Introduction to multicomponent distillation

Most of the distillation processes deal with multicomponent mixtures

Multicomponent phase behaviour is much more complex than that
for the binary mixtures
Rigorous design requires computers
Short cut methods exist to outline the scope and limitations of a
particular process

Introduction to multicomponent distillation

Rigorous methods (Aspen)

L j 1 , xi , j 1 , h
Pj 1 , T j 1

L
j 1

Short-cut methods:

V
j

V j , yi , j , h ,
Pj , T j

Wj
F j , z i , j , h Fj , PjF , T jF
Uj
L j , xi , j , h Lj ,
Pj , T j

Qj

Stage j
V j +1 , y i , j +1 , h Vj+1 ,
Pj +1 , T j +1

(See my notes on the web)

Fenske-Underwood-Gilliland
(+Kirkbride)

Multicomponent distillation in tray towers

Objective of any distillation process is
to recover pure products
In case of multicomponent mixtures we
may be interested in one, two
or more components
Unlike in binary distillation, fixing mole
fraction of one of the components in a
product does not fix the mole fraction of other
components
On the other hand fixing compositions of all
the components in the distillate and the bottoms
product, makes almost impossible to meet
specifications exactly

y1,y2,y3,y4
D

Key components
In practice we usually choose two components
separation of which serves as an good indication
that a desired degree of separation is achieved
These two components are called key components
- light key
- heavy key

There are different strategies to select these key

components
Choosing two components that are next to each other
on the relative volatility: sharp separation

Distributed and undistributed components

Components that are present in both the distillate and
the bottoms product are called distributed components
- The key components are always distributed components
Components with negligible concentration (<10-6) in one
of the products are called undistributed

light non-distributed components

(will end up in the overhead product)

key
components

heavy non-distributed components

(will end up in bottoms product)

Complete design
A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
a) Design a distillation process

n-pentane: 0.04
n-hexane:0.40
n-heptane: 0.50
n-octane: 0.06

condenser

F, z f

100kmol/h
boiler

Complete design
A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Design a distillation process

n-pentane: 0.04
n-hexane:0.40
n-heptane: 0.50
n-octane: 0.06

condenser

F, z f

100kmol/h
boiler

What is design of
a column?
- P (pressure)
- N (stages)
- R (reflux)
- D (diameter)
- auxilary
equipment
(condenser, boiler)

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Pressure consideration:
- what if you were not given P=1atm, how would you choose it?
- how do you validate that P=1atm is appropriate?

condenser

- condenser uses cooling water

(20C). Let say the exit water
temperature is 30C.
- To maintain the temperature delta
at 10C, the dew point can not be
lower than 40C.

F, z f

boiler

- Thus, the dew point of the distillate

has to be at least 40C.
- If not, will need higher pressure

Complete design
A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Design a distillation process

n-pentane: 0.04 LNK

LK
n-hexane:0.40
n-heptane: 0.50 HK
n-octane: 0.06 HNK
F, z f

condenser

100kmol/h
boiler

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance conisderations

xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

3.62

LK

Hexane

0.4

40

1.39

HK

Heptane

0.5

50

0.56

HNK

Octane

0.06

0.23

100

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations

xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

LK

Hexane

0.4

40

39.2

1.39

HK

Heptane

0.5

50

0.5

0.56

HNK

Octane

0.06

6
100

3.62

0.23

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations
- Sharp split: components lighter than the Light Key (LK) will end up completely
in the overheads
xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

3.62

LK

Hexane
LK

0.4

40

39.2

1.39

HK

Heptane
HK

0.5

50

0.5

0.56

HNK

Octane

0.06

0.23

100

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations

xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

0.092

3.62

LK

Hexane
LK

0.4

40

39.2

0.897

1.39

HK

Heptane
HK

0.5

50

0.5

0.011

0.56

HNK

Octane

0.06

0.23

100

D=43.7

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations

Bxi = FxiF Dyi

xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

0.092

3.62

LK

Hexane
LK

0.4

40

39.2

0.897

0.8

1.39

HK

Heptane
HK

0.5

50

0.5

0.011

49.5

0.56

HNK

Octane

0.06

100

D=43.7

0.23

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations
- Sharp split: components heavier than the Heavy Key (HK) will
end up completely in the bottoms
xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

0.092

3.62

LK

Hexane
LK

0.4

40

39.2

0.897

0.8

1.39

HK

Heptane
HK

0.5

50

0.5

0.011

49.5

0.56

HNK

Octane

0.06

0.23

100

D=43.7

B=56.3

Complete short cut design

A mixture of 4% n-pentane, 40% n-hexane, 50% n-heptane and 6%
n-octane is distilled at 1 atm. The goal is to recover 98%of hexane and
1% of heptane in the distillate. The feed is boiling liquid.
- Material balance considerations
- Sharp split: components heavier than the Heavy Key (HK) will
end up completely in the bottoms
xF

F xF

Moles in D

xD

Moles in B

xB

Ki

LNK

Pentane

0.04

0.092

3.62

LK

Hexane
LK

0.4

40

39.2

0.897

0.8

0.014

1.39

HK

Heptane
HK

0.5

50

0.5

0.011

49.5

0.879

0.56

HNK

Octane

0.06

0.107

0.23

100

D=43.7

B=56.3

Gilliland correlation: Number of ideal

plates at the operating reflux
R RDm
N N min

= f D
N +1
RD + 1

Gilliland correlation: Number of ideal

plates at the operating reflux
R RDm
N N min

= f D
N +1
RD + 1
Nmin
Rmin
R=1.5Rmin

Fenske equation for multicomponent

distillations
Assumption: relative volatilities of components remain constant
throughout the column

N min

xD , LK xB , HK
ln

xB , LK xD , HK

=
1
ln LK , HK

LK light component
HK heavy component

K LK (T )
LK , HK (T ) =
K HK (T )

Fenske equation for multicomponent

distillations
Choices for relative volatility:

K LK (T )
LK , HK (T ) =
K HK (T )

T
D

1) Relative volatility at saturated feed condition

LK , HK = F

LK , HK

(TF )

2) Geometric mean relative volatility

LK , HK = D

B
(
T
)

(TB )
D
LK , HK
LK , HK

LK , HK = 3 F

D
B
(
T
)

(
T
)

(TB )
F
D
LK , HK
LK , HK
LK , HK

why geometric mean?

Non key component distribution from

the Fenske equation

x D ,i
x B ,i

N min +1
i , HK

xD , HK

xB , HK

i , HK

DxD , HK
FxF ,i

BxB , HK

=
DxD , HK
N min +1
1 + i , HK

BxB , HK

Ki
=
K HK

N min +1
i , HK

DxD ,i

Convince yourself and

derive for

BxB ,i

Minimum reflux ratio analysis

At the minimum reflux ratio condition

there are invariant zones that occur
above and below the feed plate, where
the number of plates is infinite and the
liquid and vapour compositions do not
change from plate to plate

y
y1

Unlike in binary distillations, in

multicomponent mixtures these zones
are not necessarily adjacent to the feed
plate location

yB

zf

xB xN

zf

xD
x

Minimum reflux ratio analysis

At the minimum reflux ratio condition

y
there are invariant zones that occur
above and below the feed plate, where y1
the number of plates is infinite and the
liquid and vapour compositions do not
change from plate to plate
zf
Unlike in binary distillations, in
multicomponent mixtures these zones
are not necessarily adjacent to the feed
yB
plate location
xB xN

zf

xD
x

Minimum reflux ratio analysis

condenser

F, z f

boiler

Invariant zones: presence of heavy and light non-distributed

components

Minimum reflux ratio analysis

condenser

F, z f

boiler

Invariant zones: only light non-distributed

components

Minimum reflux ratio analysis

condenser

F, z f

boiler

Invariant zones: only heavy non-distributed

components

Minimum reflux ratio analysis

condenser

F, z f

boiler

Invariant zones: no non-distributed

components

Minimum reflux ratio analysis:

Underwood equations
i , HK xF ,i
(1 q ) =
i i , HK
i , HK xD ,i
V
Rm + 1 = =
D
i i , HK

For a given q, and the feed composition

we are looking for A satisfies this equation
(usually
is between LK and HK)

Once
is found, we can calculate the
minimum reflux ratio

Minimum reflux ratio analysis:

Underwood equations

i , HK xF ,i
(1 q ) =
i i , HK

Minimum reflux ratio analysis:

Underwood equations

i , HK xF ,i
0=
i i , HK

1.48

Minimum reflux ratio analysis:

Underwood equations
V

i , HK xD ,i
Rm + 1 = =
D
i i , HK

Minimum reflux ratio analysis:

Underwood equations
V

i , HK xD ,i
Rm + 1 = =
D
i i , HK

2.33

1.48

Distribution of components in
multicomponent distillation process
Feed stage
xi

hexane LK

heptane HK

Non-distributed
heavy non-key
component

pentane

octane
1

10
Non-distributed
Light non-key
component

Kirkbride equation: Feed stage location

N R xF , HK
=
N S xF , LK

xB , LK

x
D , HK

0.206

Complete short cut design:

Fenske-Underwood-Gilliland method
Given a multicomponent distillation problem:
a) Identify light and heavy key components
b) Guess splits of the non-key components and compositions
of the distillate and bottoms products
c) Calculate

LK , HK

d) Use Fenske equation to find Nmin

e) Use Underwood method to find RDm
f) Use Gilliland correlation to find actual number of ideal stages
given operating reflux
g) Use Kirkbride equation to locate the feed stage

Stage efficiency analysis

In general the overall efficiency will depend:
1) Geometry and design of contact stages
2) Flow rates and patterns on the tray
3) Composition and properties of vapour and
liquid streams

Stage efficiency analysis

Lin,xin

Vout,yout

What are the sources of inefficiencies?

Lout,xout

Vin,yin

For this we need to look at what actually happens

on the tray

Local efficiency

Emv

yn yn +1
= *
yn yn +1

Point efficiency

Actual separation
Separation that
would have been
achieved on an
ideal tray

Stage efficiency analysis

stagnation points

Depending on the location on the tray

the point efficiency will vary

The overall plate efficiency can

be characterized by the Murphree
plate efficiency:

high concentration
gradients

low concentration
gradients

When both the vapour and liquid

phases are perfectly mixed the plate
efficiency is equal to the point
efficiency

EmV

yn yn +1
= *
yn yn +1

EmV = Emv

Stage efficiency analysis

In general a number of
empirical correlations exist
that relate point and plate
efficiencies
Peclet number

N Pe

Z L2
=
Detc

eddy diffusivity

length of liquid
flow path

residence time of liquid

on the tray

Stage efficiency analysis: OConnell (1946)

(Sinnott)

Stage efficiency analysis: Van Winkle (1972)

(Sinnott)

Stage efficiency analysis

- AICHE method
- Fair-Chan
Chan, H., J.R. Fair, Prediction of Point Efficiencies for Sieve Trays, 1. Binary Systems,
Ind Eng. Chem. .Process Des. Dev., 23, 814-819 (1984)
Chan, H., J.R. Fair, , Prediction of Point Efficiencies for Sieve Trays, 1. Multi-component Systems,
Ind Eng. Chem. .Process Des. Dev., 23, 820-827 (1984)

(Sinnott)

Stage efficiency analysis

Finally the overall efficiency of the process defined as

EO =

N theoretical
N actual

If no access to the data: E0=0.5 (i.e. double the number of plates)

Column diameter, etc

Sinnott,

Jim Douglas, Conceptual design of chemical process

Column diameter, etc

Column diameter, etc