In the Laboratory
A Physical Chemistry Lab Project:
The Effect of Composition on Several Physical Properties
of Binary Mixtures of Common Liquids
Luther E. Erickson and Kevin Morris
Department of Chemistry, Grinnell College, Grinnell, IA 50112
The effect of the composition of solutions on their
physical properties has been an important focus of both
experimental and theoretical investigations in physical
chemistry for many years. For several years our physical
chemistry laboratory program included a multipart project
that involved the determination of the effect of composi-
tion on the viscosity and partial molar volume for a bi-
nary mixture and the construction of the vaporliquid
phase diagram for the same mixture. Based on the sug-
gestion of Fung et al. (1), we recently expanded the as-
signment to include determining the effect of composi-
tion on the thermodynamics of mixing of the mixture. Ap-
propriate selection of components for the mixtures insures
interesting variations in these properties with composi-
tion. Comparison of the whole range of properties invites
a consistent interpretation of the observed composition
dependence of these properties at the molecular level.
The entire project can be completed in three 3-hour
laboratory periods with students working in pairs. An ad-
ditional period is allocated to complete workup of the data
and preparation of a formal report. Coming early in the
first semester of the course, the workup of the data pro-
vides a good introduction to (or review of) the use of spread-
sheet software and computer graphing of data. This year
we added a poster presentation to encourage comparison
of results among the groups and to provide opportunities
for students to give oral reports on their investigations.
Experimental Details
Effect of Composition on Density, Viscosity, and
Molar Volume
Students are instructed to prepare 25 or 50 mL of
five mixtures and two pure liquids, a and b, and to deter-
mine the mean molar volume, density, and viscosity of
each mixture and of the pure liquids at 25 C. The mean
molar volume and density of each solution can be obtained
with satisfactory accuracy by preparing the solutions in
25- or 50-mL volumetric flasks, taking some pains to com-
pensate for the heat of mixing and for the nonadditivity
of volumes. After obtaining the mass (to nearest milligram)
of the empty flask and the mass of the flask with one com-
ponent added, the second component is added to within
about 0.5 mL of the mark and the solution is mixed thor-
oughly and placed in a 25 water bath. After final dilution
to the mark and mixing, the flask is dried carefully and
the final mass is determined. From the masses of the two
components, the mole fraction composition, Xb = n b/(na +
nb), the density, = m/V (where V is the volume of the
flask), and mean molar volume, Vm= V/(na + nb), can be
calculated. These solutions are then used for the viscos-
ity determinations, using a standard Ostwald viscometer
(2). For students working in pairs, one student can begin
making viscosity determinations for the pure liquids while
the other completes the preparation of solutions. Dupli-
cate or triplicate determinations of the times required to
Vol. 73 No. 10 October 1996 Journal of Chemical Education
empty the viscometer bulb provide an indication of the
precision in the viscosity determination.
Construction of a TemperatureComposition Boiling-Point
Diagram
Our students use the standard sidearm modified dis-
tillation flask with internal nichrome wire heater (3) to
permit convenient sampling of both the vapor and the liq-
uid with which it is in equilibrium at a particular boiling
temperature. Analysis of the liquid and vapor fractions is
most readily carried out by gas chromatography of samples
of the liquid in the flask and the condensed vapor in the
sidearm bulb. We have typically encouraged students to
work out some of the details of the separation conditions,
including the selection of the more appropriate of two chro-
matographic columns employed with a Hewlett-Packard
Model 5750 gas chromatograph. Whether electronic inte-
gration or a strip chart recorder is employed to record the
data, a calibration curve of relative response vs. mole frac-
tion may be required to get reliable concentrations from
relative peak areas (or heights) of the chromatograms.
The solutions prepared earlier for molar volume and vis-
cosity measurements can be used for this calibration. For
liquids with quite different refractive indices, nD, refrac-
tive index measurements can be employed as an alterna-
tive technique for analysis of liquid and vapor fractions
(3). A calibration curve of nD vs. Xb is also required for
this approach, though the assumption of linear variation
of nD with Xb is reasonably close.
Enthalpy of Mixing
We have employed the Parr 1451 solution calorim-
eter and the quantities of materials suggested by Fung
(1) to determine the enthalpy of mixing for 6 composi-
tions of mixtures. A simple Dewar flask equipped with an
accurate thermometer or thermistor could also be used to
get satisfactory data. The enthalpy of mixing can be ei-
ther positive (endothermic) or negative (exothermic). The
sign of H can be determined easily by adding a dropper
full of liquid a to a similar volume of liquid b in a test
tube and noting the temperature change to guide the setup
of the temperature recording device of the Parr calorim-
eter. A reasonable plot of H vs. composition can be gen-
erated by determining the heat of mixing for about 5-6
compositions. To save material, the heat of dilution from
one composition to another can be added to the heat of
mixing for the initial solution, as described by Fung (1).
Calculations and Results
Partial Molar Volume
The partial molar volume of each component in each
solution can be calculated from the mean molar volume,
Vm, by plotting Vm vs. Xb and drawing a line tangent to
the smooth curve at a particular concentration (4). The
partial molar volumes Va and Vb for the two components
971
In the Laboratory

120
1 20 1.2

100 Cyclohexane
1.0
V/mL Relative Ideal
Viscosity
80

0.8
Ethanol Experimental
60

0.6
0.0 0.2 0.4 0.6 0.8 1.0
40
Xb, mole fraction ethanol
0.0 0.2 0.4 0.6 0.8 1.0
Xb, mole fraction ethanol
Figure 2. Composition dependence of relative viscosity of
Figure 1. Composition dependence of mean molar volume, Vm, and cyclohexane(a)ethanol(b) mixtures at 25 oC.
partial molar volumes of cyclohexane (a) and ethanol (b) for cyclo-
hexaneethanol mixtures at 25 oC.

at that concentration are the intercepts of the tangent Viscosity Data

line at Xb = 0 and 1.00, respectively. The data can be
treated numerically by fitting a polynomial (typically sec-
ond degree is sufficient) to the Vm vs. Xb data. Then the
partial molar volumes, Va and Vb at Xb, where the mean
molar volume is Vm, are given by:
Va = Vm Xb (Vm /Xb) (1)
and
Vb = Vm + (1 Xb) (Vm /Xb) (2)
where (Vm /Xb) is the slope of the Vm vs. Xb curve at that
point. With a second-degree polynomial fit of the data, a
general expression for the slope can be obtained and gen-
eral expressions for Va and Vb can be formulated from eqs
1 and 2 for evaluation at each of the compositions exam-
ined.
Typical student spreadsheet data for the cyclohex-
ane (a)ethanol (b) pair are summarized in Table 1; the
mean and partial molar volumes for the pair are plotted
in Figure 1. Deviations from ideality can be displayed most
effectively by plotting Vm , Va, and Vb on the same graph.
By also including a straight line connecting the molar
volumes of the pure liquids, the direction of the deviation
from ideality can be highlighted. Both the distinct posi-
tive curvature of the Vm curve and the increase in Va and
Vb with dilution reveal the positive deviation from ideal
behavior of this system.
The relative viscosity of each mixture, r, can be cal-
culated from the densities () and flow times (t) for each
solution by
r = /o = t/too (3)
where the subscript o identifies one of the pure liquids
designated as the reference material. Relative viscosity
composition data for the cyclohexaneethanol pair, which
are included in the Table 1 spreadsheet, are plotted in
Figure 2. A straight line connecting the viscosities of the
two pure liquids has been included in the graph to em-
phasize the direction and extent of deviation of the system
from ideality. The deviation is sufficient, in this case, to lead
to a minimum in the viscosity near Xb 0.30 (b = ethanol).
The VaporLiquid Phase Diagram
Ideal solutions are typically introduced by a discus-
sion of Raoults Law and the associated vapor pressure
composition diagram (4). Extreme deviations from ideal-
ity can lead to a minimum or maximum in the vapor pres-
surecomposition diagram and the resultant maximum
or minimum in the corresponding boiling pointcomposi-
tion diagram. Binary mixtures assigned to students in
the laboratory are typically chosen to show enough devia-
tion from ideality to produce a minimum- or maximum-
boiling azeotrope (5). The temperaturecomposition phase
diagram at atmospheric pressure, constructed from stu-

Table 1. Excel Spreadsheet: Composition Dependence of Molar Volumes and Relative Viscosity for Cyclohexane(a)-Ethanol(b) Mixtures at 25 C

mass fsk/g fsk+a/g fsk+a+b/g mass a/g mass b/g moles a moles b mole frac b Vm/mL t/sec g/mL Vm/mL Va/mL Vb/mL eta
20.4923 20.4923 39.7639 0 19.2716 0 0.418311 1 59.7641 117.6 0.770864 59.7641 110.951 59.151 1
20.4925 22.4036 40.0622 1.9111 17.6586 0.022708 0.383299 0.94407 61.5753 110.5 0.782788 61.5753 110.594 59.161 0.955
21.1794 27.152 40.5552 5.9726 13.4032 0.070967 0.290931 0.803903 69.0802 102 0.775032 69.0802 109.788 59.276 0.872
20.6831 30.6171 40.1072 9.934 9.4901 0.118037 0.205993 0.635722 77.1533 96.8 0.776964 77.1533 108.992 59.586 0.829
21.2831 34.8228 40.6084 13.5397 5.7856 0.16088 0.125583 0.438391 87.2712 92.4 0.773012 87.2712 108.295 60.187 0.788
21.4391 38.8279 40.7133 17.3888 1.8854 0.206616 0.040925 0.165325 100.994 91.6 0.770968 100.994 107.753 61.441 0.779
20.4421 40.091 40.091 19.6489 0 0.233471 0 0 107.08 95.5 0.785956 107.08 107.663 62.438 0.828

972 Journal of Chemical Education Vol. 73 No. 10 October 1996


Figure 3. Vaporliquid phase diagram of the cyclohexane( a)
ethanol(b) system at atmospheric pressure (about 735 torr).
dent data for the cyclohexaneethanol system used in the
earlier illustrations, is shown in Figure 3. The minimum
in the boiling point at 74 C implies a maximum in the
vapor pressurecomposition diagram, which indicates a
strong positive deviation from ideality; that is, each com-
ponent in the mixture has a higher vapor pressure than
predicted by Raoults Law, pi = Xi pi o.
The Enthalpy of Mixing
The H of mixing in J mol1 is calculated from the
measured temperature rise, T, the molar heat capacities
of the two liquids, Ca and Cb, and the heat capacity of the
empty calorimeter, Ccal, by eq 4.
H = [(naCa + nbCb) + Ccal] T/(na + nb) (4)
The enthalpycomposition diagram for the cyclohex-
aneethanol pair, constructed from student data, is shown
in Figure 4. In addition to the experimental H data, Fig-
ure 4 includes the TS vs. composition curve calculated
for an ideal or regular solution by eq 5,
TS(ideal) = RT[Xa lnXa + Xb lnXb] (5)
the G vs. composition curve calculated from the experi-
mental H data and the ideal entropy curve [G(exp) =
H TS(ideal)], and the G vs. composition curve ex-
pected for an ideal solution for which H = 0 and
G(ideal) = TS(ideal). Since H is positive for this sys-
tem, G is calculated to be less negative than it would
have been for an ideal solution.
Analysis of the Results
The discussion portion of the student report should
be devoted largely to an interpretation of the interrela-
tion among the four phenomena that were investigated
in the project. For the cyclohexaneethanol examples
shown in Figures 14, the minimum boiling azeotrope in
the boiling point diagram (Figure 3) indicates relatively
weaker ab interactions than the average of aa and bb
interactions, where a and b are cyclohexane and ethanol,
respectively. This interpretation also accounts for the mix-
ing process being endothermic (Figure 4), as stronger aa
or bb interactions are disrupted by dilution with the other
liquid. The weaker ab interactions also permit molecules
to slide past each other more readily, accounting for the
Vol. 73 No. 10 October 1996 Journal of Chemical Education
In the Laboratory
2000
1500 T S(ideal)
1000
H(exp)
500
J/mol
0
500
1000 G(exp)
1500 G(ideal)
2000
0.0 0.2 0.4 0.6 0.8 1.0
Xb, mole fraction ethanol
Figure 4. Composition dependence of thermodynamic quantities for
cyclohexane(a)ethanol(b) mixtures at 25 oC.
large decrease in the relative viscosity of mixtures rela-
tive to values predicted by the linear dependence expected
for an ideal solution (Figure 2). Finally, the weaker inter-
actions between a and b permit a and b to get further
apart on the average, so that the partial molar volumes of
both a and b in mixtures exceeds the molar volumes of
pure a and b, respectively (Figure 1).
For the cyclohexaneethanol pair, the assumption that
the TS contribution to the G of mixing is given by eq 5
is not particularly good. Fung (1) includes experimental
data for G, as reported by Stokes and Adamson (6) from
vapor pressure data, and for TS = H G. For this sys-
tem, both a positive H of mixing and a less positive than
ideal S of mixing contribute to making G much less
negative than expected for an ideal solution and also sig-
nificantly less negative than calculated from the experi-
mental H and TS(ideal). In any case, the low-boiling
azeotrope requires a maximum in the vapor pressure vs.
composition curve with a large positive deviation from
ideality; that is, for both components i, pi > Xi pio. There-
fore, G = RT[Xa ln(pa/pao) + Xb ln(p b/p bo)] > G(ideal),
and H > 0 accounts for at least part of this non-ideality.
Alger has recently called attention to the key role of the
entropy contribution to the closely related solubility of
water in aliphatic alcohols and hydrocarbons (7).
The above explanation also nicely accounts for the
data for other alcoholhydrocarbon or alcoholester mix-
tures, such as ethanolbenzene or propanolethyl acetate,
but other mixtures dont behave quite so consistently. The
chloroformacetone mixture is the classic example of a
high-boiling azeotrope, indicative of substantial negative
deviations from ideality. However, the viscositycomposi-
tion curve for acetonechloroform mixtures shows very
little deviation from linearity, and the molar volume vs.
composition curve resembles the curve for cyclohexane
ethanol mixtures. Surprisingly, for ethanolacetonitrile
mixtures, which show a positive deviation from ideality,
and a low-boiling azeotrope such as typical alcoholhy-
drocarbon mixtures, partial molar volumes of the two com-
ponents in a mixture are, in fact, less than molar volumes
of the pure components. In general, the molar volume is
the least sensitive indicator of non-ideal behavior. Con-
versely, it is the best indicator of careful technique.
973
In the Laboratory
Assignment of Liquid Pairs
As the foregoing examples suggest, the whole project
is more interesting if each pair of students is assigned a
different liquid pair and if the class has an opportunity to
compare results. Extensive data on properties of liquid
pairs are available in the literature (5). For example, with
six pairs of students, two pairs each might be given (i)
relatively ideal mixtures, (ii) high-boiling azeotropes, and
(iii) low-boiling azeotropes. We usually have two sections
of the lab, so we assign the same set of mixtures to each
lab. The students are asked to compare results with their
counterparts in the other lab section and are expected to
resolve discrepancies before submitting their final writ-
ten report. A lunch-hour joint poster session facilitates
such an exchange.
Safety and Disposal of Hazardous Materials
Although the alcoholhydrocarbon and alcoholester
pairs can be handled without any unusual precautions,
several of the solvents suggested for investigation are suf-
ficiently hazardous to require special handling. For these
liquids, solution preparation, viscosity measurements, and
distillation and calorimetry experiments should all be car-
ried out in well ventilated hoods. By locating the gas chro-
matograph near the hood where the distillation is being
carried out, students can simply use a 10-mL syringe to
974 Journal of Chemical Education Vol. 73 No. 10 October 1996
collect a sample of the liquid or condensed vapor from the
distillation apparatus and quickly inject it into the inlet
port of the gas chromatograph a few feet away. At the end
of each experiment, the mixtures are added to the appro-
priate organic waste containerwith a separate clearly
labelled container for any mixture which contains a halo-
genated compound.
Acknowledgments
The authors wish to thank Richard Biagioni and Holly
Harris, who introduced or modified some of the procedures
employed in the projects, and the several dozen physical
chemistry students who have provided useful criticism
and feedback in the evolution of this set of experiments
over several years.
Literature Cited
1. Raizen, D.A.; Fung, B. M.; Christian, S. D. J. Chem. Ed. 1988, 65, 932933.
2. Sime, R. J. Physical Chemistry: Methods, Techniques, and Experiments; Saunders:
Philadelphia, 1990; Experiment 14, pp 522527.
3. Sime, R. J. op. cit., Experiment 6, pp 449460.
4. Atkins, P. W. Physical Chemistry, 5th ed.; Freeman: New York, 1994; Chapter 7,
pp 207237.
5. Timmermans, J. Physico-Chemical Constants of Binary Systems; Interscience:
New York, 1959. This is a rich source of data to guide selection of appropriate
mixtures.
6. Stokes, R. H.; Adamson, M. J. J. Chem. Soc., Faraday Trans. 1 1977, 73, 1232
1238.
7. Alger, D, J. Chem. Educ. 1994, 71, 281.