DEVELOPMENT OF IMPROVED ALKALIZED ALUMINA FOR SO2 CONTROL
W. S. Briggs, et al
W. R. Grace and Company
Baltimore, Maryland
September 1969
Distributed
'to foster, serve and promote the
nation's economic development
and technological advancement.'
NATIONAL TECHNICAL INFORMATION SERVICE
-------
W. K. ORAC« A CO.
pAVI«ON CH.IMICAL DIVISION
BALTIMORE . MO.
RBPLV TO:
WASHINGTON RC8CARCH CENTKM
CUARKBVILLB. MARYLAND 91O3O
PHON»:(3O1) B31-BT11
PB 199 428
DEVELOPMENT OF IMPROVED ALKALIZED
ALUMINA FOR S02 CONTROL
FINAL REPORT
BY
W. S. BRIGGS, E. EICHHORN, AND P. K. MAHER
COVERING THE PERIOD
JUNE 15, 1967 THROUGH SEPTEMBER 15, 1969
CONTRACT NO. PH-86-67-129 WITH
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
' PUBLIC HEALTH SERVICE
DEPARTMENT OF HEALTH, EDUCATION & WELFARE
/50 / (This 2U Wo. of Page.
Report) gfi
UW~ 1 M5slpffiP
20. Secartty Claai (Thi* 22. Price
"^(CLASSIFIED
••DC 4O«1»-PT1
NATIONAI*T|CHNICAL
SEPTEMBER 1969 INFORMATION SERVICE
-------
DISCLAIMER
This report van furnished to the air Foliation
Control Office by
W. R. Grace & Co.
Davlsion Chemical Division
Baltimore, Maryland
Reply to:
Washington Research Center
Clarksvllle, Maryland 21029
in fulfillment of Contract FB-86-67-129'
TABLE OF CONTENTS
PREFACE
I. Objectives
II. Summary of Results and Conclusions
Section A. Initial Investigations
I-A. Introduction
II-A. Details
.Aj Methods of Synthesis
Page
1
1
1
5
6
6
6
_d
Modification of Bureau of Mines Procedure 6
Reactions of NaA102 with NaHCO. 7
Reaction of NaA102 with O>2 7
(Table I) Evaluation of Alkalized Alumina
from Reaction of NaAlO- & CO-
Forming Methods
1. Day Company Pony Mixer
2. Pan Granulator
Pellet Mill & Continuous Pilling
Extruder
Con-
4. Auger Type Extruder
(Table II) Attrition Resistance vs.
ditioning Time, % Loss
Sorbent Production Costs
1. Alternate Raw Materials
(a) Bienstock Process
(b) Sodium Aluminate-Sodium
Bicarbonate Process
(c) Sodium Aluminate-Carbon Dioxide
Process
2. Extrapolated Commercial Plant Costs
(Figure 1) Process Flow Diagram
(Figure 2) Alkalized Alumina Plant Lay-
out
8
9
9
9
9
9
10
10
10
10
n
11
12
13
-------
E!
Page
(Table III) Estimated Capital Cost
50T/S.D. 15
(Table IV) Alkalized Alumina Order of
Magnitude Price 45 Ton/
Calendar Day Plant 16
Sorption & Regeneiation Studies 12
1. Effect of Additives on Sorption and
Regeneration 17
(Figure 3) Schematic of Sorption Unit 18
(Figure 4) Schematic of Regeneration
Unit 19
(Table V) Evaluation of Grace #1 after
Activation at 1100 F 21
(Table VI) Evaluation of Grace #1 vs.
Spray Dried Alkalized
Alumina 22
(Table VII) Evaluation of the Effect of
Additives on Sorption and
Regeneration 24
2. Study of Regeneration Parameter 23
(Figure 5) Effect of Promoters on
Hydrogen Regeneration 25
(Table VIII) Summary of Regeneration
Study 26
Endurance Testing of Formed Product 27
1. Cyclic Aging Test Development 28
(a) Introduction 28
(b) Results 28
(Table IX) Steam & Thermal Aging
Data 29
(c) Discussion 28
2. Multicycle Test 30
(Table X) Standard Sorption-Regeneration
Cycle 32
(Figure 6) Extended Aging - Grace #1 33
Page
Chemistry of the Alkalized Alumina Process 31
1. Activation Atmosphere 31
2. S02 Concentration & Space Velocity 34
3. Oxygen 34
4. Temparature 34
(Table XI) Sorption Regeneration Charac-
teristics of Grace #1 Alka-
lized Alumina Powder 35
(Table XII) Material Characteristics
Before & After High Tem-
perature Operation 35
5. Effect of C02 on Sorption 34
(Table XIII) Effects of CO, on Sorption 37
Effect of NO on Sorption
(Table XIV) £lue Gas Composition vs.
Attrition Resistance
36
38
39
39
7. Effect of Poisons
8. Effect of C02 on Regeneration
(Table XV) Effect of CO. on Regeneration40
.Gj Portable Aging Units 39
Section B. Screening New Sorbents for Improved Attrition
Resistance 41
I-B. Introduction 42
(Figure 1) % Loss vs. Crystallite Size for
Grace #2 Alkalized Alumina in
Laboratory Tests 43
II-B. Details 45
_A| Methods of Sample Preparation 45
1. Binder Incorporation & Beading Technlque45
2. Impregnation 45
J3| Evaluation for Sorption Efficiency and
Attrition Resistance
1. Single Cycle Screening Test
46
47
li
-------
2. Multicycle Aging Test
(Table I) Comparison of Results from
Laboratory & Field Units
Page
47
48
47
47
49
49
III-B.
3. Process Variables
(a) Sorption Temperature
(b) Regeneration
(c) Activation
(Table II) Effects of Sorption Temp.
on Attrition Resistance
of Grace #2 50
(Table III) Effects of Activation
Condition on Attrition
Prior to Sorption 51
(Table IV) Evaluation of Kaolin
Binders and Effect of Heat
Treatment on Attrition 52
Conclusions and Recommendations
Section C.
I-C.
II-C.
(Table V) Evaluation of Alkalized Alumina
Binders
(Table VI) Evaluation of Carriers Impregnated
with Sodium Alumlnate
(Table VII) Composition and Process Recom-
mended for Further Development
Laboratory and Pilot Plant Production of
Attrition Resistant Sorbents
Introduction
Details
jj Laboratory Preparations
(Table I) Summary of Results on Bound and
Formed Alkalized Alumina
1. Precipitation of Dawsonite In the
Presence of Binders
53
54
55
56
57
58
58
58
59
58
60
Meta-Kaolin Preparation
(Table II) Physical Evaluation of Insitu
Binder Preparation 62
Page
(Table III) Physical Evaluation of Meta-
kaolin Binder Dry Blended 63
(Tabel IV) Sorption and Attrition Eval-
uation of Various Samples 64
_fij Pilot Plant Preparations 61
III-C. Conclusions and Recommendations 65
(Table V) Evaluation of Beads Made Using the
Marumerizer 66
APPENDIX - A. Modified Air Jet Attrition Test 69
B. Sophisticated 5 Module Field Test Unit 75
C. Organization Chart 79
-------
PREFACE
The work reported here covers Che total period from June 15,
1967 through September 15, 1969 on Contract No. PH-86-67-129.
Two comprehensive reports have already been issued (May 15, 1968
and July 15, 1969) which detailed all of the work done during those
periods. It is the object of the present report to summarize in
Section A the salient features of the work reported in full earlier
and to detail in Sections B and C the work done from May 15, 1969
through September 15, 1969.
I. Objectives
The use of activated alkalized alumina for the removal of
sulfur dioxide from flue gases had been under investigation
at the Bureau of Mines, Bruceton, Pennsylvania, station both
in the laboratory and pilot plant for some time. While the
initial technical feasibility of the process appeared promising,
a number of questions regarding the performance characteristics,
method and cost of manufacture of alkalized alumina remained.
Specifically, information was required on the following topics,
which form the objectives of the program.
• Maximization of sorption rate and capacity
• Maximization of regeneration rate
• Minimization of regeneration cost with regard to
temperature and reducing gas used
• Improving of attrition resistance of the sorbent
• Development of a low cost manufacturing procedure
II. Summary of Results and Conclusions
The discovery of the adverse effect of NO on attrition resistance
materially increased the difficulty of producing a suitable sorbent.
Despite this we believe the objectives of our study have been met.
A ceramic binder system was developed which gave a sorbent with
less than 5% air jet attrition after a 46 cycle field test using
power plant stack gas. Under these same conditions the best sorbent
-------
without binder is virtually destroyed. Although the drying and
heat treating steps were done in the laboratory, the forming
and conditioning procedures used to produce this improved sorbent
were carried out using pilot plant equipment.
In the absence of definitive plant performance, the practical
utility of our improved sorbent remains unknown.
The main results of the work done are tabulated below:
1. Many of the problems associated with the process
and the equipment required for the commercial
preparation of alkalized alumina by various methods
were resolved at the pilot plant level. The most
practical commercial equipment was determined for
filtration, washing, drying and forming of the
alkalized alumina.
2. New products, showing significant improvements
in SC>2 sorption rates, regeneration rates and
attrition resistance, resulted from the pilot
plant and laboratory studies.
3. Laboratory and pilot plant studies showed that a
high quality dawsonite can be prepared by the
reaction of CO- and sodium aluminate.
4. The raw material costs were defined more precisely
for preparing alkalized alumina by several processes.
The raw materials for the CO, - sodium aluminate
process were the cheapest, $6.041/lb. of dawsonite
product. The cost of the raw materials used in the
Bureau of Mines process was $0.078/lb. of product.
5. A detailed preliminary estimate of the capital and
operating costs of a 50 T/Stream Day plant for the
C0» - aluminate process based on laboratory and pilot
plant data indicated a plant cost of $4,800,000 and
a selling price of approximately $0.20/lb.
6. Standard alkalized alumina was characterized using
up-to-date analytical techniques. These Included
the surface properties, porous structure, x-ray
analysis, chemical analysis, and rates of sorption
- 2 -
of S02 and regeneration. Dawsonite, the activated,
spent and regenerated materials were characterized
by these procedures.
7. Sulfur dioxide sorption rate studies on formed
bead products showed clearly that initial rates
increase with increase in surface area and pore
volume cf the sorbent.
8. The best promoter found for increasing the sulfur
dioxide sorption rate was copper at a concentration
of about 1%.
9. Extensive studies were completed on the effect of
promoters on the regeneration rate of spent alkalized
alumina. The data showed the small concentrations
(0.5 to 1.0%) of a hydrogen activating metal such as
iron, or nickel, or a metal oxide such as vanadium
oxide were required to obtain rapid regeneration
rates.
10. Regeneration rates using ethane or propane were found
to be superior to those using hydrogen, but possible
long term effects due to coke buildup were not evaluated.
11. Laboratory studies showed that low partial pressures
of steam should be maintained in the regeneration at
all times. Unit design, for removal of S02> using
alkalized alumina should minimize hydrothermal aging
of the sorbent during the regeneration step.
12. A reproducible laboratory cyclic aging test was
developed based on the data obtained from the hydro-
thermal treatment of alkalized alumina.
13. Investigations of various factors on the S0_ sorption
rate using real stack gas showed that the presence
of nitric oxide in the gss h;^ the r,.c3t 3ifell;fl^ai^
effect. Nitric oxide increased the rate of SO,, sorption
but produced poor attrition resistance.
14. Arsenic compounds in the stack gas and carbon monoxide
in the regenerating gas adversely affected the regene-
ration rate. An increase in regeneration rate was
observed when carbon dioxide was present in the regene-
rating gas.
-------
15. Two portable field testing units were constructed and
used for sorbent evaluation at the Dickerson station
of the Potomac Electric Power Company and at the Crane
station of the Baltimore Gas and Electric Company.
16. A comprehensive study of the attrition problem resulted
in the discovery of several binder systems which showed
much improved attrition resistance. The following
systems were the best found.
• 23% Kaolin, 2% Na silicate, 75% Dawsonite
• 25% Meta-kaolin, 75% Dawsonite
17. A larger, more sophisticated, thin bed sorbent test
unit was constructed for future use in testing new
sorbents or modified processes.
SECTION A
INITIAL INVESTIGATIONS
- 4 -
-------
I-A. Introduction
Prior to the assumption of the present contractual obligation,
a 6000 Ib. quantity of alkalized alumina was prepared by
W. R. Grace for the U. S. Bureau of Mines. This product was
made according to their specifications, using a mole ratio
of Na.CO, to AL,(SO )- of 3.6, in which A1,(S04), solution is
added to the Na'CO^sBlution forming tha precipitate of alka-
lized alumina. This material when pan granulated and tray
dried yielded the product called Grace #1. This was the
initial standard against which improvements were measured.
The object of this part of the investigation was to improve
on the method of producing alkalized alumina from the point
of view of:
• Cost
• Physical Properties (Attrition Resistance)
• Sorption Characteristics
In order to satisfy these objectives the studies were directed
toward improved methods of dawsonite synthesis and develop-
ment of more effective forming techniques.
II-A. Details
_Aj Methods of Synthesis
1. Modifications of Bureau of Mines Procedure
Spray drying of the washed product was used as
opposed to the tray drying procedure for Grace #1.
This resulted in a product having increased pore
volume and surface area. Modification of the forming
technique (extrusion followed by rolling) produced
a product of closely controlled particle size (10 -
14 U. S. mesh), and improved attrition resistance
based on the test procedures used at the time. This
process (with iron oxide additive) was used to pre-
pare the 6000 Ibs. of product sent to the Bureau of
Mines as Grace #2. Grace #2 also served as the refer-
ence point for attrition evaluations.
- 6 -
The alkalized alumina made by the reaction NaAlO
and CO, was as good as the standard alkalized
alumina with regard to Sorption and regeneration.
Although the Bureau of Mines' method of synthesis was the
best defined method from the standpoint of process econom-
ics and product purity, the carbon dioxide method utilizing
a high concentration of sodium aluminate solution in a con-
tinuous system was considered the most promising method of
synthesis. Therefore, the process economics and plant
design were prepared utilizing the carbon dioxide method
of dawsonite preparation.
_B| Forming Methods
Considerable effort was devoted to developing forming or
beading methods. Particle shape and physical characteristics
of alkalized alumina product are critical in the S0_ dispersed-
phase sorber as designed by the Bureau of Mines. Spherical
particles would be the ideal shape because of potentially
better attrition resistance.
The following types of forming equipment were investigated:
1. Day Company Pony Mixer
Dawsonite was added to the mixer and water was added
during mixing. Particles of the proper size could be,
formed but did not have the desired spherical shape.
2. Pan Granulator
Moistened dawsonite was added to a pan granulator
at a variety of moisture levels. Extensive tests
failed to produce a product having the desired
spherical shape and attrition resistance.
3. Pellet Mill and Continuous Pilling Extruder
Both types of pilling equipment used a dry feed.
The resulting pills had to be remoistened throughout
in order to be subsequently beaded by rolling.
4. Auger Type Extruder
This was the most satisfactory equipment tested. The
-------
resulting spaghetti like extrudates could be cut or fed
directly into a rotary drum where the extrudates broke
into lengths about equal to their diameter.
With regard to the auger type extruder, it was possible
to obtain bead size in the range of 8 to 20 U. S. mesh.
Other factors affecting the rounding or beading of the
product were the bed depth in the rotary forming- .drum.
and the total moisture content of the extrudates.
It was initially concluded that the best forming method
was to use an auger type extruder and then shape the
extrudates into spheres in a rotary conditioning drum.
The rotary drum treatment improved the attrition resis-
tance of the beaded product as shown in Table II.
TABLE II
ATTRITION RESISTANCE VS. CONDITIONING TIME. % LOSS
15 Min. 1 Hr. 2 Hrs. 3 Hrs. 4 Hrs. 6 Hrs.
13.0
7.4
4.0
3.3
2.3
Under 1
When the critical role of NO in sorbent attrition re-
sponse was recognized, a ceramic type binder system
was developed. However, difficulty was experienced in
reproducibly preparing attrition resistant beads of this
composition by the described technique. Since the
problem appeared in the conditioning step, other con-
ditioning equipment was employed, the most successful
of which was a Marumerizer. Use of this equipment,
details of which appear in Section C, "Laboratory and
Pilot Plant Preparation of Improved Sorbents", con-
sistently yielded beads of good attrition resistance.
C| Sorbent Production Costs
1. Alternate Raw Materials
(a) Bienstock Process
The Bienstock process can be represented by
the following equation based on a fixed ratio
of 2Al(OH)3:3.6Na2C03:
- 10 -
2NaAl(a>3)-
2Al(OH)3+3H2S04+3.6Na2C03
(OH)
Calculating the input per pound of product
and using Baltimore area costs, the raw material
costs on a 100% yield basis are $.0777/# as
shown below:
0.542# C-31 alumina hydrate @ .0404-$.0219
1.02# 100% H2SO, @ .0167-$.0170
1.33# Na2C03 (reaction) @ .0175-$.0233
0.883# Na2C03 (washing)
@ .0175=$. 0155
Per Pound of Product $.0777
(b) Sodium Aluminate-Sodium Bicarbonate Process
The sodium aluminate-sodium bicarbonate process,
using the optimum conditions, can be represented
by the following equation:
2.6NaOH+2Al(OH)3+4.4NaHO>3- » 2NaAl
(C03)(OH)2+2.4Na2C03+0.2NaOH+4.4H20
By the same procedure as above.the raw material
costs are derived:
0.542* C-31 alumina hydrate @ .0404-$.0219
0.362* NaOH (100%) @ .0330-$.0119
1.36* NaHCO.
Per Pound of Product $.0719
(c) Sodium Aluminate-Carbon Dioxide Process
The sodium aluminate-C02 process was not fully
delineated; however, the following equation was
used for initial cost comparisons:
2.8N80H+2A1(OH)3+2.6COj y NaAlC03(OH)2+
0.4NaHC03+0.2Na2C03+2.2H20
-------
Using the foregoing procedure, the cost
(100% yield basis) is:
0.542* C-31 alumina hydrate @ .0404-$.0219
0.389# 100% NaOH @ .0330-$.0128
0.396# C02 . @ .0150-$.0059
Per Pound of Product $.0406
2. Extrapolated Commercial Plant Costs
Although the experimental program did not develop
sufficient information for a rigorous evaluation,
an estimate of costs was feasible. A plant design
was prepared and estimated based on the following
factors:
• Plant Size - 50 Tons/Stream Day with 90%
operating factor.
• Product is dawsonite - NaAlCO.(OH)- -
dried, not activated.
• Battery-limits plant - located within
confines of larger plant with utilities
close by and supporting services available.
• Baltimore, Maryland, area used for pricing.
• Extrapolation of laboratory and pilot
plant data to full scale.
A process flow sheet for the operation is given
in Figure 1, and a plant layout in Figure 2. The
capital cost for the plant is shown in Table III.
Cost of the product is given in Table IV.
p) Sorption and Regeneration Studies
This work was divided into two areas. The first was a
study of:
• The effect of additives in increasing the
effective sorption capacity (rate).
-------
n
Flash Dryer ory collectors
Table Filter
So. Dr. Feed
(2 Complete Units)
vacuum
System
Spray Dryer Reslurry
Conditioning
Water
Flash Dried Mat'l
Storage
Pulverized
Over-C'nder Recycle
6
'•' Bags ' ; Additive Make & Feed
Extructors (3)
*SJ-^-:.V^^-:"-^
Rotary Conditioner Rotary Dryer
Sizing Screen
, Hammer Mill
Fin. prod. Prod. Loading
-------
CO2 Evolution
Crystallizing
.per3-sod. Al. Stg.
Make
Gas - Liquid
Reactors •
Spray Dryer Reslurry
,
Utilities:
City Water
300 GPM
(^
\Li
Saved Water
System
't n A* li > — r"™
T2EZT
^
^ -X
1
"
Steam 10,000 #/H
Power 1.100 KVA
_
Heater Gas 7 , 500 CFH
Comp. Air 100 CFM
r$ft r ;
Water
Flash Dr^Led Mat'l
Storage
CO2 Recycle System
''-'', Bags ! '-Additive Make & Feed Extructors (3
-.••••-•w
Extruders (3
-------
Hydrate Weigh Hoppers-sod. Al. Stg
Sodium Aluminate Make
j Caustic Receivina & Storage City Water
( ' . Booster Pump
Saved Water
System
Utilities:
City Water
Steam
Power
Process
Heater Gas
Comp. Air
300 GPM
10.00O #/HR.
1.100 KVA
7 , 500 CFH
100 CPM
fl 1 I I i I
Watt
:
CO2 Receiving and Storage
CO2 Recycle System
j 50 T/D ALKALIZED ALUMINA
PROCESS FLOW DIAGRAM
FIG. 1 . {/
-------
180'
, 14
Y
-1
12
Key: 1. Administration
2. 50% Na OH Tanks
3. Aluminum Hydrate Silos
4. Boiler House
5. CO2 Tanks
6. Wet Process Area
7. Spray Dryer Area
8. Flash Dryer & Collectors Area
9. Extruder Area
10. Rotary Conditioning & Drying
11. product Silos & Loading
12. Roads
13. Electrical Substation
14. Fence - Battery Limits
ALKALIZED ALUMINA PLANT - LAYOUT
SCALE 1" -40'
FIG. 2
HPZ 2/20/63
TABLE III
ESTIMATED CAPITAL COST - 50 T/S.D.
ALKALIZED ALUMINA PLANT
Installed Equipment Costs
Vessels, etc. $ 298,500
Heat Exchangers, etc. 45,800
Pumps, Fans, Blowers, etc. 72,450
Screws, Elevators, etc. 71,900
Agitators 72,000
Filters 184,500
Mills, Extruders, etc. 134,000
Heaters, Spray Dryers, etc. 1,035,000
Electrical 124,600
Instrumentation 46,000
Total Installed Equipment Cost $ 2,084,750
(Purchased Equipment Cost $1,318,450)
Structural (10,500 sq. ft.) $ 65,000
Concrete Foundations 231 CY 23,000
Building (10,500 sq. ft.) - 250,000
Piping (45% Purchased Equipment Cost) 590,000
Utility Supplies 50,000
Railroad Trackage 15,000
Site Work 15.000
Total Direct Plant Cost $ 3,092,750
Total Cost $ 3,100,000
25% Contractor's Overhead & Profit $ 750.000
Subtotal $ 3,850,000
10% Direct Plant Cost for Engineering $ 310.000
Subtotal $ 4,160,000
15% Contingency $ 640.OOP
Job Total $ 4,800,000
Basis of Estimate
, 1. Process conditions and equipment extrapolated from bench
scale work.
2. Battery-Limits manufacturing building and operation, not
including shops, analytical labs, change houses, general
offices, etc.
3. Based on current prices of equipment and labor in
Baltimore, Md.
-------
2- Reaction of NaAlO^ with NaHCO.,
This reaction of a solution of NaAlO,, with a solution
of NaHCOo was studied in detail by varying the following
o Mole ratio of Na^OrAl-O- in sodium
alurainate solution. . •
e Mole ratio of NaAlCL : NaHCO.
e Temperature of reaction
© Concentration of reactants
Using this method, dawsonite was prepared free of de-
tectable crystalline impurities.
3 . Reaction of NaAlO with CO^
As a further step in cost reduction, attempts were made
to prepare dawsonite by the following reaction:
4-
NaAl(CO )(OH)2
Several preparations were studied by varying the mole
ratio of Na?0 to Al?0., in the solution. The process
which. gave che best "product by x-ray identification
was that using a Na^O to Al,,0,. mole ratio of 1:4.
A continuous flow procedure was used as opposed to
the batch procedures of sections 1 and 2. Carbon
dioxide was made to flow through a mixing nozzle
simultaneously with a solution of sodium alumina te
at such a rate that the pH of the effluent slurry
was maintained between 9 and 9.5.
This process was optimized with regard to C09 con-
sumption and produced dawsonite according to "che
following equation:
NaAl09+1.8C09+H90— ~>NaAl(CO ) (OH)9 H-0. 8C09 t
^— ^- j— O £~ £*
Although characterization of this product by x-ray
indicated a product free from crystalline impurities,
it was necessary to obtain scrpticn. data for evaluation
of its utility in sulfur dioxide removal.
-------
TABLE I - .
EVALUATION OF ALKALIZED ALUMINA FROM REACTION OF NaAlO and CO
(Powder Testing of Oven Dried Material)
Mole Ratio Na20:Al2Oo of
Aluminate Solution
Sample
1.40:1
B
Grace #1
Standard
1.12:1
S^ulfur Dioxide Sorption'1'
Wt. 70 Sorbed at Break
Capacity
. Wt. % Sorbed at
Saturation
20.2
37.0
18.9 17.4
30.0
39.0
Regerier a t ion*
Rate % per Minute
13.1
24.0
12.4
Chemical Analysis
% Iron as Fe0
0.30
0.55
0.50
*' SorptTion and regeneration test procedures for
powders are described in section D.I. Effect
of Additive on Sorption and Regeneration.
-------
TABLE IV
ALKALIZED ALUMINA
ORDER OF MAGNITUDE PRICE
45 TON/CALENDAR DAY PLANT*
Raw Materials
Aluminum Hydrate 0.68#/# @ .041
507 Liquid Caustic Soda 0.97#/# @ .0165
Carbon Dioxide 0.55#/#
-------
Nr
v 7
H2
i'y«_ ap»
vent
CAHN
electrobalonce
temperature
regulator
I
JL
^
k
r
V
-1
»—
tf
\
s
t
4 i
*
6
VlX
t
ulfur
trap
rfl
.^
S — T
g
-, 1
(A)
Vx
cooling
water
> — a c
) oi
'•<^^^.
'/>
^
pr^
'&s5's&s's
Figure 4: Schematic of Regeneration Unit
FIXED-BED SORPTION TEST UNIT
air
vent.
-------
This system consisted of a Cahn elecCrobalance fitted
with a thermostatically controlled furnace. It was
so designed that hydrogen gas was preheated prior to
flowing over the sample to be investigated. The
weight loss of the sample vs. time at 1300 F was auto-
matically recorded on a VOM Bausch & Lomb recorder.
TABLE V
EVALUATION OF GRACE #1 AFTER ACTIVATION AT 1100°F
Grace #1 was chosen as the initial standard.
shows this evaluation.
Table V
During the course of this work alkalized alumina was
prepared under plant conditions using the modified
Bureau of Mines procedure. After forming this product
was designated Grace #2. The most significant difference
found in this material was that the pore volume and
surface areas were considerably improved. This material
served as base for evaluation of all the additives
studied, since it showed good sorption characteristics
and contained essentially no Fe impurities.
This last was important since Fe was one of the addi-
tives to be studied. The most important single dif-
ference in the modified preparations was spray drying
as opposed to a tray drying step during its preparation.
Table VI compares these two products. The spray dried
product is designated as 4928-48A.
It should be noted that the increase in pore volume and
surface area resulted in an increase in break or ef-
fective capacity. The lowering of the % regeneration
rate is attributed to the low iron content.
Using this spray dried material, a series of impreg-
nations was made with various metal ions and in some
cases combinations of ions. Iron impregnations were
carried out using ferric-ammonium oxalate solutions.
Impregnations with Cu, Ni, Co and Cr were made with
the acetate salts. In the case of vanadium, both
sodium and ammonium vanadate were used.
Mercury Pore Vol. cc/g
N_ Pore Vol. (cc/g)
o
B.E.T. Surface Area m /g
Chemical Analysis:
% Na00 - 35.40
0.69
0.51
47.50
C0
% SO,
Fe
57.30
0.55
5.46
0.36
% SO- sorbed at break capacity
% Regeneration
17.70
11.80
- 20 -
-------
- zz -
The procedure for impregnation was to add solutions
of the various salts to the sample of dawsonite and
then evaporate the slurries to dryness. These products
were then activated under Hj at 1100°F for 5 hours and
the performance determined. A total of 8 materials was
studied. Table VII shows the evaluation results of the
most promising materials.
O
rt
o o
M VO
NO
o
O TO
O
a£w
Q> It!
.&*
S
,*
o a> s*
rr H
(D 01
§
H
W
These results show:
• Copper increases the break capacity,
probably due to its oxidative activity.
• Iron and vanadium greatly improve the
regeneration rate.
• Doubly promoted samples did not show a
combining of the effects .
• Pt produces a good break-capacity and
regeneration rate but is not economical .
Based in part on these results, a plant batch (6000 Ibs.)
was produced using the spray dried dawsonite and in-
corporating Fe in its preparation. This product after
forming was sent to the Bureau of Mines for their evalu-
ation. Our laboratory test of this sample, Grace #2,
in powder form gave the following results :
Regeneration Rate
Break-Capacity(Q_) -
D
Surface Area
Pore Volume
Fe Content
35.5 (%/min.)
24.0 %
74 m /g
0.51 cc/g
0.6 %
2. Study of Regeneration Parameters
A study was made of the rate of regeneration as a function
of reducing gas, temperature, and catalyst. Table VIII
shows these relationships, and Figure 5 shows the plot of
these relationships using Ik regeneration. These results
are summarized as follows:
-------
01
•2 8
e
01
BO
O
w
I 5
I §
*
O C
0) 01
>U DO
aad % '
30
TABLE VII
EVALUATION OF THE EFFECT OF ADDITIVES ON
SORPTION AND REGENERATION
Sample No.
Grace #1
4928-48A
4607-48AF
4607-41
5245-1
5119-173
5245-19
Additive
. Fe
--
Fe
Cu
Fe
Cu
V from
NH4V03
Pt
Additive
0.4
0.0
1.0
1.0
1.0
1.0
1.0
3.0
Break Capacity
(0 ) Wt. %
16.5
24.4
23.8
31.0
25.6
19.6
24.2
7. Regeneration
Rate (%/mln)
12.0
6.9
35.9
5.5
32.3
55.2
-------
TABLE VIII
SUMMARY OF REGENERATION STUDY
Sample
Promoters
Regeneration
Grace #1 5213-34
Gas Temp . (°F)
H,
CH4
CH4 with
H_0 vapor
S'atural Gas 4
CO
=2H6
:3H8
1100
1200
1300
1300
1300
1300
1300
1100
1200
1300
1100
1200
1300
0.3% Fe 1% Fe
5103-39
1% V
5213-5
1% V and
1% Cu
Rate of Regeneration (%/min)
N.M.* N.M.
2.6 6.0
13.2 24.6
7.0 12.1
4.7 12.2
13.7
40.0
56.3
14.5
17.0
13.7
37.0
41.0
11.8
11.7
-- — — 13.3
8.1
..
..
9.0
23.8
45.8
N.M.
19.7
26.3
27.7
N.M.
26.3
58.8
18.0
44.3
50.4
* N.M. means not measurable (too slow)
4 95% CH4 and 3%
- 26 -
• The regeneration temperature can be lowered
200°F (from 1300° to 1100°F) for hydrogen
regeneration of vanadium or copper-vanadium
promoted alkalized alumina.
• The doubly promoted (copper-vanadium) alkalized
alumina is the best for use in the case of
ethane or propane regeneration.
• Low cost natural gas was a workable regeneration
gas.
The work described in this section was done prior to
the discovery of the effect of NO on the sorption of
SO,. It will be shown in SectionxB that NO in the flue
gas increases the rate of sorption but in tfie process
significantly lowers the attrition resistance of the
beads.
JEj Endurance Testing of Formed Product
Useful sorbent life is a critical economic factor in the
alkalized alumina process. It was necessary, therefore,
to establish those parameters which affected the life of
the sorbent with regard to useful capacity and resistance
to attrition.
In order to accomplish this, the following objectives were
set up:
• Development of a laboratory multicycle test
to simulate plant scale aging behavior of
various sorbents under standard conditions.
• Determination of the physical and chemical
changes accompanying the aging process.
• Laboratory evaluation of the effect on sorbent
life of a wide range of process conditions, and
chemical variations in sorption or regeneration
gas composition.
• Design and construction of two portable units to
correlate data obtained in laboratory tests with
normal catalyst aging using flue gas from coal
burning furnaces.
-------
1. Cyclic Aging Test Development
(a) Introduction
A standardized aging test was required to evaluate
the life expectancy of activated alkalized alumina
materials. First, however, it was necessary to
determine the most important aging mechanism in
order to establish the range of operating conditions
in the equipment. Viewed against a background of
experience in heterogeneous catalysis, a thermal
or steam sintering mechanism under sorbent regener-
ation conditions seemed a strong possibility.
In alumina and silica based catalysts sintering
occurs at high temperatures and is accelerated
by steam and alkali. It seemed probable that
sodium aluminate, the active sorbent in activated
alkalized alumina, would show similar reductions
in surface area and growth in crystallite size under
regeneration conditions. Such changes would be
expected to reduce sorption efficiency. These
effects could also be accentuated to provide an
accelerated test by use of higher-than-normal
regeneration temperatures coupled with steam or
thermal pre-treatment.
(b) Results
Table IX shows the results of the steam and thermal
aging evaluations. These data demonstrated that
surface area is lowered by both parameters especially
at the higher temperature. In both cases the pore
volume was increased, as was the median pore diameter.
It was also noted that steaming produced substantial
crystallite growth.
(c) Discussion
The fact that artificial steam aging of the acti-
vated alkalized alumina produced extensive crystallite
growth was considered in light of the sorption and
regeneration process chemistry. In the initial acti-
vation step, dawsonite was converted to sodium alumi-
nate.
- 28 -
ro
vo
o
o
o
o
S 5s
O 3
a
a H
c
TO
CO
Hi
S
ff
(9
CO
O
P-
rt
TO
45£t»ij£!£ro£Kiii
oooooooooo
S££;sss£ss£
i—1
wtoiotoorooooooooo
LJ Ui Ov VO h-1
1 i 1 &\ O\ ( i -P* I— ' vO
op o o o
"iK.
^^
i" en
JJ-M>
i"""*
O *fl
O •
*T3
O 2
H (D
TO O.
j>« K*
O CO
H- 3
01
a
!-*• to X 2
N It 1 ttl
fB (b yo !>
• 1-" 01 I-1
>>i-i^ o
H- ro
rt
TO
T>
n
o
a
TO
rt
rt
re
CO
H H
S >
TO DO
S "
c- i-i
X
O
i
1
-------
NaAl(C03)(OH)2 »
The crystallite growth seen was that of sodium
aluminate. In the sorption process all or part
of the sodium aluminate was converted to sodium
sulfate and aluminum oxide.
2NaA102 + S02 + %02—» Na2S04 + A12°3
The degree of conversion depended on the sorbent
loading reached in the actual process cycle.
Steam aging effects for mixtures of sodium sulfate
and alumina were not determined but are presumed
to be small relative to those for sodium aluminate.
In any event, if regeneration was complete, these
components reverted to sodium aluminate and under-
went partial crystallite size reduction in so doing.
2NaA100 + H0S + 3H^O
The high temperature regeneration process also pro-
duced a severe environment in which steam sintering
might occur. At low hydrogen space velocities
(200 VHSV) water produced in the reaction was found
to give steam concentrations of several percent.
Predicting sintering rates in practice was diffi-
cult since at low sorbent loadings the steam con-
centration was lower but the quantity of unconverted
sodium aluminate subject to attack was higher. At
high loadings more steam was produced during regener-
ation, but more of the sodium aluminate was freshly
formed from sodium sulfate and alumina.
It was predicted and then confirmed experimentally
that high reduction space velocities would minimize
steam aging effects at low sorbent loadings. Gas
flows were established in the test procedures which
insured steam concentration below 1% by volume
would be maintained.
Multicycle Test
A multicycle test unit was designed to evaluate the
aging characteristics of the sorbent in the laboratory
under simulated plant conditions.
After considerable experimentation, sorption and regener-
ation cycles were established as shown in Table X. In
twelve four-hour cycles the surface area was reduced to
31 mvg. Sorption efficiency declined from 4.5 to 3.0
wt. % sulfur sorbed in 90 minutes and appeared more
dependent on declining pore volume than on surface area.
Using this procedure an extended multicycle aging run
was carried out in order to prepare an extensively aged
material. This run was made over 61 cycles. The de-
cline in efficiency is shown in Figure 6.
The drop in effective capacity was attributed to less
of pore volume. This could have resulted from two
factors.
• Plugging up of available pores by SO,
accumulation on aging.
• Sintering of the NaAlO, during the high
temperature regeneration (hydrothermal
effects).
There was a leveling off of this decline at 2.6 wt. %
sulfur pickup. This level of efficiency might be adequate
in plant operations.
_F| Chemistry of the Alkalized Alumina Process
Knowledge of the sorption and regeneration mechanism was
necessary to determine the effect of physical structure and
chemical composition on sorption efficiency. Several factors
which would influence regeneration rates were also examined.
The results are given below.
1. Activation Atmosphere
A study of activation atmosphere was made. This showed
that there was little difference between activation in
air and H2- However, activation under N? showed much
improved sorption. Mossbauer studies showed only the
ferric state of iron to be present under all conditions
and the reason for this effect remains unexplained.
- 30 -
-------
TABLE X
STANDARD SORPTION - REGENERATION CYCLE*
Cvcle Phase
Sorption
Heat Up
High Temp.
Stabilization
Regeneration
Cooling
Low Temp.
Stabilization
* Sample
Time Lapse % Gas
Temp.,°F Minutes Composition
660 90
24
1170 12
1170 75
21
660 18
size of activated
0.28
4.71
5.00
90.01
100.00
100.00
9.99
90.01
100.00
100.00
material is
SO,
°2
N,
"2
H2
N2
N,
N,
32 cc.
ACTIVATION PROCEDURE*
Activation
1170 240
100.00
N,
* Initial volume of dawsonite Is 40 cc.
- EC -
ipsa 50 po-fiaj uoT3daos uj
paqjos s V3M
1° V *
o o o
Gas Flow
@ S.T.P.
cc/mln.
2406
520
1201
1334
1201
520
o
•*
o
sr«
1201
I
Q.
to
o
n>
-------
2. SO,, Concentration and Space Velocity
For powder samples SO- sorption occurred at a rate de-
pendent on both volume hourly space velocity (VHSV) of
the gas and S02 concentration. That Is, higher SO,
loadings were obtained in a fixed time but the pickup
efficiency dropped with higher SO, concentrations and
higher space velocity.
3. Oxygen Level
It was shown that an oxygen concentration below 1% was
too low to oxidize the S02 to SO.,, resulting in greatly
decreased efficiency of SO, pickup. This also implies
that the sorption as sulfi|e is less efficient than the
sorption as sulfate at 660 F.
4. Temperature
TABLE XI
SORPTION REGENERATION CHARACTERISTICS OF GRACE #1 ALKALIZED
ALUMINA POWDER AT 1200 F OPERATING CONDITIONS
Wt.% Wt.% Regenera-
Sorption Breakthrough S Gain Saturation Gain (? tion Time
Cycle VHSV Time. Min. Break Tlme.Min. Saturation Min.
3970
3970
380
170
16.6
7.0
554
405
20.6
12.2
185
145
A sorption run was carried out at 1200 F to examine the
effect of higher temperatures on sorption. Higher tem-
peratures were expected to more rapidly oxidize SO. to
SO, when the SO, gas concentrations were in the range
of 0.5%. Equilibrium calculations showed that 1000°F
would have been better to use as at least 25% of the SO,
remained unconverted at 1200°F. Data shown in Table XI
is for a run comprising two sorption and regeneration
cycles on -50 +80 mesh Grace #1 alkalized alumina. At
the conclusion of this run the sorbent had the properties
shown in Table XII.
X-ray analysis of the sorbed phase showed that instead
of the customary Na»SO, formation, a sodium alum Na,Al(SO,),
was formed. Under the conditions of this experiment all
x-ray detectable sodium alum was regenerable, although at a
lower rate.
TABLE XII
MATERIAL CHARACTERISTICS BEFORE AND AFTER HIGH
TEMPERATURE (1200 F) OPERATION
Pore volume, cc/gm
2
Surface Area, m /gm
Average Pore Diameter, I
Crystallite Size (NaA102), i
5. Effects of COg on Sorption
Work at CERL indicated that at 450°F, CO, competed with
SOj for the active sorption sites and that this competition
could be avoided by increasing the temperature to 660°F.
An alternative explanation is that the faster oxidative
sorption mechanism is inoperative at 450°F. To distinguish
between these possibilities, CO, was introduced into the
- 35 -
-------
sorption gas stream at various oxygen concentrations. The
results of these tests appear in Table XIII.
At low oxygen levels (ca. 0.5% 0,) where both SO. con-
version to SO, and sorption were inefficient, the addition
of 10% CO- apparently improved the sorption rate. How-
ever, where sufficient oxygen pressure was present for
stoichiometric conversion of SO, to SO,, C^ hindered
the sorption rate and reduced tfie breakthrough capacity
at 660°F. This also occurred with oxygen far exceeding
the stoichiometric requirement. It would appear, then,
that even higher temperatures than 660 F might be needed
to fully overcome the CO- retarding effect at the high
C02 concentrations expected in the flue gas stream of
power plants.
The presence of CO- also indicates a somewhat higher
sorbent attrition rate. The reasons for this are
presently unknown.
6. Effect of NO on Sorption
It was shown that NO in the flue gas substantially in-
creased the sorption rate of sulfur on formed sorbents.
A catalytic reaction with SO, to form SO- was suspected.
In addition to this beneficial effect, it was also dis-
covered that NO produced a very serious loss of attrition
resistance. Inxthe laboratory, using a flue gas compo-
sition free of NO , after 1 cycle of sorption and regener-
ation,, a second degree severity (see appendix for test
details) attrition loss of 7 to 10% maximum was obtained
on Grace #2.
In one cycle at the Crane station of the Baltimore Gas
and Electric Co. (using the same Grace #2) the attrition
loss was more than 35%.
As a result, a series of laboratory tests was performed
to find which flue gas component was responsible for the
unusually high attrition loss. Table XIV shows the gas
compositions used and the results obtained. This data
clearly showed that NO even in low concentrations severely
affected attrition resistance.
Ul !-• f* O O
o o o ui ui
o o o o o
S^* ^ ^ £*
o o o o
o o o o o
o o o o o
£* Ul 00 **J £•
00 O H* Ul O
M 10 £• U> to
lo to o O* o
• • » • •
o\ ro «*j to o
% Oxj
Concent!
•^ ^«
(0 00
§3
1
(D o^
rt O
H 0
CM
rt
s-
3
1
,
Of
H H
n
• rt
2?
h** O
a c
• 00
cr
©F
^9
ca
1 t/i
Is
•5-
©rt
I--5-3
Ul .
o cn
S '
H- n
a B>
prj
'Tj
§
H
O
O
O
8
C/l
§
Hi
H
O
a
©
So
•n
PI
X
-------
I I X)
CJ 3 B
>J O rH CO
33 CO O
i>5 B -rl c
O M B CO
• • o ci <:
•H r! E co cu
O Vj Cb
•O -rl CU CO
B JJ 43 (J
CO O. JJ CU O
O> O T3 JJ CO
rH CO CJ -*
• oo JJ cu
O T3 S co Cu
CJ O 43
UJ JJ rH JJ M
O O O 01
OJ rl 13 43
co UH a. cu jj
B UH
O CO
CO
i B co
OJ CO
0)
CO
tl OJ CO O
JJ CO tl CO U^
OJ OJ 3O
O > 4-1 JJ
B Tf CO JJ
O co 43 M C
O JJ O
B OJ 13 • iH
•H 43 B 01
JJ 3 S OJ
S iH o -3 43
co OJ UH JJ JJ
OJ Z
Vi co B 43
JJ • co O JJ
CO X 3 -rl iH
rH JJ 3
CO 01 JJ CO
CO > -rl U T3
OOlH OJ OJ
•JJ - C JJ
cu o n o) o
3 OJ OJ 00 CO
••H P, > OJ OJ
UH CO 01 M IJ
•a
!H
tr
M
jj
CO
X
rH
CO
JJ
CO
O
01
•o
OJ
O JJ
a. eg
f a
4J CO
i c
B H
01 CU
00 B
o cj
JJ 43 U
u"» O e-S H co
CJ O O co
-UH rH CO 3
O UH • -i
OJ JJ CO CO fi
41 <0 0) O
i '3( rH «
• S
-H 3
I o
I JJ
I co co
B JJ co OJ 43
i OJ co B X M
I 00 CO 01 rH CO
i Vl <0 01 O
> 00 JJ M O
I. CO HH UK
CO B rH O
!3 OJ co B
oo • B
I CNO > rl B CO
o C S; o o co
"1 "O C ' '
S£MS
IH CJ JJ
CU 43 43 CO
B JJ JJ IH
CJ -ri 01
op 3 op a.
oj B B
IH OJ -rl CJ •
B -rl |j " "H
O Q .O
o is cu
0 O
CO X
olrH
V4U
CQ-
CU
JJO
I 43 43 rH OJ
JJ
I iH (>« CO JJ rl a
I 3 O EH O
I OO 73 UH O
IX B ai >
I JJ OJ iH > 01 iH
I -rl 43 -I 43 JJ
: o 4j c co jj o
I CO 3 43
i a B o o
CO iH 43 B JJ
I O CO CO
oo C
X 01 O
§|H B
43 CO IH OJ
JJ 43 CO OJ JJ
JJ CO 3 CO
^•S^^
co Tj B JJ
OlH
B°
43 B •« 0)
O eu ft OJ
OJ o IH 43
43 OJ O JJ
CNJ3CU
3B
W O CO 4J 3 0)
JJOiH t343
OJ
*>£ CSVJ CO
UH O O Q CO B
O -H u g 3 iH
01 OJ
co 43 O. N
O JJ S
JJ CO iH
•rl OJ B
43CU «OB
JJ>OICOO
U
T3
M 4J C
O CO CO
3
CO
B -
O B
OJ -rl O
43 JJ co
JJ iH M
•O OJ
O O U
m ogj
JJ u
rH C M
rH UH O
SCO rH
O 3 B
O O CO O
iH -rl
JJ CO JJ JJ
CD .rl 43 CO
JJ 00 JJ
CO B -rl CO
O OJ
ep-ri 3 cu
iH CO X O
jj jj 43 IH
a co o
rl l>!
CJ CO CM 0)
OJ 43 O JJ X
OOP JJ B
JJ co
* CX CO CX
B x -a o
CO C Tl O O
•H CO OJ JJ
x 5. B oj o
rl B -rl U iH
£ O JJ rH JJ
UH OJ Q.
O CO o
OJ M
&£ a
01 B
•rl
•o
CO JJ
B -H
43 I T3
CO iH CJ
B I JJ B
3 O CO lH
CO IH OJ CO
CU JJ
IH -O B 43
a co co o
OJ
OJ ^r> n 05
iH CU OJ 3
O U JJ CO
OJ
O JJ
B co
43 OJ
JJ 43 -O
a co
B OJ
•r,C43
•a -ri B
OJ JJ O
JJ CO
3 CU o'
co B X
OJ 01
IH 00 MH
0) O
CO V
JJ JJ
•H "O O
B B 01
3 CO UH
UH
01 (3 OJ
CO O
OJ -rl rH
43 JJ CO
JJ (X JJ
UH O OJ
o.^
QJ JJ IH
co B JJ
3 CO OJ
JJ TJ
01 rl
> O OJ
•H 0.43
co B JJ
B -3
01 CJ
JJ P rl
1-3
C P O
•rl £ 3
•o
o
n
o co a
•d co oi
o co -a
CO CO CO OJ
CO 00 >
CJ 01 O
01 O ^lj rl
<(-l X O Q.
"4H CJ CO B
JJ iH
UH S CO
O O UH
iH CU O
O JJ 01
•rl B UH
JJ O B -iH
CO -rl O -H
rH 4J H
3 co UH jj
i M B
3 OJ 01 CJ
O B B 43 •
O OJ -rl U CO
CO 00 tH O OJ
01 O CO -rl
CJ IH -I -O
43 43 rH 3
JJ 00 O 3 JJ
C UH CO
•0 iH 43 01
B M JJ co 01
CO 3 iH 3 CO
10 s^^
K "O op o u
5 oi Er
Tj I -ri OJ B
*• C JJ JJ o
K ° ° 9 !H
« UH co B UH
TABLE XIV
FLUE GAS COMPOSITION VS. ATTRITION RESISTANCE
Sample #: Grace #2
Run #
Gas Composition
% so2
% HjO
% N2
7« C02
% HC1
% NO
Gas Flow, cc/min.
Temperature, °F
Sorption
Regeneration
Number of Cycles
Attrition Loss, %, 2nd°
1
0.28
4.71
5.00
90.01
-
-
-
2640
660
1209
1
~6.0
2
0.28
5.00
5.00
74.72
15.00
-
-
2640
660
1209
1
11.4
3
0.28
5.00
5.00
74.62
15.00
0.10
-
2640
660
1209
1
13.4
4
0.28
5.00
5.00
74.67
15.00
-
0.05
2640
660
1209
1
50.8
5
0.28
5.00
5.00
74.71
15.00
-
0.01
2640
660
1209
1
52.4
6
0.28
5.00
5.00
74.11
15.00
-
0.05
2640
660
1209
1
79.2
Time of Exposure to
Sorption Gases, Min.
90
90
90
62
60
-------
TABLE XV
EFFECT OF CO.. ON REGENERATION*
Regeneration
Gas
Composition
Wt. % Sorbed
Pore Volume, cc/g
!%H2
7.co2
!C/g
2.
m /gm
leter, 1
> 0.27. H2S in
28
80
0
20
7.85
0.42
20
622
140
Run No.
23
80
5
15
8.12
0.48
19
622
150
24
80
10
10
7.60
NA
29
NA
70
Exit Gas Stream
* Regeneration Gas, VHSV - 276;o
Regeneration Temperature 1200 F
NA - Not Available
SECTION B
SCREENING NEW SORBENTS FOR
IMPROVED ATTRITION RESISTANCE
- 40 -
-------
I-B. Introduction
Two factors were principally responsible for the severe reduction
in resistance to attrition of unbound formed alkalized alumina.
• Sodium aluminate crystallite growth during
regeneration.
• Accelerated sorption at and near the surface
due to the presence of N0x in the flue gas.
We had previously shown that growth in NaAlO, crystallite size
occurred at elevated temperatures and relatively low steam con-
centrations (3%). When sorption of SO, was oxidation rate limited
(i.e. in the absence of NO ) sulfate buildup appeared uniformly
throughout the bead. Regeneration by sulfate reduction in H2
produced H,S and steam evenly throughout the particle, thus ac-
celerating uniform crystallite growth. It was postulated that
the crystallite growth destroyed the cohesive quality of the
original particle.
The influence of crystallite size attrition resistance under
these circumstances is shown in Figure 1. Laboratory single
cycle (sorption/regeneration) experiments on Grace #2 without
NO showed a maximum attrition loss of 10%. Attrition loss
increased to 100% when N0x was introduced to the synthetic flue
gas stream.
In the presence of NO , SO, sorption became diffusion rate
limited, thus concentrating the sulfate buildup as a shell be-
ginning at the beads' outer surface. Grace #2 material ex-
amined after exposure to flue gas containing NO exhibited a
severely cracked shell which led to the 100% attrition loss
observed.
On the basis of this physical picture, the investigation of
binder systems was begun. Binder systems were investigated
which would not be severely attacked by SO, and would create
sufficiently strong interparticle bonds which could resist rup-
ture by the rearrangement and physical growth of the sodium
aluroinate crystals. Dawsonite sorbents were prepared using many
materials as binders and evaluated by test procedures adopted
during the investigation.
100$
30*
10*
I
200 400 600 800
Crystallite Size £
1000
Figure 1: % Loss vs. Crystallite Size for Grace #2 Alkalized
-------
A second approach to the attrition problem was to impregnate
with sodium aluminate carriers of good attrition resistance
and stability to an acid environment.
In addition to these approaches, other process variables which
might affect attrition were also studied.
The following materials were investigated as potential binder
materials:
• Kaolin
• Meta-kaolin (calcined kaolin)
• Avery Clay
• Bauxite
• Bentonite
• Asbestos
• Jamaica Red Mud
• Silica (gel, sol)
• Alumina (gel, sol)
• Cement (Portland Type I, Type V)
• CaS04
• CaC03
• Meta-kaolin + NaOH
• Kaolin + NaOH
• Kaolin + Na2Si03
• Avery Clay + Na-SiO-
The process variables studied were:
• Sorption temperature
• Regeneration conditions
• Activation conditions
II-B. Details
_Aj Methods of Sample Preparation
1. Binder Incorporation and Beading Technique
In all preparations using a binder, the alkalized alumina
was prepared essentially by the Bienstock procedure. The
material was spray dried and iron oxide added as the re-
generation catalyst.
The general laboratory procedure was to premix, as dry
powder, the binder, iron oxide and alkalized alumina for
a period of one-half hour in a "V" blender. 75% of the
premixed powder was then placed in a Day Company pony
mixer, and mixed, with the slow addition of water, until
cohesive beads took shape. The remaining 25% of starting
powder was used to dry back the beads formed and to pre-
vent excessive agglomeration.
In those cases where sodium silicate was used, it was
first diluted to supply about half of the water required
for the beading operation. This silicate solution was
added to 75% of the original powder. Beading then pro-
ceeded according to the general method, using the remainder
of the water as necessary.
2. Impregnation
Two slightly different methods were used as explained
in (a) and (f).
(a) A saturated solution of sodium aluminate was
prepared at room temperature (alternate procedure:
a weaker solution (40% sodium aluminate) was pre-
pared at room temperature).
- 44 -
-------
(b) A quantity of alumina beads previously activated
at 600°F for 2 hours was weighed and placed in a
500 ml round bottom flask.
(c) Using dropwise addition, sodium solution was added
to incipient wetness with thorough shaking.
(d) The wet beads were heateu at 600 F for 2 hours.
(e) The following data were obtained:
Na-0 Analysis
S.A. (m2/g)
P.V. (cc/g)
X-ray Diffraction Pattern
(f) Steps (a) through (e) were repeated until Na20
analysis showed no further change.
(g) A solution of ferric ammonium oxalate was added
to give a 1% iron content based on the total
weight of dry beads.
(h) The liquid from beads was evaporated to dryness.
The following supports were investigated:
• Kaiser ^alumina
• Pechiney ^alumina
• Norton alundum
• Norton silicon carbide
• U.O.P. yalumina
J3J Evaluation for Sorption Efficiency and Attrition Resistance
In order to properly evaluate the new materials, it was neces-
sary to develop rapid but discriminating screening tests. The
following methods were adopted.
- 46 -
• A severe single cycle test for preliminary screening.
• A multicycle aging test to simulate plant scale
aging behavior.
• A final check was then made using plant-site field
testing units.
• The effect of process conditions on activation,
Sorption, regeneration and attrition were determined.
1. Single Cycle Screening Test
Since it was shown that NO increased the efficiency
of sorption, but adverselyxaffected attrition resistance,
a single cycle test was developed in which 0.05% NO was
present in the synthetic flue gas. The concentration
of SOj was 0.84% (considerably above the normally present
plant operating conditions) to accelerate the physical
degradation of the sorbents.
2. Multicycle Aging Test
The incorporation of NO and CO- into the previously es-
tablished standard laboratory Sorption flue gas gave
multicycle test data comparable to field test data for
sorption efficiency and attrition resistance. Table I
shows a comparison of attrition results from 7 cycle
laboratory tests with those from multicycle field tests
for two different sorbents.
3. Process Variables
Several process variables were studied to determine their
effect on sorption, regeneration and attrition.
(a) Sorption Temperature
Useful SO- sorption rates were found to exist in a
temperature range of 260 F to 1000 F; however, the
sorption mechanism was dependent upon the tempera-
ture employed. From 260 F to 800 F, SO- was probably
oxidized by NO- to SO, in the gas phase with the
sorption rate fas SO,J being diffusion controlled.
-------
- 8*7 -
I D to
oo 01 ro u> i— o o
>J (6 (D
^ * *
O *• M
to en t—
jipS-
rt
ft.
a o
B) H
8
z
§3
•n fi
H CO
e
II
I
This led to surface sorption and "eggshell" cracking.
Prom 800 F to 1000 F, NO- is increasingly dissoci-
ated to NO, thus reducing the amount available to
oxidize SO,. Consequently at the higher temperatures,
Sorption or SO- became oxidation rate controlled and
a substantial reduction in attrition loss of the
beaded alkalized alumina was seen. Unfortunately
the sodium aluminate forms a less easily regencrable
compound at 1000 F. Table II shows the results ob-
tained.
(b) Regeneration
A two-step regeneration mechanism was evaluated to
determine feasibility. Carbon monoxide was shown
to be an effective reducing agent at 1200 F for the
reaction:
—I I—
00 00
4CO-
4C0
Regeneration was then most effectively completed
at 450 F with a 3/1 gas stream of CO, and H,0 vapor.
The two-step method was found to be more effective
in total sulfur removal than hydrogen alone at 1200 F.
The added mechanical complications may well restrict
the utility of this procedure.
(c) Activation
Activation conditions were evaluated for kaolin-
bound dawsonite, the precursor to the active NaAlO-
sorbent. High air sweep rates through the bed were
required to minimize hydrothermal aging accompanying
the dehydration and/or the decomposition of Al(OH),,
dawsonite and hydrated binder material (kaolin).
The most effective use of clay as a binding material
required a further calcination step to initiate re-
actions between sodium aluminate and the silica in
the clay to form a ceramic phase (carnegieite).
Table III shows the effect of activation conditions
and Table IV the effect of heat treatment and binder
type on attrition loss.
Kaolin bound dawsonite produced a marginal improve-
-------
TABLE III
EFFECT OF ACTIVATION CONDITION ON ATTRITION PRIOR TO SORPTION
Sample
RC-174
25% Kaolin
73% Dawsonite
Activation
Condition
1200°F/4 hrs.
Air Sweep
% Loss
Thru 16 Mesh
800 F/2 hrs.
followed by
1020°F/ .5 hr.
1120°F/ .5 hr.
1200 F/ 3 hrs.
all with static
air
18
1200 F/ 4 hrs.
static air
98
TABLE II
EFFECT OF SORPTION TEMPERATURE ON ATTRITION RESISTANCE OF GRACE #2
Ul
o
Sorptlon Temp.
260
600
1000*
% NO
in Flue Gas
0.05
0.10
0.05
Attrition
Loss, 7.
100
100
1-2
* Material resulting from this sorption was NaAlSO^ not
NaSO. as normally found at 660 F sorption temperature.
-------
CO
o
R>
S
O.
b
!»
f
to
ro o>
»- a. o
H- H- C*
SIS'
n
jo
o
NO
ON
ON
O
=
=
*•
b
oo
Ut
ON
ON
O
Ul
VO
b
SO
o
1
: h^
£
Ul
5"«
-- £
o
en
• H*
(B
CO
5"
8-
n
(™*
/^ •**
h- O
b b
'Of r+4
r- (D
ON 0)
o n
o
*0tf
a
-~. m
o> n
— ' rt
00 -J
i> oo
NO tsA
fU
o1 *
n> (n
a.
O9 OO
«J 00
ON N>
?
r-
Tj
CD
r^
i^
O
O> 00
w o
O Ul
a
s
g
CD
rr
74
n
2
r
5
S
0
z
o
"fl
§ s
M f
n \st
*s
o z
z a
1"
M 2
H a
M
o n
ai
a
o
X
g
H
p|
H*
ft
O
51
CO
CD
i
w
<
bound samples at 1600 F, a substantial improvement was
obtained. The reason for this is believed to be the
formation, during the heat treatment, of carnegieite
(Na20-Al203-2Si02).
The kaolin and sodium silicate bound sample showed
considerably more improvement. The reason for this
is not known but may be due to enhanced binding caused
by additional formation of carnegieite. This might
result from greater availability of soda and silicate
from sodium silicate.
III-B. Conclusions and Recommendations
The most promising of the many materials investigated as binders
are shown in Table V and are compared to binder-free alkalized
alumina (Grace #2).
Table VI shows the most promising of the six sodium aluminate
impregnated carrier systems investigated. As a result of the
screening work the composition and process shown in Table VII
were recommended for further development.
Although small pilot plant scale production of this material
gave good results (Sample No. RL-175, Table V), a large plant
scale preparation gave attrition approaching 100%. Two reasons
were advanced for the lack of attrition resistance observed
in the plant run.
• The formation of appreciable ^-alumina trihydrate in
the large scale preparation of dawsonite.
• Scale-up difficulties and lack of control in the ex-
truding and rolling steps during forming leading to
poorly compacted beads.
The laboratory and pilot plant work which led to a solution
of this problem is described in Section C.
-------
TABLE VI
EVALUATION OF CARRIERS IMPREGNATED WITH SODIUM ALUMINATE
Sample No.
5613-14
5532-36B
Alumina
Source
Kaiser
jf-Alumina
Pechiney
y-Alumina
% S
Sorbed
6.0
5.0
% Attrition
Loss Thru 16 Mesh U.S.
After Severe 1 Cycle of
Sorption and Regeneration
15.1
18.6
TABLE V
EVALUATION OF ALKALIZED ALUMINA BINDERS
Sample
No.
RC-174
RL-175
5759-16
5759-C
Binder
Kaolin
Kaolin
Na2Si03
Metakaolin
Kaolin
Na2C03
Binder
25.0
23.0
2.0
25.0
16.8
8.1
Heat
Treatment
@ 1600°F
(Hrs.J
1..0
4.0
1.0
4.0
% Attrtition Loss
%S
Sorbed
8.49
6.61
7.50
5.38
Through 16 Mesh U.S.
After Severe Cycle of
Sorptton & Regeneration
37.6
10.7
18.0
22.8
Grace #2
None
4.0
12.61
-------
TABLE VII
COMPOSITION AND PROCESS RECOMMENDED
FOR FURTHER DEVELOPMENT
Sorbent Composition
75% Dawsonite
23% Kaolin
2% Sodium Silicate
1% Fe based on total dawsonlte
Conditioning
Activation - 1200°F/4 hours with air sweep
Heat Treatment- 1600°F/4 hours static air
- 56 -
SECTION C
LABORATORY AND PILOT PLANT PRODUCTION
OF ATTRITION RESISTANT SORBENTS
-------
I-C. Introduction
As a result of work carried out in Section B, it was shown that
several binder systems had good sorption and attrition resis-
tance characteristics. The work reported in this section shows
reproducibility in the preparation of these products as measured
by their sorption and attrition values.
In addition, an improved technique was developed for producing
well formed product. The sophisticated five module automated
field test unit was completed and checked for proper operation.
(See Appendix B).
II-C. Details
_A| Laboratory Preparations
Table I gives a summary of the kaolin bound sorbents which
earlier gave good attrition resistance. The preparations of
these products were repeated as described below:
1. Precipitation of Dawsonite in the Presence of Binders
These preparations in general followed the Bienstock
method.
(a) To 4000 ml of H,0 was added 764 grams of Na-CO-
(anhydrous) and the solution was heated to 90 C.
(b) A second solution was prepared by adding 1332 grams
of Al2(SO.),-18H,0 to 2000 ml of H,0 and heating
to 90°C. * J i i
5" S3
0> in
O vo
IB l
ro
ON
vO
I
ON
n
in
vb
I
ro
o
In
VO
vo
to
3 -O 3
(6 x^
^3
ro
in
CO ON
h- CO
ro
in
ro -o
ON in
(c) To the solution of Na^CO,, was added 345 grams of
kaolin. This slurry was mixed well for 10 minutes.
(d)
(e)
slurry
The solution of aluminum sulfate was then added
slowly to the sodium carbonate-kaolin slurry.
VO H-
ON 00
h* O
ro
ro
vo
oo
This reaction mixture was then allowed to stand
with stirring for two (2) hours at 90°C.
(f) After this two hour period 14.7 grams of fe.O
were added and mixed for 10 minutes.
o c ro o O §
SB co i- H- i-
P. p. p. p. p.
5-3 0333
H- i— ro ro ro
O *• ho u in in
in in O O O o
O *« *. i_i
in
"in ON OO 00
W ON ""J In
U h- ON W
ON O ON *»J
V- -J O ON
a
H-
r(
DO
H-
BH
£
H
f_J
ON H
0 r?
O 1C
OD> X
•D n re
O 9 »
cr re rt
1 3
to rt
CO
O
It
a
H.VS
i >
C rt
rt
ON H-
rt
ff o
Co 3
3-
cET
• 01
CO CD
1
?
O
§
t-1
H
CO
O
Si
(a
O
g
5,.
2j
9
S
c
^
M
S
f
l-l
N
W
O
r1
H
?
rt
BO
5
M
-------
(g) The resulting slurry was allowed to cool and stand
at room temperature for 16 hours.
(h) The slurry was then filtered and washed three times
with 2000 ml portions of a 3% NajCO- solution.
(i) The filter cake was then allowed to air dry to a
consistency suitable for beading in a Day Co. pony
mixer.
(j) During the beading operation, the beads were surface
dried by the addition of 400 grams of a powder having
the same composition.
(k) The resulting beaded product was then allowed to
dry overnight in a 110 C oven and then sieved on
U.S. 10 - 12 mesh screen.
(1) The product at this stage was examined for surface
area, pore volume, D.T.A., x-ray, and soda content.
The mole ratio of Na_CO. to A12(SO,)- used here was 3.6/1
and the theoretical yield of dawsonite was 1036 grams.
The quantities of reactants used to prepare dawsonite
were kept constant during the several preparations and
only the quantity of binder was changed.
2. Meta-Kaolin Preparation
(a) To 1200 grams of dawsonite was added 300 grams of
meta-kaolin and 17 grams of Fe20- to give 1% Fe
based on total dawsonite used.
(b) This mixture was blended for one (1) hour in a
"V" blender.
(c) 3/4 of the blended powder was placed in the pony
mixer and beaded by spraying approximately 1500 cc
of water into the mix.
(d) The remaining 1/4 of the dry powder was used to
dry back the wet beads.
(e) This mix was placed in an oven at 110°C for 16 hours
and then sieved through 8 - 16 U.S. standard mesh
screen.
(f) The product was examined as in (e) above.
Tables II and III, attached, list each preparation
with its physical properties after drying.
In addition to the preparations tabulated, a repeat evalu-
ation was made of an impregnated alumina. As previously
reported, an alumina produced by the Davison Chemical
Division had been used to prepare an impregnated sorbent.
The attrition of this material (Sample No. 6019-8) was
found to be 89% in preliminary screening.
The severe single cycle sorption and regeneration test
was repeated. The attrition was 100%, so no further
consideration was given to this sorbent.
Table IV gives the evaluation of all samples prepared
on laboratory scale.
_BJ Pilot Plant Preparations
Laboratory and small scale pilot plant preparations of a bound-
dawsonite product (23% kaolin, 2% sodium silicate, 75% dawsonite)
demonstrated good resistance to attrition. However, this same
formulation failed to demonstrate a similar attrition resis-
tance when produced at plant scale.
Two possible reasons were advanced to account for this
difficulty.
• The formation of appreciable quantities of crystalline
p-alumina trihydrate in the large plant preparation
of dawsonite.
• Scale up difficulties in the forming method which
produced poorly compacted beads.
As a result of laboratory work reported in this section, the
forming method was implicated while the presence of P-alumina
trihydrate was demonstrated to have little influence on the
observed reduction in attrition resistance. Consequently,
work proceeded on the development of a new forming method.
A successful forming method was developed using a machine
known as a Marumerizer. This equipment operated by rotating
- 60 -
-------
TABLE III
PHYSICAL EVALUATION OF META-KAOLIN BINDER DRY BLENDED
Sample
Crystallite
Sample No. Size A
6019-11-1 130
6019-11-2 170
1 6019-11-3 123
c^
10
1
6019-11-4 130
6019-13-1 123
„ S.A.
2
Binder Binder m /g
Meta- 25 61
Kaolin
Meta- 25 67
Kaolin
Meta- 25 99
Kaolin
Kaolin 25 49
Meta- 23 51
Kaolin
Na Silicate 2
P.V. X-Eay
cc/g Identification
.21 Dawsonite &
B-trihydrate
.23 Dawsonite &
B-trihydrate
.18 Dawsonite &
B-trihydrate
.14 Dawsonite &
Kaolinite
B-trihydrate
.17 Dawsonite &
Kaolinite
B-trihydrate
D.T. A.
Temp . C
of
Endotherm % Na00
-. 140 (S) 16.1
290 (M)
360 (L)
140 (S) 14.6
290 (M)
360 (L.)
JUU \l~i/
130 (M) 12.9
290 (M)
360 (L)
130 (S) 14.6
290 (M)
360 (L)
430 (S)
140 (S) 14.7
300 (M)
355 (L)
430 (S)
S = Small, M - Medium, L = Large
PHYSICAL EVA
Sample No.
6019-10-1
6019-10-2*
i 6019-10-3
(^
M
6019-12-4
6019-12-5
6019-12-6
6019-12-7
Dawsonite
Crystallite
Size
iisi
130A
125A
nsX
130A
Binder
Kaolin
Kaolin
Kaolin
Kaolin
Kaolin
Meta-
Kaolin
Kaolin
TABLE II
LUATION OF INSITl
, S.A.
2
Binder m /g.
25 79
25
20 89
15
25
25
25
I BINDER PREPARATION
P.V. X-Ray
cc/g. Identification
0.27 Dawsonite &
Kaolinite
B-trihydrate
0.32 Dawsonite &
Kaolinite
B-trihydrate
__ M
Dawsonite
B-trihydrat3
Dawsonite
D.T. A.
Temp. °C
of
Endotherm
130°C(M)
300 C(M)
350°C(L)
430 C(S)
140°C(M)
295°C(M)
355°C(L)
~
~
%Na20
14.3
14.2
13.4
14.33
13.40
•*- In this preparation an attempt was made to extrude the product.
However, we were not successful in producting extrusions of the
proper size
-------
4J
B XI
01 i-4 CO
4J 01 CO
- C I -H CJ
B O 3 <44
i-l O -4 01
-I CO 01 Vl
O Vi > B 5
- 01 01 O I
CO
Jrf
ffl 0) 4J 4J •
6-S S U M -4 Tl
co 01 11 CO 01
r-1 C 3 jQ « C
CO II 01 CO
g >» II 45 45 U
• U 4J 41
o
« B
5L°
•ff-fc-BS
XI i4 01 O 3
C >> m 4J i4 ra
3 Vl 11 CO 4J 01
O CO .5 01 CO V4
J2 > H 4J 4J
CO 0)
01 43
fo
43 «4
ti o co
01
of t
ite
nt.
ycl
ane
s t
ons
daw
wson
ipme
le c
e Cr
show
I4J 60 01
B co
•3 B Vi C-S
4J U a O 14
CO aS M4 00
•rl 1-4 II
g-a0^
•rl 01 -rl
4-1 6-5 CO 4J 4)
co in co co r-i
Vi r-. -H o> 41
CO 45 Vl co CO
CX • U II 01 H
01 01 ' '
Vl 4J
tXffl
U 1-1
01 i4 CO 01
4J r4 3 43
CO 1-1 XJ
Vi , •
0 O
U
M Vi Vi
c o w
o
H
.3
P
4J TJ a
01
CO 4J 4J
S O 4J
3 O
"4 CO XI
M B
4J iH
CXB
01
«"§
ca
raS
CO u
44 01
(X >
O CO
II 4J 43
IS"
3 4J CO Vl
•H CO 01 01
CO -rl r4 X
I CO (X-rl
ca oi IT e
4J B
CO
14
S
O XI O
•H O (X
4JS •
5;
oi
N
X-rl
B (4
4J 01
U O
4J 4J C
43 i4 CO
60 B 4J
3 O CO
O CO i-l
43 £ ra
co oi B
<44
> 4-1
01 4J
a oi x co
xi 3 a
wss9
I 14 B CO -rl
: CO B 45 CO
4J 5 O 00 3
41 -4 01 ti
O 43 4J i4 CO
i. S
41 a
CO 01
. c?3-g
14 CO 4J
01 S II CO
. fi C 45 41
U H O U
M4 U U
4J CO B
O • 0143 CO
a y ja -i
an
Th
B f4 VI 5 r4
4J 14 Vl XI P. >. -rl
CO 4J O r4 B CX
II 4J O 0) 14 14
ra 4J CO 604J V4 Vl
co co co a xi oi
S M VJ l>> V4 I II 60
XI O Vl XI -rl Vl H
4J>,O>VlOOI
M 43 a > 43 a txr4
01
45
4J
Vl Vl XI
4J i4 01
4J 3 00
co er o
oi o ra
ao oi
CO Q 00 —I
•aS>,t
01 r4 CO
O C B co
3 O O
XI T4 X)
o 4-i M e
Vi CO - -
si
p
v
te
cti
nt wa
n bou
me
li
O 01 4J O
CO 43 CO CO
3 EH xl 45 01
O Vl
Vl 01 -H
B CX43 3
i4 4J o1
r4 4J 01
O B O Vl
a a 4J
it 4J C«
I CO MO
CO i4 O O
CO -rl O
01 Vl CS
vi a-4
O
43
-
.01
S
vC
I
TABLE IV
SORPTION AND ATTRITION EVALUATION OF VARIOUS SAMPLES
Sample No.
W
5808-74A
(6019-11-4)
5808-72B
(6019-12-4)
5808-63A
(6019-10-1)
5808-63B
(6019-10-1)
5808-62^
N (6019-11-1)
"5808-62B®
(6019-11-1)
5808-70A^
(6019-11-3)
5808-70^
(6019-11-3)
5808-73A
(6019-13-D1)
5808-73B
(6019-13-D1)
% TrSlfment
Binder Binder (31600 F(hrs)
Kaolin 25 1
S A
P.V. % S Attrition Loss
(cc/g) Sorbed on 16 Mesh-U.S. Remarks
19.4 Standard
Kaolin 15
Kaolin 25
Kaolin 25
Meta- 25
Kaolin
Meta- 25
Kaolin
Meta- 25
Kaolin
Meta- 25
Kaolin
Kaolin 16.8
Na.CO. 8.1
Kaolin 16.8
Na^CO, 8.1
4
1
4
1
4
1
4
1
4
21
25
19
.38
.43
.38
.45
7.8
3.78
7.26
6.19
7.39
6.92
7.99
8.6
16.3
41.7
41.8
6.8
15.6
13.7
15.6
26.9
13.75
Powder Blend
Insitu-
Prep.
Dry Pow-
der Blend
Dry Pow-
der Blend
All samples were activated at 1200°F for four hours prior to heat except
those noted which were activated at 800 F for four hours.
All numbers in brackets refer to the samples after preparation and drying. All
-------
o
3
•o
O
a
•o
01
e
o
*W
M
•u
0)
CO
o.
t-l
Q.
O
4J
•o
01
CO
3
01
,O
•a
3
o
CO
01
N
i-l
CU
s
3
LI
S
4 1-1 -a
CO 3*
01 A CO 01
r-l ^ 00
U C ff
•M 01 *H CO
•H U .-H
•H C T3
i-l 01 01
(U m xj Li
J3 CO 3
« 3 4J
O M •-! 3
w cu o in
0) S * ^
LI O O CO
CO 33 CO
01 -< VJ
O, IS O
O« • *W
CO CO CO
a) o s o
L| -rl -0
oi u -a e
rj flj C flJ
S Jj *D 0
ON
ON
TABLE V
EVALUATION OF BEADS MADE USING MARUMERIZER
Sample No.
5808-55B
Batch #1
5808-56B
Batch #2
5808-57B
Batch #3
5808-55B
Batch #1
Field
Tested
Binder 46 Cycle
23% Kaolin No
2% Na2Si03
" No
No
Yes
Tested
Using Severe
Single Cycle
Yes
Yes
Yes
-
% S
Sorbed
6.42
6.54
5.59
-
Attrition
Thru
16 Mesh U.S.
9.4
37.5
20.0
-------
APPENDIX
A. Modified Air Jet Attrition Test
B. Sophisticated 5 Module Field Test Unit
C. Organization Chart
APPENDIX A
Modified Air Jet Attrition Test
Because of the need for outstanding attrition characteristics of a
sorbent, a new attrition test was developed for testing alkalized
alumina beads. The test is based upon one developed to determine
attrition of catalysts used in refineries. The petroleum Industry
finds good correlation between the test and attrition in the pneu-
matic lift units.
The figure following illustrates the Air Jet apparatus.
Description of the Apparatus:
- 27 psig
- 65°F to 75°F
- -35°F to - 45°F
1: Compressed air inlet with pressure
temperature
dew point
2: 1/4" Needle valve
3: 1/4" Steel nipple
4: 1/4" x 1/2" Steel
5: 1/2" Solenoid valve - Skinner Electric Valve Div. Conn.,
115 volts, 60~, 8 watts, 150 psi. Valve #LC2DB4150,
normally closed.
6: Universal timer - Dlmco Gray Co.
7: 1/2" steel nipple
8: 1/2" x 3/4" steel bushing
9: Flowmeter - Fischer & Porter Co. #6111 A329 3B1. Precision
bore florator tube #FP-3/4-27-G-10
10: 3/4" x 1/2" steel bushing
11: 1/2" iron nipple length
12: 1/2" Globe valve
1"
- 68 -
-------
13: 1/2" x 3/8" steel bushing
14: Imperial brass hose connector - connecting 3/8"
bushing with 1/2" plastic tubing
15: 1/2" I.D. flexible plastic tubing length - 40"
16: 1/4" steel pipe length - 2"
17: 0 to 30# pressure gauge
18: 1/4" steel tee
19: 1/4" x 1/4" half union coupling, Imperial
20: 1/4" copper tubing
21: 1/4" nut. Imperial
22: 1/4" x 1/4" half union coupling, Imperial
23: 1/4" x 1/4" steel coupling
24: 1/4" x 1/4" hose connection, Imperial Brass
25: Rubber stopper No. 9 Neoprene, concave surface
7/16" deep
26: 1000 ml conical flask with 1" hole in bottom,
Pyrex No. 4984
27: No. 20 mesh support screens (upper - 2 1/2" x 2 1/2",
lower 1 3/4" x 1 3/4") bolted into position
Working Principle:
Dry air is admitted for a given period of time, through a
concave stopper at a certain flow rate. The particles of
preweighed sample are caused to rub against each other and
against the sides of the container by the air flow. At the
end of test, the material is rescreened on #16 mesh screen
and the retained material is weighed. The ratio of weight
lost divided by the original weight, multiplied by 100, equals
the percent weight loss or attrition loss.
- 70 -
ACCELERATED AIR JET ATTRITION
APPARATUS ASSEMBLY
W. R. GRACE & CO.
DAVISON CHEMICAL DIVISION
TECHNICAL SERVICE LAB.
BALTIMORE, MD.
-------
Established Conditions for Testing Alkalized
Alumina Beads with the Following Properties:
Shape
Size
Average Bulk
Density
Total Volatile
at 300°F
Spherical
8 to 14 Mesh
40 to 50 Ibs./cu.ft.
2.0%
The same procedure is followed for 1st and Ilnd Degree Severity
tests except that a time of 2 1/2 hours is used for 1st degree,
and 1 hour for Ilnd degree. However, the attrition loss is
reported on an average time of 1/2 hour. Results in a given
piece of apparatus have generally been reproducible within a
maximum deviation of + 10% of test results.
A typical result of % attrition loss of the standard alkalized
alumina has been reported in Table #1.
TABLE I
Ilnd
30
3.55
1
14
Illrd
35
4.25
0.5
14
Testing Conditions:
Degree of Severity 1st
Rotameter Reading, % 26
Volumetric Flow Rate in S.C.F.M. 3.0
Time in Hours 2.5
Screen Mesh 14
All attrition losses for I, II, and III degree are reported for
an average time of 1/2 hour.
Standard Procedure for Illrd Degree
Severity for Alkalized Alumina Beads:
A 30 gram sorbent sample (±0.1 gram) previously screened on a No.
14 U. S. Standard Screen is placed in an inverted one liter conical
flask (26). The flask has a one-inch hole centered in its bottom
which is covered by a 14 mesh screen. Dry air (-40°F D.P.) is
admitted for 1/2 hour through a concave stopper at 4.25 S.C.F.M.
flow rate. At the end of 1/2 hour, the material is rescreened on
a U. S. #14 mesh screen and the retained material is weighed (X grams).
Attrition values are calculated as follows:
Let X grams equal the material retained on #14 mesh screen after
attrition, then
30 - X
Standard Alkalized Alumina
1st Degree of Severity
Ilnd Degree .of Severity
Illrd Degree of Severity
% Attrition Loss
Converted to % hr. Basis
8
11
15
30
x 100 - % attrition loss
- 72 -
Factors Affecting A.A.J.A.
The following factors are known to influence the test results:
a. Flask and stopper configuration
b. Sorbent moisture content and air moisture content
c. Air rate - the rate must be controlled by a pressure
regulator large enough to damp out any pressure surges
in the air supply.
d. Air inlet pressure has been found to be a factor as
well as air rate.
Three Degrees of Severity
The attrition loss has been found to be proportional to the
air velocity, "higher the velocity, higher the attrition loss".
But the attrition loss is not linearly proportional to air
velocity. There appear to be three distinct degrees of attri-
tion as the air velocity is increased. These three distinct
regions are called 1st, Ilnd, and Illrd degrees of severity.
-------
1st degree of severity is at a low air velocity and the attrition
loss in this case is mainly due to rubbing of particles with one
another and with air (low degree of severity) .
Ilnd degree of severity is at medium air velocity and the attri-
tion loss is partly due to rubbing and partly due to impact
(medium degree severity) .
Html giifoe of severity is at high air velocity and the attrition
loss in this case is mainly due to the impact of particles with
the metal screen and glass surface (high degree of severity).
APPENDIX B
Sophisticated 5 Module Field Test Unit
I. Introduction
The alkalized alumina, System II, test unit consists essentially
of the following five modular units.
f Reactor Module
• Analyzer Module
• Control Module
• Synthetic Flue Gas Module
• Air Conditioning Module
The first four modules are approximately 2" x 2' x 4'. The
smaller air conditioning module provides cooling air for
proper operation of the analyzer and control modules when
testing in power plants at high ambient temperatures.
The system has been designed to manually or automatically
monitor the sorption of S02 contained in a flue gas stack
over any desired number of cycles of sorption and regener-
ation of the sorbent. In the event that it is desired to
evaluate a synthetic flue gas or alternate reducing gases,
provisions have been incorporated for doing so.
II. Description of Modules
The following will serve to describe and indicate the
operation of each module:
A. Reactor Module
The reactor module consists of:
• Furnace heated reactor unit large enough
to contain one quart of sorbent.
-------
• Syringe pump for controlled water flow,
if necessary, into gas being sorbed.
• Vacuum pump for pulling flue gas out of power
plant stacks and into and through reactor.
• Primary water cooling and removal trap
after sorption.
• Sulfur trap for use during regeneration.
B. Analyzer Module
This module analyzes the flue gas entering or exiting
from the reactor and contains the following equipment:
• F & M gas chromatograph.
• Barber Coleman recorder with disc-integrator.
• Two gas sampling valves; one of Sec capacity
for SO.
analysl
for SO, analysis and the other of Ice for H.S
' sis.
• Thermoelectric cooler having two cells for
continuous use in removing water from gas
prior to entering the chromatograph column.
C. Controller Module
This unit contains all of the necessary devices for
setting and monitoring the condition of a run.
• Timing clocks to set the automatic timing for
the following functions:
"»•"«=-»•«**,.,«.. ..,
• Sorption
• Purge #1
• Regeneration
• Purge #2
- 76 -
• Gas sampling
• Furnace controller
• 12 point Barber Coleman recorder for
monitoring the various heated portions
of the systems.
• Fischer Porter flow recorder for moni-
toring the flue gas flow.
• Switches to enable the choice to be
made between automatic or manual
operation.
D. Synthetic Flue Gas Module
This module contains the necessary devices such as flow-
meters and piping connections to permit the make-up of
a synthetic flue gas or the study of other reducing gases.
It also supplies the CO, required for purging of the system
operation of certain valves during automatic operation of
the system.
An air compressor is also located in this module and serves
to provide the necessary air to the reactor for rapid cool
down prior to the sorption portion of the cycle.
E. Air Conditioning Module
This unit supplies cooling air to both the analyzer and
control module to insure these delicate parts against
excessive heat such as is encountered in many power plant
locations.
-------
o
4J
CO
"c
v<
4J
c
o
0
c
o
CO
Ll
o
M-l
CO
o
U
CJ
•3
X
M U
Q 0
W PC
Pi O
$
PC
3
1
f-l
<
T3
OJ
N
i-l
i— t
CO
^
<
T)
0)
>
o
Ll
I
1
U
Ki
Q
C
o
T4
4.'
CO
N
•rl
c
CO
M
Li
O
(U
CO
£
•3
a
>
U
01
•?
£
DO
00
0}
S
•0
no
6
H
> 4J
C C
CO
o.
•H
U
5
0)
T-l
o
CO
M
o
01
U]
oi ra
u vi
CO 01
.-I 13
5 X
Q Q
3
U
ell
a.
a
at
CO 9)
u C
CO -H
X!
O
Pa
s
•o
CO
CJ
S
cu
I
FLUE GAS ADSORPTION FLOW DIAGRAM
FLUE GAS
ROTAMETER
REGENERATION GAS
5cc
IOOP
THERMOELECTRIC
COOLER
V _ AUTOMATICALLY OPERATED
A ~ SOLENOIDS
& _ MANUALLY OPERATED
------- |