Design of Liquid-Liquid Extraction Columns

All rights reserved, Armando B. Corripio, Ph.D., P.E., 2014

Cain Department of Chemical Engineering [email protected]
Louisiana State University, Baton Rouge, LA

Contents
Design of Liquid-Liquid Extraction Columns..................... 1

Introduction ................................................. 2

1 Liquid-Liquid Extraction ................................... 2

2 Balances on Liquid-Liquid Extraction ....................... 3

Case A. The Raffinate Product Composition is Specified ..... 7

Case B. The Solute Recovery is specified .................. 8

3 Determining the Number of Required Equilibrium Stages ...... 9

Example 1. Extraction of Acetone from Methyl-Isobutene-Ketone

(MIK)(Taken from McCabe, Smith and Harriott, Unit Operations
of Chemical Engineering, 7th ed., Example 23.3, page 784.) A
countercurrent extraction plant is used to extract acetone
from its mixture with water by means of methyl-isobutene-
ketone (MIK) at a temperature of 25°C. The feed consists of
40 weight% acetone and the balance water. Pure solvent is to
be used and the column operates at 25ºC and the feed flow is
8,000 kg/hr. How many ideal stages are required to extract
99% of the acetone fed using a reasonable solvent rate? What
is the extract composition after removal of the solvent? .. 11

4 Determining the Diameter of a Liquid-Liquid Extraction Column

............................................................ 17

4.1 Column Diameter ....................................... 17

Example 2. Sizing the Acetone-Water-MIK Extraction Column . 19

1
 4.2 Column Height ......................................... 21

5 Insoluble Diluent and Solvent ............................. 21

6 Extraction of Dilute Solutions ............................ 22

7 Parallel Extraction ....................................... 22

Example 3. Batch Parallel Extraction of a Dilute Solution. 22

Summary ..................................................... 26

Review Questions ............................................ 26

Problems .................................................... 28

Introduction
Liquid-liquid extraction is an important unit operation used to
separate liquid components when distillation is difficult and/or
expensive. Many textbook extraction problems are difficult to
solve when the solvent inlet flow is specified, sometimes as a
multiple of the minimum solvent flow; however, engineers in
industry are usually free to choose the solvent flow as a design
variable. The difficulty of solving the liquid-liquid extraction
problem is reduced when the extract product composition is
selected for the design variable instead of the solvent inlet
flow, as will be demonstrated here.

1 Liquid-Liquid Extraction

Liquid-liquid extraction consists of extracting a solute from a

liquid solution using a liquid solvent which is at least
partially immiscible with the solution. This means that the

2
extract, consisting of the solute and the solvent, forms a
separate liquid phase from the raffinate, consisting of the
solute and the diluent. The diluent is the liquid in the
original solution other than the solute.

The equipment from extraction can be a series of

countercurrent stages, each consisting of a mixer and a decanter
or settler, arranged in a countercurrent cascade, or in a tray
or packed tower, as Figure 1 shows.

Figure 1. Extraction Cascade of Mixers and Settlers, and

Extraction Tray Tower

When a tower is used one of the phases, called the

dispersed phase, bubbles through the other, called the
continuous phase. The less dense or light phase flows up the
tower while the denser or heavy phase flows down the tower. The
trays or dispersers can be designed so that either phase can be
the disperse phase.

2 Balances on Liquid-Liquid Extraction

In general, in liquid-liquid extraction all three components are

present in both phases, that is, the diluent is partially
soluble in the extract phase and the solvent is partially

3
soluble in the raffinate phase. This means that two weight
fractions are needed to define each phase (the third fraction is
obtained from the sum of the three weight fractions being equal
to 1.0). In this notes we will use the following notation:

x = weight fraction of a component in the raffinate phase

y = weight fraction of a component in the extract phase

Subscripts: A for the solute

B for the diluent

S for the solvent

Notice that these are capital letters. The lower case letters
are used for the following subscripts: a for the flows and
compositions at the end of the cascade or tower where the feed
enters and the extract leaves, and b for the other end. So, for
example, xAa is the weight fraction of the solute in the feed, yBa
is the weight fraction of the diluent in the extract. Figure 2
shows a schematic of a cascade:

Va Vb
Solvent In yAb
yAa Extract
yBa yBb
ySa ySb
Cascade
La Lb
xAa xAb
xBa Feed Raffinate Out xBb
xSa xSb

Figure 2. Schematic of a Liquid-Liquid Extraction Cascade

4
An extraction tower will be vertical with the same notation as
Figure 2. If the raffinate phase is denser than the extract
phase, the feed will enter at the top of the tower and flow
down, and the solvent at the bottom and flow up. Otherwise the
direction of flow will be reversed. The schematic of Figure 2
represents the complete battery of N stages used in the
separation.

Figure 2 shows that there are twelve variables associated

with the inlet and outlet streams to the cascade, the flow and
two weight fractions of each of the four streams (the third
weight fraction of each stream is determined by difference from
1.0). With three components, solute, diluent and solvent, we can
write three independent mass balances:

Total mass: La + Vb = Lb + Va (1)

Solute mass: LaxAa + VbyAb = LbxAb + VayAa (2)

Diluent mass: LaxBa + VbyBb = LbxBb + VayBa (3)

The solvent balance is not independent of these three. From

experimental data we can obtain

 the solubility of the diluent in the extract, yBa, and

 the solubility of the diluent in the raffinate, xBb.

With these five relationships, we need seven specifications to

define the problem.

In any design problem the feed flow La and two feed weight
fractions, xAa and xBa, are known, as well as two solvent weight
fractions, yAb and yBb, giving us five specifications. (If the

5
feed flow is not known, a basis is used, e.g., 100 kg, as the
number of stages is not dependent on the total feed flow.) The
last two specifications needed are

 either the solute recovery or the solute weight

fraction in the raffinate product xAb, and
 either the solvent rate Vb or the solute weight
fraction in the extract product yAa.

In a regular design problem the solvent flow Vb is a design

variable the engineer can select. Higher solvent flow results in
fewer required stages but a more dilute extract that is more
costly to separate, while lower solvent flow results in a higher
number of stages but a more concentrated extract. The problem is
easier to solve if, instead of selecting the solvent flow, the
extract product composition yAa is selected for the design
variable. This is so because then the diluent composition, yBa,
can be immediately determined from the solubility relationship
or chart. The procedure for selecting the extract product
composition is to look up the extract composition in equilibrium
with the feed composition xAa and select a composition that is
less that it, but not too much less. If the extract composition
is selected as the value in equilibrium with the feed, the
solvent rate will be the minimum solvent rate and the design
will require an infinite number of stages.

There are then two possible versions of the design problem,

one in which the composition of the raffinate product is
specified, and the other in which the solute recovery is
specified. Let us look at each of these in turn.

6
Case A. The Raffinate Product Composition is Specified

Let us first consider the problem where the raffinate product

composition, xAb, is specified. This is also easier to solve than
when the recovery of the solute is specified because the mass
fraction of diluent in the raffinate, xBb, can be immediately
determined from the solubility relationship or chart.

Solution. For the specifications assumed and the solubility

relationships, all of the mass fractions in Equations (1) to
(3), in addition to the feed flow La, are known. The equations
form then a set of linear equations on the three unknown flows:

[ ][ ] [ ] (4)

The solution of this system of equations is:

( ) ( ) ( )
( ) ( ) ( )

( ) ( ) ( )
( ) ( ) ( )

( ) ( ) ( )
( ) ( ) ( )

Many modern calculators, MathCad, MATLAB, and other programs are

capable of solving the system of linear equations (Equation 4).
Given the complexity of the solution formulas given above it is
probably better to use the features of the calculators and
programs to avoid making a mistake in the calculations.

7
Case B. The Solute Recovery is specified

When the recovery of solute is specified instead of the

raffinate product composition, the solubility relationship for
the solvent in the raffinate must be solved along with the three
balance equations. Given that the solute mass fraction in the
raffinate product is also unknown, the equations to be solved
are no longer linear. The solution must then be carried out by
iteration as follows.

First assume the mass fraction of solvent in the raffinate

product, xSb. This is usually a small number near zero and zero
is commonly assumed for the first iteration.

Next, calculate the mass flow of solute in the raffinate

product as the amount not recovered:

LbxAb = (1 – R)LaxAa (5)

Where R is the fraction of the solute recovered. The solute

balance can be solved for the extract product flow when the
solvent inlet stream does not contain solute, that is, when yAb =
0 or negligible:

( )
(6)

Then the remaining two balances can be solved simultaneously for

the flows of the raffinate product, Lb, and inlet solvent, Vb, as
follows:

Total: Vb = Lb + Va - La (7)

Diluent: LaxBa - VayBa = LbxBb – VbyBb (8)

8
Finally the raffinate product composition can be calculated:

(9)

For the next iteration the composition of solvent in the

raffinate, xSb, is obtained at the calculated value of xAb from
the solubility relationship or chart, and the calculations are
repeated. As the solvent concentration in the raffinate is
small, convergence is usually reached on the second iteration.

In summary, the difficulty of solving the balance equations

for a liquid-liquid extraction problem is reduced when the
extract product composition is selected for the design variable
instead of the solvent inlet flow.

3 Determining the Number of Required Equilibrium Stages

As with other countercurrent stage operations the number of

equilibrium stages can be determined by either stage-to-stage
calculations or the McCabe-Thiele graphical procedure. Both of
these methods require the equilibrium and operating
relationships or lines. The equilibrium data is obtained
experimentally and presented in either a correlation or in
graphical form. The operating relationship is obtained from mass
balances around stages 1 to n of the cascade:

9
 Va Vn+1 Vb
yAa V2 V3 Vn VN yAb
yA,n+1
yBa y2 y3 yn yN yBb
yB,n+1
Stage Stage Stage Stage
1 2 n N
La L1 L2 Ln LN-1 Lb
Ln-1
xAa x1 x2 xA,n xN-1 xAb
xn-1
xBa xB,n xBb

Figure 3. Balances Around Stages 1 to n

Total mass: La + Vn+1 = Ln + Va (10)

Solute balance: LaxAa + Vn+1yA,n+1 = LnxA,n + VayAa (11)

Diluent balance: LaxBa + Vn+1yB,n+1 = LnxB,n + VayBa (12)

These equations require an iterative solution; starting with a

value of xA,n, look up the solubility of the diluent in the
raffinate xB,n and assume a value for yB,n+1. Then solve Equations
(10) and (12) simultaneously for Ln and Vn+1 and calculate yA,n+1
(from Equation 10):

(13)

Then look up the solubility yB,n+1 and repeat the calculation

until it converges.

If the equilibrium and operating lines are approximately

straight the number of required equilibrium stages can be
estimated from the Kremser equation:

10
 ( )
(14)
( )

Where xAa* and xAb* are, respectively, the weight fractions of the
raffinate in equilibrium with yAa and yAb.

Example 1. Extraction of Acetone from Methyl-Isobutene-Ketone (MIK)(Taken from

McCabe, Smith and Harriott, Unit Operations of Chemical Engineering, 7th ed.,
Example 23.3, page 784.) A countercurrent extraction plant is used to extract
acetone from its mixture with water by means of methyl-isobutene-ketone (MIK) at a
temperature of 25°C. The feed consists of 40 weight% acetone and the balance water.
Pure solvent is to be used and the column operates at 25ºC and the feed flow is 8,000
kg/hr. How many ideal stages are required to extract 99% of the acetone fed using a
reasonable solvent rate? What is the extract composition after removal of the
solvent?
The solubility chart is shown in Figure 4 below:

11
 50
Acetone-Water MIK at 25°C
45

40

35
xA, yA, weight% acetone

30

25

20

15

10

0
0 10 20 30 40 50 60 70 80 90 100
xB, yB, weight% water

Figure 4. Solubility chart for acetone-water-MIK system at 25°C

In this diagram, typical of Type I systems, the area underneath

the dome contains compositions that form two separate liquid
phases. The point at the top of the dome is called the plait
point; at this point the two phases in equilibrium have the same
weight fractions, xB = yB = 30 weight% water and xA = yA = 48
weight% acetone.

 The curve to the left of the plait point shows the

solubility of the water in the extract phase as a function
of the weight fraction of the acetone in that phase. For
example, at yA = 40 weight% acetone, yB = 9 weight% water.

12
  The curve to the right of the plait point shows the
solubility of the water in the raffinate phase as a
function of the weight fraction of acetone in that phase.
For example, at xA = 30 weight% acetone, xB = 64 weight%
water.
 The tie lines connecting points from the extract phase to
the raffinate phase show the concentrations of the phases
in equilibrium with each other. For example, the fourth
line from the top shows that at yA = 40 weight% acetone in
the extract phase, the concentration of the raffinate phase
in equilibrium with it is xA* = 30 weight% acetone.

You are strongly encouraged to carefully study the chart and

make sure you can locate the points given above as examples.
Your objective must be to be able to read any point in the chart
by yourself.

Vb La
yAb xAa = 0
yBb xBa = 0
Extract
Solvent
99% recovery
Feed Raffinate
Lb
La = 8,000 kg/hr xAb
xAa = 0.40 xBb
xBa = 0.60

Figure 5. Schematic of Extraction Cascade for Example 1.

The unrecovered acetone goes to the raffinate:

LbxAb = (1 – 0.99)(8,000 kg/hr)(0.4) = 32 kg/hr

13
From Figure 4 we see that the composition of the extract in
equilibrium with the feed (xAa = 0.40) is yAa = 0.46. For our
design variable we must select a value less than this, say

yAa = 0.30

From Figure 4 we find: yBa = 0.05.

When the solvent is pure the calculations are simpler

because it does not contain solute, yAb = 0, or diluent, yBb = 0.
So, from the solute balance:

VayAa = LaxAa + VbyAb –LbxAb = 8,000(0.40) + Vb(0) – 32 = 3,168 kg/hr

Thus, Va = 3,168/0.30 = 10,560 kg/hr

For the first iteration we must assume a solubility of the

diluent in the raffinate product. From the chart (Figure 4)
around the area of xAb near zero, we estimate xBb = 0.96. From
Equation (8), the diluent balance:

LaxBa – VayBa = LbxBb - VbyBb

8,000(0.60) – 10,560(0.05) = Lb(0.96) – Vb(0)

Lb = (4,800 – 528 + 0)/0.96 = 4,450 kg/hr

The composition of the raffinate product is then:

xAb = 32/4,450 = 0.0072

From Figure 4, at xAb = 0.0072 we read: xBb = 0.97

Second iteration: Lb = (4,800 – 528 + 0)/ 0.97 = 4,400 kg/hr

xAb = 32/4,400 = 0.0073

From Figure 4, at xAb = 0.0073, xBb = 0.97. We have converged.

14
From the total balance we obtain the solvent rate:

Vb = Va + Lb – La = 10,560 + 4,400 – 8,000 = 6,960 kg/hr

The extract product composition after the solvent is removed is

yAa/(yAa + yBa) = 0.30/(0.30 + 0.05) = 0.857

To determine the required number of equilibrium stages we use

the McCabe-Thiele graphical procedure on the xA-yA diagram. This
requires plotting the equilibrium line, given by the end points
of the tie lines in the solubility chart (Figure 4), as follows:

yA 4.5 10 19 25 33 36 40 42 45.5 47
xA* 2.5 5 10 15 20 25 30 35 40 42.5

These are for each of the tie lines in Figure 4 from the
bottom up.

The operating line goes through the two end points (xAa, yAa)
and (xAb, yAb). To check if the operating line is approximately
straight and an intermediate point is necessary, we calculate
the slope of the operating line L/V at both ends of the
extraction cascade:

La/Va = 8,000/10,560 = 0.76 Lb/Vb = 4,400/6,960 = 0.63

So the operating line is not straight. To calculate the

intermediate point we pick a liquid stream composition near the
middle of the range, say xA = 0.20, and obtain the diluent
composition corresponding to this value from the solubility
chart (Figure 4), as xB = 0.77. For the first trial we guess a

15
diluent composition in the extract of yB = 0.03 and write the
balances around the left portion of the battery:

Total balance: 8,000 + V = L + 10,560

Diluent balance: 100(0.60) + V(0.03) = L(0.77) + 132(0.05)

Solve simultaneously to obtain: V = 8,440 kg/hr, L = 5,880 kg/hr

The solute balance gives the solute mass fraction in the extract
stream:

yA = [5,880(0.20) + 10,560(0.30) – 8,000(0.40)]/8,440 = 0.136

A check on Figure 4 shows that at yA = 0.136, yB = 0.03, as

guessed. The operating line is drawn through the two end points
and the intermediate point to complete the McCabe-Thiele
diagram, as shown below:

50%
45%
40%
35% 3.6 stages
Mass% acetone in extract

30%
25%
20%
15%
10%
5%
0%
0% 10% 20% 30% 40% 50%
Mass% acetone in raffinate

In this solution we used a slightly lower solvent rate than the

solution by McCabe, Smith and Harriott, requiring 3.6 stages as
compared to 3.4 stages in the original problem.

16
 We have seen that when the designer is free to select the
solvent rate in a liquid-liquid extraction problem, the problem
is simplified by selecting the extract composition instead. As
most courses on unit operations are design-type courses, the
problems given to the students should leave them the freedom to
select the design variable that makes it easier to solve the
problem. When the Ponchon-Savarit graphical solution procedure
was eliminated from our textbooks, liquid-liquid extraction
became one of the hardest problems for the students to solve.
The method suggested here reduces some of this difficulty.

4 Determining the Diameter of a Liquid-Liquid Extraction Column

Ref: Seader, Henley and Roper, Separation Process Principles, 3rd

ed., 2011, pp. 334-335.

4.1 Column Diameter

In a liquid-liquid extraction column there are two liquid phases
flowing past each other. The less dense phase flows upward and
the denser phase flows down. One of the phases is dispersed and
the other is continuous. Droplets of the disperse phase move
through the continuous phase.

The sizing procedures consists of determining the sum of

the velocities of the two phases at flooding and then use a
fraction, usually 50% of the flooding value, to size the column.
The velocities are superficial, that is, based on the total
cross-sectional area of the column.

The chart of Figure 6 gives the sum of the superficial

velocities at flooding versus the ratio of the velocities of the
two phases:

17
 0.45
0.40
0.35
0.30
(UD + UC)f/u0

0.25
0.20
0.15
5 4 3 2
0.10 y = -0.0002x + 0.0036x - 0.0298x + 0.1169x - 0.2271x + 0.441
0.05
0.00
0 1 2 3 4 5 6 7
UD/UC

Figure 6. Superficial flooding velocity in a liquid-liquid

extraction column.

UD and UC are the superficial velocities of the dispersed and

continuous phases, respectively. u0 is the characteristic rise
velocity for a single droplet. For systems in which one of the
phases is water, the following dimensionless number has been
determined experimentally:

( 11 )

Where C is the viscosity of the continuous phase, C is the

density of the continuous phase, is the surface tension, and
is the difference in density of the two phases.

Inspection of Figure 6 shows that if the larger of the two

flows is used as the dispersed phase the resulting flooding
velocity uo is higher and thus the column diameter is smaller.

18
Example 2. Sizing the Acetone-Water-MIK Extraction Column
Estimate the height and diameter of a column to carry out the
extraction of Example 1. Use an over-all column efficiency of
20% and a tray height of 12 inches.

Feed Extract
Va = 10,560 kg/hr
Solution: We size the column at the
La = 8,000 kg/hr
yAa = 0.30
xAa = 0.40
yBa = 0.05
top where the feed enters, because
xBa = 0.60
there the flows of the extract and
raffinate phases are highest. The
flows from Example 1 are:

Feed: La = 8,000 kg/hr

Extract: Va = 10,560 kg/hr

The specific gravities of acetone,

Raffinate Solvent
Lb = 4,392 kg/hr Vb = 6,960 kg/hr MIK and water are 0.79, 0.80 and
xAb = 0.0073 yAb = 0
xBb = 0.97 yBb = 0 1.0, respectively (Perry’s Chemical
Engineers’ Handbook, 8th ed., Tables
2-1 and 2-2). Estimate the specific gravity of the feed and
extract assuming additive volumes of acetone, MIK and water:

Feed:

Extract:

Selecting the larger flow, the extract, as the dispersed phase,

the ratio of the velocities is:

19
From the chart of Figure 6 we get, for the sum of the velocities
at flooding,

( )

To get u0, we use the viscosity of feed close to that of water at

25°C, 1 cP (0.001 Pa-s, Perry’s, 8th ed., p. 2-448), feed density
of 900 kg/m3 and extract density of 800 kg/m3 (sp.gr. of 0.90 and
0.80, respectively, determined above); the difference in density
is (900 - 800) = 100 kg/m3, and the surface tension between the
phases is 32 dyne/cm (0.032 N/m, typical of water interfaced
with organics). From Equation (11):

( )( )( )
( )( )

The sum of the superficial velocities of the phases at 50% of

flooding, using the value from Figure 3 (0.28):

( ) ( ) ( )( )( )

The total volumetric flow of the two phases:

⁄ ⁄
( )
⁄ ⁄

Required column area:

⁄
⁄

Required column diameter:

( )
√ ( )

20
As the velocity is the superficial velocity, it is not necessary
to account for the area taken by the downcomers.

4.2 Column Height

In Example 1 we determined the number of required equilibrium
stages is 3.6. With 20% over-all column efficiency and 12-in
tray spacing:

Number of trays: 3.6/0.20 = 18

Column height: 18(12 inch)(1 ft/12 inch) = 18 ft (5.5 m)

5 Insoluble Diluent and Solvent

When the diluent solubility in the extract phase and that of the
solvent in the raffinate phase are negligible, that is yB  0 and
xS  0, then the extraction problem is simplified because it is
no longer necessary to carry two weight fractions for each
stream. Also, the diluent and solvent balances become:

L(1 – xA) = La(1 – xAa) = Lb(1 – xAb)

V(1 – yA) = Va(1 – yAa) = Vb(1 – yAb)

This greatly simplifies the balance calculations. Where have you

seen these equations before?

21
6 Extraction of Dilute Solutions

When the feed and extract solutions are dilute (x << 1, y << 1)
the calculations are further simplified because then the flows
of the raffinate and extract phases are constant from stage to
stage:

L = La = Lb V = Va = Vb

7 Parallel Extraction

Countercurrent extraction is efficient in terms of obtaining the

most solute recovery with the least amount of solvent thus
producing the most concentrated extract. Under certain
conditions parallel extraction is used. In parallel extraction
the raffinate flows from stage to stage but fresh solvent is
used in each stage. Such operation requires more solvent and
produces a more dilute extract, but requires fewer stages,
particularly when the feed is dilute to start with.

Another situation where parallel extraction is used is when

the amounts to be processed are small, as in pharmaceutical
applications. Then the extraction must be carried out in batch
mode and parallel extraction is required, as in the following
example.

Example 3. Batch Parallel Extraction of a Dilute Solution.

A pharmaceutical product must be extracted from a solution
containing 5 gm/L of the product with a solvent having a

22
distribution coefficient of 15. (The distribution coefficient is
the ratio y/x of the concentration of the extract to that of the
raffinate in equilibrium with it.) For 10 L of feed, determine,

a) The amount of solvent required to recover 99% of the

product in one extraction.
b) The amount of solvent required to re recover 99% of the
product in two equal extractions using the same amount of
fresh solvent each time.
c) How many extractions are required to recover 99% of the
product using a 1.0 L of fresh solvent each time?

Solution. Assume the densities of the raffinate and extract are

not functions of composition as the solutions are so dilute.
Assume further that the solvent is not soluble in the raffinate
and the diluent is not soluble in the extract.

Schematic of each extraction:

23
 V
yo = 0

L = 10 L L
xo = 5 mg/L x

V
y

Figure 7. Schematic of Extraction for Example 3

1. One extraction:

Solute balance for one extraction: Lxo + V(0) = Lx + Vy

Distribution coefficient: y = 15x

( )
Combine and solve for V: ( )

For 99% recovery: Lx = (1 – 0.99)Lxo x = 0.01xo

Calculate V = (10/15)[(1/0.01) – 1] = 66 L

2. Two extractions:

24
Balance for each extraction: Lxi + V(0) = Lxi+1 + V(15)xi+1

So,

First extraction:

Second extraction: ( )

For 99% extraction: Lx2 = (1 – 0.99)Lxo So, x2 = 0.01xo

√ √

Solve for V:

V = (L – 0.10L)/[(15)(0.10)] = 0.6L = 0.6(10 L) = 6 L

So, we must use 6 L for each extraction or 12 L total.

Concentration after first extraction: ( )

Concentration after second extraction: ( )

3. Number of extractions for V = 1.0 L

From the formula for part (b): ( ) ( )

( )
Solve for n:
( ( )
)

So it will take five times 1.0 L or 5 L.

Therefore the total volume of solvent required is 66 L for one

extraction, 12 L for two extractions and 5 L for five
extractions.

25
Summary

These notes have presented the design of liquid-liquid

extraction operations. It has been shown that the over-all
balance calculations are simplified when the engineer selects
the extract composition as the design variable instead of the
solvent rate. As with other countercurrent stage operations the
number of required equilibrium stages can be determined by stage
to stage calculations or by the McCabe-Thiele graphical
procedure. When both the equilibrium and operating lines are
approximately linear the Kremser equation provides a good
estimate of the required number of stages.

The calculations are greatly simplified when the diluent is

not soluble in the extract phase and the solvent is not soluble
in the raffinate phase. Further simplification results when the
solutions are dilute because then the rates of the extract and
raffinate phases are approximately constant and the Kremser
equation can be used to determine the number of equilibrium
stages. The operation of parallel extraction was briefly
introduced.

Review Questions

1. Briefly describe the operation of liquid-liquid

extraction.
2. Which are the two phases involved in liquid-liquid
extraction?

26
3. Which are the three components involved in liquid-liquid
extraction?
4. What is the major requirement for the solvent in liquid-
liquid extraction?
5. How many variables are required to define each stream in
liquid-liquid extraction? What are they?
6. How many over-all independent balances can be written in
liquid-liquid extraction?
7. What other two relationships are available in liquid-
liquid extraction?
8. How many specifications are required to design a liquid-
liquid extraction operation? What are they?
9. When left as a design variable, should the solvent rate
or the extract product composition be selected? Why?
10. In the solubility chart of a Type I liquid-liquid
extraction system, what is the plait point? What region
represents the two-phase region? What are the tie lines?
11. How is the equilibrium relationship obtained from the
solubility chart?
12. How is the operating line relationship obtained in
liquid-liquid extraction?
13. What makes the extraction design problem simpler to
solve?
14. What is parallel extraction? When is it used instead of
countercurrent stage contact? What is its advantage? What
is its disadvantage?
15. In batch parallel extraction, what is the advantage of
carrying it out in successive contacts instead of a
single contact?

27
Problems

1. Extraction of Acetone from Water using MIK. You are to

design a column to extract acetone from 3,500 kg/hr of a
feed consisting of 38 weight% acetone and the balance
water. Pure methyl isobutyl ketone (MIK) is to be used as
the solvent and the column operates at 25ºC. It is desired
to extract 98% of the acetone in the feed. Draw a schematic
of the column showing all the problem data and, using a
reasonable solvent rate, determine the flow and composition
of the extract and raffinate streams leaving the column,
and the required number of equilibrium stages. Report also
the composition of the extract on a solvent-free basis
(this is the product composition after the solvent is
removed for recycle). Solubility data at 25ºC are given in
Figure 4 of these notes.
2. Sizing of Acetone-Water-MIK column. For the conditions of
Problem 1 determine the number of actual stages for an
over-all column efficiency of 20%, the column height for 6-
in spacing and 10% extra height, and the column diameter
sized for 50% of the flooding velocity. The viscosity of
MIK is 0.56 cP at 25°C.
3. Design of a Liquid-Liquid Extraction Column. You are to
design a countercurrent extraction battery to extract
acetone from 1,350 kg/hr of a feed consisting of 40 weight%
acetone and 60 weight% water. Pure tri-chloro-ethane (TCE)
is to be used as the solvent and the column operates at
25ºC. It is desired to extract 97% of the acetone in the
feed. Draw a schematic of the column showing all the
problem data and, using a reasonable solvent rate,
determine the flows and compositions of the extract and

28
 raffinate products and the required number of equilibrium
stages. Report also the composition of the extract product
on a solvent-free basis (this is the composition of the
product after the solvent is removed for recycling).
Solubility and equilibrium data at 25ºC are given in
McCabe, Smith & Harriott, 7th ed., page 794. These data
have been plotted below in the following solubility chart.

70

60

50
xA, yA, weight% acetone

40

30

20

10

0
0 20 40 60 80 100
xB, yB, weight% water

4. Sizing of Acetone-Water-TCE column. For the conditions of

Problem 3 determine the number of actual stages for an
over-all column efficiency of 20%, the column height for
12-in spacing and 10% extra height, and the column diameter
sized for 50% of the flooding velocity. The viscosity of
TCE is 0.55 cP at 25°C.

29
 5. Extraction of Sulfuric Acid from Nitrobenzene. The reactor
effluent from the nitration of benzene is to be treated
with water solvent in a counter-current extraction column
to recover the sulfuric acid catalyst. The effluent flows
at the rate of 840 gallons per minute with a composition of
13.6 weight% H2SO4 and the balance nitrobenzene (NB). It is
desired to remove 98% of the acid using a reasonable
solvent rate. The specific gravity of the feed is 1.24 and
the equilibrium chart is given below. Draw a schematic
diagram of the column showing all the problem data and
labeling all the flows and compositions. Using a reasonable
solvent rate, determine the flows and compositions of the
extract and raffinate products, and the required number of
equilibrium stages.

70

60

50
xA, yA, weight% H2SO4

40

30

20

10

0
0 20 40 60 80 100
xB, yB, weight% NB

6. Sizing of Sulfuric Acid Extraction Column. For the

conditions of Problem 5 determine the number of actual

30
 stages for an over-all column efficiency of 20%, the column
height for 12-in spacing and 10% extra height, and the
column diameter sized for 50% of the flooding velocity.
7. Di-methyl formamide (DMF) is used as a solvent for the
production of Nylon. The DMF is to be extracted from waste
water for recycle and to avoid the contamination of the
environment by the water stream. The feed enters at 30,000
kg/hr containing 32 weight% DMF and the balance water. The
solvent contains 1 weight% water and the balance methylene
chloride CH2Cl2. It is desired to produce a raffinate
product with less 0.05 weight% DMF. Draw a schematic of the
column showing all the problem data and, using a reasonable
solvent rate, determine the flows and compositions of the
extract and raffinate products and the required number of
equilibrium stages. Report also the composition of the
extract product on a solvent-free basis (this is the
composition of the product after the solvent is removed for
recycling). The solubility chart is given below.
8. Sizing of DMF extraction column. For the conditions of
Problem 7 determine the number of actual stages for an
over-all column efficiency of 20%, the column height for
12-in spacing and 10% extra height, and the column diameter
sized for 50% of the flooding velocity. The viscosity of
methylene chloride is 0.42 cP at 25°C.

31
 50%

45%

40%

35%
xA, yA, weight% DMF

30%

25%

20%

15%

10%

5%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
xB, yB, weight% water

9. Extraction of Acids from Waste Water. (Adapted from SH&LP

Problem 73.) The water used in a chemical processing plant
contains a mixture of organic acids. The acids are not
particularly toxic to the environment in small
concentrations, but the concentration in the process water
stream is 0.5% by weight leaving the process, and that
concentration is too large for discharge. The acids cannot
be used in the plant for any other purpose, and there is
insufficient acid for their recovery, purification, and

32
 sale. The acids are soluble in hydrocarbons, and it is
decided to extract the acids from the water streams by
countercurrent liquid-liquid extraction, and then use the
hydrocarbon as a fuel for one of the plant process heaters.
There is no chlorine, nitrogen, or sulfur in the acids, so
when they are burned with the fuel, they will not
contribute any additional pollution to the air. The
hydrocarbon used for the extraction is the fuel oil for the
process heaters. Equilibrium data for the acid-water-oil
system are given in the table below. Determine how much oil
will be required to reduce the concentration of acid from
the 0.5 mass percent in the feed to 0.05 mass percent,
which has been found to be acceptable for discharge. The
oil rate used in the process will be 1.5 times the minimum,
and the water to be treated will enter the extraction
system at a rate of 3500 gallons per day. The oil has a
specific gravity of 0.88. Assume the process is to be
performed in a countercurrent liquid-liquid extractor
having an overall efficiency of 20%. How many stages will
be required for the extractor?

Mass percent in water layer Mass percent in oil layer

Acid Water Oil Acid Water Oil
0.05 99.90 0.04 0.10 0.05 99.85
0.10 99.85 0.05 0.20 0.08 99.76
0.20 99.73 0.07 0.40 0.12 99.48
0.30 99.60 0.10 0.60 0.17 99.23
0.40 99.46 0.14 0.80 0.23 98.97
0.50 99.31 0.19 1.00 0.30 98.70

These data are plotted below.

33
 1.20%

1.00%

0.80%
xa, ya, weight% acid

0.60%

0.40%

0.20%

0.00%
0% 20% 40% 60% 80% 100%
xd, yd, weight% water

10. Sizing of acid extraction column. For the conditions of

Problem 9 determine the column height for 12-in spacing and
10% extra height, and the column diameter sized for 50% of
the flooding velocity. The viscosity of the fuel oil is
0.37 cP at 25°C.
11. Extraction of Dilute Solutions. Pharmaceutical products are
recovered from dilute aqueous fermentation broths by
extraction with various solvents. It is desired to recover
80% of a drug from a dilute aqueous solution. The
distribution coefficient KD is 50.

(a) For extraction on a single equilibrium stage, draw a

schematic diagram and determine the required ratio of
solvent to feed solution.
(b) If two sequential equilibrium stages are used using
half the fresh solvent from part (a) in each stage, what

34
 fraction of the drug in the feed will be recovered? Draw
a schematic of the operation.
(c) What fraction of the drug will be recovered if the two
equilibrium stages from part (b) are arranged as a
countercurrent continuous extraction battery using all
the solvent from part (a)? Draw a schematic of the
battery.
State all assumptions.
12. Extraction of Penicillin from Broth. (Based on data of
McCabe, Smith & Harriett, 7th edition, Example 23.2, page
780.) Penicillin F is recovered from a dilute aqueous
fermentation broth by extraction with amyl acetate using 6
volumes of solvent per 100 volumes of aqueous phase. At pH
= 3.2 the distribution coefficient KD is 80. For each part
of the problem draw a schematic and show all the problem
data on the sketch. (a) What fraction of the penicillin
would be recovered by a single ideal stage? (b) What would
be the recovery with two-stage extraction using half of the
fresh solvent in each stage? (c) How many ideal stages
would be required to give the same recovery as in part (b)
in a counter-flow cascade using 6 volumes of solvent per
100 volumes of feed? What if only 3 volumes of solvent per
100 volumes of feed are used in the cascade? What would be
the advantage of using less solvent? State all assumptions.
13. Parallel Extraction of a Dilute Solution. A pharmaceutical
product is recovered from 100 ml of a dilute aqueous
fermentation broth by extraction with amyl acetate. The
distribution coefficient KD is 45 and the broth initially
contains 500 mg of the drug. (a) How much solvent is
required to recover 90% of the drug in a single ideal
stage? (b) How much total solvent is required to recover

35
 90% of the drug in three extractions using the same amount
of fresh solvent each time? Draw a schematic of the
operation and state all assumptions.
14. Parallel Extraction of Dilute Solution. It is desired to
recover a solute from a dilute aqueous solution by
extraction with a solvent using a mass ratio of solvent to
aqueous phase of 25:100. The feed concentration is 0.5
weight% and the distribution coefficient KD for the solute
is 30. (a) What fraction of the solute would be recovered
by a single ideal stage? (b) What would be the recovery
with a three-stage extraction using one third of the fresh
solvent in each stage? Draw a schematic of the operation
and state all assumptions.

36