10 Mezclado y Agitacion de LIquidos
MIXING ANO AGITATION
gitation is a means whereby mixing of phases can be
A accomplished and by which mass and heat transfer
can be enhanced between phases or with externa/
surfaces. In its most general sense, the process of
mixing is concerned with al/ combinations of phases of which
the most frequently occurring ones are
1. gases with gases.
2. gases into /iquids: dispersion.
3. gases with granular so/ids: f/uidization, pneumatic
conveying, drying.
4. /iquids into gases: spraying and atomization.
5. /iquids with liquids: dissolution, emulsification, dispersion.
6. liquids with granular so/ids: suspension.
7. pastes with each other and with solids.
B. solids with solids: mixing of powders.
lnteraction of gases, liquids, and solids a/so may take place,
as in hydrogenation of liquids in the presence of a slurried
so/id catalyst where the gas must be dispersed as bubbles
and the so/id particles must be kept in suspension.
Three of the processes invo/ving liquids, numbers 2, 5,
10.1. A BASIC STIRRED TANK DESIGN
The dirnensions of the liquid content of a vessel and the dirnensions
and arrangernent of irnpellers, baffles and other internals are factors
that influence the arnount of energy required far achieving a needed
arnount of agitation or quality of rnixing. The interna! arrangernents
depend on the objectives of the operation: whether it is to rnaintain
homogeneity of a reacting mixture or to keep a salid suspended or a
gas dispersed or to enhance heat ar mass transfer. A basic range of
design factors, however, can be defined to cover the majority of
cases, far example as in Figure 10.1.
THE VESSEL
A dished bottom requires less power than a flat one. When a single
impeller is to be used, a liquid leve! equal to the diarneter is
optimum, with the irnpeller located at the center far an all-liquid
system. Econornic and manufacturing considerations, however,
often dictate higher ratios of depth to diarneter.
BAFFLES
Except at very high Reynolds numbers, baffles are needed to
prevent vortexing and rotation of the liquid mass as a whole. A
baffle width one-twelfth the tank diameter, w = D,/12; a length
extending frorn one half the irnpeller diameter, d /2, frorn the
tangent line at the bottom to the liquid leve!, but sometimes
terminated just above the leve! of the eye of the upperrnost
impeller. When solids are present or when a heat transfer jacket is
used, the baffles are offset from the wall a distance equal to one-
sixth the baffle width. Four radial baffles at equal spacing are
standard; six are only slightly more effective, and three appreciably
less so. When the mixer shaft is located off center (one-faurth to
287
and 6, employ the sarne kind of equipment; narnely, tanks in
which the liquid is circu/ated and subjected to a certain
amount of shear. This kind of equipment has been studied
rnost extensively. Although sorne unusual cases of liquid
mixing may require pilot plant testing, general rules have been
developed with which mixing equipment can be designed
somewhat satisfactorily. This tapie will be emphasized in this
chapter.
The other mixing operations of the /ist require individua/
kinds of equipment whose design in sorne cases is less
quantified and is based largely on experience and pilot plant
work. Typical equipment far such purposes will be illustrated
later in this chapter. Phase mixing equipment which
accomplishes prirnarily mass transfer between phases, such
as distillation and extraction towers, a/so are covered
elsewhere. Stirred reactors are discussed in Chapter 17.
Circulation and shear of the liquid in a vessel can be
accomplished with externa/ pumps and appropriate location of
suction and discharge nozz/es, but a satisfactory combination
of vertical and lateral f/ows is obtained more economical/y by
interna/ impellers, baffles, and draft tubes. Sorne general
statements about dirnensions, proportions, and internals of a
liquid mixing vessel can be made.
one-half the tank radius), the resulting flow pattern has less swirl,
and baffles rnay not be needed, particularly at low viscosities.
DRAFT TUBES
A draft tube is a cylindrical housing around and slightly larger in
diarneter than the impeller. lts height may be little more than the
diarneter of the impeller or it may extend the full depth of the
liquid, depending on the flow pattern that is required. Usually draft
tubes are used with axial irnpellers to direct suction and discharge
streams. An irnpeller-draft tube system behaves as an axial flow
pump of somewhat low efficiency. lts top to bottorn circulation
behavior is of particular value in deep tanks far suspension of solids
and far dispersion of gases. About a dozen applications are
illustrated by Sterbacek and Tausk (1965, pp. 283ff) and a chapter is
devoted to their use by Oldshue (1983, 469ff).
IMPELLER TYPES
A basic classification is into those that circulate the liquid axially
and those that achieve prirnarily radial circulation. Sorne of the
many shapes that are being used will be described shortly.
IMPELLER SIZE
This depends on the kind of impeller and operating conditions
described by the Reynolds, Fraude, and Power numbers as well as
individual characteristics whose effects have been correlated. Far
the popular turbine impeller, the ratio of diameters of impeller and
vessel falls in the range, d / D, = 0.3-0.6, the lower values at high
rprn, in gas dispersion, far exarnple.
 288 MIXING AND AGITATION

Side entering propellors are placed 18-24 in. above a flat tank
Baffle width, floor with the shaft horizontal and at a 10º horizontal angle with the
centerline of the tank; such mixers are used only for viscosities
w = D,I 12
below 500 cP or so.
Offset = w / 6 Draft tube In dispersing gases, the gas should be fed directly below the
impeller or at the periphery of the impeller. Such arrangements also
are desirable for mixing liquids.
H/ 3

t
H/ 2
H
10.2. KINDS OF IMPELLERS

A rotating impeller in a fluid imparts flow and shear to it, the shear
resulting from the flow of one portion of the fluid past another.
Limiting cases of flow are in the axial or radial directions so that
impellers are classified conveniently according to which of these

l
flows is dominan!. By reason of reflections from vessel surfaces and
obstruction by baffies and other internals, however, flow patterns in
most cases are mixed. When a close approach to axial flow is
particularly desirable, as for suspension of the solids of a slurry, the
Offset impeller may be housed in a draft tube; and when radial flow is
needed, a shrouded turbine consisting of a rotor and a stator may
=d / 2
T
H/ 6
be employed.
Because the performance of a particular shape of impeller
usually cannot be predicted quantitatively, impeller design is largely
an exercise of judgment so a considerable variety has been put forth
.¡... . -------0, --------..i by various manufacturers. A few common types are illustrated on
Figure 10.2 and are described as follows:
Figure 10.1. A basic stirred tank design, not to scale, showing a
lower radial impeller and an upper axial impeller housed in a draft a. The three-bladed mixing propeller is modelled on the marine
tube. Four equally spaced baffies are standard. H = height of liquid propeller but has a pitch selected for maximum turbulence. They
leve!, D, = tank diameter, d = impeller diameter. For radial are used at relatively high speeds (up to 1800 rpm) with low
impellers, 0.3 ::s d / D,::s 0.6.
viscosity fluids, up to about 4000 cP. Many versions are avail-
able: with cutout or perforated blades for shredding and breaking
up lumps, with sawtooth edges as on Figure 10.2(g) for cutting
IMPELLER SPEED
and tearing action, and with other than three blades. The
With commercially available motors and speed reducers, standard stabilizing ring shown in the illustration sometimes is included to
speeds are 37, 45, 56, 68, 84, 100, 125, 155, 190, and 320 rpm. minimize shaft flutter and vibration particularly at low liquid
Power requirements usually are not great enough to justify the use levels.
of continously adjustable steam turbine drives. Two-speed drives b. The turbine with flat vertical blades extending to the shaft is
may be required when starting torques are high, as with a settled suited to the vast majority of mixing duties up to 100,000 cP or
slurry. so at high pumping capacity. The simple geometry of this design
and of the turbines of Figures 10.2(c) and (d) has inspired
IMPELLER LOCATION extensive testing so that prediction of their performance is on a
more rational basis than that of any other kind of impeller.
Expert opinions differ somewhat on this factor. As a first c. The horizontal plate to which the impeller blades of this turbine
approximation, the impeller can be placed at 1/6 the liquid leve! off are attached has a stabilizing effect. Backward curved blades
the bottom. In sorne cases there is provision for changing the may be used for the same reason as for type e.
position of the impeller on the shaft. For off-bottom suspension of d. Turbine with blades are inclined 45º (usually). Constructions
solids, an impeller location of 1/3 the impeller diameter off the with two to eight blades are used, six being most common.
bottom may be satisfactory. Criteria developed by Dickey (1984) Combined axial and radial flow are achieved. Especially effective
are based on the viscosity of the liquid and the ratio of the liquid for heat exchange with vessel walls or interna! coils.
depth to the tank diameter, h/D,. Whether one or two impellers are e. Curved blade turbines effectively disperse fibrous materials
needed and their distances above the bottom of the tank are without fouling. The swept back blades have a lower starting
identified in this table: torque than straight ones, which is importan! when starting up
settled slurries.
Maximum lmpeller Clearance f. Shrouded turbines consisting of a rotor and a stator ensure a
Viscosity level Number of
[cP (Pa sec)) h/ D, lmpellers Lower Upper high degree of radial flow and shearing action, and are well
adapted to emulsification and dispersion.
<25,000 (<25) 1.4
g. Flat plate impellers with sawtooth edges- are suited to emul-
h/3
<25,000 (<25) 2.1 2 D,/3 (2/3)h sification and dispersion. Since the shearing action is localized,
>25,000 (>25) 0.8 1 h/3 baffies are not required. Propellers and turbines also are sometimes
>25,000 (>25) 1.6 2 D,/3 (2/3)h provided with sawtooth edges to improve shear.
h. Cage beaters impart a cutting and beating action. Usually they are
Another rule is that a second impeller is needed when the liquid mounted on the same shaft with a standard propeller. More violent
must travel more than 4 ft before deflection. action may be obtained with spined blades.
 (e)

(a) (b)

(e) (f )
(d)

(i)

(g) (h)

(j) (k) (1)

Figure 10.2. Representative kinds of impellers (descriptions in the text).

289
290 MIXING ANO AGITATION

i. Anchor paddles fit the contour of the container, prevent sticking of

pasty materials, and promote good heat transfer with the wall.
j. Gatepaddles are used inwide, shallow tanks and for materials of high
viscosity when low shear is adequate. Shaft speeds are low. Sorne §
designs include hinged scrapers to clean the sides and bottom of the 200Htttr-\-"d-+++f+itf---t-++-t+t+H---1'-+++++H+---I-...¡....¡
tank. \
k. Hollow shaft and hollow impeller assemblies are operated at high tip
speeds for recirculating gases. The gas enters the shaft above the
liquid leve! and is expelled centrifugally at the impeller. Circulation .e,
!
1
"'
ñ 20Hittt-+-t-+tti-ttf+-t++H+lt+-t+ l +1+-I 11 1 -+-+.i+
rates are relatively low, but satisfactory for sorne hydrogenations for '- i-.. P1tched-blad· turbinas
instance.
l. This arrangement of a shrouded screw impeller and heat exchange
coi! for viscous liquids is perhaps representative of the many designs
that serve special applications in chemical processing.
l
¡ 150
E
Ó
1 "'
2 Sou"e Aef 12 1 11111
1 1 11 1
1 l 1 1 111
• •110 111
..
1 1 1 1111
1
· · !ll
1 11

J 11111

50 1()2 103 10' 10• 10•

10.3. CHARACTERIZATION OF MIXING QUALITY
Agitation and mixing may be performed with severa! objectives:
l. Blending of miscible liquids.
2. Dispersion of immiscible liquids.
3. Dispersion of gases in liquids.
4. Suspension of solid particles in a s!urry.
5. Enhancement of heat exchange between the fluid and the
boundary of a container.
6. Enhancement of mass transfer between dispersed phases.
When the ultimate objective of these operations is the carrying out
of a chemical reaction, the achieved specific rate is a suitable
measure of the quality of the mixing. Similarly the achieved heat
transfer or mass transfer coefficients are measures of their
respective operations. These aspects of the subject are covered in
other appropriate sections of this book. Here other criteria will be
considered.
The uniformity of a multiphase mixture can be measured by
sampling of severa! regions in the agitated mixture. The time to
bring composition or sorne property within a specified range (say
within 95 or 99% of uniformity) or spread in values-which is the
blend time-may be taken as a measure of mixing performance.
Various kinds of tracer techniques may be employed, for example:
l. A dye is introduced and the time for attainment of uniform color
is noted.
2. A concentrated salt solution is added as tracer and the measured
electrical conductivity tells when the composition is uniform.
3. The color change of an indicator when neutralization is complete
when injection of an acid or base tracer is employed.
4. The residence time distribution is measured by monitoring the
outlet concentration of an inert tracer that can be analyzed for
accuracy. The shape of response curve is compared with that of a
thoroughly (ideally) mixed tank.
Reynolds number. D 2 Nplµ
Figure 10.3. Dimensionless blend time as a function of Reynolds
number for pitched turbine impellers with six blades whose
WI D = 1/5.66 [Dickey and Fenic, Chem. Eng. 145, (5Jan. 1976)].
vessels of different sizes and liquids of various viscosities. A review
of the literature on blend times with turbine impellers has been
made by Brennan and Lehrer [Trans. lnst. Chem. Eng. 54, 139-152
(1975)], who also did sorne work in the range 104 < NRe < 105 but
did not achieve a particularly useable correlation.
An impeller in a tank functions as a pump that delivers a
certain volumetric rate at each rotational speed and corresponding
power input. The power input is infiuenced also by the geometry of
the equipment and the properties of the fluid. The flow pattern and
the degree of turbulence are key aspects of the quality of mixing.
Basic impeller actions are either axial or radial, but, as Figure 10.4
shows, radial action results in sorne axial movement by reason of
deflection from the vessel walls and baffies. Baffies contribute to
turbulence by preventing swirl of the contents as a whole and
elimination of vortexes; offset location of the impeller has similar
effects but on a reduced scale.
Power input and other factors are interrelated in terms of
certain dimensionless groups. The most pertinent ones are, in
common units:
2
NRe = 10.75Nd S / µ, Reynolds number, (10.1)
3
Np= l.523(1013)P/N d5S, Power number, (10.2)
NQ = l.037( 105)Q / Nd 3, Flow number, (10.3)
tbN, Dimensionless blend time, (10.4)
TABLE 10.1. Blending Data for Four-Bladed 45º Turbines ª
5000 gal 10.000 gal 20.000 gal

The last of these methods has been applíed particularly to 12 JO 12 JO 12 JO

chemical reaction vessels. It is covered in detail in Chapter 17. In
7!
most cases, however, the RTDs have not been correlated with
impeller characteristics or other mixing parameters. Largely this I! 10 7!

also is true of most mixing investigations, but Figure 10.3 is an 7! 15 JO

uncommon example of correlation of blend time in terms of
I! , 15 7! 20 15 7j
Reynolds number for the popular pitched blade turbine impeller.
As expected, the blend time Ievels off beyond a certain mixing ªMotor horsepowers for various batch volumes, viscosities in cP,
intensity, in this case beyond Reynolds numbers of 30,000 or so. blend times in minutes.
The acid-base indicator technique was used. Other details of the •cenotes single four-bladed, 45º axial-flow impeller (unshaded
selections).
test work and the scatter of the data are not revealed in the t Denotes portable geardrive mixer with single 1.5-pitch propeller
published information. Another practica! solution of the problem is ("shaded" selections).
typified by Table 10.1 which relates blend time to power input to (Oldshue, 1983, p. 91).
 10.3. CHARACTERIZATJON OF MIXJNG OUALITY 291

LIQUID
LEV EL

a b d

Figure 10.4. Agitator flow patterns. (a) Axial or radial impellers without baffles produce vortexes. (b) Offcenter location reduces the vortex.
(c) Axial impeller with baffles. (d) Radial impeller with baffles.

Froude number, (10.5) 100.--r,<T""TT'T--r--r-rr-r--r-TT Tr-,----,--T7N pV7s' .-;N;-R-e-;F:nO.R !"-T+lTTl

d = impeller diameter (in.), A-4 IMPELLER ¡:::::¡:::¡
D = vessel diameter (in.),
N = rpm of impeller shaft,
P = horsepower input,
Q = volumetric pumping rate (cuft/sec),
S = specific gravity,
tb = blend time (min),
µ = viscosity (cP).

The Froude number is pertinent when gravitational effects are

significant, as in vortex farmation; in baffled tanks its influence is
hardly detectable. The power, flow, and blend time numbers change
with Reynolds numbers in the low range, but tend to level off above
NR. = 10,000 or so at values characteristic of the kind of impeller.
Sometimes impellers are characterized by their limiting NP , as an
NP = 1.37 of a turbine, far instance. The dependencies on Reynolds
number are shown on Figures 10.5 and 10.6 far power, in Figure
10.3 far flow and in Figure 10.7 far blend time.
Rough rules far mixing quality can be based on correlations of
power input and pumping rate when the agitation system is
otherwise properly designed with a suitable impeller (predominantly
either axial or radial depending on the process) in a correct
location, with appropriate baffling and the correct shape of vessel.
The power input per unit volume or the superficial linear velocity
can be used as measures of mixing intensity. For continuous flow
reactors, far instance, a rule of thumb is that the contents of the
vessel should be turned over in 5-10% of the residence time.
Specifications of superficial linear velocities far different kinds of
operations are stated later in this chapter. For baffled turbine
agitation of reactors, power inputs and impeller tip speeds such as
10 100 1000 10000 100000 1000000
REYNOLOS NUMBER
(a)
3 5
Figure 10.5. Power number, NP = Pgcf N D p, against Reynolds
number, NRe = ND 2p /µ, far severa! kinds of impellers: (a) helical
shape (Oldshue, 1983); (b) anchor shape (Oldshue, 1983); (c)
severa! shapes: (1) propeller, pitch equalling diameter, without
baffles; (2) propeller, s = d, faur baffles; (3) propeller, s = 2d,
without baffles; (4) propeller, s = 2d, faur baffles; (5) turbine
impeller, six straight blades, without baffles; (6) turbine impeller,
six blades, faur baffles; (7) turbine impeller, six curved blades, faur
baffles; (8) arrowhead turbine, faur baffles; (9) turbine impeller,
inclined curved blades, faur baffles; (10) two-blade paddle, faur
baffles; (11) turbine impeller, six blades, faur baffles; (12) turbine
impeller with stator ring; (13) paddle without baffles (data of Miller
and Mann); (14) paddle without baffles (data of White and
Summerfard). Ali baffles are of width O.lD [after Rushton, Costich,
and Everett, Chem. Eng. Prog. 46(9), 467 ( 1950)].
 292 MIXING ANO AGITATION

..._.i -'-_,_-'-
1 1 1
i 1 1 1
1 1
0.01 '-'--1'--'-.W.... '--"'-'--'--'-'
1

1 10 100 1000 10000 100000 1000000

REYNOLDS NUMBER Re
(b) (e)
Figure 10.5-(continued )
incorporation of TEL into gasoline where severa! hours may be
the following may serve as rough guides:
allowed for the operation.
Operation HP/1000 gal Tip Speed (ft/sec) Example 10.1 deals with the design and performance of an
agitation system to which the power input is specified. Sorne degree
Blending 0.2-0.5
of consistency is found between the severa! rules that have been
Homogeneous reaction 0.5-1.5 7.5-10
Reaction with heat transfer 1.5-5.0 10-15 cited.
Liquid-liquid mixtures 5 15-20
Liquid-gas mixtures 5-10 15-20 10.4. POWER CONSUMPTION ANO PUMPING RATE
Slurries 10
These basic characteristics of agitation systems are of paramount
The low figure shown for blending is for operations such as importance and have been investigated extensively. The literature is

DI SK & V ERT 1 CAL DI SK & VERTI CAL CURVEO P ITCH ED

BLADE BLADE BLADE BLADE BLADE BLADE
CURVE 1 CURVE 2 CLAVE 3 CURVE 4 CURVE 5 CURVE 6

100

50
-'"
, '"' 1
1

•+
* * r:dfn
* 1
o sti a
{l
*
1
'
1 1 1
w1ol 11a
e::.._
IJ\
w/0• 115 W/0 •1/8 W/0• 1/8
o
""
'u
:z
10 \...tl;-.. t ! 11 1

Ol
n. J t. -,

5
1 ........... -- -- 1 1 ' 1 'X
2! '' 1

:za. 3
-- ........ 1
1
1
1
1 1
'j' 1 '

i """"C
....
I.._:::: ..,¡,,1 1 · 1 1 1 1 1 11 1 1 1 1
-
-.........
4
1 1 1 1

1 1 1 11
-
' 1

ff-¡
"6
" 1

1 1
1
11
i TT ftttit
1 : 1 1 1 ¡1
1
1 1 1
•
..
 1
1 1 1 1 I'1 1 ; i 1 1
0.5
1 10

Figure 10.6. Power number against Reynolds number of sorne turbine impellers [Bates, Fondy, and Corpstein, Ind. Eng. Chem. Process.
Des. Dev. 2(4) 311 (1963)].
 10.4. POWER CONSUMPTION AND PUMPING RATE 293
1.0 1 1 1 11 11 1
Type No. baffles IV,, N.:i
0.9 DI T ' 025
1
Propeller o 0.3
08 Propeller 3-8 0.33-0.37 0.40-0.55
V 033
07 4 Tu rbine, vertical blade o 0.93-1.08 0.33-0.34
"o
' ..... 0.5 Tu rbine, vertical blade 4 3-5 0.70-0.85
lf v- Pitched turbine, 45º o 0.7 0.3
z 0.6 Pitched turbine, 45º 4 1.30-1.40 0.60-0.87
o 05
"0
)I Anchor o 0.28
z /1
D 04
/,'I A correlation of pumping rate of pitched turbines is shown as
E Vjf
:J
e

0.3 - -"' lf
Figure 10.7.
Power input per unit volume as a measure of mixing intensity
or quality was cited in Section 10.3 and in Chapter 17. From the
:J
_ .. correlations cited in this section, it is clear that power input and

-
oa...
0.2
10 20 50 100 200 500 1()().'.) 10,000
2 superficial linear velocity-the volumetric pumping rate per unit
Reynolds number, NR• 'D Np/fL
Figure 10.7. Flow number as a function of impeller Reynolds
number for a pitched blade turbine with NP = 1.37. D / T is the ratio
of impeller and tank diameters. [Dickey, 1984, 12, 7; Chem. Eng.,
102-110 ( 26 Apr. 1976)].
reviewed, for example, by Oldshue (1983, pp. 155-191), Uh! and
Gray (1966, Vol. 1), and Nagata (1975). Among the effects studied
are those of type and dimensions and locations of impellers,
numbers and sizes of baffies, and dimensions of the vessel. A few of
the data are summarized on Figures 10.5-10.7. Often it is
convenient to characterize impeller performance by single numbers;
suitable ones are the limiting values of the power and fiow numbers
at high Reynolds numbers, above 10,000-30,000 or so, for example:
Reynolds number together determine also the pumping rate of a
given design of impeller. This fact has been made the basis of a
method of agitator system design by the staff of Chemineer. The
100,000
cross section of the tank-is adopted as a measure of quality of
mixing. Table 10.2 relates the velocity to performance of three main
categories of mixing: mixing of liquids, suspension of solids in
slurries, and dispersion of gases. A specification of a superficial
velocity will enable selection of appropriate impeller size, rotation
speed, and power input with the aid of charts such as Figures 10.6
and 10.7. Examples 10.1 and 10.2 are along these lines.
The combination of HP and rpm that corresponds to a
particular superficial velocity depends on the size of the tank, the
size of the impeller, and certain characteristics of the system. Tables
10.3, 10.4, and 10.5 are abbreviated combinations of horsepower
and rpm that are suitable at particular pumping rates for the three
main categories of mixing. More complete data may be found in the
literature cited with the tables.
l. For mixing of liquids, data are shown for a viscocity of 5000 cP,
but data also have been developed for 25,000 cP, which allow for

EXAMPLE 10.1 Take N = 84 rpm.

Impeller Size and Speed at a Specified Power Input According to Figure 10.7 at d / D = 0.4,
For a vessel containing 5000 gal of liquid with specific gravity = 0.9
and viscosity of 100 cP, find size and speed of a pitched turbine
impeller to deliver 2 HP/1000 gal. Check also the superficial linear NQ = 0.61,
velocity and the blend time. Q = NQNd 3 = 0.61(84/60)(46/12) 3 = 48.1 cfs,
The dimensions of the liquid content are 9.5 ft high by
us = 48.1/[(n/4)(9.5}2] = 0.68 fps.
9.5 ft dia. Take

d = 0.4D = 0.4(9.5)(12) = 45.6 in., say 46 in., impeller, This value corresponds to moderate to high mixing intensity
according to Table 10.2.
P = 2V = 2(5) = lO HP,
From Figure 10.3, at NRe = 1720, blend time is given by
10.75SNd2 10.75(0.9)(46)2N
NRe µ 1000 20.47N,

N _ l.523(1013)P 1523(1013)(10) 821,600

P - N 3D 5 S 0.9(46) 5N3 N3
or

Solve for N by tria: with the aid of curve 6 of Figure 10.6. 17

l.67 min.
84(0.4)2·3
Trial N 114.. IV,, N (Eq. (2))
56 1146 1.3 85.8 According to Table 10.1, the blend time is less than 6 min,
84 1720 1.3 85.8 which agrees qualitatively.
294 MIXING ANO AGITATION

TABLE 10.2. Agitation Results Corresponding to Specific Superficial Velocities

ft/sec Description ft/sec Description

Liquid Systems
0.1-0.2 low degree of agitation; a velocity of 0.2 ft/sec will
a. blend miscible liquids to unifarmity when specific
gravity differences are less than 0.1
b. blend miscible liquids to unifarmity if the ratio of
viscosities is less than 100
c. establish liquid movement throughout the vessel
d. produce a flat but moving surface
0.3-0.6 characteristic of most agitation used in chemical
processing; a velocity of 0.6 ft/sec will
e. blend miscible liquids to unifarmity if the specific
gravity differences are less than 0.6
f. blend miscible liquids to unifarmity if the ratio of
viscosities is less than 10,000
g. suspend trace solids (less than 2%) with settling
rates of 2-4 ft/min
h. produce surface rippling at low viscosities
0.7-1.0 high degree of agitation; a velocity of 1.0 ft/sec will
i. blend miscible liquids to unifarmity if the specific
gravity differences are less than 1.0
j. blend miscible liquids to unifarmity if the ratio of
viscosities is less than 100,000
k. suspend trace solids (less than 2%) with settling
rates of 4-6 ft/min
l. produce surging surface at low viscosities
Solids Suspension
0.1-0.2 minimal solids suspension; a velocity of 0.1 ft/sec will
a. produce motion of all solids with the design settling
velocity
b. move fillets of solids on the tank bottom and
suspend them intermittently
0.3-0.5 characteristic of most applications of solids suspension
and dissolution; a velocity of 0.3 ft/sec will
c. suspend all solids with the design settling velocity
completely off the bottom of the vessel
d. provide slurry unifarmity to at least one-third of the
liquid level
e. be suitable far slurry drawoff at low exit nozzle
locations
0.6-0.8 when unifarm solids distribution must be approached; a
velocity of 0.6 ft/sec will
f. provide unifarm distribution to within 95% of liquid
leve!
g. be suitable far slurry drawoff up to 80% of liquid
level
0.9-1.0 when the maximum feasible unifarmity is needed. A
velocity of 0.9 ft/sec will
h. provide slurry unifarmity to 98% of the liquid level
i. be suitable far slurry drawoff by means of overflow
Gas Dispersion
0.1-0.2 used when degree of dispersion is not critica! to the
process; a velocity of 0.2 ft/sec will
a. provide nonflooded impeller conditions far coarse
dispersion
b. be typical of situations that are not mass transfer
limitad
0.3-0.5 used where moderate degree of dispersion is needed; a
velocity of 0.5 ft/sec will
c. drive fine bubbles completely to the wall of the
vessel
d. provide recirculation of dispersad bubbles back into
the impeller
0.6-1.0 used where rapid mass transfer is needed; a velocity of
1.0 ft/sec will
e. maximiza interfacial area and recirculation of
dispersad bubbles through the impeller

[Chemineer, Ca. Staff, Chem. Eng.. 102-110 (26 April 1976); 144-150 (24 May 1976); 141-148 (19 July 1976)].

llb 114.. llb

EXAMPLE 10.2 d/D d N [Eq. (1)] [Eq.(2)] (Fig. 10.7) IV,, P (HP)
Eft'ects of the Ratios of Impeller and Tank Diameters 0.25
Power and rpm requirements will be investigated and compared 28.4 300 0.637 518 0.64 1.4 45.9
0.33
with the data of Table 10.3. The superficial velocity is 0.6 ft/sec, 37.5 145 0.573 436 0.57 1.45 21.5
0.50 56.8 52 0.460 359 0.45 1.5 8.2
V = 5000 gals, Sp Gr = 1.0. Viscosities of 100 cP and 5000 cP will be
considered. With µ = 100 cP, turbulence is fully developed.
With h/ D = 1, D = h = 9.47 ft, llb
d/D d N 114.. (Fig. 10.7) N,, P
pumping rate Q = 0.6(n/4)(9.47) 2 = 42.23 cfs, 0.25 28.4 228 0.839 18,990 0.84 1.3 18.7
3 3
NQ = l.037(105)Q/Nd = 4.3793/ Nd (1) 0.33 37.5 112 0.742 16,850 0.74 1.3 8.9
2 2
NRe = l0.7Nd S / µ = 0.00214Nd , µ = 5000, (2) 0.50 56.8 40 0.597 13,800 0.60 1.3 3.2

p = NP N 3d 5S / l.523( 1013 ),
NP from Figure 10.6.
For severa) choices of di D, solve Eqs. (1) and (2)
simultaneously with Figure 10.7. With µ = 5000 cP;
(3) Table 10.3 gives these combinations of HP/rpm as suitable: 25/125,
20/100, 10/56, 7.5/37. The combination 10/56 checks roughly the
last entry at 5000 cP. Table 10.3 also has data for viscosities of
25,000 cP, thus allowing for interpolation and possibly extra-
polation.
 10.5. SUSPENSION OF SOLIOS 295
TABLE 10.3. Mixing of Liquids; Power and lmpeller Speed (hp/rpm) for
Two Viscosities, as a Function of the Liquid Superficial
Velocity; Pitched Blade Turbine lmpeller

Volume (gal)

5000 cP 25,000 cP

ft/sec 1000 2000 5000 1000 2000 5000

0.1 2/280 2/190 2/100 2/125 2/84 7.5/125

1/190 1/100 1.5/84 1.5/56 5/100
5/84
3/56
0.2 2/190 2/125 5/125 3/84 5/125 10/84
1/100 2/84 3/84 2/84 3/84 7.5/68
1.5/84 3/68 1.5/56 3/68 5/45
2/45 2/45 3/37
0.3 2/125 3/84 7.5/125 5/125 15/155 20/100
1.5/84 1.5/56 5/100 5/84 7.5/68 15/68
5/84 3/68 5/45 10/45
3/56 2/45 3/37 7.5/37
0.4 2/84 5/125 10/84 7.5/84 10/84 30/100
1.5/56 3/68 7.5/68 5/56 7.5/45 25/84
3/56 5/45 20/68
2/45 3/37 10/37
0.5 5/125 7.5/125 15/100 15/155 25/125 75/190
3/84 5/84 10/68 10/100 20/100 60/155
7.5/45 10/84 15/84 40/100
7.5/68 10/56 15/45
0.6 5/100 15/155 25/125 20/155 25/100 40/84
3/68 10/100 20/100 15/125 15/68 30/68
3/56 7.5/84 10/56 15/56 25/56
2/45 3/37 7.5/37 10/45 20/37
0.7 7.5/125 10/84 15/68 25/155 40/155 75/125
5/84 7.5/68 15/56 15/84 30/100 50/84
5/45 10/45 25/84 30/45
10/37 20/68
0.8 10/125 10/68 30/100 30/155 50/155 75/100
7.5/100 7.5/56 25/84 25/125 40/125 60/84
20/68 20/100 50/68
15/45 40/56
0.9 15/155 15/84 60/155 40/155 75/190 75/84
10/100 10/56 40/100 30/125 60/155 60/68
7.5/84 7.5/45 25/100 40/100 50/56
30/68
1.0 10/84 30/155 50/100 40/125 60/125 125/125
7.5/68 25/125 40/84 30/100 50/100 100/100
20/100 30/68 50/84 75/68
15/68 25/56 40/84 60/56
[Hicks, Morton, and Fenic, Chem. Eng.. 102-110 (26April 1976)].

interpolation and possibly extrapolation. The impeller is a
pitched-blade turbine.
2. For suspension of solids, the tables pertain to particles with
settling velocities of 10 ft/min, but data are available for
25 ft/min. The impeller is a pitched-blade turbine.
3. For gas dispersion the performance depends on the gas rate.
Data are shown for a superficial inlet gas rate of 0.07 ft/sec, but
data are available up to 0.2 ft/sec. Four bafftes are specified and
the impeller is a vertical blade turbine.
impeller and shaft must not rotate near their resonant frequency.
Such mechanical details are analyzed by Ramsey and Zoller [Chem.
Eng., 101-108 (30 Aug. 1976)].
10.5. SUSPENSION OF SOLIOS
Besides the dimensions of the vessel, the impeller, and bafftes,
certain physical data are needed for complete description of a slurry
mixing problem, primarily:

Example 10.2 compares data of Table 10.4 with calculations
based on Figures 10.6 and 10.7 for all-liquid mixing. Power and rpm
requirements at a given superficial liquid velocity are seen to be
very sensitive to impeller diameter. When alternate combinations of
HP/rpm are shown in the table for a particular performance, the
design of the agitator shaft may be a discriminant between them.
The shaft must allow for the torque and bending moment caused by
the hydraulic forces acting on the impeller and shaft. Also, the
1. Specific gravities of the solid and liquid.
2. Solids content of the slurry (wt %).
3. Settling velocity of the partides (ft/min).
The last of these may be obtained from correlations when the
mesh size or particle size distribution is known, or preferably
experimentally. Taking into account these factors in their effect on
suspension quality is at present a highly empirical process. Tables
 296 MIXING ANO AGITATION

TABLE 10.4. Suspension of Solids; Power and lmpeller Speed (hp/rpm)

for Two Settling Velocities, as a Function of the Superficial
Velocity of the Liquid; Pitched Blade Turbine lmpeller

Volume (gal)

10 ft/min 25 ft/min
ft/sec 1000 2000 5000 1000 2000 5000

0.1 1/190 2/190 5/125 2/190 2/125 5/125

1/100 3/84 1/190 2/84 3/84
3/68 1/100 1.5/84 3/68
2/45 1.5/56 2/45
0.2 1/100 2/125 7.5/125 2/125 3/84 15/155
1.5/84 5/100 10/100
5/84 7.5/68
3/56 5/45
0.3 2/190 2/84 3/37 1.5/84 5/125 10/84
1.5/56 3/68
2/45
0.4 2/155 5/155 7.5/84 2/84 7.5/155 7.5/45
1.5/100 5/56 1.5/56 5/100
3/56
0.5 1.5/84 3/84 15/155 2/68 7.5/125 15/84
2/125 10/100 2/56 5/84 10/56
7.5/68 7.5/37
5/45
0.6 2/100 5/125 10/84 3/84 5/56 25/125
1.5/68 3/68 20/100
3/56 15/68
2/45 10/45
0.7 2/84 7.5/155 15/84 7.5/155 15/155 30/100
1.5/56 7.5/125 10/56 5/125 10/100 25/84
5/84 7.5/45 5/100 7.5/84 20/68
7.5/37 3/68 7.5/68 15/56
0.8 3/84 7.5/84 25/125 7.5/125 10/84 60/155
5/56 20/100 5/84 40/100
15/68 30/68
10/45 25/56
0.9 7.5/155 15/155 40/155 10/125 15/84 75/190
5/125 10/100 30/100 7.5/100 60/125
5/100 7.5/68 25/84 50/100
3/68 20/68 40/84
1.0 7.5/125 20/100 50/100 15/155 30/155 75/125
5/84 15/84 40/84 10/100 25/125 75/100
10/84 30/68 20/100 60/84
25/56 50/84
[Gates, Morton, and Fondy, Chem. Eng., 144-150 (24 May 1976)].

10.2-10.5 are one such process; the one developed by Oldshue
(1983) will be examined shortly.
Suspension of solids is maintained by upward movement of the
liquid. In principie, use of a draft tube and an axial flow impeller
will accomplish this flow pattern most readily. lt turns out,
however, that such arrangements are suitable only for low solids
contents and moderate power levels. In order to be effective, the
cross section of the draft tube must be appreciably smaller than that
of the vessel, so that the solids concentration in the draft tube may
become impractically high. The usually practical arrangement for
solids suspension employs a pitched blade turbine which gives both
axial and radial flow.
For a given tank size, the ultimate design objective is the
relation between power input and impeller size at a specified
uniformity. The factors governing such information are the slurry
volume, the slurry level, and the required uniformity. The method
of Oldshue has corrections for these factors, as Fv F2 , and Ej.
When multiplied together, they make up the factor & which is the
ordinate of Figure 10.8(d) and which determines what combinations
of horsepower and ratio of impeller and vessel diameters will do the
required task. Example 10.3 employs this method, and makes a
comparison with the Chemineer method of Tables 10.2 and 10.3.
10.6. GAS DISPERSION
Gases are dispersed in liquids usually to facilitate mass transfer
between the phases or mass transfer to be followed by chemical
reaction. In sorne situations gases are dispersed adequately with
spargers or porous distributors, but the main concern here is with
the more intense effects achievable with impeller driven agitators.
SPARGERS
Mixing of liquids and suspension of solids may be accomplished by
bubbling with an inert gas introduced uniformly at the bottom of
the tank. For mild agitation a superficial gas velocity of 1ft/min is
used, and for severe, one of about 4 ft/min.
 10.6. GAS DISPERSION 297

TABLE 10.5. Dispersion of Gases; Power and lmpeller Speed power input as a factor is given by Treybal ( Mass Transfer
(hr/rpm) for Two Gas lnlet Superficial Velocities, Operations, McGraw-Hill, New York, 1980, 156); presumably this
as a Function of the Liquid Superficial Velocity; is applicable only below the mínimum power input here represented
Vertical Blade Turbine lmpeller
by Figure 10.11.
Volume (gal) When mass transfer coefficients are not determinable, agitator
design may be based on superficial liquid velocities with the criteria
0.07ft/sec 0.20ft/sec of Table 10.2.
ft/sec 1500 3000 5000 1500 3000 5000

0.1 2/56 5/84 7.5/68 3/56 7.5/68 10/45 SYSTEM DESIGN

15/155 10/100
0.2 2/45 7.5/125 10/84 3/45 15/155 15/68 The impeller commonly used for gas dispersion is a radial turbine
7.5/45 20/100 with six vertical blades. For a liquid height to diameter ratio
5/100 10/84 25/125 h/ D :5 l, a single impeller is adequate; in the range 1:5h /D :51.8
0.3 3/84 7.5/68 10/45 7.5/45 two are needed, and more than two are rarely used. The lower and
3/68 5/45 10/56 upper impellers are located at distances of 1/6 and 2/3 of the liquid
3/56 7.5/84 level above the bottom. Baffiing is essential, commonly with four
5/56 baffies of width 1/12 that of the tank diameter, offset from the wall
0.4 5/125 10/84 15/68 5/84 10/45 30/155 at 1/6 the width of the baffie and extending from the tangent line of
5/84 10/100 20/100 7.5/155 10/56 20/68
the wall to the liquid leve!. The best position for inlet of the gas is
5/100 10/45 15/84 5/56 15/45
5/45 10/56 20/68 15/56 below and at the center of the lower impeller; an open pipe is
0.5 7.5/125 15/155 25/125 7.5/125 15/68 25/84 commonly used, but a sparger often helps. Since ungassed power is
7.5/155 15/68 25/84 7.5/68 15/84 25/100 significantly larger than gassed, a two-speed motor is desirable to
7.5/68 15/84 25/100 7.5/84 15/45 25/56 prevent overloading, the lower speed to cut in automatically when
7.5/84 15/45 25/56 15/56 the gas supply is interrupted and rotation still is needed.
0.6 10/84 20/100 30/155 10/84 20/100 30/100
10/100 20/68 30/100 10/100 20/68 30/125
20/45 30/125 30/68 MINIMUM POWER
30/68 30/45 Below a critica! power input the gas bubbles are not affected
0.7 10/56 25/125 40/155 10/56 25/125 40/155
25/84 40/84 25/84 40/84
laterally but move upward with their natural buoyancy. This
25/100 40/100 25/100 40/100 condition is called gas flooding of the impeller. At higher power
25/56 40/56 25/56 40/56 inputs the gas is dispersed radially, bubbles impinge on the walls
0.8 15/155 30/155 50/100 15/155 30/155 50/100 and are broken up, consequently with improvement of mass
15/84 30/100 50/68 15/84 30/100 50/68 transfer. A correlation of the critical power input is shown as Figure
30/125 50/84 30/125 50/84 10.10.
50/45 50/56
0.9 15/68 30/68 60/125 15/68 30/68 60/125
60/155 60/155 POWER CONSUMPTION OF GASSED LIQUIDS
60/84 60/84
60/56 60/56 At least partly because of its lower density and viscosity, the power
1.0 20/100 40/155 75/190 25/125 40/155 75/190 to drive a mixture of gas and liquid is Jess than that to drive a liquid.
20/68 40/84 75/100 25/84 40/84 75/100 Figure 10.ll(a) is a correlation of this effect, and other data at low
75/125 75/125 values of the flow number Q/ Nd 3 are on Figure 10.ll(b). The latter
[Hicks and Gates, Chem. Eng., 141-148 (19 July 1976)]. data for Newtonian fluids are correlated by the equation

MASS TRANSFER
The starting point of agitator design is properly a mass transfer
coefficient known empirically or from sorne correlation in terms of
parameters such as impeller size and rotation, power input, and gas
flow rate. Few such correlations are in the open literature, but
sorne have come from two of the industries that employ aerated
stirred tanks on a large scale, namely liquid waste treating
and fermentation processes. A favored method of studying the
absorption of oxygen is to measure the rate of oxidation of aqueous
sodium sulfite solutions. Figure 10.9 summarizes one such
investigation of the effects of power input and gas rate on the mass
transfer coefficients. A correlation for fermentation air is given by
Dickey (1984, 12-17):
kLa = rate/(concentration driving force)
07 2
= 0.064(Pg/V ) · u· , 1/sec,
with P8/V in HP/1000 gal and superficial gas velocity u8 in ft/sec. A
general correlation of mass transfer coefficient that does not have
(10.7)
where the last group of terms is the Weber number, PL is the
density of the liquid, and a is its surface tension.
SUPERFICIAL LIQUID VELOCITY
When mass transfer data are not known or are not strictly pertinent,
a quality of mixing may be selected by an exercise of judgment in
terms of the superficial liquid velocity on the basis of the rules of
Table 10.2. For gas dispersion, this quantity is related to the power
input, HP/1000 gal, the superficial gas velocity and the ratio d / D in
Figure 10.12.
DESIGN PROCEDURES
On the basis of the information gathered here, three methods are
(10.6) possible for the design of agitated gas dispersion. In ali cases the
size of the tank, the ratio of impeller and tank diameters and the
gas feed rate are specified. The data are for radial turbine impellers
with six vertical blades.
298 MIXING ANO AGITATION

SLURRY VOL • - m 3
4 s 7 10 1s 20 30 40 so 70 100
100
80
4 .0
60

40 3.0
30 so 1-
ir

ir
o
20
30
10
s
-1'5 o
u
1-
a:
lJ._
1- 2 .0
1- 1 >-
u
a:
lJ._
CD
(/) '
10
o 1.S
8 d
(/)

d
> 6 1-
z
lJJ
u 1.0
4 ir
0.9
l JJ
3
a.. o.e
0 .7
2
0.6

O.S'---'----' -'--....1........1.... .1..---' -'

o.3 o.4 o.s o.6 0.1o.a 1.0 1.S 2.0
2000 4000 6000 10000 20000 30000 UT
SLURRY VOL. - GALLONS (b)
(al
SETTLING VELOCITY !m/MIN. )

40 0.6 1 .0 1.S 2.0 3.0 4 .0 6.0

30
20

10
B.O
ir
6.0 ir
D
1-
lJ._ 4 .0 u
a:

-
lJ._
3.0 z
u
oJ
lJJ
>
-
o
ezn
lJJ
a..
(!) en
z 1 .0 e:::n>
O.B
1-
l!l 0.6
0.4
0.3

0.2

2.0 3.o 4 .0 6.0 e.o 10.0 20.0

SETTLING VELOCITY lFT./MIN.l
DIT
(c) (d)
Figure 10.8. Suspension of solids. Power and ratio of diameters of impeller and tank, with four-bladed 45º impeller,
width/diameter = 0.2. [method of Oldshue ( 1983)]. (a) The factor on power consumption for slurry volume, F1 • (b) The
factor on power requirement for single and dual impellers at various h / D ratios, F'z. (c) The effect of settling velocity
on power consumption, Fj . (d) Suspension factor for various horsepowers: F4 = F1F2F3•
 10.6. GAS DISPERSION
299

EXAMPLE 10.3 d / D and HP is read off Figure 10.8(d).

Design of the Agitation System for Maintenance of a Slurry
These conditions are taken: HP

V = 5000 gal, d/ D Off btm Uniform

h/D = l, 0.2 20 65
settling velocity = 10 ft/rnin, 0.4 7.5 25
0.6 4 12
solids content = 10 wt %

Reading frorn Figure 10.8, Cornparing with readings from Tables 10.2 and 10.3,

Fi = 4, Superficial
liq. velocity HP/rpm
¡;; = 1.1,
0.3 (off btm) 10/45, 10/56
F; = { 3.0, off bottorn, 0.6 (uniform) 30/155, 30/125, 30/100, 30/68
10.0, uniforrn,
F = F, F F = {13.2, off bottorn,
4 1 2 3
44, uniforrn. These results correspond roughly to those of the Oldshue
method at d / D = 0.4. The impeller sizes can be deterrnined with
The relation between the ratio of irnpeller and vessel diarneters, Figures 10.6 and 10.7.

l. Start with a known required mass transfer coeflicient. Frorn a 3. As soon as a superficial liquid velocity has been selected, a
correlation such as Figure 10.9 or Eq. (10.6) the gassed power suitable combination of HP/rpm can be taken from Table 10.5.
per unit volume will become known, and the total gassed power
to the tank will be Pg. The ratio of gassed power to ungassed These procedures are applied in Example 10.4.
power is represented by Figure 10.ll(a) and the equations given As general rules, levels of 5-12 HP/1000 gal are typical of
there; at this stage the rotation speed N is not yet known. This aerobic fermentation vessels, and 1-3 HP/1000 gal of aerobic waste
value is found by tria! by simultaneous solution with Figure 10.6 treatment; concentrations and oxygen requirements of the
which relates the Reynolds and power numbers; the power here microorganisms are different in the two kinds of processes.
is the ungassed power. The value of N that results in the
precalculated Pg will be the correct one. Curve 2 of Figure 10.6 is
the one applicable to gas dispersion with the data of this section.
2. Start with a choice of superficial liquid velocity U¿ made in
accordance with the criteria of Table 10.2. With the aid of the
known gas velocity u, and d / D, find Pg/V from Figure 10.12.
Then proceed to find N by tria! with Figures 10.ll(a) and 10.6 as
in method l. o
o
o
g
.....
a.
0.1 ..:
.;
e:1 6
O/T
KGO .....
¡¡
.04 ll:
o
a.
j
LB MOLES
&
FT3/HR/l\TM .02 .§

.004 ............._. ._....... ...... -"'..... ........ 0 05 0.10 0.15 0.20 025 0.30
0.3 0 .6 !.O 2.0 4.0 B.O 10 Superficiol qos velocify, f f/s
HP I 1000 Gl\L. Gl\SSEO
Figure 10.10. Minirnum power requirernent to overcorne ftooding as
Figure 10.9. Typical data of mass transfer coefficients at various a function of superficial gas velocity and ratio of impeller and tank
power levels and superficial gas rates for oxidation of sodium sulfite diameters, d / D. [Hicks and Gates, Chem. Eng., 141-148 ( 19 July
in aqueous solution. d / D = 0.25-0.40 (Oldshue, 1983). 1976)].
300 MIXING AND AGITATION

o. p a
- = 1 - 1 26 --
2
11}
P, . NO,• 1

º' p -. .... 0.6 ..... I

'
"O A I I
a.ª ,,. 0.4 I
I'
/ '
a. ,; I I
11
>- I
0.2 - & 9,g" s
d "'
-
';:'
' M

a
102
0.1
4 • 6
1
o o o
N
d¡,
X ª 0.1 2 4 6 8 1 2
NO
X = (P/ V)(d/0) 1·85
(a)

Figure 10.12. Relation between power input, P/V HP/1000 gal,

superficial liquid velocity uL ft/sec, ratio of impeller and tank
diameters, d / D, and superficial gas velocity u5 ft/sec. [Hicks and
Gates, Chem. Eng., 141-148 (19 July 1976)].

have its own feed nozzle, as in Figure 10.13(b), but usually the
0.9 streams may be combined extemally near the blender and then
given the works, as in Figure 10.13(a).
One manufacturer gives these power ratings:

Tank size (gal) 5 10 30

0.1 Motor HP 0.5 2 3

Another ties in the line and motor sizes:

Une size, (in.) 1-4 6-8 10-12

o.a Motor HP 0.5 2

( ..)
(b)
Figure 10.11. Power consumption. (a) Ratio of power consumptions
of aerated and unaerated liquids. Q is the volumetric rate of the
gas: (0) glycol; ( x ) ethanol; (T) water. [After Calderbank, Trans.
Inst. Chem. Eng. 36, 443 ( 1958)]. (b) Ratio of power consumptions
of aerated and unaerated liquids at Jow values of Q/ Nd 3. Six·bladed
disk turbine: (O) water; (e) methanol (10%); (.Á) ethylene glycol
(8%); (.0.) glycerol (40%); Pg = gassed power input; P = ungassed
power input; Q = gas flow rate; N = agitator speed; d = agitator·
impeller diameter. [ Luong and Volesky, AIChE J. 25, 893 (1979)].
But above viscosities of 10 cP a body one size larger than the line
size is recommended.
Other devices utilize the energy of the flowing fluid to do the
mixing. They are inserts to the pipeline that force continua! changes
of direction and míxing. Loading a section of piping with tower
packing is an example but special assemblies of greater convenience
have been developed, sorne of which are shown in Figure 10.14. In
each case manufacturer's literature recommends the sizes and
pressure drops needed for particular services.
The Kenics mixer, Figure 10.14(a), for example, consists of a
succession of helical elements twisted altemately in opposite
directions. In laminar flow for instance, the flow is split in two at
each element so that after n elements the number of striations
becomes zn.The effect of this geometrical progression is illustrated
in Figure 10.14(b) and points out how effective the mixing becomes
after only a few elements. The Reynolds number in a corresponding
empty pipe is the major discriminant for the size of mixer, one
manufacturer:s recommendations being

114.. Number of Elements

10.7. IN-LINE BLENDERS ANO MIXERS Less than 10 24

When long residence time is not needed for chemical reaction or
other purposes, small highly powered tank mixers may be suitable,
with energy inputs measured in HP/gal rather than HP/1000 gal.
They bring together severa! streams continuously for a short contact
time (at most a second or two) and may be used whenever the
effiuent remains naturally blended for a sufficiently long time, that
is, when a true solution is formed or a stable emulsion-like mixture.
When it is essential that the mixing be immediate each stream will
10-2000 12-18
More than 2000 6
Besides liquid blending applications, static mixers have been
used for mixing gases, pH control, dispersion of gases into liquids,
and dispersion of dyes and solids in viscous liquids. They have the
advantages of small size, ease of operation, and relatively Jow cost.
The strong mixing effect enhances the rate of heat transfer from
viscous streams. Complete heat exchangers are built with such
 10.8. MIXING OF POWDERS ANO PASTES 301

EXAMPLE 10.4
HP and rpm Requirements of an Aerated Agitated Tank
A tank contains 5000 gal of liquid with sp gr = 1.0 and viscosity
100 cP that is aerated and agitated. The ratio of impeller to tank
diameters is d / D = 0.4. Two sets of conditions are to be examined.
From Table 10.2, a liquid velocity of 0.6-0.7 ft/sec will give
moderate to high dispersion. Table 10.5 gives possible HP/rpm
combination of 30/125, somewhat less than the value found here.
b. With Iiquid circulation velocity specified,

a. The air rate is 972 SCFM or 872 ACFM at an average uL = 0.5 ft/sec.
submergence of 4 ft. The corresponding superficial gas velocity is
0.206 ft/sec or 0.063 m/sec. A mass transfer coefficient Use Figure 10.12:
kLa = 0.2/sec is required; Dickey's equation (10.6) applies. Find
the power and rpm needed. Y = iouL(d / D)t. 2 = 10(0.5)(0.4)1. 2 = 1.67,
b. The air rate is 296 ACFM, 0.07 ft/sec, 0.0213 m/sec. The X = 0.8,
required intensity of mixing corresponds to a liquid superficial P8 /V = 0.8/(0.4) 185 = 4.36 HP/1000 gal
velocity of 0.5 ft/sec. Find the power, rotation speed, and mass
transfer coefficients for sulfite oxidation and for fermentation.
(this does exceed the minimum of 1.6 from Figure 10.11),
a. d = 0.4(9.47) = 3.79 ft, 45.46 in.,
1 2
pg = 5(4.36) = 21.8,
k La = 0.064( Pg/V ) 0· u· = 0.2,
0 2 10 7 3
P /V = [0.2/0.064(0.206) ]1 = 8.00 HP, N 3= 296/(3.79) N = 5.437/N,
8
P8 = 5(8.0) = 40.0 HP/5000 gal, NRe = 222N (part a),
3
Q/ Nd 3 = 872/(379) N = 16.02/N,
NRe = 10.75Nd S/µ = 10.75(45.46) 2N/100 = 222N.
2 K78,2P (part a).

Equation (10.2),
Solve by tria!, using Figure 10.lO(a) and curve 2 of Figure 10.6.
3
NP = 1.523(1013)P/N3d5S = 78,442P/N .
N Q/ Nd3 P,, / P 1141. N,, p P,,
Curve 2 of Figure (10.6) applies. P 8/ P from Figure 10.lO(a). Solve 100 0.0544 0.5194 22,200 4 51 26.5
by trial. 94 0.0576 0.5130 4 42.35 21.7 == 2.8

N Q/ Ntf' P,, / P IV,, p P,, The closest reading from Table 10.5 is HP/rpm = 25/100 which is a
100 0.160 0.324 22,200 4 51 16.5 good check.
150 0.107 0.422 33,300 4 172 72.6 For sulfite oxidation, at u8 = 0.07 ft/sec,
127 0.1261 0.3866 28,194 4 104.5 40.4 == 40.0

The last entry of P checks the required value 40.0. Find the P8 /V = 4.36 HP/1000 gal, from Figure 10.9,
8
corresponding superficial liquid velocity with Figure 10.12: k 8a = 0.07 lb mol/(cuft)/(hr)(atm).

X = ( P/V )( d / D)t.85 = 8.04(0.4)t.85 = 1.48, For fermentation, Eq. 10.6 gives

at uc = 0.206 ft/sec, Y = 2.0, kLa = 0.064(4.36)º 7(0.07)º 2

= Jb mol/(cuft)(sec)
0 ·105
:. uL = 2/10(0.4)1. 2 = 0.60 ft/sec. lb mol/ cuft ·

mixing inserts in the tubes and are then claimed to have 3-5 times
normal capability in sorne cases.
10.8. MIXING OF POWDERS ANO PASTES
Industries such as foods, cosmetics, pharmaceuticals, plastics,
rubbers, and also sorne others have to do with mixing of high
viscosity liquids or pastes, of powders together and of powders with
pastes. Much of this kind of work is in batch mode. The processes
are so diverse and the criteria for uniformity of the final product are
so imprecise that the nonspecialist can do little in the way of
equipment design, or in checking on the recommendations of
equipment manufacturers. Direct experience is the main guide to
selection of the best kind of equipment, predicting how well and
quickly it will perform, and what power consumption will be. For
projects somewhat out of direct experience and where design by
analogy may not suffice, testing in pilot plant equipment is a service
provided by many equipment suppliers.
A few examples of mixers and blenders for powders and pastes
are illustrated in Figure 10.15. For descriptions of available
equipment-their construction, capacity, performance, poweor
consumption, etc.-the primary sources are catalogs of manufac-
turers and contact with their offices. Classified lists of manu-
facturers, and sorne of their catalog information, appear in the
Chemical Engineering Catalog (Reinhold, New York, annually)
and in the Chemical Engineering Equipment Buyers Guide (McGraw-
Hill, New York, annually). Brief descriptions of sorne types of
equipment are in Perry's Chemical Engineers Handbook (McGraw-
Hill, New York, 1984 and earlier editions). Well-classified
descriptions, with figures, of paste mixers are in Ullmann (1972,
302 MIXING ANO AGITATION

t
(a) (b)

Figure 10.13. Motor-driven in-line blenders: (a) Double impeller made by Nettco Corp.; (b)
three-inlet model made by Cleveland Mixer Co.

(a) (b)

(e) (d)
Go o
o

Element Number
2 3 4 5

2 4 8 16 32
Number of Striations

(e) (f)
Figure 10.14. Sorne kinds of in-line mixers and blenders. (a) Mixing and blending with a recirculating pump. (b) Injector
mixer with a helical baffie. (e) Severa! perforated plates (orifices) supported on a rod. (d) Several perforated plates ftanged in.
(e) Hellical mixing elements with altemating directions (Kenics Corp.). (f) Showing progressive striations of the ftow channels
with Kenics mixing elements.
 10.8. MIXING OF POWDERS ANO PASTES 303

(a)

(b)

(e) (d)

Muller
wheels

Dnven
shoft

(e) (f)

(g) (h)

Figure 10.15. Sorne mixers and blenders for powders and pastes. (a) Ribbon blender for powders. (b) Flow pattern in a double cone blender
rotating on a horizontal axis. (c) Twin shell (Vee-type); agglomerate breaking and liquid injection are shown on the broken line. (d} Twin
rotor; available with jacket and hollow screws for heat transfer. (e) Batch muller. (f) Twin mullers operated continuously. (g) Double-arm
mixer and kneader ( Baker-Perkins !ne.). (h) Sorne types of blades for the double-arm kneader ( Baker-Perkins !ne.).
304 MIXING AND AGITATION
Vol. 2, pp. 282-300) and a similar one for powder mixers (loe. cit.,
pp. 301-311). Since this equipment industry has been quite stable,
REFERENCES
l. R.S. Brodkey (Ed.), Turbulence in Mixing Operations, Academic, New
York, 1975.
2. Chemineer Co. Staff, Liquid Agitation, Reprint of 12 articles from
Chemical Engineering, 8 Dec. 1975-6 Dec. 1976.
3. D.S. Dickey, In Handbook of Chemical Engineering Calculations, (N.P.
Chopey and T.G. Hicks Eds.), McGraw-Hill, New York, 1984.
4. S. Harnby, M.F. Edwards, and A.W. Nienow, Mixing in the Process
Industries, Butterworths, Stoneham, MA, 1985.
5. A.J. Kieser, Handbuch der chemisch-technischen Apparate, Springer-
Verlag, Berlin, 1934-1939.
6. W.J. Mead, Encyclopedia of Chemica/ Process Equipment, Reinhold,
New York, 1964.
older books are still useful, notably those of Riegel (1953), Mead
(1964), and particularly Kieser (1934-1939).
7. S. Nagata, Mixing Principies and App/ications, Wiley, New York,
1975.
8. J.Y. Oldshue, Fluid Mixing Technology, McGraw-Hill, New York,
1983.
9. E.R. Riegel, Chemical Process Machinery, Reinhold, New York, 1953.
10. Z. Sterbacek and P. Tausk, Mixing in the Chemica/ Industry, Pergamon,
New York, 1965.
11. J.J. Ulbrecht and G.K. Patterson, Mixing of Liquids by Mechanica/
Agitation, Gordon & Breach, New York, 1985.
U. V. Uh! and J.B. Gray (Eds.), Mixing Theory and Practice, Academic,
New York, 1966, 1967, 2 vols.
13. Ullmann's Encyclopedia of Chemica/ Technology, Verlag Chemie,
Weinheim, Germany, 1972, Vol. 2, pp. 249-311.