Endothermicity or Exothermicity of Water-Alcohol Miture
Endothermicity or Exothermicity of Water-Alcohol Miture
Endothermicity or Exothermicity of Water-Alcohol Miture
Abstract
Excess enthalpies of mixtures of water with methanol, ethanol, l- and 2-propanol, l-, 2- and tert-butanol
were determined by dilution calorimetry in the accessible range of mole fractions at 25°C. Ab initio RHF calcn-
lations were performed on the alcohols, their dimers and complexes with water. A full optimisation at the 6-31G
level was used to obtain the properties of interest of these systems. The calculated H-bond energies are similar for
all the studied alcohols. Nevertheless, a systematic stabilisation of the mixed dimers is found, especially for the
secondary and tertiary alcohols. The formation of mixed H-bond chains is exothermic and does not follow the
rules of the geometric mean. To a first approximation, the experimental excess enthalpies AH’ for all the systems
follow the equation
1000
.5
E
5
-1000
0,O 0,2 0.4 x 0,6 0.8 1 ,o
Fig. 1. Heat of mixing of water/methanolmixtures (J mol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
equations, in the pure liquid at 25°C a given mole- the experimental enthalpies of mixing over the
cule is liberated from the weakest bond during 6% whole range of composition, and (ii) on the results
of the time. This result is confirmed by the fact that of a quantum chemical study of the dimers and
it corresponds to the percentage of free OH vibra- 1 : 1 complexes. The literature data on the molar
tors in pure water, as determined by Luck [8] from energies of vaporisation are also used [9].
infrared data. The molecule is completely free of H-
bonding during only 0.01% of the time. Experimental
Our discussion on the endo- or exothermicity of
water/alcohol mixtures is essentially based on (i) Products used in this paper were all “for analysis”
Fig. 2. Heat of mixing of water/ethanol mixtures (Jmol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
542 D. Peeters, P. Huyskem/J. Mol. Struct. 300 (1993) 539-550
I
a
-1000 I I I I I
Fig. 3. Heat of mixingof water/l-propanol mixtures (Jmol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
quality products. Methanol and both propanols the well-known heat exchange produced by the
were provided by BDH, ethanol by Fluka, l-buta- mixture of methylacetate with cyclohexane
no1 by Prolabo, 2-butanol by Baker and tert-butanol (Cpapp) [lo]. Temperature was recorded on an x-r
by Jansen. These products were not purified further recorder and ATs were obtained graphically from
and various mixtures were prepared by weight with the charts.
distilled deionised water. Heat capacities of the pure compounds were
The heats of mixing were measured using a Parr taken from the review on alcohols by Zwolinski
solution calorimeter calibrated by the measure of [9]. Heat capacities of the mixtures were obtained,
E
E
I
a
0
Fig. 4. Heat of mixing of water/2-propanol mixtures (Jmol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
D. Peeters, P. HuyskenslJ. Mol. Struct. 300 (1993) 539-550 543
I
0.0 0,2 0,4 x o,6 o,8 1 ,o
Fig. 5. Heat of mixing of water/l-butanol mixtures (Jmol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
assuming additivity, from the pure compounds by tion of small amounts of solute starting from either
the relation: the pure compound or mixtures of known compo-
sition. In this way, the complete range of mixing
Cpmix = wCpw + nACp~ + Cpapp (4)
was scanned and the various explored regions were
All measurements were made at 25°C except for allowed to overlap to ensure the reliability of the
tert-butanol for which the temperature was raised measurements. The experimental excess molar
above its melting temperature to 28°C. enthalpies of mixing are represented by the dots
Enthalpy data were obtained by successive addi- in Figs. l-7.
Fig 6. Heat of mixing of water/2-butanol mixtures (Jmol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
544 D. Peeters. P. Huyskens/J. Mol. Struct. 300 (1993) 539-550
Fig. 7. Heat of mixing of water/terbbutanol mixtures (J mol-‘) vs. mole fraction of alcohol. The curve is obtained using Eq. (9).
Table 1
Total energies (u) at the RHF 6-31G level
0’1
----
2.873A
A4 WI
Water - - -32.80 - - -
Methanol -31.20 -34.21 -32.66 3.06 -3.08 -0.01
Ethanol -30.14 -34.88 -32.56 3.89 -4.40 -0.51
1-Propanol -30.74 -34.71 -32.46 3.78 -4.21 -0.49
2-Propanol -30.57 -35.88 -32.90 4.57 -6.06 -1.49
1-Butanol -30.68 -34.91 -32.65 4.10 -4.36 -0.26
2-Butanol -30.34 -36.23 -33.25 5.38 -6.40 -1.02
t-Butanol -29.99 -36.22 -32.58 5.40 -7.05 -1.65
546 D. Peeters. P. Huyskens/J. Mol. Strut. 300 (1993) 539-550
It is well known that theoretical reaction energies mixing by means of a function such as:
contain some imperfections due to approximations
AHe = XwXAF(Xw, XA) (7)
inherent in the theoretical approach. These are
essentially the basis set limitation leading to the because this function passes through the origin at
basis set superposition error and the lack of elec- mole fractions equal to zero and one.
tron correlation. Nevertheless, from the energetic The derived experimental function F(Xw, XA) is
point of view, these errors cancel out if some care is represented in Fig. 9. For all the alcohols, these
taken in the definition of a reference reaction. This functions show much more systematic behaviour
is the case when the number and nature of the than the direct AHe curves. Above X_,,= 0.2, F
bonds between the reactants and products is con- displays only a very limited change, although the
served. As process II (Eq. (6)) satisfies these average value in this range depends on the nature
requirements, including H-bonds, the correspond- of the alcohol. Below X, = 0.2 a sharp decrease in
ing energy may be considered with some confi- F is observed towards pure water. This means that
dence. However, from a geometrical point of in this region a formal polynomial development of
view, the structures obtained are known to be accu- F involves a dominating term of a high power for
rate enough to lie within the experimental error. Xw and that the phenomenon implies a large num-
Table 2 shows a remarkable constancy in the ber of water molecules. An initial idea of the mole-
energy of formation of homodimers from water cularity of the phenomenon may be obtained by
to tert-butanol. In heterodimers, the interaction is analysis in this region of the logarithmic beha-
always stronger when water acts as a proton donor. viour of F against Xw. One finds a value of
In this case, a systematic increase of the absolute 5.7 f 1.4 for this slope.
value is observed when going from the primary A molecularity of six can be expected on the
alcohols to secondary or tertiary ones. The reverse basis of the reinforcement by the alcohol of the
effect occurs for the complexes where water acts as weak chain in water as noted in the Introduction.
proton acceptor. In order to define a weak chain in water, six mole-
The most interesting data with respect to the aim cules are required because a strong chain involves
of this work are those of the last column. They at least three molecules.
show a systematic stabilisation when homodimers
dismutate into mixed heterodimers. Although these H i
results do not concern associates of more than two H-O H-O
molecules, they suggest that the formation of
mixed water/alcohol H-bonded chains starting H
from homogeneous chains is an exothermic pro-
cess. This makes the rule of geometric mean no H i
F q methanol
l ethanol
a
I I I I
F El propanol 1
l propanol2
-10000
-20000
0.0 0‘2 0,4 X 0.6 0.6 1
F D butanoll
l butanol2
II ter butanol
c
-20000 ! I I I I
OS0 0,2 0,4 X 0.6 0,6 1
Fig. 9. F(X) function obtained for the water/alcohol mixtures vs. mole fraction of alcohol.
548 D. Peeters, P. iiuyskens/J. Mol. Struct. 300 (1993) 539-550
Table 3
Regression coefficients of AHe following Eq. (9) (kJmcrl_‘)
molecule at the left forms two strong bonds because not taken into account and are more important
all the bonds formed by the alcohol are strong. As for the alkyl residues of the alcohol in water than
94% of the water molecules in water form weak in their own phase.
chains, the proportion will depend on pw. We shall now discuss the various coefficients of
For the primary alcohols, experimental F func- the correlation. The C,,, coefficients have impor-
tions above X, = 0.2 present a remarkable pla- tant negative values which exhibit differences
teau. For this series, one can thus envisage to a which do not seem to present any systematic corre-
first approximation a two term function: lation with the molecular weight or the structure
of the alcohols. Their mean value is
AHe = Ce,&w w, + +X,X, (8) -8.3 f 2 kJmol-‘. According to the ideas dis-
For the branched alcohols, a small positive slope cussed above, this may be interpreted as resulting
is observed which can be accounted for by an addi- from the reinforcement of the weak H-bond in
tional XwX$, term. Thus: liquid water where one alcohol molecule is
inserted in the weak chain. The H-bond formed
AH= = C&$w WA + +X,X, + c,,zXwX;
by the inserted alcohol molecule will of course
(9) preserve the strength it had in its own phase. This
The coefficients of this equation were fitted by a effect is restricted to the first insertion of an alcohol
multifactor regression analysis procedure. Curves molecule. The enforcement is estimated to have a
calculated with these coefficients are reproduced in mean value of 8.3 kJmol-‘. In this sense, the dis-
Figs. l-7. The corresponding coefficients are listed solution of alcohol in water is always markedly
in Table 3. exothermic. As a consequence, we may say that
It is evident that Eq. (9) correctly describes the the energy difference between strong and weak H-
general behaviour of alcohol/water mixtures. For bonds is close to 8.3 kJmol_’ and, if one neglects
methanol and ethanol, where the non-specific the entropic effects due to the enhancement of
effects are clearly less important, the agreement is vibrational motions, one expects between the inser-
astonishing (especially for ethanol where the inflec- tion constants in both chains a factor of:
tion of the curve is perfectly reproduced).
For higher alcohols, the overall behaviour is
satisfactory and the passage from exothermicity
to endothermicity is correctly predicted. How- This is the order of magnitude of previous estima-
ever, the intervention of non-specific forces is tions of such a ratio (see introduction in ref. 7 for a
more roughly approximated. In particular, the value of 33).
higher endothermicity at lower XA values may be
related to dipole-dipole induced forces which are The coefficient Cl/r is negative for some alcohols
D. Peeters, P. HuyskenslJ. Mol. Struct. 300 (1993) 539-550 549
and positive for others. This demonstrates that this volumes which should not be negligible in
term has a hybrid origin. Dispersion and dipole- the considered case ( qHzO) = 18 cm3 mol-’ and
dipole interactions lead to a positive heat of mix- V(autanol)= 92cm3 mol-‘). The reason is that in
ing (Scatchard-Hildebrand), whereas the previous structured mixtures such as alcohols and water,
theoretical study shows that the contrary effect is no random mixing rules the contacts between mole-
expected for H-bonding. One can therefore divide cules embedded in long molecular chains.
the C,,, term into two contributions: Finally, the Cl/2 coefficients are negligible for the
normal alcohols. For secondary and tertiary alco-
C,il = C{,, (H-bond) + Cy,, (non-specific) (10)
hols, they exhibit positive values and, in contrast to
The latter term is positive, the former negative. C,,,, concern only non-specific effects.
If we consider the sequence of primary alcohols, Differences in the non-specific cohesive forces
this coefficient exhibits a systematic increase. For are clearly shown in the energies of vaporisation
this series C,/, may be related to the number of (AUV) given in Table 4. They are compared with
carbon atoms nc by the half-vaporisation energies of the dimers
obtained from the two alkyl residues. The differ-
C,,, = -4.6 + 1.6nc (kJmol_‘)
ences are related to the cohesion of the OH group,
The value of -4.6 kJ mol-’ describes the stabilisa- (AU&). For the primary alcohols, the latter value
tion of the mixed R-OH.. - H-OH bonds which is is close to the energy obtained from the theoretical
supposed to be the same for the whole series. For calculations. However, it must be kept in mind that
the secondary and tertiary alcohols, C1,, is system- in the liquid, the OH group in the chains is also
atically more negative, the differences being of the bound to the surroundings by non-specific forces.
order of -1 and -2 kJ mol-’ respectively. This This means that the concurrence of the forces
demonstrates, as predicted by the theoretical lowers the contribution of the H-bond in the
results, that the reinforcement of the H-bonds for energy of vaporisation. Such effects can be a priori
mixing is more important in these cases. It should expected.
be noted that for the primary alcohols the non- The differences in AU’ between the isomers is
specific part of the heat of mixing is given to a essentially a consequence of the differences in the
first approximation (neglecting the C1iz contribu- non-specific forces, with a lowering of about
tion) by C~,,XwX,. This equation differs funda- 2 kJmol_’ on going from the primary to the sec-
mentally from the Scatchard-Hildebrand relation ondary isomer and 4 kJ mol-’ on going from the
because it contains only the mole fractions and primary to the tertiary isomer. A similar trend is
ignores effects due to the difference in molar observed in the C,lz coefficient.
Table 4
Vaporisation energies of alcohols and related dimer alkanes (kJ mol-‘) [9]
Comparing the isomers of butanol, one is con- to a first approximation by the geometric mean rule
fronted by the fact that Ct/t + C1lz is larger for 2- but deviate from the Scatchard-Hildebrand equa-
butanol than for tert-butanol. This can only be tion because of the non-randomness of the contacts.
explained by a competition between the non-speci-
fic forces which, on the one hand, increase the Acknowledgement
endothermicity of the mixture and, on the other
hand, increase the reinforcement of the mixed H- One of the authors (D.P.) would like to thank the
bonds, which induces an exothermic effect and Belgian National Fund for Scientific Research for
favours the tertiary butyl isomer. This competition its financial support of this research.
clearly leads to the sequence observed at X, + 1
for the various butanol isomers (Figs. 5-7 and 9(c)) References
putting tert-butanol between the primary and sec-
ondary isomers. M.I. Davis and ES. Ham, Thermochim. Acta, 190
(1991) 251.
P.L. Huyskens, J. Mol. Struct., 274 (1992) 223.
Conclusions (a) R. Furth (Ed.), in Investigation on the Theory of
Brownian Motion, Dover, New York, 1956.
The most important results of this study are on (b) G.S. Hurst and M.G. Payne, in Principles and
Applications of Resonance Ionisation Spectroscopy,
the one hand, the confirmation of the existence of
Bristol, 1988, p. 196.
two kinds of H-bonded chains in water, the weakest C. Panayiotou and I.C. Sanchez, J. Phys. Chem., 95
of which can become reinforced by the insertion of (1991) 10090.
one alcohol molecule, and on the other hand, the H. Tanaka and K.E. Gubbins, J. Chem. Phys., 97 (1992)
2626.
non-applicability of the rule of the geometric mean
(a) M. Ferrario, M. Haughney, I.R. McDonald and
to the formation of mixed H-bond chains. The M.L. Klein, J. Chem. Phys., 93 (1990) 5156
study demonstrates the usefulness of the compar- (b) G. Palinkas, I. Bako, K. Heinzinger and P. Bopp,
ison of results obtained from ab initio calculations Mol. Phys., 73 (1991) 897.
which lead to the above-mentioned conclusions, 7 P.L. Huyskens, in H. Kleeberg (Ed.), Interactions of
Water in Ionic and Non-ionic Hydrates, Springer-Ver-
with experimental quantities such as heats of lag, Berlin/Heidelberg, 1987, p. 113.
vaporisation and mixing. The theoretical method 8 W.A. Luck, Angew. Chem., 29 (1980) 92.
shows an impressive constancy in the energy of 9 (a) R.C. Wilhoit and B.J. Zwolinski, J. Phys. Chem.
dimerisation for the whole series of alcohols and Ref. Data, (1973) 2.
(b) B.J. Zwolinski and R.C. Wilhoit, Vapour Pressures
of water. This can be compared with the constancy and Heats of Vaporization of Hydrocarbons and
of the share of H-bonds in the experimental ener- Related Compounds, Thermodynamic Research
gies of vaporisation of the primary alcohols. How- Center, College Station, TX, 1971.
ever, both figures differ by a constant value. This is 10 P.L. Huyskens and L. Vanderheyden, Fluid Phase
Equilibria, 49 (1989) 271.
the consequence of the fact that other cohesive
11 M.J. Frisch, J.S. Binkley, H.B. Schlegel, K. Raghava-
forces have a concurrent effect which reduces the chari, C.F. Melius, R.L. Martin, J.J.P. Stewart, F.W.
share of H-bonding in the liquid. Bobrowicz, CM. Rohling, L.R. Khan, D.J. DeFrees,
The heat of mixing of an alcohol with water is R. Seeger, R.A. Whiteside, D.J. Fox, E.M. Fleuder and
J.A. Pople, GAUSSIAN 86, Carnegie Mellon Chemistry
ruled by the peculiarities of H-bonding as they are
Publishing Unit, Pittsburgh, PA, 1984.
predicted by the ab initio calculations, and by the 12 W.J. Hehre, R. Ditchfield and J.A. Pople, J. Chem.
intervention of non-specific forces which are ruled Phys., 56 (1972) 2257.