Role of liquid water in crystalline hydrate dehydration. Copper sulphate pentahydrate

Article abstract of DOI:10.1023/A:1010161109122

Simultaneous Differential Thermal Analysis/Thermogravimetric experiments carried out on one large single crystal, several small single crystals and powdered crystals of pentahydrate copper sulphate have been used to demonstrate the role that retained liquid water plays in maintaining crystal morphology during dehydration. Measured activation energies for stepwise dehydration in the system show the presence of solution-based transformations provide lower energy paths for the dehydration steps and stress relieving mechanisms. Skeletal anhydrous crystals from large-sized pentahydrate copper sulphate have the same morphology as the starting crystal on complete dehydration at controlled heating rates as long as a solution phase is maintained within the crystal during decomposition. The athermal activation energies, in this work, are in agreement with those obtained by isothermal methods as long as coincident reaction paths for the two techniques are maintained. The literature has been reviewed in the light of this work and a three-stage process is presented to rationalise the conflicting obtained by workers using a variety of different experimental techniques.

Full text of DOI:10.1023/A:1010161109122

Journal of Thermal Analysis and Calorimetry, Vol. 63 (2001) 233–247
THE ROLE OF LIQUID WATER IN CRYSTALLINE
HYDRATE DEHYDRATION
Copper sulphate pentahydrate
S. de Castelnuovo1*, J. B. Harness2 and I. J. McColm1**
1Department of Industrial Technology, University of Bradford, Bradford,BD7 1DP, United Kingdom
2Department of Chemical Engineering, University of Bradford, Bradford, BD7 1DP, United Kingdom
(Received May 14, 1999; in revised form October 11, 2000)
Abstract
Simultaneous Differential Thermal Analysis/Thermogravimetric experiments carried out on one
large single crystal, several small single crystals and powdered crystals of pentahydrate copper sul-
phate have been used to demonstrate the role that retained liquid water plays in maintaining crystal
morphology during dehydration. Measured activation energies for stepwise dehydration in the sys-
tem show the presence of solution-based transformations provide lower energy paths for the dehy-
dration steps and stress relieving mechanisms. Skeletal anhydrous crystals from large-sized penta-
hydrate copper sulphate have the same morphology as the starting crystal on complete dehydration
at controlled heating rates as long as a solution phase is maintained within the crystal during decom-
position. The athermal activation energies, in this work, are in agreement with those obtained by iso-
thermal methods as long as coincident reaction paths for the two techniques are maintained. The lit-
erature has been reviewed in the light of this work and a three-stage process is presented to
rationalise the conflicting information obtained by workers using a variety of different experimental
techniques.
Keywords: activation energies, copper sulphate, dehydration
Introduction
Well-crystallised, pentahydrate copper sulphate (CuSO4·5H2O(c)) is a material used to test
and sometimes calibrate Differential Thermal Analysis/Thermogravimetric (DTA/TG)
instruments. On reviewing the literature this is perhaps not surprising because from the
nineteen thirties [1] to date, the dehydration of copper sulphate has been a continuous
subject of study. The problem has been approached from as diverse directions as fine fo-
cus X-ray diffraction using topotactic methods [2]; gas chromatographic detection of
*
Temporary address: ‘Franz Viegener S.A.’, Industria de porcelana sanitaria, P.O. Box 17-03-4640,
Quito, Ecuador
Author for correspondence: E-mail: I.J.McColm@Bradford.ac.uk
**
1418–2874/2001/ $ 5.00
Akadémiai Kiadó, Budapest
© 2001 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

234
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
evolved water as a function of heating rate [3]; electrical conductivity determination
[4–7]; open pan and closed pan DTA and/or TG [8–10]; infrared measurements [11];
thermochemical measurements [12, 13]; the role of particle size; heating rate and gas at-
mosphere effects [14]. This diversity has lead to conflicting suggestions for the reaction
mechanism and has produced energy data that is not always coincident. Some confusion
arises from the apparent insulated way that separate groups have done their work. Effort
over such a long time tends to foster the belief that the peak positions in a DTA analysis
of the process and the enthalpy of the reactions are well established. Indeed, since we
were at the time investigating drug-water and clay dehydration reactions we assumed at
an early stage that the enthalpy associated with the removal of one water molecule from
CuSO4·5H2O is a fifth of the value for the overall process as quoted by earlier workers. In
the end this was not such a bad decision, but, because calibrations involving the first, that
is lowest temperature, DTA peak and associated mass loss, gave more inconsistent re-
sults, than for the other two peaks on the thermal analysis trace, we were encouraged to
consult the literature more fully. The findings of the search resulted in us performing a se-
ries of thermal analysis experiments using large, several small and powdered crystals of
CuSO4·5H2O(c).
Arguably the key work in the very diverse earlier reports appears to be the com-
bined electrical conductivity-thermogravimetric and DTA work of Wendtlandt and
his group [4–7] using both unsealed and sealed crucibles containing copper sulphate
crystals. It is against these results and the detailed study of the effect of different cru-
cible geometry of Paulik et al. [9], that the contradictions of other work can be ration-
alised in terms of whether or not equilibrium is established with water vapour pres-
sure above the decomposing hydrate. The dominant role of the water vapour pressure
and the extent to which water transport is impeded through channels of molecular di-
mensions was the focus of some of the earliest work with hydrates, such as
CuSO4·5H2O [15]. It was shown that dehydration usually leads to structural degrada-
tion except in strong covalent network systems, such as zeolites. Structural collapse
leads to impedance of water removal from the system, which probably produces some
of the variations in reported literature data. From the resultant amorphous phase a va-
riety of nucleation processes precede the appearance of a new phase. Nucleation in
the sub-surface regions produces characteristic shapes related to the original crystal-
line phase. It follows from this that a number of reaction paths relating to evaporation
of water from within hydrates exposed to increasing temperatures have been pro-
posed [9, 10].
The fact is usually overlooked that some large crystals maintain their morphol-
ogy through extremely large volume decreases. The anhydrous material is much
smaller but a close copy of the original hydrate. Hence reaction path and related rate
proposals must encompass satisfactory stress relief mechanisms. This work proposes
mechanisms that might produce the stress relief.
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
235
Proposed reaction paths
From a review of the literature we propose the following overall view of the process
in a series of reaction steps. In closed crucible conditions the process can be described
by the following five stages as the temperature increases:
Stage 1, the dehydration of the pentahydrate
CuSO4·5H2O(c) → CuSO4·3H2O(c)+2H2O(l)
(1)
This produces an endothermic event on DTA/TG examination with minimal
mass loss by evaporation from the system.
Stage 2, the formation of localised pockets of saturated solution from the liberated
water
Where the released water is in contact with CuSO4·5H2O(c) and also now CuSO4·3H2O(c),
both are highly soluble salts and a simple physical dissolution reaction can take place:
H2O(l)+CuSO4·xH2O(c) → (saturated solution)(l)
(2)
The saturated solution is at localised points within the crystal and little mass
change occurs by evaporation from the surface. The heat of solution for these pro-
cesses causes some uncertainty in the determination of the overall enthalpy change.
Stage 3, formation of the trihydrate
(Saturated solution)(l) → H2O(g)+CuSO4·3H2O(c)
(3)
This stage produces a very strong endothermic effect with of course a loss of
mass from the reaction pan.
When these three stages occur in open pan conditions it is common that only one
endothermic effect is observed, with on occasion a shoulder on the curve, dependent
on the heating rate, as localised conditions lead to a partial separation of the stages.
Topotactic X-ray experiments [2] concerning nucleation of this stage were done
without the any admitted pre-knowledge of these consecutive solution–evaporation
reactions and so no crystal structure relationship was found. Furthermore reports in
the literature of an amorphous transition from penta to monohydrate are to some ex-
tent understandable. The peak maximum temperature for this stage of the decomposi-
tion process in open pan conditions will be affected by several factors other than just
the heating rate, for example sample size and impedance to water flow, as suggested
by Topley [16], which will determine the period during which liquid water is retained
as part of the system.
Stage 4, formation of the monohydrate
CuSO4·3H2O(c) → CuSO4·H2O(c)+2H2O(g)
(4)
J. Therm. Anal. Cal., 63, 2001

236
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
This endothermic reaction leads to further mass loss from the open pan with a
heating rate dependent, peak maximum in a DTA analysis.
Stage 5, the production of the anhydrous sulphate
CuSO4·H2O(c) → CuSO4(c)+H2O(g)
(5)
This is the final stage and gives a clear third DTA peak in an open pan experi-
ment or a fourth peak in a closed pan DTA experiment.
Separation of the consecutive dehydration steps in a DTA/TG experiment will
be closely related to the heating rate and whether open or closed pan conditions are
prevailing. However, it should be stressed that the position of all peak maxima are
variable depending on the heating rate itself and this fact can be exploited to deter-
mine the activation energies associated with the various stages in the dehydration
process. For example the work of Kissinger [17] and Ozawa [18] give Arrhenius-type
relationships between the reciprocal of the peak maximum temperatures in degrees
Kelvin and natural logarithm of the heating rate, which allow estimation of activation
energies. For a DTA process an Ozawa plot [18], based on the work of Kissinger [17],
measures the slope of a graph of the reciprocal peak temperature, 1/Tp, vs. the natural
logarithm of the heating rate, α. This gives the activation energy, Q, divided by the
Universal Gas Constant, R, from the expression:
mQ
lnα =−
+C
(6)
nRTp
In this equation m and n are constants which characterise various nucleation and
growth mechanisms.
Kissinger [17] developed Eq. (6) and did verification work on re-crystallisation
of glass. Thus the method relates best to processes which involve an amorphous to
crystal transition. If Topley [16] and subsequent workers are correct, a vitally impor-
tant stage in the dehydration of ionic hydrates is the change from an amorphous near
surface layer to a crystalline product and so it was decided that Kissinger’s analysis
of the effect of the heating rate on the DTA endothermic peak positions could be a rel-
evant way to investigate the dehydration process further. Temperature gradients in
the specimen are a possible source of error when using this method and the recorded
temperature will not always be the mean specimen temperature. Since the difference
would depend upon the heating rate, which is the variable used in the method, it was
decided to minimise this problem by limiting the heating rates used to those between
2 and 20°C min–1. This produced peak maximum temperature changes of up to 50°C.
Following the literature review it was decided that the purpose of this experi-
mental work was to establish activation energies for the separate steps in the dehydra-
tion process in open pan conditions, using three forms of CuSO4·5H2O(c) in order to
test and extend the current picture of the reaction mechanism.
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
237
Experimental
Copper sulphate crystals of AnalaR grade manufactured by BDH Chemicals, Poole,
U. K., were dissolved in double distilled water, placed in a flat bottomed evaporating
dish which was covered by a filter paper and allowed to stand undisturbed until the
crystals appeared. The copper sulphate was recrystallised as some of the original crys-
tals did not have the full five waters of crystallisation. Single crystals of about 0.5 to 5
mm crystal edge length were grown and fine crystalline powder was obtained by
grinding the smaller crystals. A large crystal about 5 mm along the longest edge, or
five to seven small crystals of about 0.5 mm, or a fine powder was placed in the speci-
men pan of the DTA, and dry nitrogen gas passed at 40 ml min–1 through the equip-
ment and over the specimen. Various heating rates were used from 2.0 to 20.0°C min–1
from room temperature up to a maximum temperature of 300°C. Continuous readings
were taken of the temperature, enthalpy and mass changes. All experiments were per-
formed with open pans. The DTA/TG used was a Stanton Redcroft STA 780.
All the crystals used were examined in an Olympus BH2-UMA optical micro-
scope and micrographs were printed on a UP-3000P Mavigraph. Some were re-exam-
ined after the heating cycle.
The experiments were performed by two workers independently. Practically
identical results were obtained by both workers except for the large, single crystal
experiments. In the latter the heat transfer between the pan and crystal is not as good
as with the powders and small crystals and is very dependent upon the area of contact
between the pan and the surface of the large crystal.
Results and discussion
The experiments showed that the large single crystals of pentahydrate copper sul-
phate could be completely dehydrated without distorting the template of the original
crystal but just uniformly shrinking it (Fig. 1). Further microscopic examination of
the dehydrated single crystals showed a series of features (Fig. 2). The question is
raised as to what these are?
In Garner’s book [15] similar pictures are shown facing page 219 to those of
Figs 1and 2 and the text describes them as subsurface nuclei. They may have been in
that case but in this work they come from a completely dehydrated sample, not a
partially dehydrated one like that shown by Garner. One must assume that the nuclei
here will have become part of the new anhydrous crystalline phase and therefore the
features are more likely places where water left the crystal and so they are
aggregations of molecular channels through which water has travelled. Indeed some
non-random arrangement might just be seen in Fig. 2. These molecular-sized
channels would provide the impedance needed to establish liquid water environments
within the crystal.
Several of the micrographs shown in Garner’s [15] text could be interpreted as
etched dislocation pits. The craters here could also be such etched pits with the
etchant being water. Nuclei tend to have sharp features which mirror the topography
J. Therm. Anal. Cal., 63, 2001

238
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
of the parent crystal. The pits shown in Fig. 2 and those shown by Garner, are
spherical with fibrous projections. This feature may be indicating some degree of
liquid water interaction with the nuclei or surface traces of the emergent tunnels.
Fig. 1 A single crystal of CuSO4·5H2O(c) before heating (a); Crystal after complete de-
hydration (b). Mag. ×20 for both
Hence one suggestion is that the water left the system along well-defined crys-
tallographic paths which were related in CuSO4·5H2O(c) and the final product
CuSO4(c). This suggestion is not contradictory to the study of Zagray et al. [2] where
the effect of water removal was investigated both microscopically and by fine focus
X-rays. In their work rapid removal of water vapour into high vacuum showed
x-shaped nuclei of CuSO4(c) as vapour left the solid from CuSO4·5H2O(c). Less rapid
water removal produced elliptical nuclei as the CuSO4·3H2O(c) phase was presumed to
crystallise from local pockets of saturated solution. Even though a solution was
Fig. 2 Micrograph of the surface of the crystal. Dark spots are craters where water has
left the sample
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
239
thought to be involved, the crystal axes of CuSO4·3H2O(c) were closely related to
those of CuSO4·5H2O(c) indicating some topotactic process control. This topotactic
control may be a reason why the general shape of the crystal is maintained as shown
in Fig. 1 and the presence of liquid solutions would provide stress-relieving mecha-
nisms. This is why stages 1 and 2 (above), where solutions are formed prior to water
loss, were included in the overall mechanism.
A more detail explanation on a localised model which leaves the crystal essen-
tially intact while implying differences in behaviour between hydrates obtained by
decomposition or by addition of water is:
Stage [A]
CuSO4 3H2O(dec)(c)
xCuSO4 3H2O(c)
saturated
→ solution → (1–x)CuSO4 3H2O(dec)(c)
CuSO4 5H2O(c) →
+
H2O(1)
pockets
yH2O(l) +(2– y)H2O(g)
The materials in the square brackets are not in equilibrium and x and y can be
quite variable depending on heating rate and size of crystal. Note that (dec) means
that the material in the sample holder used for subsequent heat treatment was ob-
tained by decomposition of the previous higher hydrate.
Stage [B]
Fast heating CuSO4 H2O(dec)(c) +2H2O(g)
CuSO4 3H2O(c)
CuSO4 H2O(dec)(c) +zH2O(g)
Slow heating
concurrent with
CuSO4 H2O(dec)(c) +(2–z)H2O(l)
Again, as in stage A, z is variable and difficult to measure experimentally.
Stage [C]
CuSO4·H2O(c) → CuSO4(c)+H2O(g)
These stages are most strongly supported by the work in Wendlandt’s group
[4–7] whose combined DTA and electrical conductivity measurements definitely
show the presence of liquid water to beyond 102°C at slow heating rates for initial de-
composition of CuSO4·5H2O. A quadruple point CuSO4·5H2O, CuSO4·3H2O, liquid
water and water vapour with a partial pressure less than one atmosphere was identi-
fied. In subsequent experiments these workers showed that the presence of liquid wa-
ter depended on the heating rate and whether or not the sample was wrapped in a tight
fitting Teflon cover. This is the reason we have introduced stage B. The proposed
scheme is supported by the detailed thermogravimetric analysis of Paulik and
co-workers [9, 10]. In their work fast heating rates gave ambiguous evidence for the
J. Therm. Anal. Cal., 63, 2001

240
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
presence of liquid water but wrapped crystals always produced this evidence. Hence
hindering the diffusion of water vapour will allow liquid water to collect in the crys-
tal, and since reactions inside relatively large crystals are ‘wrapped’, different reac-
tion activation energies should come from experiments involving substantially differ-
ent particle sized material.
The balance of our work and other reports, see below, leads us to reverse the role
played by the heating rate in stage B, from that previously suggested by considering
electrical conductivity measurements. Such a technique cannot easily distinguish be-
tween water on the surface and water trapped in pockets within the crystal. Hence
confusion can arise as to where the water is situated, which in turn may lead to faulty
interpretation as to the effect of heating rate on the decomposition.
The mechanism involving stages A and B could explain the variable shapes of
the first two peaks in the DTA trace and the appearance of a skeleton CuSO4(c) from
the original hydrate. The presence of the pockets of saturated solution provides a
mechanism for stress relief, which presumably is why the crystal remains intact. The
picture in Fig. 2 suggests that water diffusion and loss can be controlled by crystal-
linity. In fact the large crystal is a ‘perfect, wrapped experiment’. The smaller values
for decomposition activation energies of all three stages, obtained when using the
large crystals, suggest solution mechanisms assist in overcoming energy barriers in
dehydration.
Experiments to assess the activation energy for such processes were focused on
measuring the DTA peak-maximum temperature as a function of heating rate. From
the above scheme more consistent results from the high temperature reaction, de-
scribed as stage C, might be expected, with some inconsistency in data from stage A,
and to a lesser extent stage B. This is because solution enthalpies are involved in step
A and may be in stage B.
Ozawa [18] plots are shown for one large crystal, Fig. 3, for 5–7 medium sized
crystals, Fig. 4 and for powders Fig. 5, for the 3 peaks occurring when open pan ex-
periments are performed. The associated activation energies are given in Table 1. In
each case distinct linear plots for peaks conforming to stages A, B and C in the model
are obtained which implies that the heating rates used are not in a range sufficient to
distinguish the two subsections of step B. It is stressed that all the results are from an
open pan configuration, with dry flowing nitrogen gas to remove the water vapour.
Good correlation coefficients were common for the experiments involving powder
and the collection of small crystals, but more variance was apparent in the single
crystal data. The results suggest that as far as stages A and B are concerned the physi-
cal form of the specimens are sufficiently close as not to produce significant differ-
ences, but the physical form does affect stage C. In all stages the trend in activation
energy values indicate that processes are assisted by the retention of a crystal struc-
ture, decrease in surface area and the presence of liquid water.
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
241
Fig. 3 Ozawa plot for single crystals of CuSO4·5H2O
Fig. 4 Ozawa plot for several small crystals of CuSO4·5H2O
J. Therm. Anal. Cal., 63, 2001

242
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
Fig. 5 Ozawa plot for powdered crystals of CuSO4·5H2O
Table 1 Activation energies for dehydration of copper sulphate crystals
Peak 1/kJ mol–1
A
Peak 2/kJ mol–1
B
Peak 3/kJ mol–1
C
Reaction path
Type of specimen
Powder
88.2
87.4
70.2
119.0
114.0
73.5
155
135
116
Small crystals
One large crystal
Sedlak and Beebe [3] used a temperature programmed desorption method that
measured water in the gas phase as the sample is purged with helium gas. The sam-
ples were collections of small crystals with maximum edge lengths of 0.5 mm. From
plots of temperature vs. concentration of water in the gas stream, at several heating
rates, they were able to make temperature of water evolution peak maxima against
square root of heating rate graphs, to obtain dehydration activation enthalpies. The
method of analysis showed a different heating rate dependence above and below
8.15°C min–1. Lines of different slope were obtained giving a difference of up to 80%
in the activation enthalpy change. Their interpretation was the intervention of a crys-
talline to amorphous phase change. From the data presented in their paper we have
calculated activation energies from their experiments, using the Ozawa [18] tech-
nique. Values of 68.4, 83.7 and 125 kJ mol–1 for the three stages A, B and C as defined
here are obtained. These results are closest to the activation energy values for the
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
243
large crystals we examined here. We were not able to detect in their results [3] any af-
fect of heating rate above and below 8.15°C min–1.
Marcu and Segal [19] used athermal thermogravimetric analysis from a constant
slow heating rate, to establish activation energies for what would be stages A and B
using powdered samples of CuSO4·5H2O. Their activation energies are 115.5 and
135.5 kJ mol–1, respectively. These are higher values than the data for our powdered
samples, as shown in Table 1, and this probably reflects the very slow heating rate
they used allowing a longer time period for equilibrium to be achieved and so, as
Paulik [10] explains, the decomposition does not involve liquid water but is the pro-
cess described by:
CuSO4·5H2O(c) → CuSO4·3H2O(c)+H2O(g)
(7)
This is in agreement with our work, which suggests that liquid water, and solu-
tion evaporation mechanisms assist the process.
Pope and Sutton [14] used CuSO4·5H2O(c) in seven size fractions, ranging from
–14+18 to –100+150 BS mesh. They proposed three distinct endothermic processes
contributing to the observed endothermic peaks for the decomposition to the mono-
hydrate. Interestingly one of these is a liquid water to water vapour transition. With
the largest particles (–14+18 BS mesh), the first stage of decomposition occurs com-
paratively slowly, due to the small surface/mass ratio and the time taken for the
evolved water to diffuse to the particle surfaces. The corresponding endothermic
peaks tend to merge, as liquid water is present throughout the process. Samples in the
range –52+72 BS mesh show three distinct peaks corresponding to the three distinct
endothermic processes. When using samples with the widest size range three peaks
always are observed. This seems to support the importance of water removal paths,
which are size and time dependent.
Langfelderova et al. [13] also used athermal thermogravimetry to find activation en-
ergies for what would be stages A, B and C here but introduced a significant change by
making measurements on freshly prepared CuSO4·3H2O(c) and CuSO4·H2O(c) crystals.
They compared the data from these fresh crystals to that obtained from a step-wise de-
composition of CuSO4·5H2O(c) down to anhydrous copper sulphate. The results show that
a well defined, freshly prepared, crystalline sample has a much lower activation energy at
the final stage, C, and a lower activation energy for the B stage, in contrast to anhydrous
CuSO4 produced by three decomposition stages. This data highlights the role that sample
crystallinity plays in the mechanism of dehydration. The progress from penta to
anhydrate by decomposition:
CuSO4·5H2O(c) → CuSO4·3H2O(dec)(c) → CuSO4·H2O(dec)(c) → CuSO4
has an activation energy of 216.9 kJ mol–1 for the final step, while the activation en-
ergy to obtain CuSO4(c) from fresh crystalline CuSO4·H2O(c) is only 126.8 kJ mol–1.
These results are summarised in Table 2 and are useful to compare with the activation
energies for the large single crystals we measured, which maintained their morphol-
ogy. Our average value for both large and small crystals is 126 kJ mol–1. From this it
could be argued that fresh CuSO4·3H2O(c) crystals are formed from crystallisation
J. Therm. Anal. Cal., 63, 2001

244
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
from the solution prior to the decomposition leading to stage C. Thus, experiments
performed at a rate able to maintain crystal morphology throughout, have a mecha-
nism which reflects closed container conditions that allow a solution – recrystallisa-
tion route. This observation fits well with Pope and Sutton’s work [14] on the effect
of particle size.
Table 2 Activation energies from Langfelderova et al., 1980 [13]
Reaction path as
Langfelderova et al., 1980
Activation energy/
kJ mol–1
Reaction*
Notes
Langfelderova et al. has
soln → (c) + (g)
here we have
CuSO4·5H2O(c)
↓
CuSO4·3H2O(dec)
A
124.9
198.8
126.8
(c) → (c)+(l)+(g)
CuSO4·3H2O(c)
↓
CuSO4·H2O(dec)
(c) → (c)+(g)
(c) → (c)+(g)
CuSO4·H2O(c)
↓
CuSO4(dec)
CuSO4·3H2O(dec)
↓
CuSO4·H2O(dec)
B
C
235.4
216.9
CuSO4·H2O(dec)(dec)
↓
CuSO4(dec)
The (dec)(dec) notation
shows that originally
this was CuSO4·5H2O
*Refers to the reactions pathways described in the text.
To a lesser extent Langfelderova et al.’s [13] data show the crystal structure ef-
fect; when the reaction:
CuSO4·3H2O(c) → CuSO4·H2O(dec)(c)
is compared to:
CuSO4·3H2O(dec)(c) → CuSO4·H2O(dec)(c)
which sequence is stage B in the above scheme but approached from different sample
histories, the decomposition of freshly prepared CuSO4·3H2O(c) gave a lower activa-
tion energy value at 198.8 kJ mol–1 than the 235 kJ mol–1 for the CuSO4·3H2O(dec)(c)
→
CuSO4·H2O(dec)(c) measured in this work. The same reaction for the small crystals we
used gave activation energies around 114 kJ mol–1, which is significantly below
Langfelderova et al.’s [13] value. We suggest that this reflects the presence of newly
crystallised CuSO4·3H2O(c) from the saturated solution pockets that are present up to
120°C. Crystals, which maintain coherence, give easier passage of water to the atmo-
sphere and it seems that Langfelderova et al.’s [13] results and those obtained here
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
245
agree over this point. The overall implication is that activation energies can be domi-
nated by the presence of liquid water within the crystal, crystal perfection, particle
size and the passage of water to the atmosphere.
Tanaka [12] using powdered, well crystallised CuSO4·5H2O(c) measured activa-
tion energies for the three stages of decomposition occurring at 52, 77 and 200 oC us-
ing both isothermal and athermal thermogravimetric methods. His values for the acti-
vation energies are 72.9, 106.5 and 157.1 kJ mol–1, which are in good agreement with
our data in Table 1 for powdered samples. Thus the use of DTA peak temperatures
and the Ozawa analysis in a variable heating rate experiment gives data close to that
obtained from accepted, isothermal methods for CuSO4·5H2O(c) dehydration. This
would seem to support our original view that a method developed for amorphous to
crystalline transitions would work.
Paulik et al. [9] and Arthur and Redfern [8] sounded a warning that care should
be taken in the interpretation and significance of TG curves when studying dehydra-
tion. They did experiments using CuSO4·5H2O(c) in open and closed pan conditions
and their data confirms the importance of water diffusion in determining values of the
activation energies and reaction enthalpies.
Table 3 Thermodynamic and derived properties of the hydrates of copper sulphate
∆Hfo/
Dehydration
enthalpy/kJ mol–1
∆H per mol H2O/
Compound
Crystal symmetry
kJ mol–1
kJ mol–1
CuSO4·5H2O
CuSO4·3H2O
CuSO4·H2O
CuSO4
–2277.4
–1688.7
–1084.4
–770.4
1506.9
301.4
triclinic
monoclinic
monoclinic
rhombohedral
m–r
1H2O–CuSO4
3H2O–1H2O
3H2O–CuSO4
5H2O–3H2O
313.5
597.0
912.5
594.4
313.5
298.5
304.2
297.2
m–m
m–r
t–m
If a large single crystal is used, can the reaction enthalpy data be used to distin-
guish between different forms of H2O to ion bond within the crystal? Several spectro-
scopic and above all X-ray crystal structure determination studies, have shown that
all the water in the copper sulphate crystal is co-ordinated and none is free. However
four water molecules are associated with the Cu2+ ion and one with the SO42– ion.
Overall the total dehydration enthalpy for the removal of five water molecules is,
from Table 3, 1506.9 kJ mol–1 which gives an average dehydration enthalpy per mol
of water of 301.4 kJ. A closer examination of data in Table 3 indicates an enthalpy
change for removal of the last water as 313.5 kJ mol–1. Does this reflect the different
bonding state or the presence of liquid water in the earlier dehydration steps? It is
hard to decide because at least two crystallographic transformations are involved.
Enthalpies of formation of the various copper sulphate hydrates are given in Table 3,
J. Therm. Anal. Cal., 63, 2001

246
CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
along with some crystallographic information. From this data the value of 301.4 kJ
(mol of water )–1 of water removed is sometimes used to calibrate thermal analysis
peaks using copper sulphates as a standard. This approach smoothes over the fact that
four of the water molecules are co-ordinated to the Cu2+ ion, one is co-ordinated to
SO24– ion and two crystallographic changes are involved in sequential water removal.
Using the data for CuSO4·3H2O dehydrating to CuSO4·H2O involves no crystal struc-
ture change and only the breaking of Cu2+–H2O bonds which gives an average value
of 298.5 kJ (mol H2O)–1. Considering the 5 to 3 H2O dehydration alongside the 3 to 1
H2O, an estimate of 2.6 kJ mol–1 is arrived at for the triclinic to monoclinic conver-
sion. An estimate for the different binding energy between H2O–Cu2+ and H2O–SO42–
comes from the CuSO4·H2O → CuSO4 and CuSO4·3H2O → CuSO4 reactions, which
have a difference in ∆H/H2O removal of 9.3 kJ (mol H2O)–1. This implies a
4.7 kJ mol–1 difference in binding water to cupric ion. Removal of the last water of
crystallisation involves a monoclinic to rhombohedral transformation and breaking
of a H2O to SO24– bond, a process involving 313.5 kJ (mol H2O)–1. This is some 4%
different from the average value frequently used. For our subsequent work on
drug–water reactions and clay hydrations the calibration peak used was that from
Stage C using the value 313.5 kJ because it is not involved with a liquid water stage at
any heating rate used.
* * *
S. Castelnuovo would like to thank the British Council for the gift of a Studentship. J. B. Harness
would like to thank the EPSRC for their grant number GRJ68557 to assist in this work.
Nomenclature
α is the heating rate in °C min–1 also °C s–1
C is a constant
∆Hfo is the heat of formation at 25°C
Q is the activation energy
R is the universal gas constant
kJ mol–1
kJ mol–1
kJ mol–1 K–1
K
Tp is the maximum temperature of the enthalpy peak
References
1 W. E. Grant and H. Tanner, J. Chem. Soc., 1930, 47.
2 A. I. Zagray, N. Z. Zyryanov, N. Z. Lyakhov, A. P. Chupakhin and V. V. Boldyrev,
Thermochim. Acta, 29 (1978) 115.
3 J. M. Sedlak and R. A. Beebe, Thermochim. Acta, 9 (1974) 261.
4 W. W. Wendlandt, Thermochim. Acta, 8 (1969) 11.
5 W. W. Wendlandt, Thermochim. Acta, 1 (1970) 419.
6 W. W. Wendlandt, Thermochim. Acta, 9 (1974) 456.
7 Z. Halmos, L. W. Collins and W. W. Wendlandt, Thermochim. Acta, 8 (1974) 381.
8 J. K. Arthur and J. P. Redfern, J. Thermal Anal., 38 (1992) 1645.
J. Therm. Anal. Cal., 63, 2001

CASTELNUOVO et al.: CRYSTALLINE HYDRATE DEHYDRATION
247
9 J. Paulik, F. Paulik and M. Arnold, J. Thermal Anal., 34 (1988) 1455.
10 F. Paulik, Special Trends in Thermal Analysis, J. Wiley and Sons, Chichester, New York, Bris-
bane, Toronto, Singapore 1995.
11 M. Guerin, M. Lallemant, N. Roudergues and P. Assire, J. Thermal Anal., 30 (1985) 1149.
12 H. Tanaka, Thermochim. Acta, 46 (1981) 139.
13 H. Langfelderova, A. Linkesova and A. Jona, J. Thermal Anal., 19 (1980) 117.
14 M. I. Pope and D. I. Sutton, Thermochim. Acta, 23 (1978) 188.
15 W. E. Garner, Chemistry of the Solid State, Butterworths, London 1955.
16 A. Topley, Proc. Roy. Soc. A, A 136 (1932) 413.
17 H. E. Kissinger, Anal. Chem., 29 (1957) 1702.
18 T. A. Ozawa. J. Thermal Anal., 9 (1976) 369.
19 V. Marcu and E. Segal, Thermochim. Acta, 36 (1980) 235.
20 CRC Handbook of Chemistry and Physics, Ed. R. C. Weast, Pub. CRC Press Inc., New York
and London 1980.
J. Therm. Anal. Cal., 63, 2001