1

Approximate shortcut methods for

multicomponent distillation
In this chapter we first develop the Fenske
equation, which allows calculation of
multicomponent separation at total reflux. Then
we switch to the Underwood equations, which
allows us to calculate the minimum reflux ratio.
To predict the approximate numbers of
equilibrium stages we then use the empirical
Gilliland correlation that relates the actual
number of stages to the number of stages at
total reflux, the minimum reflux ratio, and the
actual reflux ratio.
2
Total reflux: Fenske Equation
y A1 , y B1 , yC1  x Adist , xBdist , xCdist 
Fenske derived a rigorous
1
solution for binary and
multicomponent distillation
at total reflux. The
derivation assumes that the
stages are equilibrium
stages.
N
For an equilibrium partial
y AR , yBR , yCR 
reboiler for any two components
x AN , xBN , xCN 
x AReb , xBReb , xCReb 
 yA   xA 
   R  
(7-1)
 yB  R  xB  R
3
Material balances around the reboiler are
VR LN
VR y A, R  LN x A, N  Bx A, R (7-2a)
R
VR yB , R  LN xB , N  BxB , R (7-2b)
B
At total reflux B = 0 and LN = VR. Thus the mass
balances become
y A, R  x A , N , y B , R  xB , N (7-3)
For a binary system this means that the
operating line is the y = x line. Combining
Eqs. (7-1) and (7-3),
 xA   xA 
   R   (7-4)
 xB  N  xB  R
4
 Now we move up to stage N the
equilibrium eq. is  y A   xA 
yN xN 1    N  
 yB  N  xB  N
N The mass balances around stage N
yR xN Lx N 1  Vy R  Lx N  Vy N

Since L  V and yR  xN thus

y A, N  x A, N 1 , yB , N  xB , N 1

 xA   xA 
Combining    N   (7-5)
 xB  N 1  xB  N
Then Eqs. (7-4) and (7-5) can be combined
 xA   xA 
    N R   (7-6)
 xB  N 1  xB  R
5
Repeating this procedure until we reach the top
 xA   xA 
   1 2 3  N 1 N R   (7-8)
 xB  dist  xB  R

If we define AB as the geometric average

relative volatility
 AB  1 2 3  N 1 N  R 
1
N min (7-9)

Eq. (7-8) becomes

 xA  N min  x A 
    AB   (7-10)
 xB  dist  xB  R

6
  x A   xA  
ln     
or  xB dist  xB  R  (7-11)
N min 
ln  AB
which is one form of the Fenske equation.
Nmin is the number of equilibrium contacts including the
partial reboiler required at total reflux.
An alternative form of the Fenske equation that is very
useful for multicomponent calculation is easily derived.

 Dx A DxB dist  DxA dist is equal to the

ln  
 Bx A BxB R  fractional recovery of A
N min  in the distillate times the
ln  AB amount of A in the feed.
(7-12)
7
 DxA dist  FRA dist Fz A (7-13)

where FR A dist is equal to the fractional

recovery of A in the distillate. From the
definition of fractional recovery,

Bx A R  1 FRA dist Fz A (7-14)

Substituting Eqs. (7-13), (7-14) and the

corresponding Eqs. For component B into Eq.
(7-12) gives
 FR A distFRB bot 
ln  
        
 
1 FR A dist 1 FR B bot 
N min (7-15)
ln  AB
8
For binary system, Eq. (7-11) becomes
 x 1  x dist 
ln  
   
N min   bot 
x 1 x
(7-16)
ln  AB
Where x = xA is the mole fraction of the more volatile
component. With distillate and bottoms mole fractions
of the more volatile component specified, Nmin is easily
calculated if AB is known. If the relative volatility is not
constant, AB can be estimated from a geometric average
as shown in Eq.(7-9).
avg  1 R 1 2

where R is determined from the bottoms compositions

and 1 from the distillate composition.
9
For multicomponent systems calculation with the
Fenske equation is straightforward if fractional
recoveries of the two keys A and B, are specified.
Eq. (7-15) can be used to find Nmin. Once Nmin is
known, the fractional recoveries of the non-key
can be found by writing Eq. (7-15) for a non-key
component, C and either key component. Then
solve for (FRC)dist or (FRC)bot. When this is done,
Eq. (7-15) becomes

 CB
N
FRC dist  FR 
min
(7-17)
B bot
  CB
Nmin

1 - FRB bot

10
If two mole fractions are specified, say xLK,bot
and xLK,dist, the multicomponent calculation is
more difficult. We can’t use the Fenske
equation directly. but several alternatives are
possible. If we can assume that all non-keys
are nondistributing, we have
DxLNK ,dist  Fz LNK , xLNK ,bot  0
Bx HNK ,bot  Fz HNK , xHNK ,dist  0 (7-18)

Equations (7-18) can be solved with the light

key (LK) and heavy key (HK) mass balance
and the equations
 Dx   D  Bx   B
c c

i ,dist and i , bot (7-19)

i 1 i 1
11
Once all distillate and bottoms compositions or
values for Dxi,dist and Bxi,bot have been found,
Eqs. (7-11) or (7-12) can be used to find Nmin. Use
the key components for the calculation. The
assumption of non-keys can be checked with
Eqs. (7-10) or (7-17). If the original assumption
is invalid, the calculated values for non-key
compositions can be used to calculate the light
and heavy key compositions in the distillate
and bottoms. Then Eq. (7-11) and (7-12) is used
again.

12
Example 7-1. Fenske equation
A distillation column with a partial reboiler and
a total condenser is being used to separate a
mixture of benzene, toluene, and cumene. The
feed is 40 mole % benzene, 30 mole % toluene,
and 30 mole % cumene and is input as a
saturated vapor. We desire 95 % recovery of
the toluene in the distillate and 95% recovery of
the cumene in the bottoms. The reflux is
returned as a saturated liquid, and constant
molal overflow can be assumed. Pressure is 1
atm.

13
Equilibrium can be represented as constant
relative volatilities. Choosing toluene as the
reference component, BT = 2.25 and CT = 2.25.
Find the number of equilibrium stages required
at total reflux and the recovery fraction of
benzene in the distillate.

For A = toluene (LK), B = cumene

N
(HK), C = benzene (LNK) we have
CA  2.25,  AA  1.0,  BA  0.21
zB  0.3, zC  0.4, FRA,dist  0.95, FRB,bot  0.95
1 Find
N at Total reflux and FRC,dist = ?
14
Since operation is at total reflux and relative
volatilities are constant, we can use the Fenske
equation.
Calculate Nmin from Eq. (7-15), and then
calculate FRC,dist from Eq. (7-17)

Equation (7.15) gives

 FR A distFRB bot 
ln  
        
 
1 FR A dist 1 FR B bot 
N min
ln  AB

 0.95 0.95  
ln  
  
    3.77
0. 05 0.05
N min
ln 1 0.21

15
Equation (7-17) gives
 CB
N
FRC dist  FR 
min

B bot
  CB
N min

1 - FRB bot

FRC dist  2.25 0.213.77  0.998

3.77
0.95  2.25 
 
0.05  0.21 

which is the desired fractional recovery of benzene

in the distillate

K K C K A  CA
Note that  CB  C  
KB K B K A  BA

16
Minimum reflux: Underwood Equations

For binary systems, the pinch point (point that

top operating line intersect bottom operating
line on the equilibrium curve) usually occurs
at the feed plate.

For multicomponent systems there will be

separate pinch points in both the stripping
and enriching sections if there are non-
distributed components.

17
 The mass balances in
enriching section
1
Vmin yi , j 1  Lmin xi , j  Dxi ,dist (7-20)
D, xi ,dist
At the pinch point, where
j
compositions are constant
yi , j , V xi , j , L
xi , j 1  xi , j  xi , j 1
F, zi
and yi , j 1  yi , j  yi , j 1 (7-21)

V L Write the equilibrium eq. in term

N of K values as
yi , j 1  K i xi , j 1 (7-22)

Combining Eqs. (7-20) to (7-22) we obtain

B, xi ,bot a balance valid in the region of constant
compositions.
18
 Lmin (7-23)
Vmin yi , j 1  yi , j 1  Dxi ,dist
Ki
Defining the relative volatility  i  Ki K HKand
combining term in Eq.(7-23),
 Lmin 

Vmin yi , j 1 1    Dxi ,dist (7-24)
 Vmin i K HK 

 i Dxi ,dist (7-25)

or Vmin yi , j 1 
Lmin
i 
Vmin K HK
Summing Eq. (7-25) overall components gives
 
c  
c
 i Dxi ,dist
Vmin   Vmin yi , j 1     (7-26)
i 1  
Lmin
i 1
i 
 Vmin K HK 
19
In the stripping section a similar analysis can
be used to derive
 
c  
 i Bxi ,bot
 Vmin    (7-27)
i 1  Lmin 
 i  V K 
 min HK 
Since the conditions in the stripping section
are different than in the rectifying section, in
general  i   i and K HK  K HK
Underwood derived generalized form of Eqs.
(7-26) and (7-27) which is equivalence to
defining
Lmin Lmin
 and   (7-28)
Vmin K HK Vmin K HK
20
Equations (7-26) and (7-27) then become
polynomial in  and  and have C roots.
c
 i Dxi ,dist 
Vmin  (7-29)
i 1 i  
c  i  Bxi ,bot 
Vmin   (7-30)
i 1 i  

If we assume constant molal overflow and

constant relative volatilities  i   i ,
Underwood showed there are common values
of  and () which satisfy both equations.
Eqs. (7-29) and (7-30) can now be added.
Thus at minimum reflux.
21
 c
 i Dxi ,dist   i Bxi ,bot 
Vmin  Vmin     (7-31)
i 1   i    i   

From the overall mass balance

Fzi  Dxi ,dist  Bxi ,bot (7-32)
Substituting into Eq. (7-31)
c
 i Fz i
Vfeed  Vmin  Vmin   (7-33)
i 1  i  

Vfeed is the change in vapor flow rate at the

feed stage. If q is known
Vfeed  F 1  q  (7-34)
If the feed temperature is specified a flash
calculation can be used to determine Vfeed.
22
Equation (7-33) is known as the first
Underwood equation. It can be used to calculate
appropriate values of . Equation (7-29) is
known as the second Underwood equation and
is used to calculate Vmin. Once Vmin is known,
Lmin is calculated from the mass balance

Lmin  Vmin  D (7-35)

23
The exact method for using the Underwood
equations depends on what can be assumed.
Three cases will be considered.

Case A. Assume all non-keys do not

distribute. In this case the amounts of non-
keys in the distillate are:
DxHNK ,dist  0 and DxLNK ,dist  Fz LNK

while the amounts of the keys are:

DxLK ,dist  FRLK dist Fz LK (7-36)

DxHK ,dist  1 FRHK bot Fz HK (7-37)

24
Equation (7-33) can now be solved for the one
value of  between the relative volatilities of the
two keys, HK <  < LK. This value of  can be
substituted into Eq. (7-29) to calculate Vmin from
 i Dxi ,dist 
c
Vmin 
i 1 i  

D   Dxi ,dist 
c
Then (7-38)
i 1

And Lmin is found from mass balance Eq. (7-35)

Lmin  Vmin  D (7-35)

25
Case B. Assume that the distributions of non-keys
determined from the Fenske equation at the total
reflux are also valid at minimum reflux. In this
case the DxNK,dist values are obtained from the
Fenske equation as described earlier. Again solve
Eq.(7-33) for the  value between the relative
volatilities of the two keys. This , the Fenske
values of DxLK,dist and the DxHK,dist and DxNK,dist
values obtained from Eq.(7-36) and (7-37) are
used in Eq.(7-29) to find Vmin. Then Eq.(7-38)
and (7-35) are used to calculate D and Lmin.

26
 c
 i Fz i
Vfeed  Vmin  Vmin  
i 1  i  

Case C. Exact solution without further

assumptions. Equation (7-33) is a polynomial
with c roots. Solve this equation for all values
of  lying between the relative volatilities of all
 Dx
c 
components, V 
min
 
i i ,dist

i 1 i

 LNK ,1  1   LNK , 2  2   LK  3   HK  4   HNK ,1

This gives C-1 roots. Now write Eq.(7-29) C-1

times; once for each value of . We now have c-
1 equations and C-1 unknown (Vmin and Dxi,dist
for all LNK and HNK). Solve these
simultaneous equations and then obtain D from
Eq.(7-38) and Lmin from Eq.(7-35).
27
Example 7-2.Underwood equations
A distillation column with a partial reboiler and
a total condenser is being used to separate a
mixture of benzene, toluene, and cumene. The
feed is 40 mole % benzene, 30 mole % toluene,
and 30 mole % cumene and is input as a
saturated vapor. We desire 95 % recovery of
the toluene in the distillate and 95% recovery of
the cumene in the bottoms. The reflux is
returned as a saturated liquid, and constant
molal overflow can be assumed. Pressure is 1
atm.

28
 29

Equilibrium can be represented as constant

relative volatilities. Choosing toluene as the
reference component, BT = 2.25 and CT = 0.21.
Find the minimum reflux ratio. Use a basis of 100
kg moles/hr of feed.

N We now wish to find (L/D)min

Since the previous problem

showed that benzene does not
1
distribute, the Underwood
equations can easily be used to
estimate the minimum reflux
ratio.
29
 Let A = toluene (LK), B = cumene
N
(HK), C = benzene (LNK)
CA  2.25,  AA  1.0,  BA  0.21
zB  0.3, zC  0.4, FRA,dist  0.95, FRB,bot  0.95
1
This problem fits into Case A or Case
B. We can calculate Dx,dist values as
described in Case A or B, Eqs. (7-36)
and (7-37), and solve Eq. (7-33) for 
where  lies between the relative
volatilities of the two keys 0.21 <  < 1.0
. Then Vmin can be found from Eq. (7-
29), D from Eq. (7-38) and Lmin from
Eq. (7-35).
30
Follow Case B analysis. Since the feed is a saturated
vapor, q = 0 and Vfeed = F(1-q) = F = 100 and Eq.
(7-33) becomes

100 
2.25 40  1.0 30  0.2130 
 
2.25   1.0   0.21  

Solving by trial-and-error for  between 0.21

and 1.0 we obtain  = 0.5454
(or use Goal Seek in Excel)
Equation (7-29) is
 i Dxi ,dist 
where Dxi ,dist  Fzi FR i ,dist
c
Vmin 
i 1 i  

31
 For benzene this is
Dxben,dist  100 0.40.998   39.92

Where the recovery fraction of benzene is the

value calculated in the previous example at
total reflux. The other distillate values are
Dxtot,dist  100 0.30.95   28.5

Dxcum,dist  100 0.30.05   1.5

Summing the three distillate flow, D = 69.92

Equation (9-29) becomes

Vmin 
2.25 39 .92   1.028 .5 
0.211.5  114 .4
2.25  0.5454 1.0  0.5454 0.21  0.5454
32
From the mass balance
Lmin  Vmin  D  44.48
and L Dmin  0.636

The Case A calculation gives the same result

(not shown).

33
Gilliland Correlation for Number of Stages at
Finite Reflux Ratio
Gilliland (1940) noted that he could empirically
relate the number of stages, N at finite reflux
ratio L/D to the minimum number of minimum
of stages Nmin and the minimum reflux ratio
(L/D)min. Gilliland did a series of accurate stage-
by-stage calculations and found that he could
correlate the function (N-Nmin)/(N+1) with the
function [L/D-(L/D)min]/(L/D+1). This
correlation as modified by Liddle (1968) is shown
in Figure 7-3.

34
Figure 7-3 Gilliland correlation as modified by Liddle

35
To use the Gilliland correlation we proceed as follows:
1. Calculate Nmin from the Fenske equation.
2. Calculate (L/D)min from Underwood equation
3. Choose actual L/D. This is usually done as
some multiplier (1.05 to 1.5) times (L/D)min
4. Calculate the abscissa, x
5. Determine the ordinate value, y
6. Calculate the actual number of stage, N.

The Gilliland correlation should only be used for rough

estimates. The calculated number of stages can be off by
30% although they are usually within  7%.

36
The optimum feed plate location can also be
estimated. First, use the Fenske equation to
estimate where the feed stage would be at total
reflux. This can be done by determining the
number of stages required to go from feed
concentration to distillate concentration for the
keys.  x   z 
ln  LK   LK 
 xHK  dist  z HK  (7-40a)
N F ,min 
ln  LK  HK
Now assume that the relative feed location is
constant as we change the reflux ratio from
total reflux to a finite reflux. Thus
N F , min N F (7-40b)

N min N
37
Liddle (1968) fit the Gilliland correlation to
three equations. Let
x
L D  L D min 
L D  1
N  N min
0  x  0.01  1.0  18 .571 x (7-42a)
N 1
0.01  x  0.90
N  N min 0.002743
 0.545827  0.591422 x  (7-42b)
N 1 x
0.9  x  1.0
N  N min
 0.16595  0.16595 x (7-42c)
N 1

Naturally, the equations are useful for computer

calculation. 38
Example 7-3. Gilliland correlation
Estimate the total number of equilibrium stages
and the optimum feed plate location required for
the distillation of the previous problem if the
actual reflux ratio, L/D is set at 2.
F  100, L D  2
We wish to estimate N, NF

An estimate can be obtained from the

Gilliland correlation, while a more exact
calculation could be done stage by stage. We
will use the Gilliland correlation.
39
 L D  L D min
Calculate the abscissa 
L D 1
determine the ordinate N  N min

N 1
from the Gilliland correlation, and then find N .
L Dmin was found in the previous example. The
feed plate location, and Nmin is estimated from
Eqs. (9-41) and (9-40)

L D  L D min 2  0.636
abscissa    0.455
L D 1 2 1

40
 N  N min
ordinate   0.27 using Liddle’s curve
N 1
(Figure 9-3). Since Nmin = 3.77, N = 5.53. From
Eq.(9-40), NF,min is calculated as

 xLK   z LK 
ln     ln  0.408   0.3 
 xHK dist  z HK   0.021 
   
    1.90
0.3
N F ,min 
ln  LK HK ln 1 0.21

xLK ,dist was found from the previous example

Dxtol,dist 28.5
xLK ,dist  xtol,dist    0.408
D 69.92
41
and xHK ,dist  xcum,dist  0.021

Then, from Eq. (7-41)

N F ,min  1.90 
NF  N  5.53   2.79 or stage 3.
N min  3.77 

A check can be obtained from Eq. (7-42b) with

x = 0.455
N  N min
 0.545827  0.591422 0.455  
0.002743
 0.283
N 1 0.455

Substituting Nmin which gives N = 5.65. The 2%

difference between these two results gives an idea
of the accuracy of Eq. (7-42).
42
Multicomponent separations
Distillation
 i Dxi ,d
R
Vmin    A   B  C   D  0
i 1  i  

 A Dx A,dist  B DxB ,dist  C DxC ,dist  D DxD ,dist

Vmin      f  
 A  B  C   D 

f()

0 
4 D 3 C 2 B 1 A

43
 R
 i Bxi ,bot
Stripping section V   
i 1  i  

 A Bx A,bot  B Bx B ,bot  C BxC ,bot  D Bx D ,bot

 Vmin    
 A  B  C   D 

 A Bx A,bot  B Bx B ,bot  C BxC ,bot  D Bx D ,bot

Vmin      f  
  A  B   C  D

f()

V min

0
D 4 C B
2 A 1 
3

44
f()

0 
4 D 3 C 2 B 1 A

f()

V min

0
D 4 C B
2 A 1 
3

45