McCabe-Thiele Method Revisited – Solving Binary Distillation Problems with

Nonconventional Specifications

Luís Gonzaga Sales Vasconcelos1, José Jaílson Nicácio Alves1, Antonio

Carlos Brandão de Araújo2 and Romildo Pereira Brito1,*
1
Department of Chemical Engineering, Federal University of Campina Grande, Av Aprígio
Veloso, 882, Bodocongó, Campina Grande – PB, 58109-970, Brazil,
2
School of Engineering, Laurentian University, 935 Ramsey Lake Road, P3E 2C6, Sudbury,
ON, Canada
*Corresponding author: romildo.brito@deq.ufcg.edu.br

At undergraduate level, the separation of binary mixtures is studied using, mainly, the
McCabe-Thiele Method because of the advantage to present the results in graphical
form. However, most of the applications are limited to problems where extreme (top
and bottom) compositions and reflux ratio are specified and the objective is to
determinate the number of stages. By the use of McCabe-Thiele Method and the
advantage of a solve-package (Mathcad™) a procedure is proposed to solving problems
with different specifications. Unlike the traditional application, the number of stages can
or can not be specified. Depending on the available information, the objective can be the
one of determining top or bottom composition and reflux ratio, for instance. The
proposed procedure allows a detailed analysis of the separation, avoiding the cycle
“right solution or wrong solution”. The proposed procedure is also valid for columns
with multiple feeds and can be implemented by using other softwares.

Keywords – Distillation – McCabe-Thiele Method – Undergraduate – Mathcad™

1. Introduction mentioned above are a very expensive part of

Distillation is one of the most important
separation processes used in chemical and
petrochemical industries due to its great
efficiency in separating different types of
mixtures. Therefore, chemical engineering
curriculum covers one or more courses
involving this subject, which always consider
binary mixtures to begin with.
As computer codes have become more
popular solution of problems involving
distillation calculations has been brought to
pass by using more powerful and efficient
softwares (Aspen™, Hysys™, Design™,
etc.). However, graphic methods for dealing
with problems like those provide a
preliminary overview followed by a deeper
understanding of the distillation process,
which are fundamental pre-requisites to
initiate the studies on the separation of multi
component mixtures. Besides, the softwares
any budget such that not everyone can have
access to those tools.
Among all graphic methods the most
popular are McCabe-Thiele and Ponchon-
Savarit Methods (Henley and Seader, 1981,
King, 1971; Kister, 1990). These methods
differ one another because the later considers
the energy balance by using enthalpy
diagrams. At undergraduate level is common
to consider ideal phases, both liquid and
vapor, in the process, which makes the
supposition of constant molar flow
throughout the section true and turns the
McCabe-Thiele Method into one of the most
widespread used methods.
In most of the applications involving the
McCabe-Thiele Method if the characteristics
of the feed (composition and thermal
condition) to the distillation column are
given, top and bottom composition are
specified as well as the reflux ratio. Another line (Eq. 3) in a y-x diagram and then it
very common application of this method computes the number of stages.
occurs when the recovery ratio of one of the
components, the specification of bottom or q z (1)
y= x−
top composition, and the reflux ratio are q −1 q −1
given. That is, traditionally the McCabe-
Thiele Method is used just to determine the R x (2)
number of stages to achieve a specified y= x+ D
R +1 R +1
separation.
The fact that the McCabe-Thiele Method
L Bx (3)
has been used mainly to determine the y= x− B
V V
number of stages is due to, in a recent past,
problems involving this subject were solved
In Figure 1 a McCabe-Thiele drawing that
by hand (means of grid paper). For a given
ended up in an eight-stage column with the
problem, the equilibrium curve along with
feed at the fourth stage (the stages are
the operation lines was plotted and then the
numbered from bottom to top) can be seen. It
number of stages was obtained. All of those
is also observed that the number of stages is
tasks were done manually. For every new
more than enough to achieve the
operational condition like for example a
specification since the last stage presents a
modification in the reflux ratio, all the steps
composition higher than xD. Therefore, the
had to be repeated.
top product composition is different from
By using the commercial software
that which has been specified. In fact, if the
Mathcad™ a procedure to solve distillation
reflux ratio was manipulated, the desired
problems with different specification cases is
separation could be obtained by using seven
presented. According to this procedure the
stages.
number of stages can be fixed and depending
1
on the available inputs to the system the
xB xD
objective can be, for example, to determine
0.8
the top or bottom composition or moreover to
calculate the reflux ratio. 0.6
Vapor

The procedure can be extended to other

applications involving distillation columns 0.4

with multiple feed points. In the examples

0.2
presented in this paper the necessary
information about composition and thermal 0
0 0.2 0.4 0.6 0.8 1
condition of the feed are available. The total Liquid
condensation of the top is also assumed.
Fig. 1 Typical application of the McCabe-
Thiele Method: calculating the number of
2. Determining the Number of Stages
stages. z=0.5, q=0.5, xD=0.95, xB=0.05, α=3,
In this typical case, the information about
and R=2.5.
top and bottom composition and the reflux
are available. The objective is to determine
3. Determining the Reflux Ratio
the number of stages. In order to perform this
The first application of the proposed
task the procedure looks for the feed line (Eq.
procedure is to be used in problems in which
1), rectification line (Eq. 2), and stripping
top or bottom composition and the number of
stages are specified, and the objective is to as: once the constraints are fulfilled, find the
determine the reflux ratio. values of the unknown variables such that the
In Figure 2 the Mathcad™ code used to objective function is minimized. Needless to
deal with problems with this kind of say, the minimization problem is bound by
specification is shown. Objectively speaking, the determination of four variables, which
this code can be described as following: turns the computation sensible to initial
• Giving inputs to the problem; guesses. This means that the final value of
• Determining the top and bottom flow the objective function has to be always taking
rates; into account.
• Defining the operation lines (rectification Inputs
F := 100 xF := 0.5 q := 0.5 N := 7 α := 3 xD := 0.95 xB := 0.05

and stripping sections), feed point, Computing top and bottom flow rates

diagonal line, and equilibrium curve; D := 50 B := 50

Given
• Defining a objective function (McCabe- F D+ B F ⋅xF D ⋅xD + B ⋅xB

⎛D ⎞ ⎛ D ⎞ ⎛ 50 ⎞
Thiele Method), which is minimized by ⎜ ⎟ := Find( D , B)
⎝B ⎠
⎜ ⎟=⎜ ⎟
⎝ B ⎠ ⎝ 50 ⎠
the internal Mathcad™ Minimize Defining operational, equilibrium, and diagonal lines

function (Conjugate Gradient Method); y( x) :=

α ⋅x
+ x ⋅( α − 1) yfeed ( x) :=
q
⋅x −
xF
ydiag( x) := x
1
q−1 q−1
• Defining the constraints on the Refl xD Lb B ⋅xB
yret ( x , Refl) := ⋅x + ystr ( x , Lb , Vb ) := ⋅x −
Refl + 1 Refl + 1
intersection among the feed line, Vb Vb

Defining the objective function

stripping line, and rectification line; McCabe ( Refl, Lb , Vb , xint) := yint ← yfeed ( xint)

• Defining the constraints on the xxN ← xB

for n ∈ N , N − 1 .. 1

intersection between the stripping line yyn ←

α ⋅xxn
1 + xxn ⋅( α − 1)
and the diagonal; ⎛
xxn− 1 ← ⎜ yyn −
xD ⎞ ⎛ Refl + 1 ⎞
⎟ ⋅⎜ ⎟ if n > 1 ∧ yyn > yint
⎝ Refl + 1 ⎠ ⎝ Refl ⎠
• Defining the constraints on the ⎛
xxn− 1 ← ⎜ yyn +
B ⋅xB ⎞ ⎛ Vb ⎞
⎟ ⋅⎜ ⎟ if n > 1 ∧ yyn < yint
⎝ Vb ⎠ ⎝ Lb ⎠
intersection between the rectification line yy1 − xD

and the diagonal; xint := 0.5 Refl := 4.0 Lb := 80.0 Vb := 100.0

• Defining the constraints on the ratio Given Constraints

yfeed ( xint) yret ( xint, Refl) ystr ( xint, Lb , Vb ) Vb ( Refl + 1) ⋅( F − B) − F ⋅( 1 − q )

between the rectification flow rate and ystr ( xB , Lb , Vb ) ydiag( xB) yret ( xD , Refl) ydiag( xD)

the stripping flow rate; ⎛ Refl ⎞ ⎛ Refl ⎞ ⎛ 3.1458 ⎞

⎜ ⎟ ⎜ ⎟ ⎜
• Minimizing the objective function; ⎜ Lb ⎟ := Minimize( McCabe , Refl, Lb , Vb , xint)
⎜ Vb ⎟
⎟
⎜ Lb ⎟ = ⎜ 207.2921 ⎟
⎜ Vb ⎟ ⎜ 157.2910 ⎟
• Drawing the McCabe-Thiele plot. ⎜ ⎟
⎝ xint ⎠
⎜ ⎟
⎝ xint ⎠ ⎝ 0.4383 ⎠

McCabe ( Refl, Lb , Vb , xint) = 8.746 × 10− 6

If the previous code is to be compared to a McCabe-Thiele drawing

i := 1 .. ( 2 ⋅N + 1) yp1 := xB xp1 := xB yD := xD yint := yfeed ( xint)

traditional application then it can be said that
⎡ ⎡ α ⋅xpi ⎤ ⎤
it presents as a differential the fact that the ⎛ ypi+ 1 ⎞ ⎢
⎢ if ⎢ ypi − xD > 10− 5 ,
⎣ 1 + xpi⋅ ( α − 1)
, ypi⎥
⎦
⎥
⎥
⎜ ⎟ :=
⎝ xpi+ 1 ⎠ ⎢ ⎡ ⎛ B ⋅xB ⎞ ⎛ Vb ⎞ ⎛ xD ⎞ ⎛ Refl + 1 ⎞ ⎤ ⎥
rectification and stripping lines are drawing ⎢ if ⎢ ypi < yint, ⎜ ypi +
⎣ ⎣ ⎝
⎟ ⋅⎜ ⎟ , ⎜ ypi −
Vb ⎠ ⎝ Lb ⎠ ⎝
⎟ ⋅⎜ ⎟⎥
Refl + 1 ⎠ ⎝ Refl ⎠ ⎦
⎥
⎦
as a function of the unknown variables reflux
1

ratio, liquid and vapor flow rate in the xB xD

0.8
rectification section, and bottom
composition. That is exactly the idea of 0.6
Vapor

defining an objective function. 0.4

The reason for an objective function is to 0.2

compare the top composition obtained from 0

0 0.2 0.4 0.6 0.8 1
Liquid
determining the unknown variables with the
specified value given as an input to the Fig. 2 Mathcad™ code used to determine the
problem. In short, the procedure can be read reflux ratio.
 It can be seen in Figure 3a the situation in 4. Determining the Bottom Composition
which the minimization problem does not In this application, information about the
fulfill the constraint of matching the number distillate composition and reflux ratio is
of stages between the top and bottom available as well as the number of stages. In
compositions. Figure 3b presents the Figure 4 is presented the Mathcad™ code
McCabe-Thiele Method where the top and used to solve the problem of finding the top
bottom compositions match properly the composition.
specified number of stages. The reflux ratio
necessary to achieve the right configuration Inputs
F := 100 xF := 0.5 q := 0.5 N := 7 α := 3 xD := 0.95 Refl := 3.1457
is about 3.1458, using the fourth stage as the Defining operational, equilibrium and diagonal lines
feed stage. y ( x) :=
α ⋅x
+ x ⋅ ( α − 1) yfeed ( x) :=
q
⋅x −
xF
ydiag ( x) := x
1
q−1 q−1
The procedure can be implemented by Refl xD Lb B ⋅xB
yret ( x) := ⋅x + ystr ( x , Lb , Vb , xB , B) := ⋅x −
Refl + 1 Refl + 1 Vb Vb
writing the code in languages like Fortran,
Computing the intersection between feed and rectification lines
Matlab® or even the Mathcad™ itself. For xint := 0.5

didactic reasons the authors choose to use the Given

yfeed ( xint) yret ( xint)

internal routines of Mathcad™. xint := Find ( xint) xint = 0.438 yint := yfeed ( xint)

Defining the objective function

McCabe ( Lb , Vb , xB , B) := xx N ← xB
1 for n ∈ N , N − 1 .. 1
xB xD α ⋅xxn
yyn ←
1 + xxn ⋅( α − 1)
0.8
⎛
xxn − 1 ← ⎜ yyn −
xD ⎞ ⎛ Refl + 1 ⎞
⎟ ⋅⎜ ⎟ if n > 1 ∧ yyn > yint
⎝ Refl + 1 ⎠ ⎝ Refl ⎠
⎛
xxn − 1 ← ⎜ yyn +
B ⋅xB ⎞ ⎛ Vb ⎞
⎟ ⋅⎜ ⎟ if n > 1 ∧ yyn < yint
0.6 ⎝ Vb ⎠ ⎝ Lb ⎠
Vapor

yy1 − xD

0.4 xB := 0.5 B := 100.0 Lb := 100.0 Vb := 100.0

Given Constraints

0.2 yfeed ( xint) yret ( xint) ystr ( xint , Lb , Vb , xB , B)

ystr ( xB , Lb , Vb , xB , B) ydiag ( xB) yret ( xD) ydiag ( xD)

0 Vb ( Refl + 1) ⋅( F − B) − F ⋅( 1 − q )
0 0.2 0.4 0.6 0.8 1
⎛ Lb ⎞ ⎛ Lb ⎞ ⎛ 207.2867 ⎞
Liquid ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ Vb ⎟ := Minimize( McCabe , Lb , Vb , xB , B) ⎜ Vb ⎟ = ⎜ 157.2873 ⎟
⎜ xB ⎟ ⎜ xB ⎟ ⎜ 0.0500 ⎟
(a) ⎜ ⎟ ⎜ ⎟ ⎝ 49.9994 ⎠
⎝ B ⎠ ⎝ B ⎠
McCabe ( Lb , Vb , xB , B) = 1.392 × 10− 10
1
McCabe-Thiele drawing
xB xD
i := 1 .. ( 2 ⋅N + 1) yp 1 := xB xp 1 := xB yD := xD yint := yfeed ( xint)
0.8
⎡ ⎡ α ⋅xp i ⎤ ⎤
⎢ if ⎢ yp i − xD > 10− 5 , , yp i⎥ ⎥
⎛ yp i+ 1 ⎞ ⎢ ⎣ 1 + xp i ⋅ ( α − 1) ⎦ ⎥
⎜ ⎟ :=
0.6 ⎝ xp i+ 1 ⎠ ⎢ ⎡ ⎛ B ⋅xB ⎞ ⎛ Vb ⎞ ⎛ xD ⎞ ⎛ Refl + 1 ⎞ ⎤ ⎥
⎢ if ⎢ yp i < yint, ⎜ yp i + ⎟ ⋅⎜ ⎟ , ⎜ ypi − ⎟ ⋅⎜ ⎟⎥ ⎥
⎣ ⎣ ⎝ Vb ⎠ ⎝ Lb ⎠ ⎝ Refl + 1 ⎠ ⎝ Refl ⎠ ⎦ ⎦
Vapor

1
0.4 xB xD

0.8

0.2
0.6
Vapor

0
0 0.2 0.4 0.6 0.8 1 0.4

Liquid
0.2

(b)
Fig. 3 Non-typical application of McCabe- 0
0 0.2 0.4 0.6 0.8 1
Liquid
Thiele Method: calculating the of reflux
Fig. 4 Mathcad™ code used to determine the
ratio. z=0.5, q=0.5, xD=0.95, xB=0.05, α=3,
bottom composition.
and N=7.
 The procedure to be followed is quite
similar to that one presented in the previous
chapter. The difference in this case is that the
top and bottom flow rates can not be
determined. On the other hand, the
intersection between the feed and
rectification lines can be calculated.
By comparing the codes in Figures 2 and
4 it can be observed that in both cases there
exist four variables to be determined. The
main difference lies on the definition of the
operation lines. In Figure 3 the rectification
line has the reflux ratio as the variable to be
found while in Figure 4 all the variables are
part of the stripping line.
In terms of convergence the case shown in
Figure 4 is more stable than that of Figure 2.
In other words, when the information about
rectification and stripping lines are not
available the number of solutions appears to
be larger and not all solutions computed by
the minimization subroutine are physically
attainable.
Problems involving other specification
inputs can be handled by using similar codes
like those presented in Figure 2 and 4. For
example, determining the top composition
having the reflux ratio and bottom
composition as inputs. In any case in order to
minimize the objective function there always
be four variables to be determined regardless
of the specification inputs.
5. Conclusions
A procedure used to extend the
applications of McCabe-Thiele Method to
distillation problems was presented.
Furthermore, the basic idea was to express its
competitiveness and broadness.
By analyzing the cases presented one can
conclude that this procedure can be used in
more complex applications, instead of simply
determining numbers of stages in distillation
columns. At the end one has managed to
achieve a critical analysis of a given problem
and this is a significant part in the formation
of any chemical engineer.
The proposed procedure can be
implemented by using other softwares such
as Matlab™, Mathematica™, and Maple™.
This means there are several opportunities to
make use of computational resources in
applications like that and any effort to take
advantage of this should be always
attempted. However, it is very important that
the softwares quoted above be made more
and more public at the undergraduate level.
Symbols
B – Bottom flow rate
L – Liquid flow rate below the feed stage
N – Number of stages
q – Parameter used to define the thermal
condition of the feed.
R – Reflux ratio
V – Vapor flow rate below the feed stage
x – Liquid phase mole fraction
xD – Top product mole fraction
xB - Bottom product mole fraction
y – Vapor phase mole fraction
z – Feed mole fraction
α – Relative volatility
References
Henley, E. J. and Seader, J. D., Equilibrium-
Stage Separations Operations in Chemical
Engineering, John Wiley & Sons, 1981.
King, C. J., Separation Processes, McGraw-
Hill, 1971.
Kister, R., Distillation Design, McGraw-Hill,
1990.