United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S2-87 /105 Feb. 1988
Project Summary
The Influence of Sorbent
Physical Properties Upon
Reactivity With Sulfur Dioxide
J. A. Cole, J. C. Kramlich, W. R. Seeker, and G. D. Silcox
Sulfation behavior was measured at
1000 and 1200°C for eight calcium
oxide sorbents which were well char-
acterized in terms of particle size, pore
structure, and specific surface area.
Sulfation results were compared with
predictions of a simple mathematical
model which applied the measured
sorbent characteristics. The compari-
sons, intended to provide direction for
model development, suggest need for
model improvement in areas such as
global kinetics at short times, and
accountability for changes in structure
due to sintering during sulfation.
Subsequently, the effects of the high
temperatures on the surface areas of
the sorbents in the absence of sulfation
were also determined. Surface areas
were marginally higher for the larger
sorbents after 1000°C injection; but,
in general, no correlation between
particle size and surface area loss could
be found. Surface area decay was
shown to be very rapid in the first 200
ms, and subsequently very slow.
Differences between carbonate and
hydrate sorbents' reactivities were
investigated using aerodynamically
size-classified materials: hydrates were
found more reactive on an equal pre-
reactor size basis. Also, no thermal
comminution of the hydrate particles
was noted within the high temperature
environment. Thus, the superiority of
hydrates on a common prefiring size
basis cannot be explained in terms of
their fragmentation into smaller
particles upon firing.
This Project Summary was devel-
oped by EPA's Air and Energy Engi-
neering Research Laboratory, Research
Triangle Park. NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
This report gives results of a determi-
nation of the influence of sorbent particle
size on reactivity toward 862 in an
isothermal environment. Two aspects of
sorbent size were identified as important
for current modeling efforts and general
understanding: (1) the direct influence of
particle size on the reactivities of con-
stant surface area precalcined limes, and
(2) the relationship between particle size
and the relative reactivities of calcium
hydroxides and calcium carbonates. The
critical objective of the particle size
studies was to identify the region of
transition between pore and product-
layer diffusion control. The transition
occurs as particle size is reduced because
the maximum distance that an S02
molecule has to diffuse through the
prestructure of the calcine is limited by
the particle radius.
In the first task (Task 5), the approach
was to measure the reactivities of high
surface area precalcines (prepared from
calcium carbonate) having a wide range
-------
of particle sizes. The precalcines had
nearly constant initial surface areas of
25-32 mVg. Sorbents were injected into
an isothermal, laminar drop-tube reactor
(ITR) downfired with an S02-doped
methane flame. Reactivity was deter-
mined by gas analysis as S02 capture
vs. time at 1000 and 12000°C. Solid
samples were collected for surface area
measurement in order to test the existing
model assumptions about surface area
loss and to provide input to further model
development.
In the second task (Task 20), SO2
capture was measured for raw carbo-
nates and hydrates which had been
aerodynamically classified into narrow
size cuts. Having compared the reactiv-
ities on an equal size basis, one size cut
from a carbonate and from a hydrate
were tested for in-situ particle size using
a cascade impactor sampling system.
Precalcined Sorbent Studies
Eight precalcined sorbents were stu-
died in Task 5. These materials were
obtained as raw calcium carbonates from
Pfizer Minerals and Pigments Division.
The first five were aerodynamically
classified from Marblewhite 125 (MW
125), a commercial limestone. The
remaining three sorbents were precip-
itated calcium carbonates (PCCs). All
sorbents were calcined to high surface
areas in dry air at 700°C using a
previously developed calcination proce-
dure. Two of the MW 125 sorbents, 10
and 20 yum, were prepared by calcination
to high surface area followed by sintering
at 950°C for 30 minutes in flowing dry
air. All of the sorbents were greater than
97% calcined. Particle size distributions
determined by X-ray sedimentation are
shown in Figure 1 for the precalcines.
Arrows on each curve indicate the
estimated uncertainties in the size
distributions due to pore filling by the
sedimentation medium. Time-resolved
SC>2 capture tests with precalcined
sorbents were carried out at a firing rate
of 4.6 kW(HHV) with 4200 ppm S02(dry,
0% O2).
High Surface Area Precalcines
Time-resolved sulfation data for the
precalcines at 1000°C are shown in
Figure 2. Except for the Albagloss
precalcine, most of the sorbent reaction
occurs within the first 250 ms after
injection. Figure 3 shows similar data
generated at 1200°C. Again, most of the
700
90
80
nt Fines
» NJ
3 0
(?50
8
S.40
$
30
yn
10
0
1
0
un
iPi
eca
\
v
1
Icine \
20 urn Precalcine
\
i i
•\ >
\
4
\
^
|
~
\
-
v
\
p»
V
"S
S
E
^\
\
\
^,
Efi
3re
S.
\
?
\
rcf
calc
^^
\
\
A
u
A'
\
nes
ine ~
\
V
\
v
*\
\
\
^-'
\l
\
Mbacai
\
V
*"
\
S
Pr
"*'
—- _.
ec
j
IV
\
at
4tf
ul
\
cm
afl
^^
ifex
e
loss
Free
Prec
alri
alr.i
ie
ne
1008060504030 20 108 6543 2 10.80.60.5 03
Equivalent Spherical Diameter /jm
Figure 1. X-ray sedimentation particle size distributions for precalcined sorbent. The range
of uncertainty, based on porosimetry analyses, is indicated by the arrows at the
median diameter for each material.
sulfation occurs at very short times.
Taken together, the two data sets exhibit
the following behavior:
• In going from 1000 to 1200°C, the
larger sorbents (20, 10, 2 //m) and the
Albagloss either increased in utiliza-
tion, or stayed the same. The smaller
sorbents (Albacar, Multifex) showed a
slight decrease in utilization. (Because
of its high aspect ratio, there is some
ambiguity in defining the correct
particle size for the Albacar as dis-
cussed below.)
D
O
O
30
25
20 urn. 27 mi/g
10 fjm. 32 mz/g
2 fjm nominal, 27 m*/g
Albacar 2 /jm nominal, 27 rrf/g
Albagloss 0.8 fjm nominal. 27 m*/g
Multifex 0.07 fjm nominal, 25 m*/g
D
O
O
20 fim. 27 rrf/g
10 iim, 32 m2/g
2 fjm nominal, 27 m2/g
Albacar 2 i*m nominal, 27 rrf/g
Albagloss 0.8 fjm nominal, 27 m2/g
Multifex 0.07 fjm nominal, 25 m2/g
20
WOO°C
3100 ppm SO*
30
25
20
S75
I
3"
0.25 0.50 0.75 1.0
Residence Time, s
Figure 2. Isothermal reactivity of calcined
sorbents at 1000°C in the EER
isothermal reactor at Ca/S =
1.0.
12OO°C
3100 ppm SO2
0.25 0.50 0.75 1.0
Residence Time, s
Figure 3. Isothermal reactivity of calcined
sorbents at 1200°C in the EER
isothermal reactor at Ca/S -
1.0.
-------
• Two of the materials, the 2 fjm
Marblewhite precalcine at 1000°C
and the Albagloss at 1200°C, show
a markedly higher reactivity at longer
times relative to the other standards.
TheAlbacar hasa needle-like structure
with a diameter of 0.5-0.7 /urn and a
length of 2-4.5 fjm. Thus, the Albacar
presents a minimum dimension that is
less than that of the Albagloss. The key
question is whether the material is better
represented in terms of reactivity by its
nominal size as reported by the Sedi-
graph (2 /jm) or by its minimum diameter
(0.5-0.7 fjm). Since the models predict
that the size dependence is governed
mainly by the relationship between the
particle interior and its closest surface,
then the 0.5-0.7 fjm would be the more
appropriate dimension for characterizing
sulfation behavior. Both dimensions will
be considered in analyzing size
dependence.
The dependence of ultimate utilization
on particle size is shown in Figure 4. The
Albacar is plotted as two points to
illustrate each of the interpretations of
its size. The data support the conclusion
that finer particle size promotes improved
utilization at both temperatures. Direct
use of the grain model against the data
indicates that pore diffusion ceases
constraining the process below 1-2 fjm.
The data do not clearly replicate this
trend. Rather, there appear to be differ-
ences between the precalcines that are
not fully characterized by the particle
size, surface area, or pore size measure-
ments. This is supported by the anom-
alous behavior noted above in Figures 2
and 3.
At 1200°C the data generally show
less variation in utilization with particle
size. The grain model vastly overpredicts
the capture which illustrates the neces-
sity of including a sintering step in the
calculation to correctly predict high
temperature behavior. For both 1000 and
1200°C, there are sufficient uncontrolled
differences between the sorbents(manif-
ested as apparent data scatter) to prevent
verification of the grain model conclusion
that pore resistance becomes insignifi-
cant below about 1 /urn.
Low Surface Area Precalcines
Time-resolved sulfation data for the
low surface area precalcines are shown
in Figure 5. Grain model predictions for
each are the solid curves on each graph.
These results are significant because the
measured SC>2-capture levels were
r
CJ.
•S 30
•8
I 20
.§
<5
to
Ca/S = 1.0
4200 ppm SO2
1000°C
r=is
Grain Model
Ca/S = 1,0
4200 ppm SO2
1200°C
r=1s
Grain Model
0.01
0.1 1.0 10 0.01 0.1
Particle Diameter, fjm
1.0
10
100
Figure 4.
A summary of the sulfation vs. particle size data for precalcined materials. The
solid lines are the grain model predictions. Solid symbols represent precipitated
calcium carbonates, and open symbols represent MW 125.
to
50
40
5 30
I
: 20
o
to
10
0
20 tun. 12.5 rrf/g
1000°C
Cso2 = 4200 ppm (dry, 0% 02)
lOum, 11.7 m*/g
1000°C
CSo2 = 4200 ppm (dry, 0%
Grain Model
0.2
O.4 0.6 0.8 1.0 O 0.2 0.4 0.6 0.8 1.0
Residence time, s
Figure 5. Time-resolved SOi capture by low surface area 10 and 20 /jm precalcines. Grain-
model predictions based on sorbent properties and reactor conditions are shown.
reduced only slightly by the large
decrease in surface area (compare to
Figure 2). Parametric modeling studies
had predicted that, in the pore diffusion
control regime, surface area would have
a small impact on sorbent reactivity.
Precalcines Sintering Tests
Surface areas were measured for
precalcines sampled from the ITR at 1000
and 1200°C without S02. Time resolved
data were obtained using the 10 /urn, 32
mVg MW 125 precalcine, while samples
using the remaining sorbents were
collected at 1 s residence time. The time-
resolved data are shown in Figure 6. This
demonstrates not only a significant loss
of surface area, but also a significant
difference between the extent of sinter-
ing at the two temperatures. These data
also imply that surface area loss at these
temperatures is faster at short times and
approaches a steady state value in only
a few hundred milliseconds. Data from
the remaining sorbents also show
greater surface area loss at 1200°C.
-------
However, no correlation was found
between surface area loss and particle
size. The latter finding may simplify the
manner in which sintering is incorpo-
rated into models of the sulfation
process.
Hydrate Carbonate Studies
Intrinsic differences between the
reactivities of hydrate and carbonate
sorbents were examined using three
hydrate and two carbonate sorbents
prepared by aerodynamic classification
of (respectively) Linwood atmospheric
hydrate and Vicron 45-3 carbonate.
These studies were performed at 1100°C
with a reactor firing rate of 9.4 kW and
3000 ppm S02 (dry, 0% 02). Figure 7
shows the ITR data and compares them
with similar results obtained in the EER
Controlled Temperature Tower (CTT). The
ITR data corroborate the CTT data, but
no inference should be made about the
relative magnitudes of the two data sets
because the CTT data were obtained
under non-isothermal conditions and at
a much longer residence time. These
data are all presented as S02 capture vs.
the mass mean particle diameter (deter-
mined by X-ray sedimentation). Recog-
nizing that hydrates are more reactive on
a constant prereactor size basis, tests
were carried out to measure the in-situ
sorbent particle sizes by cascade impac-
tor sampling.
The 8.8 fjm carbonate and 9.4 /urn
hydrate, because of their proximate sizes,
were chosen for cascade impactor sam-
pling. The samples were collected at a
residence time of 475ms on the impactor
sampling system. The results of the
impactor sampling tests are shown in
Figures 8 and 9. Two tests were run with
each sorbent, indicated by the open and
closed symbols on the graphs. Also
shown are the curves obtained by X-ray
sedimentation (Sedigraph).
The agreement between the impactor
sizing and the Sedigraph is far better than
expected. Within the range and accuracy
of the cascade impactor data, no change
in particle size distribution is evident for
either sorbent. One implication, of
course, is that for this reactor operating
condition the Sedigraph particle size is
sufficient for a correlation like that
shown in Figure 7.
c
30
25
t» 20
\
"S
2
^ 15
8
<0
•t
a
<" 10
5
0
i i i l l
70 fjm. 32 rrf/g
i
\
\
\
f X
1 X
1 X _
* -^
\ **/\te^
\ ^^^^^^^^^^^^^
\ o " — ^
\
•*—»•• --n— ? 20°
V v'
• ^
• ^
1 1 1 1 1
0 200 400 600 SCO 1000 1200
Residence Time, ms
Figure 6. Surface area profiles for 10 urn. 32 rrf/g precalcine at 1000 and 1200°C.
70
60
SO
I
§. 40
30
10
1100°C
Ca/S = 2.0
Q.
Linwood Commercial
A Hydrate
• Vicron 45-3
- Solid Symbols: ITR
0.38 &
- Open Symbols: C77
12s
2345 10
Mean Particle Diameter, ftm
20 30
Conclusions
Sulfation tests using size-classified
precalcines confirmed the dependence of
calcium utilization on particle size. The
results suggested that influences beyond
Figure 7. Particle size dependence of SOz capture by Linwood atmospheric hydrate and
Vicron 45-3. The open symbols represent data taken on the Controlled Temperature
Tower.
-------
I
I
Linwood, 9.4 tun
—1100°C.9.4kW(HHVr
—Sfl, = 7.05, S/?T=7.20
100
JO 5.0 3.0
Equivalent Spherical Diameter, um
1.0
0.5
Figure 8.
Comparison of in-situ particle size for 9.4 um Linwood hydrate (determined by
cascade impactor sampling) with the prereactor particle size determined by
Sedigraph analysis. Open and solid symbols represent independently repeated
runs.
100
Vicron, 8.8 urn
—1000°C 9.4kW(HHV)*
—SR^= 1.05. SRi= 1.20
(*) Higher heating value \\ie
50 30
1.0
Figure 9.
10 5.0 3.0
Equivalent Spherical Diameter, fjm
Comparison of in-situ particle size for 8.8 um (determined by cascade impactor
sampling) with the prereactor particle size determined by Sedigraph analysis. Open
and solid symbols represent independently repeated runs.
particle size, pore structure, or surface
area had an influence on utilization. Due
to this variability, it was not possible to
verify conclusively that the size for pore/
product transition was correctly selected
by the grain model. At 1200°C the model
overpredicted the capture, a point which
indicates the necessity of including
sintering in the model. At the higher
temperature, 1200°C, the results were
less conclusive: the model overpredicted
the observed capture for small particle
sizes. In addition, two separate size vs.
capture trends were indicated by the
data, one for the PCC precalcines and
the other for the MW 125 precalcines.
Low surface area precalcines were used
to study sulfation under reduced pore
diffusion resistance. However, as the
grain model had predicted, the decrease
in surface area had only a small impact
on the reactivities of the larger diameter
precalcines.
Surface area sampling tests bring into
question the model assumption of con-
stant surface area. The data showed both
residence time and temperature effects
on surface area. The parametric tests,
however, showed no impact of particle
size on surface area loss.
The hydrate/carbonate reactivity tests
showed that, on a common size basis,
hydrates are more reactive than carbon-
ates. The impactor data demonstrated
that thermal comminution of the large
hydrate fraction does not occur during
exposure to the reactor gas. Thus, the
enhanced reactivity of hydrates is not
indicated to be due to thermal fragmen-
tation within the reactor.
-------
J. Cole, J. Kramlich, W. Seeker, andG. Silcoxare with Energy and Environmental
Research Corporation, Irvine, CA 92718-2798.
Brian K. Guf/ett is the EPA Project Officer (see below}.
The complete report, entitled "The Influence of Sorbent Physical Properties
Upon Reactivity with Sulfur Dioxide," (Order No. PB 88-143 987'/AS; Cost:
$19.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45263
Official Business
Penalty for Private Use $300
EPA/600/S2-87/105
0000329 PS
s
CHICAGO
------- |