Kura 1993
Kura 1993
Kura 1993
Bi0pr0cessEngineering
9 Springer-Verlag 1993
Abstract. Oxygen transfer in a 0.35 m diameter stirred loop fermen- Elmayergi et al. [5, 6] observed that polymer additives
tor (a stirred tank with a concentric draft tube) has been studied with (carboxypolymethylene) caused enhancement of growth
water containing a small amount of polymer(polyethylene oxide) as
a drag-reducing additive. rates and amylase production in a culture of Aspergillus
Power consumption was measured. It was found that the addi- niger. Elmayergi et al. [5] suggested that mass transfer en-
tion of polyethylene oxide causes an increase of power consumption. hancement from bubbles to liquid or from liquid to mold
This is contrary to the results reported in the literature. pellets is responsible for the increase in growth rates. Later,
Volumetric mass transfer coefficients (KLa) were measured. In
Elmayergi and Moo-Young [6] interpreted the enhancement
water the introduction of the draft tube increased the K La coeffi-
cient. The increase in K z a became larger with impeller speed. On of growth rates in terms of increased mass transfer rates into
the other hand, mass transfer in dilute polymer solutions decreased the growing mold pellets. This additional data indicated that
due to the presence of the draft tube. An empirical correlation has the mass transfer from bubbles to liquid was not the rate-
been proposed for the volumetric mass transfer coefficient in stirred controlling step.
loop fermentors. It has a general applicability.
Several studies of the turbulent drag reduction of polymer
additives such as polyethylene oxide and polyacrylamide in
List of symbols stirred tanks have been reported. They are expected to be
beneficial in certain biochemical processes. There has been
a 1/m specific surface area interest in the power consumption reduction due to the
C constant in Eq. (6) drag-reducing effect of the polymer additives. Furthermore,
g m/s 2 gravitational acceleration
KL m/s overall liquid-phase mass transfer coefficient the reduction of the small scale turbulence due to the drag-
n 1/s impeller speed reducing additives may be beneficial in systems with shear
P W aerated power input by mechanical agitation sensitive micro-organisms. It should be noted that the vis-
P. w power input by sparged air cosity of dilute polymer solutions is low and comparable
Q m3/min volumetric gas flow rate with that of water.
U,g m/s superficial gas velocity
V m3 liquid volume Quraishi et al. [7, 8] measured power consumption in aer-
ated stirred tanks in dilute polymer solutions (polyacryl-
Greek symbols amide and polyethylene oxide) and reported torque suppres-
exponents in Eq. (3) sion under turbulent conditions.
exponent in Eq. (6) However, the polymer additives are likely to reduce the
Q kg/m 3 density mass transfer rate which is one of the most important
parameters for the performance of fermentors. In fact,
Ranade and Ulbrecht [9] found a significant reduction of the
volumetric mass transfer coefficient (Kra) due to the addi-
1 Introduction tion of polyacrylamide. They showed that the influence of
the polymer additive is primarily reflected in the reduction of
A stirred loop fermentor (a stirred tank with a concentric the specific interfacial area (a). An empirical correlation for
draft tube) has been investigated by Einsele and Karrer [1], Kra was proposed. It should be mentioned, however, that
Keitel and Onken [2], and Sleichter [3, 4]. The introduction the correlation given in their paper does not correlate their
of a draft tube can be expected to be a possible way of own data. The coefficient in their correlation should be
improving the performance of stirred tank reactors. It may 1.0 x 10 -3 instead of 2.5 x 10 -4.
produce a well-defined liquid circulation and a more homo- Joshi and Kale [10] discussed the effect of drag-reducing
geneous reactant distribution in the reactors. agents on mass transfer in the gas inducing type contactors.
224 Bioprocess Engineering 8 (1993)
2 Experimental
It is seen in Fig. 3 that the above correlation agrees reason- n = 8 rps in the agitation-controlling region and Eq. (2). At
ably well with the experimental data for n < 2 rps in the this high impeller speed, mechanical agitation plays a dom-
aeration-controlling region. inant role in mass transfer.
Van't Riet [13] proposed the following correlation from In water, the presence of the draft tube gave lower volu-
extensive literature data for aerated conventional stirred metric mass transfer coefficients than those obtained in the
tank reactors with water: absence of the draft tube. This coincides with the results in
K r a = 2.6 x 10-z ( p / v ) 0 . 4 UsO.5 (2) bubble columns with and without draft tubes [14]. For
n < 2 rps, the oxygen transfer rate was only slightly affected
Fig. 3 shows good agreement between the present exPeri- by the presence of the draft tube. The influence of the draft
mental data in the conventional stirred rank reactor for tube becomes more efficient with increasing impeller speed.
226 Bioprocess Engineering 8 (1993)
tO
tO 4) = U~gg. (4)
104
This functional form is assumed by considering the following
two limiting relations:
E For no impeller agitation ( P / V = 0) or the aeration-con-
trolling region (Po >>P/V), Eq. (3) reduces to
I P ~ i ~Jrrl ~ p n ~ Irpp
10-4
10 rnS/mln 10 2
K L a = c~ U~. (5)
Gas f l o w rote 0..10 3
Fig. 4. Volumetric mass transfer coefficient in fermentor (dilute poly- In the early studies, this functional form has been used for
mer solution). the volumetric mass transfer coefficient in bubble columns.
Stirred loop fermentor Conventional stirred tank It has been shown that 7 ranges between 0.78 to 1.0 and
(8 rps) (8 rps) varies widely from 0.24 to 1.45 [15].
9 500 p.p.m. PEO [] 500 p.p.m. PEO
9 100 p.p.m. PEO o 100 p.p.m. PEO In the agitation-controlling region (P/V>>Po) , Eq. (3)
9 0 pure water o 0 pure water yields
(0 rps)
Ax 1,000 p.p.m. PEO K L a = o: (0 g) - ~ ( P / V ) p U ~ - # ,
'I> 100 p.p.m. PEO ( = C ( P / V ) ~ U,~'). (6)
0 pure water
Many previous investigators proposed correlations for con-
ventional stirred tank reactors in the above form. Data on
K L a have been correlated with the power per unit volume
In particular, there were substantial increases in K L a with and the superficial gas velocity. Moo-Young and Blanch [16]
the introduction of the draft tube at low aeration rates. The showed that the exponent for ( P / V ) ranges between 0.4-0.42
presence of the draft tube affected the liquid mixing induced for water and 0.52-0.74 for electrolyte, while the exponent
by the agitation rather than that by aeration. for U,g ranges 0.35 to 0.5 for water and 0.26 to 0.62 for
Fig. 4 illustrates the volumetric mass transfer coefficient electrolyte.
for dilute polymer solutions. It is seen that the KL a coeffi- The above discussion suggests that the proposed correla-
cient in dilute polymer solutions increased due to the pres- tion has a general applicability.
ence of the draft tube unlike the results in water. A least-squares analysis of the K L a values obtained in the
In the aeration-controlling region (n = 0 rps), the addi- 0.050 m s fermentor gives:
tion of small amounts of water soluble polymer was harmful
in gas-liquid mass transfer rate. For the conventional stirred a = 1.222, fi = 0.5, y = 1.0. (7)
tank fermentor, the K L a decreased considerably by the poly-
Measured values of K L a are plotted against the K L a values
mer additive. This finding is exactly what was obtained by
Ranade and Ulbrecht [9]. However, in the case of the stirred predicted by Eq. (3) with Eq. (7) in Fig. 5. The correlation for
loop fermentor, the K L a coefficient was enhanced by the the stirred loop fermentor has an average error of 40.1%.
water soluble polymer. This result is closely linked to the Although this error is relatively large, the error involved in
increase in power consumption of the stirred loop fermentor an estimation of the accurate oxygen concentration in the
with dilute polymer solutions shown in Fig. 2. fermentor does not exceed 9.5 %. It should be noted, further-
In the stirred loop fermentor with the PEO solutions, as more, that the prediction of the correlation obtained from
described above, a very large and stable cavity which is extensive literature data by van't Riet [13] is accurate within
ineffective to mass transfer was not formed. This attributes 20% to 40%.
to an increase in K L a due to the polymer additive. The We also obtained the constants in Eq. (3) for the conven-
polymer solutions had significant effect on K L a at low aera- tional stirred tank fermentor:
tion rates. a = 0.939, fi = 0.5, 7 = 1.0. (8)
A considerable effort has gone into the development of
the correlation for the volumetric mass transfer coefficient. The error in the evaluated K L a for the conventional stirred
Since it is affected by many variables such as impeller power tank fermentor is less than 29.4%.
S. Kura et al.: Oxygen transfer in a stirred loop reactor 227
10 -I
lations in the literature in Fig. 6. Substitution of the values
S-1 for the constants given by Eq. (8) into Eq. (5) yields the fol-
lowing correlation for bubble columns (P/V < Pg):
ul
13
Y
-g
10-2
S K L a = 0.939 Uso1'~
As shown in Fig. 6, Eq. (9) is rather close to the correlation
of Shah et al. [12] (Eq. (1)).o
Substituting Eq. (8) into Eq. (6), we get:
(9)
Specific p o w e r consumption P/V The oxygen transfer in the stirred loop fermentor has been
5 10 2 s 10 3 W/m 3 discussed. The presence of the concentric draft tube sup-
10 0 I ' I 1 pressed the volumetric mass transfer coefficient in water. On
s-l/(m/s)O.S the other hand, in dilute polymer solutions K L a was higher
='~u s-1 Stirred tank / /~/ 5
due to the presence of the draft tube. It was observed that the
K L a coefficient in the stirred loop fermentor with dilute
om polymer solutions is higher compared with that for water.
-a
The increase in power consumption in the stirred loop fer-
10 -1
h~
mentor is responsible for this result.
10-1 o=
c _ 9
5. Elmayergi, H.; Scharer, J. M.; Moo-Young, M.: Effects of poly- 14. Kawase, Y. ; Moo-Young, M.: Influence of non-Newtonian flow
mer additives on fermentation parameters in a culture of behaviour on mass transfer in bubble columns with and without
A. niger. Biotechnol. Bioeng. 15 (1973) 845-859 draft tubes. Chem. Eng. Commun. 40 (1986) 67-83
6. Elmayergi, H.; Moo-Young, M.: Effect of a polymer additive on 15. Lee, Y. H.; Luk, S.: Aeration. Ann. Rep. Fermentation Processes
mass transfer into mold pellets. Biotechnol. Bioeng. Symp. 4 6 (1983) 101-147
(1973) 507-512 16. M0o-Young, M.; Blanch, H. W.: Design ofbiochemical reactors,
7. Quraishi, A. W.; Mashelkar, R. A.; Ulbrecht, J. J.: Torque sup- mass transfer criteria for simple and complex systems. Adv.
pression in mechanically stirred liquids and multiphasc liquid Biochem. 19 (1981) 1-69
systems. J. Non-Newtonian Fluid Mech. 1 (1976) 223-245 17. Smith, J. M.; van't Riet, K.; Middleton, J. C.: Scale up of agitated
8. Quraishi, A. Q.; Mashelkar, R. A.; Ulbrecht, J. J.: Influence of gas-liquid reactors for mass transfer. Proc. 2nd. European Conf.
drag reducing additives on mixing and dispersing in agitated on Mixing 1977
vessels. A.I.Ch.E.J. 23 (1977) 487-492
9. Ranade, V. R.; Ulbrecht, J. J.: Influence of polymer additives on Received February 17, 1992
the gas-liquid mass transfer in stirred tanks. A.I.Ch.E.J. 24
(1977) 796-803 Mr. S. Kura
10. Joshi, J. B.; Kale, D. D.: Effect of drag reducing additives on Prof. Dr. H. Nishiurni
mass transfer characteristics of gas-liquid contactors-gas induc- Chemical Engineering Group
ing contactors. Chem. Eng. Commun. 3 (1979) 15-19 Department of Mechanical Engineering
11. Ogut, A.; Hatch, R.T.: Oxygen transfer into Newtonian and Hosei University
non-Newtonian fluids in mechanically agitated vessels. 66 (1988) Koganei, Tokyo
79-85 Japan 184
12. Shah, Y. T.; Kelkar, B. G.; Godbole, S. P.; Deckwer, W.-D.:
Design parameters estimations for bubble column reactors. Prof. Dr. Y. Kawase (corresponding author)
A.I.Ch.E.J. 28 (1982) 353-379 Department of Applied Chemistry
13. van't Riet, K.: Review of measuring methods and results in Toyo University
nonviscous gas-liquid mass transfer in stirred vessels. Ind. Eng. Kawagoe, Saitama
Chem. Process Des. Dev. 18 (1979) 357-364 Japan 350