21

Fluid Flow Model for a Packed-Fluidized Bed

KUNIO KATO, HIRO ARAI, and UTARO ITO

Chemical Engineering Department, Gunma University,

Kiryu, Gunma, Japan 376

To establish afluidfllowmodel in a packed-fluidized bed, mesityl

oxide was synthesized by the catalytic aldol condensation of
acetone in a fixed and a packed-fluidized bed. Reaction kinetics
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

were determined in the fixed bed. Reactant conversion in a

packed-fluidized bed was analyzed by a two-phase model with a
known reaction rate. This two-phase model has two parameters:
the gas interchange coefficient between the bubble and emulsion
phases and the fraction of catalyst in the bubble phase. Using
the known gas interchange coefficient, the fraction of catalyst in
the bubble phase in a packed-fluidized bed was experimentally
determined to be about 0.2.

packed-fluidized bed is one in which particles are fluidized with gas in a

A column suitably packed with spheres, open-ended cylindrical screens,
and the like. Since such packings prevent the growth of gas bubbles in the
bed, particles (catalyst) and gas contact each other well. Gas and particle
flow patterns are standardized by the packing, and scale-up of this reactor is
simplified.
To simulate the reaction characteristics or particle-gas transport phenom-
ena in a packed-fluidized bed, a fluid flow model is the most important factor.
Two methods are used to investigate the fluid flow pattern in such a bed. In
one, tracer gas is put into the reactor, the response curve of this gas is analyzed
by suitable model, and the model parameter is calculated. In the other method,
the reactant conversion in the bed is analyzed by a suitable model with known
reaction kinetics, and the model parameter is calculated from it. The gas
interchange coefficient between the bubble and emulsion phase in packed-
fluidized bed was obtained (I) by analyzing the residence time curve of step
response of tracer gas with a two-phase model.
Ishii and Osberg (2) isomerized cyclopropane on a silica-alumina catalyst
in fixed and packed-fluidized bed packed with open-ended screen packing.
The reaction kinetics were studied in a fixed bed. Conversion in the packed
bed was smaller than that calculated by the plug flow model and larger than
that calculated by the perfect mixing model. Thus, the reactant conversion
in packed-fluidized bed was analyzed by the diffusion model. Kato et al. (3)
hydrogenated ethylene in fixed and packed-fluidized bed packed with spheres
and open-ended screen packing. The decrease of reactant conversion in packed-
fluidized bed from fixed bed under the same contact time was explained by
catalyst contact efficiency.

271

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 272 CHEMICAL REACTION ENGINEERING II

In this work, mesityl oxide was synthesized by the catalytic aldol con-
densation of acetone in a fixed and a packed-fluidized bed with open-ended
screen packing. Reaction kinetics were studied in a fixed bed, and the reactant
conversion in a packed-fluidized bed was analyzed by a two-phase model with
known reaction kinetics. The fraction of catalyst particles in the bubble phase
was obtained.

Solid line : fixed bed o p e r a t i o n

Dotted line: packed-fluidized bed

Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

Figure 1. Schematic of experimental apparatus

1 nitrogen cylinder 11 heater
2 stop valve 12 reactor
3 deoxidation tower 13 glass beads
4 dehydration tower 14 catalyst particles
5 pressure regulator 15fluidizedbath
6 pressure gage 16 thermocouple
7 orifice 17 gas chromatograph
8 needle valve 18 product
9 constant temperature bath 19 cooling device
10 air compressor 20 evaporation bottle of acetone

Experimental
A schematic of the experimental apparatus is shown in Figure 1. Since
oxygen gas hinders this reaction, nitrogen was used as carrier gas. Taken
from a cylinder it was passed through a deoxidation tower and divided into
two streams. One stream entered the orifice for metering the flow rate and was
fed to an acetone bottle submerged in a constant-temperature bath heated to
2 5 C . Nitrogen was saturated with acetone gas in this bath. The flow rate
of another stream of nitrogen gas was measured with orifice meter, and both
streams were combined. The concentration of acetone and the flow rate of
the feed gas to the reactor were determined by measuring the flow rate in both
streams. Acetone concentration in the feed gas was usually 5 and 10 vol % .
Gas chromatography was used to analyze acetone and mesityl oxide. Poly-
ethylene glycol 1000 was packed in the chromatograph column, and hydrogen
was the carrier gas. To prevent condensation of acetone vapor between the
exit point of the constant temperature bath and the entrance point of the
reactor, the tube was heated to 4 0 C . The reactor used for this experiment
(Figure 2) was a borosilicate glass cylinder, 5.2 cm id and 40 cm high. To

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 21. KATO E T A L . Packed-Fluidized Bed 273

Solid line : fixed bed operation

Dotted l i n e : packed-fluidized bed operation

Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

Figure 2. Details of the reactor

1 reactant gas
2 heater
3 air
4 thermocouple
5 catalyst particles
6 glass beads
7fluidizedbath

control the reaction temperature, the reactor was put into the fluidized bath,
10 cm id and 50 cm high as shown in Figure 1. Catalyst particles used in
this experiment were 80-100 mesh alumina (Tekkosha Co., Ltd.). Packing
used for a packed-fluidized bed was 1 cm id and 1 cm long 16 mesh open-
ended screen packing.
To eliminate water in catalyst particles, the reactor was heated for 1 hr
at 4 5 0 C in nitrogen and cooled to reaction temperature with flowing nitrogen
gas (100 cc/min). After the reaction temperature was reached, acetone gas
of constant concentration was fed to the reactor, and the reaction was started.
After catalyst activity was constant, experimental data were taken. In the
fixed bed operation, reactant flowed upward in the reactor. Since the porous
plate of this reactor is a good gas distributor in a packed-fluidized bed operation,
only catalyst particles are put into reactor and are fluidized among packings
with reactant gas.

Results

Fixed Bed. Mesityl oxide is synthesized from acetone according to Reac-

tion 1:

2CH COCH -> ( C H ) C = C H C O C H

3 3 3 2 3 + H 0
2 (1)

When the reaction temperature was 2 8 0 - 3 7 0 C , most of the product gas was
mesityl oxide. A small amount of mesitylene and polymer were detected.
The fluid flow pattern in the fixed bed is approximately plug flow, and
the fixed bed is operated at constant temperature. From material balance of
reactant gas in the small increment of the bed, the following equation is
obtained.

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 274 CHEMICAL REACTION ENGINEERING II

If the reaction rate of the aldol condensation of acetone is expressed to

be proportional to the concentration difference between the concentration
of acetone in reaction and equilibrium concentration of acetone defined in
Reaction 1, Equation 3 results:

r = K(C - C )
eq (3)

From Equations 2 and 3,

- log (1 - X/Xed = KL/U (4)

L / U [sec]
0.1 0.2 0.3 0.4 0.5 0.6
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

J_ 0.2

Figure 3. Retionship between

J-x/xeg and L / U in fixed bed

Figure 3 shows the relationship between reactant conversion and L/U at dif-
ferent reaction temperatures. From Figure 3 the reaction rate of this reaction
is expressed as Equation 3. The reaction rate constant at 2 9 0 , 3 3 0 , and
3 6 0 C per unit volume of bed is 3.66, 5.85, and 9.1 (liters/sec), respectively.
The effect of temperature upon the reaction rate constant in this temperature
range is:

= 1.75 10 (-9500/)
4
(5)

Packed-Fluidized Bed. There are bubble and emulsion phases in a

packed-fluidized bed; the size of gas bubbles is restricted by the packing. When
the gas flow rate is increased, the number of gas bubbles per unit bed volume
and the bubble raising velocity are increased, but the size of the gas bubbles
is almost constant. Therefore, the fluid flow model based on the bubble size
in the bed seems unsuited to fluid flow model in a packed-fluidized bed, it is
better to use the same kind of model as shown in Mathis and Watson (4) and
Lewis et al. (5). The following assumption are made to calculate the reactant
conversion.
( 1 ) The bed consists of bubble and emulsion phases. The bubble phase
consists of bubbles and particles around the bubble. The fraction of catalyst
in the bubble phase is a.
(2) The total volume of gas bubbles in the bed may be expressed as
S(L-L ).
m f _
(3) Gas interchange occurs between two phases. According to Kato et al.
( I ), the gas interchange coefficient per unit volume of bubble is

F = 4.2 (6)

(4) The gas flow rate in the emulsion phase is the same as the minimum
fluidized gas velocity of particles.

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 21. KATO ET AL. Packed-Fluidized Bed 275

(5) The gas in the bubble phase and emulsion phase flows through the
bed as plug flow.
(6) Since the effective thermal conductivity in a packed-fluidized bed is
considerably large and the heat of reaction of this reaction is small, the bed
is assumed to be operating under constant temperature.
Using these assumptions, from material balance of reactant in the small
increment of each phase, the following equations are derived.

Bubble phase:

(7)
dZ ' R (U-U )
h m{ Rh(U-U ) mt

Emulsion phase:
dC _ F (R -l)
(8)
e 0 h

dZ RU (Cb-C) + ( 1
"r; r e
- 0
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org

h m{
K b U mf
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

where:

Rb =

Boundary condition:
0 Ob Ce Co

In Equations 7 and 8, the relationship between the bed expansion ratio

and operating conditions is obtained from Kato et al. (6). Therefore, if the
reaction kinetics of reactant are known, concentrations of reactant at the exit
of bubble phase and emulsion phase are calculated from Equations 7 and 8 with
fraction of catalyst in bubble phase, a, as parameter. From this value, the
concentration of reactant at the exit of packed-fluidized bed becomes:

1.0

Q=0

1
^0.5
u
U f=1.36l
m
cm
/se ^

K=0.25(V ecl S

% M F =10.0W

Parameter a Figure 4. Relationship between re

actant concentrations in the bubble
5.0 10 and emulsion phases and bed height
L(cm) when the reaction rate constant is small

_ m {
mf)r
/Q\
^ out JJ~ v e o T " JJ vy bo W /

The Runge-Kutta method is used to calculate Equations 7 and 8. Before

comparing the experimental results with calculated ones, to investigate the
reaction characteristics of this model, the reactant conversion in packed-fluidized
bed in the first-order reaction is calculated with the fraction of particle in

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 276 CHEMICAL REACTION ENGINEERING II
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

bubble phase as a parameter. Figures 4, 5, and 6 show the relation between

the reactant concentration in bubble and emulsion phase and bed height with
fraction of the particles in the bubble phase as a parameter where the reaction
rate constant is 0.1, 6.25, and 12.5 (liters/sec) respectively. From these
figures, the larger the reaction rate constant, the larger the concentration
difference between two phases. When a is increased, the reactant conversion in
bubble phase is increased. The reactant concentrations in bubble and emulsion

Figure 6. Relationship between

reactant concentrations in the bub-
ble and emulsion phases and bed
L (cm J height for large reaction rate

phases are quite sensitive to a when the reaction rate constant is large. Figure 7
shows the relationship between the reactant conversion and KL /U with the
mt

reaction rate constant and fraction of particles in bubble phase as parameters.

From Figure 7, when the reaction rate constant is small and a is large, the
reactant conversion calculated from this model is almost the same as that in
plug flow reactor. However, when the reaction rate constant is large and a is
small, the reactant conversion calculated from this model is smaller than that
in perfect mixing flow reactor. From Figure 7, the reactant conversion in

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 21. KATO E T A L . Packed-Fluidized Bed 277

packed-fluidized bed is also sensitive to a when the reaction rate constant is

large. Therefore, to obtain a numerically from this kind of test reaction,
the reaction rate constant for the test reaction must be large.
Figures 8 and 9 show the relation between the reactant conversion and
L/U when the reaction temperature is 290 and 3 3 0 C , respectively. The
dotted lines in Figures 8 and 9 show the conversion vs. L/U calculated from
1.0

J , 0.5 iM// /*
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

K=0.25 .^ ^ O O K = 1 2 5
y k=125

( tvvo phase model

] Piston flow
V Perfect mixing
0.0
io- * 10 * 10'
KLmf/yH

Figure 7. Relation between conversion and K L / / U m

L/y (sec)
0 0-1 0.2 0.3 0.4 0.5 0.6 07
1.0
0-8 ?, 0
o 0 Reaction Temperature
0.6 ^ ^ 290C

H* 0.4

.2
Fixed Bed Reactor-.^
K=3.66 (V ec)
0.1 S

0.08
Figure 8. Conversion in packed-fluidized bed reactor vs.
that infixedbed reactor
L / U (sec]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
1.0
0.8 -
0.6 - Reaction Temperature
e
330 C
0.4 -

x 0.2

X f i x e d Bed Reactor

0.1 K = 5.85(Vsec)
0.08
Figure 9. Conversion in packed-fluidized bed reactor vs.
that infixedbed reactor
Hulburt; Chemical Reaction EngineeringII
Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 278 CHEMICAL REACTION ENGINEERING II

this words K(1/ ) sec

3.66 -
5.85 [
0.8 t> 9.10 )
Kato,Takeuchi,lto, Kubota
= 1-1
0.6- = 3.35

10A ' Ishi and Osberg

K(Vsec)
0.03
A 0.09
0.2 V 0.27
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org

0.652
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

4 1.36

0.2 0A 0.6 0.8 1.0

Xcai/* or X | c a

Figure 10. Experimental vs. calculated conversion

plug flow model under the same reaction rate constant. Figure 10 compares
the experimental results with calculated ones in this model. From Figure 10,
both the reactant conversion of this experiment and that of Ishii and Osberg
(2) and Kato et al. (3) agree well with the value calculated from this model
when the fraction of catalyst in bubble phase is 0.2. From Figure 10, the
reactant conversion in a packed-fluidized bed with almost the same size of
packing as that in this experiment is calculated from this model when the
fraction of catalyst in bubble phase is 0.2.

6 8 10" 2 t 6 8 10 2 e
t 6 8 10' 2
KLmf^H

Figure 11. Conversion calculated from this model vs. that in plug flow
and perfect mixing reactors for first-order reaction rate

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 21. KATO ET A L . Packed-Fluidized Bed 279

Discussion

In a packed-fluidized bed, the size of the gas bubble is restricted by the

packing. No matter how the bed diameter and bed height are changed,
the fluid flow pattern in the bed is almost the same as that in this kind of
experimental apparatus. The fluid flow model and the fraction of catalyst
particle obtained in this experiment may be applied to the packed-fluidized
bed with almost the same type and the same size packing ( D = 1.0-3 cm) p

as these of this experiment. Figure 11 shows the relationship between KL /U mt

and reactant conversion calculated from this model for a first-order reaction
with reaction rate a constant parameter when the minimum fluidized gas
velocity is 1.0 cm/sec. From Figure 11, when the reaction rate constant is
small (0.1 liter/sec), the reactant conversion in packed-fluidized bed is almost
the same as that in the plug flow reactor. When the reaction rate constant is
intermediate (1.0 liter/sec), the reactant conversion calculated from this model
Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

is smaller than that in plug flow reactor but usually larger than that in perfect
mixing flow reactor. When the reaction rate constant is large (10.0 liters/sec),
most of the reactant conversion calculated from this model is smaller than that
in perfect mixing flow reactor. When the reaction rate constant is large, the
reactant conversion in a packed-fluidized bed is affected by U/U at the same
mf

value of K L / U .
m f

6 8 10" 2 4 6 8 10 2 A 6 8 10' 2 4
KCoLmf / U ( - J

Figure 12. Conversion calculated from this model vs. that in plugflowand
perfect mixing reactors for second-order reaction rate

Figure 12 shows the relationship between KC L /U and reactant con-

0 mt

version calculated from this model for the second-order reaction with KC as Q

parameter. The same results as those described above are found in Figure 12.
Generally, the reactant conversion in a packed-fluidized bed may be larger
than that in fluidized bed reactor under the same reaction condition as shown
in Figures 11 and 12.

Conclusions

A fluid flow model in packed-fluidized bed is proposed. Comparing the

reactant conversion in packed-fluidized bed with that in fixed bed, the fraction

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.
 280 CHEMICAL REACTION ENGINEERING II

of catalyst in bubble phase is obtained. The fraction of catalyst in bubble

phase is about 0.2 for 1.0-3.0 cm open-ended screen packing.

'Nomenclature

a = fraction of particles in bubble phase

C = concentration of reactant (gram-mole/cc)
C = concentration of reactant in bubble phase (gram-mole/cc)
b

Cbo = concentration of reactant at the outlet of bubble phase (gram-mole/cc)

C = concentration of reactant in emulsion phase (gram-mole/cc)
e

^eo concentration of reactant at the outlet of emulsion phase

(gram-mole/cc)
Cea equilibrium concentration of reactant (gram-mole/cc)
=

C = concentration of reactant in feed gas (gram-mole/cc)

0

C = concentration of reactant at the outlet of packed-fluidized bed

Downloaded by UNIV LAVAL on July 12, 2016 | http://pubs.acs.org

o u t
Publication Date: June 1, 1975 | doi: 10.1021/ba-1974-0133.ch021

(gram-mole/cc)
F = gas interchange coefficient between bubble and emulsion phase
0

(liters/sec)
= chemical reaction rate constant (liters/sec)
L = bed height (cm)
^mf bed height at the minimum fluidized gas velocity (cm)
R = bed expansion ratio
b

r = reaction rate per unit volume of bed (gram-mole/cc sec)

rb = reaction rate in bubble phase (gram-mole/cc sec)
re = reaction rate in emulsion phase (gram-mole/cc)
S = cross sectional rate (cm ) 2

= absolute temperature ( K )
17 = superficial gas velocity (cm/sec)
U = the minimum fluidized gas velocity (cm/sec)
mt

X = conversion of reactant gas

^ea equilibrium conversion of reactant gas
= axial length (cm)

Literature Cited
1. Kato, K., Imafuku, K., Kubota, H., Chem. Eng. Japan. (1967) 31, 967.
2. Ishii, T., Osberg, G. L., A.I.Ch.E. J. (1965) 11, 279.
3. Kato, K., Takeuchi, N., Ito, U., Kubota, H.,J.Chem. Eng. Japan (1969) 2, 8, 204.
4. Mathis, J. F., Watson, C. C., A.I.Ch.E. J. (1956) 2, 518.
5. Lewis, W. K., Gilliland, E. R., Glass, W., A.I.Ch.E. J. (1959) 5, 419.
6. Kato, K., Imafuku, K., Hattori, K., Kubota, H., Chem. Eng. Japan. (1966) 30, 8,
748.
RECEIVED January 2, 1974.

Hulburt; Chemical Reaction EngineeringII

Advances in Chemistry; American Chemical Society: Washington, DC, 1974.