Journal of Solution Chemistry, Vol. 26, No.

10, 1997

The Freezing Process in Methanol-, Ethanol-, and

Propanol-Water Systems as Revealed by
Differential Scanning Calorimetry
K. Takaizumi, 1'* and T. Wakabayashi 2

Received April 21, 1997; revised August 11, 1997

Solid-liquid phase diagrams, including metastable phases, have been obtained

by differential scanning calorimetry (DSC) for methanol-, ethanol-, and propa-
nol-water systems. The metastable solid phases which are initially formed on
cooling were detected for these three systems, in order to analyze the freezing
processes and to correlate the formation of the metastable phases with solution
structure.

KEY WORDS: Lower alcohol-water systems; phase diagram; inflection point;

metastable solid phase; solution structure; supercooled liquid.

1. INTRODUCTION
Studying metastable solid phases which are formed on cooling liquid
solutions is important for understanding of solute-solvent interactions, espe-
cially in water-soluble organic compound-water systems. Although the lower
alcohol-water systems are of particular interest because they are very vis-
cous~1) at low temperatures and provide an opportunity to study metastable
solid phases, only a few reports are known to us. ~I-s)
In a previous paper, ~9)we reported on a differential scanning calorimetry
(DSC) study of the freezing process and the solid-liquid phase diagram of the
water--ethanol system, focusing attention especially on the relation between
solution structure and solid phases that separate from the supercooled liquid.

~Department of Applied Biological Chemistry, Faculty of Agriculture, Tohoku University,

Kawauchi, Aoba-ku, Sendai 980-77, Japan.
2present address: 7-6-52, Dainohara, Aoba-ku, Sendai 981, Japan.
927
0095-9782/97II000-0927512,50/0 9 1997 Plenum Publishing Corporation
928 Takaizumi and Wakabayashi

We discussed the physical meaning of the inflection point on the liquidus

line of the phase diagram appearing at an ethanol mole fraction X of 0.17.
This is a remarkable feature common to solid-liquid phase diagrams for
lower alcohol-water systems, and we shall denote the inflection point by X*
in all these systems. We assumed that the liquid structure would change
abruptly near X*; that is, at concentrations lower than X*, the liquid structure
will be such that the clustering of water molecules is favorable for nucleation
of ice from the supercooled liquid, while at concentrations above X*, it
is unfavorable.
The present paper reports the liquid-solid phase diagrams, including
metastable solid phases, for the water-methanol and water-propanol systems
obtained by DSC. The water-ethanol system was also reconsidered and
studied more carefully than before. Detection of metastable solid phases that
initially formed on cooling and determination of their phase compositions
were attempted, in order to analyze the freezing process and to correlate the
formation of the metastable phases with the solution structure, as was done
previously. It was shown that the first solid separated from the supercooled
liquid mixture on cooling is ice if X < X*, but a metastable hydrate if X >
X*, in all three systems. The inflection point on each liquidus line therefore
was understood as a transition point from a liquid structure for aqueous
solution to that for alcohol solution, and X* corresponds to the composition
where maxima are found in heat capacity Cp and apparent molar volume
V, (~5)at room temperature. Other than the first separated metastable hydrates,
there are many metastable solid phases above X* which seem to increase
with the increase in alkyl chain length of alcohol.

2. EXPERIMENTAL
2.1. Materials and Methods
The methanol (MeOH) and propanol (PrOH) used were both of Infinity
Pure Grade, 99.8% from Wako Junyaku Co. Ltd., and were used without
further purification. Details of the experimental procedures have been
described previously. ~

2.2. Temperature Scanning Program

Five programs for temperature scanning were applied to obtain both
metastable and stable phases. The temperature scanning rates were the same;
1 K-min -1 for cooling and 0.7 K-min -1 for heating, except for Prog. 4 where
scanning was at 0.7 K-rain -I for cooling and 0.5 K-min -l for heating. The
programs were as follows: Prog. 1 (normal): cooling from 30 ~ to -150~
holding there for 10 rain, then heating to 30~ Prog. 2 (annealing): cooling
Freezing of Aqueous Alkanol Systems 929

to -160~ annealing at 130~ for 2 h followed by heating from -150~

and another one in which annealing is carded out at a temperature about 10~
lower than the peritectic point. Prog. 3 (partial): a partial melting program
by which the separated alcohol hydrate on cooling is almost completely
thawed, leaving ice, then again cooling to -150~ and then heating. Prog.
4 (halO: this is for examining the first solid phase separated from a supercooled
liquid on cooling; cooling to a temperature where first freezing is completed,
and then heating to 30~ Prog. 5: combination of Prog. 3 and Prog. 2; first
Prog. 3 is applied, and then Prog. 2.

2.3. Determination of Melting Points

Two cases are distinguished: the phase transition points for the eutectic
and peritectic are the onset of melting and of decomposition, respectively,
as shown in Fig. 2. For melting of a solid phase in the presence of the liquid
phase except for the peritectic, the temperature where melting just finished,
i.e., very close to the temperature of the endothermic peak, was assumed to
be the melting point. In principle, some corrections should be made, mainly
for the time lag due to the finite rate of thermal conduction; however, at the
low scanning rates applied here, such minor corrections could be neglected.
The accuracy of temperature determination is believed to be ___0.3 K,
for alcohol concentrations X < 0.65 for all three alcohols, while at higher
concentrations, 0.65 < X < X~ut~ct~c,the uncertainty is greater owing to the
very high viscosities of these systems at low temperatures.

2.4. Determination of Solid Phase Composition

The composition of the peritectic compound was determined as follows.
The enthalpy of decomposition to ice Ih and a liquid solution of the compound
at a near-equilibrium state which was attained by annealing for a few hours
at about 10~ below the decomposition point, was measured from the thermo-
gram. The enthalpy per unit mass of alcohol was plotted against the mole
fraction of alcohol. We assume that the mole fraction where the enthalpy
plot shows a break corresponds to the composition of the peritectic as seen
from Fig. la. The composition of the metastable peritectic compounds, which
were obtainable by applying the normal program, was also estimated from
the peak in the plot of the enthalpy of decomposition per unit mass of solution
vs. mole fraction of alcohol. With this program, equilibrium was not attained
(Fig. lb). This method gave much clearer results in determining the composi-
tions than possible alternatives, such as detection of the mole fraction where
the eutectic just vanishes. On the other hand, the composition of other hydrates
which melt on heating was assumed to be given by the highest melting point
in the plot of melting point vs. mole fraction, as shown in the figures.
930 Takaizumi and Wakabayashi

300

EtOH

A d A A ,,Ilk

200 u-It- w ~~O

'7

1OO

0 i i I

0 10 20 30 40 50
Wt%
Fig. la. Determinationof compositionof the peritecticstable compoundin the
ethanol-watersystemobtainedwiththe annealingprogram.

3. RESULTS AND DISCUSSION

Typical examples of DSC melting thermograms for methanol-water and
propanol-water systems are shown in Fig. 2. Solid-liquid phase diagrams
are also shown in Figs. 3a-3c for the three lower alcohol-water systems.
Dotted lines in the thermograms show the graphical procedure used to deter-
mine the phase transition points. A previously unknown metastable solid
phase of methanol was found in the thermograms giving curve b in Fig. 2.
The symbols labelling the phase boundaries have the following meanings:
A, melting of ice; B, melting of metastable solid other than ice, that is first
formed on cooling; D, melting of alcohol hydrate (metastable); El, stable
eutectic point; ~ , metastable eutectic point; P1, decomposition of stable
peritectic; P2, decomposition of metastable peritectic. Of the three systems,
the water-propanol system was the most difficult to work with. We could
not obtain the melting point of pure propanol or the eutectic point, presumably
because of the unusually high viscosity of the solution. This incompleteness
is, however, not crucial for the present purpose of detecting metastable solid
phases. The invariant temperatures, and their corresponding compositions are
listed in Tables I and II for all three systems.
Freezing of Aqueous Aikanol Systems 931

100

EtOH

.=

50

.o" ', A'S.~TH2O

A-I7H20
i

0 I I1 I
0 10 20 30 40 50

Wt%
Fig. lb. Determination of composition of the metastable peritectic ( 9 obtained by the normal
DSC program, and stable peritectic at near-equilibrium( 0 ) obtained by the annealing program.

3.1. Composition of Peritectics

The peritectic hydrates of EtOH and PrOH are listed in Table I. There
are some uncertainties in their compositions which were estimated from plots
of the enthalpy change per unit mass of solution v s . concentration of alcohol,
as shown in Fig. lb for ethanol. There we can distinguish one large peak
and two other peaks (or shoulders), corresponding to EtOH'5.67H20,
EtOH. 7.67H20, and EtOH. 17H20, for the highest, second highest, and lowest
peak, respectively. Therefore, the results of the present study suggest that
both types of clathrates, I and II, are formed with all cages occupied by ethanol
guest molecules. Jeffrey and McMullan (~~ did X-ray studies of clathrate
and clathrate-like hydrates and found that the composition of clathrate I is
M'5.75H20, M being the guest molecule, when all cages of two types (12
hedra and 14 hedra) are occupied by M, and M'7.67H20 if only the larger
cages are filled. With clathrate II (12 hedra and 16 hedra), the composition
is M. 17H20 if only the larger cages are occupied. Thus, the EtOH peritectic
may exist as a mixture of clathrate I and clathrate II. Potts and Davidson (it)
confirmed the existence of the clathrate II hydrate EtOH'I7H20 below
-73.5~ by thermal analysis and dielectric measurements. EtOH'5H20 with
a decomposition temperature of -74~ was reported by Vuillard and
932 Takaizumi and Wakabayashi

a)MeOH(46.7wt%)
A

b)MeOH(95.3wt%)

t,t) e-/ A,..j

c)PrOH(36.7wt%)

I I !

-135 -98.8 -62.5 -26.3 10

T/~
Fig. 2. TypicalDSC thermograms:(a) methanol-water(X = 0.33, annealing,at -113~ 2
hr); (b) methanol-water(X = 0.92, normalprogram);(r propanol-water(X = 0.15, annealing,
at -70~ 3 hr).

Satragno(12) from thermal analysis, while Ott et al. (l) deduced EtOH'2H20
by thermal analysis. Calvert and Sdvastava03~reported from X-ray diffraction
the formation of two types of ethanol hydrates, clathrate I and "modified
clathrate II", the former possibly being stable when a large fraction of its
smaller cages are filled with air. Recently Boutron and KaufmannrE) reached
the conclusion from their X-ray diffraction and thermal data that there exist
clathrate II hydrate EtOH-17H20 and, at higher concentrations, another
hydrate of composition 46H20 per 6 to 8 EtOH molecules. This is not
inconsistent with the present finding of a metastable hydrate. Thus, there
remain questions to be addressed in the future.
For PrOH, results similar to those for EtOH were obtained, as can be
seen from Table I. From the plot of the enthalpy change, the existence of
clathrate I with cages all occupied by propanol molecules and clathrate II of
composition PrOH'17H20, with only the larger cages occupied, are con-
firmed, despite an expectation that the formation of PrOH-5.7H20 would be
unlikely owing to the difficulty of fitting the propanol molecule into the
smaller 12 hedra cage. Apparently there are no reports to compare with the
present results.
Freezing of Aqueous Alkanol Systems 933

a) MeOH

A
-50

,u
[.,

-100
' '-1
o,
t"4 ,,

-15o..,
0
t /i' J

I ll~

Xl'I-~
O
I'--
9

,
0.5
,
i

t
)
[

,I ,
]
1.0
X( MeOH )
Fig. 3. Solid-liquid phase diagram for lower alcohol-water systems: O, obtained by normal
program; O, solid first formed from supercooled liquid on cooling; ~, peritectie compound
obtainedby meltingof solid D leaving ice (by partial meltingprogram);-, equilibriumline; ---,
metastable line. A: solid phase first formed from supercooledliquid, ice, B, solid first formed
other than ice; D, intermediate compound; P, peritectic; E, eutectic, (a) methanol-water, (b)
ethanol-water, (c) propanol-water.

The composition of peritectic methanol hydrate obtained here is

MeOH-2H20, which is inconsistent with the monohydrate reported by Ott
et al. (~) from their phase diagram.

3.2. Freezing Process

The present study systematically revealed the existence of metastable
solid phases, which have so far been known only partially, in the middle
range of concentration as shown in Figs. 3a-3c. (1'2'9)
934 Takaizumi and Wakabayashi

0d

b) EtOH

~_ A

-50 tl q~ w

U
[-, ! I
OI ; I
= ;:
,,Q 10

-1OO
. i ~ i ,,-'I

':i I
;i
|1
,~ E-'
zl : /-4
Ii I l
ii
", r 1
"'*~ II ,
-150
O O.5 l .O

X(EtOH)
Fill. 3 Continued.

The metastable solid phases begin to appear around X*, which is the
composition for the inflection point on the liquidus line; common to all
systems examined: the first separated solid phase from supercooled liquid
on cooling is ice Ih in the concentration range below X*, while at higher
concentrations, metastable hydrate B as marked in Figs. 3a-3c appears, though
its composition depends on the alcohol and alcohol concentration as shown
in Table II. Except for the methanol system, the hydrate B is transformed to
another metastable (but kinetically very stable) solid phase D during the
temperature decreasing scanning. Thus, to obtain a thermodynamically stable
peritectic hydrate, a special temperature scanning program was essential, that
is, a partial melting program such that the hydrate D formed would melt
leaving ice. The hydration number of hydrate D seems to be 1.2 for both
Freezing of Aqueous Alkanol Systems 935

'J•l•
I I I w I I I ! I

II .o-eoo"~O-o.. "x'.,
-i!S o .. -.,:
-50
...... -~2..
tl
,I
II \
~I ~
- t-~ ,i
~.q ..
1-~11

-100" "
ml
II
tl
Ii

If
tl
I

It
II

- 15o ,~{ , , i L I I I I I
0 0.5 1.0
XlPiOH)
Fig. 3 Continued.

EtOH and PrOH. From the above we can conclude that the peritectic hydrate
is formed only by the reaction of ice Ih with the residual liquid solution. In
the methanol-water system, the second hydrate D apparently does not exist,
and the peritectic hydrate is easily formed without annealing or without
applying a partial melting program. This suggests that the formation of the
hydrate D is characteristic of the alcohols, such as ethanol and propanol,
which form clathrates or clathrate-like hydrates. The lower alcohol-water
systems studied here are unusual in the sense that thermodynamically unstable
solid phases have higher melting points than the transition points (peritectic
points) of the stable solids. The reason for this may be that these metastable
solids are prevented from rearrangement to form the thermodynamically
stable peritectics by a potential barrier arising from the extraordinarily high
viscosity of their solutions at low temperatures.
936 Takaizumi and Wakabayashi

Table L Composition of Solid Phases and Invariant Points in Lower Alcohol-Water

Systems

MeOH EtOH PrOH

ma Stable Ice Ice Ice

B Metastable Solid phase first formed from supercooled liquid on cooling
(Table II)
D Metastable c a A . 1.2H20 b c a . A 9 1.2H20
- 65.0~ - 35.20(:
Pi Stable A 9 2H20 A 9 5.67H20 A 9 5.75H20
Peritectic - 102.5~ - 73.9~ - 53.6~
P2 Metastable A . 17H20 ~ A - 17H20
Peritectic A 9 7.67H20 A 9 7.67H20
A 9 5.75H20
El Stable X = 0.80 X = 0.86 Not detected
Eutectic - 117.3~ - 124.3~
E: Metastable X = 0.82 X = 0.84 Not detected
Eutectic - 118.5~ - 125.0~

"Symbols in column 1 correspond to those in the figures.

bSymbol A in the hydrate formulas denotes the alcohol.
r hydrates are assumed to contain air.

Table II. Composition and Melting Point of the First Solid Formed on Cooling~

MeOH EtOH PrOH

A 9 1.7H20 A 9 4H20 A 9 7.7H20

-91.3~ -47.5~ -24.0~
B metastable A 9 2H20 A 9 5.7H20
-52.5~ - 26.0~
A 9 3H20
-23.0~
A 9 2H20
-26.3~
A " 1.5H20
-27.5~

"The symbol A in the hydrate formula denotes the alcohol.

3.3. R e l a t i o n B e t w e e n F r e e z i n g P r o c e s s a n d S o l u t i o n S t r u c t u r e

The results described above are very interesting, taking into account
the following:

(1) T h e c o i n c i d e n c e b e t w e e n t h e c o m p o s i t i o n at t h e i n f l e c t i o n o f t h e
liquidus line and the appearance of the first hydrate.
(2) T h e i n c r e a s i n g o r d e r in X * , 0.1 f o r P r O H < 0 . 1 7 f o r E t O H < 0.3
for MeOH, w h i c h is i n t h e s a m e o r d e r as t h e m o l e f r a c t i o n o f a l c o h o l at
Freezing of Aqueous Alkanol Systems 937

which extrema appear in the partial molar volume and heat capacity of alcohol
in solution. (m
(3) The increasing prominence of the inflection in the order MeOH <
EtOH < PrOH parallel to that in the extrema in volume and heat capacity. (m
These facts strongly suggest the occurrence of some abrupt change in solution
structure near these compositions. Thus, the alcohol-water systems can be
divided into three concentration regions according to the freezing processes,
each of which reflects the solution structure.

3.3.1. Region 1
Water-rich region: alcohol in water as solvent, X < X*. Alkyl groups
are completely surrounded by a water layer where water molecules adjacent
to alkyl groups will somehow be modified in their orientation and energy in
comparison with bulk water (i.e., will undergo hydrophobic hydration). In
this situation ice nuclei can form on cooling.

3.3.2. Region 2
Transient region: a narrow range of concentration above X*. Here, direct
contact between alkyl groups begins to occur and a population of small
clusters of alcohol molecules increases rapidly with increasing alcohol con-
centration. In this concentration range, ice and hydrate B can be formed on
cooling, and the freezing point depression of ice decreases sharply because
contact among alkyl groups results in a decreasing probability for water
molecules to interact with alkyl groups. Thus, the inflection on the liquidus
line appears. Considering the difference in the number of water molecules
required to surround the alkyl groups, X* should depend on the size of alkyl
group, as we find in this study. Our data are inconsistent with the results of
Mashimo and Umehara, (15)who concluded that the solution structures change
at the same concentration X = 0.17 for MeOH, EtOH, and PrOH.

3.3.3. Region 3
Water in alcohol as solvent: X > X*. Further increase in alcohol concen-
tration causes water molecules to interact mainly with hydroxyl groups of
the alcohol through hydrogen bonding, and water molecules do not "see"
enough water molecules to form ice nuclei on cooling. Here the solid first
formed is not ice, but the hydrate B.
Mizuno et aL 07-19) made careful NMR and FT-IR measurements on
alcohol-water mixtures to determine hydrogen-bonding interactions. As can
be seen from their results for the ethanol-water system, the composition
dependence of the chemical shift of the water proton and the frequency of
938 Takaizumi and Wakabayashi

the H--O-H bending vibrational band, each show three turning points near
X equal to 0.15, 0.65, and 0.85 in the liquid state, which are then interpreted
as transition points in solution structure. These points are essentially identical
with the discontinuities labelled X*, P, and E in Fig. 3b, at compositions X*
= 0.17, Xp = 0.63, and XE = 0.86. The solid first deposited from supercooled
liquid on cooling is pure ice (as already discussed) for 0 < X < X*; for
EtOH hydrates of various compositions depending on X* < X < Xp; c a.
EtOH. 1.2H20 for Xp < X < XE; and pure ethanol for X > XE. Similar
correlations are seen in methanol-water and propanol-water systems.
Recently Takei e t aL (2~ studied dielectric properties of water-EtOH
solutions by measuring the resonance frequencies fR. The plot of fR and
conductivity g vs. mole fraction of EtOH shows a clear break point at X
0.15 (close to X* = 0.17), indicating that the structure of a binary solution
changes at X ~- 0.15; that is, below X ~- 0.15 the relative dielectric constant
er and tr of the solution show perturbations from the properties of pure
water, and above X ~- 0.15 they are determined largely by the properties of
pure alcohol.
There are many discussions concerning the clustering structure of mole-
cules in solution based on various kinds of measurements, especially X-
ray and neutron diffraction. ~2~-25)Mashimo and Umehara ~15-~6)inferred from
dielectric relaxation of ethanol-water mixtures that chainlike clusters of alco-
hol and water are formed at alcohol concentrations above X = 0.17 and,
finally, that the chain-like clusters form a network through hydrogen bonding
between chains.
With regard to solution structure, we believe that hydration in solution
is reflected in the compositions of hydrate B, hydrate D, and the peritectic
only indirectly. Thus, we say nothing about the size and shape of ethanol
and water clusters in the liquid state at higher concentrations in the present
work. We think, however, that chain-like clusters of ethanol will be the basic
structures, as in the solid ethanol, and water molecules will be hydrogen-
bonded with hydroxyl groups of alcohol molecules at the end of chain-like
clusters. The size of clusters will depend on the concentration of alcohol.
The present results will obviously have to be confirmed by other methods.

REFERENCES
1. J. B. Ott, J. R. Goates, and B. A. Watte, J. Chem Thermodytt 11, 739 (1979).
2. P. Boutron and A. Kaufmann, J. Chem. Phys. 68, 5032 (1978).
3. P. Bouton, A. Kaufmann, and N. V. Dang, Cryobiology 16, 372 (1979).
4. E L. Oetting, J. Phys. Chem. 67, 2757 (1963).
5. G. A. Miller and D. K. Carpenter, J. Chem. Eng. Data. 9, 371 (1964).
6. G. Vuillard and M. Sanchez, Bull. Soc. Chim. Fr. 1877 (1961).
7. J. C. Rosso and L. Carbonnel, Comptes rendus. Series C, 1012 (1969).
Freezing of Aqueous Alkanol Systems 939

8. J. C. Rosso and L. Carbonnel, Comptes remus. Series C, 4 (1968).

9. K. Takaizumi and T. Wakabayashi, Cryo-Letters 16, 41 (1995).
10. G. A. Jeffrey and R. K. McMullan, Prog. Inorg. Chem. 8, 43 (1967).
11. A. D. Potts and D. W. Davidson, J. Phys. Chem. 69, 996 (1965).
12. G. Vuillard and N. Satragno, Compt. Rend. 250, 3841 (1960).
13. L. D. Calvert and P. Srivastava, Acta Cryst. Sect. A25 S131 (1969).
14. E Franks and J. E. Desnoyers, Alcohol-Water Mixtures Revisited, in Water Science Reviews,
Vol. 1, E Franks, ed., pp 171-227, (Cambridge University Press, Cambridge, 1985).
15. S. Mashimo and T. Umehara, J. Chem. Phys. 95, 6257 (1991).
16. S. Mashimo, T. Umehara, S. Kuwabara, and S. Yagihara, J. Phys. Che~ 93, 4963 (1989).
17. K. Mizuno, K. Oda, and A. Okumura, J. Phys. Chem. 99, 3056 (1995).
18. K. Mizuno, Y. Miyashita, and Y. Shindo, J. Phys. Chem. 99, 3225 (1995).
19. K. Mizuno, Y. Kimura, Y. Shindo, in 18th Symposium on Solution Chemistry, Japan,
Kusatsu, 1995.
20. S. Takei, Y. Sugitani, T. Amano, and Y. Nishimoto, Bunseki Kagaku 45, 903 (1996).
21. A. H. Narten and H. A. Levy, J. Chem. Phys. 55, 2263 (1971).
22. A. H. Narten and A. Habenshuss, J. Chem. Phys. 80, 3387 (1984).
23. K. Nishikawa and T. lijima, J. Phys. Chem. 97, 10824 (1993).
24. N. Nishi, S. Takahashi, M. Matsumoto, A. Tanaka, K. Muraya, T. Takamuku, and T.
Yamaguchi, J. Phys. Chem. 99, 462 (1995).
25. M. Matsumoto, N. Nishi, T. Furusawa, M. Saita, T. Takamuku, M. Yamagami, and T.
Yamaguchi, Bull. Chem. Soc. Japan 68, 1775 (1995).