-------
                                                  4)
                                                  Q.
                                                  Q.
                                                  (A

                                                   I
                                                  o
                                                  o

                                                  In
                                                 o

                                                  0)
22

-------
                          S02



!
II
<

1
l(

\

\
1
1




D1


D1

i

I
0'
•

J^


^

\/\
Y
A
/\

F
*Y
1 IACIDULATOR
- — i
i
.^— <^UI FURIC ACID
i
ABSORBER
_ ,. . - ppr>nnr;T
— txi J LIQUOR
\ STEAM 	 1
ALTERNATE " ~~ WATPR RATH
STRIPPING 1!-JL WATER BA™
GAS ~ —
/CONNECTIONS
-txf
4"DIA
— tXKSTRIPPING AIR
AMMONIUM
       SULFATE
       (SOL'N)
Figure II. Final acidulator-stripper
               23

-------
    As shown in Figure 9, the ammonium sulfate solution from the
stripper flowed by gravity to a surge-pump tank and was metered
to the evaporator-crystallizer (Goslin Inc.,  Birmingham).   The
evaporator-crystallizer, made of 316L SS, was designed to remove
200 Ib H20 per hour from a saturated ammonium sulfate solution at
about 170°F and under a 22-in. mercury vacuum.  Heat was added to
the system as steam in a tube-and-shell heat exchanger in the
external recirculation loop.   A constant-speed centrifugal pump
was used to recirculate the brine.   A water-cooled condenser and
a steam ejector were used to condense the water vapor and to
maintain the internal pressure in the system.

    A slurry of ammonium sulfate crystals was removed from the
evaporator-crystallizer and pumped to a 6-in. screen bowl
centrifuge  (Bird Machine Co.).  The crystals were separated and
could, if desired, be dried in a rotary gas-fired dryer.  The
centrate was returned to the evaporator-crystallizer.  Crystals
from the ammonium sulfate separation step were to have gone to
the ammonium sulfate decomposer which was designed but not
constructed.

Vent System

    Exit flue gas from the absorber was vented to the atmosphere
for plume observation.  The recovered S02 from the acidulator and
stripper was vented into the power plant stack.

Instrumentation

    The pilot plant was instrumented throughout so that all
pertinent liguid and gas flows were monitored and values were
recorded.  All signals were electrically transmitted from the
sensing elements to the recorder-controllers.

    The gas flow through the absorber system was monitored with a
differential pressure cell (The Foxboro Company), which sensed
the pressure differential across a flange orifice in the gas duct
from the absorber and sent a signal to a recorder-controller.
Any deviation from the preset values on the controller-recorder
caused a signal to be sent to the variable-speed drive mechanism
on the induced-draft blower to correct the deviation.  This
arrangement assured that a constant gas flow through the absorber
system was maintained.

    The SO2 concentrations in the gas to the absorber and after
each stage were monitored with an ultraviolet analyzer  (DuPont's
460 Photometric analyzer).  The analyzer has three ranges of S02
values:  0-4000, 0-1000, and  0-100 ppm full-scale reading.  The
sample selection was changed manually from station to station to
avoid the possibility of leaks from an automatic sample
seguencing system.  Periodic checks by wet-chemical analysis
                                24

-------
methods confirmed the analyzer readings.  Steam-traced and
insulated sample lines (DeKoron)  made of Teflon were used to
bring samples to the analyzer.

    A smoke detector was used to monitor the opacity of the plume
at the stack exit.  The instrument, manufactured by Photomation,
Inc., used a light source and a photocell to measure the plume
opacity.  The digital readout was in Ringlemann units.

    Gaseous NH3 was metered to the system as required with a
Foxboro differential pressure cell coupled with a recorder-
controller and a flow control valve.  Liquid flows were sensed
with magnetic flow meters which sent electronic signals to
recorder-controllers.  The required flows of recirculating liquor
to the first three absorber stages were maintained with variable-
speed pumps.  Variable-speed pumps instead of valves were used
for flow control because flyash removed in the bottom stages
could cause plugging and erosion of control valves.  Automatic
flow control valves were used to control the flow of the
remaining liquid streams.  Air for S02 stripping was metered to
the stripper through rotameters.  Temperatures throughout the
system were sensed with thermocouples and recorded on strip
charts in the control room.   (See Appendix D, Equipment
Evaluation.)
ABSORBER TEST PROGRAM AND RESULTS

Plume Theory and Control Program

    Emission of a dense plume from the absorber has plagued all
the EPA-TVA pilot-plant work and has been present in full-scale
operations that use NH3 to remove S02 from process gases.  Figure
12 shows a typical plume from the pilot plant during routine
operation.  The plume contained fume particles having a mean
diameter of 0.25y.  About 10* particles per cubic centimeter are
in the size range of 0.005 to O.Oly.  Figure 13 shows the number-
size distribution and Figure 1U shows the mass-size distribution.

    The plume particles were identified chemically as
predominantly ammonium sulfate with varying amounts  (up to 30%)
of ammonium chloride.  The contribution of ammonium chloride
 (NH4C1) to the plume problem was not determined although bench-
scale tests, discussed later, did show that chlorides as well as
SO3 could contribute to the problem.  Flyash also was present in
samples of the particulate caught in a modified cascade sampler
 (Brink sampler) .

    Conventional wet scrubbers are ineffective in removing
extremely small particulate from gas streams.  Proposals have
been made and actual installations have been built that use
electrostatic precipitators or mechanical collectors  (impaction
                                25

-------
Figure 12. Typical  plume from pilot plant,
                   26

-------
or
UJ
o
o:
-------
                            0.4        0.6        0.8
                         PARTICLE  DIAMETER, yu.
1.0
Figure 14. Moss-size distribution of particulote in  absorber plume
                              28
-------
 devices  and bag houses)  to remove plume from the washed gas.
 Unfortunately,  the  electromechanical methods are expensive and
 account  for a large fraction (about 25%)  of the gas cleaning
 cost.

     An approach to  plume control that avoids the need for fine
 particulate removal equipment was proposed by M. L. Spector and
 P.  L. T.  Brian (21)  of Air Products and Chemicals, Inc., and
 modified  later by N.  D.  Moore of TVA (see Appendix E).   The basis
 of  the proposal is  that  a gas-phase reaction will produce a solid
 (fume) whenever the product of the partial pressure of the gases
 involved  exceeds the equilibrium constant for the reaction.  The
 problem  then is to  avoid absorber operating conditions that allow
 the gas-phase reaction to occur.

     The  reaction constant k for a gas-phase reaction can be
 expressed as a product of the vapor pressures of the
 constituents.   For  a compound of the form A2B, the equation for
 the reaction constant is
               k =  (PA)MPB)
                                        (8)
     The  equilibrium constant also can be related to the heat of
 reaction by
               d (In k)   AH                            (9)
                  dT     RT2
where  /\ H =  heat of  reaction,  calories/g mol
         R =  gas  constant,  1.987  calories/(g-mol)(  K)
         T =  temperature,  °K
Assuming AH  is constant,  integration of equation 9 expressed in
terms  of Iog10 gives
               logtok = -  AH
                                        (10)
                         TR(lnlO)

where I is the integration constant.

      The  above  equation shows  that the value for the equilibrium
 constant  for  the  gas  phase  reaction is a function of temperature.
 Spector and Brian evaluated data  from in-house experiments and
 concluded that  the gas  phase reaction product—fuiae particle—is
 solid ammonium  bisulfite, NHi»HSOs, and equation 10 is of the form
Iog10k = -17,300
             T
                                  31. U
(11)
 where T is in degrees Rankine and k is in mm of mercury.
                                29
-------
    Moore evaluated data by St. Claire, Earhart, U.S. Bureau of
Standards, and Air Products and Chemicals, Inc.  He concluded,
from the typical ammonia absorption pilot-plant operating
conditions, that the fume particles were ammonium sulfite
monohydrate [ (NH4) 2S03.H20] and that equation  8 has the  form
              k =
and equation 11 has the value

              Iog10k = -16,520/T + 53.35               (13)

where T is in degrees Kelvin and k is in mm of mercury.   Moore's
treatment of the data and the analysis leading to this conclusion
is included in Appendix E of this report.

    Taking the logto of both sides of equation 12 and solving  the
log10P   ,  equation 12 becomes
                         Ok - 2 logto  (PNR3)

where, from Johnstone  (6) ,

              PNH = Sly 2A)(C  - S)                  (15)
                "" 3     "  / -50 Z7*^ 1
                                A

              P H n=     100 Pw	                (16)
                "2U  (100 +. CA +S + 3A)


A in the above equations represents the concentration  of
 (NH4)2S04  (mols per 100 mols of water)  in the  liquor.

    Moore derived an expression for ^, the vapor  pressure of
pure water over the temperature range  of 35 to 60°C  as

              lo
-------
Absorber Tests

Pilot Plant—

    A study plan was developed to test the concept of avoiding
the fume problem by controlling the vapor pressures of the
gaseous constituents necessary to form the ammonia and sulfur
compound fume, namely, ammonia, S02, and water.  The study was
carried out under operating conditions indicated by Phase II work
to be necessary for fume control.  These conditions were:

         •  A water wash ahead of the absorber

         •  A water wash after the absorber

         •  Reheat of the scrubbed flue gas

    The water wash ahead of the absorber (prewash section) was
used to cool and humidify the gas to prevent localized
evaporation of absorber liquor, which would increase the vapor
pressure of S02 or NH3 to a point where gas-phase precipitation
would occur.  The prewash operation also would remove chlorides
and flyash from the flue gas.

    In the first prewash unit, which was made of type 316L SS,
flue gas entered through a duct in which it was cooled and
humidified with a water spray.  It then impinged on the surface
of a water recirculation sump from which it flowed to the
absorber.  The unit proved unacceptable because of severe mist
carryover and because of excessive corrosion by the recirculated
liquor  (pH 1-3).

    Condensed data from these first fume control tests appear in
Appendix Fr Table F-l.

    A second prewash unit made of corrosion-resistant material
and equipped with a mist eliminator was developed and installed.
A settling tank also was installed to remove undissolved solids
from the recirculating liquor in the closed loop prewash section.

    The prewash section was tested with "clean" gas  (0.05-0.29 gr
flyash/scf) at L/G ratios of 10 and 20 gal/1000 ft3 and at Ap's
of 5, 10, and 15 in. H20 across the venturi element.  However,
because humidification was not complete at the lower L/G ratio,
most of the tests were made with L/G of 20 and ftp's of 10-15 in.
H20.  Under the conditions, humidification was essentially
complete.  Chlorides  (as HCl) were decreased in the prewash
section from about H5 ppm to 3-7 ppm.  This level of chlorides
was not expected to cause chloride fuming.  "Clean" gas from
downstream of electrostatic precipitators was used in all tests.
The exit particulate loading from the prewash section was 0.05-
0.2 gr/scf, essentially the flyash loading to the prewash
section.
                                31
-------
    Both dissolved and undissolved solids in the recirculating
prewash liquor increased with operating time.  (Figure 15 shows
the undissolved solids versus time, and Figure 16 shows the
dissolved solids versus time for the liquor from the settling
tank clarifier to the prewash loop.)   The solids content had not
leveled off at the end of the 9-day sample period, the longest
continuous operating period without dilution of prewash liquor.
However, replacement of the water lost in the periodic purge of
flyash from the settling tank was expected to cause the solids
content to level off near the uppermost values shown in Figures
15 and 16.  The pH of the liquor in the system was typically 1.0,
Chemical analysis of "typical" samples of the prewash liquor is
shown in Table 1.
          TABLE 1.   CHEMICAL ANALYSIS OF PREWASH LIQUOR
                   Test NO.            FGW-5
                   Analysis, g/1

                   Total Fe             3.4
                         Fe203          1.0
                         FeO            3.5
                   Total S as SOi*       21.1
                         Cl             4.5
                         Na             0.08
                         K              0.19
                         Ca             0.73
                         Al             0.75

                   Solids, %

                   Dissolved            3.1
                   Undissolved          0.01
    The mist eliminator, which had been installed in the
horizontal duct between the venturi section and the absorber, was
highly effective; only traces of water carryover could be found
during air-water tests.  Tests by EPA personnel during operations
with flue gas found mist carryover to be 1 ml/m3 of gas, which
was an acceptable level.

    The low degree of mist carryover also would lessen the
contamination of absorber product liquor by the heavy metals
                                32
-------
en
o

_j
O

co co

o
       1.5
1.0
0.5
                             T
          • TO SETTLING TANK


          O  FROM  SETTLING TANK
                                                 8
                                                    10
                          DAYS SINCE START-UP
 Figure 15. Undissolved solids in prewash  liquor to and from  settling

         tank vs. time.
                             33
-------
     120
     100
     80
 to
 o
      60
 0
 UJ


 o
 CO

 S    40
      20
       0
                   O
                                  o
o
                             468


                        DAYS SINCE  START -UP
     10
Figure 16. Dissolved solids in prewash liguor from settling tank vs. time.
                                  34
-------
contained in the flyash.  In Phase II work it had been found that
the presence of ferrous ammonium sulfite hindered the separation
of crystalline  (NH4)2S04 in the regeneration step of the process.

    Plume control tests were made following the prewash section
tests.  In these tests the concentration of the absorber product
liquor was to be 12 mols active NH3 per 100 mols total water (C A
= 12) and a sulfur to active NH3 mol ratio of 0.80  (S/CA = 0.8) .
These values were selected because:

         The concentration was sufficiently high for regeneration
         purposes; about 2.5 Ib of water is evaporated for each
         pound of recovered S02.

         The ratio of NH4HS03 to (NH4) 2S03 was 4:1, which
         minimized the amount of NH4HS04 needed for acidulation
         [2 mols of NH4HS04 are required per mol of (NH4)2S03 and
         1 mol is required per mol of NH4HS03--see equations 5
         and 6 ].

         The equilibrium vapor pressure of S02 above the absorber
         product solution is such that a large driving force
         exists between the solution and incoming gas to permit
         good S02 removal on the first stage (the equilibrium
         vapor pressure of S02 was 0.93 mm Hg while the vapor
         pressure of S02 in the inlet flue gas was 2.12 mm Hg).

         The S02 equilibrium vapor pressure is below the fume
         level as predicted by equation 18.

    The desired S/CA of the liquor on stage G-2 of the unmodified
four-stage absorber was 0.72.  This value was selected because it
is compatible with the product liquor composition and, because
under the conditions expected on stage G-2, equation 18 predicts
fumeless operation.

    The absorber product liquor CA was controlled by the amount
of makeup water added to stage G-4.  The S/CA's of stages G-l and
G-2 were controlled by the amount and point of addition of makeup
ammonia; ammonia could be added to either G-l or G-2 or both.
The liquor composition on G-3 was determined by the requirements
of G-l and G-2.  Water only was added to G-U.  However, through
absorption of S02 and ammonia from the gas stream and by
collecting mist containing S02-NH3 salts from G-3, the liquor on
G-U at steady-state operating conditions had a CA of about 2 and
an S/CA of about 0.9.  The desired or expected liquor
compositions on absorber stages G-l and G-2 are summarized below:
                                35
-------
Stage
G-l
G-2
c.
12
10
S/C
0.80
0.72
A"
1.75
1.50
              a.   "A" values, mols of sulfate
                  per 100 mols water, were
                  assumed for calculation
                  purposes.


    Using equation 18 the calculated S02 concentration for
formation of the ammonium sulfite monohydrate fume particle at
130°F on stage G-l is 6,255 ppm.

    Similarly, the equation predicts that a fume can form on G-2
when the S02 concentration to G-2 exceeds 2,399 ppm.  Figures 17
and 18 Show the calculated fume lines and equilibrium lines for
the above-listed absorber liquors over a range of temperatures
for stages G-l and G-2.

    For fumeless operation,  the vapor pressure of S02 must be
controlled within the area bounded by these lines.

    Most of the fume control tests were made at the high liquor
concentrations.  Some tests were made at low liquor
concentrations because, as can be shown by equation 18, formation
of a fume is less likely to occur at the low C^ *s than at the
high CA'S.  Condensed data from the tests appear in Appendix F.

    Control of the absorber liquor composition was difficult in
the four-stage valve-tray absorber  (before the use of mobile
spheres on the G-l and G-2 trays).  in one series of tests before
the absorber was modified (series AX) the absorber liquor CA
varied from a low of 11.6 to a high of 15.5, with 8 of the 11
tests having a CA of between 11.6 and 12.6.  A fluctuation of
+0.5 unit of CA does not greatly affect the vapor pressure of S02
and was considered acceptable for the test operation.  A
fluctuation of + 0.02 unit of S/CA has a large effect on S02
vapor pressure and can move the predicted fumeless operation to
the fume region.  For instance, the S02 concentration required to
cause fuming as predicted by equation 18 for a solution having a
CA of 12 and an S/CA of 0.80 and at 125°F is 5,800 ppm.  At the
same condition, with the S/C Adecreased to 0.78, the predicted
fume value is lowered to about 4,200 ppm.  Figure 19 shows the
effect of deviations of 0.5 unit c A and 0.02 unit S/C,  from the
fume line for a liquor having CA of 12 and S/C  of O.»0.

    Two of the 11 tests in the AX series  (tests AX-4 and AX-5)
had observed plumes of 5% opacity or less; all of the others had
stack opacities greater than 5%.  According to the fume
prediction equations, all tests should have produced fumes

                                36
-------
   8000
cc
o
Q.
     1000
                   EQUILIBRIUM SOz

                   VAPOR  PRESSURE
        125
      130

TEMPERATURE,°F
135
 Figure 17. Equilibrium SOa vapor  pressure with respect to the fume

          line  for G-l stage.
                            37
-------
6000



5000



4000





3000
 CM

O
O  2000
tu
o:
ID
en
-------
UJ
cr
UJ
cc
QL

or
o
0.
   4000
   3000
2000
        120
                              D
                                  i
                                           OPERATING CONDITIONS
                O   CA = 12.0, S/CA = 0.80
                        (DESIRED)

                *   CA = 11.5, S/CA =0.80

                A   CA = I2.5,S/CA= 0.80


                O   CA =  12.0, S/CA= 0.82


                D   CA =  12.0, S/CA= 0.78
      125
TEMPERATURE, °F
                                                       130
 Figure 19. A theoretical fume line with points demonstrating the sensitivity

          of fume values to  deviations of  CA and S/CA •
                                 39
-------
(Appendix F).   No explanation for the low opacity in tests AX-4
and AX-5 could be determined from the data.  The temperature of
the gas to the prewash section and the temperature of the
saturated gas to the absorber were lower in these tests than is
normal—less than 200°F as compared with a "normal" temperature
of 250-300°F.   Several tests were run with inlet gas temperature
of 200°F or less without reproducing the results of tests AX-U
and AX-5.

    The calculated fuming occurred on stage G-2 in all tests
because the S/CA was below the desired value on G-2 or above the
desired value on G-l.  A high S/CA (>0.80) on G-l results in less
SO2 removal on G-l and more SO2 to G-2.  A low S/CA  (>0.72) on G-
2 results in a high ammonia vapor pressure and increases the.
likelihood of fume by increasing the product of the vapor
pressures of the constituents.  The overall S02 removal
efficiency in the four-stage valve-tray absorber was poor, and
additional NH3 was introduced to G-2 in order to reach the
arbitrarily set minimum of 80% S02 removal for this test series.
Murphree tray efficiencies for G-l and G-2 trays ranged from 13.2
to 95% with an average Murphree efficiency of 43.6% on G-l and
66.1% on G-2.

    Prior to the next series of tests  (BX series), the two bottom
stages of the valve-tray absorber were modified to improve S02
removal efficiency and to improve control of the liquor
concentrations on these two stages.  As stated earlier the
modification involved adding a i2-in. depth of 1-in.-diameter
hollow sphere  (5-g weight) to each stage to improve the mass
transfer of the stages.  The spheres were simply poured onto each
stage.  A 6-in.-thick wire mesh mist eliminator was anchored 1 ft
above the bed of spheres.  During operation, liquor from the
valve tray irrigated the spheres.  The wire mesh pads caught the
large volume of mist to prevent its being carried to the next
higher stage and disrupting control of the liquor concentration
on that stage.  The modifications increased the pressure drop by
2 in. of H20 on both stages G-l and G-2  (overall pressure drop
for the modified absorber was 12 in. H20).  Sulfur dioxide
removal efficiency in the modified absorber was improved--the
Murphree stage efficiency averaged 90% for G-l and 92% for G-2 in
the BX series.

    The desired values of CA and S/CA for the BX series were:

              Stage	CA__S/C, 	Aa
G-l
G-2
12
10
0.78
0.72
1.75
1.50
               a.   "A" values  assumed for
                   calculation purposes.
                               40
-------
    It was expected that, under these conditions, the liquor of
G-3 would have a CA of 2 and an S/CA of 0.80.

    The decrease in S/CA on G-l for the BX series was made to
increase S02 removal on G-l and decrease the quantity of S02 to
G-2.  Under these conditions and at a temperature of 125°F, 3,954
ppm of SO 2 are required to cause a fume, well above the maximum
inlet SO2 concentration of 3,UOO ppm observed during the test.

    The S/C, selected for stage G-2 was 0.72, the same as for the
AX series.  The calculated minimum S02 concentration for fuming
on G-2 under these conditions is 1,979 ppm.  This gives a safe
margin if reasonable Murphree tray efficiencies are achieved
because the equilibrium concentration of S02 to G-2 is 338 ppm.

    Data from the BX series also appear in Appendix F.  Control
of liquor composition was improved over previous work, although
in some tests the S/C,^ varied more than the 0.02 unit considered
acceptable for the tesh:.

    Stack opacities of 5% or less were observed during 6 of the
12 sampling periods.  However, inconsistencies in the test
results were apparent.  Shown in Table 2 are selected data from
three tests in the BX series.  Test BX-11 predicts no plume (5%
or less opacity) and the observed opacity was 5%.  A plume was
predicted for test BX-12 and a plume of 20% was observed.  In
test BX-10 a plume was predicted but the observed opacity was 5%.
No explanation has been found for the inconsistencies.

    One series of tests was run in the unmodified absorber in
which the washed gas was reheated from the absorber outlet
temperature  (about 135°F) to 225°F in 10° intervals.  The gas was
reheated with an in-line indirect tube-and-she11 heat exchanger
with 350 psig steam on the tube side.   (The overall heat-transfer
coefficient for the heat exchanger ranged from 1U.8 to 27.6
Btu/(hr) (ft2) (°F).  Data from these tests appear in Appendix F,
Table F-2.

    In none of the five tests with plumes at 185°F was the stack
opacity made acceptable by increasing the stack gas temperature
to 225°F.  These tests show that stack gas reheating is not the
answer to the ammonia-sulfur compound plume problem.


    Typical operating data for the absorption section are shown
in Table 3.
                              41
-------
   TABLE  2.   SELECTED DATA FROM BX SERIES
Test No.
Liquor concentrations
G-l
In
CA
S/CA
Out
CA

G-2 A
In
CA
S/CA
Out
CA.
S/CA
G-3
In
CA
S/CA
Out
CA
S/CA
G-4
In
CA

Out
CA
S/CA
Gas temperatures, °F
To prewash
To G-l
To G-2
To G-3
To G-4
Stack
Liquor temperatures, °F
G-l out
G-2 out
G-3 out
G-4 out
SO j concentrations, ppm
To G-l
From G-l
Calculated fume on G-l
To G-2
From G-2
To fume on G-2
To G-3
From G-3
To fume on G-3
To G-4
From G-4
To fume on G-4
Overall S02 removal, %
Plume opacity, %
Observer
BX-10



13.33
0.80

11.38
0.85


12.63
0.69

12.82
0.68


1.76
0.95

2.17
0.85


0.69
0.95

0.69
0.94

232
122
121
116
115
185

122
120
116
115

2,240
1,160
4,361
1,160
280
752
280
260
a
260
240
a
88.4

5
BX-11



12.53
0.79

-
-


10.53
0.72

-
-


1.52
0. 91

-
-


-
-

-
—

225
123
121
116
115
178

124
118
116
115

2,640
1,180
4,232
1,180
300
1,533
300
290
a
-
*~

89

5
BX-12



12.72
0.80

11.92
0.82


10.06
0.69

10.25
0.70


1.35
0.89

1.17
0.93


-
-

-
-

224
124
121
117
115
176

123
119
116
115

2,320
1,200
4,483
1,200
360
1,032
360
340
a
-
""

84

20
a.   Theoretical calculations show that  it  is
    impossible to fume at these tray  concentrations.
                          42
-------
     TABLE 3.  TYPICAL ABSORBER TEST DATA
Flue gas to prewash
  Flowrate, scfm at 32 F                 2358
  Temperature, °F                        225
  S02, ppm                               2640
  Flyash, gr/scf                         2
Flue gas to absorber
  Flowrate, acfm                         2800
  Temperature, °F
    Wet bulb                             123
    Dry bulb                             124
  S02, ppm                               2640
  Flyash, gr/scf                         0.5
Flue gas leaving G-l stage
  Temperature, °F                        121
  S02, ppm                               1180
Flue gas leaving G-2 stage
  Temperature, °F                        116
  S02, ppm                               300
Flue gas leaving G-3 stage
  Temperature, °F                        115
  S02, ppm                               290
Flue gas leaving G-4 stage
  Temperature, °F                        115
  S02, ppm                               290
Flue gas leaving stack
  Temperature, °F                        178
  S02, ppm                               290
  NH3, ppm                               20
Overall S02 removal, %                   89
Plume opacity, %                         5
Absorber liquors
  Temperature, °F
    To G-l                               123
    To G-2                               118
    To G-3                               116
    To G-4                               115
  Composition
    To G-l
      CA                                 12.53
      S/CA                               0.79
      pH A                               6.0
      Sp.gr.                             1.220
    To G-2
      CA                                 10.53
      S7c                                0.72
      pH A                               6.2
      Sp.gr.                             1.190
    To G-3
      CA                                 1.52
      S/Ca                               0.91
      pH A                               5.7
      Sp.gr.                             1.060
    To G-4
      CA                                 0.70
      S/CA                               0.90
      pH                                 4.8
      Sp.gr.                             1.025
                       43
-------
    The absorber was operated with low liquor concentrations to
determine whether fume would form with low ammonia vapor pressure
in the absorber.  (These tests were made at the beginning or end
of a high C* test series.)   Liquor with CA'S in the range of 0.5-
1.5 was proauced.  From equation 18, it would be predicted that
no fume is possible when operating with a CA of 1 and an S/CA as
low as 0.58.  Except for one isolated sampling period (test AX-
12), the observed plume opacities were 5% or less for all four-
staqe absorber operations at the low concentrations.

    The results of the plume tests at concentrations acceptable
for the ammonium bisulfate process  (or for production of
crystalline ammonium sulfate) indicate that the thermodynamic
equations are useful in predicting regions of fumeless operation.
However, satisfying the equations is a necessary but not total
requirement for fumeless operation.  Also, the equations do not
consider fumes that may occur from chlorides and S03 reacting
with ammonia.  Further developmental work on a large scale does
not appear warranted.

Bench Scale--

    A bench-scale sample train of six glass wash bottles was set
up to investigate fume formation.  The first two wash bottles
were used as a prewash section, the third was dry for mist
fallout, the fourth bottle was used as an ammoniacal liquor
absorption section  (NH4OH), and the last two were filled with
hydrogen peroxide for removing NH3 and S02 from the gas before it
passed into the gas flow meter.  Fuming sulfuric acid (20%
oleum), 20% hydrochloric acid, deionized water, and 80% isopropyl
alcohol solution were used as scrubbing media.  Clean flue gas
 (0.2 gr/scf) , dirty flue gas  (5 gr/scf) , bottled gas of known
concentration  (S02 span gas, 950 ppm S02), or air was pulled
through the sample train.  A filter system for particulate
removal was inserted at various points in the sampling train.  A
Gelman Instrument Company fiberglass "absolute" filter, type E,
was used in the filter system.  The filter removed  99.7% of 0.3y
particles and 98% of O.OSy particles.  Table n lists the various
test conditions and test results.
    A fume was formed in all tests with flue gas whenever the
absolute filter was excluded from the sampling train.  Whenever
the absolute filter was used anywhere in the system, a fume did
not leave the system; if the filter was located before the
ammoniacal solution, no fume formed; if the filter was after the
ammoniacal solution, a fume formed but was removed by the filter.
The prewash section prevented fume from forming for  15-20 min
                             44
-------
o
U
W
H
a
Cfl

»
u
,Z
Ed
sr



r-l
r»

C
o
• H
4J
in
O
a.


vo
C
0
•H
4->
•H
to
O
(X.


in
C
O
•H
4J
•H
n
o
04


c
0
•H
4J
••4
in
O
Cu


M
C
0
•H
4J
-H
W
o
CL,


CM

O
•H
4-1
• rH
I/I
O
ft.
r~t
C
o
•H
4-*
•H
Ul
O
a.


•o
Ul
3
10
<0
O


4J •
V) O
o >:
fr*



^^
0)
cTl
(MW
X
0
*> d
in — *


^-,
Q)
fSjg
O 3
.
fl)
g
3
M-l

>i O
M C
Q —

•C -~^
o 4)
•wg
X 3


dP O
0 3
<*> —


>-»
O
c
3
r-l M
fsj
O 01
IS VI
X C
01
dP-O
in —
^^
g
3
IM

41
n
c
>i 01
M3
Q —
I
EC 3
0»M
W 0
2 w
c
d» 0)
o »O
n "—


,^H
O
£
3
M m
4J O
ffl C
S-'


(U
P*
3
M ^H
o>
4J O
10 C




i




01
3
>M
10
41 in
r-l 10
O tn
03
r-*
1
in
0.
01
g
3
^^ _-. *w *— «
01 41 01
fMg CMg CM4-> fMg
o3 o3 ox o3
r*n-i CM*W (N O1 oju-i
X X X -rl X
O O t-4. O
d»r: dPC dPM dPC
in — m — in — • in —
0
g
*-^ ^^ IM >-^
41 4) 4)

o 3 o 3 ox o 3
(NJU-I CM**-! CM Ol CMU-4
X X X -H x
O O <-l o
dPC OP C OP W dPC
in — in— in— in —
•«
g
>— » .»-% *i-l ,— »
(1) Q) (U
6 G •*•* £
3 3 .45 3
*M IM QFV t^
•rl
^1 O »*1 O i>ir-4. >t O
V4C M C H Ul r4C
0— Q— Q— Q—
I
r, >H
X 41 X
O g O 4J
14 v9 •*£ M
4) X«H X tsn o>
•P Z Z -H 4J
rH O •-! r-l
•H dP C dP W -H
CL, in — in — CK
-J
B 3
X 3 *-- ^^ V4
O >M 41 41 X
f £ £ o 4J
X 0) 3 3 ^*X
ZVl MM-I r4^M XO*
CO 4) Z -H
dP 0) 4-> O 4-1 O r-l
0 T3 IOC IOC dPin
r>— S— S— in —


0) 01 4) 01
C F P g
3339
M*M 1M **-! M 4-^ m^M
0) O O Q)
•P O -UO -U O 4->O
«JC nJC «JC nJC
OJ ^)
§g
M 3
V| M-l (D ^4 H 1
41 4J 1 41
4J O -t 4-» O
10 C -H Ifl C
S — U. S —

(11 0) 41 Ol
3333
•W ^ *4 M^
C C C C
fl *0 flj ftj
OIVI 4IU1 0) Ul 0) U)
r-4IO r-llfl rHIO r-llfl
OCP OCP 00» 00-
U < CO O
r-t PM (N (N
1 1 1 1
ui ui en 03
Cu O. CL, CL,



^^
01
o 3
(NU-I
S
0
«> C
m *—


^^
o
{N 6
o 3
(NM-I
X
o
<*> c
in «-*


*— ,
Q)
g*
3
*M

>i O
M C
a^-

^^
X 4)
°J
X4 *4-l
41
4J 0
n) C
S —


41
pi
3
M U-l
o
4J O
n) c


r^
01
4J
r-l
•H
b.
x
o
10
in
4-1
»H in
•H 10
Q b>
<
n
1
CO
a
01


U4

) Ql
X -H
rH
dA 0)
in —
7
g
If4

CA| 4J
o x

X -r|
r-l
dp m
in ~-
»
3
*H

4J
£

•H
r^rH
M in
Q —
1
*w
X
O 4J
EC trP
.Z-rl
r-l
dP in
in —


^-^
4)
g
3
r4 *H
41
4J 0
10 C



41

3
r>4 M-l
41
4J O
10 C




1




41

IH
4J
r4 M
•r4 10

n
n
1
CO
CL,



^r^
41
(NE
O 3
C*M
X
o

in —


Jn
41

0*3
fMt-l
X
o
dP C
m —


^-.
41

3
IM

^ ft
M C
Q —


M
01
4J
I-l
•rl
Cu
"o
3
*M
X
0 4J

X O*
Z-rl
dp in
in —


4)
g
3
r-l 4-4
01
4J 0
fl C
01

3
t. n i
01
4J O
ID C


4)
3
.M
4->
rJ VI
•r4 Ifl

o
ro
1
IA
CL.



^v^
41
o 3
CNJI44
X
0
dP C
in —


^,
4)
IN g
o 3
om-4
23
0
dP C
in —


^~.
4)

3
114

>i O

Q —



0)
4-1
r-4.
• r4
t.
"w
§
*M
23
0 4J

X O1
Z -H
dp in
in —
r-l
a
o
a oi
o g
in 3
•H 4 U-4
01
4J O

S —

01
3
UJ
4-1
kj in
•rl 10
Q tP

in
1
en
a.



f*
4)
INg
o 3
(\4lt4
X
0
dP C
in —


^-^
4)
(N g
o 3
OjU-4
X
0
dP C
in —


^-.
41

3
'I 1

>i O
M C
Q —

^^.
X 41
o g
•* 9
M Ul
z
o
dP C
in —


^^
01

3
M t-l
41
4-1 O
10 C



41
g
3
M IM
0)
4J O
lO C




1



tN
o
CO
en
•o e
t7* (i
a
•a o

Ul 
<
\D
1
u>
(X.



J,^^
4)
o 3
f^M
X
0
dr C
in —


^^


_ .
(SJtl'J
X
0
dP C
in —


,v*
Q)
g
13

a
VD

CO
a.







1







r>4
o —
CM Q>
X g

dp IH
in —




04
o ~.
(M4I
X £

dP M-l
in ~^



01
g
^ 3
t4 
-------
r-
c
o
•H
4J
•H
U)
O
(X


to
C
o
•M
•P
•H
10
0



c
o
•H
4J
•H
01
0
(X


£ „

OP-O
in —
|
3 3
01 IM
o -a
•H
OP U
o id
CN —







^
.(H
•^
rf
tr>
i
in
IX







i










i



1
IM

CN-P
o x:
CNOl
X-H
H
OP 01
in —
~


IM

CN+J
ox:
ac*-H
rH
OP ca
in —
-J
E
IM
X
0 4>
^fXI
X Cr
Z -H
rH
OP 01
in — '
«
§
IM

4J
r^
M CTi
fl)-H
*J rH
id en
s —
"of
3 3
0) IM
OTJ
•H
OP u
o id
IN —







t|
•H
•*
a
01
i
CO
(X







1



«
3
IM

CNP
o x:
CNtP
X -H
rH
OP ca
m —
1

i
in
(X







i










i



1
*M

csM
ox:
CNOl
X-H
rH
OP 01
in —
•J
g
9
m

rsyJ
ox:
X^-H
rH
OP 01
m —
17
I
IM
X
O V
^"X!
X 0>
Z-H
rH
OP 01
in —





M
ID
4J
rH
•H
U.
g
3 3
0) IM
o-o
•M
OP O
o id
CN —







^4
•H
•*
a
Ot
1
CA
IX







1










1




, —



CNIM
X
o
OP C
m —



*-*
0)
CNE
CNM
X
O
OP C
in —





M
0)
P
rH
•H
fci
, 	
p
3
X IM
o
^* 0)
rr> (Q
5C C
01
""3,

iS
3 3
OHM
O T)
•H
OP O
o id
CN —







r^
•H
*
W
Ot
1
CO
0.







1



„
g
3
IM
CN
0 0)
CNO]
X C
01
OP-O
in —
I
3
IM
CN
O 01
CNOI
X C
CD
OP-O
in —

"w

3
X IM
o
K en
Z C
0)
-O
in —
_
"o?
3
IM

01
M ea
a) c
•P 0)
id rd
SE-
	 ,
p
•*r3
o IM
to
CN 01
X ca
C
OP ID
r** TJ
0> —
gS
9 3
0) IM
OtJ
•H
OP U
o id
CN —







M
•H
«*
O
rH
1
CO
IX







1



9
g
^4* IM
o
W -P
CNX;
x en
•H
OP rH
r- en
01 —
1
3
IM

X ca
Z C
0)
OP -a
in —





__^
M CD
01 E
•P 3
id MH
» —
|
p 3
01 
-H
OP rH
r- ca


i
g
X MH
O
••rtu
X ea

OP -d
in —





• ^^
n oi
0) 6
1 1 3
id MH
» —
8
3 3
ai MH
O "O
-H
OP U
o id
fN —







rj
•H
**
O
rH
1
in
(X
46
-------
while clean flue gas was used,  when silver nitrate was added to
the water, a white precipitate formed, which indicated the
presence of chloride or sulfate ions or both.

    Fume did not form when S02 span gas was pulled through a
train of deionized water, ammonium hydroxide, and hydrogen
peroxide.  When the deionized water was replaced with 10%
hydrochloric acid solution, a fume formed with either span gas or
air.  An absolute filter inserted into the train downstream from
the ammonium hydroxide impinger removed the fume.

     When 20% oleum was substituted for the deionized water and
air was pulled through the train, a dense fume formed.  A water
wash between the oleum impinger and the remainder of the train
decreased the severity of the fume.  An absolute filter also
decreased the severity of the fume but did not completely clear
the gas stream.

    The bench-scale studies showed that the flue gas contained
one or more materials that may cause the formation of fume.  The
material could be chloride or S03.  Other material such as flyash
 (which could serve as sites on which the plume particles grow) or
organic materials may also cause formation of fume.  The tests
also showed that the material required for fume formation can be
removed ahead of the absorber with a filter that removes
submicron particles.  Further bench-scale work is needed to
identify the fume agent.
REGENERATION TEST PROGRAM AND RESULTS

Acidulation and^Stripping

    As previously stated, the acid source in the ABS process is
NH4HS04 obtained by decomposing  (NH4)2S04.  Because the thermal
decomposer was not constructed, H2S04 was used in the pilot-plant
study of the acidulation step.  Chemically, this substitution is
valid because in solution, a mixture of H2S04 and  (NH4)2S04 would
differ from an NH4HS04 solution only by the S04:NH3 ratio.  This
difference was not expected to alter the test results at the
concentrations used in the study.

    The purpose of the acidulation and stripping tests was to
develop techniques that would result in removing essentially all
of the absorbed S02 from the absorber product liquor.  A limit of
0.5 g/1 of S02 remaining in the stripped liquor  (NH4)S04 solution
was arbitrarily set for the test program.  This low limit was set
because any S02 remaining in the solution to the evaporator-
crystallizer would be stripped from the solution there and cause
pollution problems in the condensate and gases leaving this unit.
By the same token, under-acidulation would increase the sulfites
to the evaporator-crystallizer where they could either be
                               47
-------
decomposed to S02 and stripped from the solution or else be
disproportionsted to sulfates that would have to be removed from
the system.  A series of tests was made in the "short"  (1-ft-
diameter by 6-ft high)  acidulator and stripper.  The acid ion to
ammonia ion mol ratio was varied from 1.04 to 1.85 to determine
its effect on S02 removal.  The acid and product liquor were
combined in the cone mixer in the top of the acidulator.  The
retention time in the acidulator was approximately 10 min.  The
liquor overflowed into the stripper which contained a U-ft packed
bed.  The stripping gas  (air) flow rate was varied from 5 to 15
ft3/min.  The data from these tests are shown in Appendix H,
Table H-5.  Figure 20 shows the overall S02 removal efficiency
verus acid to ammonia ion mol ratio in the acidulation and
stripping equipment.  At a ratio near 1.0, the overall S02
recovery efficiency was about 50%.  When the ratio was increased
to approximately 1.8, the efficiency was about 96%; at this
ratio, the solution contained 5.1 g S02 per liter.  Extrapolation
indicated that a ratio of 2.0 would te needed to reach a removal
efficiency of 100%.  Over the range tested, the stripping air
flow rate had little effect on removal of S02 from the acidulated
solution.

    Extrapolation of the data obtained in tests with this
equipment indicated that an excessively high acid ion to ammonia
ion ratio  (about 2.0) would be required to reach the 0.5 g/1
limit set for the test program  (99.5% of the absorbed S02 must be
removed from the solution to meet this goal).  The excess acid
would place an added load on equipment and present severe
corrosion problems, particularly in the evaporator-crystallizer.
If the acid were neutralized ahead of the evaporator-
crystallizer, the added  (NH4)2S04 would have to be removed and
decomposed to the acid NH»HS04, which would place added energy
requirements on the system.

      Much better results were obtained with a  redesigned
 acidulator-stripper  unit.  The  original acidulator was  replaced
 with  a  simple mixing vessel.  The  new stripper was 4  in.  in
 diameter  and  contained  30 ft of dumped Tellerette packing.   The
 initial tests with  the  unit were made in  a batchwise  mode.
 Thirty  gallons of  absorber product  liquor was  acidulated  to  a
 final acid ion to  ammonia ion ratio of 1.05 over  a period of
 15  min.   The  acidulated material was held in the  reaction vessel
 for an  additional  15 min before it was fed to  the stripper.  The
 acidulate was heated to 140°F,  the  calculated  equilibrium
 temperature, when  acidulating with NH^HSC^ , and metered to  the
 stripper  at a  flow  rate of  0.5  gpm to approximate the rate  that
 liquor  is produced  in the absorption section.  The 0.5-gpm  rate
 corresponds to a packing irrigation rate  of 5.7 gal/(min)(ft2)  of
 packing cross-sectional area.   Air was fed to  the bottom  of  the
 stripper  at flow rates  of 5, 10,  and 15 ft3/min,  corresponding  to
                                48
-------
UJ
o:

 
-------
stripping gas flow rates of 10, 20, and 30 ft3/(min) (gal)  of
acidulated material.

    The results of these tests, AS-1, AS-2, and AS-3,  are shown
in Appendix F, Table F-6.  The acid to ammonia ion ratio was
approximately 1.0 in the first two tests and 1.15 in the third.
The stripping gas flow rate was 5, 10, and 15 ft3 for tests AS-1,
AS-2, and AS-3 respectively.  The stripper effluent from these
three tests contained free S02 in excess of the 0.5 g/1; in test
AS-3, the S02 was decreased to 0.68 g/1.  The acidulated and
stripped liquor was stripped again in test AS-3A using an air
feed rate of 15 ft^.  The S02 was decreased to 0.12 g/1 on second
passage through the stripper, which indicated the need for
additional packing height or a higher stripping gas flow rate.

    In these batchwise tests, the H2S04 was added slowly (about
0.3 gpm) to a large volume of absorber product liquor (30 gal).
Near the beginning of the acidulation process, the acid ion to
ammonia ion ratio was near zero because of the differences in
volumes.  The reaction at the point of acid-solution contact was
violent and gases flashed off rapidly.  As the amount of acid
increased, the ratio approached the desired value of 1.05.   The
violent release of gases decreased with the increase in acid
level in the solution and little evidence of gas release was
noticeable after about 1/2 to 3/U of the acid had been added.
Since the acidulation was not mechanically agitated, thorough
acidulation may not have occurred, even at acid ion to ammonia
ion ratios greater than 1.0.

    Continuous acidulation was attempted in test AS-U, in which
absorber product liquor and sulfuric acid were fed continuously
and directly to the stripper.  The feed rates approximated full-
scale continuous pilot-plant operation.  The direct acidulation
in the stripper was tried because  (1) it permitted continuous
operation,  (2) it eliminated a piece of process equipment, and
(3) gases flashing from the solution would be swept away instead
of being reabsorbed in the solution.  A violent reaction at the
point of addition resulted in the liberation of much heat.  In
one instance, the temperature rose to 195°F, which is above the
design temperature for the plastic stripper.  Heavy foaming at
the point of addition resulted in surges of flow through the
stripper.  Sulfur dioxide retained in the stripper effluent was
34.3 a/1.

    The procedure for continuous acidulation and stripping was
tested further in the final system, Figure 11.  The acid and
absorber product liquor were fed simultaneously to the bottom of
a mixing pot, which overflowed to the top of the stripper.  The
pot had an effective volume of 1.5 gal, which gave a residence
time of about 3 min.
                              50
-------
    Introduction of acid and liquor into a heel of partially
acidulated material reduced the violence of the reaction although
some degasing was occurring as the material entered the top of
the stripper.  Tests were run using an airflow rate of 5 ft3/min
(10 ft3/gal of liquor) and the full 30 ft of stripper packing.
Data from these tests, AS-6, AS-8, and AS-9, appear in Appendix
F, Table P-7.  The S02 retained in the stripped liquor was below
0.5 g/1 for each of these tests (O.U4, 0.10, and O.UO g/1
respectively).  The acid ion to ammonia ion ratio was 1.05 for
test AS-6, 1.13 for test AS-8r and 1.15 for test AS-9.

    Since the limit of 0.5 g/1 S02 in the stripped liquor could
be met in the 30-ft stripper at the 5 ft3 airflow rate, tests
were made to determine whether the tower could be shortened and
still meet the S02 limit with stripping gas rates of up to 15 ft3
(30 fta/gal of liquor) .

    Tests AS-10, AS-11, and AS-12 were made with 5, 10, and 15
ft3 airflow rates, respectively, with an effective tower packing
height of 20.3 ft.  Data from these tests appear in Appendix F,
Table F-7.  in each of these tests, the retained S02 exceeded the
0.50 g/1 limit although in test AS-12 with the 15 ft3 stripping
gas flow rate, the retained S02 was 0.5H g/1.  The effect of
stripping gas flow rate on retained S02 is shown in Figure 21.

    Since the stripping operation did not reach the desired limit
with the 20-ft packing, no tests were made using 10 ft of
packing.

    The results of acidulation and stripping tests showed that a
mixing-pot acidulator coupled with a 30-ft stripper is adequate
to reach 0.5 g/1 retained S02 in the stripped liquor with as low
as 5 ft3/min stripping gas flow rate.  Operation with a 20-ft
tower and a 15 ft3/min stripping gas flow rate would be marginal.

    The lower stripping gas flow rate results in a higher S02
concentration in the off-gas  (56% for the 5 ft3 rate and 27% for
the 15 ft3 rate) though either gas would be acceptable for H2S04
manufacture.  The nominal S02 concentration in the sulfur burner
off-gas feed stream to a contact H2S04 plant is 8.5%.

    Table 5 shows typical data from tests meeting the 0.5 g/1
limitation for S02 retained in the stripped liquor.

Ammonium Sulfate Crystal Separation

    The ammonium sulfate solution from the acidulation and
stripping step was concentrated in an evaporator-crystallizer to
produce crystals of  (NH4)2S04.  The evaporator-crystallizer,
manufactured by Goslin-Birmingham, was designed to remove 200
Ib/hr of water from the solution.  The performance of the single-
                              51
-------
    3.0
z
UJ
UJ

tt
UJ
Q.
a

a:

CO
 ca
    2.0
UJ
UJ
a:
1.0
      0
        0
                                   RACKING HEIGHT =20.3 FT

                                   STRIPPER OIA.  = 4 IN.
                      I
I
                     5                   10                  15

                 STRIPPING GAS RATE, FT3OF AIR/MIN
  Figure 21.  Effect of  stripper gas rate on SOg in stripper effluent.
                                52
-------
     TABLE 5.  TYPICAL REGENERATION TEST DATA
Flowrates
  Product liquor, gpm                     0.45
  Sulfuric acid, gpm                      0.076
  Stripping gas  (air), ft3/min            5.0
Temperature,  °F
  Product liquor                          123
  Sulfuric acid                           61
  Stripping gas                           61
  Stripper effluent                       89
  Stripper exit gas                       104
Liquor analyses
  Product liquor to acidulator-stripper
    Sulfite sulfur, g/1                   30.76
    Bisulfite sulfur, g/1                 105.23
    Sulfate sulfur, g/1                   32.59
    Total sulfur, g/1                     168.58
    Specific gravity, g/ml                1.242
    pH                                    5.7
  Sulfuric acid, % H2SO4                  91.4
  Stripper effluent
    Sulfite sulfur, g/1                   0.0
    Bisulfite sulfur, g/1                 0.0
    Sulfate sulfur, g/1                   100.46
    Bisulfate sulfur, g/1                 22.42
    Free S02, g/1                         0.12
    Total sulfur, g/1                     122.95
    Specific gravity, g/ml                1.230
    pH                                    1.7
Acidulation stoichiometry                 1.164
Acidulation efficiency, %                 100
Percent of released S02 that
 is stripped                              99.9
Stripper packing height                   30

iuAcidulation stoichiometry refers to the mol
    ratio of the acid ions from sulfuric acid to
    the ammonium ions  (from ammonium sulfite and
    bisulfite) in the absorber product liquor.
                       53
-------
effect unit was acceptable while operating at a temperature of
170°F and an absolute pressure of about 8 in. Hg (22 in. Hg
vacuum).   The 170°F limitation was set to minimize corrosion in
the type 316L SS unit.

    The concentrated  (NH4)2S04 slurry from the evaporator-
crystallizer was fed to a continuous belt filter.  The filter
(Eimco Model 112) had a 12-in.-wide belt with a total of 10 ft*
filtering area.  The unit proved to be greatly oversized for
continuous operation, even at the lowest belt speed and with the
(NH4)2S04 slurry entering the filter at the point that used the
least amount of filtering area.  Sufficient material could not be
maintained on the belt to prevent vacuum breaks.  In a batchwise
operation, the filter system removed crystals sufficient to
balance the pilot-plant production rate of 200 Ib/hr.

    Approximately 1,000 Ib of crystals were removed on the belt
filter.  The crystals were sized about 70% plus 35 mesh and
contained 5-10% moisture.  The crystals were dried in a gas-fired
rotary dryer to 2% or less moisture and bagged in standard
fertilizer bags.  The bags were left open and stored 9 mo in an
open-air shed.  The material was free-flowing at the end of the
storage period.

    Centrifuges are used in most commercial  (NH4)2S04 production
facilities.  The belt filter was replaced with a 6-in. screen-
bowl centrifuge manufactured by Bird Machinery Company.  Slurry
from the evaporator-crystallizer was pumped continuously to the
centrifuge.  A crystal separation rate of 200 Ib/hr was achieved
when the ammonium sulfate solids in the feed to the centrifuge
was 10%.  The moisture content of the cake was 3%.  Line pluggage
occurred when the solids content was 15%.  When the solids
content was decreased to 5%, the cake moisture content increased
and eventually became "mud." Variation of the centrifugal force
(760, 1,040, and 1,350 Ib force/lb mass) had little or no effect
on the cake moisture.

Ammonium Sulfate Decomposer Design

    EPA had the responsibility for developing the design of the
decomposer to be used in the ammonium bisulfate study.  Some work
had been done by others on the decomposition step although none
used solutions generated on power plant stack gases.  In the
1920's a fertilizer process  (22) was developed that required
decomposition of  (NH4)2S04 to drive off ammonia and produce
ammonium bisulfate, which was then used as an acidulate to
release S02.  More recently, an engineering company  (23) used the
bisulfate process in various fertilizer flowsheets.  However, the
decomposition step has not been demonstrated in a continuous
operation in any of these facilities.  An objective of the Phase
III work was to operate the complete absorption-regeneration
                              54
-------
system including an (NH4)2S04 decomposer.  Dr. Richard M. Felder,
Associate Professor of Chemical Engineering at North Carolina
State University, was contracted (EPA purchase order No. 4-02-
04510) to survey the literature and develop the necessary
reaction kinetics and rate equations.  These equations and data
were used to design the reactor vessel with appropriate
temperatures and reaction times.  Felder's work covering the
reaction kinetics and the engineering design basis for the
reactor is included in this report as Appendix G.  He recommended
that the reactor be designed to operate at 700°F with a melt
retention time  (melt mass divided by feed rate) of 7.1 hr when
using a steam to  (NH4)2S04 mass ratio of 0.2.  (The steam is
necessary to sweep the released NH3 from the melt and to prevent
decomposition of NI^HSO* to ammonium pyrosulfate and water.)

    Ajax Electric Company, in Philadelphia, a manufacturer of
molten salt bath furnaces, was contracted to furnish a workable
design complete with detailed drawings that met the design
criteria as specified by Felder.  The Ajax contract was handled
by Research Triangle Institute, Research Triangle Park, North
Carolina  (RTI contract No. 1006), as a part of Research Triangle
Institute's service contract with EPA.

    The decomposer design specified an inside dimension of 3 ft
diameter by 6 ft high and had melt outlets at the 15- and 24-in.
levels corresponding to retention times of 4 and 7 hr at the base
feed rate of 200 Ib/hr of ammonium sulfate.

    The decomposer wall had a 9-in. thickness of acid-resistant
brick plus 5 in. of insulating material.  The outer shell was
aluminized steel.

    Heat input to the system was by current flowing through the
electrically resistant melt.  The power source was single-phase,
60-hertz, 460-V to the primary side of an 80-kW-rated
transformer.  The secondary voltage was infinitely variable from
30 to 50 V to maintain the desired temperature of the melt up to
a maximum temperature of 750°F.  Two 6-in. carbon electrodes
carried current to and from the melt.  Both the spacing and the
immersion depth of the electrodes could be varied.  The test
program was to determine the effect of electrode spacing,
electrode immersion depth, voltage and steam flow on one or more
of the following response variables:  power input, melt
temperature, ammonium sulfate decomposition rate, electrode
consumption, rate of formation of pyrosulfate, and ammonia
concentration in the melt and vapor space.

    The decomposer development program was stopped short of the
construction phase because of unfavorable economics for a full-
scale stack gas desulfurization process employing an electrical
thermal decomposer.
                              55
-------
ECONOMIC EVALUATION

    Cost estimates were made to permit comparison of the
investment and net equivalent unit revenue requirements
(mills/kWh) for the ammonium absorption - ammonium bisulfate
regeneration (ABS) process with several advanced FGD processes.
Each process was designed to desulfurize the flue gas from a 500-
MW, new, coal-fired power plant unit burning coal with 3.5% S
(dry basis).  The study was based on 1975 costs and available
technology.  The results of the estimates are summarized in Table
6 and given in detail, including flowsheets, in Appendix H.
Brief descriptions of the processes follow:

         Process 1 is the ABS process as originally envisioned,
         in which the S02 is recovered from the ammoniacal
         absorber product liquor by acidulating with NH4HS04 and
         stripping with air; it then is used in the production of
         sulfuric acid.  Ammonia is evolved in the  (NH4)2S04
         decomposition step, recovered and recycled to the
         absorption step.  It was assumed for this study that the
         absorber could be operated without a plume by control of
         operating conditions.  This assumption may be proven
         with limited further experimentation.  Should it not be
         verified, elimination of plume may be obtained through
         use of a wet electrostatic precipitator as an absorber
         and plume collector at some added cost, as described
         later.

         Process 2 is a noncyclic adaptation of the ABS process
         in which the S02 in the absorber product liquor is
         recovered as  (NH4)2S04.  Since there is no regeneration
         section, NH3 leaves the system as a part of the
          (NH4)2S04.  Again it was assumed that the plant could be
         operated without a plume by control of operating
         conditions.  Should this not be acceptable, an absorber-
         wet electrostatic precipitator could be used at added
         cost.

         Process 3 is the basic limestone slurry absorption
         process with simple sludge throwaway.  No salable or
         useful byproducts are redeemed in the process.  It is
         recognized that fixation of sludge may be necessary to
         meet disposal requirements.  Fixation would improve its
         compaction characteristics and result in longer
         disposal-pond life and better landfill capability.

         Process U is a regeneration process that involves
         scrubbing with a slurry of magnesia to absorb the S02.
         The product stream of the absorbing slurry is dewatered,
         dried, and calcined.  Magnesium oxide is regenerated and
                               56
-------
•PXI
-p q x:
•H CU CU tS
C 3 g.*
3 C 0) \
0) VH to
(3 .p > -H rH
cn CU CU 3 rH
U 25 ^-t tJ^-H
cn cu g
M
u
o
PS rH 4J
pj (0 CU C
3 3 CU
2 q c e
o c cu cu
M (0 > M -
EH CU -H a
< -p n 3 a
N cu tr
M SO
PS rH
D
Cn

£D i — I •.
cn to -u
W 3 C
Q q o cu
q 3 g
cn  -H a
to cu 3 a
W 0 rH CJ1
D k O

Cn

cn
D
O rH
H (0 -
PS -P -P
i< -H C
> ft O
(0 g 
PS o -p
O CO ~
Cn rH 0 §
(0 > a
cn -PC
E-* O -H
cn E-I
o
u

p4
o

^H
PS
frf
*r
s!
£j
cn
s
VO

W
«
^
H W
to
O
O
0
^_l
&4




CN
^3*
•
n










CN
•
CN
rH











O
•
*•£>
rH













rH
•
CN
*3*









U
-H
^1
1 3
4-1
C CU rH
O -P 3
•H (0 cn u
-P 4-1 q
ftrH 1 O
rH 3 -H
o to c -p
CO -H O O
Xl CO. -H 3
< 4-> T3
g 10 0
(0 3 k M
•H -H CU P^
q c q
O O CU T3
C g (31 .,-)
§ g o u
<3 
•
o
rH











V^)
•
O
rH













r-
•
o
ro



1 . 	 .
^i
C cd
0 5
•rH (0
-P S
ft O
rH rH
0 X!
co E-i
Q • — '
^
0)
rH ti
rH 3
3rH
rH cn
cn
M-l
O O
c
O C7i
-P C
CO-H
OT)
g C
•H 0
JPJ
ro



00 CTi
*3* rH
• •
CN CO








•
1-1 T3
in o • O CU
x: -P -a
OO rH |S (0 3
rH X. S --H
x^ O
3 -a x
-P O CU U
ffl C -H
•H x: n
o (0 to 3
O in (04-1
0 -P >1 rH
- C rH 3
CT> O 4-1 to •
O
CT> in ^ * 4-1 *> -P
• • • , — 4-J • O O (0
rH T rH CO (0 CO O4-I
rH rH (0-HCU 4JrHrHrH
> to x: c (0 3
O (0 O O CO C tfl
g X! M gO O
O CU ft -P g
1-1 >i g VH tfl \ 3
M (0 -H -H 0 -H
CM rQ* O 3 r^ CO T3
o — ' -P CT  O
CO CO 0) M W
CO • rH O
#1 4J M 4-1 . 4J C
OdPU >iO MUO
o\ino \3to>i3-P
co i — i • M VH q -P rO *\
« • ^ CO -H (~j O C O O ^*
CM r— -pro >OCO!-ICN
CO CO -H*-COCUg ft
qi-l-H OrHO^>1
3 >i o l-i -P XI -
\ >, r^ to -rH c cu
rHtOXl r^3(04-H-P
1 O C - CTi CTH O 10
1 SOCn OrHOft4-i
C 0 -PO • g H O rH
OU CU ft in^-rH- MrH3
•HC OC or-tfl-PCOO(Otfl
-PO -HO TiOrH-H O3Stfl
ft-H -P-H om tng-rHCo g
rH -P ft-P 1-1 -O(OS-PCUftg3 •
OU 1-4U -HCN-PX1O(0> O-HU3
£03 O3 4-HrH rHOO4-)rHCtO
XiTi COT! in4J-P-POMO4-lOCU
ipj P^ O x;~-'C-PC4-l3(0>-l
rH O U^CU OC(0-HC ft
MT) -PTi TSrH-P (0 rHOC
3-H -H-rH SO S-PrH-P >OC
rHU 4-lU CUCCO -HftCtTiO-PO
cn< rHcc qrH(0>iC ocu \-H
3 3CJiX13-Pg-rH 'S'-P
 M
tfll-l grH ••IrHUinCUSOlOrc) O
CU3 33 cOO(0(0-S'O>g3-q
q4-l -H4-4 -HOO-POO-rHCOrHrdO
CTlrH rrJrH COinCJW P-iaHPSU-HtJl
(03OP(0 CUCU
acncnwpq H.OPS
•* m id xi u
57
-------
         recycled to the absorption step, and S02 is evolved and
         is used in the production of sulfuric acid.

         Process 5, also a regeneration process, uses a solution
         of sodium salts and makeup sodium carbonate (Na2C03)  to
         absorb the S02.  Sodium sulfate (Na2S04), resulting from
         oxidation in the absorber, is purged from the absorber
         product solution, crystallized and sold.  The remaining
         sodium bisulfite (NaHS03) solution is evaporated and the
         resultant crystals thermally decomposed to give sodium
         sulfite (Na2S03) and gaseous S02.   The Na2S03 is
         recycled to the absorption step and the S02 is used in
         the production of sulfuric acid.

    As shown in the cost tabulation, the ABS process has the
highest capital costs, $42.1 million, and the highest net unit
revenue requirement, 3.42 mills/kWh.  (Use of a wet electrostatic
precipitator would increase these costs by 14% and 4%
respectively.)

    Process 2,  ammonia absorption of S02 with (NH4)2S04
production has a capital requirement of $31.5 million and a net
unit revenue requirement of 2.87 mills/kWh.   (Use of an absorber-
precipitator system would increase these costs by about 20% and
5% respectively.)

    The limestone slurry process with ponding (throwaway) of
sludge  (Process 3)  has an estimated capital requirement of $30.7
million and a unit revenue requirement of 2.97 mills/kWh.  If
sludge fixation is necessary, the unit revenue requirement will
increase by about 17%.

    The magnesia slurry process  (No. 4), which produces sulfuric
acid, has the lowest unit revenue requirement, 2.48 mills/kWh,
and one of the lowest capital requirements $32.3 million.

    The sodium sulfite process with production of sulfuric acid
was indicated to be more costly than the magnesia process;
capital requirement is $37.1 million and net unit revenue
requirement is 3.19 mills/kWh.

    A study, sponsored by EPA, investigated the  economics of
using  (NH4)2S04 from an ammonia scrubbing FGD process as a
replacement for anhydrous NH3 for direct application of nitrogen
to the soil.  It was assumed that the ammonia planned for direct
application would be routed to the power plant,  used for
absorbing S02 from the flue gas, and recovered as  (NH4)2S04,
which then would be transported and applied to the soil as a
replacement for anhydrous NH3.
                              58
-------
    Two midwestern coal-fired power plants were chosen for the
study.   The first, in Kincaid, Illinois, is in an area of high-
density agricultural NH3 consumption.  The power plant at Kincaid
is rated at 1,300 MW and has the potential to produce U02,396
tons of (NH4)2S04 annually at a rate of 1,219 tons/day.  The
second plant is at Paradise, Kentucky, a region of low-density
agricultural NH3 consumption.  The plant is rated at 2,U50 MW and
has the potential to produce 961,251 tons of (NH4) 2S04 annually
at a rate of 2,913 tons/day.

    The results of this study indicated that the sum of the costs
of handling, transporting, storing, and applying a ton of NH3 to
the soil as  (NH4)2S04 may be about $28 less than that for NH3 as
anhydrous NH3 in the high-use area (Kincaid, Illinois)  and about
$8 less in the low-use area  (Paradise, Kentucky).  The cost of
FGD is not included.
                              59
-------
                CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS

     Phase I of the pilot-plant study demonstrated that ammonia
absorption as applied to SO2 removal from coal-fired power plants
was effective (90% or higher) over a wide range of operating
conditions.  Specific conclusions from the Phase I work were:

     1.  The absorption efficiency can be reliably predicted for
         given operating parameters.

     2.  Ammonium sulfate levels in the absorber loop have only
         slight influence on S02 absorption.

     3.  Flyash has a negligible effect on SO2 removal.

     4.  Temperature of the inlet flue gas has little effect on
         S02 removal in the range covered by the study  (180-300°F).

     5.  Corrosion was not a problem in the absorption loop when
         using SS and certain nonmetals.

     The favorable results from the Phase I study led to the
recommended decision to expand the work to include a study of
a regeneration scheme to make the process cyclic.  This work,
called Phase II, began the investigation of the ammonium bisulfate
process for recovery of ammonia to be used in regenerating the
ammoniacal liquor from the absorber section.  The conclusions
drawn from the Phase II study were:

     1.  The plume identified and only casually examined in the
         Phase I work is persistent, and precise control of the
         absorber operation is required in order to meet an
         opacity limit of 5% in the pilot plant  (about 20% in
         commercial stack).  Equipment modifications were made
         in the Phase III work in an effort to obtain the
         necessary precise control of the absorber operation.

     2.  Formation of a plume is less likely to occur at low
         liquor concentrations  (CA = 1) than at higher
         concentrations  (CA = 12).
                                60
-------
     3.  Reheat is required to dissipate the steam plume from the
         absorber.  Under certain ambient conditions (temperature
         and humidity) dissipation of the steam plume by reheat
         is either impossible or impractical.

     4.  Regeneration of the absorber liquor by acidification
         and stripping removes and recovers up to 99% of the
         absorbed SO2 .

     5.  Corrosion in the regeneration loop requires use of low-
         carbon SS and plastics.

     Though difficulties were identified (for instance, separation
of flyash and ammonium sulfate crystals from liquors in the system)
they did not appear to be technically insurmountable.  A recom-
mendation was then made to extend the work into Phase III to study
the complete, closed-loop regeneration system.

     The plume problem was not overcome in the third phase of the
pilot-plant study.  Also, the economics of the process were
unfavorable.  Conclusions drawn from the work are:

     1.  Absorption was adequate (90% or higher) using scrubbing
         liquors of moderate-to-high salt concentration (C^ =
         10-15) .

     2.  While operating at these liquor concentrations, effective
         and consistent plume control (pilot-plant stack opacity
         5% or less) was not achieved by methods and equipment
         tested in the pilot plant.

     3.  An in-line indirect steam-heated reheater dissipated
         the water vapor in the scrubbed flue gas but did not
         significantly reduce the opacity of the ammonia- sulfur
         compound plume .

     4.  Bench-scale  studies identified chloride and SOa (both
         found in the inlet flue gas) as fuming agents.

     5.  Predictions  of the formation of the ammonia-sulfur
         fume, presumably ammonium sulfite monohydrate, can
         be made from a thermodynamic equation derived from
         the equilibrium constant for the reaction:
              2NH3 + 2H20 + S02  ->   (NHO 2S03 'I^O

         This study indicated that the fume prediction equations
         must be satisfied as a necessary, but not limiting,
         condition for fumeless operation; for instance, chlorides
         and SO 3 are not considered  in the equation.
                                 61
-------
     6.   Essentially complete acidulation of the absorber
         product liquor was accomplished using sulfuric acid.

     7.   The stripper removed 99.9% of the free S02  from the
         acidulated absorber product liquor.  The liquor effluent
         from the stripper ammonium sulfate solution was
         essentially free of S02.

     8.   The combined off-gas stream from the acidulator and
         stripper contained approximately 60% S02/ which is
         more than adequate for a  feed gas stream to an H2SO ^
         plant.

     9.   The evaporator-crystallizer produces an ammonium
         sulfate slurry suitable for separation of crystalline
         ammonium sulfate.

    10.   Standard ammonium sulfate separation techniques appear
         to be acceptable for removing crystalline ammonium
         sulfate from the evaporator-crystallizer slurry.

    11.   A comparative economics study of the NH3-ABS process
         and other more advanced regenerable and nonregenerable
         processes (500-MW units)  showed that the NH3-ABS process
         had the highest fixed investment cost and next to the
         highest annual revenue requirement (operating costs).


RECOMMENDATIONS

     1.   Developmental work on the NHa-ABS process should cease.
         The prime drawbacks to the process are the plume problem
         and the unfavorable economics.  Attempts to control the
         fume have met with little success.  The ABS process is
         a high energy-consuming process.  The electrical thermal
         decomposer in the regeneration section requires nearly
         50% of the power requirement for the entire FGD system
         (see Table H-lA).  The cost of power, which has more
         than doubled since the study began, is the major factor
         in forcing the ammonium bisulfate process into an
         unfavorable economic situation.  Calculations based
         on information obtained from vendors indicate that use
         of wet electrostatic precipitators to collect the fume
         would add about  15% to the capital cost of the ABS
         system, further  adding to its already untenable economic
         position.

     2.   The emphasis of  any further ammonia absorption pilot-
         plant work should be directed toward a nonregenerable
         process to eliminate the costly decomposition step.
                                62
-------
The ammonia absorption system can produce  (NH 4)2 SO 4, a
nitrogen source in the formulation of some fertilizers.
Indications are that  (NH 1^2 SO it produced during 862
absorption can be sold as a fertilizer at a price high
enough to recover the cost of ammonia.  A comparative
economics study showed that a system to produce
(NH ij 2SO i» during S02 removal has a much lower unit
revenue requirement than does the ABS process.  The
study also showed that the ammonia absorption -
(NH ij) 2SO it process is at least as attractive economi-
cally as the limestone slurry process with simple
sludge throwaway (2.87 and 2.97 mills/kWh respectively)
Even with a mechanical or electrical particulate col-
lector added to the system, it remains competitive.  A
marketing study showed that the cost of handling and
applying ammonia to the soil as ammonium sulfate is
competitive with applying the ammonia as anhydrous
ammonia.
                       63
-------
                           REFEPENCES
1.   Tennessee Valley Authority.  "Sulfur Oxide Removal from Power
    Plant Stack Gas:  Sorption by Limestone or Lime - Dry
    Process." TVA report S-439 (prepared for EPA).  NTIS PB 178
    972, 1968.

2.   Tennessee Valley Authority.  "Sulfur Oxide Removal from Power
    Plant Stack Gas:  Use of Limestone in Wet-Scrubbing Process."
    TVA report S-440 (prepared for EPA).  NTIS PB 183 908f 1969.

3.   Tennessee Valley Authority.  "Sulfur Oxide Removal from Power
    Plant Stack Gas - Ammonia Scrubbing:  Production of Ammonium
    Sulfate and Use as an Intermediate in Phosphate Fertilizer
    Manufacture." TVA Bulletin Y-13 (prepared for EPA).  NTIS PB
    196 804, October 1970.

4.   Ramsey.  "Use of the NH3-S02-H20 System as a Cyclic Recovery
    Method." Brit. Pat. 1,427, 1883.

5.   Lepsoe, R., and w.  S. Kirkpatrick.  "S02 Recovery at Trail, A
    General Picture of the Development and Installation of the
    Sulfur Dioxide Plant of the Consolidated Mining and Smelting
    Company of Canada,  Limited, at Trail, B.C." Trans._ Can. Inst.
    Mining Met... 4.0  (in Can. Mining Met. Bull. No. 304) , 399-404,
    1937.

6.   Johnstone,  H. F.  "Recovery of S02 from Waste Gas:
    Equilibrium Partial Vapor Pressures Over Solutions of the
    Ammonia - Sulfur Dioxide - Water System." Ind. Eng. Chem^ 27.
    (5), 587-93, May 1935.

7.   Johnstone,  H. F., and D. B. Keyes.  "Recovery of S02 from
    Waste Gases:  Distillation of a Three-Component System
    Ammonia - Sulfur Dioxide - Water." Ind. Eng. Chem. 27  (6),
    659-65, June 1935.

8.   Johnstone,  H. F., and A. D. Singh.  "Recovery of S02 from
    waste Gases: Design of Scrubbers for Large Quantities of
    Gases." Ind. Eng. Chem. 29 (3), 286-98, March 1937.

9.   Johnstone,  H. F.  "Recovery of S02 from Waste Gases:  Effect
    of Solvent Concentration on Capacity and Steam Requirements
    of Ammonium Sulfite - Bisulfite Solutions." ^ncL. Eng. Chem.
    29  (12) 1396-98, December  1937.
                              64
-------
10.  Johns-tone, H. F.  "Recovery of Sulfur Dioxide from Dilute
    Gases." Pulp Paper Mag. Can. jv3  (4) , 105-12, March 1952.

11. Tennessee Valley Authority.   "Removal of Sulfur  Dioxide from
    Power  Plant  Gases."  Annual  Report  of the Development Branch,
    FY 1955, pp. 20-21,  1955.

12. Nakagawa, S.   (Japan Engineering Company,  Tokyo,  Japan).
    Private communications,  1968.

13. Hamelin, R.   (Ugine  Kuhlmann,  Paris, France).  Private
    communications, May  1969.

14. Tennessee Valley Authority.   "Pilot-Plant Study  of  an Ammonia
    Absorption - Ammonium Bisulfate  Regeneration Process, Topical
    Report Phases I and  II."  EPA-650/2-74-049a (NTIS PB  237 171),
    June 1974.

15. Cominco Ltd.  "Cominco's  Fertilizer Operation."  Nitrogen 35,
    22-27, 29, May  1965.

16. Lepsoe, R.,  et  al.   "Process  for the Production  of  Ammonium
    Sulphate and Elemental Sulphur." U.S. Pat. 2,359,319, 1944.

17. Tennessee Valley Authority.   Internal Progress Report, July
    1954 to March 1955.

18. Hixon, A. W., and R.  Miller.   "Sulfur Dioxide from  Flue
    Gases." U.S. Pat. 2,405,747,  August 13,  1946.

19. Jordan, J. E.,  and G.  M.  Newcombe.  "Sulfur Dioxide  Removal
    from Stack Gases." U.S.  Pat.  3,927,178,  December 16, 1975.

20. Lazarev, Vladimir I.,  Deputy  Director,  NIIOGAZ  (Moscow,
    Russia).  Private communications,  March 29-31,  1976.

21. Specter, Marshall, and P. L.  Thibaut Brian.  "Removal of
    Sulfur Oxides from Stack  Gases." U.S. Pat. 3,843,789, October
    22, 1974.

22. Alabama Power Company.   "New  Process of Fertilizer
    Manufacture  Announced."  Mfr.  Rec.  92  (26), 53, December 29,
    1927.

23. Rubin, Allen G.   (Bohna  Engineering and Research, Inc., San
    Francisco, California).   Private communication,  1973.
                                65
-------
                           APPENDIX A

            ANALYTICAL AND GAS SAMPLING PROCEDURES


                           CONTENTS
lodimetric Method for Analysis of Total Sulfites ....    68
Alkimetric Method for Analysis of Total Bisulfite
 Sulfur and Total Sulfur 	    68
Analysis of the Acidulator and Stripper Liquors for
 Bisulfite, Bisulfate, and Total Sulfur  	    69
Silver Nitrate Method for Analysis of Chloride 	    70
Ammonia in Exit Flue Gas Sample (Direct Nesslerization
 Method)	    71
Preparation of Ammonia Reagents (for Nessler Method) . .    73
Preparation of Standard Ammonium Chloride and Ammonium
 Sulfate Solutions for Calibrating Spectrophotometers. .    73
Procedure for Sampling Inlet or Exit Flue Gas for
 Particulate and Sulfur Dioxide  	    74
Procedure for Sampling Exit Flue Gas for Ammonia ....    80

Figures
A-l  Gas Sampling Apparatus for S02 and Particulate  . .    75
A-2  Gas Sampling Apparatus for Ammonia	    81
                              67
-------
                           APPENDIX A

             ANALYTICAL AND GAS SAMPLING PROCEDURES



IODIMETRIC METHOD FOR ANALYSIS OF TOTAL SULFITES


Mani pu1ation s

    Aliquot 25 ml of the scrubber solution into a 1,000-ml
volumetric flask containing about 300 ml of condensate water.
Use caution to see that the discharge end of the pipette is under
the water.  Dilute to volume.  In a 250-ml wide-mouth erlenmeyer
flask add 10 ml of 1-10 HC1 and 35 ml of 0.1 N I2 (more may be
used if necessary) .   Aliquot 20 ml of the diluted sample under
the surface of the iodine.  Allow time to react and back titrate
the excess I2 with 0.1 N Na2S203.

Calculations

g/1 total sulfites =

ml Iy -  (ml Na?S?03 x N NagS-.O^/N !„) x N I-, x 0.0160321 x 1000 =
                        ml sample analyzed
ALKIMETRIC METHOD FOR ANALYSIS OF TOTAL BISULFITE SULFUR AND
TOTAL SULFUR

Manipulations

    Aliquot 10 ml of sample to a 250-ml volumetric flask
containing about 100 ml of deionized water and 10 ml of 30%
hydrogen peroxide.  Make to volume with deionized water and allow
to cool and make to volume again, mix thoroughly.

    Take a 20-ml aliquot of the diluted sample into a 250-ml,
wide-mouth erlenmeyer flask.  Add 8 drops of methyl red-methylene
blue mixed indicator to the flask.  Titrate with approximately
0.2 N NaOH to the first green end point.  Record the ml of NaOH
used as the titer Ta for calculating the grams per liter of
bisulfite sulfur.

    Add 10 ml of formaldehyde to the same sample.  Add one
dropper of phenolphthalein-methylene green indicator, continue to
titrate with the 0.2 N NaOH through the blue color, through the
green color, and to the first dark blue color.  Record the ml of
NaOH used; this is titer Tb.

                              68
-------
    Establish a blank on each new bottle of formaldehyde.  Add 20
ml of deionized water to a 250-ml erlenmeyer flask, add 10 ml of
formaldehyde to the flask, add eight drops of methyl red-
methelene blue indicator.  Titrate to the green end point with
0.2 N NaOH.  The ml of the 0.2 N NaOH used to titrate 10 ml of
formaldehyde is the blank.

Calculations

    g/1 HS03 = ml NaOH (Taj x N NaOH x 0.032064 x 1000
                   ml of sample analyzed

    g/1 total sulfur = (Tb - blank) x N NaOH x 0.0160324 x 1000
                             ml of sample analyzed
ANALYSIS OF THE ACIDULATOR AND STRIPPER LIQUORS FOR BISULFITE,
BISULFATE, AND TOTAL SULFUR

    The total sulfites are determined by the iodimetric method.
It is assumed that when the pH of the acidulator is 2.0 or below,
that all of the sulfites are in the form of ammonium bisulfite.
The acidulator and stripper also contain ammonium bisulfate and
ammonium sulfate.  The acidulator and stripper are analyzed
essentially the same way as the samples from the ammonium
scrubber pilot plant, as described above; however, the
calculations are somewhat different.

    The total sulfites are calculated to ammonium bisulfites.  A
separate determination is made for the ammonium bisulfate and
total sulfur alkimetrically.

Manipulations

    Aliquot 10 ml of the sample into a 250-ml volumetric flask
containing about 100 ml of deionized water and 10 ml of 30%
hydrogen peroxide; make to volume with deionized water.  Take a
20-ml aliquot into a 250-ml erlenmeyer flask.  Add to it 10 drops
of methyl red-methylene blue mixed indicator.  Titrate with 0.2 N
NaOH to the green end point, and record ml used as the titer Ta
for calculating the ammonium bisulfate.  Add 10 ml of
formaldehyde to the sample, then one dropper of phenolphthalein-
methylene green mixed indicator, continue to titrate through the
blue color, through the green color, and to a dark blue color.
This is the end point for the total sulfur, record mis used as
titer Tb.

    Establish a blank on each new bottle of formaldehyde.  Add 20
ml of deionized water to a 250-ml erlenmeyer flask, add 10 ml of
                               69
-------
formaldehyde to the flask, add eight drops of methyl red-
methylene blue indicator.  Titrate to the green end point with
0.2 N NaOH.  The ml of 0.2 N NaOH used to titrate 10 ml of
formaldehyde is the blank,

Calculations

Ammonium Bisulfate--

    g/1  (NFUHSO*) (ml sample) = ml 0.2 N NaOH due to NH4HS03 = a
           N NaOH x 99.112

                   g/1 NH4HS04 =  (Ta-a) x N NaOH x 115.112
                                    ml sample analyzed

Total Sulfur--

         g/1 total sulfur =  (Tb - blank)  x N NaOH x 16.0324
                                  ml sample analyzed
SILVER NITRATE METHOD FOR ANALYSIS OF CHLORIDE

    The pH of the sample to be tested is adjusted between the
limits indicated by methyl orange and phenolphthalien indicators.
The chloride ion is titrated with silver nitrate solution in the
presence of potassium chrornate.  The silver reacts with the
chloride forming silver chloride which precipitates.  When all
the chloride has precipitated, red silver chromate is formed thus
indicating the end point.

Manipulations

    The sample to be tested should have been previously filtered
through Whatman No. U2 filter paper or similar grade paper.  From
previous analysis, estimated concentration, and the tabulation below
determine size sample to analyze.  The concentration of chloride
ion should be between 5 and 200 ppm in the portion titrated.

         Estimated ppm chloride   Sample size, ml

                  0-150                100
                150-300                 50
                300-650                 25
              Above 650                 10
                               70
-------
    Pipet size sample selected into a porcelain casserole.  When
size sample selected is less than 100 ml, dilute the sample to
about 100 ml with distilled water.  Add a few drops of 0.1%
phenolphthalein indicator to the sample and discharge the pink
color by careful addition of a few drops of 0.5 N sulfuric acida.
Add about 1 ml of potassium chromate solution to the sample.
Titrate the sample with standard silver nitrate until one drop
produces a faint reddish color that does not disappear upon
stirring.  Record the ml of silver nitrate as titer Ta.  Using
reagents used in the analyses, make a blank determination by
titrating a volume of distilled water equal to that used to
dilute sample.

Calculations

Cl-mg/l= Fml AgNO*  (Ta) - blank] x  (mqCl~/ml silver nitrate) x 1000
                             ml analyzed
AMMONIA IN EXIT FLUE GAS SAMPLE  (DIRECT NESSLERIZATION METHOD)

Method

    Nessler reagent and Rochelle salt solution are added to the
sample to be tested.  The resulting color intensity is determined
with a spectrophotometer by taking the light transmittance at 425
millimicrons through a 2.5-cm cell.

    The NH3 concentration in the unknown sample is determined by
comparing its color intensity with the color intensities of
samples containing known concentration of NH3.  The comparison is
made from a graph previously prepared by plotting the light
transmitted through the color developed in standard samples
against the concentration of ammonia in them.

Manipulations

    1.   From previous analyses of the same type samples,
         estimate the concentration of NH3 in the sample.  Then
         use the tabulation below to determine aliquot to use.
a.  If for any reason the  sample  of water  is  acid, add 0.5 N
    sodium hydroxide solution until a pink color is obtained
    with phenolphthalein solution.  Then add  just enough  0.5 N
    sulfuric acid to discharge the pink color.
                               71
-------
        Sample concentration,  dilution,  and aliquot: to use	
     Approximate     ml of    Volume    ml of diluted     ml
   concentration,    original  diluted      sample      original
       ppm NH3        taken   to,  ml      analyzed     analyzed
0.0-1.4
1.5-2.9
3.0-7.3
7.4-14.5
14.5-29.0
29.0-58.1
-
-
50
25
25
-
-
500
500
500
-
-
100
100
50
100
50
20
10
5
2.5
2.    Transfer selected aliquot of filtered samples into
     separate 100-ml Nessler tubes.  If aliquot is less than
     100 ml,  dilute to 100 ml with ammonia-free distilled
     water.  Always test distilled water for ammonia before
     using it.  If sample is colored or turbid and not water
     clear, transfer a duplicate aliquot into another Nessler
     tube.  Add 1 ml of Rochelle salt and determine light
     transmitted through it to make sure it is not darker
     than reagent blank used to adjust instrument as
     described below.  If its color is darker than reagent
     blank, adjust instrument with it and determine light
     transmitted through portion of same sample reacted with
     Nessler at the new instrument setting.

3,    Prepare a reagent blank by adding 100 ml of ammonia-free
     distilled water to another 100-ml Nessler tube.

4.    Add 1 ml of Rochelle salt solution to each sample and
     reagent blank.  Stopper each Nessler tube with clean
     polyethylene stopper and mix by inverting two or three
     times.  Never use rubber stoppers in this step and step
     5 because a color other than ammonia reaction may
     result.

5.    Add 1 ml of Nessler reagent solution to each sample and
     reagent blank; again, stopper and mix as in step 4.
     Allow color to develop 30 minutes.

6.    Transfer sample containing Rochelle salt and Nessler
     reagent into spectrophotometer test tube and read
     percent transmittance.

7.    From transmittance reading determine mg NH4 and/or ppm
     from a prepared chart.
                          72
-------
Calculations

         	mg NH» x 1000	= ppm NH4
         ml of orig. sample used for comparison

         ppm NH4 x 0.94U = ppm NH3

         ppm NH4 x 0.777 = ppm N



PREPARATION OF AMMONIA REAGENTS  (FOR NESSLER METHOD)

Nessler Reagent

    Dissolve 100 g mercuric iodide  (HgI2) and 70 g potassium
iodide, (KI) in a small quantity of ammonia-free distilled water
and add this mixture slowly, with stirring, to a cool solution of
160 g NaOH in 500-ml ammonia-free distilled water.  Dilute to 1
liter with ammonia-free distilled water.

    Stored in Pyrex glassware and out of sunlight, this reagent
is stable for periods up to a year under normal laboratory
conditions.

    The reagent should give the characteristic color with mg/1
ammonia nitrogen within 10 min after addition but should not
produce a precipitate with small amounts of ammonia within 2 hr.

CAUTION:  This reagent is very toxic; take care to avoid
ingestion.

Rochelle Salt Reagent

    Dissolve 500 g of reagent grade KNaC4H406-4H20 in 1 liter of
distilled water.  Boil off 200 ml or until free from ammonia.
Cool and dilute to 1 liter with ammonia-free distilled water.
PREPARATION OF STANDARD AMMONIUM CHLORIDE AND AMMONIUM SULFATE
SOLUTIONS FOR CALIBRATING SPECTROPHOTOMETERS

Ammonium Chloride and Ammonium Sulfate Stock Solutions

    Dissolve 1.1862 g anhydrous reagent grade ammonium chloride
or 1.U652 g ammonium sulfate, dried at 100°C, in ammonia-free
distilled water and dilute to 2000 ml with NH3 free distilled
water.  Mix well.
                               73
-------
Standard Solution Containing 0.002 mg NHA per ml

    Pipet 20 ml of either stock solution and transfer into a 2000
ml volumetric flask.  Dilute to 2000 ml with ammonia-free
distilled water and mix well.  This solution is used for
calibrating spectrophotometers.
PROCEDURE FOR SAMPLING INLET OR EXIT FLUE GAS FOR PARTICULATE AND
SULFUR DIOXIDE

Sampling

Apparatus (Figure A-l) —

    A.   Environeering dust filter.
    B.   Four 600-ml gas scrubber bottles arranged in order
         listed below.   (An additional scrubber bottle of
         peroxide may be required for inlet determinations.)

         1.    Dry trap with short open-end sparger.
         2.    6% hydrogen peroxide solution with fritted glass
              imping er  (250 ml) .
         3.    6% hydrogen peroxide solution with fritted glass
              impinger  (250 ml).
         4.    Dry trap with short open-end sparger.

    C.   Dry test meter.
    D.   Vacuum supply.

Procedure --

    A.   Insert Environeering  sample nozzle into gas duct.
    B.   Pull approximately 0.5 cfm sample for about 120 min
         (increase vacuum to maintain flow).
    C.   Record pressure and temperature readings at meter.
    D.   Record pressure and temperature readings of duct  (wet
         and dry bulb for exit gas sample).
    E.   Combine the peroxide  bottles and  analyze for S02.
    F.   Dry the filter  paper  at 110°F and weigh.
                               74
-------
                                            3
                                            U
                                            O
                                            Q.

                                            TJ
                                            C
                                            O
                                            O
                                            in
                                            O
                                            0.
                                            Q.
                                            O

                                            C3>
                                            C

                                            "5.
                                            6
                                            o
                                            U)

                                            tf)
                                            o
                                            O
                                            Q>

                                            3
                                            O»
75
-------
Sample Calculation Sheet

Particulate and Sulfur Dioxide in Inlet Flue Gas—

     Test Data—

     Dry test meter readings
       Total volume through meter, ft3           23.2
       Temperature of meter, °F                  58
       Vacuum on meter, in Hg                    7
     Sampling time, min                          80
     Dust collected during sampling period, g    5.3320
     Sulfur caught in peroxide during sampling
      period, g                                  1.89
     Average moisture of inlet gas during
      sampling period, %                         7.75

     Calculations—

     1.  To convert the volume at meter conditions to  the volume
         at standard conditions, use the pressure-volume-temperature
         relationship expressed in the ideal gas law.
         where
              P  = initial pressure, in Hg
              V  = initial volume, ft3
              T  = initial temperature,  R or  (460 + °F)

              and PI, Vi, and TI  = above values at standard
                                    conditions
         Transposing,
                      o

         Inserting test data values,

              vi =  (29.92 - 7.0)a  (23.2)(460 +  70)
                        (460 + 58)  (29.92)

                 =  (22.92) (23.2) 530
                      (518) (29.92)

                 = 18.18 ft3
a.  Vacuum on meter.

                               76
-------
Converting from wet volume to dry volume, use the formula

     V,   = VTTO+.  (100 - avq % moisture by volume)
      ary    wet               1QQ


          = (18.18)  (100 - 7.75)
                         100

          = (18.18)  (0.9225)

          = 16.78 ft3  (dry at 29.92 in Hg and 70°F)

To convert the weight of dust collected on the filter during
the sampling period  into dust concentration in gr/dscf70/
merely convert the weight in g to gr by multiplying  by  15.43
(the number of gr in 1 g) and divide by the dry volume  of gas
at standard conditions.

     Dust loading =  (5.3320) (15.43)
                         16.78

                  =  4.9030 gr/dscf70

To convert the weight of sulfur caught in the peroxide
bottles into its equivalent volume of S02, divide the weight
of the sample by the gram molecular weight of sulfur and
multiply the quotient by the mol volume  (in liters)  divided
by 28.32  (the number of liters in 1 ft3).

     Vol. S02 sampled = samplg weight x  22.4
                         g mol wt of S   28.32

                      =  1.89 x 0.791
                        32.06

                      = 0.04664

To convert the volume of S02 sampled to  ppm in inlet gas,
multiply the volume  by 106 and divide by the combined volume
of S02 plus inlet gas.

     S02 in inlet gas = vol of S0y sampled x 10*	
                        vol of S02 sampled + vol of  gas

                      = 0.04664 x 10«
                         0.04664  +  16.78

                       =  2772  ppm
                            77
-------
Particulate and S02 in Exit Flue Gas—

     Test Data—

     Dry Test meter readings
       Total volume through meter, ft3              17.8
       Temperature of meter, °F                     58
       Vacuum on meter, in Hg                       7
     Sampling time, min                             60
     Wet-bulb temp, exit scrubber, °F               105
     Dry-bulb temp, exit scrubber, °F               106
     Dust collected during sampling, g              0.005
     Sulfur caught in peroxide during sampling, g   0.33

     Calculations—
         To convert the volume of gas from meter conditions^to  dry
         standard cubic feet
         the formulas below.
standard cubic feet at atmospheric pressure and 70 F, use
              Vi =  (29.92 - 7) (17.8) (460 + 70)
                          (460 + 58) (29.92)

                 =  13.95

              V,    = V    (100 - avg % moisture by volume)


                    = 13.95  (100 -  7.9)
                           100

                   = 12.86 (dscfyo)

         From the psychrometric chart, gas with  a wet-bulb
         temperature of  105°F  and  a  dry-bulb  temperature of
         106bF contains  0.0508 Ib  of water per  Ib of dry air.

         Then

              % moisture by volume = (Ib H20/lb  dry  air)  x  100
                                     _      ,__         j.

                                      (Ib H20/lb  dry  air  +_1	)
                                      (      18             30.4)

         where, 30.4 is the average molecular  weight  of the  flue
         gas.

                    =   (0.00282) x 100
                       (0.00282 + 0.0329)

                    =  7.9%
                               78
-------
To calculate the dust loading use the formula

     Dust loading, gr/dscf70 = q collected x 15.43
                               vol of gas, dscf70

     = 0.005 x 15.43
           12.86

     = 0.0060 gr/dscf70

To convert the weight of sulfur caught in the peroxide
bottles into its equivalent volume of S02, divide the weight
of the sulfur by the gram molecular weight of sulfur and
multiply the quotient by the mol volume  (in liters) divided
by 28.32  (the number of liters per ft') .

     Vol S02 sampled = 	sample wt     x 22.4
                        g mol wt of S    28.32

                      =  0.33 x 0.791
                        32.06

                      = 0.008142

To convert the volume of S02 sampled  to ppm in the exit gas,
multiply the volume by 106 and divide by the combined volume
of S02 plus exit gas.

     SO2 in exit gas = 	0.008142	
                         0.008142 + 12.86

                     =633 ppm
                           79
-------
PROCEDURE FOR SAMPLING EXIT FLUE GAS FOR AMMONIA

Sampling

Apparatus (See Figure A-2)—

    A.   Stainless steel sampling nozzle.
    B.   Four 600-ml gas scrubber bottles arranged in order
         listed below.

         1.    Dry trap with short open-end sparger.
         2.    Distilled water with fritted glass impinger  (250
              ml) .
         3.    Distilled water with fritted glass impinger  (250
              ml) .
         4.    Dry trap with short open-end sparger.

    C.   Dry test meter.
    D.   Vacuum supply.

Procedure —

    A.   Insert sampling nozzle into gas duct.
    B.   Pull approximately 0.5 cfm sample for about 30  min
          (increase vacuum to maintain flow).
    C.   Record pressure and temperature readings at meter.
    D.   Record pressure and temperature readings of duct  (wet
         and dry bulb).
    E.   Combine the water bottles and analyze for NH3  (see
         analysis procedure).

Sample Calculation Sheet

Ammonia in Exit Flue Gas--

    Test Data—

     Dry test meter readings
       Total  volume through  meter,  ft3           7.0
       Temperature of meter,  °F                 58.0
       Vacuum on meter,  in Hg                   3.0
     Sampling time,  min                         60
     Wet-bulb temp,  exit scrubber,   F            105
     Dry-bulb temp,  exit scrubber,   F            106
     NHs caught in water during sampling,  g     0.01037
                               80
-------
                                                       c
                                                       o
                                                       £
                                                       e
                                                       o
                                                       (A
                                                       3
                                                       O
                                                       O.
                                                       a.
                                                       o

                                                       o>
                                                       c

                                                      "5.
                                                       E
                                                       o
                                                       (A

                                                       at
                                                       o
                                                      CV
                                                       I
                                                      
-------
Calculations—

1.   To convert the volume of gas from meter conditions to
     dry standard cubic feet at atmospheric pressure and
     70°F, use the formulas below.

          vi = (29.92 - 3.0) (7.0) (460 + 70)
                    (460 + 58) (29.92)

             = 6.44 ft3 at 70°F

2.   From the psychrometric chart, gas with a wet-bulb
     temperature of 105°F and a dry-bulb temperature of 106°F
     contains 0.0508 Ib of water per Ib of dry air.  Then

          % moisture by volume =  (Ib H»0/lb dry air) x 100
                                  (         18         )
                               -  (Ib H,0/lb dry air  + JL	)
                                  (        18           30.4)

                               =  (0.00282) x 100	
                                  (0.00282 + 0.0329)

                               = 7.9%

3-        Vdry = Vwet  (100 - % moisture)
                              100

               = 6.44  (92.1)
                       100

               = 5.92 dscf70

     To convert the weight of ammonia caught in the  sample
     into its equivalent volume, divide the weight by the
     gram molecular weight of ammonia and multiply the
     quotient by the mol volume  (22.4 1) divided  by  the
     liters per ft*  (28.32).

          Vol NH3 = 0.01037  x 22.4
                     18        28.32

                  = 0.000459
     To  convert  this volume  to  ppm in  the  exit gas,  multiply
     the volume  by  106  and divide  by the combined volume of
         plus exit  gas.

          ppm NH3 = 0.000459 x  106
                    0.000459 +  6.44

                  =71  ppm


                            82
-------
                          APPENDIX B

                      SAMPLE CALCULATIONS


                           CONTENTS
Typical Absorber Product Liquor Composition 	   84
Vapor Pressures of SOa/ NHa, and H20 Calculated by
 Modified Johnstone Equations 	   88
Calculating the Vapor Pressure of SOa (ppm) Necessary
 to Cause a Fume Above or Below a Tray Using Moore's
 Fume Equation	   89
                              83
-------
                           APPENDIX B

                       SAMPLE CALCULATIONS



TYPICAL ABSORBER PRODUCT LIQUOR COMPOSITION

Solution

Sulfite sulfur (S0§ as S)              33.83 g/1
Bisulfite sulfur (HSO? as S)          109.18 g/1
Sulfate sulfur (SOf as S)              20.18 g/1
                          Total       163.19 g/1
pH - 5.7
Specific gravity - 1.226

Mols of Sulfur per Liter

Grams sulfur/I/molecular weight of sulfur = mols  of sulfur/1
Sulfite sulfur = 33.83/32 = 1.06 mols/1
Bisulfite sulfur = 109.18/32 = 3.HI mols/1
Sulfate sulfur = 20.18/32 = 0.63 mols/1

Total Grams of Salt per Liter

Mols  of sulfur x molecular weight of salt = grams of  salt/1
Sulfate sulfur     1.06 x 116 = 122.96 g  (NHJzSOs/l
Bisulfite sulfur   3.41 x  99 = 337.59 g NHi^SOs/l
Sulfate sulfur     0.63 x 132 =  83.16 g  (NH 0ZSO*/l
    Total salt                  543.71 g/1


Total SO-,, Mnls of SOg/Liter as Sulfite and Bisulfite

    S03 as Sulfite

     Mols  S02/l as sulfite = mols  sulfite sulfur/1 x  1.0

                           = 1.06 x 1.0

                           = 1. 06 mols
                               84
-------
    SOj,  as Bisulfite



   Mols   S02/l as bisulfite = mols  bisulfite sulfur/1 x 1.0



                             = 3.41 x 1.0



                             = 3.41 mols



    Total SQg



    Total S02 = S02 as sulfite + S02 as bisulfite



              = 1.06 + 3.41



              =4.47 mols/1



Active NHAf Mols of NHa/Liter as Sulfite and Bisulfite



    NH3  as sulfite



       Mols NH3/1 as sulfite = mols  sulfite sulfur/1 x 2.0



                             = 1.06 x 2.0



                             = 2.12 mols/1



    NHr,  as Bisulfite



      Mols  NH3/1 as bisulfite = mols  bisulfite sulfur/1 x 1.0



                               = 3.41 x 1.0



                               = 3.41 mols/1



    Active NH^



      Active NH3 = NH3 as sulfite + NH3 as bisulfite



                 = 2.12 + 3.41



                 = 5.53 mols/1






Total NH^, MnLs  of NHa/Liter as Sulfite, Bisulfite, and Sulfate



    NHa  as Sulfate



      Mols  NH3/1 as sulfate =  mols  sulfate/1 x 2.0



                             = 0.63 x 2.0



                             = 1.26  mols./l
                                85
-------
Total NH*

    Total NH3 = NH3 as sulfite  + NH3  as bisulfite + NH3 as sulfate

              = 2.12 + 3.U1  + 1.26

              = 6.79 mols/1

Reaction water f Grams H-,0/Liter Combined with S0? and EIH-,

         2NH3 + S02 + H20       *      (NH4) 2S03

         NH3 + S02 + H20        -»•      NH4HS03

         2NH3 + S02 + H20  +  1/2 02  -»•
     Reaction Water in Sulfite  (grams H20/l)

       Reaction water (SO") = mols S02 as sulfite x  1.0
                              x 18gH20/g mol

                            = 1.06. x 1.0 x 18

                            = 19.08 g/1

     Reaction Water in Bisulfite  (mols H20/l)

       Reaction water (HSOi) = mols SO 2 as bisulfite  x 1.0
                              x 18 gH20/g mol

                            = 3.41 x 1.0 x 18

                            = 61.38 g/1


     Reaction Water in Sulfate  (mols H20/l)

       Reaction water  (S0=^) = mols S02 as  sulfate  x  1.0
                              x 18 gH20/g  mol

                            = 0.63 x  1.0 x 18

                            = 11.34 g/1
                                86
-------
Free water, Grams HgO/Liter - Unreacted

    Free water =  (specific gravity x 1000)  - total salt g/1

               =  (1.226 x 1000)  -  543.71

               =  1226 - 543.71

               =  682.29 g/1

CValue  ,(mols total NH-,/100 mols
  C =	(Mols  total NH,) (1800)	
      (free water) + [reaction water (SOf)
     + reaction water  (HSOf)  * reaction water (SOfjjj

  C = 	6.79 (1800)	
              (682.29)  +  (19.08 + 61.38 * 11.34)

  C = 	12,222	  =  12,222
      682.29  + 91.80       774."09

  C =  15.78  mols total NH3/100 mols  H20

C  Value  (mols  active NH,/100 mols  H90)
 A	'	
  CA =	(mol s  active NHa)  1800	
        (free  water) +  [reaction water (SOf)
     * reaction water  (HSOf)  + reaction water

  CA = 	5.53 (1800)	
        (682.29) +  (19.08 + 61.38 +  11.34)

  CA = 	9,954        = 9,954
       682.29  +  91.80      774.09

  CA = 12.85  mols active  NH3/10Q mols  E9Q

A Value [mols  (NHA) gSO^/100 mols  Hj>01
  A = 	(mol s  sulfate sulfur/1)  1800	
       (free water)  +  [reaction water (SO,)
       + reaction water  (HSO^) + reaction water  (SO1?)]

  A =  I0«63) (1800)     =  1,134
        682.29  •«•  91.80 774.09

  A = 1.46 mols   (NHi)gSO>/100 mols  H,0
                                87
-------
S Value (mols  SO-,/100 mols  HgO)

            (Total SO,,  Mols/1) (1800)
      (free water) + [reaction water (S03)  + reaction water(HS03)
      + reaction water  (S04)]

  S =	(4.47) (1800)	
          (682.29) +  (19.08 +  61.38  +  11.34)

  S = 8,046
      774.09

  S = 10.39  mols.  S02/100 mols  H20

S/C- Ratio (mols   SOg/mols  active NH3)

    S/CA = mols  total S02/mols   active NH3

    S/C. = 4.47/5.53
       A

    S/CA = 0.81
VAPOR PRESSURES OF S02, NH3, AND H2O CALCULATED
BY MODIFIED JOHNSTONE EQUATIONS  (6)

  SOg Vapor Pressure of Absorber Liguor jmm Hg) at 126°F

  S02 Vapor Pressure = PSQ   =  MS (2 S/C  - 1) a
                          2      S/C  l-V<
  where  logtoM =  5.865  -	2369
                         liquor temperature, °K

                =  5.865  -  2369
                          325.37

       logl0M  = -1.416
            M  = 0.0384

  Pqn  = {0.03841 (10.39U (2 x 0.81) - l]2
    bU2        0.81 (1-0.81)

  P   = 0.1534
    S02   0.1539


  Pso2 = °-996
                                88
-------
  a Vapor Pressure of Absorber  Liquor (nun Hg)  at 126 °F
  NH3 vapor pressure = PNH   =  NC (1-S/C,) _
                          3     2 S/CA - 1

         where log10N =  13.680 - _ 4987 _ __
                                  liquor temperature, °K

         log10N = 13.680  - U987
                           325.37

         log10N = -1.647

              N = 0.0225

  PNTH  = (0.0225) (15.78) (1-0.81)
     3   (2 x 0.81) -1

  P    = 0.0675
         07620

  PNH  = 0.108 mm Hg

Hj.0 Vapor Pressure of Absorber Liquor (mm Hg) at 126°F

  H20 vapor pressure =  PH  Q  = PW x 	(100)	
                         2          CLOO  +  c"A + s + (3 x A)]

    where PW = vapor pressure of pure water at 126°F

  PH n = 103.31 x                 100
   n_U
                   {IQO + 12.85 +  10.39  +  (3  x 1~.46£J
  PH n = 10 • 3 31
    2°   127762
  P  _ = 80.95 mm Hg
   H2°
CALCULATING THE VAPOR PRESSURE OF SO2 (ppm) NECESSARY TO CAUSE
A FUME ABOVE OR BELOW A TRAY USING MOORE'S FUME EQUATION

   logioPs02 = -2102  + 4.3134 + 2 logio  [(100 + Ca + S + 3A)  (2S-CA)1
           2     T                       I     (CA + 2A)  (CA  -  S)      J

   where T is the tray liquor temperature in  °K = 325.37
                                89
-------
log10Ps()2
          = -2102   + 4.3134

            325.37
+ 2 log 10   ClOO + 12.85 +  10.39  + (3 x 1.46)11(2 x 10.39) -12.85:

                       [12.85  +  (2 x 1.46)] (12.85 - 10.39)



          = -2.1469 +  2 logic 26.087



   men  = 2.1469 + 2.8328
      oU 2


   ioPgQ  = 0-6859


Pe_  = 4.852 mm Hg
 SU 2
 S02
Pcn  (ppm) = 4.852 mm Hg  x  10'

              760 mm Hg


     (ppm) = 6,384
                           90
-------
                          APPENDIX C

                        CORROSION DATA


                           CONTENTS


                                                          Page

Tables

C-l  Nonmetallic Materials Tested in the Pilot Plant
      for Removal of Sulfur Dioxide by the Ammonia
      Absorption Process 	     93

C-2  Corrosion Tests in the Ammonia Absorption -
      Ammonium Bisulfate Regeneration Pilot Plant  ...     94

C-3  Corrosion Tests in the Ammonia Absorption -
      Ammonium Bisulfate Regeneration Pilot Plant  ...     95
                              91
-------
                           APPENDIX C

                         CORROSION DATA

    Corrosion tests were made in the pilot plant during the
following periods of operation:  December 3-14, 1973 (period A)
and July 2-18, 1975 (period B).  The test specimens were immersed
in the prewash sump liquor and in the gas duct leading from the
prewash section.  The first prewash section (period A)  was made
of 316L SS and was operated in an open-loop mode.  The unit was
susceptible to the highly corrosive prewash liquor  (pH = 2.7)
even with a purge rate of 15 gallons per minute of fresh water
through the prewash.  The metal surfaces exposed to the gas phase
were severely pitted (40 to 60 mils deep).  Based on these
factors a second prewash was designed and constructed of
fiberglass-reinforced plastic  (period B) .  The FRP prewash was
operated in a closed-loop manner and was impervious to the
prewash liquor  (pH = 1.0).

    Table C-l lists the trade name, base type, and manufacturer
of each of the nonmetallic materials tested.  Tables C-2 and C-3
list the various corrosion data and material evaluations.

    During period A, specimens of Type 316L, Type 201, Type 304-
L, USS 18-18-2 stainless steels, Carpenter 20, mild steel,
neoprene, and Bondstrand 4000 were tested.  The test results are
listed in Table C-2.

    The corrosion rates for stainless steels in the venturi sump
liquor (with the exception of 18-18-2) were less than 2 mils/yr.
Either pitting or crevice corrosion or both occurred on each
specimen.  The corrosion rate for USS 18-18-2 was 54 mils/yr with
minute pitting and attack in the heat-affected zone of the weld.

    The corrosion rates for the stainless steels tested in the
gas duct ranged from 42 to 146 mils/yr with crevice corrosion
occurring on each specimen.  The specimens in the gas duct were
wetted by about 2.5 gallons per hour of mist from the venturi
sump.  The mist contained sulfur dioxide in equilibrium with the
flue gas stream containing about 2400 ppm sulfur dioxide and had
a pH of about 1  (compared with an average pH of  2.7 in the sump).
The lower pH accounts for the higher corrosion rates experienced
in the gas ducts.  Mild steel corroded at excessively high rates
in both locations  (580-2200 mils/yr).
                                92
-------
      TABLE  C-l.  NONMETALLIC MATERIALS TESTED IN THE PILOT PLANT  FOR

        REMOVAL OF SULFUR DIOXIDE BY THE AMMONIA ABSORPTION PROCESS
     Trade name
    Base type
   Manufacturer
Bondstrand 4000
Kerolite, blue
Keroseal
Polypropylene
Neoprene  (sheet)

Rubber, butyl
(covered mild steel)
Rubber, natural gum
(covered mild steel)
Rubber, neoprene
(covered mild steel)
                         Epoxy,  fiberglass
                         reinforced
Polyethylene
                         Polyvinyl  chloride
Polypropylene,
rigid

Chloroprene polymer

isobutylene-
isoprene. Gates
No. 26,666

Polyisoprene,
Gates No. 1375
Ameron
201 N. Berry Street
Brea, CA   92612

Kearny Fluid Equipment,  Inc.
Raritan, NJ  08869

B.F. Goodrich Industrial
  Products Company
Tuscaloosa, AL  35403

American Viscoe Corporation
Philadelphia, PA  35403
Chloroprene polymer,
Gates No. 9150
Gates Rubber Company
999 S. Broadway
Denver, CO  80217

Gates Rubber Company
999 S. Broadway
Denver, CO  80217

Gates Rubber Company
999 S. Broadway
Denver, CO  80217
                                   93
-------
     TABLE C-2.  CORROSION TESTS  IN THE AMMONIA ABSORPTION - AMMONIUM

                    BISULFATE REGENERATION PILOT PLANT

                          (December  3-14,  1973)

                                       	Location of specimen
                                       Immersed in
                                       prewash sump
Sump outlet
  gas duct
Operating time, hr
Exposure conditions
Test medium
Solids, % by wt
PH
Flow rate
Gal/min
Acfm
Temperature, °F
Chemical analysis, g/1
SO 4 (total sulfur)
Cl
Corrosion rate of metals, mills/yrc
Carpenter 20 C6-3 welded to
Carpenter 20 C6-3
Type 316L welded to type 316L
Type 201 welded to type 316
Type 304L welded to type 30 8L
USS 18-18-2 welded to Inconel 82
Mild steel A-283 welded to E6012
Condition of nonmetallic specimens^
Bondstrand 4000
Neoprene (sheet 0.223 in. thick)
217

Sump liquor
0.002*
2.4 (2.0-3.9)

30a
-
117(101-124)

0.15(0.09-0.22)
0.1(0.04-1.5)


<1, Pm
-------
    TABLE  C-3.  CORROSION TESTS IN THE AMMONIA ABSORPTION - AMMONIUM

                  BISULFATE REGENERATION PILOT PLANT

                        (July 2-18, 1975)


                                  	Location of specimen	
                                  Immersed in
                                  prewash sump
               Downstream of prewash
                  mist eliminator
Operating time, hr
Exposure conditions
  Test medium
    Solids, % by wt
    pH
    Chemical analysis13
      Undissolved solids, %
      Total dissolved solids, g/1
      Sulfur as S04r g/1
      Total iron, g/1
      Cl, g/1
      Ca, g/1
      Al, g/1
      K, g/1
      Na, g/1
    Temperature, °F
Corrosion rate of metals, mils/yrc
  Carpenter 20 C6-3 welded
   to Carpenter 20 C6-3
  Duriron, not welded
  Hastelloy G welded to Hastelloy G
  Illium P, not welded
  Inconel 625 welded to Inconel 625
  Inconel 800 welded to Inconel 82
  Inconel 825 welded to Inconel 135
  Type 316L welded to type 316L
  USS 18-18-2 welded to Inconel 82
Evaluation of nonmetallic materials 8
  Kerolite, polyethylene
  Koroseal, polyvinyl chloride
  Polypropylene
  Rubber, butyl
  Rubber, natural gum
  Rubber, neoprene
     221"

Sump liquor
     0.5
     1.0

     0.031
    72.0
    50.4
     7.U
     6.0
     0.9
     0.8
     0.2
     0.1
127(125-130)
1,
 3
P-ll
 le
29d
                    190'
            Saturated flue gas
   1,126,  P-l
   4,d p-20
    8,d p-15
     <890f

     Poor
     Good
     Fair
     Good
     Fair
     Good
                 136 (124-150)
                    38, d  P-8
                    5, P-25
                       5e
                    U3,d  P-9
               65,
                 86
                   Good
                   Good
                   Good
                   Good
                   Good
                      P-ll
                      P-9
a.  Because water  (used 29 hr) or air  (used 14 hr) is not corrosive
    to the alloys tested, the rates for the alloys in the sump
    were determined on the basis of 192 hr exposure and those
    in the duct 176 hr.
b.  Analysis of liquor from clarifier near end of test.
c.  "P" preceding a number indicates pitting during the exposure
    period to the depth in mils shown by the number.
d.  Crevice corrosion at Teflon insulator.
e.  Attack of weld.
f.  Specimen corroded to failure during test.
g.  Evaluation:  good, little or no change in condition of
    specimen; fair, definite change—probably could be used;
    poor, failed or severely damaged.
                                  95
-------
    Both nonmetallic materials showed good resistance to
corrosion in both locations.  The hardness of neoprene did not
change appreciably during the test period.  These short duration
tests indicate that neoprene-lined equipment resists attack
downstream from the venturi element where temperature is not a
factor (the maximum temperature in the gas duct was 133°F; the
maximum recommended operating temperature for neoprene is 150°F).
Long duration tests are needed to evaluate neoprene fully.

    Corrosion test specimens of nine alloys, three plastics, and
three rubbers were exposed during particulate removal efficiency
tests carried out in period B.  The test results are listed in
Table C-3.

    The specimens in the sump were immersed a total of 221 hr
(192 hr in scrubber liquor and 29 hi" in water) .  Specimens in the
gas duct downstream from the mist eliminator were exposed for 176
hr of operation with flue gas and for 14 hr with air for a total
of 190 hr.  The pilot-plant equipment and the test specimens were
washed clean at the beginning of each idle period.

    Corrosion of alloys by water or air at temperatures of 125°
to 150°F is usually negligible.  Therefore, the rates of attack
of alloy specimens in the sump were determined on the basis of
192 hr of exposure to scrubbing liquor and those in the duct on a
basis of 176 hr of exposure to treated flue gas.  However, the
total exposure periods were considered in evaluation of the
nonmetallic materials.

    The corrosion rates for the alloys ranged from less than 1
mil/yr for Inconel 625 in both tests to 1126 mils for Incoloy 800
in the sump liquor.  Six of the nine alloys had lower rates in
the liquor than in the flue gas.  Two alloys, Incoloy 800 and USS
18-18-2 were corroded at higher rates by the liquor; their rates
were excessively high  (<890 and 1126 mills/yr).  Duriron and
Hastelloy G each had rates of 1 mil/yr in the liquor and 5 mils
in the gas.  Carpenter 20 Cb-3 had a low rate, 3 mils/yr, in the
liquor, but the rate was 38 mils with pitting in the gas.
Inconel 625 was the only alloy not affected by localized attack
in both test locations.

    The three rubbers and two plastics tested in the treated flue
gas duct showed good resistance to deterioration.  The conditions
were more severe for the specimens immersed in the liquor sump.
Three materials were in good condition-Koroseal  (PVC), butyl, and
neoprene.  Polypropylene and natural rubber were in fair
condition.  Kerolite, a polyethylene coating, failed.  No change
was detected in the Durometer A hardness of the three rubbers and
of Koroseal exposed in the sump or in the gas duct.
                              96
-------
                          APPENDIX D

                     EQUIPMENT EVALUATION


                           CONTENTS
Equipment for the Pilot Plant	    98
  Ductwork to Pilot Plant 	    98
  Gas Prewash Section No. 1	    98
  Gas Prewash Section No. 2	    98
  Settling Tank	    99
  Absorbers	    99
  Exit Gas Reheat System	107
  Blower System 	   107
  Pumps	107
  Acidulator-Stripper No. 1	Ill
  Acidulator-Stripper No. 2	Ill
  Evaporator-Crystallizer 	   113
  Crystalline Ammonium Sulfate Separation Equipment .  .  .   113
  Piping Materials  	   117
Instrumentation 	   117
  Gas Flow Rate Measurement	117
  Flowmeters for Liquid Flow Measurements 	   118
  Process Recorders and Controllers 	   118
  Gas Analyzers for the Pilot Plant	118

Figures
  D-l  Second Prewash Section (Fiberglass Reinforced
        Plastic)  	100
  D-2  Flyash Settling Tank in Prewash Section  	   101
  D-3  Absorber Configuration A 	   102
  D-4  Absorber Configuration B 	   103
  D-5  Absorber Configuration C 	   104
  D-6  Koch Flexitray	105
  D-7  Module for Reheating Stack Gas with Indirect
        Steam	108
  D-8  Fouled Reheat Module 	   109
  D-9  Final Acidulator-Stripper  	   112
  D-10 Oslo-Type Evaporator-Crystallizer  	   114
  D-ll Ejector-Condenser  	   115
  D-12 Eimco Extractor Horizontal Belt Filter 	   116
  D-13 Typical Foxboro Magnetic Flowmeter 	   119
  D-14 Pilot-Plant Control Board  	   120


                              97
-------
                           APPENDIX D

                      EQUIPMENT EVALUATION
EQUIPMENT FOR THE PILOT PLANT

Ductwork to Pilot Plant

    No maintenance was required on the 16-in.-diameter rigid mild
steel ductwork used for transporting the flue gas to the pilot
plant.  A 1/2- to 1-in. scale and flyash buildup protected the
duct from the condensing S03 acid mist in the inlet flue gas.

Gas Prewash section No. 1

    The initial venturi section was a 1-ft-square duct construc-
ted of 1/4-in. type 316L stainless steel.  Rods (3/4-in. type
316 SS pipe) were laid across the venturi throat perpendicular
to the gas flow.  The number of rods was changed to vary the
pressure drop across the venturi throat.  The venturi section
was mounted on a sump constructed of 10-gage type 316L SS.  The
venturi sump was the reservoir for the recirculating prewash
liquor and also channeled the conditioned gas into the type 316L
SS ductwork (14-in. O.D.)  leading to the absorber.

    The venturi housing was severely damaged by erosion and
corrosion and was replaced, after 650 hr of operation, with one
coated with urethane.  Urethane was used to protect the steel
from the low pH  (1.5 to 3.0) liquor and the abrasive flyash in
the recirculating liquor.  The urethane coating in the upper
portion of the venturi throat was completely destroyed after 700
hr of operation.  The rods  (316 SS) across the venturi throat
were badly pitted and needed replacing.  The sump was inspected
after 2,000 hr of operation.  The walls of the sump were corroded
and showed heavy pitting above the normal gas-liquid interface.
The SS ductwork from the sump to the absorber also was severely
pitted.

    The initial gas prewash unit was not equipped with a mist
eliminator between the venturi section and the absorber.  Mist,
which contained flyash solids and dissolved materials, was
carried by the gas stream to the absorber.  The absorber product
liquor was diluted and its contaminant level increased.

Gas Prewash Section No. 2

    A second gas pretreatment section was designed and
constructed of fiberglass reinforced plastic  (FRP)  (Atlac 382)
and coated internally with an epoxy resin paint.  The venturi
                              98
-------
throat was lined with a 1/8 in.-thick neoprene skirt while the
rod supports were made from I/2-in. neoprene.  The unit was
equipped with a plastic chevron mist eliminator (Heil Process
Equipment Company) mounted internally in the horizontal run.  A
schematic drawing of this prewash unit was shown in Figure 7  (see
text).  Figure D-l is a photograph of the prewash unit installed
in the pilot plant.

    The FRP and neoprene rubber were impervious to the low pH
(1.0)  corrosive sump liquor and the abrasive flyash in the
liquor.  Stainless steel fittings and spray nozzles failed and
had to be replaced periodically.  The original rods in the
venturi throat, made from 3/4-in. type 316 SS pipe, were replaced
with CPVC pipe.

    With a pressure drop of 10 in. of water across the venturi
throat and a liquor/gas  (L/G) ratio of 20  (approximately 55 gpm
of recirculating wash liquor), the prewash humidified and cooled
the flue gas before the gas entered the absorber.  The chevron
mist eliminator decreased the mist carryover into the absorber to
1 ml/m3 (2.09 x 10-« gal/1000  ft3).

Settling Tank

    A settling tank constructed of FRP (Atlac 382) was installed
in the prewash liquor loop to remove the undissolved solids from
the recirculating sump liquor.  The sump liquor was purged to the
settling tank at 0.5 or 1 gpm.  These rates correspond to
clarified liquor residence times of 8 and 4 hr, respectively.
The settling tank is shown in Figure D-2.  The unit performed
well during short test runs.  Long-term operations are needed to
properly evaluate the unit.

Absorbers

    The basic absorber was comprised of 4-ft sections each 32 in.
square.  The sections were constructed of either type 304 or  316
SS.  Three different configurations were tested as shown in
Figures D-3, -4, and -5.  The absorber of configuration D-3
originally had three beds of 3/4-in. glass marbles approximately
6 in. in depth.  It was built by NDC, Environeering, Inc.  The
marbles on the lower stage were subject to thermal shock and
cracking whenever cooler liquor came into contact with the heated
bed.  Liquor fall-through was a problem with the marble beds; in
some instances the level was completely lost on a bed.  Solids in
the scrubbing liquor would agglomerate the marbles causing
channeling of the liquor and gas streams resulting in liquor  fall
through and poor S02 removal efficiency.  Therefore, this marble
bed was replaced with a valve tray element  (Koch Flexitray).  A
typical valve tray element is shown in Figure D-6.  The
efficiency of the valve tray element was similar to that of the
                              99
-------
                                               FLUE  GAS INLET i
    ABSORBER
     INLET
                                                        PREWASH
                                                        SECTION
Figure  D-l.  Second prewash  section  (fiberglass  reinforced plastic).
                            100
-------
Figure D-2. Flyash settling  tank in  prewash section.
                      101
-------
                    FLUE GAS
                     OUTLET
       LIQUOR
        INLET
       LIQUOR
        INLET
                  wSssSsssssss
                                    G-3
                               -(MARBLE BED)
        LIQUOR
         INLET
  FLUE GAS
FROM VENTURI
 SECTION (V-1)
                                 CHEVRON MIST
                                  ELIMINATOR
LIQUOR
OUTLET
                                    G-2
                                (MARBLE BED)
                                 -»• LIQUOR
                                    OUTLET
                                   G-l
                                (VALVE TRAY)
                       LIQUOR
                       OUTLET
 Figure D-3. Absorber configuration A

                        102
-------
                    FLUE GAS
                    OUTLET
        LIQUOR
        INLET
        LIQUOR
         INLET
        LIQUOR
         INLET
        LIQUOR
        INLET
    FLUE GAS
  FROM VENTURI
   SECTION(V-I)
                    SECTION
                      6
                    SECTION
                      5
                  -L-i_<*i_f^f^-n^f^j^-f^Tti
Ljf
\       '
                             CHEVRON MIST
                              ELIMINATOR
                               G-4

                             (VALVE TRAY)
                                LIQUOR
                                OUTLET
                               G-3
                             (VALVE TRAY)
                               6-2
                             (VALVE TRAY)
                                G-l
                             (VALVE TRAY)
                             SECTION
                                I
                     LIQUOR
                     OUTLET
Figure D-4. Absorber configuration B
                        103
-------
            LIQUOR
            IN
             LIQUOR
             IN
             LIQUOR
             IN
             LIQUOR
             IN
    GAS  FROM
PREWASH SECTION
                         Lrun-nj-LTt-nr •
                          G-4 TRAY L
                          G-3 TRAY
                        XXXXXX
                          G-2 TRAY
                         XXXXXX
                                                                    GAS TO
                                                                 REHEAT SECTION
                                                      CHEVRON MIST ELIMINATOR

                                         -PLASTIC YORK PAD MIST ELIMINATOR
 G-4 LIQUOR
     OUT
 G-3 LIQUOR
     OUT


-S.S.  YORK  PAD MIST ELIMINATOR
  G-2 LIQUOR
     OUT

-S.S. YORK PAD MIST ELIMINATOR
                Figure  D-5. Absorber  configuration  C.

                                   104
-------
Figure D-6. Koch flexitray
          105
-------
marble bed; approximately 35% of the S02 to the absorber was
removed on the first stage.

    The two remaining marble beds later were replaced with valve
tray elements, and a fourth valve tray was added as a water wash
to decrease the quantity of ammoniacal salts in the entrained
mist leaving the absorber  (see Figure D-4, configuration B).
Entrained mist would be evaporated in a shell-and-tube reheat
element leaving scale deposits on the reheater tubes.

    Each valve tray element has an adjustable dam in the liquor
outlet weir box.  The liquor level on each tray could be varied
from 0 to 3 in.  A series of air and water tests determined that
a liquor depth of 2 in. on each tray and a gas flow rate of 2,800
acfm (125°F) were necessary to minimize liquor transfer from
stage to stage.

    The four-stage absorber was adequate to produce ammoniacal
liquors for regeneration test programs.  However, several
problems existed in the absorber performance area.  The valve
tray elements were prone to transfer liquor from one stage to
another, therefore, true stage separation was not achieved.  The
SOZ removal efficiency of the bottom two stages was very poor;
the Murphree tray efficiencies for stages G-l and G-2 averaged
43.6% and 66.1%, respectively.  A three-pas chevron mist
eliminator, mounted above the fourth stage in the vertical run,
was ineffective in preventing mist carryover from the absorber.

    Further modifications to the absorber resulted in the final
tower arrangement  (configuration C) shown in Figure D-5.  Mobile
plastic spheres (1-in.-diameter) were poured onto stages G-l and
G-2 to a depth of 12 in.  These spheres increased the average
Murphree efficiencies to 90% and 92% for stages G-l and G-2,
respectively.  Also, installation of SS wire mesh pads prevented
excessive mist carryover from stage to stage.  A chevron mist
eliminator  (Heil Process Equipment Corporation), mounted in a
horizontal run after the fourth stage, gave improvement but the
mist carryover still exceeded the standard limit of 110 mg/m3.
After a plastic York mesh pad mist eliminator was installed
between the fourth stage and the Heil mist eliminator, mist
carryover decreased to 70.6 mg/m3, well below the standard.

    The final absorber configuration improved control of absorber
liquor concentrations and S02 removal and decreased mist
carryover to acceptable levels.  However, liquor fall through
 (weepage) continued to be a problem in the vertical configuration
and resulted in less than true stage separation.  It is possible
that a horizontal-packed absorber would solve this problem.
                             106
-------
Exit Gas Reheat System

    The exit flue gas was reheated to 175°F in most tests with an
in-line, indirect steam-heated reheater.  The heat exchanger
element contained 234 (12 rows) l-in.-O.D. by 20-in.-long tubes.
The heat transfer area was 102.4 ft2.  The tubes were constructed
of the following materials:  Inconnel 625, Incoloy 825, 316L SS,
Cor-Ten A, and Hastelloy C-276.  A reheat module is shown in
Figure D-7.

    The calculated overall heat transfer coefficient  (Uo) ranged
from 14.8 to 27.6 Btu/(hr) (ft*) (°F).  The pressure drop across
the tube banks averaged 0.9 in. of water.  The unit required no
maintenance except for removing scale deposits.  Before the
plastic mesh-pad mist eliminator was installed ahead of the exit
chevron mist eliminator, heavy mist carryover resulted in the
buildup of scale on the tubes as shown in Figure D-8.  The scale
consisted largely of ammonium sulfate and ammonium chloride.  The
reheat element dissipated the water vapor in the scrubbed flue
gas but did little to reduce the opacity of the ammonia-sulfur
plume.

Blower System

    The flue gas was moved through the system with three fans;
two constant-speed drive fans ahead of the absorber and a
variable-speed drive fan after the absorber.  The two constant-
speed fans were installed in series.  All of the fans were
manufactured by American Standard and were constructed of 1/4-in.
type 304L SS plate.  The fans were V-belt driven by 40-hp motors
and were rated for 4,000 acfm at 300°F.  They were operated so
that the bottom of the absorber was under a slight pressure and
the top under a slight vacuum.

    The blowers were relatively maintenance free.  Flyash
deposits in the blower housings were minimal since the flue gas
was drawn from downstream of the electrostatic precipitators.
One set of fan bearings was replaced after approximately 3,500 hr
of operation.  The fluid drive on the variable-speed fan required
priming after lengthy periods of inactivity.
    Scrubbing liquor was circulated to three of the four absorber
stages  (G-l, G-2, and G-3) with Allen-Sherman-Hoff  (A-S-H)
centrifugal pumps.  The A-S-H pumps were of split housing
construction with removable neoprene rubber linings and neoprene-
coated impellers.  The pumps were coupled to their respective
motors  (15-hp General Electric) by American Standard  (Gyrol)
fluid drives.  The fourth absorber stage  (G-4) was fed with a
Wilfley centrifugal pump.  The casing and impeller were
constructed of type 316L SS.  The Wilfley pump was V-belt driven
with a 15-hp General Electric motor.
                             107
-------
Figure D-7. Module  for reheating stack gas  with indirect  steam
                           108
-------
Figure D-8.  Fouled reheat module
              109
-------
    All of the A-S-H and the Wilfley pumps gave excellent service
with minor problems.  No noticeable wear was observed in the
rubber linings or on the rubber-coated impellers.   The main
problem area with the A-S-H pumps was in the automatic control
loop.  The control motor would not regulate the speed of the pump
motor.  Changes in the pumping rates were made manually.  Some
shaft leakage occurred with the Wilfley pump while circulating
liquors with high specific gravities (1.25).  Some of the
mechanical seals stuck because of solids.  Other Wilfley pumps
gave excellent service in general use.

    The A-S-H pump unit supplied with the evaporator-crystallizer
failed in two areas.  The pump motor was undersized  (3 hp) and
burned out within 200 hr of operation.   The motor was replaced
with a 15-hp motor which functioned satisfactorily.  The neoprene
lining on the suction and shell side of the pump was damaged by
erosion and chemical attack.   (Original specifications called for
Hypalon linings but neoprene was supplied inadvertently.)  The
erosion resulted from contact with the  (NH4)2S04 crystals and the
chemical attack from the high temperature sulfate solution.  The
literature indicates that neoprene is resistant to  (NH4)2S04
slurries at temperatures to 212°F though the maximum recommended
temperature is 150°F.  The temperature in the crystallizer
reached 205°F for short periods of time.  The suction side of the
impeller was completely destroyed and the underlying mild steel
severely corroded.  The neoprene liners were replaced with
Hypalon liners and the impeller was replaced with a Hypalon-
coated impeller.  The operating temperature of the crystallizer
was restricted to 175°F and the pump performed satisfactorily
during the remaining test programs.

    An A-S-H pump  (split housing, 3-in. suction, neoprene lined,
and V-velt driven) was used to recirculate the prewash liquor in
the venturi sump.  There was no corrosion from the low pH  (1.0)
liquor nor any erosion from the undissolved solids  (flyash).

    A Tuthill gear pump, used  to pump absorber product liquor to
the storage tanks, performed satisfactorily in all tests.

    A Jabsco pump  (i-in. suction, air driven) failed to
consistently pump the  (NH4) 2S04 slurry  (15% solids)  from the
evaporator-crystallizer to a filter.  The pump was unable to
maintain suction against the high vacuum  (20 in. mercury) in the
crystallizer.  The plastic impeller blades broke off during the
infrequent periods when the pump was able to move  the slurry.
With  a motor-driven Jabsco pump inserted into a recirculation
feed  loop at near atmospheric  suction pressure, it was able to
pump  a 10%  (NH4)2S04 crystal slurry continuously to  and  from a
centrifuge in the sulfate separation section.
                             110
-------
 Acidulator-Stripper No. 1

    The first acidulator-stripper unit tested was used in both
Phase II and Phase III.  It was shown in Figure 10  (see text) .
The acidulator, 6 ft by 1 ft diameter, was constructed of type
316L SS schedule 10 pipe  (wall thickness = 0.180 in.)  and coated
internally with Teflon.  A mixing cone  (316L SS) located near the
top of the acidulator received the sulfuric acid and absorber
product liquor streams.  The mixed stream dropped from the cone
to a pool of retained acidulant in the lower portion of the
vessel.  The stripper was the same size as the acidulator and was
made of the same materials.  It was packed with 54 in. of dumped
2-in. Tellerette packing rings.  The rings rested on a Teflon-
coated type 316 SS screen located approximately 6 in. from the
bottom of the stripper.  Both the acidulator and the stripper
were oversize.  Intimate mixing of the acid ion source (sulfuric
acid) and the absorber product liquor to obtain complete
acidulation was not achieved in the acidulator.  Also, in the
stripper, the packing irrigation rate (2.04 gal of acidulated
liquor/min/ft2 of packing cross sectional area) was insufficient
to decrease the amount of free S02 in the effluent to 0.5 g/1.

    The extremely corrosive acidulated material and the elevated
liquor temperature caused material failure.  The reaction of the
sulfuric acid and the absorber product liquor is highly
exothermic with temperatures at the point of mixing reaching as
high as 190°F.  The SS mixing cone was destroyed.  The Teflon
liner in the acidulator deteriorated and eventually separated
from the vessel wall.

Acidulator-Stripper No. 2

    A corrosion resistant acidulator-stripper was constructed
from 4-in. I.D., schedule 40 plexiglass and PVC tubing.  This
unit was used during most of the Phase III work.  By trial and
error, the acidulator evolved as a mixing pot connected to the
stripper by a gravity overflow tube.  A schematic drawing was
shown in Figure 11 (see text).  Figure D-9 is a photograph of the
unit.  The effective volume of the acidulator was 1.5 gal and the
liquor residence time approximately 3 min at a combined liquid
flow rate of 0.5 gpm.  The S02 flashed in the acidulator was
combined in a common vent system with the S02 released in the
stripper.

    The stripper design resulted from meetings with
representatives of Cominco.  The stripper contains 30 ft of
dumped Tellerette packing.  Stripping gas inlets were provided so
that 10, 20, or 30 ft of packing could be used.  With the reduced
cross sectional area (0.087 ft2), a packing irrigation rate of
5.7 gal of acidulated material per square foot of packing cross
sectional area and 5 ft3 of stripping gas per minute, the amount
of free S02 remaining in the stripper effluent was decreased to
                            111
-------
STRIPPED r*»
 LIQUOR
       ABSORBER
        PRODUCT
        LIQUOR
    Figure D-9. Final acidulator-stripper.
                  112
-------
0.5 g/1 or less.  The overall performance of acidulator-stripper
was excellent with no maintenance problems.

Evaporator-Crystallizer

    The evaporator-crystallizer was designed and constructed by
Goslin Division of Envirotech Corporation, Birmingham, Alabama.
This Oslo-type single-effect crystallizer is shown in the
schematic drawing in Figure D-10,  The vaporizer chamber was 1 ft
I.D. by 8.5 ft tall and rests on the crystallizer chamber which
was 2 ft I.D. by 13-ft tall.  A downcomer, extending from the
vapor chamber down into the crystallizer, was 5 in. I.D. by 12 ft
8 in. tall.  The mother liquor was heated externally in a tube-
and-shell heat exchanger with low-pressure steam (50 psig).  A
direct contact  (barometric) condenser and a steam ejector
connected in series maintained the vacuum in the vapor chamber
and also removed any chemical contaminants in the vaporizer off-
gas.  The steam ejector was powered with high-pressure steam  (250
psig).  The condenser and ejector are shown in Figure D-ll.  The
entire unit and its piping were constructed from type 316L SS.
The operating specifications required that the unit evaporate 200
Ib/hr of water from the (NH4) 2S04 solution at 170°F and 22 in. of
mercury vacuum.

    The overall performance of the unit was acceptable.  The
primary problem area was in establishing the operating parameters
for the unit.  The recommended steam pressure to the ejector  (250
psig) was insufficient to maintain the desired 22 in. of mercury
vacuum in the vaporizer section.  As a result the mother liquor
temperature exceeded 200°F and caused chemical attack on the heat
affected zones  (welds)  in the unit.  Increasing the steam
pressure to 270 psig resulted in the desired vacuum and operating
temperature  (170°F).  Solids buildup and eventual plugging of the
downcomer and heat exchanger tubes resulted from operating the
unit with a crystal loading of 20-30% by wt.  At lower loading
rates  (10-15%)  plugging did not occur and crystals of adequate
size  (70% plus 35 mesh) were produced.

    The (NH4) 2S04 crystals produced would not flow out of the
evaporator by gravity and had to be pumped to the crystal
separation equipment.  An insulated 1-in, pipeline to the
separation equipment plugged frequently during intermittent
operation.  A continuous recirculation feed loop was installed
and eliminated plugging problems.

Crystalline^Ammonium Sulfate Separation Equipment

    Two types of solids separation equipment were tested, a
vacuum belt filter and a screen bowl centrifuge.  The belt filter
was an Eimco model 112 extractor horizontal belt filter as shown
in Figure D-12.  The continuous belt filter  (10 ft2 of vacuum
                             113
-------
                                   f- WATER

                                        STEAM
                                 4" % CELL CONN
                                    PRODUCT
                                    (SLURRY)
Figure  D-IO. Oslo-type evaporator-crystallizer
                     114
-------
Figure D-ll.  Ejector-condenser.
             115
-------
Figure D-l2.Eimco extractor horizontal  belt filter.
                       117
-------
filter area)  was skid mounted.  The unit was self-contained with
a variable-speed belt drive, a Nash vacuum pump, two supernatant
liquor receiving tanks, and a pump to move the supernatant back
to the crystallizer.  The belt filter was too large for
continuous operation.  Sufficient material could not be
maintained on the belt to prevent vacuum breaks even at the
lowest belt speed (3 ft/min).  In batch-wise operation, the unit
removed (NH4) gSO* crystals at a rate sufficient to balance the
production rate of 200 Ib/hr.  The crystals produced were sized
about 10% plus 35 mesh and contained 5-10% moisture.  The
crystals were dried in a gas-fired rotary dryer to 2% moisture or
less.  The filter was maintenance free.

    The centrifuge was a 6-in. continuous screen bowl centrifuge
manufactured by Bird Machinery Company.  The unit was constructed
from type 316 SS.  The screen bowl contained 0.008-in.
circumfrential slots.  Slurry from the evaporator-crystallizer
was pumped continuously to the centrifuge through a recirculation
feed loop.  Sheaves were provided to permit operating the
centrifuge at 3,000, 3,500, and 4,000 rpm corresponding to g-
forces of 760, 1,040, and 1,350 Ib force/lb mass.

    The centrifuge gave acceptable service with few operational
problems.  A crystal separation rate of 200 Ib/hr was achieved
when the  (NH4)2S04 solids in the feed to the centrifuge was 10%
and the feed rate was 9 gpm.  The moisture content of the product
crystals was 3%.  The screen bowl would be blinded by "mud" when
the solids content of the feed decreased to 5%.  Varying the g-
force had little effect on the centrifuge performance.

Piping Materials

    Type 304 and 316 SS pipe and rubber hoses provided excellent
service in piping ammoniacal liquors.  Rubber hose was used in
all corrosive material handling applications  (prewash liquor and
acidulator-stripper liquor and gas effluent streams) and was
maintenance free.  The connectors for the hoses were Kam-Loc
fittings, a type of quick connect fitting.  Metal Kam-Locs  (SS
and black iron) failed in corrosive liquor streams.
Polypropylene fittings were corrosion resistant but had poor
impact strength.
INSTRUMENTATION

Gas Flow Rate Measurement

    The gas flow rate through the pilot plant was measured with a
8-1/2-in. I.D. sharp-edged orifice mounted in the ductwork
downstream from the absorber.  A Foxboro differential pressure
                               118
-------
(d/p)  cell used to sense the pressure differential across the
orifice did not perform reliably.  Pressure taps led from the
orifice to manometers mounted in the plant and control room.  The
orifice was calibrated and the gas flow determined from a graph.
The orifice gave reliable service with only occasional cleaning
of the pressure taps and manometer leads.

Flowmeters for Liquid Flow Measurements

    Foxboro magnetic flowmeters  (Figure D-13) were used in
measuring the flow rates of the following streams:

              Recirculating liquor to the absorber stages
              Absorber product bleedoff
              Humidification water to the prewash
              Forward feed water to the fourth-stage feed tank
              Absorber product liquor to the acidulator
              Forward feed to the evaporator-crystallizer

    All of these units gave reliable service.  A magnetic
flowmeter used to measure the recirculating prewash sump liquor
(pH = 1.0) failed.  The type 316 SS electrode was eaten away,
which allowed the corrosive liquor to penetrate the internals of
the metering tube.  The unit was replaced with one containing a
platinum--10% iridium electrode to withstand highly corrosive
liquids.  The forward feed  (water) magnetic flowmeters along with
the absorber product flowmeter were coupled with flow integrators
for detailed accounting of the flows.

Process Recorders and Controllers

    Foxboro electronic flow recording and controlling instruments
were used in the pilot plant.  The instruments were mounted in
shelf units installed in the control board as shown in Figure D-
m.  The shelf units contained wiring terminal boards to which
the instrument and the field-mounted flowmeters were connected.
The electronic instruments were reliable and required no
maintenance over a 3-yr period.  The data recorded on the strip
chart were easily read and provided quick access to past
operating conditions.  The 12-point temperature recorders, also
from Foxboro, were satisfactory.  The only maintenance required
was an occasional cleaning of the slide-wires.  Each motor had
both a board-mounted and field-mounted control station.  No
problems were encountered with the Cutler-Hammer magnetic
starters used throughout the plant.  Ammeters were used on all
major motors as a check of the loading.

Gas Analyzers for the Pilot Plant

    A DuPont 460 analyzer was used to monitor S02 in the plant
gas streams.  The analyzer was equipped with an automatic zero
                               119
-------
Figure D-13.  Typical  Foxboro magnetic  flowmeter.
                    120
-------
                                       t  *
Figure D-14.  Pilot-plant control board
             121
-------
sequence, but the automatic sequence was bypassed and the
instrument zeroed manually.  Sample stations also were manually
selected.  The unit gave acceptable service.  The prime
maintenance area was keeping clean the sample lines and the light
path in the measuring tube clean.

    Stack opacity was measured in Ringlemann numbers with a
Photomation Smoke Monitor.  The instrument had a photocell to
measure light transmittance from a single source through the
plume.  Initially, the unit performance was acceptable.  Readout
agreed with observations of trained visual emission observers.
The lens faces required frequent cleaning.  During the later test
runs, the unit failed to zero and readings were taken by visual
observers only.
                               122
-------
            APPENDIX E

FUME FORMATION IN AMMONIA SCRUBBERS
                By

           Neal D. Moore

       Power Research Staff

            August 1975
  The Tennessee Valley Authority
  Environmental Research Section
          Office of Power
        524 Power Building
   Chattanooga, Tennessee  37401
               123
-------
              FUME FORMATION IN AMMONIA SCRUBBERS



                           ABSTRACT
     The published thermodynamic equations for the gas phase
reaction of sulfur dioxide, ammonia, and water are reviewed and
revised.  The formation of a fume is predicted based upon the
revised equations and compared to actual fuming conditions.  The
ammonia salt most likely to be formed is identified as ammonium
sulfite monohydrate.
                               124
-------
                           CONTENTS
                                                          Page

Introduction 	   126
Background	   126
Analysis	   128
Application	   128
Conclusions	   130
References	   132

Appendices
     I.  Calculations for Comparing Thermodynamic
          Equations	   133
    II.  Analysis of Data Published by Hillary
          St. Clair	   135
   III.  Boundary for Heat of Formation of Ammonium
          Pyrosulfite	   136
    IV.  Calculations of Equilibrium Constants 	   139
     V.  Typical Scrubbing Stages at TVA's Colbert
          Pilot Plant	   143
                             125
-------
               FUME FORMATION IN AMMONIA SCRUBBERS

Introduction

    Ammonia scrubbing has been employed for many years to remove
sulfur dioxide from waste gases.  With the recent enactment of
the Clean Air Act, several processes using ammonia scrubbing have
been proposed which produce a saleable byproduct from the ammonia
scrubbing process.  These products are elemental sulfur, sulfuric
acid, and ammonium sulfate.  Considerable pilot plant work on one
of these processes has been carried out by the Tennessee Valley
Authority. (1)

    Ammonia scrubbing investigations, such as TVA's pilot plant
investigations, have resulted in defining a problem of ammonia
salt formation (fuming)  as reported in Reference 1.  Several
other organizations have also reported this problem.  A recent
patent(2) issued to Air Products, Inc. deals specifically with
this situation and claims certain techniques for controlling the
formation of the fume.  In this patent there are thermodynamic
equations relating the concentrations of sulfur dioxide, ammonia,
and water, and temperature to equilibrium constants.  Previous
work by Hillary St. Clair, (3) reviewed by Jonathan Earhart(4)
also contains thermodynamic equations for the same ammonia salts
as Air Products, Inc. has investigated.

    An analysis and comparison of the available information was
conducted to gain an understanding of the phenomenon of fume
formation and to attempt to resolve differences in analyses which
have been published.

Background

    Basically a gas phase reaction will produce a solid whenever
the product of the partial pressures of the gases involved exceed
the equilibrium constant for the reaction.  The equilibrium
constant for a reaction can be related to the heat of reaction as
follows:

               d  (log, k)   =  AH                         (1)
                  d(TT        RT2"

 where  AH = heat  of reaction  (calories/gram/mol)
         R = gas constant = 1.987  (calories/gram/mol/°K)
         T = temperature  (  K)

      An alternate expression which  is  equivalent  is

           d (log!Ok)   =     -AH                          (2)
            d  (1/T)         (loge!0)R
                              126
-------
 Assuming that  AH  is  constant,  integration of equation  (2) gives

       logiok =      -AH       + A                         (3)
                 TR log  10
                       e

      The above  form is used in all the calculations  in  this
 paper.

     The heats of formation for (NH4)2S03(s),  NH4HS03(s) and
 (NH4)2S03-H20(s) are published in "Circular of National Bureau of
 Standard 500,  Selected Values  of  Chemical Thermodynamic
 Properties" issued February 1,  1952.   The heat of formation of
 (NH4)2S205(s)  was not published by the Bureau of Standards.   Air
 Products, Inc.  (2)  and St. Clair(3)  report two different values
 for the heat of reaction for  (NH4) 2S20S (s) +2NH3 (g)  + 2S02(g)  +
 H20 (g).  Earhart(4) did not change the heat of reaction reported
 by St. Clair.

     However, in his analysis Earhart did  revise some of St.
 Glair's equations.  These revisions were  based upon the heat of
 formation of ammonium bisulfite which apparently was not
 available when St. Clair did his  analysis in 1937.   Table 1 is a
 comparison of Air Products thermodynamic  equations and St.
 Glair's equations as changed by Earhart.   Appendix I contains the
 calculations necessary to arrive  at the values in Table 1.

                TABLE  1.  THERMODYNAMIC  EQUATIONS
       Reaction
                                                Source
St.Glair/EarhartAir  Products
(NH4) 2S2Os«-2NH3-»-2S02-i-H20       logkt:

NH4HS03JNH3+S02+H20            logkz

(NH4) 2S03. H20^2NH3+S02+2HZ0    Iogk3

(NH4) 2S03^2NH3+S02+H20         Iogk4

*T -  temperature  (°K),  k = atm5
-17050/T+U1.26

-9620/T+23.42

-16520/T+40.73

-13370/T+32.26
-16611/T+39.20

-9611/T+22.76

-16556/T+39.80

-13500/T+32.08
     An examination of Table 1 shows  some  differences and some
 marked similarities.  An investigation  and  analysis of the
                               127
-------
available data was conducted in an attempt to resolve the
differences.   If a resolution could be obtained, then a
comparison of the resolution to known conditions for fume
formation would be conducted.

Analysis

    The first question or difference to be reviewed was what is
the value to assume for heat of reaction for the following:

         (NH4)2S20S(S)   ->  2NH3(g) + 2S02(g) + H20   (g)

              Iog10k = 	-AH      + A
                       (log 10) RT

    St. Clair reported 78 kcal and Air Products, Inc. reported 76
kcal for this value.  A least squares analysis of St. Glair's
data performed by the author showed 76 kcal as the value
(Appendix II) .  However,  the value of 78 kcal could not be
rejected based on the analysis of Appendix II.  Therefore, a
ssrrond approach, namely determining a lower bound for the heat of
reaction, as shown in Appendix III rejected a AH less than 76.45
kcal.  Therefore it was assumed that a AH of 78 kcal was
appropriate and Air Products and the least squares analysis AH
values were not used.  The second step was to define the value of
the constant. A, since a difference exists in the published
material.  To accept St.  Glair's value for the constant would
lead to the same equations as Earhart obtained.  Air Products
value appeared questionable since the heat of reaction was in
error, yet Air Products value could not be rejected based on the
analysis contained in Appendix II.  So, an independent method was
used based on other data(s), namely the solubility diagram for
the system NH3-S02-S03-H20.  The details of the method and
results are contained in Appendix IV.  The results showed that
Air Products constant is valid.  Table 2 is a compilation of
Table 1 and the thermodynamic equations developed from Appendix
IV.

Application

    Consider the following reactions:

         NH4HS03 ^NH3 + S02 + H20

           log k2 = -9611/T +  31.24

          (NH4) 2S03 -H20 J2NH3+S02+2H20

           log k3 = -16556/T + 54.204
                              128
-------
rH
CM











































CM

H
rJ
§













































































































































M
O
r*
-M
rijj







(0
4J
O
^
T)
O

a-

M
-H
CU <
O
3
0 -P
CO M
id
r^
^1
id
U
x.
• r-l
HJ
rH
O

4-J
CO








































c
o
-H
4J
O
It)
cu
K




*
Ov
ro
^-
EH
0
in
o
r^
iH
1
O
CM
•
ej\
ro
.j.
EH
X.
rH
rH
VC
M>
rH
1


>£>
(N
•
rH
^
+
E-«
0
in
o
r^
rH
1
II

££

Cn
O
rH









*-,
en
o
33

+

	 .
Cn
s —
CM
O
CO
CM

+

^-^
Cn
— '
oo
33
Z
CM

t
^^
Ul
' —
in
o
CM
co
fM
—
X
£
rH

\fj
t—
•
CM
CM
J
— ^
O
CM
\D
O*>
1

V0
[•^
•
CM
CM
Jf
EH
X
iH
rH
VD
g^
i



CM
^T

ro
CM
-\-
EH
O
CM
vo
c^
1
II

M

en
O
rH












"en
"• —
o
CM
a:

4-

^^
Cn
•*~*
CM
O
W

+

^-^
Cn
^-^
ro
33
Z

t
	 .
cn
- —
CO
o
CO
33
a:
z
CM
O
rH
•
CTi
ro
^.
t:
o
CM
m
VO
rH
1
O
00
•
c^
ro
^-
EH

V£>
in
m
t^
rH
1


ro
r-~
•
o
^*
jf.
EH
O
CM
1/1
V0
rH
1
II

X

CP
o
•H






	 ,
Cn
*--•
o
TM
CM
+

. 	 .
5

CM
O
co

IT

	 .
Cn
— '
00
33
Z
CM

t

,-—
cn
o
CM
33

CO
o
CO
CM
-~
X
Z
ro
CM
l£>

O
ro
-^
EH
CTl
U3
ro
ro
rH
1
GO
O
•
CM
CO
.f-
EH

o
0
tn
CO
rH
1


>£)
CM
•
CM
CO
+
EH
O
r-
ro
ro
rH
1
II

£4

en
o
rH











5
o
CM
33

+

	 .
-?

CM
O
CO

4-

^^
Cn
^^
CO
X
z
CM


	 .
Ul
^->
M
o
CO
CM
—
33
Z
rr


































CO
EH
§
EH
W
z
o
L)
£
D
H
P-*
ffi
H
J
H
D
a
w
z
o
H
EH
H
Q
Q





































           O
           •H

           O
           id
           cu
                 EH

                 O
                 CM
                  I
                  Cn
                  O
                 3
                 O
                   CM
                 33
       GO
       VC
       ro
       I
81
                  cn      ro
                 "-^     O
                   in   to
                 O     " CM
                   CM   ^

                   CM   33
                 ^~*     J3
                   •sr   '-^
 ro
o
CO
33

33
Z
CM
                        Ul
                              00

                              E-i
      ro
       I
              Di
              0
        CM
       O
       CO
                   V£>


                   EH
                                     o
                                     CM
             en
             O
                              O
                               CM
                              O
                              CO
                               IM
             33
             Z
                   S

                     CM
                   O
                   CO
             o
             CO
              •CM
       O
        CM

 in
o      «o
 fM    O
cn     co
 CM     CM
                              X
                              z
             in

              co
            o
            W


            X
            Z
            CM
                                    •a
                             Ul
                             cu
 a
 Ul
 o

B
 id

 c
•H

 cu
 id
 in
                                                     

 CU  CU
 cn «
 id
.a  w
    cu
 cu  cu
43  M
JJ  Cn
    CU
 O TJ
4J -~-

 CU  CU
 )H  M
 Id  3
   -U
 ui  id

£,  cu
•P  a

 M  CU
 id -M
 cn
 O  II

*  EH
129
-------
    These are two of the reactions with the associated
equilibrium constants which are published in Air Products patent.
The question now arises as to whether one could operate a
scrubber as outlined in Air Products patent and still fume.

    Assume that we have isothermal operation (125°F), a
prescrubber to humidify the gas and remove HC1 and a three stage
scrubber with the compositions on each stage as shown in Figure
1.

         log k2 (324.82°K) = 1.6513    k2 = 44.80 atm5

         log k3 (324.82°K) = 3.234     k3 = 1714.87 atm5

    From this typical example, one can observe that at no point
in the scrubber is the product of Pn20 •  p NHa •  Ps02 i° violation
of Air Products' equation for ammonium bisulfite; however, using
the ammonium sulfite monohydrate equations, One would predict a
fume.  It is interesting to note that at approximately 125°F, the
patent does not contain any experimental work above approximately
73 mm H20.  Therefore, although one may concede that the fume
formed in Air Products work was indeed the bisulfite,
extrapolation of the work to actual scrubber conditions, as shown
in Figure 1 would indicate that the patent is not operable.  The
equations developed in this paper are even more critical than Air
Products with respect to the formation of ammonium sulfite
monhydrate.  Appendix V shows the typical scrubber conditions
employed at TVA1s Colbert pilot plant, the point of fume
formation and the recommended method for avoiding a fume using
the equations developed here.  Reference 1 confirms that the fume
is formed under conditions very similar to those depicted in
Figure 1.

Conclusions

    For the scrubbing conditions employed at TVAfs Colbert pilot
plant, a fume can form under isothermal operations, and with a
prewash section.  The fume is most likely ammonium sulfite
monohydrate and not ammonium bisulfite.

    Operation of a scrubber as proposed in Air Products patent
may not prevent the formation of a fume.

    Further pilot plant testing is needed to define regions of
fume-free operation.
                              130
-------
     Figure I. Comparison of Ka and KS values for liquors in
             three-stage absorber operation.
 K2(stage 3) = .295
 K3(stoge 3) = 2.17
K2(stage 2)=43.08
K3(stage 2) = I830.46
K2(stage l)= 15.374

K3(stage I) =127.07
K2
   •IS]
  CA=5
S/CA=.7
                          PS02=0.04(.23)*
                                = 12
S/CA = 67
                               r
                        Pc  = 1.00(1.00)*
 CA = '5

S/CA= 8
                             )=(l.824)*mm
                                              P  =.08 mm
                                               NH3
                                                  = 83.89
                                              PNH3 s •'06 mm


                                              PH20=78-74
                     [PH20_
           The number in parenthesis is the
            typical concentration obtained in
            TVA's pilot plant.
                               131
-------
                           REFERENCES

1.  Hollinden, Gerald A., Moore, Neal D., Williamson, P. C.
    "Removal of Sulfur Dioxide from Stock Gases by Scrubbing with
    Ammoniacal Solutions: Pilot Scale Studies at TVA."
    Proceedings: Flue Gas Desulfurization Symposium  (1973) pp.
    961-96, Environmental Protection Technology Series, EPA -
    650/2-73-038 (December 1973) .

2.  Spector, Marshall L.  and Brian, P. L. Thibaut, "Removal of
    Sulfur Oxides from Stack Gas." U.S. Patent 3,843,789  (October
    22, 1974)  - assigned to Air Products and Chemicals, Inc.,
    Wayne, PA.

3.  U.S. Bureau of Mines; Report of Investigations #3339; May
    1937, pp.  19-29 "Vapor Pressure and Thermodynamic Properties
    of Ammonium Sulphites" Hillary W. St. Clair.

4.  Earhart, J. P. (National Air Pollution Control
    Administration, U.S.  Department of Health, Education, and
    Welfare, Cincinnati,  Ohio).  Private communication to C.C.
    Shale of the Morgantown Coal Research Center, May 5, 1969,
    enclosure entitled "Discussion of Gaseous Ammonia for Flue
    Gas Desulfurization," 19 pp.; copy of enclosure received by
    N. D. Moore, January 30, 1973.

5.  Tennessee Valley Authority Sulfur Oxide Removal From Power
    Plant Stack Gases; Ammonia Scrubbing Conceptual Design and
    Cost Study Series, Study No. 3 Prepared for National Air
    Pollution Control Administration  (U.S. Department of Health,
    Education, and Welfare) 1970.

6.  Johnstone, H.F.  "Recovery of S02 from Waste Gases:
    Equilibrium Partial Vapor Pressures Over Solutions of the
    Ammonia - Sulfur Dioxide-Water System." Ind. Eng. Chem 27
    (5) , 587-593  (May 1935)

7.  Ishikawa, F. and Murouka, T.; "The solubility and Transition
    Point of Ammonium Sulfite." Scientific Report, Tohoka
    Imperial University, Vol. 22  (1933), pp. 201-19.

8.  Divers, E. and Ogawa, M.  "Products of Heating, Ammonium
    Sulfites, Thiosulfate and Trithionate;" Transactions, Journal
    Chemical Society, Vol. 77  (1900), p  340.
                               132
-------
                           APPENDIX I

       CALCULATIONS FOR COMPARING THERMODYNAMIC EQUATIONS



    Air Products equations (2) are as follows:

         1/2 (NH4) 2S205 J NH3 + S02 + 1/2H20

         Iog10kt = (-14,950/T) + 26.8

         NH4HS03 J NH3 + H20 + S02

         Iog10k2 = (-17,300/T) + 31.4

         1/2 (NH4) 2S03 -H20 i NH3 + H20 + 1/2S02

         log! Ok3 = (-14,900/T) + 27.1

         (NH4)2S03 Z 2NH3 * S02 + H20

              *4 = (-24,300/T) + 43.6
The temperature T is in degrees Pankine and the partial  pressures
are in millimeters of mercury.  To convert to degrees k  and
atmospheres the following equations were used.

         T(°K) = 5/9 T  (°R)

                    mm = 1 atmosphere) = 2.8808
Also the equations for ammonium  pyrosulfite  and ammonium  sulfite
monohydrate are multiplied by two in order to place all equations
on a mol  basis for the salt.  Applying these transformations  to
Air Products equations, the following are obtained.

         Iog10ki = -16611/T + 39.196

         Iog10k2 = -9611/T * 22.758

         Iogi0k3 = -16555/T + 39.796

         Iog10k4 = -13500/T + 32.077

Earharts1 corrections (4) to St.  Clair's data led to the following
equations.


                               133
-------
               S205 + 2NH3  +  2S02  + H20                        (1)

         log kt = -17050/T + 41.255

         2NH4HS03 ->(NH4)2S205 +  H20                           (a)

         log ka = -2190/T  +  5.578

         (NH4)2S205 + (NH4)2S03 +  S02                          (b)

         log kb = -3680/T  +  8.997

         (NH4)2S03  H20  -V(NH4)2S03 + H20                       (c)

         log kc = -3150/T  +  8.476

Tr in this case, is in  degrees Kelvin and partial pressures  are
in atmospheres.  Combining equations  (1)  and  (a),  (1) and  (b) ,
(c)  and  (1) and  (b) the following was obtained.

         NH*HS03 ?  NH3  + S02 + H20

         log k2 = -9620/T  +  23.4165

         (NH4)2S03 -H20  t 2NH3 *  2H20 * S02

         log k3 = -16520/T -•• 40.734

         (NH4) 2S03  * 2NH3  •»•  S02  + H20

         log k4 = -13370/T + 32.258
                                134
-------
                            APPENDIX II

       ANALYSIS OF DATA PUBLISHED BY HILLARY ST.  CLAIR(3)
Temperature
°C
60
70
80
90
100
110

OK
333.15
343.15
353.15
363.15
373.15
383.15
logt
W
grams/liter
0.044
0.102
0.158
0.295
0.534
0.946
nP = -AH
P
atmospheres
0.030
0.070
0.108
0.204
0.365
0.646
+
Y = log10P

-1.5229
-1.1549
- .9666
- .6904
- .4377
- .1898
C
X = 1000
T(°k)
3.00165
2.91418
2.83166
2.75368
2.67989
2.60994

                     T(1.987) (Ioge10) (5)

         loglok = 5  log10P  + log (16/3125)

         Ex2 = 47.097   Ey* =  5.29144   Ex  =  16.791
   W
            6
            16.791
        16
        47
                       .79l]    TT1   -1
                       .097]    he xl
  -4.9622
 -14,2431
47.097
-16.791
                                           Ey =  -4.9622
-16.791
   6
                                                0.644319
                                     B =
              log10P =  -3318,7/T +  8,46

              log10K =  -16593.5/T + 40.01
         T~I
        x y]
Source
Sum of
Squares
Mean (b^   4.1039

x (bt)       1.1830

Residual     .0045

Total       5.2914

         Var(b0)  =
                      Analysis of Variance
                        Degrees  of
                         Freedom




el
1
A V
tf
1
1
4
6
Ex 2
nExa - (Ex) 2

n
                                         Mean
                                        Square
                                        1.1830

                                         .0011
       F-Ratio
                                         1075
                                                 .3289


                                                 0419
                               -  (IX) 2
                              135
-------
                          APPENDIX III

    BOUNDARY  FOR HEAT OF FORMATION OF AMMONIUM PYROSULFITE


Consider the  following reactions

    (NH4)2S20S(S)  + 2NH3(g)  + 2 S02 (g) + H20 (g)             (1)

    (NH4)2S205(S)  + (NH4) 2S03 (S) + S02 (g)                   (2)

    AHj (kcal)  =  2 (-11. 04)  + 2 (-70. 96) - 57.80 -  AH
                                                   (NH4)2S20S

    AH2 (kcal)  =  -212.0  - 70.96 - AH
Assuming that the  heats  of formation are constant,  then

         loglokt = _ -AH.  (1000)    +  a
                    T(loge10) (1.987)

         logt Ok2 =    -AH? (1000) _ +  b
                    T(loge10) (1.987)

         AHt =-221.8  - AH

         AH2 =  -282.96 -AH
                                                    (3)
K'  ' [PNH3]
                      [
                                PH2o]
(5)
Substituting the  above in equation 5, we obtain  an  expression for
the vapor pressure  of  S02 due to  (NH4)2S20S decomposition
kt = 1/2
    A
log PSQ
    A  2
log Pso
                    A
                     =  1/5
og kt + 1/5 log 2
-AH, -1000 * a
T-loge10- 1.987
(6)
+ 1/5 log 2 (7)
                               136
-------
Similarly the following expression represents the vapor pressure
of S02 due to reaction (2)

         log PSO?   = -AH2 -1000	   , ,                      (8)
                      T-loge 10-1.987    D


Now at some temperature, say T, there will be an equilibrium  such
that log Pso2 = log PSO2-  st- Clair in reference 1 cites 120°C
as very nearly this temperature.  For the moment let's not specify
the exact temperature other than to agree that one exists.  What
we wish to do is to establish a AH for  (NH02S205 such that for some
temperature T greater than T,  (NH i+) 2S03 (S) will be the stable
compound and for a temperature less than ^  (NH 0 2S2C>5 (S) will be
the stable compound.

     Combining equations 3, 4, 7, 8 we obtain
         A                   r~     ~i
     log Pso2 -log Pso2 =  a*|   T  I + I + 1/5 l09 2           (9)

     + *[^]- b

         A                   r.H      AH I
     log Pso2  ~ 1°9 PSO2 = ap^2-  -  *&?} + a/5 + 1/5 Iog2 - b


     = a["238'6 ~ -8AH] + a/5 + 1/5 Iog2 - b

     a/5 + 1/5 Iog2 - b =  a P+.8 AH + 238.6~| si
                                         since

                                T
         A                           A
     log Pso2 ~ Io9 Pso2 = ° for T = T
Now
         log Pso2 ~ log Pso2 = ~ a/T [238.6 +  .8AH]
     + a/T [238.6 +  .8AH]

             A       A
     For T < T  log  Pso2 > 1°9 PSO2' therefore
       (238.6 +  .8AH)       (238.6 +  .8AH)
             T         >          A
                                 T

      (238.6 + .8AH)       (238.6 +  .8AH)
            T                   A
                                T
*a =       1000
      (Ioge10) (1.987)


                                 137
-------
                                    A
For the inequality to hold with T