Chem. Eng. Technol.

14 (1991) 333 - 347 333

SlBub reduced bubble velocity due to interactions with reactor wall [6] Peters, M.H., Fan, L.-S., Sweeney, T.L., Chem. Eng Sci. 37 (1982)
UMF gas flow through emulsion at minimum fluidization velocity p. 553.
ZND gas distributor zone according to BAM model [7] Kobayashi, H . , Arai, F., Chiba, T., Tanaka, Y., Chem. Eng (Tokyo)
10 KBE tenfold enhanced mass transfer coefficient 31 (1967) p. 239.
[81 Follmer, G . , n e s i s , Ruhr-Univ., Bochum 1988.
191 Behie, L.A., Kehoe, P., AZChEJ. 19 (1973) p. 1070.
References 1101 Rowe, P.N., in: Yates J.G.: Fundamentals of Gus-Solids Fluidiza-
tion, Butterworth, London 1987.
[I] Wen. C.Y., Chen, L.H., Handbook of Fluids in Motion [ I l l Werther, J., Ger. Chem. Eng I (1978) p. 243.
(Cheremisinoff, N.P. and Gupta R., Eds) Ann Arbor Science, 1121 Kunii, D., Levenspiel, O., Fluidization Engineering, J. Wiley &
Michigan 1983. Sons, New York 1969.
[2] Cheremisinoff, N.P., Cheremisinoff, P.N., Hydrodynamics of Gas- 1131 Follmer, G., Lehmann, L., Baerns, M., ACSDiv. Petrol. Chem. 33
Solids Fluidization, Gulf Publishing Company, Houston 1984. (1988) p. 453.
131 Kato, K.. Wen, C.Y., Chem. Eng Sci. 24 (1969) p. 1351. 1141 Andorf, R . , Mleczko, L. Baerns. M., DECHEMA Mono,?/.. 122, Kutulwe,
[4] Mori, S.. Wen, C.Y., AZChEJ. 21 (1975) p. 109. VCH, Weinhrim 1991.
[ S ] Davidson. J.F., Harrison, D., Fluidised Particles, Cambridge 1151 Mleczko, L., Andorf. R., Baerns, M., The AIChEAnnual Meeting,
University Press, Cambridge 1963. Chicago, 1 1 - 16.1 I . 1990, Paper 57c.

Role of Sparger Design in Mechanically Agitated Gas-Liquid Reactors

Part I: Power Consumption

Vilas B. Rewatkar and Jyeshtharaj B. Joshi*

Power consumption (Part I) and liquid phase mixing time (Part 11) were measured in 0.57, I .O
and 1.5 m i.d. vessels. A pitched blade downflow impeller (PTD) was used. Design details of
the PTD impeller such as diameter (0.22T to 0.5T), blade width (0.250 to 0.40) and blade
thickness (2.8, 4.3 and 6.4 mm) were studied. The effect of sparger type and geometry on power
consumption has been investigated. For this purpose, pipe, ring, conical and concentric ring
spargers were employed. Design details of the ring sparger such as ring diameter, number of
holes and hole size were also studied in depth. Sparger location with respect to the impeller was
found to be the most important parameter and was therefore varied for nearly all the spargers
studied. A correlation for the power consumption has been developed.

1 Introduction
The hydrodynamics inside the stirred vessel depends mainly on
Many chemical processes involve the dispersion of a gaseous the designs of impeller and sparger and their locations within
phase in a liquid phase. An easy method to achieve such a the reactor. Operating parameters such as impeller speed and
dispersion is to bubble a gas into a liquid through one or several superficial gas velocity are also important. A careful analysis
orifices (spargers, perforated rings, perforated discs, etc.) fit- of previous work indicates that most of it was devoted to the
ted under a mechanical agitator. Mechanical agitation disperses optimization of impeller design so as to avoid flooding of the
the gas phase, increases the contact time of bubble with liquid impeller. Little attention has been paid to the optimization of
and increases the heat transfer with the wall or coil. sparger design. The aim of this contribution is to find a
favourable arrangement of stirrer and sparger, which would
* V.B. Rewatkar and Prof. J.B. Joshi, Department of Chemical Techno- prevent flooding. The stirrer is said to be flooded, when it is
logy, University of Bombay, Matunga, Bombay-400 019, India. COI incapable of dispersing the gas. As a result, the gas passes up-
respondence should be addressed to Prof. J.B. Joshi. wards through the reactor without being dispersed and the stir-

@ VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991 0930-751619110510-0333 S 03.50+ ,2510

334 Chem. Eng. Techno]. 14 (1991) 333 - 347

rer thus loses its effectiveness. The speed at which the impeller tween the sparger and the impeller was selected arbitrarily.
becomes effective for dispersion is called the critical impeller Therefore, it was considered desirable to carry out a systematic
speed for gas dispersion. The impeller characteristics were investigation into the effects of sparger design and location on
studied by measuring the power consumption in gas-liquid power consumption and liquid phase mixing time.
dispersions. The hydrodynamic instability of the impeller in the
presence of gas has also been studied.
3 Experimental
The mechanism of gas dispersion is discussed on the basis of
visual observation, power consumption (Part I) and liquid The experimental set-up is shown in Fig. la. Experiments were
phase mixing time (Part 11). carried out in 0.57, 1.0 and 1.5 m i.d. mechanically agitated
contactors (MAC), fitted with four baffles of widths equal to
10% of the tank diameter. Pitched blade downflow turbines
2 Previous Work with six blades (at 45" inclination) were used in this study.
Details pertaining to MAC'S and impellers are listed in Tables
The dispersion of gas in liquid by mechanical agitation has been 1 and 2, respectively. The impeller speed was varied in the
the subject of numerous investigations. Joshi et al. [ I ] , Pandit range of 0.4 - 10.5 rls. Air and tap water were used as gas and
and Joshi [2] and Midoux and Charpentier [ 3 ] reviewed the liquid phases, respectively. The superficial gas velocity was
work on mechanically agitated gas-liquid contactors. War- varied from 1.5 to 30 mmis. Different sparger designs, namely
moeskerken and Smith [4, 51, Frijlink et al. [6] and Breucker pipe, ring, concentric ring and conical spargers were
et al. [7] investigated this topic. Warmoeskerken and Smith [ 5 ] employed. These spargers are referred to as SP, SR, SRR and
found that the effect of sparger separation (distance between SC, respectively. Their construction details are illustrated in
impeller and sparger) on flooding is somewhat less pronounced Fig. lb. The spargers are described in accordance with their
in the case of D T (disc turbine) than for PTD (pitched blade design details as follows":
downflow turbine) impellers. They observed the displacement
of the power curve (power number versus impeller speed) with Pipe sparger [SP 60 - n h - dh - 11
the changes in distance between the sparger and the impeller. Ring sparger [SR ro.d. - n h - d h - 11
Breucker et al. 171 did not observe impeller flooding when the Concentric ring sparger [SRR rL,d - nh - dh - I]
sparger ring (diameter) was nearly as wide as the vessel. Conical sparger [SC 95 - nh - dh - I ]
Therefore, they recommended the use of large ring spargers.
Bruijn et al. [8] found that, with an increase in the number of I ) List of symbols at the end of the paper.
gas cavities behind the blade, the ratio PglPodecreases. In their
study, they used pipe and ring spargers of different diameters.
They observed instabilities in the power consumption when the
ring sparger and the impeller were of the same diameter (for
disc turbine). Botton et al. [9] confirmed the importance of the
sparger hole diameter and the number of holes for achieving the
same hole velocity ( v h ) for a given V , in vessels of different
sizes. Hassan and Robinson [lo] systematically studied mass
transfer in gas-liquid contactors and found that PJP, ratio and
gas hold-up correlations were highly specific to a particular im-
peller type, tank size and physicochemical properties of the li-
quid phase. They observed no effect of sparger design in
smaller vessels. For large vessels, they recommended the use
of ring spargers rather than single orifice spargers.

In most of the mentioned studies, the impellers were disc tur-

bines. However, the pitched blade downflow turbine is known
to be more energy efficient in various applications, such as
homogenization of liquids [l 1 , 121, solids suspension in solid-
liquid systems [13 - 151 and gas dispersion in gas-liquid system
[ 161. The energetic efficiency of the PTD impeller is attributed
to its low power number and relatively uniform power dissipa-
tion in the tank 117, 181. In spite of this, very little attention has
been given to the PTD design. Furthermore, previous work
placed the emphasis on the understanding of the process of gas
dispersion and optimization of the impeller design, so as to
Fig. la. Experimental set-up, I D.C. motor, 2 cooling fan for motor,
avoid flooding. Little attention was paid to the systematic op- 3 bearing supoort, 4 tank, 5 impeller, 6 torque table, 7 weigths and pan,
timization of sparger design in gas-liquid o r gas-liquid-solid 8 conductivity meter, 9 recorder, 10 conductivity probe, I I thyristor panel
systems. Many workers used a single pipe for gas supply in- for conversion of A.C. to D.C., 12 gas sparger, 13 sparger pipe, PI pulse
stead of a ring sparger. More often than not the distance be- addition point.
Chem. Eng. Techno]. I4 (1991) 333-347 335

( A ) CONTROL ( 9 ) CONCENTRIC R I N G
[SC 981 (C CONCENTRIC R I N G
S R R 200 1 S R R 4101

( E l R I N G f S R r.o.d.1 not t o scale

I Fig. lb. Sparger design.

where, in the case of pipe sparger, 60 mm is the sparging length the outer diameter of the largest ring and 95 is the outer
(length of pipe with drilled holes), nh is the number of holes, diameter of the cone base in mm.
dh the hole diameter in mm, 1 the distance between the top
orifice of pipe sparger and the centreline of impeller blade. In Other design details of the spargers, such as tube diameter, hole
the case of ring spargers, is the outer ring diameter, rA.cl. size, number of holes and location are given in Table 3. The
design of the ring sparger was studied in detail. Ring diameter
was varied from 0.5 to twice the impeller diameter. In addition,
Table 1. Design details of mechanically agitated contactors. the effects of the size and number of holes were studied using
a ring 0.8 D in diameter in the 1.5 m i.d. vessel. For this pur-
Vessel diameter 0.57 m, 1.0 m, 1.5 m pose, hole sizes of 2, 3 and 6 mm were used, maintaining the
HIT 1 free area (total area of holes) for gas sparging constant. Fur-
Baftle width (percent of tank diameter) 10 thermore, the effect of free area was also studied by increasing
No. of baffles 4 the number of the 3 mm holes (42, 84, 168).
Material of construction Transparent acrylic
Geometry Cylindrical with flat
bottom The design details of pitched bladed downflow turbine were in-
Impeller clearance of tank bottom TI3 vestigated thoroughly. The ratio of impeller to tank diameters
-

Table 2. Impellers studied in this work.

__
Impeller No. of Diame- Blade Hub Horizontal Blade Blade
tY P C blades ter width 0.d. blade pitch thick-
length (degree) nesa
b D [ml W [ml do [ml I Iml 4 k In11

PTD 6 0. I9 0.057 0.04 0.075 45 0.0023

PTD 6 0.19 0.057 0.05 0.07 45 0.0028
PTD 6 0.33 0.099 0.05 0.14 45 0.0028
PTD 6 0. I9 0.038 0.05 0.07 45 0.0028
PTD 6 0. I9 0.0475 0.05 0.07 45 0.0028
PTD 6 0.19 0.057 0.05 0.07 45 0.0028
PTD 6 0.19 0.0665 0.05 0.07 45 0.0028
PTD 6 0.19 0.076 0.05 0.07 45 0.0028
PTD 6 0.33 0.0825 0.05 0.14 45 0.0023
PTD 6 0.33 0.01 16 0.05 0.14 45 0.0023
PTD 6 0.19 0.0665 0.05 0.07 45 0.0043
PTD 6 0.19 0.0665 0.05 0.07 45 0.0064
PTD 6 0.19 0.057 0.04 0.075 45 0.0028
PTD 6 0.5 0. I5 0.05 0.225 45 0.0028
PTD 6 0.75 0.225 0.05 0.35 45 0.0028
~ -~
336 Chem. Eng. Technol. 14 (1991) 333-347

Table 3. Details of sparger dcsign.

Sr. Sparger Sparger Ring Orifice Number Sparger

model type diameter size of location
i.d. o.d. orifices (below = +ve
[mml [mml above = -ve)
[mml [mml

1. [SR95 6 3 - 1001
- - ring 67 95 3 6 100
2. [SR95 - 6 - 3 - 1521 ring 67 95 3 6 152
3. [SR95 - 6 - 3 - ( - 35)] ring 67 95 3 6 - 35
4. [SR 152 - 12- 3 1001 - ring 114 152 3 12 I00
5. [SR 190-6-3-l00] ring 158 I90 3 6 100
6. [SR 280-21-6-1001 ring 245 280 6 21 100
7. [SR 280 - 21 - 6 - 3901 ring 245 280 6 21 100
8. [SR 400 - 42 - 3 - 1001 ring 350 400 3 42 I00
9. [SR 400 - 42 - 3 - 3901 ring 350 400 3 42 390
10. [SR 400 84 3 - 1001
- - ring 350 400 3 84 100
1I . [SR 400- 84 3 3901 - - ring 350 400 3 84 390
12. [SR 400 - 168 3 - 1001
- ring 350 400 3 168 I00
13. [SR 400 - 168 - 3 - 3901 ring 350 400 3 168 390
14. [SR 400 - 189 - 2 - 1001 ring 350 400 3 I89 100
15. [SR 400-189-2-3901 ring 350 400 2 I89 390
16. [SR 400-21-6- 1001 ring 350 400 6 21 100
17. [SR 400 - 21 - 6 - 3901 ring 350 400 6 21 390
18. [SR 600 - 21 - 6 - 1 101 ring 550 600 6 21 110
19. [SR 600 21 6 3901
- - - ring 550 600 6 21 390
20. [SR 1000- 168-3-3701 ring 950 I000 3 168 370
21. [SRR 200 - 28 - 2 - 1521 concen- 200 2 28 I52
tric rings
22. [SRR 410- 189-2-3901 - 410 2 189 290
23. [SRR 410 - 198 - 2 - ( - 45)J 410 2 I89 - 45
24. [SP 60- 14-3.6-701 Pipe 3.6 14 70
25. [SP 60- 14-3.6- 1521 Pipe 3.6 14 152
26. [SP 60-14-3.6-1001 Pipe 3.6 14 100
27. [SC 98-28-2-951 conical 2 28 95

(DiT) was varied from 0.27 to 0.5 in the I .5 m i.d. vessel. The condition also arises on increasing the number of holes or
ratio of blade width to impeller diameter (WID) was varied decreasing the hole size at the same value of V,.
from 0.25 to 0.4 in the 0.57 m i.d. vessel and from 0.25 to 0.35
in the 1.0 m i.d. vessel. The effect of blade thickness (2.8, 4.3, It was considered desirable to develop a criterion for ensuring
6 mm) was also studied. Since the sparger location was found that all the holes shouldbe effective. It was observed that, when
to be the most important parameter for the gas dispersion pro- the orifice Reynolds number [d,, V,, @ G / p G ]exceeded 2100, gas
cess, its effect was studied for nearly all the sparger designs. was sparging through all the holes. This criterion was found to
hold for all the sparger designs and in both the 0.57 and 1.5 m
An attempt was made to clarify the mechanism of gas disper- i.d. vessels. The criterion includes the variations in V, as well
sion by visual observations as well as from the variation of as the number and diameter of holes. The orifice Reynolds
power input (Part I) and liquid phase mixing time (Part 11) with numbers are listed in Table 4. The criterion of minimum
respect to the impeller speed. Power consumption was Reynolds number can be successfully applied for determining
measured by using a frictionless torque table. Details pertaining the hole size and the number of holes for a given gas flow rate.
to the measurement of power consumption were given by
Rewatkar et al. [19].
4. I .2 Liquid Circulation Generated by Gas Sparging

4 Mechanism of Gas Dispersion In the absence of impeller action, the sparged gas develops a
liquid phase flow pattern similar to that in a bubble column, as
4. I In Absence of Impeller Action shown in Fig. 2. The liquid flow develops because of the non-
uniform hold-up profile in the radial direction. When the gas is
4.1.1 Criterion for Effective Gas Sparging sparged in the central region, the gas hold-up is high in this
region and very low near the column wall. Therefore, the static
When a gas is sparged in a vessel, it rises upwards because of pressure below the sparger is lower than that near the column
the buoyancy force associated with it and escapes from the li- wall which results in the pressure driving force for liquid flow.
quid surface. It was observed that, at a very low gas flow rate This produces intense liquid circulation which is directed up-
(low VG),only few orifices of the sparger were effective. This wards in the central region and downwards near the column
Chem. Eng. Techno]. 14 (1991) 333-347 337

Table 4. Orifice Reynolds numbers of different sparger designs.

-
Orifice Reynolds number (Re,) When the gas sparger is large so that the gas is sparged close
Sparger
VG [mm/s] : 1.5 2.4 4.8 9.4 to the column wall, the direction of liquid circulation is revers-
ed to downward in the central region and upward near the col-
A) T = 0.57 m, H/T = 1 umn wall. For this case, the average liquid circulation velocity
SR,, [SR95-6-3- 1001 1754 2806 5613 10991 is also given by Eq. (1). The liquid circulation generated by gas
SR,,, [SR152- 12-3- 1001 877 1403 2806 5496 sparging has several implications:
SR,, [SR190-6-3-100] 1754 2806 5613 10991
sc,, [SC98 28 - 2 - 951
- 564 902 1804 3532
- When the liquid circulation is directed upwards in the cen-
SF,, [SP60- 14-3.6-701 626 1002 2005 3926
SRR,, [SRRZOO- 28 - 2 - 1521 564 902 1804 3532 tral region, the direction of liquid flow is opposite to that
SR,, [SR95-6-3- 1001 1754 2806 5613 10991 generated by the impeller if the latter is in action. When the
B) T = 1.5 m, H/T = 1
impeller (PTD) is not energized, the liquid circulation
SR,,, [SR400-42-3-100] 1734 2775 5550 10866 generated by sparging induces impeller rotation in the
SR,,, [SR400 - 84 - 3 - 1001 867 1387 2775 5434 reverse direction. Under these conditions, the torque re-
SR,,, [SR400- 168-3- 1001 434 694 1387 2717 quired for starting the impeller action is much greater than
SR,,, [SR400- 189-2- 1001 578 925 1850 3623 in the case of stationary liquid (point L in Fig. 2 ) . This
SR,B, [SR400-21-6- 1001 I734 2775 5550 10867 results in a higher power consumption by the impeller than
SRR,, [SRR410- 189-2-3901 578 925 1850 3623
in pure liquid. The increase in the power consumption ob-
SR,, fSR280- 2 1 - 6 - 1001 1734 2775 5550 10867
SR,,, [SR600-21 - 6 - 1001 1734 2775 5550 10867 viously depends on the intensity of liquid circulation by gas
[SRIOOO- 168 - 3 - 3501 434 694 1387 2717 sparging as given by Eq. (1).
SR,,.

- The liquid circulation developed by gas sparging also

governs the relative importance of impeller action and gas
I sparging. At low impeller speeds, the liquid flow generated
by the impeller is small, compared to that generated by gas
MN- CAVITY FORMATION
sparging. Therefore, gas sparging controls the flow pattern
and the behaviour of mechanically agitated reactors. At a
E F - RECIRCULATION certain critical impeller speed, the flows generated by im-
v) peller and gas sparging are comparable. At high impeller
speeds, the impeller action takes over and the behaviour of
z 2a
- the MAC is then controlled by this action.
a-

4.2 In Presence of Impeller Action

3
0
a The flow pattern generated by pitched blade downflow turbine
in the absence of gas flow was discussed by Ranade and Joshi
IMPELLER SPEED, N ( r / s 1 [17, 181. When gas is sparged, visual observations indicate
various dispersion patterns. The dispersion pattern (flooding,
cavity formation, complete dispersion of gas, recirculation of
gas-liquid mixture) was found to depend on the hydrodynamic
characteristics of the impeller which, in turn, depend upon the
impeller speed, superficial gas velocity, sparger design and
sparger location. The different flow patterns can be observed
visually. In addition, these flow patterns can be related to the
variation of power number and liquid phase mixing time with
change in the impeller speed at a constant superficial gas velo-
city. The variation of power number depends strongly on the
sparger design and its location. This aspect will be discussed in
detail. The variation of mixing time with impeller speed is dealt
with in Part 11. Fig. 3 shows the variation of power number for
FLOW PATTERN AT VERY LOW IMPELLER SPEED (RANGE L(L) different sparger designs.

Fig. 2. Variation of power number with impeller speed in gas-liquid system

(at constant V G ) . 4.2.1 Sparger Below and Close to Impeller

wall. The average liquid circulation velocity can be obtained Curve I represents the variation of N,, for the case of ring
from the force or energy balance and is given by the following sparger located below and close (100 mm) to the impeller. At
equation: low impeller speeds, the liquid flow generated by the impeller
338 Chem. Eng. Technol. 14 (1991) 333 - 347

break-up does not occur through the impeller action in this

SYMBOL SPARGER
1 region, the impeller is said to be flooded with gas, and the
region L , N , is called the flooding region.
SR,, [ SR400-168- 3 - 390 1
SR,, [SRl000-168-3-370 1
Impeller rotation generates a low pressure region behind the
impeller blades. The extent of pressure reduction depends on
the kinetic head [ ( T N D ) ~ / At
~ ] a. certain impeller speed, the
reduction in pressure is sufficient to hold the gas phase in the
low pressure region, against the buoyancy force. The gas ac-
cumulated behind the impeller is referred to as a cavity. Cavity
formation starts somewhere at the point M , and its size in-
0
5 t
'
creases with increasing impeller speed along the curve MIN , .
c- 1.8-
At point N , , the cavity size reaches its maximum. The flow pat-
8 ,
tern under these conditions is shown in Fig. 4B. It is seen that
the impeller action distributes the gas bubbles in the radial
direction above the plane of the impeller. Because of cavity for-
mation, the intensity of eddy motion behind the impeller
decreases and the turbulent energy dissipation there is reduced.
Therefore, the impeller power consumption and thus the power
number decreases along the curve MIN , .

At an impeller speed corresponding to point N , , the shear stress

field generated by the impeller is sufficiently strong to break up
the cavity, starting from point N , . The flow pattern is shown
in Fig. 4C. At this stage, the impeller action starts contributing
to the bubble size. The gas bubbles now occupy more liquid
Fig. 3. Variation of power number with impeller speed in mechanically volume and the gas phase extends up to the impeller plane.
agitated gas-liquid contactors: effect of sparger location.
As the impeller speed increases along the curve N I P , ,the cavi-
is weak. Therefore, the gas bubbles generated at the sparger ty break-up increases continuously and is completed at point
rise vertically through the impeller without any disturbance as P , . As a result of continuous reduction of cavity size along the
shown in Fig. 4A. Under these conditions, the variation of the curve N I P , , the power number increases continuously. At im-
power number occurs along the curve L , M , in Fig. 3. peller speeds beyond point N , , the gas bubbles penetrate into
the region below the impeller plane and, at a certain critical
It is seen that the power number at a point such as L , is high, speed (point O , ) , they even reach the vessel bottom (Fig. 4D).
compared to that in the absence of gas. This is due to the liquid Under these conditions, the impeller action predominates over
circulation generated by gas sparging. As the impeller speed in- the effect of gas sparging. This means that both the average
creases along the curve L , N , , the flow generated by the im- bubble size and liquid flow in the vessel are controlled by the
peller increases continuously and provides increasing resistance impeller action. The speed at point 0, is referred to as the
to the gas flow. The gas hold-up increases steadily and the critical impeller speed for gas dispersion (NcD).
reduction in the power number occurs along the curve L , N , .
The flow pattern still remains as shown in Fig. 4A. It shows At impeller speeds beyond point P I , the liquid circulation
that the impeller action is not strong enough to effect a marked generated by the impeller increases continuously and the bub-
change in the flow behaviour of the gas phase. Since bubble bles become entrained in the downflow above the impeller. In-

(A) ( 8 ) (C) (Dl (El

N = N,, N = N,
Fig. 4. Flow patterns of gas-liquid dispersion at increasing impeller speeds (at constant Vc) for sparger close to impeller.
Chern. Eng. Technol. 14 (1991) 333-347 339

itially. only small bubbles are entrained by this downflow. peller. It has been observed that the range N 2 P 2 is wider in
However, at sufficiently high impeller speeds, larger bubbles the 1.5 m i.d. vessel than in the 0.57 m i.d. vessel. The
are also entrained. The corresponding flow pattern is shown in foregoing discussion leads to the conclusion that the cavity
Fig. 4E. Because of gas recirculation into the impeller region, size and the value of N,, depend on gas accessibility to the
the gas hold-up increases continuously along the curve P , Q,. impeller region. When the sparger is close to the impeller,
This results in a reduction of dispersion density in the impeller gas accessibility is better than for a sparger at a distance
region and the power number decreases along the curve P I Q,. from the impeller. Therefore, the gas cavity is larger and
This regime is called the recirculation region. N , , is higher in the former case.

4.2.2 Sparger Below and Distant from Impeller 4.2.3 Sparger Above Impeller

The sparger location in this case was 152 mm below the im- The flow pattern is shown in Fig. 6 and the power number
peller in the 0.57 m i.d. vessel and 390 mm below impeller in profile is represented by curve 3 in Fig. 3. The liquid circula-
the 1.5 ni i.d. vessel. A typical power number profile is il- tion generated by gas sparging is less intense because the rate
lustrated by curve 2 in Fig. 3 and the flow pattern is shown in of energy input is relatively small since the gas is sparged
Fig. 5 . The variation in the power number in this case was against a relatively small static head. Therefore, the initial in-
found to be similar to that when the sparger was close to the crease in the power number (Npo - N P ) is smaller than in the
impeller. This means that the flooding, cavity growth, cavity case of spargers below the impeller. At low impeller speeds,
break-up and recirculation regions appear in the impeller speed the gas leaving the sparger remains above the impeller and im-
ranges of L,M,, M 2 N 2 , N , P , and P2Q,, respectively. peller flooding does not occur. Therefore, the power number
However. there are some differences in the power consumption remains practically constant along the curve L,M,.
and gas dispersion patterns: these are,
The cavity formation starts at the point M,.At this speed, the
- when the sparger is distant from the impeller, the gas bub- downward liquid flow in the sparger region is sufficient to en-
bles also remain at a distance from the impeller. Therefore, train the gas to the impeller region. At speeds beyond point M,,
the reduction in the power consumption is in this case much the gas accessibility to the impeller region is greater when the
lower than when the sparger is close to the impeller. The gas sparger is above the impeller than below the impeller. For the
cavity is also small. former case, the sparged gas is entrained towards the impeller
while, in the latter case, the gas phase is diverted away from
- In case one. gas dispersion occurs mainly by break-up of the the impeller. As a result of good gas accessibility, the cavity
gas cavit) . However, when the sparger is further away, gas size and the value of N,, are larger. The cavity growth region
dispersion occurs in the region between the sparger and the ( M 3 N 3 )is larger than in the former two cases, namely M , N ,
impeller as shown in Fig. 5C. This happens at speeds and M 2 N 2 .
beyond the point N , . The condition of critical impeller
speed for gas dispersion ( N c D )is shown in Fig. 5D. In this It may be noted that, for a given impeller, the speed at which
case, N , , is lower than in case one. This is because the gas the cavity formation starts (M,, M,,M,)is more or less con-
is sparged close to the bottom of the vessel and gas disper- stant for the three discussed sparger locations.
sion occurs in the region below the impeller.

The range of impeller speeds in which cavity break-up oc- 4.2.4 Large Ring Sparger
curs is somewhat wider in case two than in case one. This
means that the rate of increase in power consumption is It was considered desirable to study the dispersion
gradual in case 2. This effect becomes more pronounced on characteristics of ring spargers with diameters larger than the
increase in the distance between the sparger and the im- impeller diameter. A typical flow pattern is shown in Fig. 7 and

Fig. 5. Flow patterns of gas-liquid dispersion at increasing impeller speeds (at constant Vc) for sparger distant from impeller
340 Chem. Eng. Technol. 14 (1991) 333 - 347

>
w Y w
IA (8) IC) ID) (€1

Fig. 6. Flow patterns of gas-liquid dispersion at increasing impeller speeds (at constant V,) for sparger located above impeller.

the variation of power number by curve 4 in Fig. 3 . In this case, ferent sparger designs in order to minimize this instability of
at low impeller speeds, the gas bubbles generated at the sparger power consumption during gas dispersion. Four types of
rise vertically without coming into contact with the impeller, as spargers, namely pipe, conical, ring and concentric ring
shown in Fig. 7A. As the impeller speed increases, gas disper- spargers were used in the 0.57 m i.d. vessel. The ring and the
sion occurs in the sparger region (Fig. 7 C ) and small bubbles concentric ring spargers were also used in the 1.5 m i.d. vessel.
rise upwards. At high superficial gas velocities, some gas is The variation of the power number (N,,) with impeller speed
diverted to the impeller region as shown in Fig. 7B. At a certain at a constant V, is shown in Fig. 8, for different types of
impeller speed, in excess of that at C, the gas bubbles appear spargers.
even below the sparger. This is the critical impeller speed for
gas dispersion (point 0, in Fig. 3 ) At still higher speeds, gas The plots of N,, vs N were found to be similar for all four types
recirculation begins and the power number decreases con- of sparger design. However, the speeds (at which flooding,
tinuously in the region 0, P, Q,. The flow pattern in the case cavity formation, cavity break-up, complete gas dispersion and
of this sparger is quite different from those generated by the recirculation of gas occur) were found to vary with sparger
other three spargers. When a large sparger ring is located type. These changes are due to the varying amounts of gas
below and at a distance from the impeller, the gas is not easily available in the impeller region for the different sparger
accessible to the impeller region. Therefore, the regions of designs, at a given impeller speed and V,.
cavity formation, cavity growth and cavity break-up are prac-
tically eliminated in the case of large ring spargers. The It is seen from Fig. 8 that, for the concentric ring sparger
behaviour of power consumption (curve 4)is also correspon- (SRR,), the sequential transition from flooding to recirculation
dingly different from that shown by the other three curves. occurs much sooner than for the other sparger types. The
power number at the end of the cavity formation zone (point N
in Fig. 2 ) for the concentric ring sparger is also the highest, in-
5 Results and Discussion dicating the smallest size of cavity. The jump in the power
number (A N p G ) in the vicinity of N,, is less steep than for the
5.1 Effect of Sparger Design other sparger designs.

The typical variation of power number with impeller speed is For conical (SC,) and ring (SR,) spargers, the critical impeller
shown in Fig. 2. The maximum instability in the power number speed required for complete gas dispersion is the same but
was found to occur in the vicinity of NCD.This is attributed to lower than that for the pipe sparger (SP,). It may be observed
changes in the hydrodynamic conditions of the impeller as that the power numbers are practically the same for all
discussed earlier. It was considered desirable to study the dif- spargers, after complete gas dispersion.

Fig. 7. Flow patterns of gas-liquid dispersion at increasing impeller speeds (at constant V,) for large ring sparger
Chem. Eng. Techno]. 14 (1991) 333-347 34 1

SYSTEM : GAS-LIQUID
plained by a corresponding increase of the fractional gas hold-

-
c3
1.8
up with V,. The other sparger designs also show a similar
behaviour.

5.3 Effect of Sparger Location

The effect of sparger location was studied for pipe spargers

(SP,, SP,) and ring spargers (SR,, SR,, SR,) in the 0.57 m i.d.
vessel and concentric ring spargers (SRR,, SRR,) in the 1.5 m

I
i.d. vessel. The plots of power number against impeller speed,
at constant V,, are presented in Figs 10 to 12.
SP, [SP60-14-3.6-1521

sc, [SC9B-ZB-2-95 I
These diagrams show that the variations of N p , with N at con-
: 0.8
0 l . 1
I 0 1SR2 l S R 9 5 - 6 - 3 - 1 5 2 1 11 stant VG but different sparger locations are similar. The rise in
the power number in the vicinity of N,, (A NPGin Fig. 2) was
1 I I J 1 I I I I I found to decrease with increasing distance between sparger and
z 3 4 5 6 7 8 9
IMPELLER S P E E D , N Ir / s I impeller. This behaviour was prominent in the case of the 1.5
m i.d. vessel. The value ANpGwas very low when the gas was
Fig. 8. Effect of sparger design on N,, - N relationship.
sparged from the top of the impeller (spargers SR,, SRR, in

r----~ ~____.~__
2,0- SYSTEM: GAS-LIQUID 5 I M 8 0 1 :OARGER ~ ~ ~ 1 0 ~
The superficial gas velocity (V,) was varied in the range from T = O . Z ~ ~ , O = C ~ T I ~ , W 0.3
/ O = SP? 1 5 P b O - l L - 3 . 6 - I S l I 6.8

1.5 to 30 mm/s. The variation of NpGwith N at different super- 9.6

15.0
ficial gas velocities for different sparger designs is shown in SP, ISPbO-I- 3.6-701 1.0

Fig. 9. It was observed that the impeller speed, at which max- F ,.K-
9.L
15.0
imum gas cavity (point N in Fig. 2) occurs, increases with in-
creasing superficial gas velocity. The corresponding power $ 1.4-

number decreases with increasing V,. This behaviour is due to 2

an increase in the amount of gas reaching the impeller zone 5

with increasing V,. This increases the cavity formation zone as

I 1.0-
well as the maximum gas cavity formed during the process of 2
gas dispersion. The large cavities formed explain the reduction o.O
of N,, at the onset of cavity break-up as V , increases. These 2'
large cavities require a higher shear rate for break-up which ex- 0.6 I
2 1 1 5 6 7 0 9 10

I-
l,bl 2 3 L 5 5 7 0 9

2.4

2- 2.2
w
1

Y)
;2.0 1.6

e
:1.8 l.L

Y)

Z I 2
1.2
-1.2
m w
LI
I
-1.0 2 $ 1.1 1.0

a
Lu
-0.8 1.2 0.0
1 c 0.57m ,O = C = l / 3 , W / D = O . 3 a
1.0
2 3 ' 0.6
IMPELLER S P E E D , N ( r l S ) IMPELLER SPEED, N l r 1 1 1

Fig. 9. Variation of power number with impeller speed for conical (SC,), Fig. 11. Effect of distance of ring and concentric ring spargers from im-
ring (SR,) and concentric ring (SRR,) spargers. peller on N , , - N relationship.
342 Chem. Eng. Technol. 14 (1991) 333 - 347

W
I-
>
m
VI 1.6-
e3
?
? 1.1-
VI
4,
z
- 1.2-
a
m
W
T
z 1.0-
a
W
%
:0.8-
I I I I I 1 I I I I I
2 3 L 5 6 7 8 9 10
IMPELLER SPEED, N I r f r l IMPELLER S P E E D , N l I / I l

Fig. 13. Effect of ring sparger diameter on N,, -N relationship

Fig. 12a. Variation of power number with impeller speed for sparger SR,.

diameter. For this geometric condition, the values of N C D are

also practically the same.

A distinctly different behaviour was observed in the case of

ring spai-ger 2 0 in diameter (SR,,). The power number was
found to decrease continuously with increasing impeller speed.
This behaviour was observed at all superficial gas velocities
(Fig. 14). In this case, gas accessibility to the impeller region
is difficult. Therefore, the regions of cavity formation, growth
and break-up are practically eliminated in the case of a large
ring sparger which prevents a rise in the power number.
Therefore, the speed required for gas dispersion was lower in
this case.

5.4.2 Number of Holes and Hole Size

IMPELLER SPEE0.N l r l i l As discussed earlier, the variations in the hole size and the
number of holes change the values of orifice Reynolds number
Fig. 12b. Variation of power number with impeller speed for spargers at constant V,, which leads to a change in the bubble size in
SRR, and SRR,. the liquid bulk. For this study, a ring sparger of diameter equal

Fig. 11). This behaviour was observed at all gas velocities

(Figs 12a and 12b) and is explained by the mechanism of gas
dispersion in section 4.

NCD was found to increase as the sparger approaches the im-

peller from below. The highest value of N C D was obtained
when the sparger was located above the impeller.

5.4 Effect of Ring Sparger Design

5.4.1 Ring Diameter

A PTD impeller with diameter of 1/3 the tank diameter and

b l a d e s 0 . 3 0 w i d e was u s e d in t h i s study. Ring spargers of outc'l
diameters 0.5D,0.80, 1 0 a n d 0.80, 2 0 were employed in tlir.
0.57 and 1.5 m i.d. vessels, respectively. The results are shoM I I
IMPELLER S P E E D , N I r / r l
in Fig. 13. The N,, versus N curves are similar for all thc
spargers when the ring diameter is smaller than the impeller Fig. 14. Variation of power number with impeller speed for sparger SR,,.
Chem. Eng. Technol. 14 (1991) 333-347 343

to 0 . 8 times the impeller diameter (D = T/3) was used in the

1.5 m i.d. vessel. The number of 3 mm holes was increased
(42, 84, 168). Holes of 2 mm (189 in number), 3 mm (84) and
6 mm (21) were used for studying the effect of hole size, at a
constant area for gas sparging. Since location of the sparger
plays an important role, these effects were studied at two dif-
ferent locations (100 mm and 390 mm from impeller). The
results at constant superficial gas velocity (V, = 9.4 mmis) are
shown in Figs 15 and 16.

The N,, versus N curves are similar for all the spargers. It can
be seen from Figs 15 and 16 that the effect of the number of
holes (spargers SR,,, SR,,, SR,,) and hole size (SR,,, SR,,,
SR,,) on the shape of the N,, versus N curve is negligible when
the sparger is nearer to the impeller. The differences were
0
found to be significant (Figs 15 and 16) when the sparger was a 1.2 I I I
1 2 3
distant from the impeller (SR,,, SR,,, SR,, and SR,,, SR,,, IMPELLER SPEED,Nlr/r)
SR,,). Similar behaviour was observed at all investigated
superficial gas velocities. If the gas is sparged close to the im- Fig. 16. Effect of orifice size on NPG - N relationship.
peller, the impeller action is stronger than the sparger action
which explains the negligible effect of the number of holes and
the hole size. However, if the sparger is located at a distance 5.5.1 Impeller Diameter
from the impeller, the sparger action becomes more important
in controlling the bubble size, as discussed earlier, resulting in The effect of impeller diameter on power consumption was
significant effects of the number and size of holes. studied in the 1.5 m i.d. vessel. Impellers of 0.33 m, 0.5 m and
0.75 m diameters, with a WID ratio of 0.3 were used for this
purpose. Ring spargers (SR,, SR,, SR,,, SR,,, SR,,, SR,,) of
5.5 Effect of P72) Impeller Design outer diameters equal to 0.8 times the impeller diameter were
employed at two locations for each impeller (100 mm and 390
Design details of the PTD impeller such as diameter, blade mm from the impeller). The variation of power number with
width and blade thickness were found to be important for mix- impeller speed at constant V, for different impeller diameters
ing [ 121 and solids suspension in solid-liquid [15] as well as is shown in Fig. 17.
gas-liquid-solid 122, 231 systems. Therefore, a systematic in-
vestigation of the effect of these design details on the power It was observed that the rise in the power number in the vicinity
consumption profile ( N p , - N curve) in gas-liquid systems was of N,, (ANpc in Fig. 2 ) decreases with increasing impeller
carried out in the 0.57, 1.0 and 1.5 m i.d. vessels. diameter. This was observed at all values of Vc (Figs 18 to 20).
It is seen from these variations that the cavity size (before NcD)
decreases with increasing impeller diameter. For an impeller,
0.75 m in diameter (TID = 2), the power number decreases
continuously with increasing impeller speed (at a constant V,).
3-0 This behaviour was observed at all superficial gas velocities as

2.8

2.6

Y
,2.0
E 1.L

m
T

= 1.1
Y

0
a T.0
+.6
I I I
1 i 3 4
0
IMPELLER S P E E D , N ( r / s )

Fig. 15. Effect of number of holes on N,,G - N relationship. Fig. 17. Effect of impcller diameter on N,, ~ N rebdtionship
344 Chem. Eng. Techno]. 14 (1991) 333 - 347

.-

2.‘.

* 2.2-
- A I
2.0

: I
1
SYSTEM:

z
GAS-LIQUID

I s m , 0 : 0.75rn ,W/O: 0.3 , C : T i 3

z
f 2.6’
I
w 2.0-
VI
r
VI

g 1.0-
3
z
u
1.6-
\
\ d
5
P; 1.4-
m SR) [SRZ80-21-6-lWl
I
3

= 1.2-
w
iz
0
SR, ISR280-21-6-3901
1.4 I I I I I I I I
a 1.0- 0.6 0.7 0.8 0.9 1.0 1.1 1.1 1.3 1.4 1.5
IMPELLER S P E E D , N l r l r l

0.8
Fig. 20. Variation of power number with impeller speed for Ti2 impeller.
IMPELLER S P E E D , N I r l r l

Fig. 18. Variation of power number with impeller speed for Ti4.5 impeller.

5.5.2 Blade Width

shown in Fig. 20. Similar observations were made by Raghava The WID ratio was varied from 0.25 to 0.4 in the 0.57 m i.d.
Rao and Joshi [16], for 0.33 rn diameter impeller in a 0.57 m and between 0.25 and 0.35 in the 1.O m i.d. vessel. An impeller
i.d. vessel. This behaviour can be explained by the mechanism of diameter TI3 was used. The variation of power number with
of cavity formation behind impeller blades. The cavities are impeller speed was similar for all the WID ratios (Fig. 21). It
formed because of a low pressure region there. The size of a was observed that the impeller speed, at which gas cavity starts
gas cavity depends on the extent of pressure reduction behind breaking up, decreases with increasing WID ratio and, hence,
the blade which, in turn, depends on the kinetic head the value of NCD also decreases. ANpGwas found to decrease
[a(ND)’12] of the impeller. The values of NcD for different im- slightly with increasing WID ratio.
peller diameters are given in Table 5. It was observed that the
value of NcD varies with D - 1.85. This means that the reduction
in NcD is considerably overproportional with respect to the in- 5.5.3 Blade Thickness
crease in the impeller diameter. This affects the kinetic head of
the impeller at NCD. The values of (NCDD)’ are given in Table The effect of blade thickness on the NPGvs N relationship was
5. It can be seen that, at an identical experimental condition of studied by using an impeller 0.19 m in diameter (WID = 0.35)
gas dispersion, the kinetic head decreases continuously with in- in the 0.57 m i.d. vessel. Thicknesses of 2.8 111111, 4.3 mm and
creasing impeller diameter. As a result, the reduction in 6 mm were investigated. In general, it was observed that the
pressure behind the impeller blade decreases and the cavity size tendency to cavity formation decreases with increasing blade
decreases with increasing impeller diameter. thickness (Fig. 22). In the impeller speed range between L and
N (Fig. 2 ) , the variation of power number was found to be
greater for thicker blades. It is interesting to note that the dif-
ferences in the power number, attributable to blade thickness,
1 _1 ,.n in the liquid alone (ordinate) are not very significant. This in-
dicates that the reduction in the power number on account of
cavity formation is not the same for different impeller
thicknesses. If we compare the power numbers at maximum gas

Table 5 . NcDD2 and (NCDD)’ values.

0.33 0.0048 4.5 0.49 2.205

0.33 0.0094 5.23 0.57 2.98
0.5 0.0048 2.13 0.533 1.134
0.5 0.0094 2.4 0.6 1.44
0.75 0.0048 0.85 0.478 0.406
0.75 0.0094 0.97 0.546 0.53
Fig. 19. Variation of power number with impeller speed for Ti3 impeller.
Chem. Eng. Technol. 14 (1991) 333-347 345

, jo.7
2 3 4 5 6 7 8 9 Fig. 23. Effect of tank diameter on N,, -N relationship.
IMPELLER SPEED,NIrIrl

Fig. 21. Effect of blade width on N,, -N relationship.

discussed in a previous section, the increase in the power
cavity (point N in Fig. 2), the power number increases with in- number in the vicinity of NCD (ANpG)was found to depend on
creasing blade thickness. This indicates that the maximum gas the actual distance between the sparger and the impeller and to
cavity formed behind the blade also depends strongly on blade be independent of the vessel size. Comparing the behaviour of
thickness, decreasing with its increase. Consequently, the value spargers distant from the impeller (Fig. 23, 152 mm in 0.57 m
of N c , was found to decrease with increasing blade thickness. i.d. vessel and 390 mm in 1.5 m i.d. vessel), it is seen that the
increase in N P G is much smaller in the case of the 1.5 m i.d.
At impeller speeds beyond point P , the power number remain- vessel. This indicates that the actual distance between the
ed practically constant, with higher values occurring with sparger and the impeller is an important parameter for the
thicker blades. This indicates that the tendency of the gas-liquid variation of power number with impeller speed. However, on
mixture to recirculate (which is the cause of the decrease in placing the sparger at constant distances from the impeller (100
N p G ) above N , , decreases with increasing impeller blade mm in 0.57 m i.d. and 1.5 m i.d. vessels), the increase in the
thickness. Similar behaviour was observed in the case of gas- power number (ANpG) was similar in both cases (Fig. 23).
liquid-solid systems.
6 Correlations
5.6 Effect of Vessel Size
As discussed earlier, the dependence of the power number on
Fig. 23 shows that the variation of the power number with im- impeller speed and superficial gas velocity is complex when a
peller speed is similar in the 0.57 and 1.5 m i.d. vessels. As wide range of speeds is considered. However, in practice, the
desired impeller speed is above NCD so as to ensure effective
gas-liquid contacting. Therefore, it was considered desirable to
develop correlations for power consumption in the range of
speeds in excess of N,, (320 data points).

T = 0.57, 1.0, 1.5 m, 0.22 IDIT 5 0.5,0.25 i WID 5 0.4,

CIT = 0.3, k = 2.8 mm, 1.5 I VG I 30 mrnls.

PG
-

P
= 0.434 (2)"
(T) (7)(T) N 2 0 -8 W 0.174 D -0.64
, (2)

5 6 9
3 4
IMPELLER SPEED. N I r l r l
7 8
S.D = 6.2% ,

Fig. 22. Effect o f impeller blade thickness on N , , -N relationship where, a = 0.064 T-0.94 and fl = 0.148 T-1.42
346 Chem. Eng. Technol. I4 (1991) 333 -347

PG
-
P
= 1.324 -0.71 (s)(F) -0.2 +
The research was supported by a grant under the Indo-US Col-
laborative Materials Science Program (CE- 1).

+ 0.35 (+) - 1.26 ($) , (3) Received: October 12, 1990 [CET 3281

S.D = 5.5%
Symbols used
The above correlations are expressed in terms of dimensionless
numbers. The following dimensional correlation was also ob- impeller clearance
tained: bubble diameter in the bulk of liquid
hole diameter
impeller diameter
gravitational acceleration
distance between sparger and impeller
where y = 0.05 T - ’ . 5 6 .
number of sparger holes
impeller rotational speed
The correlation coefficient was 0.996 and the standard devia- critical impeller speed for gas dispersion
tion was 4 % . power number, P l ( e L N 30’)
power number in presence of gas, P G / ( e L N 30’)
The exponents on V, were found to be - 0.1, and - 0.027 i n increase in N,, shown in Fig. 2
0.57 and 1.5 m i.d. vessels, respectively, which are lower thaii power consumption
those observed by Mihel and Miller [22], Luong and Volesk! power consumption in presence of gas
[23], Hassan and Robinson [lo] and Hughmark [24]. These in- gas flow rate
vestigators determined the exponent as - 0.24, - 0.38, orifice Reynolds number (dh Vhp,ip,)
- 0.38, and - 0.25, respectively. This difference is attributed
tank diameter
bubble rise velocity in MAC
to the fact that these workers covered the entire range of im-
bubble rise velocity
peller speeds, including those below NcD. At low speeds, par-
liquid circulation velocity in bulk
ticularly below NcD,the role of V, is prominent. We were of
superficial gas velocity
the opinion that, as already mentioned, gas-liquid contacting gas hole velocity
becomes attractive only at speeds above N,, and, accordingly. blade width
Eqs (2) and (3) were developed.

7 Conclusions
Greek symbols
The variation of the power number with impeller speed
(NpG vs N) at constant V, can be explained on the basis of Q [-I fractional gas hold-up
corresponding visual observations on hydrodynamics 6 [degree] blade pitch
around the impeller. PG [kgims] gas viscosity
NPG vs N was found to depend on superficial gas velocity, ec [kgini’] gas density
sparger design, sparger location and PTD impeller design. eL [kdm’l liquid density
Hole size and number of holes in the ring sparger exert a
negligible effect on the Np, - N relationship when the
References
sparger is located near the impeller. However, these
variables become important when the sparger is at a distance
[I] Joshi, J.B., Pandit, A.B., Sharma, M.M., Chem. Eng Sci. 37 ( I 982)
from the impeller. pp. 813-844.
The increase of N,, in the vicinity of NcD(ANpG)was found [2] Pandit,A.B..Joshi,J.B., Chem. EngSci. 38(1983)pp. 1189- 1215.
to decrease with increasing PTD impeller diameter (at con- [3] Midoux, N . . Charpentier, J.C.. Int. Chem. Eng 24 (1984) pp.
stant WiD). 249 - 287.
At impeller speeds in excess of N,,, the power number re- [4] Warmoeskerken, M.M.C.G., Speur, J., Smith, J.M., Chem. Eng
Cummun. 25 (1984) pp. I 1 - 29.
mains practically constant, in the presence of thicker blades.
[5] Wamoeskerken, M.M.C.G., Smith, J . M . , Chem. Eng Sci. 40 (1985)
The NPG- N curves were found to depend on the absolute pp. 2063 - 2071.
distance between the sparger and the impeller and, at the [6] Frijlink, J.J, Kalijn, M . , Smitg, J.M., Insr. Chem. Eng Symp. Ser.
same distances, the effect of size was found to be negligible. (1984) NO. 89, pp. 49-58.
Correlations have been proposed for the power consumption [7] Breucker, Ch.. Steiff, A., Weinspach, P.M., Proc. of 6th European
in the presence of gas. Conference on Mixing BHRA, Cranfield, 1988, pp. 399 -406.
[8] Bruijn, W . , Riet, K.V., Smith, J.M., Trans. Inst. Chem. Eng 52
(1974) pp. 88- 104.
[9] Botton, R . , Cosserat, D., Charpentier, J.C., Chem. Eng Sci. 35
Acknowledgement ( 1 9801 ..
DD. 82 - 89.
[lo] Hassan, I.T.M., Robinson, C.W., AIChEJ. 23 (1977) pp. 48-56.
We are grateful to the University Grants Commission of the [ l l ] Shiue, S . J . , Wong, C . W . , Can. J . Chem. Eng 62 (1984) pp.
Government of India for the award of a fellowship to V.B.R. 602 - 609.
Chem. Eng. Techno]. 14 (1991) 347-352 347

[I21 Rewatkar. V.B.. Joshi, J.B., Chem. Eng Commun. 91 (1991)pp. [IS] Ranade, V.V., Joshi, J.B., Chem. Eng Commun. 81 (1989)pp.
322 353.
- 224 - 247.
[I31 Oldshue, J.Y.. Fluid Mixing Technology, McGraw-Hill Co., New [I91 Rewatkar, V.B.. Raghava Rao, K.S.M.S.. Joshi. J.B., Chem. Eng
York 1984. Commun. 88 (1990)pp. 69-90.
[I41 Gray. J.B.. Oldshue, J.Y., in: Mixing-Theor?; andPrucrice, Vol. I11 [20] Rewatkar, V.B., Joshi, J.B., Ind. Eng Chem. Res. 30 (1991)pp.
(Uhl, J.B. and Gray, V.W., eds.), Academic Press, New York 1986, 51 1 - 526.
pp. 1-61 [21] Rewatkar, V.B., Joshi, J.B.. Ind. Eng Chem. Res. 30 (1991)pp.
[ 151 Raghava Rao. K.S.M.S., Rewatkar, V.B.. Joshi, J.B., AlChE J . 34 527 - 534.
(1988)pp. 1332 - 1340. [22] Michel, B.J.. Miller, S.A., AlChE J . 8 (1982)pp. 262-266.
[I61 Raghava Rao. K.S.M.S., Joshi, J.B., lnt. J . Eng Fluid Mech. 2 [23] Luong, H . T . , Volesky, B., AlChE J . 25 (1979)pp. 893 895.
-

(1989)pp. 645-663. [24] Hughmark. G.A.. Ind. Eng Chem. Process, Des. Det,. 19 (1977)pp.
1171 Ranadc. V . V . . Joshi, J.B.. Chetn. Eng Commun. 81 (1989)pp. 638 641.
-

197-224.

Theoretical Analysis of Gas Stripping Performance of Spray Towers

Shafkat A. Beg, Syed Obaidur Rehman and Mirza M. Hassan"

A simplified model has been developed to describe ammonia stripping process in spray towers.
The effect of various operating parameters including air flow rate, contact time, droplet size, li-
quid to gas ratios and tower height on the performance of such systems has been studied in detail.
The results obtained for ammonia stripping have been compared with the limited experimental
data available in literature. The predicted results agree reasonably well with the reported ex-
perimental data.

1 Introduction point chlorination, selective ion exchange and air stripping.

Among the various biological and physico-chemical processes,
Among the various forms of nitrogen (ammonia-nitrogen, air stripping is considered to be the simplest and the most
organic nitrogen, nitrates and nitrites), removal of ammonia- economical process with a high efficiency of removal and no
nitrogen is at present receiving considerable attention because solid or liquid wastes to be disposed of.
of the various deleterious effects associated with its presence in
water. These include eutrophication, dissolved oxygen deple- The most widely used air-water contactors, employed by the
tion, reduction in chlorine disinfectant power and toxicity to water industry are packed columns. Various research workers
fish and other aquatic species. studied ammonia stripping in packed columns on bench, pilot
and commercial scales [I]. However, various operational pro-
Various physico-chemical and biological processes are avail- blems associated with ammonia stripping in packed columns
able for ammonia-nitrogen removal. The biological processes have been reported. These include calcium carbonate scale
include nitrification and combined nitrification-denitrification deposition on packings, resulting from dissolved calcium com-
processes, while the physico-chemical processes include break- pounds reacting with carbon dioxide in the air passing through.
In addition, biological oxidation of ammonia to nitrate may also
..
..
occur in such columns.
* Prof. Dr. S.A. Beg. S . Obaidur Rehmann and Dr. M.M. Hassan,
Department o f Chemical Engineering, King Fahd University of Therefore, spray towers are at present the most extensively us-
Petroleum & Minerals. Dhahran 3 126liSaudi Arabia. ed systems for ammonia stripping. The severity of problems

0 VCH Verlag\ge\ellschaft mbH, D-6940 Weinhcim, 1991 0930-7516/91/0510-0347 $03.50+.25/0