AN. TNVESTEGATTON OF THE ABSORPTTOR
0F SULFUR DIOXIDE BY AN AMMONIACAL
SOLUTION EN. A PACKED COLUMN.

Thesis for the Degree of M. S.
MICHIGAN STATE UNIVERSTTY
STEVEN R. AUVIL
1971'

9;";me

TUTTITTTTTTTTTWT W” L m 1. R Y “T

3 1293 00830 8110
M bigan Stat

Univeis it)!

    
   

 

gm E‘ -:
110:8“sz ‘

   

‘89" RNDBE .RY INF.

 

 

 

-.—___ .. ‘-

PLACE n RETURN BOX tc rorhavo this checkout from your record.
TO AVOID FINES Mum on or baton duo due.

“[3
fiCflT—T
_J-L_J
E—TLJ—j
[—7 i 7

MSU To An Afflrmdivo ActionlEqunl Opportunity Institution
ammo-9.1

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tl

 

 

 

 

 

 

 

ABSTRACT

AN INVESTIGATION OF THE ABSORPTION OF SULFUR
DIOXIDE BY AN AMMONIACAL SOLUTION IN A PACKED COLUMN

By

Steven R. Auvil

A study was made of the absorption of sulfur dioxide
by a solution containing a mixture of ammonium sulfite and
ammonium bisulfite in a continuously operated packed column.
The column used was 3.75 inches inner diameter packed to a
depth of 4 feet with %-inch unpolished porcelain Raschig
rings. Operating conditions used in this work are summarized
in Table 1. The column was operated centinuously by
maintaining a large liquid recycle rate, 94-97%, with the
remainder taken from the system to maintain a constant
sulfur and ammonium ion content. Ammonia and water were
added to the recycle liquid to maintain solution pH and
volume. No oxidation inhibitor was used to prevent the
formation of sulfate ion.

The equilibrium relationships used for the (NH4)zSO3-
NHAHSO3-H20 system were those derived by Johnstone (1) and
later modified by Chertkov (2) who gave an additional
correction for the presence of sulfate ion.

Data were obtained by measuring sulfur dioxide

absorption by first holding the gas rate constant and

Table 1: Operating Conditions for Experimental Tests

Gas rates 272-544 lb/hr-ft2
Gas relative humidity 75-80%
Oxygen content in gas 10%
Sulfur dioxide concentration

in gas 2100-2300 ppm
Gas and absorbent temperature 5000 2
Liquid rates 300-560 lbs/hr-ft
Liquid/gas mass flow rate .56-1.36
Ammonium ion concentration 9-16 moles/100 moles H20
(S/C)6ffeCtiVe 062’076
Sulfate concentration]

total sulfur 5-16%

varying the liquid rate. Secondly, the liquid rate was
held constant and the gas rate varied. Along with the
variations observed in sulfur dioxide absorption,
variations were seen in the ammonia losses and the
oxidation of dissolved sulfur dioxide to sulfate ion.

It has been reported by Chertkov (3) in the range of

(3/0)

dioxide is gas phase controlling and the mass transfer

effective studied that the absorption of sulfur

coefficient can be represented by the following equation:

kg d
0 e = .0035 Re Re '4
0- GL

 

That is. it is proportional to the first power of the gas
phase Reynold's number and proportional to the four-tenths
power of the liquid phase Reynold's number. The results of
this work showed very good agreement with this relationship
with the only difference being in the constant term. The
losses of ammonia were correlated to this same type of

expression and found to depend on the gas phase Reynold's

number to the .62 power and the liquid phase Reynold's
number to the five-tenths power.

The formation of sulfate ion was compared to that
predicted by a relationship developed by Chertkov (4)
with very good agreement. This was not expected, however,
since Chertkov's equation was developed for industrial
packed scrubbers using gases and liquids contaminated
with compounds which promote oxidation. This work was
done with clean gases and liquids and hence was expected

to show less oxidation.

 

 

AN
DIOXIDE;

AN INVESTIGATION OF THE ABSORPTION OF SULFUR
DIOXIDE BY AN AMMONIACAL SOLUTION IN A PACKED COLUMN

By

Steven R. Auvil

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering

1971

To Pam

iii

grins
.. ~ ”2 II II

5%

r—V—<

"J

E . ‘1 ' I" v;
m' a
m ' W

 

The a
Dr. Bruce
for his gt
Work and ;
eXtended t
financial

The a
Pollution
Health, f.
equipment
Valued,
Michigan
0f the Sc

of Parts

ACKNOWLEDGMENTS

The author would like to express deep appreciation to
Dr. Bruce W. Wilkinson, Department of Chemical Engineering,
for his guidance and encouragement during the experimental
work and preparation of this thesis. Appreciation is also
extended to the Lansing Board of Water and Light for their
financial support.

The author is deeply indebted to Mr. John Shaffer, Air
Pollution Control Section, Michigan Department of Public
Health, for the many hours spent helping build and run the
equipment. His assistance and suggestions were greatly
valued. Special thanks is given to Mr. Allen Croff,
Michigan State University, for his assistance in the analysis
of the solutions and to Mr. Don Childs for the fabrication

of parts for the equipment.

iv

 

 

Introducti:
Theory....

EXperiment.‘
The F
The F
Experiment
Data, , . . . .
Results . . .
Anal)
Anal}
Anal}

Analysis .

TABLE OF CONTENTS

Introduction..........................................l
Th80fYooooooooooooooooooone.no...oooooooooooooobouooooé

Experimental Equipment...............................16
The Flow Of the Gas Stream......................22
The Flow Of the LiQUid Stream...................Z3

Experimental MQEhOdSoooooooooooooooo00000000000000.0025

DataOOOOOOOOOOOIOOO0.0....00.000.000.000...0.00.00.0031

Results..............................................35
Analysis of the Absorption of 802...............36
Analysis of the NH3 Stripping...................41
Analysis of the Oxidation of 803: to 804:.......46

Analysis of Results..................................49
$02 Absorption..................................49
NH3 Stripping...................................SS
Solution Oxidation..............................56

conclusions and Recommendations......................60
Basic Observations..............................60
Suggested Research..............................63
Suggested Experimental Improvements.............64

.Bibliography.........................................66

>‘u ' a)
x . a.

 

 

Appendix A.
Analys

I

I]

II]

1\

Appendix B.
Appendix C
Appendix D
Appendix E

Appendix F

Appendix A...00000000000000.00000000000090:0000.00.00.68
Analysis of Liquid Solutions
I. Determination Of (NH4)HSO3000000000000.068

11. Determination Of (NH ) SO 0.00000000000069
4 2 3+

111. Determination of Total NH4 Present.....70

IV. Determination of $04: Present...........7l

Appendix B............................................73
Appendix C............................................77
Appendix D............................................81

Appendix E............................................85

Appendix Fooouoooooooootoooooooooooooo00.000000000000088

vi

r?

I. 1.th .

 

 

Table 2.

Table 3,

Table 4,
Table 5.

Table 6,

Table 7.

Table 8,

Table 9.

l

Table

Table

Table
Table

Table

Table

Table

Table

3.

4.

5.

8.

9.

LIST OF TABLES

Values of (S/Ceff)out Calculated from
Equation 9 with 80% of the 802
Removed from a Gas Stream
Containing 2000 ppm 802 and an
Inlet Liquid with S/Ceff = .7.........13

Typical Results from the Concentration
vs. Time Measurements Taken during
EaCh Experiment.......................29

Experimental Data Taken at a Constant
635 Rate 0f 3 ft/SeCoooooocoooooooo00033

Experimental Data Taken at a Constant
Liquid Rate of 185 ml/min.............34

HOG Values for the Absorption of SO2
Calculated from the Experiments

at a conStant Gas Rate................37

HOG values for the Absorption of $02,
Calculated from the Experiments

at a CODStant LiQUid Rate.............38

HOG Values for the Stripping of NH3
Calculated from the Experiments
at a ConstantGas Rate................42

HOG Values for the Stripping of NH3
Calculated from the Experiments
at a COUStant LiQUid Rat3000000000000043

vii

(

Table 100

Table 11.

 

. u, .u u E g. . ,
15...» II I 1...? . '1‘“
{Handy ,. L
. .

Table 10. Comparison between Observed and
PrediCted Oxidation RateSooooopoooooo48

Table 11. Properties of the Various Solutions
Used during the Experiments..........84

viii

 

a:

 

 

Figure 1.

Figure 2 .

Figure 3.
Figure 4 .

F1Sure 5'

FiSure 6

Figure 7

Figure 1

F1gure

Figure

Figure

Figure
Figure

Figure

Figure

Figure

Figure

Figure

5.

6.

9.

LIST OF FIGURES

Ammonia Scrubbing: Regeneration or
PdeUCtion 0f Ammonium SUlfateoooooooo3

Equilibrium Partial Pressures of SO2
and NH3 over Ammonium Sulfite-
BiSUlfite SOIUtiODSo00000000000000.0008

Front View of the Experimental

EQUipmentooooooooooo0000000000000000019

Rear View of the Experimental

EQUipmentoo00000000oooooooooooooooooozo

Flow Sheet of the Experimental
Apparatus Used for the
Absorption Of $02.0000000000000000.0021

A Plot of Log HOG vs. Log R8 for the
Absorption of Sulfur Dibxide at
a conStant Gas Rateooooooooooocoooooo39

A Plot of Log HOG vs. Log G for the
Absorption of Sulfur Dioxide at
a COHStant LiQUid Rate...............40
A Plot of Log HOG vs. Log R for the
eL
Stripping of Ammonia at a
Constant Gas Rate.0.0000000000900000044

A Plot of HOG for the Stripping of
Ammonia at a Constant Liquid Rate....45

.ix:

A
'5

a w...-
. ..

 

 

Figure 10.

Figure 11 .

Figure 12.

Figure 1 3.

Figure

Figure

Figure

Figure

10.

11.

12.

13.

The Operating and Equilibrium Lines
Depicting the Absorption of 80%-of
the Inlet SO2 from the Inlet Gas.....52

A Plot of N0G vs. S/Ceff during the
Integration up the Column............53

Computed Operating and Equilibrium
Lines for the Stripping of NH3
During the Simultaneous Absorption
and Partial Oxidation of 802.........57

A Plot of Percent 802 Absorption, NH3
Lost and Percent 802 Absorbed at that
Point that is Oxidized vs. Position
in the Column........................6l

' "xiii—421

,_ .
I; an: 4:“. “‘J
‘m

 

The
of sulft
With the
being c;
require.
chemisr
Contint
This 11
by the
emissi
Plants
10W St

1
Syste1
Such
DH'S
Works
howe‘
W1 th
eV6111
to p
add:

£011

INTRODUCTION

The adaptability of an ammonia process to the removal
of sulfur dioxide from gas streams has long been recognized,
with the earliest work recorded, other than test tube scale,
being carried out in Japan (4) in 1926. The demands
required by the system for a thorough understanding of the:
chemistry involved and evaluation of various parameters have
continually intrigued chemists and engineers to the present.
This interest has been further promoted in the last decade
by the great concern over air pollution by sulfur dioxide
emissions from power plants. The emissions from power
plants will steadily be increasing since the availability of
low sulfur content coal is dwindling.

A large amount of the technology of the scrubbing
system has been developed with packed absorbers. Parameters
such as variations of the liquid and gas rates, solution
pH's and packing depths have been investigated. Two major
works, Tennesse Valley Authority (5) and Chertkov, USSR, (6),
however, are essentially the only ones which were published
with enough variations of the system's parameters to allow
evaluation for design and the establishment of an equation
to predict the gas phase mass transfer coefficient. In
addition, a quantity of work has been done with the
following types of scrubbers: sieve tray, venturi, spray,

l

A.::2§

\ ’ i:
£135.»: 3mywq... myriauvswi'

 

mobile-bCd
studied wit

In 655
that must I
effluent g2
there are ;
of gas evo
the follow

1. L

2. i

Fig:
with the
Or the r
dioxide,
extremes

by 1t3e1

2
mobile-bed and orifice. These systems have also been
studied with variations of the above mentioned parameters.

In essence, any type of scrubber used, especially one
that must be able to handle the tremendous amount of
effluent gases from a power plant, where, for example,
there are approximately two million cubic feet per minute
of gas evolved from a 1000 megawatt boiler, should have
the following characteristics (7): ,

1. Low pressure drop (much less than 12 inches water)

2. High efficiency for the removal of sulfur
dioxide (90%) '

3. A means to control the formation of sulfate,
eSpecially if it is not wished as an end
result

4. Resistance to plugging from fly ash

5. High turn down volume (ratio of design gas volume
rate to minimum volume rate)

6. Provision for stage wise contact of gas and
liquid to get good sulfur dioxide absorption
with minimum ammonia 1039.

Figure 1 contains a flow sheet of an ammonia process
with the final step being.the oxidation to ammonium sulfate
or the regeneration of the absorbent which yields sulfur
dioxide. These two end results essentially represent the
extremes of the process with each product accepted as an end

by itself or one to be used as an intermediate in some other

Now

 

XUNUW OH

 

filliei'kiivr'fifi
W” Hhh NI. . . . II,

‘

oquHam enacoee< mo coauonuoum no coauouocowom

Aaaomv oommxomzv

MHC

mmz

 

nouaonxo

 

 

 

 

 

coaumuocewom

a

 

 

lfll

 

momNAemzv + nommomz

o~ml1.

mmzn...

Ho>oaom
roe

 

.wannsuom wacan¢

 

 

 

 

 

 

 

oHozoom

on

 

 

Hennauom

 

 

H

 

aboum OH

 

 

 

 

 

uoaooo
moo

 

.fi ouowau

TI.

 

uoaaom
Bonn new

 

 

producti
fertiliz
The
both to
Past wo:
The son
operate
of sulf'
small 5
the sul
the qua
the Syg
ammonii
reWell
0f the
nature
that n
I

and t1
Produ.
for '1
be th
Of til
Sste
In th
“Mst
the

a

4
production process, whether it be in the production of
fertilizers or the manufacture of sulfuric acid.

The process consists of a cooling tower that serves
both to cool and remove particulate matter from the gases.
Past work (8) indicates a spray chamber will suffice.

The scrubbing tower, regardless of the type used, will
operate with a large recycle rate due to the small amount
of sulfur dioxide contained in a pound of flue gas. A
small stream will be continuously withdrawn to represent
the sulfur dioxide absorbed in the scrubber and represents
the quantity that must be processed in the end step of

the system. If the end result wanted is the formation of
ammonium.sulfate, ammonia and water must be added to the
recycle to maintain solution pH and volume. Regeneration
of'the effluent liquid stream from the scrubber will, by its
nature, mean that the only additions of ammonia and water
that must be made are those necessary to counter losses.

At present, the market for ammonium sulfate is poor
and techniques to use it as an intermediate in the
production of nitric phosphate fertilizer are being persued,
for instance, by TVA (9). The regeneration scheme looks to
be the best alternative due to the versatility and usability
of the sulfur dioxide evolved. .This allows the absorbent
system to be a “closed loop" With very little make-up cost.
In this case, however, the oxidation of sulfite to sulfate
must be controlled, since continued oxidation would destroy

the absorbent. Regeneration, however, presently lacks the

75-6. .4‘
.‘..
.

\‘S—m "cm“ “1‘“

I;
Afiw.au

 

technolog
economica
It w
demonstra
data take
appearing
out using
the flow
regenerat
theref ore
necessita
Water to
The
the heigh
dioxide a
°f 838 an
°°Curing
for the c

predicted

5
technological advances of the scrubber system for any
economically feasible operation.

It was the purpose of this study to build a
demonstration model of the scrUbbing process and to compare
data taken from its operation to that predicted by equations
appearing in the literature. The investigation was carried
out using a laboratory scale packed column which followed
the flow scheme presented in Figure 1. Neither the
regeneration nor the oxidation step was investigated, and
therefore, the drawn-off stream was merely collected, thus,
necessitating the continuous addition of ammonia and
water to the system.

The Specific items of interest were the variations of
the height of a transfer unit for the absorption of sulfur
dioxide and those for the los5es of ammonia as a function
of gas and liquid rates. Data on the rate of oxidation
occuring in the column were not taken for correlation but
for the observation of trends and for the comparison with

predicted values.

Ihef

 

Z mtg-1-? V
5‘...“ _.-

vapor pres
fl mixture of
E

major wo r1

 

following

me thes
can be WI

H20 over

' Wile

 

THEORY

The first, and accepted, study of the equilibrium
vapor pressures of 802 and NH3 over solutions containing a
mixture of NH4HSOB, (NH4)2803 and (NH4)ZSO4 appeared in a
major work by Johnstone (10). He established that the

following equilibria could be written:

: ~++ ‘
802(g) + H20 + 303 «- 21—1503

" H : 1’ ° +

From these equilibrium expressions the following equations
can be written giving the vapor pressures of $02, NH3 and

H20 over the solutions:

 

 

 

2
(ZS/C -l)
Pso = MCeff (l-sigf ) (1)
2 eff
(l-S/C )
“ . 100
PH20 = Pw(100+c+s+A) - (3)
- Where P P

NH , and PH 0 = the partial pressures of
‘ 3 2
802, NH3, and H20 over the
solutions in mm Hg
Log M = 5.865 - 2369/T
Log N = 13.680 - 4987/T
T = Temperature in 0 Kelvin

41'-
u.

r
‘ i‘m "flaw-sat. and

 

List of V6

A fu
for the ;
presence

expressic

List of variables (cont'd)

A a moles $04=/100 moles H20
c = moles NH4+/100 moles H20
Ceff a C-Zfi = moles reactive
NH4 /100 moles H20
moles of (SO3=+H802')/
moles reactive NH4

S/C

eff

P = vapor pressure of pure
water at temperature T
A further correction was made to the above equation
for the prediction of the 802 vapor pressure due to the
presence of sulfate by Chertkov (11). The following

expression was derived from his data.

( ) ‘( ) Ceff + A (4)
P true a P calc'd -
so2 SO2 Ceff

 

Where P (calc'd) is the value of P computed
$02 302

by the expression by Johnstone, given above.

In the expressions derived by Johnstone, the term given

by Ceff(ZS/Ceff-1) is the H803" concentration and the term
given by Ceff(1'S/Ceff) is the $03: concentration. It can
be seen that the value of S/Ceff ranges between .5

(pure $033) and 1. (pure HSO3'). Observation of the vapor
pressure equations show that a value of S/Ceff close to .5
gives the smallest vapor pressures of SO2 and the largest
‘vapor pressures of NH3. A value of S/Ceff close to 1.
gives just the opposite results. Figure 2 contains a plot
502 and PNH3
tabove equations, for two values of Ceff’ which illustrates

of P vs. S/Ceff at 50°C, prepared from the

 

.4 1

03 "’

Asa—v

2

thum mmhnu H Md Uhflnw

.1

0.0

Fig.

 

1t.! -DRikN ”mi
. a n

u wand”! .

Partial Pressure (mm)

.4

.3

.2

.1

0.0

 

 

Curves a are for

C = 5 moles NH4+/
100 moles H20

Curves p_are for

C e 13 moles NH4+I
100 moles H20

IP

4»

u.
D

.65 .7 .75 .8 .85 .9

S/Ceff 2 moles SOZImole reactive NH4+

'Figure 2. Equilibrium Partial Pressures of 302

and NH3 over Ammonium Sulfite-bisulfite

Solutions

Temperature a 50°C
304‘ concentration a l mole/100 moles

H20

731.444"
.
' wk

 

 

dds faC'

The

that in

low soon
(90% of
with any
This lea
literatw
stage t'
have a
Promote
into a
and a 3
first 1
0r low
1

of 802
contrc

expreg

this fact.

The foregoing observation leads to the conclusion
-that in a single stage scrubber having a value of S/Ceff
low enough to achieve a high degree of SO2 absorption
(90% of 2000 ppm inlet gas stream) necessarily conflicts
with any attempt to keep ammonia losses to a minimum.
This leads to the conclusions drawn in the current
literature of using a multi-stage absorber. In the first
stage the gas would come in contact with liquid that would
have a relatively low value of S/Ceff (about .65) to
promote very good SO2 recovery. The gas would then pass
into a stage having a greater value of S/Ceff (about .9)
and a low value of S/Ceff to trap the lost NH3 from the
first stage but still maintain the SO2 content in the gas
or lower it further.

It was reported by Chertkov (12) that the absorption
of $02 by a NH4H803-(NH4)ZSO3 solution was gas phase

controlling in the range of S/C up to .78 at 30°C and

eff
that the gas phase mass transfer coefficient could be

expressed as:

 

 

G e 4
. .0035 R R ° (5)
DG eG eL .
36OOW d PG
. Where R = Gas phase Reynold's no. = 0 e r
eG Wang
4QPi
R 2 Liquid phase Reynold's no. = -——-

prL
d = equivalent diameter of the packing (ft)

23?).
-. ",1

. if

 

 

List of

The cor
Chertkc
of 30 1

Si
Soluti.
(less .
t0 rea
Surfac
mas 0n
the SO
be tru

List of

10
variables (cont'd)
D = diffusivity of 802 in air (ftZ/hr)
kG 2 gas phase transfer coefficient (ft/hr)
. 96 a density of the gas and liquid (lbs/ft3)
a viscosity of the gas and liquid (lbs/ft-hr)
= void fraction of the packing

JL
<1»
w = superficial gas rate (ft/sec)
Q = liquid rate (ftB/ftZ-hr)

f

= Specific surface of the packing (ft2/ft3)

ix
The constant term contains the Schmidt number (P.%.) which
G G

Chertkov assumed constant at 1.4 over the temperature range

Of 30 to 1000C.

Since the equilibrium vapor pressure of 802 over a

solution of (NH4)2803-NH4HSO3 is very low in this range

(less than 12 mm for Ceff=10) and since the $02 is assumed

to react as soon as it comes in contact with the liquid

surface

the conclusion of gas phase controlling seems

reasonable. As S/Ceff gets above .9 the assumption that

the 802
be true
is very

m
discuss

for'the

reacts upon contacting the liquid surface may not
since the partial pressure of SO2 over the solution
large.

the study of absorption columns it is common to

the variations of the height of a transfer unit

gas or liquid phase. Since the absorption in the

above range is gas phase controlling the value for the

overall gas phase height of a transfer unit (HOG) can be

Ian-vain». ..'.

 

 

written E

or

Ch

fins EX}
a trans:
sYStem
In
are use

may be

then me

Wm

The QM

11
written as

H =—-——--—— (ft) (6)

0G

specific surface of the packing (ftZ/ft3)

§
(0
"S
(1)
DJ
ll

G a superficial gas mass velocity (lbs/hr-ftz)
k = gas mass transfer coefficient (ft/hr)

9 2 gas density (lbs/ft3)

or substituting the expression for kg given by

Chertkov gives

JULGQ -.‘4

= I , R
0G .OJJSPDGa eL

 

H (7)
This expression says that the overall gas phase height of
a transfer unit is independent of the gas rate for the
system in question.

In absorption processes in which lean gas mixtures.
are used, as is the case in this work, the values of “QC
may be determined by using the following relationship and

then making a correlation to an expression such as (6).

Hm a Z/NOG (8)

Where NOG = number of overall gas phase transfer
units

2 a depth of the packing (ft)
3’3?
, _l__2 ' . (9)

N a
OG (y‘yfl )1!“

3:1, 3'70 = inlet and exit 802 mole fraction in gas

The quantity in the denominator of the above expression for

12
NOG is the log-mean driving force for the transfer of 802
from the gas to the liquid based on the SO2 inlet and exit
concentrations in the gas and the equilibrium concentrations
of 802 in the gas over the inlet and exit liquid streams.
The assumptions being made are that there exists a linear
relationship between the concentrations of 802 in the gas
and the liquid as it passes through the column and that
there exists a linear relationship between the equilibrium
gas composition of $02 and the liquid composition under

isothermal conditions. The log-mean driving force is

given by the following expression:

J

 

 

 

N a: rd N 7': ,
(~,§*) _= (yl-y Q)- (yo- Y1 ) (10)
1'“ 1 0.16" so")
n s
(370- “1’5" .)

3(-

* .
Where'yi , - the 802 mole fraction in the gas in

<2
0
I

equilibrium with the inlet and exit
liquid streams, reSpectively
,¢*

i.e. y = PSOZIPT

PT = total pressure (mm Hg)

Writing an expression for the relationship between
the 802 concentration in the gas and the liquid in the
column allows one to evaluate the validity of the above
assumptions for the use of C§J§*)l . A material balance

m
on the sulfur dioxide gives the following expression:

”N” ‘V
G(yi-y) = (So-S)L

S/Ceff ‘ (S/Ceff)o ‘EEEQ"" (71:7) (11)-

13

Where the subscripts "i" and "0“ refer to "in" and

“out" of the column, respectively.
This expression assumes isothermal operation and does not
take into account the loss of ammonia or the change in S
due to the oxidation of sulfite to sulfate, since these
changes at this point are assumed to be small relative to
the $02 absorbed. Since 3 and '1: are constant through the
column, the relationship between S/Ceff and‘y is seen to
be linear. The variation of the SO2 equilibrium over
solutions versus S/Ceff as seen from Figure 2 can be
assumed linear only if S/Ceff changes by a small amount
upon passing through the column. Table 2 contains the
predictions of Equation 9 for the absorption of 80% of the

$02 from an inlet gas stream containing 2000 ppm 802 and

an inlet liquid with S/Ceff equal to .7.

Table 2: Values of (S/C Calculated from Equation 9

eff)out
with 80% of the 302 Removed from a Gas Stream

Containing 2000 ppm 302 and an Inlet Liquid with

 

G/L peff 9 13 15

.50 .708 .706 .705
1.00 .718 .712 .711
1.75 .727 .718 .716

It is evident from Table 2 and comparison with Figure 2
that the assumption of a linear equilibrium relationship

is valid.

The oxidation of sulfite-bisulfite solutions has been

14
the object of many studies with as many different results
presented. The most recent major study was performed by
Chertkov (13) who proposed a mechanism for the oxidation
rate and was able to explain the observations of many

investigatbrs. The mechanism proposed by Chertkov is as

follows:.
Sulfite oxidation 02+ZSO3:-9'ZSO4: a
Bisulfite oxidation 02+2H303'—-§I2HSO4' b
ZHSO4 +SO3- ---:> 211803 +2804“ c

where reaction "c" occurs because bisulfate is a much
stronger acid than bisulfite (l<==1.2x10"2 compared to
6.4x10'8). The net effect of "b" and "c" is for the
bisulfite concentration to remain constant and S/Ceff to
rise. Chertkov found that for a total concentration of
sulfite-bisulfite less than 2-3 gram-moles/liter the
oxidation was reaction controlled, and above these v
concentrations the oxidation is oxygen diffusion controlled.
He found that a maximum in the oxidation rate occurred at
a sulfite-bisulfite concentration of 2-3 g-moles/liter for
various values of S/Ceff. In all cases he found the
oxidation to increase steadily with increasing values of
S70eff above .73.

. Chertkov, in a later work (14), collected data on the
oxidation kinetics of sulfite-bisulfite solutions under
industrial conditions and by correlation developed an
emperical equation whereby it is possible to calculate

the oxidation rates of various sulfite-bisulfite solutions.

15
The equation was presented as

G 0.8 Q’7a(S/Ceff)6 (12)
O2 ‘5’ ”‘L

 

where a = T/SO, TM: temperature in oC

G0 = grams oxygen absorbed/hr-m2 contact area
2

S/Ceff as defined previously

9 == solution density, kg/m3

.LL 2 solution viscosity, kg-sec/m2
All data were taken from packed absorbers in which the mass
transfer surface was arbitrarily chosen to be the geometeric
surface of the packing. Chertkov warns that the equation
was developed using contaminated gases and solutions under
industrial conditions and may not agree with laboratory

StUd 1.88 o

EXPERIMENTAL EQUIPMENT

The eXperimental equipment was designed with the

following requirements and objectives in mind:

1.

2.

3.

5.

6.

The column was to be 3.75 inches in diameter and
was to be packed with Raschig rings. .
Liquid and gas temperatures would be kept at 500C,
to simulate actual industrial operating conditions.
The gas used would contain approximately 10% 02,
.2% 802, and the remainder N2.

The column was to be designed to handle gas flow
rates of 2-3 ft/sec and liquid rates which
correspond to values of L/G equal to approximately
one.

The column system would be divided into sections
so that each section could be operated
independently.

The equipment should be mobile enough so that it
could be easily moved to an actual power plant to

test its operation on flue gases.

' Calculations of the maximum allowable velocities (14)

with L/G=1 showed that Specifying a column diameter of

3.75 inches and packing it with %-inch porcelain Raschig

rings would permit operation at gas rates of 2-3 ft/Sec.

16

17

These gas rates corresponded to 50 and 75 percent of the
maximum allowable.

To decide on a reasonable height of a column that
might be expected to absorb 90% of the $02 from a gas
stream containing 2000 ppm, calculations were made using
Chertkov's mass transfer coefficient in Equations 6, 7, and
9 at a temperature of 50°C. The calculations were made
assuming a value of C=9 moles per 100 moles H20 with inlet
and exit values of S/Ceff equal to .68 and .7, reapectively.
A value of L/G equal to 1 was used with G=540 lbs/hr-ftz.
The results of the calculations Showed that under the above
conditions a column ranging from approximately 7 to 10 ft
in height might be sufficient with the height variation
arising from changing the amount of sulfate present from
1 to 2 moles per 100 moles H20.

A total column height of 8 feet was chosen for the
construction of the experimental system. The column was
divided into two 4-ft sections and piped so that they were
in series arrangement with respect to the gas flow and
could either be connected so that there was series or
independent flows of liquid to each section. Preliminary
experimentation with the equipment led to the conclusion
that only one of the columns was necessary to accomplish
the goals of this work (giving absorption of about 75%) and
the use of both columns in any arrangement was a hinderance

experimentally.

18

The columns were constructed from equal 4%-ft
sections of 3.75 inch 1. D. by 1/8-inch wall plexiglass
tubing. The columns were wet packed with constant shaking
of the columns and stirring of the top layer of packing as
it was dumped in to insure against void spaces along the
walls of the columns. All the piping, valves and fittings
used for the liquid streams were plastic. The only metal
in contact with the liquid solutions were the pumps used
for circulation, which had stainless steel heads. The
gas stream prior to the addition of 802 was piped with
1%-inch galvanized iron pipe. After the humidification
and addition of $02, 1%-inch polyvinyl chloride pipe was
used.

Since the surface to volume ratio for the column was
rather large (3.2:1) and since the column was to operate
at 50°C with an ambient temperature of 23°C, the column
was insulated with a 3/4-inch layer of spun glass insulation.

Inlet liquid distribution over the packing was
facilitated with the use of a “sprinkler“ type head from
which the incbming liquid was spread over an area of
packing equivalent to a circle with a diameter of 2% inches.

Figures 3 and 4 contain photographs of the equipment
as it was constructed. Both columns are seen in the
photograph as part of the equipment, although, as stated
previously, only one was used. The column used is clearly
evident as the one on the left which is covered with

insulation in Figure 4. Figure 5 contains a flow sheet of

l9

 

I IIIIIIIIII

IIIIIIII
. I 1'
am, I

 

‘71:! 7.

».-¢_-vfl' Q

oak-“000"”

.KII:

 

 

Front View of the Experimental Equipment

Figure 3.

20

ucoEQHSUM Hmucoswuoaxm ecu mo 3oa> umom .q ouswwm

-fiu‘. A.
.-._ - I-

1.00....

o. .00 Ian ..0 v I I.» I.
.IOOCIOoIcO.IQOOIOOOCOOOO
I. I

I. .v
.\

 

.
-p- .

u...fl4.u3.u.u.u....n53.3". Jaw.

 

21

wow mo coauanomo< can you com: msumuwae< “mucoaauoaxm on» no woocm 30am .m unawam

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

noume 30am n B

mama oedemm

ONE Emouum.mmo
Emouum mmo woman
Emouum poem

Bowman oaoxoom
house: noduaaom
xcuu wtaxaz

GESHoo coduauomn<
GESHoo wcammavaeam
House: new

.m
.a
.r
.w
.u
.0
.e
.0
.n
.M

22

the apparatus as it was used, that is, it does not Show

the column, piping or flow meters for the column section

not used.

The following is a description of the path

traveled by the gas and the liquid streams as presented

in Figure 5.

The Flow of the Gas Stregm'

 

1.

4.

5.

6.

Air and N2 are admitted through rotameters to the
system so that the mixture gives a 10% oxygen
concentration and the desired bulk flow rate.
This mixture then passes into a 2-inch iron pipe
containing heating elements whose temperature is
controlled via a variac.

The gas is then passed into a 2-ft high by 3.75
inch I. D. column, which is packed with %-inch
Raschig rings, to be humidified to 75-80%. The
water used for humidification was at 50°C.

The variac on the gas heater is adjusted so that
the exit temperature from the humidification
column is about 50°C.

802 is then metered into the gas stream at such

a rate to yield a gas mixture containing
approximately 2200 ppm by volume 802.

Gas then enters the absorption column through a
crowned riser to prevent liquid from running back
into the gas line. After passing through the

column, the gas is exited to the hood.

23
Gas samples are drawn off prior to and after

passing through the column for analysis by an

 

Liquid is pumped from the mixing tank which is

kept at 50°C by a glass covered heating element

The feed stream passes through a rotameter and

then onto the sprinkling head at the top of the

The liquid drops from the bottom of the column
into a swell hold-up of liquid kept there to
prevent the gas from escaping out the liquid exit
and to maintain a level from which a pump can draw.
Part of the liquid passed through the column is
directed off as product at a controlled rate by

a rotameter. This Stream removes enough liquid

containing dissolved $02 to equal that Picked up

The remainder of the liquid passes through a
rotameter, used only to help maintain the level

at the bottom of the tower, and flows back to

7.
infrared Spectrophotometer.
The Flow of the Liquid Stream
1.
connected to a variac.
2.
column.
3.
4.
in the column.
5.
the mixing tank.
' 6.

NH3 is continuously Sparged into the mixing tank
to convert the H803' formed in the column back to
803', to make up the NH4+ carried off in the
product stream and to make up NH3 losses in the

exit gas.

24

7. H20 is added to the mixing tank to make up the
H20 lost in the product stream so that the volume

of liquid in the system is constant.

EXPERIMENTAL METHODS

Prior to performing the experiments, all the r0tameters
were calibrated. The air and nitrogen rotameters were
calibrated with air. The SO2 rotameter was calibrated
by making a plot of ppm 802 at a constant inlet gas rate
to the column vs. rotameter setting as analyzed by an
infrared spectrophotometer (see Appendix B). The NH3'
rotameter was calibrated with air and corrected by equations
(21) for the flow of NH3. The calibrations for the NH3
rotameter were undoubtedly in slight error. However, the
NH3 was always added to the system at a rate to maintain
steady state solution pH with little regard to the actual
setting of the rotameter other than for comparisons between
different runs. The rotameters used to measure the flow
rates of the liquid streams were calibrated using a 4 molar
solution of NaSO3-NaHSO3 at 50°C. The water rotameter was
calibrated with water at room temperature just as it would
be fed to the system.

Preliminary running of the equipment to learn its
various characteristics and responses to variations in
operation quickly led to an efficient technique of
starting the equipment prior to making an experimental run.

A must for the quick start-up of the equipment was having

25

26
a supply of pregenerated feed stock with a NH4+
concentration and a value of S/Ceff that could be used in
the experiment. This eliminated the time necessary to
generate this solution in the equipment which would alone
take over an hour. The feed stock was made by adding a
predetermined amount of concentrated NH40H to distilled
water in a 6 liter plastic bottle so that the concentration
of NH4OH was approximately 5 molar. After placing the
bottle in an ice bath, $02 was bubbled in the solution
.from a cylinder. Samples of the solution were taken off
every 15 minutes at the beginning to measure the pH and
then at closer intervals as the pH drew close to the value
wanted in the feed stock (about 6.4). The proceedure lasts
about 1% hours. The amount of heat evolved is great and
limits the rate of addition of 802. I

The start-up routine proceeded as follows:

1. The air was turned on to equal the tetal flow
(air + NZ) to be used in the eXperiment. No
nitrogen was used in start-up as long as there
was no liquid flowing in the column.

2. The heater and humidifying water were turned on
with the variac on the heater set at maximum.
Humidifying water was taken from a hot water
tap at 50-53°c.

3. 3000 ml of feed was placed in the mixing tank and
the mixer turned on along with the heater. The

heater was set 3/4 of maximum to prevent boiling

4.

5.

6.

8.

9.

27
on the surface of the heating element.
Constant surveilance was kept on the liquid as
it reached a temperature of 50°C in about 10 min.
At this time the heater was turned down low or
shut off to maintain the temperature at 50°C.
The heater on the gas line was adjusted after
about 15 minutes operation to give a gas
temperature leaving the humidifier 0f 50°C.
It was found that it required an additional 30
to 45 minutes after the temperature of the gas
leaving the humidification column could be
maintained at 50°C for the gas leaving the
absorption column to settle at about 48°C.
After the exit gas from the column reached this
temperature, the air and the nitrogen were
adjusted to give a 10% 02 content and desired
total gas flow rate.
To flush out of the column the condensate collected
on the packing during warm-up, 1000 ml of the 50°C
feed from the mixing tank was surged into the
column causing immediate flooding at the top. As
this surge worked its way down the column, the
condensate was replaced by the concentrated feed
solution. This liquid was pumped from the system
as it collected at the base of the column.
The feed was then turned on at the desired

setting with full recycle.

10.

11.

28
$02 was now admitted to the gas stream so that
its concentration was 2200 ppm.
The NH3 was turned on to a setting of 30 on the
rotameter as a starting point for further

adjustments to maintain solution pH.

At this point the equipment was ready for an experimental

run to be made. After a few preliminary experiments, under

various solution conditions, were performed, it was observed

that the average amount of 302 absorbed in the column would

be close to 75%. On this basis the necessary flow rates of

the product and input water streams were computed with the

corresponding rotameters set accordingly.

The routine followed after Start-up to complete an

experiment was as follows:

12.

13.

14.

'15.

16.

Product and water streams were turned on to the
calculated settings and the recycle adjusted to
maintain a level in the bottom of the column.

A sample of inlet gas was drawn off every 15
minutes to observe the SO2 input concentration.
Samples of the liquid from the mixing tank were
taken every 10 minutes and teSted on a pH meter
with the correSponding adjustments then made on
the NH3 inlet.

Every 20 minutes a liquid sample was taken

(100 ml) from the mixing tank to be later analyzed.
Every 10-15 minutes gas samples were taken from

the exit gas streams and analyzed for $02

29
concentration on the infrared Spectrophotometer.
The eXperimental runs were 60-70 minutes long, measured
.from the last step of the start-up proceedure. At the end
of a run, samples of the input and exit liquids of the
column were taken.

By taking a liquid sample every 20 minutes an analysis
of the concentrations of the various components could be
made as a function of time. This was used to determine
the degree of steady-state operation during the experiment
and enable the planning of appropriate changes to be made,
if necessary, for the next experiment. For illustration,
Table 3 contains the results of these analyses for two
typical experiments. A complete table of this data from
all the experiments performed is presented in Appendix C.

Table 3: Typical Results from the Concentration vs. Time
Measurements Taken during Each Experiment

Experiment No. 4-27 Concentrations in gram-moles/liter
- + - : :
Time (min) NH4 HSO3 $03 $04 S/Ceff
20 5.458 1.208 1.931 .194 .620
40 5.359 1.219 1.883 .188 .622
*Inlet 60 5.347 1.233 1.868 .190 .624
*Exit 60 5.325 1.445 1.723 .197 .647

Experiment No. 4-3

20 6.029 1.900 1.752 .312 .676
40 6.014 1.870 1.729 .343 .676
Inlet 60 6.001 1.829 1.704 .381 .675
Exit 60 5.981 2.095 1.544 .400 .702

Inlet and exit liquid streams, reSpectively

30

AS can be seen from Table 3 and more completely in
Appendix C, the concentrations were changing very slowly
over the 60 minute experiment with a very good

maintainance of a constant value of S/Ceff.

DATA

All the eXperiments were performed with liquid and
gas inlet temperatures of 50°C. The particular data taken
from each experiment to analyze the performance of the
system was the concentrations of all species in the liquid
streams entering and leaving the column at the end of the
60-70 minute run, and also the infrared Spectrophotometer
(I.R.) read out of the $02 concentrations in the inlet and
exit gases. The concentration of 302 in the exit gas was
also computed via a sulfur mass balance on the inlet and
exit liquid streams (Appendix E). This value of the SOé
concentration in the exit gas was used in preference to
that taken by the I.R. due to unresolved technical problems
in analizing accurately a mixture of NH3. 802, H20 and air.
Appendix B contains a discussion of the analysis of the gas
stream with the I.R. and some of the problems involved.

The physical properties of the liquid solutions used
in each experiment (density and viscosity) were not measured
but calculated by empirical equations presented by Chertkov
(15). The values of these properties and the equations used
to compute them are presented in Appendix D. Also, in I
Appendix D are the equations used to compute the solution

molecular weight. ratio of the moles of water to a mole of

31

32

solution, conversions to change flow rates in ml/min to
moles of water/hr and an equation to convert concentrations
from gram-moles/liter to moles/mole H20.

The_experiments were divided into two major groups.
The first series of eXperiments were performed holding the
gas rate through the column constant at 3 ft/sec (543 lbs/
hr-ftz) varying the liquid rate from 150 ml/min to 280 ml/min
(300-560 lbs/hr-ftg). The second series of experiments
were performed holding the liquid rate constant at
185 ml/min (380 lbs/hr-ftz) and varying the gas rate from
1.5-3 ft/sec. In all these experiments the 802 concentration
in the inlet gas was between 2100-2300 ppm. One odd
experiment was performed at a liquid rate of 250 ml/min
(490 lbs/hr-ftz) and a gas rate of 2.25 ft/sec (408 lbs/hr—
ftz) to test the correlation of this data point to the other
experiments.

Tables 4 and 5 contain the data taken for analysis
of the system from the constant gas rate experiments and
constant liquid rate experiments, reapectively. The odd
experiment is included in Table 5 (Run No. 5-5). The
pressure drop across the column for the various experiments
is not included in Tables 4 and 5. They were purposely
omitted because they were felt to be unrepresentative of
the true pressure drop across the column, since included
in the pressure drop data was the pressure drop across the
packing support. This pressure drOp was felt to be much

greater than the pressure drop across the column.

33

.Am xaocooa< ommv Emmuum weaved map so mocmawn meme usmaam m up oouaano «

Nan. 000. 000. 005. 00a. ~00. 0N5. Non. 0H5. ~00. u0080

000. com. 0NN. 000. ass. 0NN. NHN. 000. ass. NNN. .eom

“om.“ ~a0.H 000.H 000. Nfiow. 0Ha.H NNO.~ 000.“ aam.H 000.“ "now

00a.H 050.H 000.s ~50." s0N.H ~05.H ~50.“ 0ao.~ NNH.~ ~H0.fl _ -nomm

000.0 500.0 005.0 00N.0 0mm.s ~00.0 amfi.0 Hwa.0 005.0 ”~0.0 +0mz
mcoaumuucoocoo banded uwxm

000. 000. ~50. use. 005. a~0. a00. 050. 050. ~00. u00o\m

000. 00a. osN. “N0. one. mam. mas. H00. mam. ~s~. "sow

hum.“ 0H0.H 0N0.H mas. “Na. oao.~ “AH.” son.“ 000.” ~00.H "mom

000.H flas.s 00¢.H ~00.“ m0m.fi a00.fl 0H0.H 0N0.“ 000.” ~am¢.~ -momm

000.0 mao.0 00a.0 ma~.0 N0m.0 000.0 omfi.q “00.0 aaa.0 000.0 +0mz
mcouumuucoocoo cased; uoHGH

omm omm omm 00N 00N 00a 00a 00s 00“ ans Acaa\sav mums 0s2Usq

owe 000 ass oaq 000 000 000 000 0H0 own «000 usxm

comm comm ooHN ooHN OONN QONN ooHN OONN oo- oomN mmmamwsau

macauwuucmocoo New

mmsq mflse n~-¢ one n-q «flue NH-¢ nuc Nuq mane .oz ucmfiauoaxm

.uouHH\mmHoE-Emum :a one mcoaumuucoocoo HH<
.ooom u wusuwuoaemu UHSUAH one now

.oom\um m we comm moo ucmumcoo m on aoxMH moon Housmaauoaxm .e oHQMH

34

own.
5H0.
wmm.H
mmm.m
Hno.m

ma“.
mom.
¢w¢.H
B~N.N
mnm.m

mN.N

oqo
CONN

00-0

“on.
owm.
hm~.~
mmm.~
«00.0

emu.
mam.
nm~.H
mNm.N
omo.o

No.N

oww
oom~

Hum

woo.
one.
oqo.H
mno.~
mow.o

mqo.
BNO.
woo.~
onm.H
me.o

mN.~

00¢
ooHN

dNum

new.
baa.
men.fi
mq¢.~
mNm.m

«No.
00H.
wow.H
Nm~.H
mqm.m

m~.~

owe
ooHN

nmuq

mew.
mew.
m~n.~
cqo.~
h~0.m

mac.
nmN.
omw.H
Hm¢.H
awo.m

m.H

mac
oomw

~sm

GHE\HE omN a moon canvaq 0
Am awocoaa< oomv emouum oazcaa map so mocwamn mama unmanm scum vousnEoo s

000. 0000\0
000. ”000
000.0 .000
000.0 -momm
0N0.0 +0mz
macaumuucoocoo vdsvaa uaxm
0N0. mumo\m
0N0. «000
000.0 "000
000.0 -0000
000.0 +002
mcoaumuucooaoo banded uoHcH
m.~ Aoom\umV oumu new
000 «000 0000
oo- 000 000:0
acoaumuucoocoo Now
owuq .oz uGoEauoaxm

.umuaa\mmaoeuemum cw mcoaumuucoccoo HH<
.ooom I ousumuoaeou oasaaa one moo

.cae\ae me no spam vane“; ucmumcoo m um :oxmw mama AmocoEAHmaxm .m manna

RESULTS

For each experimental run the value of HOG was computed
by two methods. The first method was to use Equation 8 with
the value of N0G defined by Equation 9. The second method
was to evaluate NCG by the numerical integration of its 1

defining equation given as

9’00

.. “92’.
”06 S $.39? (13)
Yin

and then use Equation 8 to compute HOG' The second method
is the most accurate since there is not a linear
relationship between S/Ceff and y'as presented in Equation 11.
This is due to the simultaneous loss of NH3 and the
oxidation of $03: as so2 is being absorbed, all of which
are affecting the value of S/Ceff' The method of computing
N0G by the above method is shown in Appendix F.

The values for'§*, in all cases, were computed using
Equation 4 in the following way since the column was run

at atmOSpheric pressure.

~a
Y~ "' P302 (tme)/760.

3S

 

3o
Reynold’s numbers for the gas and liquid streams were
computed using the equations presented with Equation 5.
The constants used for the packing parameters appearing in
these equations were taken from Sherwood and Pigford (16)
and have the following values for %-inch Raschig rings:
f 0 114 ftzlft3
¢ = .53
de = 4¢/f = .0186 ft.

Analysis of the_Absorption of $02-

 

Table 6 contains the results of the experiments for
the absorption of 802 performed at a constant gas rate and
Table 7 contains the results of the experiments at a
constant liquid rate. Making the assumption that the value
of ”CG was dependent on the Reynold's number for the liquid
stream and gas mass velocity as illustrated by Equation 6,
or simply as

H = function (R3 Rb )
OG eG eL
a plot of Leg H0G vs. Log ReL was made from the constant
gas rate experiments. The slope of the resulting straight
line was the power on R81’ and the intercept was the

combined value of all the constant terms. This plot is

presented in Figure 6 and yielded the following equation:

(H = 3.89 Re"39. (14)

OG)G=constant L

Figure 7 contains a similar plot where Log HOG is now

plotted vs. Log G from the constant liquid rate experiments.

37

Nfim.H mmn.fi 0n.o ewm com m.ow ONq CONN m~u¢
00w.fi mmn.~ an.m mm” mmH m.mw own comm mane
moo.H mow.H Ho.o NNN mofi m.om qu oo- sane
nmm.H www.H wo.o cos 5mm p.05 ooq oo- mus
mmw.~ mmw.H mo.o one mmN w.mn mmm comm hue
qo~.~ NwH.N mm.¢ omfi HHH H.mh cod ooNN «Hus
mHH.~ mwo.N o¢.¢ Nwm hmfi H.nn owe ooHN Nan:
mo~.~ HNH.N mm.q «mm «mm m.mh oqm oowm mnc
Nnm.N omm.~ om.m cum NHN N.~n ofio comm Nue
mNm.N mm~.N nm.m 0N~ MNH ¢.nn owm oomN mane
-

mwmwummwm mmeMMWde Aeoav Assay
Scum Eoum “fixes.0 ”fiancee.A _ .
vouSQEoo MMuooEoo go UAmVHH Sufism). mfimav Amway .oz

3.3 m m 05.333500 :0 om “00930002 N m. um. passauwaxm
Amum-u£\mnH memv .oem\pm m n mumu mow
muom mow unnumcoo m um mucmawuoaxm mnu
some ooumasoamo Now mo coauauomne emu mom monam> wow .0 manna

A

38

0m.m u om unassumaxo 002a cH «
000.H m0w.fi 0.000 H00 00m m.~0 0H0 00mm «mnm
mNN.N 00H.N m.m00 000 com 0.00 0N0 00mm finm
00H.N m0H.N 0.000 mmm 00H 0.00 000 00H~ «Nun
mHN.N m-.~ 0.000 0NH mm 0.00 0N0. 00- .0N-0
m0~.~ HOH.~ 0.HON MOH Hmfi 0.00 000 000m mum
0N~.N 0~N.~ 0.H0~ mm” 00 m.00 000 00- 0~-0
.mmeumme wwwfluwwcm wfimmv uwfimwv
scum eoum Numoun fi% 0%. 00 00
nmusaeoo GWeusasoo Allmflllv UHDUHH Spas AWDOV A” v .02
nut m 0 0.3.3.0325 :0 Now 0onu000< N 000.. MM ucoEauoaxm
00.0 a 000
oumm 003004 uCMumcoo m um mucosahmaxm onu
soup woumHSOHmo Now mo coauauomn< map you moDHm> 00m .0 magma

39

 

 

Amum-nn\mna 000 a 00 ovum 000 pawumcou 0 ud 000x000 unmanm
0
mo coauauomn< ecu Mom mm 004 .m> 00m woq mo uoam < .0 ouanm
, a a
a 0\00 a mm ‘00 004
mo mam. pm. . MN. $0 “#0. w. mm. m.
a a a a 0 u d . 1 q ON.
: NN.
, 1H .u 6N.
000. + 00 000 000. - n 000 004
i. ON.
L. wNo
000.
..~m.
4 0m.
..0m.
0 s.

 

(33 HI 90H) 90H 901

40

A
A0m.0iu_ omv mumm 005004 UCQUmcoo w um 000x000 HSMHDm mo

..dowu0uomn< onu now 0 wag .m> mom woa mo uoam < .0 enamdm

ANuu-un\0na ca 00 o 000
m0.~ 0.N m0.~ 0.~ mm.~ m.~ m0.~ 0.~

.

J a 1 a

0

mm.

J.

n:- on.

00

000. + 0 000 000. a m 004

H 901

.. 00. w

o 0

M o 1.. «m.

a 00.

(a; “I 90H)

4 mm.

 

L. O‘No

41

The slope of the resulting line is, for all practical
purposes, zero. Taking the slope as zero means no
dependency of HOG on the gas rate as was also shown by
application of Chertkov's equation for kg in the equation
for HOG,(Equation 7). Since the value of HOG does not
depend on the gas rate, the value of HCG found in the odd
experiment should compare to the HOG values obtained at
that same liquid rate in other experiments. This was in
fact found to be the case, The value of HOG from the odd
experiment was equal to 1.900 compared to values of 1.885,
1.927, and 1.905 for HOG found in other experiments at
this liquid rate.

By using the physical properties of the air,
diffusivity of 302 in air (see Appendix D) and the packing
parameters an equation can be written in the form of

Chertkov's equation (5) which is

-8—2 = .0013 Re Re '39 (15)

G G L
Analysis of the NH3 Stripping

An approach identical to that used for the analysis

of the absorption of SO2 was used to analize the stripping
of NH3. The values of N0G for the loss of NH3 were
computed using Equation 9 and then used in Equation 8 for
the computation of HOG‘ Table 8 contains the results of
the eXperiments performed at a constant gas rate and

Table 9 contains similar results from the constant liquid

42

.m 000cmaa< mom .memwuum nHSUHH uaxe 0:0.ueacw so mocmHmn mmms mmz m Scum vouSQEoo a

 

00.0 00.0 000 000 000 000. 000. 00... 00-0
00.0 00.0 000 00m 00m 000. 000. 00.00 00-0
00.H H0.0 00H H00 00N 000. N00. 0m.0~ 0H-0
mm.“ 00.0 ofifi 00H mNH 000. 0H0. 00.0 0-0
Nm.H 00.0 00H 00H 0HH 000. 0N0. N0.0 0-0
00.H 00.0 000 000 000 000. 0N0. 00.00 00-0
m0.H 00.0 0NH 000 00H mN0. 000. H0.0 Na-0
00.0 00.0 000 N00 00m ~00. 000. H~.0H 0-0
N0.~ 00.0 00N 00m 000 000. 000. 0m.0H N-0
00.0 00.0 0mm 000 000 N00. N00. N0.~H mH-0
00 .00
mcmwwmsam wflmmws um%mmw0 on 00005 000
wouSQEOO 7r 3. 05000 . mz meaoe .
AumV 00$ 400 angumemwwocstmmz Junom. unofluumo\0v aaAmuou\mV .>mo uceEAmmaxm
0 u 000 map :0 coaumuucoocoo 0:2 uoHSH
Amum-u5\mna 000 u 00 .omm\um 0 0 open 000
oumm 000 ucmumcoo e um mucoEHuoaxm onu
Eoum woumHSOHwo mmz mo waflamwuum onu Mom 003Hm> 00m .0 manww

43

4m

 

 

00.0 a m ucmEOHmaxm 00:0 00 a
00.0 0.000 000 NNN 00m 000. 000. 00.00 «0-0
00.0 0.000 000 000 000 000. 000. 00.00 0-m
00.0 0.000 000 000 000 000. 000. 00.00 «N-m
00.0 0.000 000 000 000 000. 000. 00.00 0N-0
00.0 0.00m 000 000 000 000. 000. 00.00 N-m
00.0 0.000 000 000 000 000.. 000. 00.00 0N-0
0 .0
mm00u0WUm - AEQQV AEQQV N
5000 mum-0; 00x00 000000 0 x 00005 000
00030500 A 0.00 0 000300000 503 um. 05000 0mz 000.9: .02
003 000 0 0300000200 5. 002 050.0 “500000800 50.000300 >00 000500000
4
00.0 n 00

0000 003004 00000000 0 pm mucoE0Hoaxm 0:0
5000 0000030000 0mz no 000000000 0:» non 00300> 00$ .0 00000

Amuw-pn\mn0 000 u 0V mumm mmo achmdoo m um
m0coEE< mo MC0am0uum msu 000 0mm w00 .m>.oo
. 0

m meg no uo0m < .m muaw0m
«\0q n mm 0mm 000

mm. om. m0. 00. mo. 00. mm. om.

 

44

0 A . 0 . ‘4 n 00.

Av NH.

; «0.

0m. + 4mm 004 00.- a mom mag

: 00.

a 00.

: ON.

It NN.

nv +VNO

-- ON.

5 ”N.

 

 

: on.

(a; U1 90H) 9OH 901

45

00
000.0 u 00
mu0m 005000 00000000 0 00 0050550 00 0:0000Hum map 000 00m no u00m 0 .0 003000

um-uc\mn0 :0 0 0 000

 

N .
00.~ m0.~ 00.0. $0 00.~ mm.~ 00:0 00.0 0.0.0
.7 0 A u o b. u . 0 OH.
0. N“.
00. - 0 000 00. n 000 000 s 00.

0. 0N.
I. NN.
.. QNO

.0 ON.

 

 

03.

90H 801

(31 H? 9OH)

46

rate experiments. Again a plot of Log ”H0 vs. Log Re was
. L
made as shown in Figure 8. The following equation could

be written for the best straight line through the points

using a least squares fit:

3.7 R "5 (A)

(HOG)G=constant = eL

Figure 9 shows the plot of Log HOG vs. Log G for the
constant gas rate experiments and gave the following
equation:

(H = .162 G’38 (B)

OG)R =constant
eL

Dividing Equation A by (543.4)'38

or multiplying Equation B
by (4.37)'S, in which the quantities in the parentheses are

the values of G and Re held constant to get Equations A and
L
B, gives the following equation for HOG:
038Re '05
L
Using this equation to compute the value of HOG expected

HOG = 0338 G

for the odd experiment gave 1.37 compared to the observed
value of 1.38. From Equation 6 the following equation can
be written for the partial gas mass transfer coefficient
of NH3:

k d .62

ii?- = .0035 Re '5

R
G eL

Analysis of the Oxidation of 803: to 804:

The amount of oxidation occuring during a pass through

the column for each of the experiments was compared to

47

that predicted by Chertkov's equation (12). Since the value
of S/Ceff changes as the liquid passes through the column

and since the oxidation rate predicted by Chertkov's equation
depends very heavily (6'th power) on this value, an
integrated average value of S/Ceff was used. The average
value used was given by

7

7
out - (S/Ceff)in

(S/Ceff)

 

6
s C = fl

A further illustration of the calculations is in Appendix B.
Table 10 contains the rates of oxidation experienced
in the various experiments along with the corresponding

predicted values.

-13 ‘ ‘TTF'T‘ZF -! , .

48

 

a
I IIII rig-«l ill-km

0.90

:00posuo m.>oxuhoco 5000 onmHGUHwo «

m00.

 

00. 00. 0.00 00.0 000. 0-0
00. 00. 0.00 0.00 00.0 000. 000. 0-0
00. 00. 0.00 0.00 00.0 000. 000. 00-0
00. 00. 0.00 0.00 00.0 000. 000. 00-0
00. 00. 0.00 0.0 00.0 000. 000. 0-0
00. 00. 0.0 0.0 00.0 000. 000. 00-0
00. 00. 0.00 0.000 00.0 000. 000. 00-0
00. 00. 0.00 0.0 00.0 000. 000. 00-0
00. 00. 0.00 0.0 00.0 000. 000. 00-0
00. 00. 0.00 0.00 00.0 000. 000. 0-0
00. 00. 0.00 0.00 00.0 000. 000. 0-0
00. 00. 000, 0.0 00.0 000. 000. 00-0
00. 00. 0.00. 0.0 00.0 000. 000. 00-0
00. 00. 0.00 0.0 00.0 000. 000. 0-0
00. 00.0 0.00 0.00 00. 000. 000. 0-0
00. 00.0 0.00 0.0 00. 000. 000. 00-0
000 x 000 x 000000xo 003000 u0xm on
«wouo0ooum 0mucoE0Honm wmmuwmmw :0 .000 AmwflmQHMAv .
u: 0 N no uSMHSm oumm use mum :0 000 oz
000005-000 005000 u 00 00 0 00000 00 0 002000 0 o\0v 0 o\mv ucoE0uoaxm

moumm GO0u0p0xo pou00poum use vo>umnno cmozuon nau0umafioo .00 oaan

ANALYSIS OF RESULTS

§92 Absorption

The equation for the mass transfer coefficient

" ”7'11

(Equation 14) derived from the experimental data is almost

m...-

identical in form to that proposed by Chertkov, Equation 5. -_
The fact that the power on the liquid stream Reynold's -
number is essentially the same, however. may be coincidental.
The variations of the mass transfer coefficient with the
liquid rate through the packing is undoubtedly due to
changing amounts of wetted area resulting from changing
characteristics of the fluid films surrounding the packing
pieces and variations in the interfacial area. Since
Chertkov's work was done with an absorber packed with
bundles of rods 3 and 5 mm thick with 20 and 25 mm between
the rods as compared to this work which was dOne with %-inch
Raschig rings. the assumption that the wetting phenomenon
would be the same for each work is weak.

It has been shown (17) that below the flooding point
countercurrnet flow of gas to liquid in a packed column has
no effect either on the volume of liquid adhering to the
packing or on the wetted area. Thus. the power on the

Reynold's number for the gas stream is only a function of

49

50
the absorption dependency on the eddy diffusion promoted
by changing gas rates. It was shown in Equation 7 that the
substitution of Chertkov's equation for the mass transfer
coefficient into Equation 6 for HOG gave an equation for
HOG which was independent of the gas rate. This is also
seen to be the case with the equation for HOG (Equation 14) .
obtained in this work. This means physically that in the

ranges studied at a constant inlet value of S/Ceff' constant

1’. D“ U.n“u‘_l ‘ “fl!

liquid rate, and for a given depth of packing. a constant
percentage of the incoming 802 will be absorbed for any gas
rate in the range. More generally, lifting the criteria of
a constant inlet value of S/Ceff’ the percentage of SO2
absorbed will always be such that the values of N06 are
equal for any gas rate in the range. This was also observed
in a work by TVA (18) in which the packing depth, liquid
rate and the inlet value of S/Ceff were held constant,
variations in the gas rate gave a constant percentage
absorption of $02.

To analyze this phenomena a computer program (see
Appendix F) was developed which computed points on the
operating and equilibrium lines for a y vs. S/Ceff plot
depicting the absorption of $02. The program used the
values of HOG determined in this work for the absorption
of $02 and the stripping of NH3 and also the oxidation of
$03= to $04: via Chertkov's equation. TWO values of L/G

were used. .7 and 1.4 (G=1.5 and 3 ft/sec). with an inlet

value of 0:14 lb-moles NH4+/100 moles HZO- S/Ceff

= .667

 

0}. ‘ka n. win-“ll
310. .

51

and A=1.4 lb-moles 8043/100 lb-moles H20. In both cases
80% of the incoming SO2 was absorbed. The results of these
calculations are presented in Figure 10. It is interesting
to note that the operating lines as well as the equilibrium
line are essentially straight with very little curvature,
which seems to validate the assumptions for using the log-
mean driving force in calculating N0G by Equation 9. For

this illustration, however, Equation 13 was used for the

‘ Kill». ‘h‘nlhl Ill:

computation of NOG' letting the computer do the numerical

'1'
I

integration. The integrations were started at the bottom
of the column in each case which is depicted by a value of
SIC

0685 for L/G = 104 and S/G 0700 for L/G '3 .7.

eff = eff =
N0G is plotted as a function of S/Ceff in Figure 11 for
both cases as the integration proceeds up the column to the
top. depicted by S/Ceff = .667. For L/G = .7 the value of
NOG for the absorption was 2.21 compared to 2.18 for

L/G = 1.4, a change in N00 of 1.35% compared to a change of
100% in L/G. Such a change is well within experimental
error, and hence not conclusive. Thus, it appeared. upon
the analysis of the data in this work as in the works of TVA
and Chertkov. that the value of HOG is constant with

changing gas rates at a constant liquid rate. It may be

worthwhile to point out that, upon closer observation of

 

Figure 10. one sees that the slope of the operating lines
are much greater than the slope of the equilibrium line.
Thus. to the operating lines. the equilibrium line may look

to be horizontal. i.e., have zero slope, which means the

Gas Phase 802 Concentration in ppm by Volume

2.1T

1.9.

1.7..

1.5.

193'

101'

.9-

 

52

 

 

'7‘ a = top of column
b = bottom of column
c = operating lines
’5‘ d = equilibrium line
a
,3 .. /d’//
.1 0 a : 0 : s a
.665 .67 .675 .68 .685 .69 .695 .70
Slceff
Figure 10. The Operating and Equilibrium Lines

Depicting the Absorption of 80% of
the Inlet 802 from the Inlet Gas

W

I?! ma 0‘. n
_4 #L_'i_:

53

204‘"

2.0...

 

 

 

1.2..
.81.
04"
0.0 ; x : . 4 f :
.665 .67 .675 .68 .685 .69 .695 .70
Slceff

Figure 11. A Plot of NOG vs. S/Ceff during the

Integration up the Column

54
integral for NOG will always be nearly constant for a
particular percent absorption and any value of L/G choosen.
One would expect to begin seeing changes in NOGwith changes
in L/G as the slope of the operating line becomes similar to
that of the equilibrium line. However, research present in
the literature and this work have all used values of
L/G = .7 or higher and consequently HOG is always seen to
be independent of G.

It is noted that the major difference between the
equation for the gas phase transfer coeffieient obtained by
. Chertkov and that arrived at in this work is in the constant
term. This is to be expected, since the constant was‘

defined by the following equation:

Juom Q (Schmidt No.)
COHStant ’3 (3089—)f PGDG= f 3.89

which is a function of the packing void fraction (©) and:
Specific surface (f).

It can be seen in Tables 6 and 7 that the value of
HOG computed by using Equations 8 and 9, which involves the
log-mean driving force in the computation of NOG’ are
always slightly less than thoSe obtained by the numerical
integration of NOG' This can be explained by observing
Figure 10 and noting that the equilibrium line is curved
slightly away from the operating line. This tends to
reduce the value of NOG slightly, which in turn slightly
raises the value of HOG. Using the log-mean driving force

approach assumes the equilibrium line to be a straight line,

55

connecting the end points, which puts most of the points

on the equilibrium line closer to the operating line,
-giving a slightly higher value of NOG and a correspondingly
lower value of HOG.

The spread in the data points seen in Figures 6 and 7
containing the plots of Log HOG vs. Log R61‘ and Log G is
undoubtedly due to the errors involved in making titrations
'of the liquid solutions (see Appendix A) and the lack of an

accurate technique to measure SO2 in the exit gas.

3&3 Stripping

 

The stripping of NH3 was measured, as was the
absorption of $02, by using the difference between the
inlet and exit liquid concentrations of N114+ ions for the
case of NH3 losses and the differences in total sulfur
concentration for the case of SO2 absorption (see Appendix E).
The differences calculated for NH3 losses were always about
10 to 20 times smaller than those for SO2 absorption. This
method was used for the determination of NH3 losses, since
a good technique to analyze the exit stream for NH3 was
not available. I

The exit stream, which contained a calculated average of
150 ppm of NH3, made the obvious techniques of bubbling the
NH3 containing gas through a colorimeteric solution or acid
solution, at a known rate, impractical and inaccurate due
to the large quantities (prolonged flow rates) of gas to be

-passed through the solution. There is also the problem of

56

assuring that the very dilute gas was sufficiently Sparged
into the solution to trap all the NHB' Again a small
difference would be worked with, thus leaving either
technique with something to be desired.

To make the plots of Log (HOG) vs. Log Re

G=constant L

and tog (HOG)Re vs. Log G the assumption had to
L=constant

be made that the stripping of NH3 is gas phase controlling.
Since the data correlates quite well to a straight line in
these plots, it is felt that this assumption may be
justified.

Since the data points for the losses of NH3 were felt
to be less accurate than those found for the $02 absorption,
the values of HOG for the various experiments were computed
using only Equations 8 and 9, which involves the log-mean
driving force. The applicability of these equations seems
justified by the observation of Figure 12 which contains a
plot of the operating and equilibrium lines for the stripping
of N83 which were obtained from the forementioned computer
program. The lines are very slightly curved due to the
simultaneous absorption and partial oxidation of 802 in the
solution. Since the curve of the lines in this simulation

is slight, the use of the log-mean driving force to compute

NOG is a fair assumption.

§olution Oxidation
The oxidation of the $02 desolved in the solution as

803' and H803- to $04: compared very well to those

NH3 in the Gas Stream (pmeby volume)

350-. .

300»

250»

2000

150"

1000

500

 

0

.665 .670
Top of
Column

Figure 12.

e = equilibrium line

0 = operating line

57

       

.675 .00 .685 .690 .695 .700
Bottom

of
Column

SlCeff
Computed Operating and Equilibrium Lines
for the Stripping of NH3 During the

Simultaneous Absorption and Partial

Oxidation of 802

C = 14 moles/100 moles H20

58

predicted by Chertkov's equation for the oxidation. This
was not to be expected, however, since Chertkov himself
warned that the equation he gave was derived from data
taken from industrial scrubbers. These solutions are
contaminated with various elements and compounds which
can promote the oxidation. Thus, Chertkov's equation
would be expected to predict relatively high rates of
oxidation as compared to those experienced using clean
gas and solution. This result can be explained in two
ways. One explanation may be that the solutions used in
the lab experiments were as contaminated with iron (which
is a great oxidation catalyst) as those solutions Chertkov
encountered when he developed his equation. A second
possible explanation is that the oxidation rate is
controlled more heavily by system parameters other that
the degree of contamination. Neither one of these hypotheses
can be refuted at this point due to the lack of understanding
of the oxidation mechanism.

It is noteworthy, however, that the solutions used
in this experiment, although not contaminated with the
many different compounds as would be found in an industrial
scrubber, contained a trace of iron. The iron came from
the pumps in the system and gave the solutions a definite
yellow—orange tint. If iron is the major oxidation
promoter in a contaminated industrial solution then the

results obtained in this work would be expected.

59

The analysis of the rate of oxidation of the solution
has the same disadvantages as that of the analysis for
ammonia losses. To make a mass balance of the sulfate
formed on a pass through the column the difference between
inlet and exit concentrations of sulfate must be taken.
These differences are small and border on the accuracy
of the titrations used to determine the total sulfate
present in the inlet and exit solutions themselves. To
get a meaningful analysis means very strict monitoring of
the concentrations of the standard solutions used to make
titrations and a well organized and precise titrationm
proceedure. This type of control was thought to be followed
in this work, however, the data still has a quantity of

scatter.

CONCLUSIONS AND RECOMMENDATIONS

Basic Observations

 

‘In the analysis of the absorption of 802 by
ammoniacal solutions in any type of scrubber, three
phenomena are occuring simultaneously:

1. SO2 absorption

2. NH3 desorption
3. Oxidation of 803: to 804=
Each has the same effect of increasing the value of S/Ceff
which, in turn, determines the rate at which eaCh proceeds.
To illustrate these simultaneous events Figure 13 was
prepared from the data obtained from the computer program
mentioned previously (see Appendix F) in which L/G = 1.4

(G = 271.6 lb/hr-ftz), (s/c .667, c = 14 moles/

eff)in =
100 moles H20 and A = 1.4 moles/100 moles H20. The computer
began at the bottom of the column and calculated upward
until the column height was such that 80% of the 802 in the
inlet stream containing 2000 ppm 302 was removed. The
column was assumed to be packed with %-inch Raschig rings.
In Figure 13 the percent absorption of $02, the percent of
the total ammonia lost, and the percent of the 802 being

absorbed at that point that is oxidized is plotted vs.

the position in the column.

60

IOU--

90..

80-»

70..

60..

50..

40..

Percentage

30 +-

20)’

10.]

0

 

0
Bottom

Fig‘lre 130

61

Position in Column (ft)

_.I
1

2

n l P
I I l

3 4 5
Top

A Plot of Percent SO2 Absorption, NH3 Lost
and Percent 802 Absorbed at that Point that
is Oxidized vs.

Curves:

Position in the Column.

3.

b.
C.

Percent SO2 input that

is absorbed

Percent of total NH3 lost
Percent of 802 being
absorbed at that position
that is oxidized

62

The interesting point here is that as one approaches
the top of the column where the amount of 802 being
absorbed is decreasing very rapidly due to its approach
to equilibrium, about 72% of the 502 being absorbed at
that point is oxidized. Noting the rate of climb of the
oxidation curve. c. itncan be seen that if the height of
the column was extended so that perhaps 90% of the total
incoming 802 could be removed, curve 0 would raise above
100%. This would mean that some of the absorbant itself
is being oxidized which would cause the gradual build-up
of 804: in the system, destroying its capacity to absorb
802.

It has been observed in the literature that a typical
absorber would operate in stages with the first stages
contacting the gas with liquid that has a low value of"
S/Ceff (.7). This will promote good 802 absorption, high
ammonia losses and a rather low rate of oxidation,
remembering that according to Chertkov the oxidation rate
is proportional to (S/Ceff)6. In the top stages, however,
it is stated that the liquid should have a high value of
-_S/Ceff (.9+). This would promote very little absorption
of $02» but would absorb the NH3 lost in the first stages.
The oxidation. however, would increase substantially.
essentially oxidizing the solution, since little SO2 is

being absorbed. It seems then that a more general

investigation of this technique may be in order with the

63
problem of oxidation considered along with curbing NH3
losses.

If the end result of the absorption process is the
manufacture of (NH4)ZSO4 the approach may be acceptable
since any oxidation in the absorber just means less load
on the actual oxidation unit providing the oxidation in
the absorber is not so great as to destroy the absorbent.
If the absorbing solution is to be regenerated. however.
the oxidation should be stopped as much as possible. Thus,
the use of stages, which seems to promote oxidation,
counters what is wanted. It would seem that in order for
a system of stages to be used, an oxidation inhibitor is
mandatory.

It seems justified to conclude that even though the
scrubbing of flue gases for the removal of $02 by .
ammoniacal solutions is a very efficient method. the system
must be looked at more generally than just what has to be
done to get the best 802 absorption. Two other simultaneous
phenomena are occuring that can literally "make or break"

the feasibility of the process.

Suggested Research

The next step in carrying on the work started here
would be to test the equipment on actual flue gas. It
would be very beneficial to know if the absorption
characteristics observed in the laboratory could be

duplicated using contaminated gases and solutions. The

64
following is a list of major topics that must be generally
investigated:

1. The resistance to absorption of $02 by the liquid
phase at higher values of S/C with possible
developement of an equation for the liquid phase
mass transfer coefficient.

2. A further study and verification of the results
obtained in this work for the losses of NH3.

3. A thorough study of the oxidation is needed,
analyzing it as an absorption process determining
and developing correlations for the gas and liquid
phase mass transfer coefficients.

4. The regeneration step in the process must be
investigated as existing methods are either
uneconomical or lack enough technological study

to be evaluated.

Suggested Experimental Improvements

There are four major areas for improvement that
should be made prior to further investigation. First,
modification of the infrared Spectrophotometer or
switching to another technique to very accurately monitor
the 802 concentration in the exit gas is needed. Second.
an instrumental technique should be developed to analyze
the exit gas from the column for NH3 content. Third. a
better analytical technique should be developed to

eliminate some of the errors involved in the present type

65
of analysis used to determine the rate of formation of
$043. Forth, the tap on the column that is used to
measure pressure drop should be moved above the packing

support plate.

BIBLIOGRAPHY

3.

4.

5.

6.
7.
8.

9.
10.
11.
12.
13.

BIBLIOGRAPHY

Johnstone, H. F. “Recovery of 802 from Waste Gases."
Ind. Eng. Chem. ;Z(5), 587-93(May 1935).

Chertkov, B. A., and Dobromyslova, N. S. "The
Influence of Traces of Sulfate on the Partial Pressure

of SO Over Ammonium Sulfite-Ammonium Bisulfite Solutions.“
J. Appl. Chem. USSR 31(8), 1707-11(Aug. 1964).

Chertkov, B. A. “Coefficients of Mass Transfer in

Absorption of 80 from Gases by Ammonium Sulfite-

Bisulgite Solutions." J. Appl. Chem. USSR.§Z, 2404-10
1964 .

Tennessee Valley Authority. "Sulfur Oxide Removal from
Power Plant Stack Gas: Ammonia Scrubbing“ (1970), 23-4,
Report No. Not Yet Assigned. Prepared for National Air
Pollution Control Administration (U. S. Department of
Health, Education, and Welfare) '

Hein, L. B., Phillips, A. B., and Young, R. D.

"Recovery of $0 from Coal Combustion Stack Gases.“

In Problems and Control fo Air Pollution (Frederick

S. Mallette, ed.), Reinhold, New York (1955) pp. 155-69.
Ibid. NO. 30

Ibid. NO. 40 (pp. 66).

Showa Denko K. K. (Tok o) "302 Removal Plant",
Publication DA-251(1.0 , July 1967.

Ibid. No. 4. (pp. 43-46).

Ibid. No. 1.

Ibid. No. 2.

Ibid. No. 3.

Chertkov, B. A. "General Equations for the Oxidation
Rate of Sulfite-Bisulfite Solutions in the Extraction

of $0 from Gases." J. Appl. Chem. USSR 33(4), 743-47
(19613.

66

 

lul'll

‘14.

15.

16.
17.
18.
19.
20.

21.

67

Sherwood, Thomas K. and Pigford, Robert L. Absorption
aggugxtrgctigg; McGraw-Hill Book Company, Inc., New

York, 195?. p. 248.

 

Chertkov, B. A., and Pekareva, T. I. "Density and
Viscosity of Ammonium Sulfite, Ammonium Bisulfite, and
Ammonium Sulfate Solutions." J. Appl. Chem. USSR 33(1),
135-41(1961).

Ibid. NO. 140 (p. 239).

Ibid. N00 140 (po 230).

Ibid. No. 5.

Ibid. NO. 150

Wilke and Lee. Industrial Engineering Chemistry. 47,
1253(1955).

Perry. Robert H., Chilton, Cecil H. and Kirkpatrick,
Sidney D - 9111111 .051). -_ 152m: i7.1.53:22:52..iirzaIgzifitiqqk... McGraW-Hil 1-
Book Company, Inc., New York, 1963, Forth Edition, pp. 5-13.

APPENDIX A

Analysis of Liquid Solutions

The following techniques were used to analyze the
liquid samples taken during the various experiments. These
techniques are standard quantitative techniques found in any
elementary quantitative analysis text. The portions of
reagents cited to be added at various steps are based on
the fact that all solutions analeed were to be less than
6 molar in NH4+ ions. In cases where the concentration was
greater than 6 molar the appropriate changes were made..
Standard KOH and HZSO4 were used with concentrations of

approximately 2 and 1 molar, respectively.

I. Determination of (NH423303_

1. To a 25 ml sample add 150 ml distilled water. Add
25 ml 30% H202. In order to avoid excessive and
dangerous heating, this proceedure must be carried
out in an ice bath, with the H202 added in smaller
portions.
‘Reactions: (NH4)ZSO3 + H202-9 (NH4)ZSQ4 + H20

(NH4)HSO3 + H202-—4 (NH4)HSO4 + H20

2. Allow the solution to stand out of the ice bath

for ten minutes to insure the above reactions

are complete.
68

69

3. Boil the solution gently to destroy the excess
H202.

4. Cool the solution and add methyl red indicator
(approximately 10 to 15 drops, depending on the
strength of this indicator).

5. Titrate with standard KOH to the bright yellow
methyl red end point.

Reactions: ZKOH + 2(NH4)HSO4-—->|(NH4)ZSO4 +
K2304 + ZHZO
, ((NH4)HSOB) =

(ml KOH soln used in titration)(KOH)
(ml sample used)

II. Determination of (NH4)2§Q3_

1. Take a 100 ml aliquot of a standardized KI/IZ
solution (.3 molar) and dilute with distilled water
to 200 ml.

2. Titrate with sample solution until a pale orange/
yellow color is present. Constant stirring must
be employed, otherwise SO2 will be evolved during
the process.

3. Add starch indicator (approximately 10 drops).

4. The solution, which now is deep blue, is titrated
to a clear end point. The end point is very sharp
and even the slightest excess will cause the
solution to turn pale yellow. I
Reactions: (NH4)HSO3 + 12 + H20-* (NH4)HSO4 + 2HI

(NH4)2803 + 12 + H20-—%’(NH4)ZSO4 + 2H1

((NH ) so ) a (mi KILL )(Molarity I 1 - ((NH )HSO )
4 2 3 (Volumezsample used)2 4 3

1A.. «a: «Q
~—

III. Determination of the Total NH

70

+ -
4 Present

The technique used here will be to add an excess of

KOH to a aliquot of sample and boil the solution. The

vapors containing NH3 will be collected in a known quantity

of acid which can then be titrated to determine the NH3

absorbed.

1.

2.

4.

5.

A 25 ml sample is placed in a 500 ml boiling
flask and diluted with 100 ml distilled water.
Into a 300 ml flask (Erylenmyer) place 100 ml
standardized H2804 and dilute to 150 ml. A
condenser must be set up so that the boiling
flask can be attached very quickly. Also, a
delivery tip must be attached to the condenser so
that the tip can be placed under the surface of the
acid.

Put the flask containing the acid in position
with the tip well under the surface.

Add boiling chips to the boiling flask and then
add 100 ml of 2 molar KOH and very Quickly connect
the condenser and turn on the cooling water.

Boil the solution until about % of the liquid has
passed over. Caution must be taken when the
solution first starts to boil as this is the time
when most of the N83 is driven off and the acid
solution may be drawn up the delivery tip into the

condenser and into the boiling flask.

71

Reactions in the boiling flask:
2 KOH + 2(NH4)Hso3-—+.Kzso3 + (NH4)2803 + ZHZO
- A
Then 2 Ken + (NH4)2303-+.K2303 + NH3 + ZHZO
Reactions in the collecting flask:
2NH3 + HZSO4-~—)»(NH4)ZSO4
6. After about % of the liquid has boiled over

momentarily disconnect the condenser and test the~ _M4

vapor leaving the flask with Hydrion Paper for any

evolving NH3. If the test is positive, continue
the boiling and rechecking.

7. When the test is negative titrate the acid
solution with KOH using methylred as the indicator.
Reactions: HZSO4 + ZKOH-w+ KZSO4+ H20

 

(ml sample)
IV. Determinatign of SO43 Present

There are essentially two techniques available to
determine the 804= concentration. The first involves simply
finding the $04= ion necessary to satisfy the stoichiometery
of mixture. This was obviously a very simple method. The
second method is to oxidize the solution converting all the
H303“ and sog'to H304 and $04: and precipitating with BaClz.
Knowing the H803- + $03: concentrations from parts 1 and 2
would allow by difference the computation of the $04: ‘
concentration originally present.

The first technique was used through out all the

analyses since both techniques required the use of a

 

72

a difference of two separate titrations making either less

than ideal.

A.

B.

This first technique requires no further titrations
but only to satisfy stoichiometry.
S‘:==}[;"'_ " _ z]
()04 ) 2 (N1-4 ) (HSO3 ) 2(SO3 )

Therefore, the NH4+ ion tied up with the 304= is

+ :
(NH4 ) _ 2(so4 ).

Determine total sulfur present, i.e. (3042 +

1. To the solution resulting from the H303
determination (Part I), add an excess of
saturated BaClZ solution.

2. Filter on to a tared piece of filter paper,
dry and weigh. .
Reactions: K2804 + (NH4)ZSO4 + ZBaClzrrs

ZKCI + 2NH4Cl + ZBaSO4l

(804 z) = (96 07 %) (Net weight of precipitate)
233. 43 (m1 of sample)

- (Hso3‘) - (803:)

1"!“ "_‘1““"‘ "._“—~ 1‘

APPENDIX B

It was observed that in the frequency range of 1400 to
1500 cm'l, 802 exhibited a quantitative absorption peak when
viewed with an infrared spectrophotometer (I.R.). This
fact was made use of to calibrate the I.R. to analyze the
inlet and exit gases from the column during the experiments.

Prior to the calibration of the I.R., a zero of 90%
transmittance was taken as a base setting at a frequency
of 2000 cm’l. The I.R. was always adjusted to give this
reading at this frequency before its use. Calibration of
the I.R. was accomplished by first very carefully measuring
the volume of the I.R. gas cell. The gas cell was then
purged thoroughly with dry air and then injected with a
measured amount of 802. A scan of the absorption range was
made and the percent transmittance recorded. The gas cell
was again purged and injected with a measured amount of
802. This proceedure was repeated until a plot of percent
transmittance vs. volume percent 802 in the gas could be
made.

With the calibration of the I.R. the calibration of
the $02 rotameter on the experimental equipment could be
made. The gas rate to the column was set and the 802

turned on to a particular setting. A tap in the gas line

73

1h h.‘8="‘i".’. ‘
. F

k

74

to the column was connected via rubber tubing to the gas
cell since the back pressure of the column was sufficient
.to cause gas flow to the cell. After about a minute the
reading of 802 content in the gas remained steady and the
value of the transmittance and the SO2 rotameter setting
recorded. The above calibration was done with dry gases.
To check these calibrations under aetual operating
conditions the gas was humidified by the humidification
column on the equipment and then the SO2 added. Samples
were again taken off, but now had to be passed through a
water trap and CaClZ chamber to dry the gas before it
could be directed to the I.R. gas cell. There were no
observed variations in the calibrations. The calibrations
were also checked if the gas prior to passing into the
CaCl2 was passed through concentrated H2804. Again no
change in the calibration was observed. I

The gas samples taken from the exit gas from the
column presented insurmountable problems when an attempt
was made to analyze them with the I.R. In the exit gas
stream, along with air there are three components: N33,
SO2 and H20. It is believed that the reaction of these
species and/or their dissolution in condensed moisture
caused the following three phenomina to be observed:

1. In one case the gas sample was passed through

a water trap and then through a CaClz drying
chamber. This gave very low 802 content in the

exit gas. The values were much lower than the

 

erld ISFEJBB‘I‘

LO
0

75
equilibrium concentrations of 802 over the
solutions being used. Thus, these values were
not acceptable.
In another case the gas after passing through a
water trap was bubbled through a concentrated
solution of H2804 and then passed into the CaCl2
drying chamber. For the same exit gas as used in
“1" the 302 concentrations recorded were greater
than those observed in “1" and agreed at times with
the SO2 concentrations in the exit gas computed
via a sulfur mass balance on the liquid. Most of t
the time, however, the values of the SO2
concentrations in the exit gas were still too small,
differing by more than 50%.
In a third case the water trap was removed, leaving
only the concentrated HZSO4 trap and the CaClz
chamber. The $02 concentrations observed in this
case were still about 25-50% low compared to those
computed from a sulfur mass balance on the liquid
stream. The transmittance observed, however,
would make large jumps at times even exceeding
the $02 concentration observed in the inlet gas.
It was speculated that this was due to condensation
of the water vapor in the sample gas and allowing
NH3 and 802 to react and to disolve in it. As
this liquid would drip into the HZSO4 in the first

trap, the 802 would be driven off giving an

.1

gr. "--’A. m
a

76

observed large SO2 concentration in the exit gas.
With these problems in the analysis of the exit gas
stream with the I.R., it was decided to compute the SO2
content in the exit gas solely by a sulfur mass balance

on the inlet and exit liquid streams from the column.

i:"_“‘""."“."""£!

APPENDIX C

Presented here is a complete listing of the
concentration vs. time data for the liquid stream from
all the experiments performed. All concentrations are

in gram-moles/liter.

EXperiment No. 4415

Feed = 150 ml/min
Gas = 12.88 cfm

+

_ “2.41

Time (min) NH4 H803 ' 803‘ 804' S/Ceff
20 5.731 1.397 1.999 .168 .630
40 5.618 1.417 1.917 .184 .635
In 60 5.548 1.459 1.832 .212 .642
Out 60 5.521 1.819 1.583 .272 .682

Experiment No. 4-2

Feed = 150 ml/min
Gas = 12.88 cfm

20 5.811 1.828 1.688 .304 .676

40 5.796 1.825 1.688 .296 .676

60 5.779 1.808 1.656 .379 .675
Out 60 5.758 2.122 1.374 .419 .719
Experiment No. 4-12

‘ Feed = 185 ml/min
Gas = 12.88 cfm

20 4.176 1.393 1.230 .161 .681

40 4.163 1.402 1.210 .170 .683

60 4.150 1.419 1.171 .195 .689

77

I Ill-'11 II.“ I I
Joli ‘JJI'III 1|

Experiment No. 4-3

Feed = 185 ml/min

Gas =

Time (min)

12.88 cfm

NHT-

4

6.029
6.012
6.001
5.981

H303

1.900
1.870
1.829
2.095

Experiment No. 4-14

Feed = 185 ml/min
Gas = 12.88 cfm

20 6.181 1.385
40 6.134 1.427
60 6.066 1.459
Out 60 6.032 1.752

Experiment No. 4-7

Feed = 250 ml/min
Gas = 12.88 cfm

20 4.401 1.555
40 4.379 1.559
60 4.362 1.563

4.356 1.754

Out 60

Experiment No. 4-9

Feed = 250 ml/min

. Gas = 12.88 cfm

20
4O
60

4.277
4.273
4.273
4.264

1.449
1.466
1.487
1.672

Out 60

Experiment No. 4-17

Feed = 250 ml/min
Gas 2 12.88 cfm

20 4.842 1.444
40 4.792 1.459
60 4.755 1.489
Out 60 4.743 ‘ 1.688

78

803‘

1.752
1.729
1.704
1.544

2.228
2.143
2.090
1.916

.955
.938
.921
.801

1.036
.994
.972
.864

1.519
1.475
1.424
1.304

80 r

.312
.343
.381
.491

.170
.210
.213
.224

.468
.472
.478
.499

.379
.410
.421
.433

.180
.191
.210
.224

S/Ceff
.676
.676
.675
.702

.618
.625
.629
.657

.724
.727
.730
.760

.706
.712
.717
.746

.661
.666
.672
.696

79
Experiment No. 4-19

Feed = 280 ml/min
Gas = 12.88 cfm

Time (min) NH4+ n303' 303: $64: s/ceff
20 5.154 1.475 1.690 .150 .660
40 5.096 1.475 1.641 .170 .655
60 5.072 1.471 1.613 .188 .657
Out 60 5.057 1.673 1.492 .201 .680
Experiment No. 4-28
Feed = 280 ml/min
Gas = 12.88 cfm
20 4.933 1.533 1.392 .308 .678
40 4.913 1.548 1.360 .322 .681
60 4.894 1.581 1.317 .327 .688
Out 60 4.884 1.768 1.199 .358 .712
Experiment No. 4-26
Feed = 185 ml/min
Gas = 6.44 cfm
20 5.961 1.323 1.874 .445 .631
40 5.907 1.316 1.862 .433 .631
60 5.847 1.284 1.854 .427 .629
Out 60 5.825 1.436 1.762 .433 .645
Experiment No. 5—2
Feed = 185 ml/min
635 = 6.44 Cfm
20 5.625 1.554 1.790 .246 .651
40 5.637 1.512 1.832 .230 .646
60 5.644 1.491 1.820 .257 .645
Out 60 5.626 1.644 1.723 .269 .663
Experiment No. 4-27
Feed = 185 ml/min
Gas = 9.66 cfm
20 5.458 1.208 1.931 .194 .619
40 5.359 1.219 1.882 .188 .622
60 5.349 1.233 1.868 .190 .624
Out 60 5.325 1.445 1.743 .197 .647

an“. -—~ .3"-
I

Experiment No. S-ZA

Feed = 185 ml/min
Gas = 9.66 cfm

. 3 . ._ + -
Time (min) RNA H803
20 6.856 1.910
40 6.841 1.845
60 6.825 1.776
Out 60 6.803 1.975
Experiment No. 5-1

Feed = 185 ml/min
Gas = 11.28 cfm

20 5.934 2.423
40 5.966 2.384
60 6.020 2.328
Out 60 6.010 2.533

Experiment No. 5-5

Feed = 250 ml/min
Gas 2 9.66 cfm

20 5.756 2.375
40 5.875 2.297
60 5.963 2.233
-“ 80 5.979 2.216
Out 80 5.971 2.359

80

803‘

2.040
2.074
2.098
1.949

1.254
1.266
1.297
1.157

1.372
1.426
1.467
1.484
1.388

SO4

.433
.424
.427
.439

.501
.524
.510
.580

.318
.363
.398
.398
.417

lCeff

.659
.654
.649
.668

.746
.742
.736
.761

.732
.723
.716
.713
.729

APPENDIX D

Methods to Calculate the Physical Properties of

NHAHSO3-(NH4)ZSO3-(NH4)2804 Solutions

1. Specific gravity

The specific gravity was computed from an empirical
equation presented by Chertkov (19) and was given as
follows:

SP gr = 1.0 + .482(803=) + .4(H803') +,474(304:)

where all the concentrations are in grams/liter.
2. Viscosity
The viscosity was computed by an empirical equation

also developed by Chertkov (19) and was given as follows:

ViSCOSity (Cp) = .065(SO3:)(1.+.016K)+.036(H303-)
(1.+.100K)+.051(804:)(1.+.015K)

K e (303:) + (HSOB') + (304:)

where the concentrations are in gram-moles/liter.

3. Solution Molecular Weight
The following method was used to compute the solution
molecular weight. A basis of 1000 ml of solution was

taken. All concentrations are in gram-moles/liter.

81

82

Grams solute/1000 ml (GS) = (NH4HSO3) x 99.
+ ((NH4)2804) x 132.

Grams H20 (GHZO) = 1000 Sp gr GS
Solution Molecular Weight (SMOLWT) a
1000. Sp gr/(K + GH20/18)

K is defined under Viscosity. g

 

Moles H20 per mole Solution (RMOL) = f i
GHZO/18./(K + GHZO/18.) . 7

4. Conversion of solution concentrations from g-moles/liter
to moles/mole H20
Conversion factor (CONF) = SMOLWT/1000./sp gr/RMOL

5. Conversion of solution flow rates from ml/min to
lb-moles/hr of water
Conversion for Liquid (CLIQ) = Sp gr/SMOLWT x 60/
454 x RMOL

Table 11 contains the values of the solution properties

as used in all calculations for the various experiments.

6. Computing the diffusivity of 302 in air and NH3 in
air at 50°C
. The method of Wilke and Lee (20) was used.to make this

calculation which was done as follows:

Eggmple SO2 in Air

2 LI
DG (cm /scc

Giving

For NH3-air

Do

)

ll

83

 

3 2
BT “11/141 + 1mZ

 

2
P r1,2 I1)

(10.7-2.46 AJl/M1 + 1/M2 ) x 10'

 

64

29.2

10.15 x 10"4
3.954

1 atm

.5374

.1564 cm2/sec

.3524 cm2/sec

7. Air viscosity was used as .091 cp at 50°C and 1 atm'

pressure

4

9.- —-—-¢_—.—-.___..._. ._
- _.'
‘

84

Table 11. Properties of the Various Solutions Used

during the Experiments

 

5-5

Experiment Sp gr Viscosity Molecular Moles H20
No. (0P) Weight Moles

Solution
4-15 1.18 1:23 24.95 .924
4-2 1.25 1:19 25.57 .917
4-3 1.26 1:19 25.86 .914
4-12 1.17 1.14 23.24 .942
4-14 1.26 1.19 25.55 .918
4-7 1.18 1.14 23.68 .938
4-9 1.17 1.14 23.49 .940
4-17 1.20 1.15 23.97 .934
4-19 1.21 1.16 24.32 .931
4-28 1.20 1.16 24.26 .931
4-26 1.24 1.18 25.16 .923'
5-2 1.24 1.18 25.03 .923
4-27 1.22 1.17 24.45 .930
5-24 1.30 1.22 26.86 .904
5-1 1.27 1.20 .26.39 .907
1.27 1.20 26.19 .909

APPENDIX E

A. Method used to compute the 802 present in the exit
gas stream
S02 taken up by the liquid (M802)
= F x 60 x (Kout - Kin) (g-moles/liter)
F = liquid flow rate in liters/min
K.= (H303 + 303‘ + 304’) p
where all concentrations are in gram-moles/liter

In and out refer to in and out liquid streams,

 

respectively.
80 in the exit gas (ppm) = SO in the inlet gas
2 2
(ppm)
.4- M302
. gram-moles total
gas fed/hr

B. Method used to compute the NH3 concentration in the

 

exit gas
NH3 lost by liquid (MNHB) = F x 60 x (Cin - Gout)
C = gram-moles NH4+/liter
NH3 in the exit gas (Ppm) = MNH3 I
gram-moles total gas
fed/hr

C. Computing the predicted oxidation rate and the observed

 

oxidation rate 6
gramsr O Q”as/Ceff
10G:( 2) a 08 e“
at 50°C a .- 1/50 .. 1

85

 

86

Q a liquid flow rate, m/hr
Q = solution density, kg/m

.U-= solution viscosity, kg-SCC/mz

The following conversions were made on the data taken

from experiments to use the above equation:

1. Measured Q in lbs/hr-ftz, to convert to m/hr use.

I . 1 meter
Q a Qmeasured I P (3. 29 ft)

9 is the density in lbs/ft3

2. In the experiments 9 was used in lbs/ft3 to

convert to kg/m2

3. In the eXperiments.u.was used in lbs/ft-sec, to

convert to kg-sec/m2

2 .
J$=L$L888d x (;£§%_Eg)x sec ) x (§;Z%_££)2

x(32.2 ft
To convert G from grams OzlmZ-hr to lbs-moles SO43]

ftZ-hr the following was used:

 

 

(lb-moles) G. grams O2 1 gram-mole 02
GM = x___
hr-ft2 nl-hr 323

2 moles SO

4

2
1 mole 02 x (3. 29m ft )

 

X

lb-moles 804'I
AA ( hr ) G:(Specific surface of the

packing) x volume of packing

11A. a G a1; D2 H

 

where a a specific surface, ftz/ft3

D a column diameter, ft
H a column height, ft

I-n_umq

87

2. Computing AA from the experimental data

AAex = Fx60x(SO

SO43: the 804: concentration in gram-
‘ moles/liter

F = flow rate in liters/min

APPENDIX F

A. The following computer program was used to compute

the values of HOG for the absorption of $02 from the

various experiments. The following values of the column

variables have to be read in:

CFM
YM
T

CB
AM

DTST

DFSO3

D504

CBCL

FEED

WMDL
' DENS

ft3/min of gas fed to the column at temperature T
802 cone. in the inlet gas mole fraction
temperature of the gas and liquid streams in
degrees Kelvin
the molar cone. of NH4+ leaving the column
the molar conc. of'NH4+ leaving the column tied
up with the $04:
molar increase in the total sulfur conc. on a
pass through the column
sum of the molar cones. of 803: and HSO3' leaving
the column .
molar cone. change of NH4+ cone. on passing
through the column that is tied with 804:
molar conc. change of N114+ cone. on passing
through the column
feed to the column in ml/min
solution molecular weight
solution Specific gravity

88

0000000

983

92
12

90

89

N=]

N EQUALS THE NUMBER OF EXPERIMENTAL

DATA POINTS TO BE COMPUTED

D0584NSTART=1,N

READ(60,2)CFM,YM,T,CB,AM,DTST

FORMAT(5F8.4)

READ(60,983)DFSO3,CBCL,DSO4,FEED

FORMAT(4 f10.6)

GRML=SOLUTION CONC IN GRAMS/LITER

GHZO=GRAMS HZO/LITER

wM0L=NOLECULAR WEIGHT OF THE SOLUTION

HZOMOL=MOLES HZO/MOLE SOLUTION

CONF=CONVERSION FROM MOLAR CONC

To MOLEs/NOLE H20

CLIQ=FLOW RATE IN MOLEs HZO/HR .

GRML=(2.*DFSO3-CB+AM)*99.+AM/2.*132.+(CB-AM-DFSO3)*116.

DENS=1.0+(CB-AM-DFSO3)*.116*.482+(2.*DFSO3-CB+AM)
*.099*.4+AM/2.*.132*.474

GHZO=1000.*DENS-GRML

WMOL=1000.*DENS/(DFSO3+AM/2.+GHZO/18.)

H2MOL=GH20/18./(DFSO3+AM/2.+GH20/18.)

CONF=WMOL/1000./DENS/HZOMOL

CLIQ:FEED*DENS*60./WMDL/454.*H2QMOL

CB=CB*CONF

AM=AM*COMF

CBCL=CBCL*CONF

DSO4=DSO4*CONF

TSO3=DFSOB*CONF

BLB=CLIQ

ZCCB=CB

CCB=CB

AMM=AM

SB=T803

FM=EXP(2.303*(5.865-2369/T))

FEED=FEEDI1000.

ABSOZ=DTST*FEED*60./454.

BzYM/llOO o

AIRMOL=CFM*45.68/T

SOZMOL=AIRMOL*YM

YO=YM~ABSOZ/SOZMOL*YM

BEGIN TRAPAZOIDAL INTEGRATION

YSZ=YM

N=1

CB=ZCCB-AM

C2=CB

SZ=SB~SOZMOL*(YM-YSZ)/YM/BLB+(YM-YSZ)/(YM-YD)/2.*DSO4

‘YSEzFM*.1315*(2.*SZ-C2)**2/(CZ-SZ)*(C2+AM/2.)/C2

IF(YSZ-YSE)90,90,91

B=B/1.25

CB=CCB

AM=AMM

GO T092

90

91 HFS=1./(YSZ-YSE)
IF(N-1)10,10,11
10 HFSS=HFS
YSZ=YSZ-B
N=10
ZCCB=CCB+CBCL*(YM-YSZ)/(YM-YD)
AM=AMM-DSO4*(YM-YSZ)/(YM-YD)/2.
GO TO 12
11 HS=HS+(HFS+HFSS)*B/2.0
HFSS=HFS
YSZ=YSZ~B
ZCCB=CCB+CBCL*(YM-YSZ)/(YM-YO)
AM=AMM-DSO4*(YM-YSZ)/(YM-YO)
IF(YSZ-1.00I*YO)188,12,12
188 IF(.999*YO-YSZ)222,90,90
222 HOGS=4.0/HS
QLIQ=CLIQ/H20MOL*WMOL/3.14159*4.*144./3.75**2
QGAS=AIRMOL*29.2/3.14159*4.*144./3.75**2
OPER =QLIQ/QGAS
QLIQ AND QGAS Is THE LIQUID AND
THE GAS MASS VELOCITY RESPECTIVELY
WRITE(61,171)HOGS,QLIQ,QGAS,OPER
171 FORMAT(4F20.5)
END

91

B. The following computer program was used to compute

the value of N

0G for the absorption of 80% of the SO2 from

.an inlet gas containing 2000 ppm 802 for values of S/G=.7

and 1.4.

Calculations had to be started at the bottom of

the column since here all the variables could be fixed.

The flow of the program is as follows:

1.

The value of N06 for L/G=.7 is computed with the
following variables read in:

a) cfm = 12.88 ftB/min total gas flow at T=500C
b) YM = .002 mole percent 802 in inlet gas

c) T = gas and liquid temperature = 3230K

d) CB = .14 moles NH4+/mole H20

e) BLB = 1 lb mole H20 fed/hr

f) AM = .014 moles 804:/mole H20

g) R2 = .7 = S/Ceff

These variables represent the concentrations of the

various species at the bottom of the column. The calculations

ending at the top of the column give the concentrations of

all Species in the inlet liquid and exit gas.

2.

3.

The values of the concentrations of the various
species in the inlet liquid are stored and the
calculations of N0G were started for L/G=1.4.

A first guess has to be made at the exit
compositions from the column with L/G=1.4 and then
the calculations proceed up the column until 80%
of the $02 was absorbed. The inlet concentrations

of the various species are then compared to those

00000

000

666

92

12

92

computed for L/G=.7. If these agree the
computations are done, However, if there isn't
agreement, the concentrations of these Species
at the bottom of the column are adjusted
accordingly and the calculations repeated and
checked until the inlet concentrations of the

various Species matched the L/G=.7 case.

WRITE(61,666)

FORMAT(*1*)

D0584NSTART=1,2

THE FIRST TIME THROUGH THE DO LOOP GIVES THE
VALUE OF NOG FOR L/G=.7. THE SECOND .
TIME GIVES THE FIRST GUESS AT THE EXIT
COMPOSITIONS FOR THE COMPUTATIONS OF NOG
FOR THE L/G=1.4 CASE.
READ(60,2)CFM,YM,T,CB,BLB,AM,R2
FORMAT(7F8.4) '

YO=.0004

SB=R2*(CB-2.*AM)

STCB=CB

STSB=SB

STAM=AM

FM=EXP(2.303*(5.865-2369./T))
FN=EXP(2.303*(13.68—4987./T))
PNH3=FN*CB*(CB-SB)/(2.*SB-CB)*100.
A=.00001*PNH3/760.

B=YM/400.

THE VALUES OF A AND B ARE THE

VALUES THAT wILL BE USED TO

INCREMENT THE NH3 AND THE 302
AIRMOL=CFM*45.6/T

SOZMOL=AIRMOL*YM

CB=STCB

SB=STSB

AM=STAM

R2=SB/(CBo2.*AM)

HSSn0.0

INITIALIZE CONCS. OF 802 AND NH3 AT
THE BOTTOM OF THE COLUMN

YSZ=YM

YNH2=0.0

N=l

.BEGIN TRAPAZOIDAL INTEGRATION FOR NOG
ASBT=SB

BBB=YSZ

00

817
90
91

10

11

31
93
94

30

40

93

AAMTeAM

C2=CB+AIRMOL*YNH2/BLB
SZ=SB-SOZMOL*(YM-YSZ)/YM/BLB
SB=SB+.0007188LB88.7*(52/(C2-2.4AM))*868(HS-Hss)*HOGS
AM=AM~.OOO71*BLB**.7*(SZ/(CZ-2.*AM))**6*(HS-HSS)*HOGN
YSE=FM*.1315*(2.*SZ-C2+2.*AM)**2/(CZ-2.*AM-SZ)*(C2-AM)
. /(C2- 2.*AM)
YNE=F Nm' 1315 C2 (C2- 2. kAM-SZ)/(2. >82 CZ+2. *AM)
RT=SZ/(C2- 2. kAM)

SOZOXID=(SB ASBT)/(.000005*AIRMOL)
WRITE(61,817)YSE,YSZ,YNE,YNH2,RT,C2,SZ,AM,HS,SOZOXID
FORMAT(zx,10F13.6)

GUARD AGAINST THE INCREMENT B BEING

TOO LARGE

IF(YSZ YSE)90, 90, 91

B=B/1. 25

GO TO 92

HFS=1./(YSZ-YSE)

IF(N-1)10,10,11

HFSS=HFS

YSZxYSZ-B

N=10

GO TO 12

HSS=HS

HS=HS+(HFS+HFSS)*B/2.

HFSSxHFS

AAA=YSZ

IF(N.EQ.20)YNH2=YNH2+A

C2T=C2

SZT=SZ

YSZ=BBB

SBT=SB

AMT=AM

SB=ASBT

AM=AAMT

BEGIN TRAPAZOIDAL INTEGRATION FOR

NH3 LOSSES

C2=CB+AIRMOL*YNH2/BLB
SZ=SB-SOZMOL*(YM-YSZ)/YM/BLB
YNE=FN*.1315*CZ*(CZ-2.*AM—SZ)/(2.*SZ-CZ+2.*AM)
IF(YNE-YNH2)93,93,94 ' _
A=A/1.25

GO TO 92

HFN=1./(YNE-YNH2)

IF(N-10)30,30,40

YNH2=YNH2+A

HFNN=HFN

N=20

GO TO 31

HAA=HA

HA=HA+(HFNN+HFN)*A/2.
IF(HA-HOGS*HS/HOGN)50,51,51

000

00000

000

50

51

22

584

94

NH3 LOSSES ARE COMPUTED UNTIL THE

FOLLOWING CRITERIA IS MET

NOGNH3AHOGN=NOGSOZAHOGS

YNHZ=YNH2+A

YSZ=BBB~(BBB-AAA)*HAAHOGN/HOGS/HS

SB=SB+(SBT-ASBT)AHAAHOGN/HOGS/HS

AM=AM+(AMT-AAMT)AHAAHOGN/HOGS/HS

HFNN=HFN

GO T031

YNH2=YNH2-A

SB=SBT

AM=AMT

HA=HAA

YSZ=BBB

IF(YSZ.LE.YO)GO To 22

YSZ=AAA-B

BBB=AAA

GO TO 12

WRITE(61,666)

IF(NSTART.EQ.1)FCZT=C2T

IF(NSTART.EQ.1)FSZT=SZT

IF(NSTART.EQ.1)FAMT=AMT

THE ABOVE THREE STATEMENTS STORES

THE INLET CONCS OF THE SPECIES IN THE

LIQUID STREAM. THESE VALUES MUST

BE MATCHED WHEN THE INTEGRATIONS ARE

CARRIED OUT FOR L/G=1.4

NC2T=FCZT*10.**6

NSZT=FSZTA10.*A6

NAMT=FAMT*10.**6

IF(NSTART.EQ.1)GO TO 584

CONVERENGE METHOD TO CORRECT

EXIT LIQUID COMPO SO THAT INLET

LIQUID COMPO WILL BE MET

KCZT=CZT*10.**6

KSZT=SZTA10.*A6

KAMT=AMTA1O.**6

1F(KCZT.NE.NCZT)STCB=STCB+FCZT-CZT

IF(KSZT.NE. NSZT)STSB=SISB+FSZT-SZT

IF(KAMT.NE. NAMT)STAM=STAM+FAMT-AMT -

IF(KC2T. EQ. NCZT. AND. KSZT. EQ.NSZT. AND. KAMT. EQ. NAMT)
GO TO 584

GO TO 92

CONTINUE

END

HICHIGRN STQTE UNIV. LIBRQRIES

312930083081 10