures) at reduced liquid and air flow rates simulating the L/G and
space velocities used in the laboratory reactor. These conditions are noted in Table XXVI and
the results presented in Tables XXVII and XXVIIL
As can be seen from the data of the two experiments, average nitrosylsulfuric acid
stripping efficiencies averaged 99% + for Experiment 1 and 95% for Experiment 2. Excellent
total material balance agreement was obtained in the second run where more suitable liquid flow
and temperature control was maintained. Although of limited operating duration, the material
balance agreement of the second run provides considerable credibility to the data obtained.
Visual examination of the catalyst after the experimentation showed no particulate deterioration
even though the catalyst had been subjected to greater than 200 hr of operation.
Although the operating conditions of the stripper have not yet been optimized, examina-
tion of the data in Tables XXVII and XXVIlI suggests two disquieting factors which may influence
future work on the stripper:
1. Significant quantities of NO (averaging 20 to 34% by volume of the total NO
x
removed from the acid) were present in the effluent gas.
2. The concentration of NO in the exit gas averaged 1.1 to 2.0% by volume.
x
Analysis of the data makes it apparent that catalytic denitration of the nitrose solutions
did occur. If the stripper in the experiment described in Table XXVII were denitrating the acid
only by thermal decomposition of the nitrosylsulfuric acid, the maximum concentration of N203
in the effluent gas would be given by the equilibrium vapor pressure. Under the conditions of
this experiment, this would be equivalent to about 8500 ppm of N203 or 17000 ppm of total NOx'
In the second experiment (Table XXXI!), the equilibrium vapor pressure would have been about
4500 ppm of N203 or 9000 ppm of NOx' In both cases, the amount of NOx given off was
slightly higher than this.
It should be noted that, although the total NO evolved was significantly higher than the
x
equilibrium vapor pressure, the amount of NO was consistently lower than would be expected
from equilibrium considerations while the N02 was consistently higher. This tend to verify the
contention that NO is emitted only because of its vapor pressure while the N02 is mainly being
formed by catalytic oxidation of the nitrose.
Fig. 45 is a schematic diagram of a stripper that would effectively denitrate the nitrose
and provide an effluent gas that would contain a high enough concentration of N02 to permit the
economical production of nitric acid while containing a minimum amount of NO. This would be
accomplished while conforming to the results of both the laboratory experiments and the mini-
plant operation. The stripper would be operated at a lower L/G with the bulk of the gas passing
through the column being recycled back to the bottom of the unit. Fresh air would be fed to the
bottom of the column as in previous operation.
125
-------
Table XXVIL Miniplant Stripper Verification Experiment No.1
Acid Air Stripper Effluent
Pumping Flow,
Time Rate, eel m in Temperature, 0 F Gas, ppm Effluent Acid % HNS05
min. cc/15 min. at STP Top Middle Bottom NO N02 % H2S04 % HNSO Denitrification
5
*
o 0.667
15 623 4732** 134 243 224 4400 12,000 84.04 0.011 98.3
30 644 4732 282 298 262 4900 24,200 82.93 0.008 98.6
.... 45 616 4732 294 324 286 4575 23,200 82.12 0.004 99.4
1>.:1
CJ) 60 456 4732 246 310 286 3350 12,400 82.52 0.004 99.4
75 480 4732 172 290 261 2725 9,400 81.71 0.004 99.1
Average: - 3990 16,240 82.66 0.006
Total acid pumped: 2819 cc
*
Initial HNS05 Concentration
**0. 167 SCFM
-------
Table XXVIIL Miniplant Stripper Verification Experiment no. 2
Acid Air
Pumping Flow, Stripper Efflu ent
Time Rate, eel m in Temperature, 0 F Gas, ppm Effluent Acid
min. cc/15 min at STP ------- ----------- ---------------
Top Middle Bottom NO N02 % H2S04 % HNS05
*
o 4732** 188 282 276 78.2 ( 0.385*
~ 0.374*
0.374
15 1000 194 270 268 2600 3650 78.2 0.011
30 780 245 258 258 3275 5350 78.2 0.011
45 913 268 248 242 3275 6500 78.2 0.007
60 900 275 248 236 3525 7200 78.2 0.048
75 870 290 258 240 3850 8100 78.2 0.055
90 915 296 270 250 4200 8800 78.2 0.037
105 870 292 282 260 4400 9200 78.4 0.015
120 865 294 292 282 4325 8800 78.4 0.015
135 910 295 302 296 4400 8700 78.4 0.018
150 935 290 308 312 4250 8200 78.4 0.011
165 910 291 310 320 4325 8000 78.2 0.007
180 905 290 310 320 4200 7650 78.2 0.007
Average 898 3885 7513 78.2 0.0201***
Total Acid Pumped: 10773
*Initial HNS05 Concentration
**0.167 SCFM
***Equivalent to 94.8% denitration
The advantages of this approach to the stripper are that the recycling of the gas stream
permits the buildup of the concentration of N02 while operating the column at low L/G. At the
same time, the concentration of NO in the effluent gas stream is limited to the equilibrium
vapor pressure, and it is unlikely that the conditions at the top of the stripper would approach
equilibrium. (Laboratory experience suggests that effluent concentrations would be in the range
of about 50% of the equilibrium concentration.) This would mean that the gas leaving the stripper
would have a composition of about 7 to 10% N02 and about 0.4% NO. This amount of NO could
easily be accounted for in the total process system by recycling an equivalent amount of N02
to the reactor loop to permit total absorption of NO in the scrubber on a 1: 1 NO?: NO basis.
x "
127
-------
To
reactor.
10% N02
"'0% NO
6000 F
Preheat air
Fig. 45. Modified catalytic stripper - recycle concept
128
Acid
Charcoal
catalyst
Scrubber
acid
Charcoal
catalyst (2900 F)
Oenitrated acid
-------
There is one problem that might be encountered in the operation of a stripper as shown
in Fig. 45 and that is the possibility of the acid leaving the column picking up some NO from the
x
concentrated gas stream entering at that point. This is somewhat unlikely because the stripper
column is very short and should be very inefficient as a scrubber, but the possibility can be
anticipated.
There are two approaches to this problem which would prevent the loss of NO and conse-
x
quent contamination of the product acid. First, another short, charcoal-packed stripper section
could be added at the bottom of the column shown in Fig. 45. In that section, the effluent acid
would come in contact with fresh air at a very high L/G and could be effectively denitrated in
the usual manner.
Another approach would be to inhibit the absorption of NO by reducing the amount of NO
x
in the recycle gas stream. If the NO is being absorbed only in equimolar quantities of NO and
x
N02' then the maximum absorption would be all the NO plus the equivalent amount of N02 or
about 0.2% of each. To reduce this absorption, the NO concentration can be decreased by
oxidizing it before it reenters the stripper. This could be accomplished by catalytic air oxida-
tion.
There is considerable evidence in the literature that NO can be oxidized catalytically at
relatively high temperatures at rates almost one thousand times greater than the homogeneous
reaction. Burdick2 shows that activated charcoal, in particular, is an excellent catalyst in the
air oxidation of NO. This suggests that a way to reduce the amount of NO in the stripper recycle
gas stream would be to add a catalytic oxidizer section using charcoal as the catalyst.
It is clear that more work is necessary in order to fully define the configuration and
operating characteristics of the catalytic stripper. The first order of business is to determine
the optimum operating conditions: temperature, L/G, space velocity, bed depth, etc. When
these parameters have been optimized the stripper can be modified to produce a product gas of
the proper concentration of NO and N02.
11. Miniplant operation conclusions
As previously noted, the miniplant effort is the result of less than 1 month of
round-the-clock operation on an integrated and semi-integrated basis. All phases of the work
requirements were completed. Based on the experimental work performed, the following con-
clusions can be drawn:
1. All systems invoIced in the Catalytic Chamber Process were shown to be
technically feasible.
2. A 84 to 97% oxidation of S02 in the concentration range of 3000 ppm (0.3%)
was found to occur at 3000 F in approximately 20 to 23 sec. This oxidation time was somewhat
higher than anticipated from previous laboratory studies.
129
-------
3. Operation of the high temperature scrubber was verified and good scale-up
agreement was obtained. The theoretical number of transfer units was calculated to be about
4.4, whereas extrapolation of the experimental data showed a requirement of about 3.7 units.
Calculations of HTU indicated a requirement for 8.3 ft of column per transfer unit. Gas side
mass transfer coefficients were calculated to be 1.18 moles/hr/ft2/ atmosphere. These data
are sufficient for initial construction of a column to demonstrate the desired absorption; i. e.,
a 30-ft packed column to absorb 97% of the NO in the feed stream at l/G ratios used. The
x
column is well within the limits of feasibility from size and material considerations. Data was
obtained on gases containing no S02'
4. The catalytic stripper was operated and demonstrated to denitrate the scrubber
acid at greater than 97% yields, as required for suitable stripper operation. Preliminary scale-
up criteria for the stripper have been obtained. Denitration has been found to yield sizeable
quantities of nitrogen dioxide and less than the equilibrium amount of nitric oxide.
5. On the basis of present results, it is desireable to continue to study the
Cat'llytic Chamber Process on the current miniplant scale. Stepwise, this would entail:
(1) modification of the miniplant equipment as required, (2) statistical optimization of operating
and design conditions, (3) verification of commercial feasibility and (4) pilot plant design.
130
-------
VIL ACKNOWLEDGEMENTS
We gratefully acknowledge the assistance of Irving Frutkoff, Mourad Ghobrial, Robert
Fragala, and David Cogley in performing the experimental research and in preparing this
report. We also acknowledge the advice and assistance of William Elder and Gerald
McGlamery of the Tennessee Valley Authority. The contributions of Stanely Bunas of the
Office of Air Programs of the Environmental Protection Agency are acknowledged with gratitude.
131
-------
VIII. REFERENCES
1.
A. C. Stern, "Air Pollution, " Vols. 1 and 3, Academic Press, New York, 1962.
NAPCA Report, Sulfur Oxides Pollution Control, PB 180769, April 1968.
2.
3.
Anon., Chemical and Engineering News, July 8, 1968.
D. Bienstock, J. H. Field, S. Katell, and K. O. Plants, Journal of Air Pollution Control
Association, ~, 459-464 (1965).
4.
5.
E. Berl and H. H. Saenger, Z. Anorg. Allgern. Chern., 202, 113-134 (1931).
W. F. Giaque, E. W. Hornung, J. E. Kunzler, and T. R. Rubin, J. Arner. Chern. Soc.,
82, 62-70 (1960).
6.
7.
Greenwatt, Inc. Eng. Chern., ,!2, 522 (1925).
A. Hantzch and K. Berger, Chern. Ber., 63, 321-336 (1930).
8.
9.
I. E. Ioshpa and L. E. Zlotnik, J. Appl. Chern. USSR, 29, 937 (1956).
10. T. I. Kunin, J. Appl. Chern. USSR, 27, 248-57 (1954).
11. T. I. Kunin, and N. A. Sorov, J. Appl. Chern.USSR,23, 136-139 (1950).
12. K. Malin, "The Technology of Sulfuric Acid, " Moscow (1941).
13. M. Matsui, J. Soc. Chern. Ind. Japan, 30, 134 (1927).
14. W. J. Mueller, D. M. Forbes, and R. Fort, Angew. Chern., 45, 782-785 (1932).
15. A. Sanfourche and L. Rondier, Cornpt. Rend., 30, 815 (1928).
16. K. Stopperka and F. KHz, Z. Anorg. Allg. Chernie, 348, 58-70 (1966).
17. L. E. Topol, K. B. Oldham and R. G. Adler, J. Inorg, Nucl. Chern., 30, 2977-90 (1968).
18. L. E. Topol, R. A. Osteryoung and J. H. Christie, J. Electrochern. Soc., 112, 861-864
( 1965).
19. A. N. Tseidin and V. T. Yavorskii, J. Appl. Chern. USSR, 39, 698-699 (1966).
20. A. N. Tseidin and V. T. Yavorskii, J. Appl. Chern. USSR, 39, 977-981 (1966).
21. M. L. Varlarnov, J. Appl. Chern. USSR, 23, 127-135 (1950).
133
-------
22. C. L. Burdick, J. Amer. Chern. Soc., ,!!, 244 (1922).
23. R. E. Treybal, Mass-Transfer Operations, McGraw-Hill Book Co., Inc., 1955 .
24. T. K. Sherwood and R. L. Pigford, Absorption and Extraction, McGraw-Hill Book Co., Inc.,
1952.
25. R. E. JohnstOne and M. W. Thring, Pilot Plants, Models, and Scale-Up Methods in Chemi-
cal Engineering, McGraw-Hill Book Co., Inc., 1957.
134
-------
APPENDIX I
PROCESS ECONOMICS
-------
Table XXIX. Estimated Major Equipment Cost*
Item Task No. Size Cost
Cyclone separator Remove coarse fly 3 487,000 dm $ 245,000
ash
Scrubber Remove N201' 3 35-ft diameter 1,500,000
~S04' and f y ash 70 ft high
8 ft/ sec
Stripper Recover N203 2 32-ft diameter 250,000
60 ft high
2 ft/ sec
N03 absorber Recover NO at 52% 44-ft diameter 450,000
HN03 x 60 ft high
Filter Remove fly ash from 3 250,000
acid
Product storage Store ~S04 3 100,000
tank
Exhaust fan Provide differential 3 12 in. ~O 60,000
pressure 87 mm CFH
Gas cooler Cool stripper gas 1 2.5 mm BTU/hr 20,000
2100 ft sq. area
Gas cooler Cool reactor gas 6 124 mm BTU/hr 525,000
105,000 ft sq.
area
Pumps Move acid from 8 3800 gal/ min 50,000
scrubber to stripper
Pumps Move acid from 8 3800 gal/min 50,000
stripper to scrubber
$3,500,000
* 800 megawatt (e) power plant producing 1.44 X 106 SCFM flue gas.
1-1
-------
Table XXX. Estimated Capital Cost Summary*
Item
Factor
1. Major equipment (see Table Ill) 1.00
2. Erection labor and foundation 0.38
3. Piping 0.55
4. Instruments 0.05
5. Insulation 0.03
6. Electrical 0.05
7. Building 0.20
8. Plant facilities 0.05
9. Plant utilities 0.08
10. Receiving, shipping, ete. ~
11. Physical plant costs 2.44
12, Engineering and construction ~
13. Direct plant cost 3.04
14. Contractor's fee 0.10
15. Contingency ~
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 gas.
1-2
-------
Table XXXL Annual Operating Costs*
Raw materials and chemicals
Catalyst replacement
Water (3500 GPM at $0.10/M gal)
Direct labor
Supervision
Maintenance at 0.05 of fixed capital
Supplies at 15% of maintenance
Utilities
Power
Heat
$ 100,000
168,000
108,000
27,600
653,000
98,000
145,700
ot
Total direct cost
$1,300,300
Payroll burden (20% of direct labor and
supervis ion)
Plant overhead (50% of direct labor, super-
vision, maintenance, and supplies)
Packing and shipping
Waste disposal
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)
Total fixed cost
Total operating cost
$1,151,500
230,300
115,150
$1,496,950
$3,670,450
(or $ 1. 64/ ton of coal)
* 800 megawatt (e) power plant producing 1.44 x 106 SCFM flue gas.
t Heat is needed only for startup and process adjustment situations.
1-3
-------
APPENDIX II
NITROSYLSULFURIC ACID ANALYSES
-------
APPENDIX IL NITROSYLSULFURIC ACID ANALYSES
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 NHSO", remaining. This procedure depends on the hydrolysis of the N °
de 2 3
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 equimolar 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 N203 is hydrolyzed.) This will
probably not be a significant problem since very little N02 should be dissolved in the acid. In
any case, any discrepancy in the amount of nitrogen oxides will show up in the complete material
balance around the various stages in the process. 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% nitrosyl sulfuric acid in 98% H2S04 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 N02 fumes which were lost to the analysis. When
diluted with 80% sulfuric acid so that the acid and nitrogen oxide concentrations were lower
(about 2% HNS05)' the temperature did not rise as much during the titration and no N02 was
lost. Care must be taken to add the permangate solution sufficiently slowly so as not to cause
siginficant heating of the nitrosyl solutions. The use of a surrounding ice bath may be em-
ployed.
11-1
-------
APPENDIX III
MINlPLANT EXPERIMENTAL PROCEDURES
-------
APPENDIX IlL MINIPLANT EXPERIMENTAL PROCEDURES
A. Miniplant Operating Concept
The goal of the miniplant operation was to start the plant, achieve steady state and then vary a
few operating parameters to obtain experimental verification of the basic concept of the Catalytic
Chamber Process and to obtain preliminary order-of-magnitude engineering data. Since only 2
weeks were alloted to this task, it was realized that this examination would be very limited in
scope and there would be no time to go into the operation of anyone particular process stage in
great depth. This work was to provide initial data and would be followed up by more extensive
optimization studies at a later time.
B. Startup Procedure
A startup procedure was carefully worked out so that the plant could be slowly brought
to steady state operation without sudden changes occurring in the system that might be dangerous
to either equipment or personnel. Since the integrated plant had never been run it was necessary
that each stage of the plant be brought into operation individually for observation before the next
stage of operation was started up. Table XXXII shows the startup procedure used to begin the
operation of the miniplant.
Table XXXIL Miniplant Startup Procedure
1. Start blower with stack and vent by-passes open. Adjust stack by-pass to give
10 SCFM. Start air through NO oxidizer.
2. Start acid pumps for scrubber and stripper.
3. Start air through stripper. Control to desired level.
4. Turn on cooling water to condensers and gas burner.
5. Start heating system. Set all variacs at 40: reactor, scrubber, stripper and
sample lines.
6 Turn on instrumentation. Allow at least 15 min warm - up for recorder. UV and IR
. photometers are left on at all times. For initial startup allow 2 hr warm-up for
both instruments. Two hours should be allowed for trap cooling bath to reach -40°C.
7. Adj ust heating levels as necessary. Raise heating by increasing variacs. Sample
lines are kept on 40.
8. Turn on gas burner. Allow about 30 min to reach steady state. Readjust air flow.
9. Zero and span instruments.
10. Introduce S02 and NO into system.
11. Close vent bypass, sending gas flow through scrubber. Readj ust all flow rates for
new pressure drop condition. Readjust temperature as necessary.
III-I
-------
C. Safety Considerations
Safety precautions were carefully followed throughout the operation of the miniplant. All
personnel were equipped with protective gear which included acid resistant coveralls, rubber
gloves, safety glasses, face shields, hard hats and boots. Other safety precautions that were
followed included having two operators in attendance at all times, keeping lime available for
acid spills, avoiding placement of tools or other objects on the apparatus where they could fall
and break glass piping, and keeping a gas mask convenient at all times.
The plant was set up so that in case of an accident all the electrical equipment (heaters,
pumps, blower) could be turned off remotely. A self-contained breathing apparatus was kept
outside the plant area in case of a major gas leak. A large exhaust fan was installed in the area
which could change the air in the room within 1 min. Although there were a few minor gas leaks
and acid spills, there was never any occasion for emergency action.
D. Data-Taking Procedures
During miniplant operation, all aspects of the plant were monitored on a regular basis
and operating data taken at specific intervals. This was done, not only to obtain the data for
calculations and future reference, but also to provide a check list for the operators so they
would be forced to observe all operating equipment at regular intervals and make sure operating
levels were maintained throughout an experiment.
Forms were prepared for recording data in addition to a log book wh ich was maintained
for general comments and discussion. Copies of these forms are shown in Tables XXXII
through XXXVI of Appendix IV.
1II-2
-------
APPENDIX IV
MlNlPLANT DATA SHEETS
-------
Table XXXlll
S02:
N02:
NO:
Date:
Span
Zero
Zero
Gain
Page:
Gain
,.....
<
I
-
N02 S02
Air No N203 Low Air
Condition Time Rotameter Rotameter ppm ppm Rotameter ppm Rotameter Rotameter Comments
-------
<:
I
~
ACID SCRUBBER STRIPPER
SCR. STR TOTAL GAS NO AIR AIR ACID S02 --
IN (NO) (STR) SCR STR ML % ML %
! I
I I
!
I I
I
I
i
I
,
I
!
!
I
I
I
1
I
)
,
MANOMETERS
FLOWMETERS
Table XXXIV
ACID ANALYSIS
DATE:
-------
Table XXXV
TEMPERATURES
DATE:
.....
<
i
t.)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
TIME
1. Hot water heater
2. Blower i_nlet
3. Blower outlet
4. Reactor inlet
5. Reactor
6. Reactor outlet
7. Scrubber inlet gas
I 8. Scrubber
, 9. Scrubber gas outlet
I 10. Scrubber acid feed
I
(at pump)
11. Scrubber acid feed
(at top)
12. Scrubber acid outlet
13. Stripper acid feed
(at pump)
14. Stripper acid feed
(at top)
15. Stripper - top
16. Stripper - middle
17. Stripper - bottom
18. Stripper acid reservoi
19. Stripper gas exit
20. NO oxidizer outlet
r
-------
Table XXXVI
......
<
I
,.!>.
TIME 1 2 3 4 5 6 7 8 9 10 11 12 13 14
LOG AND VARIAC SETTINGS
DATE:
------- |