Formation of super disperse phase and its influence on equilibrium and thermodynamics of thermal dehydration

Article abstract of DOI:10.1016/j.tca.2007.11.005

New data on the dehydration and rehydration processes of calcium, manganese and copper dichlorides are presented that reveal surprising, in a certain sense, behaviour difficult to be explained for the last two chlorides in terms of the usual conception of thermodynamic equilibrium. A substantial role of a super disperse phase at studying the equilibrium of the thermal decomposition of a hydrate is postulated to explain the experimental results for manganese and copper dichlorides. It is shown that the formation of such a phase of the hydrate is able to change appreciably the experimental results, causing the increase of water vapour pressure and the decrease of the derived enthalpy of a reaction. The results obtained allow to understand the reasons for considerable differences of some literature data. They enable to receive more precise and reliable data for thermal dehydration and probably for some other decomposition processes.

Full text of DOI:10.1016/j.tca.2007.11.005

Available online at www.sciencedirect.com
Thermochimica Acta 467 (2008) 44–53
Formation of super disperse phase and its influence on equilibrium and
thermodynamics of thermal dehydration
a,∗
a
a
b
O.G. Polyachenok , E.N. Dudkina , N.V. Branovitskaya , L.D. Polyachenok
a
Department of Chemistry, Mogilev State University of Foodstuffs, 212027, Belarus
Department of Chemistry, Mogilev State University of A.A. Kuleshov, 212022, Belarus
b
Received 22 July 2007; received in revised form 31 October 2007; accepted 11 November 2007
Available online 19 November 2007
Abstract
New data on the dehydration and rehydration processes of calcium, manganese and copper dichlorides are presented that reveal surprising,
in a certain sense, behaviour difficult to be explained for the last two chlorides in terms of the usual conception of thermodynamic equilibrium.
A substantial role of a super disperse phase at studying the equilibrium of the thermal decomposition of a hydrate is postulated to explain the
experimental results for manganese and copper dichlorides. It is shown that the formation of such a phase of the hydrate is able to change appreciably
the experimental results, causing the increase of water vapour pressure and the decrease of the derived enthalpy of a reaction. The results obtained
allow to understand the reasons for considerable differences of some literature data. They enable to receive more precise and reliable data for
thermal dehydration and probably for some other decomposition processes.
©
2007 Elsevier B.V. All rights reserved.
Keywords: Super disperse phase; Thermodynamics; Thermal dehydration; Water vapour pressure; TG; Metal chloride
1
. Introduction
Such an approach is theoretically absolutely correct, except
for the two obvious and important circumstances. The first, the
reality is that the experiment cannot continue infinitely long,
and the second, the process of equilibration may proceed very
slowly. Therefore, in practice we encounter very often a quasi-
equilibrium state of a system, which is not always easy to
distinguish from the true equilibrium. Thus, there is no simple
and definite reply to the above-asked question.
In this paper, we present the experimental results on the dehy-
dration and rehydration processes of calcium, manganese and
copper dichloride hydrates. The data obtained for the last two
chlorides permit us to suppose a substantial role of the interfacial
processes in special cases of the thermodynamic investigations
of thermal dehydration. The choice of these substances was not
accidental – all of them have a rather high desiccating ability,
anhydrous CaCl2 is widely used in industry and laboratories;
MnCl2 is a promising desiccant, and CuCl2 was recently shown
[2,3] to be a desiccant with a very low temperature of regenera-
tion:
In a detailed and comprehensive analysis of numerous exist-
ing data on thermal dehydration of crystalline solids Galwey [1]
introduced and developed new approaches for understanding
the nature of this important group of heterogeneous reactions.
Great attention was paid to the structural changes and mecha-
nisms of interfacial processes and their influence on the thermal
dehydration kinetics; but the problems of equilibrium and ther-
modynamics have not been practically touched.
Do these interfacial processes have any relation to equi-
librium and the received thermodynamic properties of the
reactants?Atfirstsighttheydonot, asequilibriumisastablestate
of a system, not depending on its history and the rate of equili-
bration. That is, the results of a thermodynamic study of thermal
dehydration should not be influenced by any peculiarities of its
mechanism – all the participating phases should have their equi-
librium structure, and the system should be characterized by the
minimum of the Gibbs energy.
1
2
1
CuCl2 · 2H2O(s) ↔ CuCl2(s) + H2O(g).
(1)
2
∗
The latter chloride has an important peculiar property – it
forms the only one hydrate stable at ordinary conditions, the
Corresponding author.
E-mail address: polyachenok@mogilev.by (O.G. Polyachenok).
0
040-6031/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.tca.2007.11.005

O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
45
dihydrate. Therefore, in this case we may expect the easiest
interpretation of the experimental results. Indeed, the most con-
clusive proof on the problem concerned was received for this
chloride in particular.
2
. Experimental
2
.1. Chemicals
All used chemicals were of analytical or chemical grade
purity. Anhydrous CaCl2 and MnCl2 were obtained by careful
and stepped heating of the recrystallized hydrates in the oven up
to the temperature 520 K. Special attention was paid to prepare
pure CuCl2·2H2O, since there are many differences in litera-
ture concerning its characteristics. This hydrate was prepared by
double recrystallization of a commercial sample from hot water,
slightly acidified with HCl. The crystals were washed with cold
water on a glass filter and dried in air; they form small blue
needle-like crystals. They are noticeably hygroscopic, probably
due to the adsorption of water vapour on the surface of the crys-
tals. Anhydrous CuCl2 was obtained by heating the hydrate in
the oven at the temperature 390 K. All the concerned anhydrous
salts are strongly hygroscopic.
2
.2. Experimental procedure
In order to reveal the desired information, the experimen-
Fig. 2. Reaction vessel for the gas transpiration technique: 1, pyrex glass flask;
tal methods should be used allowing to receive not only the
equilibrium knowledge but also some features of the dynamics
of the concerned process. Therefore, two methods of this kind
were chosen to measure the equilibrium water vapour pressure
for the systems, including the anhydrous salts and their lowest
hydrates.
The first one was the gas transpiration technique. A schematic
diagram of the arrangement used is shown in Fig. 1. The flow
of nitrogen was passed through a filter bed of the chloride and
its hydrate (Fig. 2), and the partial water pressure at the outlet
of the apparatus was found gravimetrically.
2
, substance; 3, inner glass tube to withdraw gas; 4, glass fiber filter; 5, sealed
off tube for loading substance; 6, water level in thermostat; 7, gas supply; 8, gas
outlet.
A special procedure was used to prepare granulated chlorides
applied on the surface of silica gel. Silica gel was ignited in the
oven for 10 h at 1270 K to produce 1–2 mm granules with closed
porosity. It was denoted by the considerable growth of bulk den-
sity, from 0.41 g cm for the initial silica gel reduced to the
similar particle size, to 1.28 g cm for the ignited product. The
specific water vapour sorption was determined gravimetrically
−
3
−
3
Fig. 1. Schematic diagram of the arrangement for the gas transpiration technique: 1, nitrogen cylinder; 2, manostat; 3, capillary rheometer; 4, CaCl2 or water saturator
in a Dewar; 5, saturation block; 6, CaCl2 protection; 7, large water manometer; 8, weighed absorption vessel; 9, small water manometer; 10, large bottle to collect
nitrogen; 11, bottle to collect water.

4
6
O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
for both forms of silica gel in the same conditions (indoor air,
94.5–295.3 K, water vapour pressure 15.2–16.1 hPa); it was
2
−
1
equal to 55.2 mg g for the initial silica gel (for the sample
8
−
1
.1719 g) and 0.044 mg g for the ignited sample (13.4890 g).
It is important to note that this small limiting value of water
sorption was reached relatively quickly, in 1.5 h, whereas for
the initial porous sample it lasted for more than 70 h. The main
conclusion that follows from the above data is that silica gel
particles had too little surface area to produce any noticeable
influence on the results of the transpiration experiments. In addi-
tion it is worth to mention that, for instance, the total receptivity
of copper chloride in the reaction flask (76 g) to water vapour
was about 20 g, while for the ignited silica gel particles (100 g)
Fig. 3. Water vapour pressure for the system CaCl2–CaCl2·H2O at 334.05 K.
–
only 4.4 mg.
or CaCl2, depending on the value of the vapour pressure. The
volume of the flowing gas and the time of the experiment were
chosen such that the increase of the mass of the absorption vessel
The procedure of making the chloride granules may be illus-
trated by the example of copper (II) chloride. Hundred grams
of both the ignited silica gel and the ground CuCl2·2H2O were
mixed in a porcelain evaporating dish together with 20 ml of
water, and thoroughly stirred with a glass rod to give a dense
pulp. The latter was heated on the hot plate at continuous stirring
untilallthe solution was evaporated, and then dried at395 K. The
received granules (197 ml and 176 g) were 2–3 mm in diameter,
and contained the average 43 mass% of anhydrous copper (II)
chloride. These granules were put into the reaction flask (Fig. 2)
through the side tube 5 in such a way that all the small particles
of CuCl2, torn off the granules and stuck to the glass, should
be placed below the water level in the thermostat. For the same
reason, glass fiber filters were used (4) to prevent these particles
of CuCl2 from being carried out to the cold zone by the gas flow.
The flow of nitrogen was either dried with calcined CaCl2 (4)
(
8) was never less than 10 mg, to ensure sufficient measurement
accuracy.
The flow rate of the gas has been chosen for each substance
and for each temperature to achieve equilibrium. As is seen in
Figs. 3 and 6, horizontal lines are observed for CaCl2 in a rather
−1
widerange, from55to145andfrom40to170 ml min , andthis
is usually interpreted as indicating the equilibrium attainment.
More convincing is the argument based on coincidence of the
results obtained “from below” and “from above”, and these are
the results received for CaCl2 at 334.05 K (Fig. 3).
It is worth to note that the contact time in these experiments
varied from 1.2 to 5 min, and the total duration of an experiment
was usually 30–170 min, depending on the value of water vapour
pressure, rate and volume of the gas passed. The latter was usu-
ally 1.4–7 l, but in special cases, when the water vapour pressure
was very low, it increased to 99–117 l and then the experiment
lasted for 10–11 h (for CaCl2 at standard temperature).
(Fig. 1) or instead, saturated with water vapour at the appropriate
temperature. In the first case, the partial water pressure in nitro-
gen was lower than the decomposition pressure of the hydrate.
Therefore reaction (1) proceeded from left to right, and we could
achieve the equilibrium state “from below”. In the second case,
on the contrary, reaction (1) proceeded from right to left, and we
could reach the equilibrium state “from above”. Temperature of
the water saturator was chosen in this case from the condition,
that the saturation water pressure should be somewhat higher,
than the decomposition one. For instance, at studying the sys-
tem CaCl2–CaCl2·H2O, the water saturator was placed into a
Dewar flask with melting ice. Delivery of moist and dry gas was
alternated, and reproducible results were achieved.
At the same time, the above-mentioned saturation water pres-
sure was always somewhat lower than the saturation water
pressure at room temperature to avoid condensation in the con-
nections. The latter were made of pvc-tubes and tested not to
pass any perceptible amount of water vapour.
The saturation block (5) (Fig. 1) consisted of a water ther-
mostat with a mercury-contact regulating thermometer, capable
to keep temperature constant to 0.01 K, and the reaction ves-
sel (Fig. 2). Water temperature was measured to 0.01 K with a
calibrated mercury thermometer using a magnifying glass. The
overall uncertainty of the temperature measurements, taking into
consideration all the necessary corrections, was ± 0.05 K.
The mass of water vapour evolved was determined by weigh-
ing (± 0.1 mg) the absorption vessel (8) filled with Mg(ClO4)2
The volume of the flow gas was found gravimetrically with
accuracy to ± 0.1%. As it is seen in Fig. 1, the aspirator
(Marriott’s bottle) was used to collect the gas at the outlet of
the apparatus. It allowed to maintain a constant gas flow and
included two bottles – large ten-litre bottle (10) for the gas and
one-litre bottle (11) for the dripping out water. The latter was
periodically weighed to 1 g and the found mass was converted to
the gas volume V using the water’s density. Water in bottle (10)
was saturated with nitrogen for 0.5 h at the preliminary work of
the apparatus without the absorption vessel (8) to achieve the
steady-state conditions.
Calculations of the water vapour pressure Pw from the exper-
imental results were made using the equation
Pw = (P + ꢀ1P)N,
(2)
b
where (P + ꢀ1P) is the total pressure of the gas not far from
b
the saturation block (Fig. 1), P – the atmospheric pressure,
b
measured with a mercury barometer to 0.1 hPa, ꢀ1P – reading
of the manometer (7) (Fig. 1) and N is the molar fraction of water
vapour in the mixture with nitrogen. It was calculated from the
mass increase of the absorption vessel (8) due to the sorption of
watervapour, andthereceivedamountofnitrogen. Thelatterwas
calculated according to the ideal gas law from V, temperature T

O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
47
of water in bottle (10), pressure P = P − Psat + ꢀ2P, where Psat
were immediately closed with short pvc-tubes with glass beads.
The vessel was cooled in a desiccator and slightly opened before
weighing to equalize pressure.
b
is the saturation vapour pressure of water at the temperature T,
ꢀ
2P – reading of the manometer (9).
Use of eq. (2) may be illustrated with a typical example
for the copper chloride dihydrate at 328.30 K. The measured
The experiments showed that in our conditions the rate of
saturation in the case of water was lower, than for the chlo-
ride systems studied, it is well seen in the diagrams of the
received pressure versus gas flow rate, since the plateau was
usually narrower. It visibly shortened at temperature rise, and
this was the main reason for the upper limit of such measure-
ments. Special experiments have been made to replace nitrogen
with helium. Indeed, the rate of saturation increased noticeably,
but new problems arose due to the rapid diffusion of helium
through the pvc-tubes – it was difficult to obtain a precise mass
of the absorption vessel. Therefore, all the measurements for the
chloride systems were carried out with the nitrogen gas flow.
The following values of the water saturated vapour
pressure were received as a result (obtained/standard val-
ues, hPa) at the temperatures 273.15 K (6.17 ± 0.21/6.11),
290.13 K (19.53 ± 0.16/19.35), 301.98 K (39.65 ± 0.44/39.65)
and 313.31 K (74.4 ± 1.3/74.38), where the confidence interval
(0.95) is given. These results permit us to conclude that the real-
ized variant of the gas transpiration technique enables to receive
sufficiently accurate and reliable data. Thus, we may hope that
all the necessary precautions have been undertaken to ensure the
exact values of the dehydration water vapour pressure obtained
using this method.
quantities were: the atmospheric pressure P = 970.3 hPa; vol-
b
ume of the gas V = 1853 ml; water temperature T = 293.6 K
Psat = 24.0 hPa); ꢀ1P = 20.9 hPa; ꢀ2P = 3.5 hPa; increase in
mass of the vessel (8) 25.4 mg. The resulting value of the water
vapour pressure Pw was calculated to be 19.0 hPa.
(
Precision of the results and certainty, that all the methodical
factors had been taken into proper consideration, were verified
by measuring the equilibrium values of the saturated vapour
pressure for the standard substance – pure water. In this case a
special device was constructed, instead of shown in Fig. 2, to
receive the perfect saturation of the flowing gas, as is illustrated
in Fig. 4.
Glass beads in Fig. 4 played two roles: the first, they slowed
down lifting of the gas bubbles to ensure more complete sat-
uration, and the second, they inhibited water spattering at its
surface. The absorption vessel (4) was filled with Mg(ClO4)2 or
CaCl2, and its upper part was positioned above the thermostat
water to ensure the best absorption of water vapour. The protec-
tive skirt (5) was made of filter paper and prevented the trickling
down water from getting into the inlet of the vessel (4).
At the end of the experiment, the upper rubber stopper (Fig. 4)
was withdrawn together with the vessel (4) and both openings
The other method was the weight modification of the static
method, using the thermobalance of the Paulik–Paulik–Erdey Q-
derivatograph (TG). Its sensitivity allowed observing a change
in mass of a sample down to ± 0.1 mg. The gas flow of dry argon
or argon saturated at fixed temperature with water vapour passed
−1
over the substance at the rate of 15–20 ml min . If this water
pressure was equal to the pressure of thermal dissociation of
the hydrate, the indication of the balance was constant. Other-
wise it showed the increase or loss of the substance mass that
was determined visually on the scale of the balance. Measure-
ments were made at the constant water vapour pressure, and
the temperature was varied, until both became constant – the
mass and the temperature. Thus, we could observe the equilib-
rium state at the fixed temperature and water vapour pressure,
or, more precisely, the quasi-equilibrium, for it lasted no longer
than 10–15 min.
Pure argon was dried with anhydrous Mg(ClO4)2. Otherwise
it was saturated with water vapour by passing through a spe-
cial vessel, containing distilled water and placed into the Dewar
flask, also filled with water. The temperature of this water was
measured with a mercury thermometer to 0.1 K and was chosen
according to the two conditions. The first, it should be as high
as possible since the sensitivity of this method increase with
the water vapour pressure. The second, it should be somewhat
lower than the room temperature to avoid the vapour condensa-
tion in the connections. Thus, it was chosen in a narrow interval
of 286–290 K, providing the water vapour pressure in the range
of 15–20 hPa. The overall uncertainty of the water vapour pres-
sure measurements in these experiments was estimated to be
Fig. 4. Device for measurement of saturation water vapour pressure: 1, two
bubblers for preliminary saturation; 2, water with glass beads; 3, thermometer;
±
0.5 hPa. The degree of saturation was sometimes checked by
weighing the absorption vessel, placed instead of the device
4
, weighed absorption vessel; 5, protective skirt; 6, water level in thermostat; 7,
gas supply; 8, gas outlet.

4
8
O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
tion of argon. Microscope photographs for this system were
obtained using Axio Imager.A1, Carl Zeiss, and powder X-ray
diffraction (XRD) patterns were recorded on a diffractometer
DRON-3 with Ni-filtered Cu K␣ radiation.
3. Results and discussion
Water vapour pressure for the system CaCl2–CaCl2·H2O
was measured using the gas transpiration method. The results
obtained at 354.30 K are shown in Fig. 6 as a function of the gas
velocity. It is well seen that no substantial velocity dependence
is observed, indicating the equilibrium achievement.
At the same time there are two distinct horizontal lines for
different states of the system. The upper line cannot corre-
spond to the higher known hydrate (CaCl2·2H2O), because the
equilibrium pressure for the dihydrate is about 33 hPa at this
temperature. What are these states then? Our experiment does
not specify their nature. One can only conclude that the differ-
ence in thermodynamic stability of these states is rather small,
◦ −1
rG of the transition is at the level of about 1.6 kJ mol . This
ꢀ
effect may be quite explained in terms of physical differences
between the participating phases, for instance, if there are two
different crystal modifications, or if one of the phases is either
bulk or super disperse.
We use the term “super disperse” because only the particle
size at the rate of about few nanometres may appreciably change
the Gibbs’s energy of the phase (ꢀrG), as it is shown in Fig. 7
for the water droplets. These results were calculated using the
Kelvin equation [4]
Fig. 5. Device for measurement of water vapour pressure (weight variant): 1,
pyrex glass ampoule; 2, substance; 3, Pt–Pt/Rh thermocouple of derivatograph;
ln Pr = ln Po + 2γVr 1(RT)
−
−1
,
(3)
(4)
4
, glass tube to deliver gas to the surface; 5, small non-inertial transparent
rG = 2γVr−
1
,
ꢀ
furnace; 6, gas flow; 7, to the balance of derivatograph.
where Pr is the saturation vapour pressure of the water droplets
(Fig. 5) and filled with fresh calcined CaCl2, after passing 1 l of
with the radii r, P is that for the bulk phase, γ is the excess sur-
o
argon. It was collected in a volumetric flask, filled with water and
turned upside-down. This water was preliminarily saturated with
argon; the correction was made for the saturated water vapour
pressure. In a typical experiment the value of the water vapour
pressure was obtained, using the ideal gas law equation, to be
equal to 19.2 hPa in comparison with the exact value 19.1 hPa
at 289.9 K.
face energy of water, V is its molar volume, R is the universal gas
constant and T is the temperature. It is as usually assumed, as the
first approximation, that the values of γ and V are independent
of the dispersity of the substance.
Does the thermodynamic theory give any chance to distin-
guish the two above-mentioned variants of explanation of the
Fig. 5 illustrates the device used for these measurements.
The temperature of the transparent furnace (5) was set with
the electrical power system, including a voltage stabilizer, an
autotransformer and an ammeter. The furnace was almost non-
inertial, so that change of the electric current gave a new constant
temperature in few minutes. Thus, we succeeded in a rapid
increase and lowering temperature over and over again, and
measuring the water vapour pressure in its dynamics.
The Pt–Pt/Rh thermocouple of the derivatograph was cali-
brated in the same furnace by using the precise thermometer,
melting temperature of ice and boiling temperatures of pure
organic liquids. The resulting uncertainty of the temperature
measurements was obtained to be ± 0.8 K.
The specific surface area for the system CuCl2–CuCl2·2H2O
sampleswasmeasuredbythemethodoflowtemperatureadsorp-
Fig. 6. Water vapour pressure for the system CaCl2–CaCl2·H2O at 354.30 K.

O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
49
In Fig. 8a, lines (1) and (2) correspond to the stable forms
ꢀ
ꢀ
of the monohydrate, and lines (1 ) and (2 ) – to the metastable
forms, and at the temperature Ttr a change of the stable form
may be observed. Therefore, the usual form of such a functional
dependence is just shown in Fig. 8a – crossing firm lines. As it is
ꢀ
ꢀ
noted above, the metastable equilibrium lines (1 ) and (2 ) may
also be observed.
In Fig. 8b, an unusual form of such dependence is represented
for the second above-mentioned case, when one of the phases
(monohydrate) is either bulk or super disperse:
CaCl2·H2O(s,bulk) ↔ CaCl2(s) + H2O(g),
CaCl2·H2O(s,disperse) ↔ CaCl2(s) + H2O(g).
(7)
(8)
Fig. 7. Change of the Gibbs’s energy with the radii of water droplets.
We shall assume for the beginning that the dispersity of the
second phase (r in Eq. (4)) is constant and does not change with
time. Use of Eq. (4) may raise well-founded questions in case of
the quantitative calculations for solids, but we may undoubtedly
use it for the deduction of qualitative relationships. For that we
may assume the spherical shape of the concerned particles.
Turning back to Fig. 8b we may assert that, as it is shown
below, in this case line (1) represents the stable equilibrium
for the bulk phase (Eq. (7)), and line (2) – equilibrium for the
unstable super disperse phase (Eq. (8)).
observed phenomenon (Fig. 6)? In the following we shall try to
prove that it does.
Let us briefly review possible behaviour of a three-phase and
two-component system, undergoing in a rather narrow inter-
val of temperature two different transformations with closely
◦
◦
related, but not exactly equal values of ꢀrH and ꢀrS . The reac-
tions (5) and (6) for the two different forms of calcium chloride
monohydrate (␣ and ␤) give a proper example of such a system:
The following thermodynamic relations for the system
concerned allow to make more clear some features of the phe-
nomenon discussed. The super disperse phase of the hydrate,
as is well known, has a higher Gibbs energy, so that the value
CaCl2·H2O(s,␣) ↔ CaCl2(s) + H2O(g),
CaCl2·H2O(s,␤) ↔ CaCl2(s) + H2O(g).
As it is well known, the temperature dependence of the cor-
responding equilibrium constants KP = Pw,
where Pw is the equilibrium water vapour pressure for the
reaction either (5) or (6), will be, in the coordinates ln P–T
two crossing straight lines as is shown in Fig. 8a. At the unique
temperature of transition Ttr the system will be four-phase and
non-variant, and at temperatures higher or lower only one form
(5)
(6)
◦
of ꢀrG for the thermal dissociation of this phase is, as shown
in Fig. 9, lower, than for the bulk stable phase. Therefore the
water vapour pressure Pw, calculated from the thermodynamic
equation
−
1
,
◦
ꢀ G = −RT ln P ,
r
w
(9)
should be higher for the disperse phase and increase as the
reduction range rise.
(␣ or ␤) will be stable, while the other – thermodynamically
unstable. It does not mean that a phase cannot be obtained exper-
imentally at temperatures where it is unstable, but it does mean
that at this temperature it will be metastable, i.e. may exist tem-
porarily, for more or less time without any transformation. We
repeat here this well-known truth for it is very important for our
further discussion.
Analogous diagram may be designed also for enthalpy, so
that the enthalpy of the process (8) should be lower, than for
the process (7). Thus, we obtain line (2) in Fig. 8b for the
thermal decomposition of a super disperse phase. Therefore,
the formation of a super disperse hydrate in the process of this
Fig. 8. Different forms of dependence of ln Pw versus reciprocal temperature:
(a) usual form for two crystal modifications of the hydrate and (b) unusual form
for the super disperse phase of the hydrate.
Fig. 9. Change of the Gibbs energy for the thermal dissociation of CaCl2·H2O.

5
0
O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
type leads not only to the increase of the “equilibrium” water
vapour pressure but also to underestimation of the enthalpy of
this process.
It is very important to understand, that the upper line (2) in
ꢀ
Fig. 8b has a double nature (the same is for line (2 ) in Fig. 8a)
–
on the one hand, it corresponds to the true equilibrium of the
super disperse phase with water vapour, and the value of Pw in
this case will be constant until the particle size is unchanged; on
the other hand, the nano-size particles themselves are thermo-
dynamically unstable and should sooner or later grow and turn
into the bulk phase. From this point of view line (2), as well as
ꢀ
line (2 ) in Fig. 8a, corresponds to the unstable equilibrium.
The most striking difference between these two figures lies
ꢀ
in the character of lines (2) (Fig. 8a) and (2 ) (Fig. 8b), and
ꢀ
also of the appropriate lines (1) and (1 ). Prolongation of line
2) in Fig. 8b, that is line (2 ), is obviously impossible, for in
Fig. 11. Water vapour pressure for the system MnCl2–MnCl2·H2O.
ꢀ
(
such systems the disperse phase should always be thermody-
namically unstable. Hence, line (1) in Fig. 8b corresponds to the
true equilibrium both before and after its intersection with line
Our experiments have shown that the data in Fig. 6 (and
Fig. 3) were reproduced over and over again at repeated change
of pressure of the incoming water vapour (measurements from
below and from above). Consequently at 354.30 K (and also
at 344.05 K) equilibrium was established sufficiently rapidly
in both directions. At the highest temperature (364.85 K) we
observe (Fig. 10) only one value because it is situated near
the lines crossing, and the differences are apparently within the
range of the data scattering. At 334.05 and 298.25 K, the rate of
equilibration in one direction may be already so slow that we
also observe only one value.
Water vapour pressure for the system MnCl2–MnCl2·H2O
was measured using both the static method with a glass mem-
brane null-manometer and the dynamic gas transpiration method
(Fig. 11). The data obtained with the static method give a
very important thermodynamic information concerning the cor-
responding equilibrium and are worth of special discussion
elsewhere. What is more important now, is the difference
between the true equilibrium data (static method) and the data
obtained with the gas transpiration method (Fig. 11). The lat-
ter received “from above” and “from below” are in a good
agreement with one another. Thus, in Fig. 11 we observe two
different equilibrium states – line (1) corresponds to the stable,
true equilibrium, and line (2) – to the unstable one.
(
2).
Returning to Fig. 6 we should mention again, that our exper-
iments do not allow to distinguish these two variants – Fig. 8a
and b. For that we should use the data received by Lannung [5],
who has measured the equilibrium of calcium chloride monohy-
drate decomposition with a static method. Lannung [5] observed
ꢀ
indeed two lines (1) and (2 ) (Fig. 8a), and drew a conclusion
that two crystal modifications of the calcium chloride monohy-
drate were formed – ␣ and ␤, as it is shown in Eqs. (5) and (6).
The data of Lannung [5] are presented in Fig. 10 together with
our data, received using the transpiration method.
In order to decide between the two cited variants, we should
know the position of the line in Figs. 8 and 10 at temperatures
higher than Ttr. There is an old source of such information – the
data of Roozeboom [6], also given in Fig. 10. It is well seen that,
in spite of low precision of these data [6], the plotted value is
line (1) continuation. Therefore we should choose, following to
Lannung [5], the variant shown in Fig. 8a, that is to acknowledge
the formation of the two crystal modifications (␣ and ␤) of the
calcium chloride monohydrate.
Now if we compare Fig. 11 for the system
MnCl2–MnCl2·H2O with Fig. 8, then it is well seen that
it closely resembles Fig. 8b. Therefore, we may conclude that
line (2) in Fig. 11 corresponds to the unstable equilibrium of the
super disperse phase of the manganese chloride monohydrate
with water vapour. The data received “from above” and “from
below” are reproduced over and over again and are in good
agreement with one another, and this indicates a rather high
stability of the nano-size particles towards their growth in our
transpiration experiments. As in case of the calcium chloride
monohydrate, the equilibrium pressure for MnCl2·2H2O is at
these temperatures much more than for the monohydrate.
At present we have no information concerning the behaviour
of line (2) in Fig. 11 near to the intersection with line (1) because
of insufficient precision of the experimental data. But from gen-
eral considerations we may suppose that this line should crook
Fig. 10. Thermal decomposition of calcium chloride monohydrate.

O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
51
Fig. 12. Water vapour pressure for the system CuCl2–CuCl2·2H2O.
Fig. 13. Water vapour pressure for CuCl2·2H2O and its saturated solution.
upwards and approach to line (1) more gradually. When these
lines merge, it should mean that in conditions of the given
dynamic experiment nano-size particles of the hydrate have
enough time to grow into the sufficiently coarse-grained, i.e.
bulk phase.
Water vapour pressure for the system CuCl2–CuCl2·2H2O
was measured using three methods – the static with a glass mem-
brane null-manometer, the dynamic gas transpiration method
and the weight modification of the static method. Special X-
ray investigation confirmed the absence of any intermediate
hydrate in this system – the received XRD patterns for CuCl2,
CuCl2·2H2O and their mixture (1:1), heated at 390 K for 6 h, are
shown in Tables S1–S3 of the Supplementary Data.
These experimental data are shown in Fig. 13. One can see
that our data for reaction (1), designated as circles, are in good
agreement with the equilibrium data of [7] (the right line in
Fig. 13). They were received “from below” and never “from
above” for the hydrate with a substantial degree of decomposi-
tion (30–60%) at the relatively high temperature (340–350 K),
when the supposed super disperse phase should completely van-
ish. Obviously they correspond to the stable, bulk phase of
CuCl2·2H2O. The data for the saturated solution of CuCl2·2H2O
in water are also quite close to the data of [8,9] (left line in
Fig. 13). Therefore, reliability of the results obtained by the
used experimental method is surely verified. However, the most
surprising results are those designated as squares and placed in
the middle of Fig. 13.
The tensimetric results resemble those shown in Fig. 11,
except for the difference between the lines is much smaller
(
Fig. 12). This may mean that the rate of the nano-scale par-
Three left values were received “from above” at a slow sub-
stance cooling with a subsequent stop of the temperature. We
surmise that the multiple supersaturation caused the formation
of the nucleation centers of the hydrate at the surface of CuCl2.
We may imagine that the initial nanometric particles of the dihy-
drate with variable characteristics depending on their size are
colloidally dispersed on the surface of relatively large grains of
anhydrous copper dichloride. In presence of water vapour, a suf-
ficiently rapid growth or aggregation and formation of grosser
particles occur in one way or another, so that after some time
their properties practically do not depend on their size.
It was most difficult to obtain the values nearer to the right
line (Fig. 13). They were caught after cooling the substance to
the beginning of water sorption with a subsequent slow rise of
the temperature until the quasi-equilibrium was reached.
Now the idea of a super disperse phase formation supposed
on basis of the thermodynamic considerations and the above
given experiments, receives new indirect evidence. We observe
in Fig. 13 a real change of the thermodynamic properties of the
unstable phase with time, i.e. rise of its temperature of decom-
position, and the corresponding values in Fig. 13 tend to shift to
the right. We may say that after a short time amounting to some
minutes, the conversion to the more and more stable phase is
observed, i.e. to the more and more coarse hydrate.
ticles growth is much higher, than in the previous case, so that
the transpiration data are more close to the true equilibrium. The
data received “from above” and “from below” are well repro-
duced. The main conclusion resulting from these experiments is
that in case of copper (II) chloride we may also suppose the for-
mation of the super disperse phase of the hydrate CuCl2·2H2O
in the process of water absorption by the anhydrous salt. But
the most interesting results for this last system were obtained
using the weight variant of the static method. Here, the above-
mentioned higher rate of the particles growth enabled us to
observe a real change of these phase thermodynamic properties
during the experiment.
Approximately 0.4 g of ground CuCl2·2H2O was placed into
the ampoule (1) (Fig. 5), and the gas flow of dry argon passed
over the surface of the substance at room temperature. The bal-
ance immediately showed loss of mass due to reaction (1). The
rate of this process increased with rise of temperature, and soon
the surface turned brown, indicating that the thermal dissocia-
tion spreads little-by-little deep into the substance. In that way,
passing argon dry or saturated with water over the substance, we
could achieve any desired fractional conversion in both direc-
tions. Moreover, at room temperature we could obtain a desired
amount of the saturated solution of CuCl2·2H2O in water.

5
2
O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
Many questions may arise concerning the phenomenon con-
der, for its lifetime is short. Moreover, such a phase should
quickly decompose under the conditions of such measurements
(dry helium + argon). Microscope photographs for this system
(Fig. S1a and Fig. S1b) cannot give any information about nano-
scale particles for their invisibility in an optical microscope.
However, as may be seen, these photographs do not contradict
to the theory of nucleation and growth – the brown surface of
the large anhydrous CuCl2 particles (55–100 ␮m) is dotted with
small blue crystals of CuCl2·2H2O (0.9–2.5 ␮m). It is necessary
to mention that electron microscope is also useless at study-
ing these objects for they should immediately decompose in
vacuum.
sidered, and by no means all of them may be earnestly answered
at present. One of the questions, that we are able to make clear
now is the following. In view of the above shown results the data
in Ref. [10] may relate to the unstable equilibrium of the reac-
tion concerned, including the formation of the super disperse
phase of the dihydrate CuCl2·2H2O. Is it possible to obtain non-
equilibrium data skilfully using the most precise and rigorous
static method? We do not know all the experimental details of
the work [10], but we can suppose a situation, that may lead to
such results.
Let us assume that a sample of pure hydrate CuCl2·2H2O
is taken for the investigation. Then the amount of the received
anhydrous salt in conditions of the static method might be negli-
gible. Therefore the process of evolution of the system towards
the stable equilibrium “from above” will be accordingly very
slow, as the rate of the nucleation and growth process for the
hydrate should be proportional to the surface area of the anhy-
drous salt. So in this case it is quite possible to obtain values of
the water vapour pressure relating to the unstable equilibrium.
In our experiments there was always a sufficient amount of both
products.
A simple rule may be formulated from the foregoing for
studying a process of thermal dehydration, and obviously other
processes of decomposition, using the static method. In contrast
to vaporisation, no pure samples might be used; they should be
necessarily a close mixture of the substance studied and the solid
product of its decomposition in approximately equal quantities.
This simple rule may help one to avoid different experimental
complications and obtain more reliable thermodynamic infor-
mation.
4. Conclusion
Rather surprising findings have been received as a result
of the thermodynamic study of the dehydration and rehydra-
tion processes of calcium, manganese and copper dichlorides,
resembling each other within certain limits. In case of CaCl2
these results are accounted by the formation of the two differ-
ent crystal modifications of the monohydrate, but for the rest a
hypothesis is advanced of a hydrate super disperse phase forma-
tion.
It is necessary to emphasize that we have received no con-
vincing and actual proof of the super disperse phase formation,
and it is not clear whether it could be received for the substances
discussed at all. However, this idea enables us to suppose a uni-
fied and sequential interpretation of the main results obtained.
It seems to be a boundary area between the chemical kinet-
ics and thermodynamics, their dominant role being defined by
the chemical rate and the time of thermodynamic measure-
ments.
Inaccuracy of the derived reaction thermodynamics and of
the standard thermodynamic constants of the hydrate, arising
from the influence of this phenomenon, may be illustrated with
the data for the hydrate CuCl2·2H2O, since the authors of dif-
ferent reference books [11,12] proceed in this case from not
the same experimental data. The most convincing are the val-
It is shown that such a super disperse phase formation is able
to change appreciably the experimental results. More precise
and reliable data for the thermal dehydration processes may be
obtained taking into consideration this phenomenon.
◦
ues of its standard entropy S (298.15), equal to 167 [11] and
Acknowledgements
−1
−1
1
90.63 J mol
K
[12], the difference being much more than
any expected experimental error. Therefore, reinvestigation is
in some cases necessary to achieve more precise and more reli-
able thermodynamic data for the hydrates taking into account
the above stated considerations.
It is necessary to emphasize the practical significance of the
discussed phenomenon, regardless of its theoretical details. It
may undoubtedly influence the results of experimental study of
such reactions using the dynamic methods like the transpiration
one, but it should also be taken into account using all the other
methods including, as we have seen, the most accurate, static
one.
The authors express thanks to N.V. Akulich, Regional Scien-
tific Centre of Kuleshov State University, Mogilev, for help in
the microscopic measurements, and to Yu.G. Zonov, Institute of
General and Inorganic Chemistry, Belarus Academy of Science,
for help in the X-ray investigations.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.tca.2007.11.005.
Can we obtain any direct evidence of the super disperse phase
formation? For that we have tried to use, but without any success,
two methods – specific area measurements of different samples
of the system CuCl2–CuCl2·2H2O (Table S4) and microscopical
analysis of this system (Fig. S1a and Fig. S1b of the Supple-
mentary Data). None of the samples studied gave any evidence
of the super disperse phase formation; and it is no won-
References
[
[
1] A.K. Galwey, Thermochim. Acta 355 (2000) 181–238.
2] O.G. Polyachenok, L.D. Polyachenok, E.N. Dudkina, Method for gas dehy-
dration, Belarus Patent No. 7795 2002-09-10.
[
3] O.G. Polyachenok, L.D. Polyachenok, E.N. Dudkina, Method for con-
trolled gas dehydration, Belarus Patent No. 8998 2004-05-19.
[4] A.W. Adamson, Physical Chemistry of Surfaces, Wiley, New York, 1976.

O.G. Polyachenok et al. / Thermochimica Acta 467 (2008) 44–53
53
[
[
[
[
[
5] A. Lannung, Z. anorg. allg. Chem. 228 (1936) 1–18.
6] B. Roozeboom, Z. Phys. Chem. 4 (1889) 31–53.
7] R. Perret, Bull. Soc. Chim. France 2 (1966) 755–758.
8] M. Diesnis, Bull. Soc. Chim. France 2 (5) (1935) 1901–1907.
9] M.H. Lescoeur, Ann. Chim. Phys. (7) 2 (1894) 78–117.
[11] D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, I. Halow, S.M.
Bailey, K.L. Churney, R.L. Nuttall, J. Phys. Chem. Ref. Data 11 (1982)
2–392.
[12] V.P. Glushko (Ed.), Thermal Constants of Substances, Reference Book in
10 Issues, VINITI, Moscow, 1965–1982 (in Russian).
[
10] I.H. Derby, V. Yngve, J. Am. Chem. Soc. 38 (1916) 1439–1451.