COURSE OUTLINE

Learning Outcomes

At the end of the course, students are expected to understand the followings:

• Properties of particles.

• Motion of particles in a fluid

• Stoke’S and newton’S LawS

• Flow through packed beds

• Fluidization

• Sedimentation and flocculation

• Filtration

• Screening

• Classification and grinding

2
 MODULE ONE (1)
Introduction

Particle technology is a term used to refer to the science and technology related to the handling

and processing of particles and powders. Particle technology is also often described as powder

technology, particle science and powder science. It can be defined as the study of materials

dispersed within a continuous fluid. The particles may be solid, but they can also be oil

droplets in water, water droplets in air etc. Hence by a particle we mean any dispersed material

within a fluid. In communal with most engineering subjects, particle technology requires a

basic competence in calculus and algebraic manipulation. A first course in fluid mechanics

would also be helpful, as it will be assumed that the student is acquainted with concepts such

as Reynolds number, Bernoulli’s equation, Hagen-Poiseuille, Newton's law, Friction Factor

and Flow Regimes. Knowledge of very basic statistics would be useful for a complete

understanding of solid/ solid mixing, likewise, the concepts of mean, median and mode are

used in discussing particle size distributions.

PROPERTIES OF PARTICLES

Single Particles: The simplest shape of a particle is the sphere. Frequently the size of a

particle of irregular shape is defined in terms of the size of an equivalent sphere. A measure

of particle shape which is frequently used is the sphericity, 𝜓, defined as:

Surface area of sphere of same volume as particle

𝜓=
Surface area of particle

Another method is to use the factor by which the cube of the size of the particle must be

multiplied to give the volume. Other properties of the particle are crystalline or amorphous,

porosity and the properties of its surface, such as roughness and presence of absorbed films.

Measurement of particle size: Measurement of particle size and of particle size distribution

is a highly specialized topic, and considerable skill is needed in the making of accurate

measurements and their interpretation.

3
A wide range of measurement techniques are available both for single particles and for

systems of particles. The principles of some of the chief methods are considered together with

an indication of the size range to which they are applicable.

Sieving (> 𝟓𝟎 𝝁𝒎)

The efficiency of screening is defined as the ratio of the mass of material which passes the

screen to that which is capable of passing. This will differ according to the size of the material.

It may be assumed that the rate of passage of particles of a given size through the screen is

proportional to the number or mass of particles of that size on the screen at any instant. Thus,

if 𝜔 is the mass of particles of a particular size on the screen at a time 𝑡, then:

𝑑𝜔
= −𝑘𝜔
𝑑𝑡

Where,

𝑘 = Constant for a given size and shape of particle and for a given screen.

Thus, the mass of particles (𝜔1 − 𝜔2 ) passing the screen in time, 𝑡 is given by:

𝜔2
ln = −𝑘𝑡
𝜔1

Or

𝜔2 = 𝜔1 𝑒 −𝑘𝑡

Microscopic analysis (𝟏 − 𝟏𝟎𝟎 𝝁𝒎): Microscopic examination permits measurement of the

projected area of the particle and also enables an assessment to be made of its two-dimensional

shape.

4
 Figure 1: Particle profiles and comparison circles

Sedimentation and elutriation methods (> 𝟏 𝝁𝒎): These methods depend on the fact that

terminal velocity of a particle in a fluid increase with size. The sedimentation methods are of

two types – the pipette method and the use of sedimentation balance. The elutriation method

is really a reverse sedimentation process in which the particles are dispersed in an upward

flowing stream of fluid.

Permeability methods (> 𝟏 𝝁𝒎): This methods depend on the fact that at low flow rates the

flow through a packed bed is directly proportional to the pressure difference, the

proportionality constant being proportional to the square of the specific surface

(𝑆𝑢𝑟𝑓𝑎𝑐𝑒 ∶ 𝑉𝑜𝑙𝑢𝑚𝑒 𝑟𝑎𝑡𝑖𝑜) of the powder. It is possible to obtain the diameter of the sphere

with the same specific surface as the powder from this method. And the method is dependent

upon the care with which the sample of powder is packed.

Particle Size Distribution

Mean particle size: A mean size will describe only one particular characteristic of the powder

and it is important to decide what that characteristic is before the mean is calculated. Thus, it

may be desirable to define the size of particle such that it’s mass or its surface or its length is

the mean value for all the particles in the system.

5
Considering unit mass of particles consisting of 𝑛1 particles of characteristic dimension 𝑑1 ,

constituting of mass fraction 𝑥1 , 𝑛2 particles of size 𝑑2 , and so on, then:

𝑥1 = 𝑛1 𝑘1 𝑑13 𝜌𝑠

And

∑ 𝑥1 = 1 = 𝜌𝑠 𝑘1 ∑(𝑛1 𝑑13 )

Thus;

1 𝑥1
𝑛1 =
𝜌𝑠 𝑘1 𝑑13

If the size distribution can be represented by a continuous function, then:

𝑑𝑥 = 𝜌𝑠 𝑘1 𝑑 3 𝑑𝑛

Or,

𝑑𝑥
= 𝜌𝑠 𝑘1 𝑑 3
𝑑𝑛

And
1
∫ 𝑑𝑥 = 1 = 𝜌𝑠 𝑘1 ∫ 𝑑3 𝑑𝑛
0

Where,
𝜌𝑠 = The density of the particles, and
𝑘1 = Constant whose value depends on the shape of the particle.

Mean Size Based on Volume

The mean abscissa in Figure 2 is defined as the volume mean diameter 𝑑𝑣 , or as the mass

mean diameter, where:

1 1
∫0 𝑑 𝑑𝑥
𝑑𝑣 = 1 = ∫ 𝑑 𝑑𝑥
∫0 𝑑𝑥 0

Expressing in finite difference form, then:

6
 ∑(𝑑1 𝑥1 )
𝑑𝑣 = = ∑(𝑑1 𝑥1 )
∑ 𝑥1

which, in terms of particle numbers, rather than mass fractions gives:

𝜌𝑠 𝑘1 ∑(𝑛1 𝑑14 ) ∑(𝑛1 𝑑14 )

𝑑𝑣 = =
𝜌𝑠 𝑘1 ∑(𝑛1 𝑑13 ) ∑(𝑛1 𝑑13 )

Figure 2: Size distribution curve – cumulative basis

Another mean size based on volume is the mean volume diameter, 𝑑𝑣′ . If all the particles are

of diameter, 𝑑𝑣′ , then the total volume of particles is the same as in the mixture.

Thus,

𝑘1 𝑑𝑣′ 3 ∑ 𝑛1 = ∑(𝑘1 𝑛1 𝑑13 )

Or,

3 ∑ 𝑛1 𝑑13
𝑑𝑣′ = √( )
∑ 𝑛1

3 ∑ 𝑥1
⟹ 𝑑𝑣′ = √( )
∑(𝑥1 ⁄𝑑13 )

3 1
= √( )
∑(𝑥1 ⁄𝑑13 )

Where,

7
 1 𝑥1
𝑛1 ≡
𝜌𝑠 𝑘1 𝑑13

Mean Sizes Based on Surface

In the previous figure, instead of total mass, the surface in each fraction is plotted against size,

a similar curve is obtained although the mean abscissa 𝑑𝑠 is then the surface mean diameter.

Thus,

∑[(𝑛1 𝑑1 )𝑆1 ] ∑(𝑛1 𝑘2 𝑑13 )

𝑑𝑠 = =
∑(𝑛1 𝑆1 ) ∑(𝑛1 𝑘2 𝑑12 )

∑(𝑛1 𝑑13 )
=
∑(𝑛1 𝑑12 )

Where,

𝑆1 = 𝑘2 𝑑12 , 𝑎𝑛𝑑 𝑘2 is a constant whose value depends on the particle shape, 𝑑2 is also

known as the particle as the Sauter mean diameter.

Substitution for 𝑛1

∑ 𝑥1 1
𝑑𝑠 = 𝑥1 = 𝑥
∑ ( ) ∑ ( 1)
𝑑1 𝑑1

The mean surface diameter is defined as the size of particle 𝑑𝑠′ which is such that if all the

particles are of this size, the total surface will be the same as in the mixture.

Thus,

𝑘2 𝑑𝑠′ 2 ∑ 𝑛1 = ∑(𝑘2 𝑛1 𝑑12 )

∑(𝑛1 𝑑12 )
𝑑𝑠′ = √
∑ 𝑛1

Substituting for 𝑛1 gives

∑(𝑥1 ⁄𝑑1 )
𝑑𝑠′ = √( )
∑(𝑥1 ⁄𝑑13 )

8
Mean Dimensions Based On Length.

A length mean diameter maybe defined as:

𝑥1
∑[(𝑛1 𝑑1 )𝑑1 ] ∑(𝑛1 𝑑12 ) ∑ (𝑑1 )
𝑑1 = = =
∑(𝑛1 𝑑1 ) ∑(𝑛1 𝑑1 ) ∑ 𝑥1
( 2)
𝑑1

A mean length diameter or arithmetic mean diameter may also be defined by:

𝑥
∑ ( 12 )
∑(𝑛 𝑑
1 1 ) 𝑑1
𝑑1′ = =
∑(𝑛1 ) 𝑥1
∑ ( 3)
𝑑1

Worked Example 1: The size analysis of a powdered material on a mass basis is represented
by a straight line from 0 per cent mass at 1 𝜇𝑚 particle size to 100 per cent mass at 101 𝜇𝑚
particle size as shown in the Figure below. Calculate the surface mean diameter of the particles
constituting the system.

Figure 3: Size analysis of powder

Solution: Surface mean diameter is given by the relationship:

∑ 𝑥1 1
𝑑𝑠 = 𝑥 = 𝑥
∑ ( 1) ∑ ( 1)
𝑑1 𝑑1

And since the size analysis is represented by a continuous curve

𝑑 = 100𝑥 + 1

Then:

9
 1
𝑑𝑠 =
1 𝑑𝑥
∫0 𝑑

And substituting for 𝑑 we have

1
𝑑𝑠 =
1 𝑑𝑥
∫0 100𝑥 + 1

= (100⁄ln 101)

= 21.7 𝜇𝑚

Worked Example 2: The equations giving the number distribution curve for a powdered

material are 𝑑𝑛⁄𝑑𝑑 = 𝑑 for the size range 0 − 10 𝜇𝑚 and 𝑑𝑛⁄𝑑𝑑 = 100,000⁄𝑑 4 for the size

range 10 − 100 𝜇𝑚 where 𝑑 is in 𝜇𝑚. Sketch the number, surface and mass distribution

curves and calculate the surface mean diameter for the powder. Explain briefly how the data

required for the construction of these curves may be obtained experimentally.

Solution:

For the range,

𝑑 = 0 − 10 𝜇𝑚, 𝑑𝑛⁄𝑑𝑑 = 𝑑

On integration:

∫ 𝑑𝑛 = ∫(𝑑) 𝑑𝑑

𝑛 = 0.5𝑑2 + 𝐶1 𝐸𝑞. 1

Where, 𝐶1 is the constant of integration

For the range:
𝑑 = 10 − 100 𝜇𝑚, 𝑑𝑛⁄𝑑𝑑 = 100,000⁄𝑑 4
On integration:

∫ 𝑑𝑛 = ∫ 105 (𝑑 −4 ) 𝑑𝑑

𝑛 = 𝐶2 − (0.33 × 105 𝑑 −3 ) 𝐸𝑞. 2

Where, 𝐶2 is the constant of integration
𝑤ℎ𝑒𝑛 𝑑 = 0, 𝑛 = 0 𝑎𝑛𝑑 𝑓𝑟𝑜𝑚 𝐸𝑞. 1: 𝐶1 = 0

10
𝑤ℎ𝑒𝑛 𝑑 = 10 𝜇𝑚 𝑖𝑛 𝐸𝑞. 1: 𝑛 = (0.5 × 102 ) = 50
𝐼𝑛 𝐸𝑞. 2: 50 = 𝐶2 − (0.33 × 105 × 10−3 ), 𝑎𝑛𝑑 𝐶2 = 83.0
Thus for 𝑑 = 0 − 10 𝜇𝑚: 𝑛 = 0.5𝑑 2
And for 𝑑 = 10 − 100 𝜇𝑚: 𝑛 = 83.0 − (0.33 × 105 𝑑 −3 )
Applying these equations, the following values of 𝑛 are obtained

𝒅(𝝁𝒎) 𝒏 𝒅(𝝁𝒎) 𝒏
0 0 10 50.0
2.5 3.1 25 80.9
5.0 12.5 50 82.7
7.5 28.1 75 82.9
10.0 50.0 100 83.0

100

80

60
n

40

20

0
0 20 40 60 80 100
d (μm)

Figure 4: Calculated data values of 𝒙 and 𝒔 are plotted as functions of 𝒅.

From Figure 4, values of 𝑑 are obtained for various values of 𝑛 and hence 𝑛1 and 𝑑1 are

obtained for each increment of 𝑛. Values of 𝑛1 𝑑12 and 𝑛1 𝑑13 are calculated and the totals

obtained. The surface area of the particles in the increment is then given by:

11
 𝑠1 = 𝑛1 𝑑12 ⁄∑ 𝑛1 𝑑12

And 𝑠 is then found as ∑ 𝑠1 , similarly the mass of the particles, 𝑥 = ∑ 𝑥1 , where

𝑥1 = 𝑛1 𝑑13 ⁄∑ 𝑛1 𝑑13

The results are:

𝒏 𝒅 𝒏𝟏 𝒅𝟏 𝒏𝟏 𝒅𝟐𝟏 𝒏𝟏 𝒅𝟑𝟏 𝒔𝟏 𝒔 𝒙𝟏 𝒙
0 0
20 3.1 192 596 0.014 0.014 0.001 0.001
20 6.2
20 7.6 1155 8780 0.085 0.099 0.021 0.022
40 9.0
10 9.5 903 8573 0.066 0.165 0.020 0.042
50 10.0
10 10.7 1145 12250 0.084 0.249 0.029 0.071
60 11.4
5 11.75 690 8111 0.057 0.300 0.019 0.090
65 12.1
5 12.85 826 10609 0.061 0.0361 0.025 0.115
70 13.6
2 14.15 400 5666 0.029 0.390 0.013 0.128
72 14.7
2 15.35 471 7234 0.035 0.425 0.017 0.145
74 16.0
2 16.75 561 9399 0.041 0.466 0.022 0.167
76 17.5
2 18.6 692 12870 0.057 0.517 0.030 0.197
78 19.7
2 21.2 890 18877 0.065 0.582 0.044 0.241
80 22.7
1 24.1 581 14000 0.043 0.625 0.033 0.274
81 25.5
1 28.5 812 23150 0.060 0.685 0.055 0.329
82 31.5
1 65.75 4323 284240 0.316 1.000 0.670 1.000
83 100
13641 424355

Values of 𝑥 and 𝑠 are plotted as functions of 𝑑. The surface mean diameter,

∑(𝑛1 𝑑13 )
𝑑𝑠 =
∑(𝑛1 𝑑12 )

12
 ∑ 𝑥1 1
= 𝑥 = 𝑥
∑ ( 1) ∑ ( 1)
𝑑1 𝑑1

= ∫ 𝑑3 𝑑𝑛⁄∫ 𝑑 2 𝑑𝑛

𝑓𝑜𝑟 0 < 10𝜇𝑚, 𝑑𝑛 = 𝑑 𝑑𝑑

𝑓𝑜𝑟 10 < 𝑑 < 100𝜇𝑚, 𝑑𝑛 = 105 𝑑 −4 𝑑𝑑

10 100 10 100
𝑑𝑠 = (∫ 𝑑 4 𝑑𝑑 + ∫ 105 𝑑 −1 𝑑𝑑 )⁄(∫ 𝑑 3 𝑑𝑑 + ∫ 105 𝑑 −2 𝑑𝑑 )
0 10 0 10

= ([𝑑 5 ⁄5]10 5 100 4 10 5 −1 100

0 + 10 [ln 𝑑]10 )⁄([𝑑 ⁄4]0 + 10 [−𝑑 ]10 )

= (2 × 104 + 2.303 × 105 )⁄(2.5 × 103 + 9 × 103 ) = 21.8 𝜇𝑚

13
 MODULE TWO (2)

Motion of Particles in a Fluid: This module will take a look at the motion of single solid
particles in fluids. According to Stokes, 1851 the drag force resisting creeping motion between
a rigid sphere of diameter x and a fluid of infinite amount, of viscosity 𝜇 is composed of two
components:

𝑎 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝑑𝑟𝑎𝑔 𝑓𝑜𝑟𝑐𝑒, 𝐹𝑃 = 𝜋𝑥𝜇𝑈 2.1

𝑎 𝑠ℎ𝑒𝑎𝑟 𝑠𝑡𝑟𝑒𝑠𝑠 𝑑𝑟𝑎𝑔 𝑓𝑜𝑟𝑐𝑒, 𝐹𝑆 = 2𝜋𝑥𝜇𝑈 2.2

𝑇𝑜𝑡𝑎𝑙 𝑑𝑟𝑎𝑔 𝑓𝑜𝑟𝑐𝑒 𝑟𝑒𝑠𝑖𝑠𝑡𝑖𝑛𝑔 𝑚𝑜𝑡𝑖𝑜𝑛, 𝐹𝐷 = 𝐹𝑃 + 𝐹𝑆 = 𝜋𝑥𝜇𝑈 + 2𝜋𝑥𝜇𝑈 = 3𝜋𝑥𝜇𝑈 2.3

where,

𝑈 = Relative velocity.

This is known as Stokes’ law. Experimentally, Stokes’ law is found to hold almost exactly for
single particle Reynolds number, 𝑅𝑒𝑝 ≤ 0.1; within 9 % for 𝑅𝑒𝑝 ≤ 0.3; within 3 % for𝑅𝑒𝑝 ≤
0.5 and within 9 % for𝑅𝑒𝑝 ≤ 1.0 ; where the single particle Reynolds number is defined in
Equation 2.4.

𝑆𝑖𝑛𝑔𝑙𝑒 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒 𝑅𝑒𝑦𝑛𝑜𝑙𝑑𝑠 𝑛𝑢𝑚𝑏𝑒𝑟, 𝑅𝑒𝑝 = 𝑥𝑈 𝜌𝑓 ⁄𝜇 2.4

𝑅′
𝐴 𝑑𝑟𝑎𝑔 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡, 𝐶𝐷 𝑖𝑠 𝑑𝑒𝑓𝑖𝑛𝑒𝑑 𝑎𝑠 𝐶𝐷 = 2.5
1/2𝜌𝐼 𝑈 2

Where,

𝑅′ = Force per unit projected area of the particle.

Thus, for a sphere:

𝜋𝑥 2
𝑅′ = 𝐹𝐷 ⁄( ) 2.6
4

And, Stokes’ law, in terms of the drag coefficient, becomes

24
𝐶𝐷 = 2.7
𝑅𝑒𝑝

14
 Figure 5: Standard Drag Curve for Motion of a Sphere in a Fluid

Table: Reynolds number ranges for single particle drag coefficient correlations

Region Stokes Intermediate Newton’s law

𝑅𝑒𝑝 range < 0.3 0.3 < 𝑅𝑒𝑝 < 500 500 < 𝑅𝑒𝑝 < 2 × 105
𝐶𝐷 24⁄𝑅𝑒𝑝 24 ~0.44
(1 + 0.15𝑅𝑒𝑝0.687 )
𝑅𝑒𝑝

At higher relative velocities, the inertia of the fluid begins to dominate (the fluid must
accelerate out of the way of the particle). Analytical solution of the Navier–Stokes equations
is not possible under these conditions. However, experiments give the relationship between
the drag coefficient and the particle Reynolds number in the form of the so-called standard
drag curve. Four regions are identified: the Stokes’ law region; the Newton’s law region in
which drag coefficient is independent of Reynolds number; an intermediate region between
the Stokes and Newton regions; and the boundary layer separation region. The Reynolds
number ranges and drag coefficient correlations for these regions are given in the table above.

𝐶𝐷 expression for the intermediate region shown in the table above is that of Schiller and
Naumann (1933).

Haider and Levenspiel (1989), correlation is proposed for 𝐶𝐷 over the entire range i.e.

24 0.4251
𝐶𝐷 = (1 + 0.15𝑅𝑒𝑝0.687 ) + ( ) 2.8
𝑅𝑒𝑝 6880.95
1+
𝑅𝑒𝑝

15
Particles Falling Under Gravity Through A Fluid: The relative motion under gravity of
particles in a fluid is in general, the forces of buoyancy, drag and gravity act on the particle:

𝑔𝑟𝑎𝑣𝑖𝑡𝑦 − 𝑏𝑜𝑢𝑦𝑎𝑛𝑐𝑦 − 𝑑𝑟𝑎𝑔 = 𝑎𝑐𝑐𝑒𝑙𝑒𝑟𝑎𝑡𝑖𝑜𝑛 𝑓𝑜𝑟𝑐𝑒 2.9

A particle falling from rest in a fluid will initially experience a high acceleration as the shear
stress drag. As the particle accelerates the drag force increases, causing the acceleration to
reduce. Eventually a force balance is achieved when the acceleration is zero and a maximum
or terminal relative velocity is reached. This is known as the single particle terminal velocity.

For a spherical particle the general equation above becomes:

𝜋𝑥 3 𝜋𝑥 3 𝜋𝑥 2
𝜌𝑝 𝑔 − 𝜌𝑓 𝑔 − 𝑅′ =0 2.10
6 6 4
Combining this equation and that of the drag force 𝐶𝐷 i.e.

𝜋𝑥 3 𝜋𝑥 3 𝜋𝑥 2
𝜌𝑝 𝑔 − 𝜌𝑓 𝑔 − 𝑅′ =0
6 6 4
And,

𝑅′
𝐶𝐷 =
1/2𝜌𝐼 𝑈 2

We have,

𝜋𝑥 2 𝐶𝐷 𝜌𝑓 𝑈𝑇2 𝜋𝑥 2
(𝜌𝑝 − 𝜌𝑓 )𝑔 − =0
6 8

Where, 𝑈𝑇2 is the single particle terminal velocity. The equation for the drag coefficient under
terminal velocity is given as thus:

4 𝑔𝑥 (𝜌𝑝 − 𝜌𝑓 )
𝐶𝐷 = [ ]
3 𝑈𝑇2 𝜌𝑓

Thus, in the Stokes’ law region, with 𝐶𝐷 = 24⁄𝑅𝑒𝑝 , the single particle terminal velocity is
given by:

𝑥 2 (𝜌𝑝 − 𝜌𝑓 )𝑔
𝑈𝑇 =
18𝜇

NOTE

In the Stokes’ Law Region: Terminal Velocity is directly proportional to the square of the
particle diameter.

16
 𝑈𝑇 ∝ 𝑥 2

In the Newton’s Law Region: with 𝐶𝐷 = 0.44, the terminal velocity is given by:
1⁄
2
𝑥(𝜌𝑝 − 𝜌𝑓 )𝑔
𝑈𝑇 = 1.74 [ ]
𝜌𝑓

In this region, the terminal velocity is independent of the fluid viscosity and proportional to
the square root of the particle diameter.

In the Intermediate Region: No explicit expression for the terminal velocity can be found.
However, in this region, the variation of terminal velocity with particle and fluid properties is
approximated to be:
0.7
𝑈𝑇 ∝ 𝑥 1.1 , (𝜌𝑝 − 𝜌𝑓 ) , 𝜌𝑓 −0.29 , 𝜇−0.43

Generally, when calculating the terminal velocity for a given particle or the particle diameter
for a given velocity. One way to know which region of operation is relevant is to formulate
the dimensionless groups, 𝐶𝐷 𝑅𝑒𝑝2 𝑎𝑛𝑑 𝐶𝐷 ⁄𝑅𝑒𝑝 .

1. To calculate 𝑈𝑇 , for a given size 𝑥. Calculate the group

4 𝑥 3 𝜌𝑓 (𝜌𝑝 − 𝜌𝑓 )𝑔
𝐶𝐷 𝑅𝑒𝑝2 =
3 𝜇2
4
Which is independent of 𝑈𝑇 . And nothing that, 𝐶𝐷 𝑅𝑒𝑝2 = 𝐴𝑟, where 𝐴𝑟 is the Archimedes
3

number.

For given particle and fluid properties, 𝐶𝐷 𝑅𝑒𝑝2 is a constant and will therefore produce a
straight line of slope −2 if plotted on the logarithmic coordinates (𝑙𝑜𝑔 𝐶𝐷 versus 𝑙𝑜𝑔 𝑅𝑒𝑝 ) of
the standard drag curve. The intersection of this straight line with the drag curve gives the
value of 𝑅𝑒𝑝 and hence 𝑈𝑇 .

2. To calculate size 𝑥, 𝑓𝑜𝑟 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑈𝑇 . Calculate the group

𝐶𝐷 4 𝑔𝜇(𝜌𝑝 − 𝜌𝑓 )
=
𝑅𝑒𝑝 3 𝑈𝑇3 𝜌𝑓2

Which is independent of particle size 𝑥. For a given terminal velocity, particle density and
𝐶𝐷
fluid properties, is constant and will produce a straight line of slope +1 if plotted on the
𝑅𝑒𝑝

logarithmic coordinates (𝑙𝑜𝑔𝐶𝐷 versus 𝑙𝑜𝑔𝑅𝑒𝑝 ) of the standard drag curve. The intersection
of this straight line with the drag curve gives the value of 𝑅𝑒𝑝 and hence, x. An alternative to

17
this graphical method, but based on the same approach, is to use tables of corresponding
𝐶𝐷
values of 𝑅𝑒𝑝 , 𝐶𝐷 , 𝐶𝐷 𝑅𝑒𝑝2 𝑎𝑛𝑑
𝑅𝑒𝑝

Figure 6: Method for Estimating Terminal Velocity for a given Size of Particle and
Vice Versa

Figure 7: Drag coefficient 𝑪𝑫 versus Reynolds number 𝑹𝒆𝒑 for particles of

Sphericity 𝝋 ranging from 0.125 to 1.0

Boundaries Effect on Terminal Velocity: When a particle is falling through a fluid in the
presence of a solid boundary the terminal velocity reached by the particle is less than that for
an infinite fluid. In practice, this is really only relevant to the falling sphere method of
18
measuring liquid viscosity, which is restricted to the Stokes’ region. In the case of a particle
falling along the axis of a vertical pipe this is described by a wall factor, 𝑓𝑤 , the ratio of the
velocity in the pipe, 𝑈𝐷 to the velocity in an infinite fluid 𝑈∞ . The correlation of Francis
(1933) for 𝑓𝑤 is given in the Equation below:

𝑥 2.25 𝑥
𝑓𝑤 = (1 − ) 𝑅𝑒𝑝 ≤ 0.3; ≤ 0.97
𝐷 𝐷

Worked Example 1

Calculate the upper limit of particle diameter 𝑥𝑚𝑎𝑥 as a function of particle density 𝜌𝑝 for
gravity sedimentation in the Stokes’ law regime. Plot the results as 𝑥𝑚𝑎𝑥 versus𝜌𝑝 over the
range 0 ≤ 𝜌𝑝 ≤ 8000 𝑘𝑔/𝑚3 for settling in water and in air at ambient conditions. Assume
that the particles are spherical and that Stokes’ law holds for 𝑅𝑒𝑃 ≤ 0.3.

Solution:
The upper limit of particle diameter in the Stokes’ regime is governed by the upper limit of
single particle Reynolds number:
𝜌𝑓 𝑥𝑚𝑎𝑥 𝑈
𝑅𝑒𝑃 = = 0.3
𝜇
In gravity sedimentation in the Stokes’ regime particles accelerate rapidly to their terminal
velocity. In the Stokes’ regime the terminal velocity is given by the Equation:
𝑥 2 (𝜌𝑝 − 𝜌𝑓 )𝑔
𝑈𝑇 =
18𝜇
Solving these two equations for 𝑥𝑚𝑎𝑥 we have
1⁄
18𝜇2 3
𝑥𝑚𝑎𝑥 = [0.3 × ]
𝑔(𝜌𝑝 − 𝜌𝑓 )𝜌𝑓
1⁄
𝜇2 3
= 0.82 × [ ]
(𝜌𝑝 − 𝜌𝑓 )𝜌𝑓
Thus, for air (density 1.2𝑘𝑔/𝑚3 and viscosity 1.84 × 10−5 𝑃𝑎 𝑠):

1⁄
1 3
𝑥𝑚𝑎𝑥 = 5.37 × 10−4 [ ]
(𝜌𝑝 − 1.2)

And for water (density 1000 𝑘𝑔/𝑚3 and viscosity 0.001 𝑃𝑎 𝑠):

19
 1⁄
1 3
𝑥𝑚𝑎𝑥 = 8.19 × 10−4 [ ]
(𝜌𝑝 − 1000)

𝒌𝒈 𝒙𝒎𝒂𝒙 𝒇𝒐𝒓 𝑨𝒊𝒓 𝒙𝒎𝒂𝒙 𝒇𝒐𝒓 𝑾𝒂𝒕𝒆𝒓

𝝆𝒑 (𝟎 ≤ 𝝆𝒑 ≤ 𝟖𝟎𝟎𝟎 ) 𝟏⁄ 𝟏⁄
𝒎𝟑 𝟏 𝟑 𝟏 𝟑
𝒙𝒎𝒂𝒙 = 𝟓. 𝟑𝟕 × 𝟏𝟎−𝟒 [ ] 𝒙𝒎𝒂𝒙 = 𝟖. 𝟏𝟗 × 𝟏𝟎−𝟒 [ ]
(𝝆𝒑 − 𝟏. 𝟐) (𝝆𝒑 − 𝟏𝟎𝟎𝟎)

0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
5500
6000
6500
7000
7500
8000

20
Figure 8: Sketch (a) Limiting Particle Size for Stokes’ Law in Water. (b) Limiting
Particle Size for Stokes’ Law in Air

Worked Example 2: A sphere of diameter 10 𝑚𝑚 and density 7700 𝑘𝑔⁄𝑚2 falls under
gravity at terminal conditions through a liquid of density 900 𝑘𝑔⁄𝑚3 in a tube of
diameter 12 𝑚𝑚. The measured terminal velocity of the particle is 1.66 𝑚𝑚⁄𝑠. Calculate the
viscosity of the fluid. Verify that Stokes’ law applies.

Solution: To solve this problem, we first convert the measured terminal velocity to the
equivalent velocity which would be achieved by the sphere in a fluid of infinite extent.
Assuming Stokes’ law we can determine the fluid viscosity. Finally we check the validity of
Stokes’ law.
Applying the Francis wall factor expression:

21
 𝑈𝑇∞ 1
= = 56.34
𝑈𝑇𝐷 𝑥 2.25
(1 − )
𝐷
Thus, terminal velocity for the particle in a fluid of infinite extent,

𝑈𝑇∞ = 𝑈𝑇𝐷 × 56.34 = 0.0901 𝑚⁄𝑠

Equating this value to the expression for 𝑈𝑇∞ in the Stokes’ regime, we have that:
𝑥 2 (𝜌𝑝 − 𝜌𝑓 )𝑔
𝑈𝑇∞ =
18𝜇
Becomes,

(10 × 10−3 )2 × (7700 − 900) × 9.81

𝑈𝑇∞ =
18𝜇
Solving the above equation gives the fluid velocity 𝜇 = 4.11 𝑃𝑎. 𝑠
Checking the validity of Stoke’s law:
Single particle Reynolds number:
𝜌𝑓 𝑥𝑈
𝑅𝑒𝑃 = = 0.197
𝜇
𝑅𝑒𝑃 Is less than 0.3 and so the assumption that Stokes’ law holds is valid.

22
 MODULE THREE (3)

Flow through Packed Beds

Pressure Drop–Flow Relationship

Laminar Flow: Darcy (1856) observed that the flow of water through a packed bed of sand
was governed by the relationship:

𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝐿𝑖𝑞𝑢𝑖𝑑 (−∆𝑝)

( )∝( ) 𝑜𝑟 ∝𝑈 3.1
𝐺𝑟𝑎𝑑𝑖𝑒𝑛𝑡 𝑉𝑒𝑙𝑜𝑐𝑖𝑡𝑦 𝐻

Where 𝑈, is the superficial fluid velocity through the bed and (−∆𝑝) is the frictional pressure
drop across a bed depth 𝐻.

𝑓𝑙𝑢𝑖𝑑 𝑣𝑜𝑙𝑢𝑚𝑒𝑡𝑟𝑖𝑐 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒

𝑆𝑢𝑝𝑒𝑟𝑓𝑖𝑐𝑖𝑎𝑙 𝑉𝑒𝑙𝑜𝑐𝑖𝑡𝑦 =
𝑐𝑟𝑜𝑠𝑠 − 𝑠𝑒𝑐𝑡𝑖𝑜𝑛𝑎𝑙 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑏𝑒𝑑

𝑄
𝑖. 𝑒. 𝑈=
𝐴

The starting point of the flow of tube through a packed bed is the Hagen-Poiseuille equation
for laminar flow through a tube:

(−∆𝑝) 32𝜇𝑈
= 3.2
𝐻 𝐷2

Where 𝐷 is the tube diameter and 𝜇 is the fluid viscosity.

Consider the packed bed to be equivalent to many tubes of equivalent diameter 𝐷𝑒 following
tortuous paths of equivalent length 𝐻𝑒 and carrying fluid with a velocity 𝑈𝑖 . Then, from
Equation 3.2,

(−∆𝑝) 𝜇𝑈𝑖
= 𝐾1 2 3.3
𝐻𝑒 𝐷𝑒

𝑈𝑖 is the actual velocity of fluid through the interstices of the packed bed and is related to
superficial fluid velocity by:

𝑼𝒊 = 𝑼⁄𝜺 𝟑. 𝟒

23
Where 𝜀 is the voidage or void fraction of the packed bed.

Since the paths of the tubes are twisting, we assume that their actual length is proportional to
the bed depth, that is

𝑯𝒆 = 𝑲𝟐 𝑯 𝟑. 𝟓
The tube equivalent diameter is defined as
4 × 𝑓𝑙𝑜𝑤 𝑎𝑟𝑒𝑎
𝑤𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟
Where,
𝑓𝑙𝑜𝑤 𝑎𝑟𝑒𝑎 = 𝜀𝐴
Where, 𝐴 is the cross-sectional area of the vessel holding the bed.
𝑤𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 = 𝑆𝐵 𝐴
Where, 𝑆𝐵 is the particle surface area per unit volume of the bed.
That this is so may be demonstrated by comparison with pipe flow: Total particle surface area
in the bed = 𝑆𝐵 𝐴𝐻.
For a pipe,
𝑤𝑒𝑡𝑡𝑒𝑑 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝜋𝐷𝐿
𝑤𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 = =
𝑙𝑒𝑛𝑔𝑡ℎ 𝐿
And so for the packed bed,
𝑆𝐵 𝐴𝐻
𝑤𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 = = 𝑆𝐵 𝐴
𝐻
Now if 𝑆𝑣 is the surface area per unit volume of particles, then

𝑆𝑣 (1 − 𝜀) = 𝑆𝐵 3.6

Since

𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑜𝑓 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑜𝑓 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠

( )×( )=( )
𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑏𝑒𝑑 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑏𝑒𝑑

And so,

𝟒𝜺𝑨 𝟒𝜺
𝒆𝒒𝒖𝒊𝒗𝒂𝒍𝒆𝒏𝒕 𝒅𝒊𝒂𝒎𝒆𝒕𝒆𝒓, 𝑫𝒆 = = 𝟑. 𝟕
𝑺𝑩 𝑺𝒗 (𝟏 − 𝜺)

Substituting Equations 3.4, 3.5 and 3.7 in 3.3:

24
 (−∆𝑝) (𝟏 − 𝜺)𝟐
= 𝐾3 𝜇𝑈𝑆𝑣2 3.8
𝐻 𝜺𝟑

Where 𝐾3 = 𝐾1 𝐾2 . Equation 3.8 is the Carman – Kozeny equation describing laminar flow
through randomly packed particles. The constant 𝐾3 depend on particle shape and surface
properties and has been found experimentally to have a value of five (5).

Taking 𝐾3 = 5, for laminar flow through a randomly packed bed of monosized spheres of
diameter 𝑥. (for which S = 6/x ), the Carman – Kozeny equation becomes:

(−∆𝑝) 𝜇𝑈 (1 − 𝜀)2
= 180 2 3.9
𝐻 𝑥 𝜀3

This is the most common form in which the Carman – Kozeny equation is quoted.

Turbulent Flow: For turbulent flow through a randomly packed bed of mono – sized spheres
of diameter 𝑥 the equivalent equation is:

(−∆𝑝) 𝜌𝑓 𝑈 2 (1 − 𝜀)
= 1.75 3.10
𝐻 𝑥 𝜀3

General Equation for Turbulent and Laminar Flow: Ergun (1952) suggested the following
general equation for any flow conditions for a wide range of size and shape of particles:

(−∆𝑝) 𝜇𝑈 (1 − 𝜀)2 𝜌𝑓 𝑈 2 (1 − 𝜀)
= 150 2 + 1.75 3.11
𝐻 𝑥 𝜀3 𝑥 𝜀3

𝑙𝑎𝑚𝑖𝑛𝑎𝑟 𝑡𝑢𝑟𝑏𝑢𝑙𝑒𝑛𝑡
( ) ( )
𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡

This is known as the Ergun equation for flow through a randomly packed bed of spherical
particles of diameter 𝑥. Ergun’s equation additively combines the laminar and turbulent
components of the pressure gradient. But with the constant 150 rather than 180, the difference
in the values of the constants is probably due to differences in shape and packing of the
particles. In terms of the Reynolds number defined, fully laminar condition exist for 𝑅𝑒 ∗ less
than about 10 and fully turbulent flow exists at Reynolds numbers greater than around 2000.

𝑥𝑈𝜌𝑓
𝑅𝑒 ∗ = 3.12
𝜇(1 − 𝜀)

25
Ergun also expressed flow through a packed bed in terms of a friction factor defined in
Equation 3.13:

(−∆𝑝) 𝑥 𝜀3
𝑓𝑟𝑖𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟, 𝑓 ∗ = 3.13
𝐻 𝜌𝑓 𝑈 2 (1 − 𝜀)

Comparing this with the familiar Fanning friction factor for flow through pipes. The general
equation for turbulent and laminar flow becomes:

150
𝑓∗ = + 1.75 3.14
𝑅𝑒 ∗

With

150
𝑓∗ = 𝑓𝑜𝑟 𝑅𝑒 ∗ < 10 𝑎𝑛𝑑 𝑓 ∗ = 1.75 𝑓𝑜𝑟 𝑅𝑒 ∗ > 2000
𝑅𝑒 ∗

Non-spherical Particles: The Ergun and Carman–Kozeny equations also accommodate non-
spherical particles if 𝑥 is replaced by 𝑥𝑠𝑣 . The relevance of this will be apparent if equation
3.8 is recalled. Thus, in general, the Ergun equation for flow through a randomly packed bed
of particles of surface-volume diameter 𝑥𝑠𝑣 becomes:

(−∆𝑝) 𝜇𝑈 (1 − 𝜀)2 𝜌𝑓 𝑈 2 (1 − 𝜀)
= 150 2 + 1.75 3.15
𝐻 𝑥𝑠𝑣 𝜀3 𝑥𝑠𝑣 𝜀3

And the Carman - Kozeny equation for laminar flow through a randomly packed bed of
particles of surface-volume diameter 𝑥𝑠𝑣 becomes:

(−∆𝑝) 𝜇𝑈 (1 − 𝜀)2
= 180 2 3.16
𝐻 𝑥𝑠𝑣 𝜀3
Filtration
Analysis of flow through packed beds of particle can best be considered cake filtration. It is
widely used in the Industry to separate solid particles from suspension in liquid. It involves
the build-up of a bed or ‘cake’ of particles on porous surface known as the filter medium,
which commonly takes the form of a woven fabric.
In cake filtration the pore size of the medium is less than the size of the particles to be filtered.
The filtration process is analysed in terms of the flow of fluid through a packed bed of
particles, the depth of which is increasing with time. In practice the voidage of the cake may

26
also change with time. Now let’s first consider the case where the cake voidage is constant
(an incompressible cake) and secondly the case where it changes with time (compressible
cake).
Incompressible cake: Cake voidage remains constant with time in this case. Ignoring the
filter medium and considering only the cake alone, the pressure drop versus the liquid flow
relationship is described by the Ergun Equation:
(−∆𝑝) 𝜇𝑈 (1 − 𝜀)2 𝜌𝑓 𝑈 2 (1 − 𝜀)
= 150 2 + 1.75
𝐻 𝑥𝑠𝑣 𝜀3 𝑥𝑠𝑣 𝜀3
The particle size and range of liquid flow and properties commonly used in industry give rise
to laminar flow; hence the second (turbulent) term vanishes. So for a given slurry with particle
properties fixed the resulting cake resistance is defined as:

150 (1 − 𝜀)2
𝐶𝑎𝑘𝑒 𝑟𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒, 𝑟𝐶 = 3.17
𝑥𝑠𝑣 2 𝜀 3

Substituting into Ergun equation (Equ. 3.15) yields

(−∆𝑝)
= 𝑟𝐶 𝜇𝑈 3.18
𝐻

𝑑𝑉
If 𝑉 is the volume of filtrate (liquid) that passed at time 𝑡 and is the instantaneous
𝑑𝑡

volumetric flow rate at time 𝑡, then:

1 𝑑𝑉
𝑆𝑢𝑝𝑒𝑟𝑓𝑖𝑐𝑖𝑎𝑙 𝑓𝑖𝑙𝑡𝑟𝑎𝑡𝑒 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦 𝑎𝑡 𝑡𝑖𝑚𝑒 𝑡, 𝑈 = 3.19
𝐴 𝑑𝑡
Each unit volume of filtrate is assumed to deposit a certain mass of particles, which form a
certain volume of cake. This is expressed as 𝜑, the volume of cake formed by the passage of
unit volume of filtrate:
𝐻𝐴
𝜑= 3.20
𝑉
Thus, from the above equation:
𝜑𝑉
𝐻=
𝐴
Substituting this into equation 3.18 we have:
(−∆𝑝) 1 𝑑𝑉
= 𝑟𝐶 𝜇
𝜑𝑉 𝐴 𝑑𝑡
𝐴
Rearranging the above expression, we have
27
 (−∆𝑝). 𝐴. 𝐴 𝑑𝑉
=
𝜑𝑉𝑟𝐶 𝜇 𝑑𝑡

𝑑𝑉 𝐴2 (−∆𝑝)
⟹ = 3.21
𝑑𝑡 𝑟𝐶 𝜇𝜑𝑉
Constant rate filtration
𝑑𝑉
If the filtration rate = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡. Thus,
𝑑𝑡

𝐴2 (−∆𝑝)
=𝑘
𝑟𝐶 𝜇𝜑𝑉

Then the pressure drop (−∆𝑝) across the filter cake will increase in direct proportion to the
volume of filtrate passed.

Constant pressure drop filtration

If (−∆𝑝) is constant then
𝑑𝑉 1
∝
𝑑𝑡 𝑉
On integration of equation 3.21 i.e.
𝑑𝑉 𝐴2 (−∆𝑝)
=
𝑑𝑡 𝑟𝐶 𝜇𝜑𝑉
Yields
𝑡
= 𝐶1 𝑉 3.22
𝑉
Where,
𝑟𝐶 𝜇𝜑
𝐶1 = 3.23
2𝐴2 (−∆𝑝)
Combining the Resistance of the Filter Medium
The total resistance to flow is the sum of the resistance of the cake and the filter medium;
Hence:
𝑇𝑜𝑡𝑎𝑙 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝑑𝑟𝑜𝑝 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝑑𝑟𝑜𝑝
( )=( )+( )
𝑑𝑟𝑜𝑝 𝑎𝑐𝑟𝑜𝑠𝑠 𝑚𝑒𝑑𝑖𝑢𝑚 𝑎𝑐𝑟𝑜𝑠𝑠 𝑐𝑎𝑘𝑒

(−∆𝑝) = (−∆𝑝𝑚 ) + (−∆𝑝𝑚 )

If the medium is assumed to behave as a packed bed of depth 𝐻𝑚 and resistance 𝑟𝑚 obeying
the Carman – Kozeny equation, then:
1 𝑑𝑉
(−∆𝑝) = (𝑟 𝜇𝐻 + 𝑟𝐶 𝜇𝐻𝐶 ) 3.24
𝐴 𝑑𝑡 𝑚 𝑚
28
The medium resistance 𝑟𝑚 is usually expressed as the equivalent thickness of 𝐻𝑒𝑞 .
𝑟𝑚 𝐻𝑚 = 𝑟𝐶 𝐻𝑒𝑞 ∗
𝑟𝐶 𝜇𝜑𝑉𝑑𝑉 = 𝐴2 (−∆𝑝)𝑑𝑡
𝐴2 (−∆𝑝)
⟹ 𝑉𝑑𝑉 = 𝑑𝑡
𝑟𝐶 𝜇𝜑
On integration:
𝐴2 (−∆𝑝)
𝑉 ∫ 𝑑𝑉 = ∫ 𝑑𝑡
𝑟𝐶 𝜇𝜑
This will yield

𝑡 𝑟𝐶 𝜇𝜑
= 2
𝑉 2𝐴 (−∆𝑝)

𝐻𝐴
Hence combining with Equation 3.20 (𝜑 = ) we have
𝑉

𝜑𝑉𝑒𝑞
𝐻𝑒𝑞 = 3.25
𝐴
Where,
𝑉𝑒𝑞 is the volume of filtrate that must pass in order to create a cake of thickness 𝐻𝑒𝑞 . The
volume 𝑉𝑒𝑞 depends only on the properties of suspension and the filter medium. Thus equation
1 𝑑𝑉
3.24 [(−∆𝑝) = (𝑟𝑚 𝜇𝐻𝑚 + 𝑟𝐶 𝜇𝐻𝐶 )] now becomes:
𝐴 𝑑𝑡

1 𝑑𝑉 (−∆𝑝)𝐴
= 3.26
𝐴 𝑑𝑡 𝑟𝐶 𝜇(𝑉 + 𝑉𝑒𝑞 )𝜑
Considering operation at constant pressure drop, which is the most common case, integrating
the above equation we have:
𝑡 𝑟𝐶 𝜇𝜑 𝑟𝐶 𝜇𝜑
= 2 𝑉+ 2 𝑉 3.27
𝑉 2𝐴 (−∆𝑝) 𝐴 (−∆𝑝) 𝑒𝑞

Compressible cake (Varying particle voidage): Many materials gives rise to compressible
filter cakes. A compressible cake is one whose cake resistance 𝑟𝐶 increases with applied
pressure difference (−∆𝑝). Changes in the cake resistance 𝑟𝐶 is due mainly to the effect on
the cake voidage. See equation below:

150 (1 − 𝜀)2
𝐶𝑎𝑘𝑒 𝑟𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒, 𝑟𝐶 =
𝑥𝑠𝑣 2 𝜀 3

29
References/Further reading: Principle of Mass Transfer and Separation Processes. By
Binary K. Dutta; University Technology Petronas Malaysia

30