Chem.

Educator 2002, 7, 349–353 349

Chemical Equilibrium, Free Energy, and Entropy of Mixing

Mauricio Gomes Constantino* and Gil Valdo José da Silva

Departamento de Química, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São
Paulo, Av. Bandeirantes, 3900, 14040-901, Ribeirão Preto – SP, Brazil, mgconsta@usp.br
Received June 3, 2002. Accepted August 13, 2002.

Abstract: In this paper we describe an approach for teaching the relation between chemical equilibrium and free
energy that is, according to our experience, very efficient and enlightening. This approach has a strong visual
appeal and can be used at different levels, from simple presentation of the results as graphs of free energy versus
reaction mixture composition, up to full derivation of formulas to find the equations of the curves.

Introduction abstraction, not easily attainable to the average junior student;

definition of quantities such as the extent of reaction (see note
Teaching chemical equilibrium as a comprehensible [7]) are also required, resulting in a dispersion of the students’
fundamental concept is a challenging task. While the students minds and opening the doors for mistaken interpretations (the
usually understand quite well the reasons for the chemical student can, for instance, start to think that the extent of
equilibrium based on reaction kinetics considerations, the reaction is the fundamental quantity that explains the existence
relationship between equilibrium and thermodynamic of equilibrium).
quantities, particularly the entropy of mixing, is often poorly On the other hand, when we consider a simple system with
understood, largely due to the difficulty of understanding the strongly limited variations, we can go fast and straight to the
connection between mathematical expressions and the physical heart of the problem, and a satisfactory first comprehension
phenomena. (that should later be complemented by considering more
A considerable improvement in clarity has been attained in general approaches) can be easily attained.
this matter by textbooks in the last few decades [1]. A curve of A simple general chemical transformation will be considered
Gibbs free energy versus reaction progress (at constant here:
temperature and pressure), showing a minimum that
corresponds to equilibrium, is now commonplace, and students A + B C + D
can easily understand why the reaction progress brings the
1-x 1-x x x
system to the equilibrium (minimum G) point, regardless of the
side from which it is started. Why this curve has a minimum,
To simplify the calculations, we will say that
however, is still not easy to understand for most students, and
how to plot the curve that would demonstrate the desired (a) all compounds are ideal liquids (∆Hmixing = 0), and they
aspects is still a hard task for many lecturers. Different are all miscible with each other;
approaches have been proposed in articles on this matter: (b) all mixtures to be considered contain no other substance
Cohen and Whitmer [2] give a demonstration largely based on besides A, B, C, and D;
chemical potential of reactants and products; MacDonald [3] (c) all mixtures contain a total amount of 2 mol; and
shows clearly the importance of entropy of mixing in these (d) The Gibbs free energy of (A + B) is higher than (or equal
processes; Shultz [4] also emphasizes the role of entropy to) the free energy of (C + D), or
discussing a specific reaction.
In this paper we propose an approach that requires a (G0
C ) ( )
+ GD0 − GA0 + GB0 ≤ 0 .
minimum of concepts. We found that if we choose a simple
If we start from A (1 mol) and B (1 mol), all possible
system, as the one described ahead, there is no real need to use
mixtures will have the composition summarized in Table 1.
the quantity defined as “extent of reaction,” or “reaction
The value of x can vary from 0 to 1, including the extremes.
progress”; the more intuitive concentration (mole fraction, to
Clearly, the corresponding mole fractions (XA, XB, etc.) are
keep the mathematical transformations to a minimum) can be
obtained by dividing the corresponding amount (mol) by 2
used instead. The concepts of entropy, entropy of mixing, and
(which is the total amount).
Gibbs free energy are also required; particularly important is a
We will say that the free energy of the mixture of reactants
previous knowledge of the connection between Gibbs free
(A + B) (x = 0) is G1, and the free energy of the mixture of
energy and spontaneity of transformations.
products (C + D) (x = 1) is G2 [5]. The question we intend to
answer is “What is the free energy of the other possible
The Proposed Approach compositions?”
Chemical transformations can involve a variable number of A simple way to achieve this objective occurs if we imagine
substances, both as reagents and as products, and a variable that we can prepare mixtures of A, B, C, and D without any
molar proportion among the substances. Derivations based on reaction taking place in the meanwhile. As this is an
general chemical equations usually demand a high level of

© 2002 Springer-Verlag New York, Inc., S1430-4171(02)06613-1, Published on Web 10/4/2002, 10.1007/s00897020613a, 760349mg.pdf
350 Chem. Educator, Vol. 7, No. X, 2002 Constantino and da Silva

Table 1. Composition of the Reaction Mixture What will happen now if we remove the membrane of one
Substance Amount (mol) Mole fraction of these combinations? As the substances are all miscible
A 1–x (1 – x)/2 (assumed as a hypothesis), a complete homogenization will
B 1–x (1 – x)/2 occur.
C x x/2 The spontaneity of the homogenization process already
D x x/2 shows that the resulting mixture has a free energy value (Gm)
that is lower than the corresponding Gt value (observe the
points in Figure 2). As there is no enthalpy of mixing (∆Hmixing
= 0, as assumed), the difference in free energy is due
exclusively to the entropy of mixing (A + B) with (C + D) (a
rather elaborated notation [6] for the entropy of mixing was
A+B used here for the sake of clarity, particularly in the following
1-x mol section).
of each Membrane
A+B+C+D
Gm = Gt − T ∆S(A+B) ↔ (C+D) (2)
C+D
x mol
of each The entropy of mixing, then, is the reason why the graph of
Gibbs free energy versus x is not a straight line like in Figure
2, but a curve with a minimum as discussed in the Conclusion
Figure 1. Mixture vessel with the membrane in place. section. We should remind the reader that this is valid only for
ideal liquids or gases, otherwise the curve would include a
more or less important contribution from ∆Hmixing.
Gt At this point students should have understood the main
points, and in a course for beginners it could be more advisable
to jump from here to the section “Conclusions,” because the
G1 next section involves too many mathematical operations to
(A+B) demonstrate quantitative aspects that are not essential for a
first comprehension of the subject.

Gt The Equation of the Curve

T∆S G2 We could derive expressions to give Gm as a function of XA,

(C+D) XB, XC, or XD; all these mole fractions, however, can be easily
Gm calculated if we know the value of x (see Table 1) [7]; we can,
0.0 0.2 0.4 0.6 0.8 1.0
therefore, find Gm as a function of x. After combining eqs 1
x A+B+C+D
and 2, we have still to find ∆S(A + B) ↔ (C + D) as a function of x.
The entropy of mixing for ideal liquids (or gases), which can
Figure 2. Energy of the several possible combinations of be found in most physical chemistry texts, is
(A + B) with (C + D), before mixing (with the membrane
still in place). ∆Smixing = − nR ∑ X i ln X i
i
exclusively mental operation, it is easy to imagine that
mixtures can be prepared in so short a time that no significant where n is the total amount of substances (in mol) and Xi is the
chemical transformation can occur during the process. mole fraction of component i.
Under these premises, we prepare now in a vessel an Now we will consider the following points:
equimolar mixture of A and B and, in another vessel, an
equimolar mixture of C and D. Then we take an amount of the (a) A and B were already mixed together, the corresponding
first mixture containing (1 – x) mol of A (and, thus, containing entropy being [8] (we remind the reader that there are
also (1 – x) mol of B), and an amount of the second mixture always 1 – x mol of each component)
containing x mol of C (and, thus, containing also x mol of
D).We put these two solutions in the same vessel, which is 1
∆S AA+B
↔ B = −2 R(1 − x )ln
provided with a membrane that keeps the two solutions 2
SEPARATED from each other (see Figure 1). What is the free
energy of this system? Of course, it will be the sum of the free (b) C and D were also already mixed together (x mol of each),
energies of each solution, the corresponding entropy being

Gt = (1 − x ) G1 + xG2 or Gt = (G2 − G1 ) x + G1 (1) 1

∆SCC+D
↔ D = −2 Rx ln
2
In a graph, this equation corresponds to the straight line
between G1 and G2, as in Figure 2.

© 2002 Springer-Verlag New York, Inc., S1430-4171(02)06613-1, Published on Web 10/4/2002, 10.1007/s00897020613a, 760349mg.pdf
Chemical Equilibrium, Free Energy, and Entropy of Mixing Chem. Educator, Vol. 7, No. X, 2002 351

S If the curve has a minimum, in this minimum the value of the

first derivative of the function is zero. The first derivative is
A+B+C+D

dGm
= ∆G + RT ln x 2 − ln (1 − x ) 
2
A + B+ C + D
∆S (A + B) ↔ (C + D) Desired value dx  

+ B +C + D (A + B) ↔ (C + D) In the minimum, x has the equilibrium value xe:

∆S AA↔ B ↔C ↔ D

0 = ∆G + RT ln xe2 − ln (1 − xe ) 
C+D
∆S C↔ D
2

(A + B) ↔ C ↔ D
 
and then
+B
∆S AA↔ B
−∆G
A↔B↔C↔D xe2
=e RT

Figure 3. Relationship among values of entropy of mixing considered

(1 − xe )2
in the text.
where the left expression can be regarded as a representation of
the equilibrium constant Kx, because

∆G = 0 (∆G = G2 - G1) xe xe
XCXD ⋅
G1 G2
Kx = = 2 2
(A+B) (C+D)
X A X B 1 − xe ⋅ 1 − xe
2 2

1 kJ mol
-1 (XA, XB, XC, and XD are, here, the mole fractions of the
substances in the equilibrium state).

Conclusions
0.0 0.2 0.4 0.6 0.8 1.0
x
xe= 0.5 (Kx=1) The equation of the curve of Gibbs free energy versus x for
this system is very simple:
Figure 4. Curve of free energy for ∆G = 0 (at 25 °C).
Gm = x∆G + G1 + 2 RT (1 − x ) ln (1 − x ) + x ln x 
(c) The entropy of the whole mixture, starting from pure (4a)
( ∆G = G2 − G1 ); (0 < x < 1) [10]
components (1 – x mol of A, 1 – x mol of B, x mol of C,
and x mol of D) is [9]
This can be easily introduced in any suitable computer
∆S AA+B+C+D = −2 R[(1 − x )ln(1 − x ) + x ln x ] program for plotting graphs with the desired values of ∆G and
↔ B↔ C ↔ D
T. The lecturer can then use these curves to show how the free
1 1 energy difference affects equilibrium.
− 2 R(1 − x )ln − 2 Rx ln
!" "#"" 2
$ ! "#" $ 2 Observing eq 4a the student can realize that the shape of the
∆S AA+B
↔B ∆SCC+D
↔D curve depends exclusively on the value of ∆G and on the
expression 2 RT (1 − x ) ln (1 − x ) + x ln x , this expression
These three values are interrelated in the way depicted in being independent of energy values; G1, whose value can be
Figure 3, from which we can find the desired value of the only arbitrarily assigned, can just move the curve up or
entropy: downward, without shape modifications (it would be the same
as if we just moved the vertical scale). This means that we can
(1 − x ) ln (1 − x ) + x ln x 
A + B+C+D
∆S(A + B) ↔ (C + D) = −2 R  (3) plot curves, for the sake of observing its shape, for any
reaction with known ∆G (remembering all restrictions),
Now, combining eqs 1 through 3: regardless the impossibility to know the values of G1 (and G2);
we can assign any arbitrary value to G1 because the shape of
the curve would be the same anyway.
Gm = (1 − x ) G1 + xG2 + 2 RT (1 − x ) ln (1 − x ) + x ln x Comparing eq 4a with its equivalent, eq 2, the student can
(4)
0 < x < 1 [10] also realize that the curve representing Gm can be interpreted as
the addition (point by point) of two curves: a straight line
This is the desired equation of the curve and can also be given by the free energy of the extreme cases (Gt, or x∆G +
written as G1) and a curve given by the entropy of mixing
(1 − x ) ln (1 − x ) + x ln x  ; cf. also
A+B+C+D
( −T ∆S(A+B) ↔ (C+D) , or 2 RT 
Gm = x∆G + G1 + 2 RT (1 − x ) ln (1 − x ) + x ln x  eq 3). This curve has the shape depicted in Figure 4, and its
(4a) depth is independent of the energy values; for this reason,
( ∆G = G2 − G1 ); (0 < x < 1) [10] when │∆G│ is large, we have a steep straight line and the

© 2002 Springer-Verlag New York, Inc., S1430-4171(02)06613-1, Published on Web 10/4/2002, 10.1007/s00897020613a, 760349mg.pdf
352 Chem. Educator, Vol. 7, No. X, 2002 Constantino and da Silva

curve due to the entropy of mixing put on the same scale looks
G
-1
much less deep; the addition results in a shallow curve, as can
kJ mol be seen comparing Figures 4–7.
G1 10
(A+B) -1 Figure 5 is the graph of eq 4 (or eq 4a) for values of G1 and
∆G = - 6 kJ mol
9
G2 taken arbitrarily as G1 = 10 kJ mol–1 and G2 = 4 kJ mol–1 at
8 25 °C (298 K). The straight line G1G2 was included as a visual
(∆G = G2 - G1)
7
reference.
Comparing Figures 4 to 7, it is easy to see that, when the
6
free energy difference is high (|∆G| > 40 kJ mol–1, for
5 instance), it is difficult to notice the occurrence of an
Arbitrary equilibrium: the concentrations of the reactants are too low at
values! 4 G2
(C+D) equilibrium and, thus, are hard to detect. The appearance is
3
that the reaction has proceeded to the end, transforming all
2
0.0 0.2 0.4 0.6 0.8 1.0 reactants into products.
x In Figure 4 we see that when ∆G = 0 (rather uncommon for
xe= 0.7706 (Kx=11.28)
chemical reactions but it can occur, for instance, when A is a
chiral organic iodide, C its enantiomer, and B and D are I–), the
Figure 5. Curve of free energy for ∆G = –6 kJ mol–1 (at 25 °C). The equilibrium point is in the center, with equal amounts of
shape of the curve is strongly affected by the value of ∆G. In Figures reactants and products.
5, 6, and 7 are the curves for ∆G = 0, ∆G = –20 kJ mol–1 and ∆G = – Finally, this approach offers the student an opportunity to
40 kJ mol–1. realize the importance of entropy as a determinative factor for
the occurrence of chemical equilibrium, itself one of the most
fundamental concepts in chemistry.

G1 References and Notes

(A+B) ∆G = -20 kJ mol
-1

1. For instance, see Atkins, P. W. Physical Chemistry, 6th ed.; Oxford

(∆G = G2 - G1) University Press: Oxford, 1998; Kotz, J. C.; Treichel Jr., P.
Chemistry & Chemical Reactivity, 3rd ed.; Saunders College
Publishing: Fort Worth, 1987.
2. Cohen, R. W.; Whitmer, J. C. J. Chem. Educ. 1981, 58, 21–24.
3. MacDonald, J. J. J. Chem. Educ. 1990, 67, 380–382.
-1
5 kJ mol 4. Shultz, M. J. J. Chem. Educ. 1999, 76, 1391–1393.
G2
(C+D) 5. Values G1 and G2 are the sum of free energies of the corresponding
substances considering the entropy of mixing: G1 = GA0 + GB0 − T ∆S A+B
0.0 0.2 0.4
x 0.6 0.8 1.0
xe= 0.9827
and G2 = GC0 + GD0 − T ∆SC+D (at constant temperature and pressure). In

(Kx=3217) the considered extreme cases, however, we would be mixing one mol
of each compound (A + B or C + D), and ∆SA+B = ∆SC+D; hence, G2 –
G1 = ∆G0 (the ∆G0 value for the reaction, or the difference of free
Figure 6. Curve of free energy for ∆G = –20 kJ mol–1 (at 25 °C). energy between reactants and products).
6. The solutions mentioned in this text were prepared in steps (first
mixing A with B, then C with D, etc.), each step having its own value
of entropy of mixing. It is, therefore, necessary to distinguish
between entropy values belonging to different steps. The notation
G1 explained below intends to provide an indexing system that would
(A+B) ∆G = - 40 kJ mol
-1
enable the reader to identify to which step each value of entropy
N
(∆G = G2 - G1) belongs: the symbol ∆S M represents the entropy of mixing for a
system in the initial state M and in the final state N; the double
headed arrow Y↔Z indicates that Y and Z are separated, while the
sum signal Y+Z means that Y and Z are mixed together.
7. For this particular system, considering all restrictions, x has, in fact,
-1 the same value as the extent of reaction as defined, for instance, by
5 kJ mol
MacDonald (see ref. 3):
G2
0
(C+D) nM − nM
x =ξ =
0.0 0.2 0.4
x 0.6 0.8 1.0 νM
xe= 0.9997
0
(Kx=1.035 x 10 )
7
nM : initial amount of component M;

nM : amount of component M after some reaction has taken place;

Figure 7. Curve of free energy for ∆G = –40 kJ mol–1 (at 25 °C).
ν M : stoichiometric coefficient of M in the particle equation for the
reaction (negative for a reactant, positive for a product); ν M is

a pure number, ξ is expressed in moles.

© 2002 Springer-Verlag New York, Inc., S1430-4171(02)06613-1, Published on Web 10/4/2002, 10.1007/s00897020613a, 760349mg.pdf
Chemical Equilibrium, Free Energy, and Entropy of Mixing Chem. Educator, Vol. 7, No. X, 2002 353

8. Summary of operations: first and last rows. These values should then be typed in the proper
boxes (they are the same as the column B values for these extreme
cases). After that, we have only to set the program to plot the graph
with column A values in the x axis and both column B and column C
 1− x 1− x 1− x 1− x 
↔ B = − 2 (1 − x ) R 
∆S AA+B +
ln ln  values in the y axis.
!"#" $  2 (1 − x ) 2 (1 − x ) 2 (1 − x ) 2 (1 − x ) 
mol 
1 1 1 1
↔ B = −2 R (1 − x )  ln + ln 
∆S AA+B
2 2 2 2
9. Summary of operations:

∆SAA+B+C+D
↔B↔C↔D

 
 (1 − x) (1 − x) (1 − x) (1 − x) x x x x 
= −2R  ln + ln + ln + ln 
2 2 2 2 2 2 %2 2
!""#""$ !""#""$ % 
 A B C D 


= −2 R (1 − x ) ln
(1 − x ) + x ln x 

 2 2
 1 1
= −2 R (1 − x ) ln (1 − x ) + (1 − x ) ln + x ln x + x ln 
 2 2
Figure 8. Appearance of the computer monitor screen while setting
 1 1
= −2 R (1 − x ) ln (1 − x ) + x ln x  − 2 R (1 − x ) ln + x ln  column B values.
 2 2
Equation 4 or 4a can easily be introduced in a computer program for
plotting graphs (see ahead), but they cannot be used to calculate Gm
for the extreme points (x = 0 and x = 1), because ln 0 would appear in
A+B+C+D
both cases. We know, however, that T ∆S(A+B) ↔ (C+D) = 0 , in
these cases, therefore,
for x = 0, Gm = G1
for x = 1, Gm = G2
10. The graphs on this paper were plotted using the computer program
Microcal Origin 6.0, by Microcal Software Inc.
(http://www.microcal.com/). An example is as follows for the graph
shown in Figure 5 (observe in Figure 8 and Figure 9 the appearance
of the monitor screen during the operations of setting column
values). First, we set column A (x axis) values (of the data table), for
row number (i) 1 to 101, as (i – 1)/100; this will give 101 points for
the curve, evenly distributed in the horizontal direction. Then, we set
column B values (y-axis values for the straight line), also for row
number (i) 1 to 101 as G1+(G2 – G1)*col(A) or, in this specific case,
10–6*col(A). Finally, we set column C (y-axis values for the curve) Figure 9. Appearance of the computer monitor screen after all the
as –col(A)*6 + 10 + 2 *0.00831*298*((1–col(A))*ln(1– column values have been set, but the first row of column C has not yet
col(A))+col(A) *ln(col(A))). If this is set for row number (i) 1 to 101, been corrected.
the program will calculate the values, but will give an error for the

© 2002 Springer-Verlag New York, Inc., S1430-4171(02)06613-1, Published on Web 10/4/2002, 10.1007/s00897020613a, 760349mg.pdf