9.

DISTILLATION
9.1 Introduction
The method of separating the components from a solution depending on its
distribution between a liquid phase and vapour phase is termed distillation. This is
applied to mixtures which distribute in both the phases.
This can also be defined as an operation in which a liquid or vapour mixture
of two or more components is separated into its component fractions of desired
purity, by the application of heat. Thus, in this process, a vapor is obtained from a
boiling mixture which will be richer in components that have lower boiling points.
9.2 Vapour liquid Equilibria (VLE)
The vapor liquid equilibrium data is the basis for distillation operations.
9.2.1 Constant pressure Equilibria
A typical VLE at constant pressure is shown below in Fig. 9.1

B.P of B P
J
I H
G y vs. T
T F
E K
D
C
B.P of A
x vs. T
M

0 x, y 1.0
Fig.9.1: VLE diagram at constant pressure

The upper curve is the dew point curve which provides the relationship between
temperature and mole fraction of the more volatile component in vapor phase(y) and
the lower curve is the bubble point curve which gives the relationship between the
temperature and mole fraction of the more volatile component in liquid phase (x) at a
particular pressure. The horizontal tie lines CD, EF and GH at different temperatures
provide equilibrium compositions of liquid and vapour phase at each temperature.
Any mixture lying on the lower (bubble point) curve will be a saturated liquid and a
mixture lying on the (dew point) upper curve will be a saturated vapor. A mixture
located in between the two curves, say K, will be a two phase mixture of liquid and
vapour with compositions C and D respectively. Their relative amounts are given by

239
 moles of C Length of line KD

moles of D Length of line KC
Consider a mixture at point M. It is only a liquid. If it is kept inside a cylinder fitted
with a frictionless piston and heated, its temperature will increase till it reaches „E‟
when it will become a saturated liquid. The vapour in equilibrium with it will have a
composition of F. As heating is further continued more vaporization takes place, the
liquid phase composition will move towards G and the associated vapour will have a
composition of H. The effective composition of the entire mass comprising both
liquid and vapour continues to remain at M. Finally, when the last droplet of liquid as
indicated at point „I‟ is vaporized, the vapour generated would have a composition of
„J‟. Further application of heat results in super heating of the vapor. During the entire
operation, the pressure is kept constant.

9.2.2 Effect of pressure

As pressure is increased the boiling points of components increase and the
looped curves become more and more narrow. As the critical pressure is exceeded for
one of the components there is no longer a distinction between vapor and liquid for
that component, and for mixtures the looped curves are therefore shorter as depicted
in Fig.9.2, for case (C). Distillation is possible only in the region where a looped
curve exists. It is also clear that relative volatility, α , also changes in such cases.

T B

0 1.0
x, y
Fig.9.2: Effect of pressure on VLE

9.2.3 Constant temperature Equilibria

A typical VLE at constant temperature is shown in Fig. 9.3
As in the case of constant pressure equilibria, lines CD, EF and GH are tie
lines indicating the equilibrium compositions of liquid and vapor phase at various

240
pressures. A liquid defined at point M is a liquid below its bubble point and as the
pressure is reduced at constant temperature, at point „N‟ on the upper (bubble point)
curve, a saturated liquid is obtained. As the pressure is brought down further, at point
Q on the lower (dew point) curve, a saturated vapor forms and a further reduction in
pressure gives a fully superheated vapor as defined by R.

T: constant
Bubble point
Liquid curve
V.P of A
x vs. P
M G
H
E
F Dew
P point
N curve
O
C
D y vs. P
V.P of B Q

0 x, y 1.0

Fig. 9.3: VLE at constant temperature

9.3 Relative volatility ( α )

This is defined as the ratio of vapor pressure of more volatile component to
that of less volatile component. If PA and PB are the vapor pressures of A and B
respectively, the relative volatility of A with respect to B, αAB is defined as the ratio
of vapor pressure of A to that of B.
(i.e.) αAB = PA / PB (9.1)

Raoult’s law states that when a gas and a liquid are in equilibrium, the partial
pressure of „A‟, „pA‟ is equal to the product of its vapor pressure, PA at that
temperature and its mole fraction „xA‟ in the liquid.
(i.e.) pA = PA. xA (9.2)
Similarly pB = PB. xB (9.3)
When the gas and liquid behave ideally, Raoult‟s law holds good.

241
 We know that sum of the partial pressures of components in a gas mixture is
equal to the total pressure, PT. The composition of a component y, in gas phase is
given by Dalton‟s law,
pA p
yA  and yB  B (9.4)
PT PT
 p A   PT y A   y A 
     
PA  x A   x A   y B 
 α AB    
PB  p B   PT y B   x A 
     
 xB   xB   xB 
 yA 
 
α AB  1 yA  (9.5)
 xA 
 
1 xA 
Rearranging, we get
α AB x A
yA  (9.6)
1  x A (α AB  1)
and more simply as
αx
y= (9.7)
1  x(α  1)
9.4 Computation of VLE data (Equilibrium data)
The vapour pressure of the components involved is the basis for the
computation of VLE data.
From Eq. (9.2) and (9.3) pA = xA PA
pB = xB PB = (1 – xA) PB (9.8)
For a binary system,
pA + pB = PT = xA PA + (1 – xA) PB = PB + xA (PA – PB) (9.9)
PT  PB 
xA  (9.10)
PA  PB 
From the vapor pressure data at each temperature, xA can be computed using
Eq. (9.10). After computing xA, the partial pressure pA can be estimated by using
Eq.(9.2). The mole fraction of A in gas phase, yA is then determined by using Eq.
(9.4). Thus, for the whole range of boiling points of components involved, we can
compute VLE data.
Whenever α lies in a narrow range, y can be computed by assuming various
values of „x‟ using Eq. (9.7).

242
9.5 Deviation from ideality:
A mixture whose total pressure is either greater or lesser than that computed
using Raoult‟s law is said to exhibit either a positive deviation or a negative deviation
from ideality.

9.5.1 Positive deviation from ideality

When the total pressure of a mixture is greater than that for ideal mixtures
computed using Raoult‟s law, the mixture is said to exhibit positive deviations from
ideality and such mixtures are called minimum boiling azeotropes. (i.e.) at some
composition the mixture shows minimum boiling point (at constant pressure) and
maximum pressure (at constant temperature) as shown in Fig. 9.4 and 9.5. A typical
x-y diagram is also shown in Fig 9.6. Most of the azeotropic mixtures fall under this
category.

T: constant PT vs. x

PT vs. y PA
Pressure

PB

0 x, y 1.0

Fig. 9.4: Minimum boiling azeotrope at constant temperature

P: constant

BPA
y vs. T
T

BPB

x vs. T
0 x, y 1.0

Fig.9.5: Minimum boiling azeotrope at constant pressure

243
 1.0
P: constant

0 x 1.0

Fig.9.6: VLE of minimum boiling azeotrope

9.5.2 Negative deviations from ideality
When the total pressure of a system is less than the ideal value as computed
using Raoult‟s law, the system is said to deviate negatively. Such systems are very
rare and they are also called as maximum boiling azeotropes (i.e.) at some
composition the mixture shows maximum boiling point. Typical P–x–y, T–x–y and
x–y diagrams are shown in Fig. 9.7, 9.8 and 9.9.

PA
Pressure

PT vs. x

PB
PT vs. y

0 x, y 1.0

Fig.9.7: Maximum boiling azeotrope at constant temperature

244
 y vs. T

A x vs. T

Temperature
B

0 x, y 1.0

Fig. 9.8: Maximum boiling azeotrope at constant pressure

1.0
P: Constant

0 x 1.0

Fig. 9.9: VLE of maximum boiling azeotrope

9.6 Types of distillation columns
Based on the nature of operation, distillation columns have been classified as
batch and continuous columns.
9.6.1 Batch columns
In batch columns, the feed to the column is introduced batch wise and the
distillation is carried. When the desired quality is reached or when the desired
quantity is distilled out, the operation is stopped and next batch of feed is introduced.
9.6.2 Continuous columns
These columns have a continuous feed stream and are capable of handling
high throughputs. These are further classified based on,
 The nature of the feed they are further processing
 Binary columns – Feed has only two components
 Multicomponent column – Feed has more than two components
 The number of product streams they have
 Two product streams
 Multi product streams

245
  The use of additional components in distillation
 Extractive distillation – use of solvent
 Azeotropic distillation – use of entrainer
 The type of columns:
 Tray columns – use of sieve plate columns/Bubble cap trays/Valve trays
for better vapour – liquid contacting
 Packed towers – use of packings in columns for better vapor – liquid
contacting.

DISTILLATION

Batch Continuous

Simple/ Flash/
Differential Equilibrium Stage wise contact Differential
distillation distillation (Tray tower) contact
(Packed
tower)

Sieve Bubble Valve

Tray Cap tray tray

Fig.9.10: Types of distillation and equipments

9.7 Steam distillation:

Some systems have very high boiling points and some of these substances are
unstable at high temperatures. Especially when such systems are completely insoluble
with each other, then steam distillation can be a useful method of separating such
mixtures.
For example, consider a mixture of hydrocarbon and water which are
immiscible. The vapor pressure of either component cannot be influenced by the
presence of the other and each exerts its own vapor pressure at the prevailing
temperature. When the sum of the vapor pressures is equal to the total pressure, the

246
 mixture boils. With vapor pressure data of the individual components, one can also
estimate the temperature at which such distillations take place.
PT = P A + P B (9.11)

Fig 9.11 : Steam Distillation

Insoluble
Water (A)
Compound (B)
Pressure PT

Fig.9.11: Steam Distillation

TB BP of A BP of B
Temperature

Fig.9.11. Steam DIstillation

It is clear from the above Fig 9.11, that this type of distillation takes place at a
temperature which will be even less than that of water. This method suffers from poor
efficiency in its operation, as large quantity of water has to be evaporated. However,
one can introduce effectiveness in such operations by
 Operating at different total pressures in which case the ratio of vapor pressure
of the substances may be more favorable.
 Sparging the mixture with superheated steam or other insoluble gas.

9.8 Differential or simple distillation.

Consider a feed F containing xF mole fraction of more volatile component fed into a
batch still. Let L be the total moles present in the still at any instant„t‟ and „x‟ be the
mole fraction of more volatile component. Let „dL‟ be the moles distilled out. The
concentration of vapor leaving is „y*‟.The moles left behind in the still is (L–dL).
During this process the concentration of more volatile component left behind is
(x – dx).
Total moles of more volatile component present initially is Lx
Total moles of more volatile component in distillate is y* dL
Total moles of more volatile component in residue is (L – dL) (x – dx)

247
 Fig.9.12: Differential Distillation

Making a component balance we get

Lx = y* dL + (L – dL) (x – dx) 0 (9.12)
Lx = y* dL + Lx – Ldx – [Link] + dx dL (9.13)
( Product of two very small quantities)
Then, dL(y* – x) = Ldx (9.14)
dL dx
  (9.15)
L y *  x 
Integrating between limits x = xF L=F
x = xW L=W
F XF
dL dx
   (9.16)
W
L Xw (y *  x)
XF
F dx
ln   (9.17)
W Xw (y *  x)

Eq. (9.17) is called as Rayleigh’s equation.

The right hand side cannot be integrated as y* is a function of „x‟.
Hence, the right hand side of Eq. (9.17) can be evaluated either graphically or
numerically with the help of x – y data.
For systems where the relative volatility lies in a narrow range, then we can use Eq.
αx
(9.7) which states that y =
1  α  1x 
Hence, replacing y in terms of Eq. (9.7) we get

248
  
F  xF
dx 
ln    (9.18)
W xw  αx
 x
 1  α  1x  
 
 dx 
RHS of Eq. (9.18) =    (9.19)
 αx
 x
 1  α  1x  
On simplification,


1  α  1x dx 1  α  1x dx

αx  x  αx  x  xα  1  αx  x
2 2


1  α  1x dx  1  α  1x dx
x1α  1  x α  1  [x1 - x α  1]
 A   B 
 dx  ln 1  x 
dx dx 1
     
x1  x α  1 1  x  α  1  x   1  x 
A1  x   Bx  1  dx dx 
   ln 1  x 
1

α  1  x1  x 
dx  ln 1  x    
α  1  x 1  x 


Substituting the limits for „x‟ as xF and xw

1  
lnx  ln 1  x   ln 1  x   1 ln x F

  ln
1  x F   1  x F 
   ln
α  1 α  1   x W  1  x W  1  x W 
 F 1   xF   1  x F   1  x F 
(i.e.) ln    ln    ln    ln
 W  α  1   x W   1  x W   1  x W 
 xF 
 1  x F    x 
 F  
1
 
ln    ln    α  1 ln  1  x  
W

W  1  x 
W  F
 1  x  
 W 

F1  x F  1  x F   1  x W 
ln  ln   
W1  x W  α  1  1 - x F   x W 
 
α  1ln F1  x F   ln  x F   1  x W 

W1  x W   1 - x F   x W 
 xF 
F1  x F   x   F1  x F  
α ln  ln  W
  ln  
W 1  x W   1  x W    W1  x W 
 1  x  
 F 

 F1  x F    x F  F 
α ln    ln   
 W 1  x 
W   W  W 
x
 F1  x F    Fx F 
α ln    ln  
 W1  x W   Wx W 

249
 α
Fx F  F(1  x F ) 
(i.e.)   (9.20)
Wx W  W(1  x W ) 

The above equation (9.20) is very useful in the estimation of the amount of residue
(alternatively the estimation of the quantity to b distilled) in the case of systems of
constant relative volatility. This is also used in the estimation of relative volatility
for such systems.

9.9 Equilibrium or flash distillation

Consider a feed at a flow rate of „F‟ (moles per hour) , containing the more
volatile component with a composition of ZF and an enthalpy of HF (per mole of feed)
entering a preheater. Let the heat added in the preheater be „Q‟. The mixture then
enters a flash chamber where a distillate leaves at a rate of „D‟(moles per hour) with a
composition of „yD‟ and an enthalpy of HD (per mole of distillate). The bottoms
(residue) leave at a rate of W, with a composition of „xW‟ and an enthalpy of HW (per
mole of residue). The entire process is shown below in the following Fig.9.13
Distillate Vapor
D, yD, HD

(HF + Q/F)
F, ZF, HF

Pressure
Reduction valve
Residue liquid
W, xW, HW
Q

Fig.9.13: Flash distillation

A total material balance, F = W + D (9.22)
A component balance gives FZF = W x W + DyD (9.23)
An enthalpy balance gives FHF + Q = WHW + DHD
 Q
(i.e.) F H F    WH W  DH D (9.24)
 F
From Eqs. (9.22) and (9.23) we get
(W + D)ZF = W x W + DyD (9.25)

 W [ZF – x W ] = −D [ZF – yD] (9.26)

W Z F  y D 
  (9.27)
D Z F  x W 
Similarly from Eqs. (9.22) and (9.24) we get

250
  Q 
 H F  F   H D 
W   
  (9.28)
D  Q 
 H F  F   H w 
  
Dividing Eq. (9.26) by „F‟ we get,

W
Z F  x W    D Z F  y D  (9.29)
F F
Let „f‟ be the fraction of vapor that has been condensed and removed
(1 – f) is the fraction of residue left behind.
 (1 – f) (ZF – x W ) = f (yD - ZF) (9.30)

ZF – x W – fZF +f x W = fyD - fZF

 ZF – x W = f (yD – x W )

ZF + x W (f – 1) = fyD

 yD 
ZF
 xW
f  1 (9.31)
f f
So the Eq. (9.31) can be called as an operating line drawn with a slope of [(f – 1)/f]
and simplified as
 ZF   f  1 
y     x  (9.32)
 f   f 
The Feed point is x = y = ZF
Having seen the principles involved in flash distillation, let us now see how
compositions are estimated in flash distillation operation
9.9.1 Steps:
There are two methods available to estimate the composition of products.
They are explained in detail below.
[Link] Case I
When the equilibrium data and the quantity of distillate or residue and feed are
available, the following procedure shall be adopted :
 Draw the equilibrium curve
 Draw the diagonal.(x = y line)
 Locate feed point corresponding to xF on the diagonal (xF = yF = ZF)
 W
 Draw the operating line with a slope of - 
 D

251
  The intersection of this line with equilibrium curve gives xW and yD as shown in
Fig.9.14.
[Link] Case II
When the enthalpy – concentration data (HL vs x and HG vs y) and heat
added „Q‟ are available, the following procedure shall be adopted.
 Plot the enthalpy concentration data and also equilibrium curve below it.
 Locate the feed point corresponding to F‟ (ZF, HF + Q/F)
 Draw a line by trial and error, passing through „ F‟ ‟ such that it will be a tie
line.
 The points of intersection of this line with enthalpy – concentration. Plot gives
the enthalpy and concentration of both the distillate and residue.
Figures (9.14) and (9.15) represent the procedures followed to determine the product
concentrations through case I and case II respectively.

yD

y ZF
Slope = - W/D

xW ZF
x
Fig.9.14: Estimation of composition of products in Flash distillation

252
 HG vs. y
HD HG vs. y
HF + Q/F
F‟ HL vs. x
HL
HL vs. x
H
F
ZF x, y 1.0

Slope = –W/D

ZF x

Fig.9.15: Enthalpy – concentration diagram

9.12 Continuous rectification

A schematic sketch of a typical distillation column with a feed stream and a
distillate and residue stream is shown below in Fig. 9.16 along with its main
accessories.
9.12.1 Ponchon – Savarit method
There are two methods by which the design of the continuous fractionator can
be established. Let us first consider Ponchon – Savarit method where it requires both
enthalpy and concentration data.

253
 G1, y1, HG1

I
QC

L0, x0 Distillate
G1, y1 D, ZD, HD
Reflux L0
1 x0
HL0
L1 G2
2
Heat L2
Loss

n III

Feed
Ln Gn+1
xn Yn+1
HLn HG, n+1
II

Gm+1
Lm
Ym+1
xm
Heat H IV
HLm G, m+1
Loss
Np

QB Reboiler

W, xW, HW

Fig.9.16: Continuous Fractionator

Envelope: I: Condenser section
Envelope: II: Full distillation unit
Envelope: III: Enriching/Rectifying section
Envelope: IV: Stripping/Exhausting section
The numbering of plates or trays is accounted from the top to bottom. Suffix
denotes the properties of streams leaving a particular plate or tray. Let „n‟ and „m‟,
denote general plates in the enriching section and stripping section respectively.

254
 Let G be the molar flow rate of vapor in enriching section, G the molar flow
rate of vapor in stripping section, L the molar flow rate of liquid in enriching section,
L the molar flow rate of liquid in stripping section, HG the Enthalpy of vapor, HL the
Enthalpy of liquid, y the mole fraction of more volatile component in vapor and x the
mole fraction of more volatile component in liquid.
Let R be the external reflux ratio L0/D, QC the load on condenser, QB the heat
supplied in reboiler and QL the total heat loss.
Considering envelope I and making a mass balance,
G1 = D + L0 (9.45)
G1 = D + RD = D(R + 1) (9.46)
A component balance gives
G1y1 = DZD + L0x0 (9.47)
Making an energy balance gives,
G1HG1 = L0HL0 + DHD + QC (9.48)
 QC = G1HG1 – L0HL0 – DHD (9.49)
Substituting for G1 from Eq. (9.46) get,
QC = [D(R + 1) HG1] – RDHL0 – DHD
= D [(R + 1) HG1 – RHL0 – HD] (9.50)
Considering envelope II and making an energy balance,
Heat in = Heat out
QB + FHF = DHD + WHW + QC + QL (9.51)
 Heat added in reboiler „QB‟ = DHD + WHW + QC + QL – FHF (9.52)
Now, let us consider envelope III, the enriching section and make mass and energy
balance.
A total mass balance yields,
Gn + 1 = Ln + D (9.53)
A component balance gives,
Gn + 1 yn +1 = Lnxn + DZD (9.54)
An energy balance gives,
Gn + 1 HG, n + 1 = [Link] + DHD + QC (9.55)
 (Q  DH D ) 
Let Q‫( = ا‬Net heat out/Net moles out) =  C  (9.56)
 D 
Then, Eq. (9.55) becomes
Gn + 1 HG, n + 1 = [Link] + D Q‫ا‬ (9.57)

255
Eliminating „D‟ in Eq. (9.54) using Eq. (9.53) gives,
Gn + [Link] +1 −Lnxn = (Gn + 1 – Ln) ZD (9.58)
(Gn + 1)[ZD – yn +1] = Ln (ZD – xn)
Ln

ZD  yn  1 (9.59)
Gn  1 ZD  xn 
where (Ln/Gn + 1) is defined as internal reflux ratio
Similarly Eq. (9.53) and (9.57) yields
Ln

Q'HG, n  1 (9.60)
Gn  1 Q'HLn 
Equating Eq.(9.59) with Eq.(9.60) gives,
Ln

ZD  yn  1  Q'HG, n  1 (9.61)
Gn  1 ZD  xn  Q'HLn 
Eq. (9.61) represents a straight line passing through
(HG, n + 1, yn +1) at Gn + 1 (HLn, xn) at Ln and (Q‫ا‬, ZD) at ∆D
where ∆D is called the difference point and it represents
Q‫ا‬: Net heat, out/Net moles, out and
ZD: Net moles of more volatile component, out/Net moles, out
Let us consider Eq. (9.60)
Ln

Q'HG, n  1
Gn  1 Q'HLn 
Substituting for Gn + 1 in the above expression from Eq. (9.53) gives us,
Ln

Q'HG, n  1 (9.62)
Ln  D Q'HLn 
(i.e.)
Ln  D  Q'HL, n 
Ln Q'HG, n  1
D
(i.e.) 1    
Q'HL, n 
 Ln  Q'HG, n  1
D HG, n  1  HL, n 
  (9.63)
Ln Q'HG, n  1

Ln

Q'HG, n  1 (9.64)
D HG, n  1  HL, n 
When n = 0, it indicates the condenser and then for n=0 we get,
L0

Q'HG,1 (9.65)
D HG,1  HL,0 

256
 L0 Lenth of line ΔD.G1
(i.e.) „R‟, the external reflux ratio = = (9.66)
D Lenth of line G1L0
as indicated in Fig. 9.17
Hence, if the reflux ratio „R‟ is known, then it will be easy for us to locate ∆D point
(ZD, Q‟)
Ln ZD  yn  1
Also,  from Eq.(9.59)
D yn  1  xn 
Let us consider envelope IV, in the stripping section
A mass balance yields
Lm = Gm + 1 + W (9.67)
A component balance yields
Lmxm = Gm + 1ym + 1 + WxW (9.68)
 Lm xm − Gm  1 ym + 1 = WxW (9.69)
An energy balance yields
[Link] + QB = Gm + 1HGm + 1 +WHW (9.70)

Let Q" 
WHW  QB (9.71)
W
 [Link] − Gm + 1HG, m + 1 +WQ‫اا‬ (9.72)
Eliminating „W‟ from Eq. (9.67) and (9.69) yield
Lm

ym  1  xW  (9.73)
Gm  1 xm  xW 
Similarly eliminating „W‟ from Eq. (9.67) and (9.72) gives us,
Lm

HG, m  1  Q" (9.74)
Gm  1 HL, m  Q"
Hence from Eq. (9.73) and (9.74) we get
Lm

ym  1  xW   HG, m  1  Q" (9.75)
Gm  1 xm  xW  HL, m  Q"
From Eq. (9.67)

Lm  W 
 1   

Gm  1  m
G  1 
Hence rearrangement of Eq. (9.75) using Eq.(9.67) gives,
W

ym  1  xW  (9.76)
Gm  1 xm  xW 

257

Lm

HG, m  1  Q"  ym  1  xW  (9.77)
W HG, m  1  HL, m  ym  1  xm 
Eq. (9.77) represents a line passing through (xW, Q‫)اا‬, (xm, HL, m) and
(ym + 1, HG, m +1)
where Q" represents: Net heat, out /Net moles out and
xW: moles of A out/Net moles out. Now let us consider the fractionator as single unit
and make mass and energy balances;
Total mass balance gives,
F=D+W (9.78)
A component balance gives
[Link] = DyD + WxW (9.79)
An enthalpy balance gives
[Link] = (DQ‫ ا‬+ WQ‫)اا‬ (9.80)
(Neglecting QL, the heat loss)
Eliminating „F‟ from Eqs. (9.78), (9.79) and (9.80) we get,
D ZF  xW  HF  Q"
  (9.81)
W ZD  ZF  Q'HF 
Eq. (9.81) represents a line passing through (ZD, Q‫)ا‬, (ZF, HF) and (xW, Q‫)اا‬
In other words F = ΔD + ΔW (9.82)
The schematic representation of enthalpy concentration and distribution diagrams for
determination of number of stages is shown in Fig. 9.17.

258
Fig. 9.17 (a) : Determination of number of stages by Ponchon Savarit Method

259
Fig. 9.17 (b) : Determination of minimum reflux in Ponchon Savarit Method

260
[Link] Steps involved:
1) Draw the H vs x, y diagram and the equilibrium curve
2) Locate xF, yD and xW in both the diagrams and draw vertical lines from yD in
„+‟ve y axis direction and from xW in the „-„ve y axis direction.
3) Locate F (xF, HF) in the H – x, y diagram

4) Obtain Q' using the given reflux ratio, R 

Q'HG,1
HG,1  HL,0
where HG1, HL0 corresponds to enthalpy of vapour and liquid both at distillate
composition.
5) In cases where the reflux ratio is not given an optimum reflux ranging from
1.5 to 2 times the minimum reflux can be chosen.
6) To determine the minimum reflux, several lines can be drawn through the feed
point „F‟ in the entire range of x and projected downwards from both H vs x
and H vs. y curves to the x-y diagram as shown in Fig 9.17 (b) and one such
horizontal line in x-y diagram will be a tie line. This line is extended to cut the
vertical line drawn at yD and this cutting point corresponds to the value of Q'
at minimum reflux and the value of R estimated is Rmin.
7) Locate ΔD (yD, Q' ) using the HG1, HL0 and reflux ratio
8) Join ΔD with F and project it to cut the vertical line at xW and that point is ΔW
9) Draw arbitrarily several lines both from ΔD and ΔW to cut both the curves. The
values taken from H vs y give y‫ ا‬and the line from H vs x gives x‫ا‬. For each
line drawn we will have a set of x' and y' values with which we can construct
the operating line in the distribution diagram for both enriching and stripping
sections.
10) Draw the equilibrium curve and plot x' , y' data obtained from step (9)
11) By step wise construction starting from the point D, between the equilibrium
curve and operating line up to W, the number of stages for the desired
separation is determined.
9.12.2 McCabe – Thiele method
When systems exhibit ideal behavior, the time consuming Ponchon – Savarit
method of determing the number of ideal stages, can be replaced with the following
technique. Let us begin our analysis by considering enriching section shown in Fig.
9.18.

261
 G1, y1

L
x0 y1 D, yD
x0 = yD
1
x1
y2

2
x2

yn
n

L
xn yn+1

Fig.9.18: Enriching section of a Fractionator

Assuming the application of equimolar counter diffusion, (i.e.) the molar flow
rates are assumed to be same for both the vapour and liquid steams irrespective of the
stages, we get
G1  G2  Gn + 1  G
A total material balance gives,
G=L+D (9.83)
Let the external reflux ratio „R‟ be given by
L0 L
R  ,
D D
then, G = DR + D = D(R + 1) (9.84)
A component balance for „A‟ gives
Gyn + 1 = Lxn + DyD (9.85)
L D
(i.e.)  yn  1    xn    y D (9.86)
G  G 

262
 L
 
L

L
 D 
R
(9.87)
G L  D  L   D  R 1
  
 D  D
D D 1
  (9.88)
G L  D R 1
 R   1 
 yn  1    xn    yD (9.89)
 R  1  R  1
The equation (9.89) represents the operating line for enriching section, which has
 R   1 
a slope of  and an intercept of  R  1 . If xn = xD= yD then substituting in the
 R  1
above equation (9.89) gives,
 1 
yn  1    RyD  yD   y D (9.90)
 R  1
(i.e.) when, xn = yD, yn + 1 = yD.
Hence, this line passes through x = y = yD. (i.e.) it lies on the diagonal. This point on
 y 
the diagonal and the y intercept  D  permit us the construction of operating line
 R  1
for enriching section.
Let us consider the stripping section shown in Fig. 9.19
ym
m
L G
xm ym+1
m+1
xm+1

yNp
NP

xNP

W, xW

Fig. 9.19 Stripping section

Material balance gives
LGW (9.91)

263
A component balance gives
L xm  G y m1  Wx W (9.92)

L W
 y m1    x m    x W (9.93)
G G

 L 
x m  
W 
(i.e.) y m1    x W (9.94)
LW LW
The above equation (9.94) describes the operating line for stripping section. The
 L 
 and an intercept of – 
W 
operating line has a slope of   .
LW LW
Let us assume that xm = xW (Reboiler)
 L 
 x W  
W 
 y m1     x W  xW (9.95)
LW LW
(i.e.)  xm = ym + 1 = xW Hence, the operating line passes through the point x = y = x W
(i.e. it lies on the diagonal). Having seen the analysis of enriching section and
stripping section separately, let us analyze the feed plate, f, shown in Fig.9.20.

L G
H L, f -1 HG, f
Feed, F
f
GF, yF, HGF
F LF, xF, HLF
ZF, HF
L G
HL, f HG, f +1

Fig.9.20 Feed plate section

A Mass balance on feed plate gives,
F+L+ G =G+ L (9.96)
L  L  G  G  F (9.97)

(i.e.) L  L  G  G   F (9.98)
Enthalpy balance on feed plate gives,
[Link]  [Link], f  1  [Link], f  1  [Link], f  [Link], f (9.99)

264
As an approximation, HG, f  HG, f + 1 = HG and HL, f – 1  HL, f =HL (9.100)
 
 L  L HL  G  G HG  FHF   (9.101)

i.e. L  L HL  G  G  HG  HF (9.102)
F F

Substituting for L  L from Eq. (9.98) we get 



 G  G  F
 HL 

GG
HG  HF
  (9.103)
 F  F

(i.e.)
G  G  H L  HL 
G  G  HG  HF (9.104)
F F
G  G  H L  HG   HF  HL  (9.105)
F

 
 G  G  HF  HL 
  (9.106)
 F  HL  HG 

Substituting for G  G from Eq. (9.98) we get, 
L  L  F  H  H  F L
(9.107)
F HL  HG 

(i.e.)
L  L  H  H   1 F L
(9.108)
F HL  HG 
=
HG  HF  (9.109)
HG  HL 
Let us now, define
HG  HF  as q, where „q‟ is the quantity of heat required to
HG  HL 
convert one mole of feed at its thermal condition to a saturated vapor to the molal
latent heat of vaporization.

q 
L  L  H  HF 
G
(9.110)
F HG  HL 

Similarly
G  G   L  L  1  q  1 (9.111)
F F

(i.e.) G  G  Fq  1  (9.112)
A solute balance above feed plate gives,
Gy  Lx  DxD (9.113)
A solute balance below feed plate gives,

265
Gy  Lx  WxW (9.114)
Subtracting Eq. (9.113) from (9.114) gives,
G  G y  L  L x  Dx D  Wx W  (9.115)
Total component balance for the distillation column gives,
FZF = DxD + WxW (9.116)
Substituting Eqs. (9.110), (9.112) and (9.116) in Eq. (9.115) gives,
F (q – 1) y = Fq.x - FZF
q ZF
y  x (9.117)
q  1 q  1
q
The above Eq.(9.117) is the equation for feed line. It has a slope of and passes
q 1
through y = x = ZF.
The various slope values obtained under different thermal conditions of feed are
given below and shown in Fig. 9.21.

1
2 1. Cold feed
3
2. Saturated liquid
4 3. Liquid + vapor
y 4. Saturated vapor
5 5. Super heated vapor

x
Fig.9.21: Feed line for different Thermal conditions of feed

Feed condition GF LF HG,F HL,F Enthalpy  (H  H F )  q

q G 
of feed, HF  (H G  H L )  q 1

Liquid below 0 F - HF HF<HL >1.0 >1.0

boiling point
Saturated 0 F - HF HF = HL 1.0 
liquid
Liquid + vapor GF LF HG* HL* HG*>HF >HL
*
1.0 to 0 LF
LF  F
Saturated F 0 HF - HF = HG 0 0
vapor
Superheated F 0 HF - HF>HG <0 1.0 to 0
vapor

* indicates HG and HL are enthalpies per mole of individual phases

266
Determination of „q‟ is as follows:
i) Cold feed
From Eq. (9.110)

q
HG  HF
HG  HL 
Let Tb is boiling point of mixture and TF is feed temperature
λ is latent heat of vaporization, CP,L is the specific heat of feed liquid and T0 is the
reference temperature then,
HG  CP, L Tb  T0   λ, HF  CP, LTF  T0  and HG  HL   λ, (9.118)

q
CP, L Tb  T0   λ  CP, LTF  T0  CP, L Tb  T0 - TF - T0  λ (9.119)
λ λ
CP, L Tb  TF 
(i.e.) q  1  (9.120)
λ
ii) Saturated liquid

q
HG  HF (9.121)
HG  HL 
For saturated liquid HF = HL  q = 1.0
iii) Mixture of liquid and vapor
Let „x‟ be the mole fraction of liquid in feed in the case of liquid + vapor mixture
Then,
HF = xHL + (1 – x) HG (9.122)

Therefore, q 
HG  xH L  HG  xH G 
HG  HL 
1.0
x HG  HL 
 (9.123)
HG  HL 
iv) Saturated vapor

q
HG  HF  (9.124)
HG  HL 
For saturated vapor HF = HG  q = 0
iv) Super heated vapor
Let CP,V is the specific heat of feed vapor
HG = HG
HF = HG + Cp, v (TF – Tb) (9.125)

267
 HG  HG  CP, vTF  Tb  CP, vTF  Tb 
q  (9.126)
λ λ

[Link] Steps involved in the determination of number of trays:

The distribution diagram along with the operating lines for both enriching and
stripping sections to determine the number of stages is shown in Fig. 9.22
1.0
D

T
y F
S

yD
R+1 W

0 xw x xF yD 1.0

Fig. 9.22 Determination of number of stages by McCabe-Thiele method

1) Draw the equilibrium curve and diagonal.
2) Locate F, D and W corresponding to feed, distillate and residue compositions
based on more volatile component.
yD
3) Estimate and locate it on y axis as “S”.
R 1
4) Join SD this is the operating line for enriching section.
5) From „F‟ draw q – line depending on feed condition. Let it cut the operating line
for enriching section at T.
6) Join TW – operating line for stripping section.
7) By stepwise construction from „D‟ up to „W‟ the steps are constructed which will
give the number of stages.
9.13 Location of feed tray:
For an optimal design or when a column is designed first (wherein one goes
for optimal design) the feed tray is located at the intersection of operating lines of
enriching and exhausting sections of the tower.
However, we may at times use a column which has been designed with some
other objective. Whenever, the quality and condition of feed is fixed along with reflux
ratio, yD and xW, the operating lines are fixed. It may so happen that in an existing
column, the location of feed nozzle is fixed and it may not really lie at the optimal
point as shown in Fig.9.23 (a; b and c)

268
 Fig. 9.23 (a) Optimal feed location

Fig.9.23 (b) Delayed feed entry

269
 Fig.9.23 (c) Early feed entry
The point of intersection of the two operating lines is generally believed to be the
point that demarcates enriching and exhausting sections. This normally occurs in the newly
designed columns fro a specific xw, yD, xF and condition of feed. However, in an existing
column, designed for a different utility, the feed point location is fixed and may not be at the
optimum location. Further, the feed entry point will not demarcate the enriching and
exhausting sections.
Generally when the reflux ratio and the yD values are fixed, the operating line for
enriching section is fixed. Further when the xw and the condition of the feed are fixed the
operating line for exhausting section is fixed.
Once a feed enters a specific plate, below the point of intersection of operating lines
and q-line (in an existing column), from the top plate to feed entry point, the operating line
for enriching section is to be used and subsequently the operating line for exhausting section.
Such an arrangement indicates a delayed feed entry. If the feed enters at a specific plate,
above the point of intersection of operating lines and q-line (in an existing column), from the
top plate to feed entry point, the operating line for enriching section is to be used and
subsequently the operating line for exhausting section . This arrangement indicates an early
feed entry. In both the cases the number of stages estimated will always be more compared

270
to the number of stages estimated with feed entering exactly at the point of intersection of
operating lines and q-line.
Consider the above three figures,
Fig. (a): Optimal design 9 plates and 5th plate is feed plate.
Fig. (b): An existing column with 10 plates and feed enters at 7th plate.
Fig. (c): An existing column with 10 plates and feed is introduced at 3rd plate.
9.14 Reflux ratio
It is one of the important operating parameters in distillation, by which the
quality of the products can be changed. Let us deal with the relationship between
reflux ratio and number of trays in the tower.
9.14.1 Determination of minimum reflux ratio:
To determine the minimum reflux ratio, draw the q – line from F to cut the
equilibrium curve at T‫ا‬. Join DT‫ ا‬and extend it to intersect on y axis to meet at S‫ا‬. OS‫ا‬
 yD 
gives   from which Rmin is estimated as shown in Fig. 9.24. Normally at
 Rmin  1
Rmin condition the number of stages will be infinity as the equilibrium curve and
operating line get pinched.
1.0
T‫ا‬ D
‫ا‬
S q line

F
y
W

0 xw xF yD 1.0
x
Fig.9.24: Determination of minimum reflux ratio
[Link] Steps involved in the determination of minimum reflux:
1) Draw the equilibrium curve and diagonal.
2) Locate F, D and W corresponding to feed, distillate and residue compositions
based on more volatile component.
3) Draw the q line from F and allow it to intersect the equilibrium curve at T
4) Join T‟D and allow it to intersect the y – axis at S‟.

271
  yD 
5) OS‟ corresponds to   from which Rmin , the minimum reflux ratio is
 Rmin  1
estimated.
9.14.2 Total reflux:
At total reflux, all the distillate is returned back to the column and no distillate
is taken out as distillate product.
(i.e) D = 0
L
R  
D
Hence, the operating line [Eq.(9.89)] for enriching section which is,
 R   1 
yn  1    xn   xD
 R  1  R  1
becomes, yn +1 = xn (9.127)
(i.e) it merges with the diagonal (x = y line) for both enriching and stripping sections.
Under such circumstances, minimum number of theoretical stages can be estimated
by the same graphical procedure described in section [Link].. For systems where the
relative volatility is constant and under total reflux conditions the theoretical number
of stages needed could be estimated analytically.
  yA     yA  
   
  A    B 
x y
We know that for a binary system, α AB
 y  x  
 B   A  
  x B     x B  

yA x
  α AB A
yB xB

yA xA
 α AB
1  yA 1 xA
Let us apply this relationship to (n + 1)th plate
y n 1 x n 1
 α (9.128)
1  y n 1  1  x n 1 
L R
At total reflux D = 0 and   1.0 ,
G R 1
Hence, from Eq.. (9.27)

 (yn + 1) = xn (9.129)

272
 When n = 0 (i.e) at the top of the column, xn = x0
y1 = x0 = yD , when total condenser is used
Substituting for yn +1 in terms of xn in Eq. (9.129) we get
xn x n 1
  α AB (9.130)
1  x n  1  x n 1 
x0 x1
When n = 0,  α AB (9.131)
1  x 0  1  x 1 
x1 x2
When n = 1,  α AB (9.132)
1  x 1  1  x 2 
x n -1 xn
When n = n -1,  α AB (9.133)
1  x n-1  1  x n 
(i.e) Substituting for x1, x2 … xn – 1 from Eqs. (9.132) and (9.133)
x0 xn
 α AB 
n
(9.134)
1  x 0  1  x n 
When n = NP + 1, (Last stage, (i.e.) reboiler) we get,
x0 x NP 1
 α 
NP 1
(9.135)
1  x 0  1  x NP1 
As (NP + 1)th stage accounts for reboiler, x NP 1 = xw

yD xW
 α 
NP 1
(i.e) (9.136)
1  y D  1  x W 
Eq. (9.136) is called Fenske equation. To apply this equation, α , the relative
volatility must be fairly constant and the column has to be operated under total reflux
conditions. This may not be possible in industries, but has theoretical importance.
9.14.3 Optimum reflux ratio:
At minimum reflux ratio the column requires infinite number of stages or
trays. However, as reflux ratio increases from minimum, for a given feed and
specified quality of distillate and residue, the number of stages or trays decrease. At
minimum reflux ratio when the stages or trays are infinite, the fixed cost and
maintenance cost are also infinite. However, the operating cost for operating
condenser, reboiler etc is the least. When the reflux increases, the trays or stages
reduce but the column diameter is to be increased to handle larger capacities of liquid
being recycled. The size of other accessories like condenser and reboiler increase
which will result in a higher requirement of cooling water or heating. Ultimately this
will result in a higher operating cost. Thus, the total cost which includes both

273
 operating cost and fixed cost, vary with reflux ratio and reach a minimum value for a
certain reflux ratio which is called the optimum or economic reflux ratio. This value
is normally in the range of 1.2 Rmin to 1.5 Rmin.

Total cost

Operating cost

Annual cost
Fixed cost

Ropt
Reflux ratio

Fig.9.25: Effect of Reflux ratio on cost

9.19 Azeotropic distillation:

This is a technique which is used for the separation of binary mixtures which
are either difficult or impossible to separate by ordinary fractionation. This happens
when either the mixture to be separated has a very low relative volatility, in which
case one may require high reflux ratio and more number of trays, or when the mixture
forms an azeotrope. Under such circumstances, a third component called an
„entrainer‟ is added to the binary mixture to form a new low boiling hetero azeotrope
with one of the components in the original mixture whose volatility is such that it can
be separated from the other original constituent.
A typical example for this operation is presented in the flow diagram of Fig.
9.31 where the separation of acetic acid (BP: 118.1°C) and water (BP: 100°C)
mixture. This mixture has a low relative volatility and hence separation by
conventional methods is not economical. Here Butyl acetate, which is slightly
soluble in water, is added to the mixture from the top of the column as an entrainer. It
forms a hetero − azeotrope with all the water in the feed and readily distills out from
the high boiling acetic acid and the acetic acid leaves as a residue product. The hetero

274
 azeotrope on condensation forms two insoluble layers which can easily be separated.
The water layer obtained is saturated with ester and vice versa. The ester layer
saturated with water is returned back to the column as a source of entrainer for further
separation. The aqueous layer is also sent to another column to separate water and
ester. The separated ester is also sent back as entrainer.
Sometimes the new azeotrope formed contains all the three constituents. In the
dehydration of ethanol water mixture, benzene is added as an entrainer which gives a
ternary azeotrope containing benzene (53.9 mol %), water (3.3 mole %) and ethanol
(22.8 mol %) boiling at 64.9°C as distillate and ethanol (BP: 78.4°C) as residue.
Benzene is separated and sent back to the top of the column as entrainer. Since water
– ethanol are equally present in distillate, the mixture should be given a preliminary
rectification to produce an alcohol rich binary azeotrope.

9.19.1 Desired properties of an entrainer for Azeotropic distillation

i) Should be cheap and easily available.
ii) Chemically stable and inactive towards the solution to be separated.
iii) Non corrosive
iv) Non toxic
v) Low latent heat of vaporization
vi) Low freezing point to facilitate storage and easy handling
vii) Low viscosity to provide high tray efficiency and minimum pumping cost.

275
 Fig. 9.31 Azeotropic Distillation
9.20 Extractive distillation
This method is also used under similar circumstances as in the case of
azeotropic distillation. Here a third component called solvent is added, instead of
entrainer, which alters the relative volatility of the original constituents, thus
permitting the separation. The added solvent should have low volatility and not
vaporized in the fractionator.
One such example is the separation of toluene (BP: 110.8°C) from isooctane
(BP: 99.3°C). Their separation is relatively difficult. In the presence of Phenol (BP:
181.4°C), the relative volatility of isooctane increases, so that with increase in phenol
content, the separation becomes more and more easy. A typical flow diagram of the
process is shown in Fig. 9.32.
Here, the toluene – isooctane binary mixture is introduced in the middle of the
column and phenol near the top of the column. Isooctane is readily distilled as an
overhead product, while toluene and phenol are collected as residue. The residue from
the tower is rectified in the auxiliary tower to separate toluene and phenol as distillate
and residue respectively. Phenol is returned back to the main column as solvent.
Similarly a mixture of acetone (BP: 56.4°C) and methanol (BP: 64.7°C) can be
separated by using Butanol (BP: 117.8°C) solvent.

276
 Fig. 9.32 Extractive Distillation

9.20.1 Desired properties of solvent for extractive distillation

i) High selectivity and capability to alter VLE for easy separation.
ii) Ability to dissolve the components in the mixture
iii) Low volatility in order to prevent vaporization of solvent
iv) Easy separability, for easier removal of solvent
v) Non corrosive
vi) Non toxic
vii) Cheap and easily available
viii) Low freezing point
ix) Low viscosity
x) Chemical stability and inertness towards the components to be separated.
9.21 Comparison of Azeotropic and extractive distillation
In both the processes an additional external agent is added, which is
undesirable. Solvent to feed ratio in extractive distillation greater than 3 or 4 is found
to be effective. Proper choice of material of construction and recovery technique are
to be examined. However, of the two, extractive distillation is said to be more

277
 favoured than azeotropic distillation since i) there is a greater choice of solvent ii)
smaller quantity of solvent to be volatilized. In spite of the above advantages, the
azeotropic distillation is said to be more effective in the dehydration of ethanol from
an 85.6 mole% Ethanol – water solution. In this case water is azeotroped with n –
pentane and then separated rather than using extractive distillation with ethylene
glycol as solvent.

9.22 Low pressure distillation:

Whenever, the heat sensitive materials are to be separated, as in the case of

many organic mixtures, low pressure distillation will be effective. In this the time of
exposure of the substances to high temperature is kept minimum. Packed towers can
be used for distillation under pressures of 50 to 250 mm Hg. Bubble cap and sieve
trays can be used for pressure drops around 2.6 mm Hg and shower trays for pressure
drops around 0.75 mm Hg. This is used in the separation of vitamins from animal and
fish oils as well as the separation of plasticizers.

9.22.1 Molecular distillation:

This is a very low pressure distillation where the absolute pressure is in the
range of 0.003 mm Hg to 0.03 mm Hg. On reducing the absolute pressure, the mean
free path of the molecules becomes large. If the condensing surface is placed at a
distance from the vaporizing liquid, surface not exceeding few cms, only few
molecules will return to the liquid. The composition of the distillate will now be
different from that given by normal vaporization and the ratio of the constituents is
given by
  pA  
 0.5 

NA moles of A   MA  
  (9.153)
NB moles of B   pB  
 
  MB0.5  

In molecular distillation this ratio is maintained by allowing the liquid to flow

in a thin film over a solid surface thus renewing the surface continually and at the
same time maintaining low holdup of liquid.
A schematic arrangement of a device used for this type of distillation is shown
below in Fig.9.33. The degassed liquid to be distilled is introduced at the bottom of

278
the inner surface of the rotor, rotating at 400 to 500 rpm. A thin layer of liquid 0.05 to
0.1 mm spreads over the inner surface and travels rapidly to the upper periphery
under centrifugal force. Heat is supplied to the liquid through the rotor from radiant
electrical heaters. Vapors generated are condensed and collected in the collection
troughs. The residue liquid is collected in the collection gutter and removed. The
entire unit is maintained at low pressure good enough for molecular distillation to
occur. Normal residence time is of the order of 1 second and hence decomposition of
mixture does not take place. Multiple units can be used to have multistage separation
effects.

Fig. 9.33 Molecular Distillation

279