Simultaneous differential scanning calorimetry and reflected light intensity (DSC-RLI) in the study of inorganic materials

Article abstract of DOI:10.1016/S0040-6031(99)00274-9

This paper re-examines work on inorganic materials and oxalates, but combines differential scanning calorimetry (DSC) measurements with simultaneous observation and recording of the reflected light intensity to obtain more detailed information about the p

Full text of DOI:10.1016/S0040-6031(99)00274-9

Thermochimica Acta 340±341 (1999) 285±292
Simultaneous differential scanning calorimetry and re¯ected light
intensity (DSC±RLI) in the study of inorganic materials
Peter J. Haines
Oakland Analytical Services, 38 Oakland Ave., Farnham, Surrey, GU9 9DX, UK
Accepted 6 August 1999
Abstract
This paper re-examines work on inorganic materials and oxalates, but combines differential scanning calorimetry (DSC)
measurements with simultaneous observation and recording of the re¯ected light intensity to obtain more detailed information
about the physical and chemical changes occurring. The method indicates the nature of the changes recorded by the DSC, both
physical and chemical and gives additional con®rmation of the nature of complex and overlapping DSC peaks for ammonium
salts, copper carbonate and the potassium oxalato-complexes of aluminium, copper, iron and chromium. # 1999 Elsevier
Science B.V. All rights reserved.
Keywords: Inorganic chemicals; Thermal analysis; Thermomicroscopy; Thermochromism
1. Introduction
heated between 358C and 4508C and showed that
allotropic changes, dehydration and decomposition
can all cause colour changes. They stress that the
nature of the sample, the rate of heating and the
surrounding atmosphere may affect the temperature
at which these changes occur and hence caution
should be used if they are to be employed as tem-
perature indicators. They also note the value of ther-
mogravimetry in clarifying certain of the observed
changes. Their paper was in honour of the 70th birth-
day of Prof. Fritz Feigl, which gives an added rele-
vance here.
Dollimore and his co-workers have made vital
contributions to the subsequent work using thermo-
analytical and other techniques on a very wide range
of materials through their papers on oxalates [3±8],
formates [9], sintering and surface studies [10,11] and
a wide range of other topics. At the conclusion of his
Battersea paper [1] he comments that ``to derive the
In September 1964, at a Symposium on Thermal
Analysis held at Battersea College of Technology, Dr.
David Dollimore gave a lecture entitled ``The use of
TGA in studying the Thermal Decomposition of
Oxalates'' [1]. The methods and investigations des-
cribed at this symposium encouraged many workers to
try to apply them in other ®elds of chemistry and
materials science. Both the instrumentation and the
applications of thermal analysis have multiplied in the
intervening years, but the fascination of experiment-
ing with thermal methods on an industrial or labora-
tory sample remains.
The changes in colour and texture which may occur
in inorganic substances when heated have been inves-
tigated by many workers, and an especially important
paper by Harmelin and Duval [2] published in 1962
describes the results for 32 inorganic substances
0040-6031/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 9 9 ) 0 0 2 7 4 - 9

286
P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
maximum information from thermal decomposition
studies, all possible methods must be used''. This is
still as true today, but frequently some methods such as
thermomicroscopy remain under-used.
aluminium sample pans. The heating rate was gen-
erally 16 K/min and the atmosphere static air. Since
the maximum temperature which could be reached on
this instrument was 5008C, some changes did not
occur to completion, but direct observation indicated
the probable course of reaction. The primary objective
of this work did not include the quantitative determi-
nation of the enthalpies of the various changes
observed. The sensitivity of the DSC corresponded
to a full-scale de¯ection of around 100 mW.
The experimental curves from the 3-pen recorder
were scanned into a computer and digitised into a
spreadsheet. Since open pans needed to be used for
sample observation and measurement, the baseline of
the DSC trace tended to be curved. While this was
certainly due to reaction in some cases, it was possible
to correct for the curvature of a ``no-sample'' baseline
using a simple computer program on the digitised
DSC data. The offset between the DSC and RLI traces
was also corrected.
It has been shown that by combining DSC with
observations using a microscope and recording the
light intensity re¯ected from the sample re¯ected light
intensity (RLI), signi®cantly more information may be
obtained [12]. Other instruments study DSC and the
light transmitted through a transparent sample [13] to
give similar observations. The advantage of using
re¯ected light becomes apparent when the sample is
opaque. The DSC±RLI method has been used to study
phase changes of both organic and inorganic materi-
als, the decomposition of copper sulphate pentahy-
drate and manganese formate dihydrate [12], tin
formate [14] as well as polymers [15], foods [16]
and ®re-retarded materials.
While the full and formal investigation of decom-
positions requires a large amount of time, sophisti-
cated equipment and the study of many relevant earlier
studies, a simple quantitative investigation of some
thermal decompositions may be a valuable educa-
tional exercise [17]. The direct observation of mass
loss and of colour change will give both quantitative
and qualitative analytical information on the structure
and decomposition mechanism of inorganic com-
pounds. Although direct observations of optical
changes have often been noted, there are advantages
to recording the optical and energetic events simulta-
neously.
3. Results and discussion
3.1. Ammonium nitrate NH4NO3
The sample was a laboratory-grade material. In
a previous paper, the DSC±RLI results for potassium
nitrate were reported [12] and for that sample the
small alteration in the intensity which occurs at the
transition at 1288C was enhanced when a pre-melted
sample was run. A sample of ammonium nitrate
recrystallised in the pan from just above the melting
point gave very clear changes in RLI exactly corre-
sponding to the endothermic DSC peaks as shown in
Fig. 1.
The ®rst transition, at around 418C with a small
increase in RLI corresponds to the IV ! III phase
change. The second transition at 838C is the III ! II
phase change and shows a small decrease in RLI. The
third change gives a very large DSC peak at 1298C, but
on the RLI trace it is indicated by a change of slope.
This is the II ! I change and is followed by the major
event of the melting of phase I at 1708C. Factors
affecting these transitions have been discussed by
Dellien [18]. The possible causes for alterations in
re¯ectance have been considered [12] and the change
in the molar refraction, RL, is one of the reasons. This
2. Experimental
The apparatus used in these investigations has
already been described [12]. It comprises a Perkin-
Elmer DSC 1-B ®tted with a Beck stereo microscope.
A light±tight cover replaced the normal DSC cell
cover. The RLI detector was a CdS photoconductive
cell, and the ampli®ed RLI signal was recorded on the
same chart recorder as the DSC temperature and heat-
¯ow signals. The DSC was calibrated in accordance
with the manufacturers instructions with high-purity
metals. Con®rmatory TG runs were performed under
similar conditions using a Perkin-Elmer TGA 7 sys-
tem.
Samples as originally received were ®ne powders
and in each case about 15 mg was placed in open

P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
287
Fig. 1. DSC±RLI of recrystallised ammonium nitrate.
Fig. 2. DSC±RLI of ammonium metavanadate.
quantity is a composite of the molar volume and the
refractive index:
much lower temperature of about 1008C turning gra-
dually through light lemon-yellow to orange. From
1808C to 2808C, the powder darkened ®rst to deep
orange, and then to dark brown. During the second
endotherm, the colour lightened to a strong brick-red
by 4508C. This is shown in Fig. 2.
RL ˆ Vmfn2 1g=fn2 ‡ 2g:
Szabo et al. [19] report that there is an expansion at
the ®rst transition and assuming that the refractive
index n of the phases is similar, this would correspond
to an increase in the molar volume, Vm and hence an
increase in RL. The contraction at the second transition
reverses this. More complex behaviour occurs at the
third transition, while the melting produces a clear
¯uid and allows a very large increase in re¯ection from
the base of the aluminium pan. It is evident that quite
small changes in morphology may be observed and
complement the measurement of heat changes by the
DSC even if no change in the ``colour'' of the sample
accompanies the transition.
Although it has been suggested that the initial
yellowing is due to the presence of potassium vanadate
impurity [19] analysis failed to detect any potassium
in this sample. The two-stage endotherm around
2308C has been attributed to the production of
HVO3 [2] with a 14.5% loss or [(NH4)2O(V2O5)2 ]
(11.1% loss) and then [(NH4)2O(V2O5)3] (14.8% loss)
[20] or several other possible intermediates. Since the
mass losses by TG in the early part of the ®rst stage
total 11% and the loss in the complete stage is 14.8%,
the ammonium oxyvanadates seem more likely and
this is con®rmed by analysis [19,21]. The colour
changes also suggest this since the oxyvanadates
are described as being ``golden yellow, ruby red
and garnet red'' [22]. The second stage is the decom-
position to vanadium pentoxide, V2O5, a brick red
powder. The total loss amounts to 22.3% although it
has been suggested that some reduction by ammonia
3.2. Ammonium metavanadate NH4VO3
This sample was an ``AnalaR'' grade white powder.
The DSC trace shows two broad endothermic peaks at
around 2178C and 3778C with indications of possible
overlapping reactions. The RLI starts to decrease at a

288
P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
to V2O4 may occur [23]. It is of interest that a very
de®nite colour change at low temperatures is detected
by RLI before any loss on TG or endotherm on DSC.
This suggests that there is minimal bulk reaction and
that the colour is con®ned to the surface of the powder.
Recent work investigating ammonium metavanadate
supported on alumina has con®rmed this behaviour
[24]. The many possible intermediates [22] have very
similar colours and mass losses, so that distinguishing
between them is complicated.
3.3. Basic copper carbonate CuCO3ÁCu(OH)2
Dollimore and Taylor [25] have published a com-
parative thermal analysis study of the decomposition
of copper carbonate. In an investigation of various
copper carbonates, the combined DSC±RLI technique
has given useful additional information. The green
mineral, malachite has the formula CuCO3ÁCu(OH)2
while azurite is 2CuCO3ÁCu(OH)2 and deep blue.
Other basic carbonates of copper have also been
reported. The sample used here was prepared by
mixing solutions of copper (II) sulphate and sodium
bicarbonate, ®ltering, washing and drying the preci-
pitate which was dark green in colour. It was shown by
X-ray powder diffraction to have the malachite struc-
ture, although the broadened diffraction peaks sug-
gested lower crystallinity than the mineral sample.
The DSC and RLI curves are shown in Fig. 3. There is
a small, broad DSC endotherm below 1008C, which
corresponds to a mass loss of 2±3%. The RLI started
to decrease at low temperature and continued to do so
at an increasing rate as the temperature rose. The
major endothermic decomposition occurred around
2708C after which the sample was dark brown or
black. Above 3308C the RLI increased sharply and
observations indicated that this was due to contraction
of the sample, exposing the re¯ective base of the
sample pan.
Fig. 3. DSC±RLI of basic copper carbonate.
®ndings of Henmi et al. [26,27]. The ®nal shrinkage
appears to be due to sintering of the very ®ne particles.
3.4. Complex oxalates
The study of the thermal behaviour of simple and
complex oxalates using thermal analysis and comple-
mentary methods has been greatly enhanced by the
many papers of Dollimore and his co-workers. The
effects of catalysis, atomic number, bond length and
thermodynamic parameters and the multiplicity of
techniques are all excellent examples of the power
of thermal analysis [6]. The complex oxalates used
here were prepared by methods reported in the litera-
ture [17,28,31].
The ®nal stages of the decomposition were not
complete until about 3508C. These results generally
agree with other work [25±27]. The small peak at low
temperature is probably due to retained moisture.
Since the development of the brown colour starts at
low temperature also, this seems to form a part of the
total structure. The single peak for DSC and for the
complementary TG run, suggests that the loss of CO2
and H2O occurs at the same time which agrees with the
3.4.1. Potassium trioxalato aluminate (III) trihydrate
K3Al(C2O4)3Á3H2O
The thermal analysis of this sample presents a more
complicated situation than might be expected. Firstly,
the water content can be variable and even ``dry''
samples are reported as having 3 [17] or 2.8 [28] moles
of water. Secondly, hydroxo-oxalato aluminates may
be present as a contaminant [28]. The thermal analysis

P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
289
was signalled by a broad, shallow exotherm peaking
around 4008C and followed by a strong exothermic
de¯ection above 4508C. The RLI shows a small
increase, as would be expected if the highly-re¯ective
alumina is formed.
The changes observed and projected are as identi-
®ed by Dollimore and co-workers [30].
K3AlꢀC2O4†3 Á nH2O ˆ K3AlꢀC2O4† ‡ nH2O
3
2K3AlꢀC2O4† ˆ 3K2C2O4 ‡ Al2O3
3
‡ 3CO ‡ 3CO2
K2C2O4 ˆ K2CO3 ‡ CO
These steps may be followed at higher temperatures
by reaction to form the aluminate:
K2CO3 ‡ Al2O3 ˆ K2Al2O4 ‡ CO2
and also the formation of K2O. This accounts for the
further mass losses up to 13008C
3.4.2. Potassium trioxalato chromate (III) trihydrate
K3Cr(C2O4)3Á3H2O
Fig. 4. DSC±RLI of potassium trioxalato aluminate (III) trihydrate.
Wendlandt et al. [31] performed TG and DTA runs
on this dark green compound and found that it lost
water around 1008C and decomposed in several stages
between 3008C and 4508C and that the ®nal residue at
around 6008C in air was the yellow potassium chro-
mate K2CrO4. House and Blumthal [32] detected ®ve
steps in the decomposition of this compound, Horvath
and Kristof [33] report a rather simpler behaviour. The
DSC±RLI curves for the potassium oxalato chromate
(III) trihydrate and shown in Fig. 5. The endothermic
dehydration gives several shallow peaks up to about
2008C and is accompanied by a decrease in RLI . The
original greeny-blue crystals become less crystalline
and darker. Between 1508C and 3008C, the RLI con-
tinues to decrease and a small exothermic peak at
about 3008C, also reported in [33], coincides with a
lessening of the rate of decrease. A strong exothermic
trend commences around 3508C and continues to the
upper temperature limit of 5008C and the RLI
decreases until 4508C and then gradually increases.
In order to achieve the highest conversion, the sample
was held isothermally at 5008C for about 30 min,
when the ®nal colour was bright yellow and the
product was fully soluble in water. Classical wet-
analysis tests con®rmed the presence of carbonate
and chromate.
curves shown in Fig. 4 present a rather incomplete
picture of the decomposition. The sample used in this
investigation was analysed by ICP and found to con-
tain 25.8% K (theory: 25.3%) and 5.5% Al (theory:
5.8%). It also appeared to pick up atmospheric moist-
ure easily and contained slightly more than 3H2O. The
DSC shows several dehydration endotherms between
room temperature and 3008C, of which the largest
occurs around 1008C. This is con®rmed by TG and is
similar to the reported behaviour [28±30] of the loss of
adherent water followed by a series of poorly-resolved
dehydration steps at increasing temperature. The RLI
changes little until the later part of the dehydration
when it decreases, indicating that the largest structural
change occurs with the elimination of the ®nal H2O.
The sharp, exothermic phase change reported by
Verdonk [28] around 2508C was not observed on
the DSC, although a broad change in slope was noted
and very slight changes in the RLI did occur around
this temperature. The decomposition of the oxalatoa-
luminate structure starts around 4008C and overlaps
the oxalate decomposition which continues until
above 5008C, that is, above the range of this instru-
ment. However, the commencement of these changes

290
P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
Fig. 5. DSC±RLI of potassium trioxalato chromate (III) trihydrate.
Fig. 6. DSC±RLI of potassium trioxalato ferrate (III) trihydrate.
The dehydration occurs over a fairly wide tempera-
ture range and involves at least two overlapping steps.
The major RLI change coincides with the ®rst stage.
The anhydrous complex oxalate starts to decompose,
possibly accompanied by a phase change, around
3008C, and the decompositions of the oxalates occur
over the range 300±5008C and are followed in the later
stages and at higher temperatures by the reaction in air
to form the chromate:
anhydrous material started to decompose around
2808C and changed through orange to dark red by
3608C. This matched the large exothermic peak with a
maximum at 3508C. The exotherm peaking at 4008C
corresponded to a very small increase in the RLI the
sample appearing to be more matt. The change to a
dark red during the ®rst exotherm suggests that this is
corresponds to the decomposition of the complexed
iron to its oxide, whereas the higher temperature
exotherm would be due to the potassium oxalate
decomposition. This agrees exactly with the ®ndings
of Dollimore and co-workers [4,5] and complements
their ®ndings with simultaneous observation and
recording of the colour changes. The small exotherm
at 2608C noted by Horvath and Kristof [33] was not
observed with this sample.
K3CrꢀC2O4†3 Á 3H2O ˆ K3CrꢀC2O4† ‡ 3H2O
3
K3CrꢀC2O4† ˆ 3=2K2CO3 ‡ 1=2Cr2O3
3
‡ 3CO ‡ 3=2CO2
K2CO3 ‡ 1=2Cr2O3 ‡ 3=4O2 ˆ K2CrO4 ‡ CO2
3.4.3. Potassium trioxalato ferrate (III) trihydrate
K3Cr(C2O4)3Á3H2O
3.4.4. Potassium dioxalato cuprate (II) dihydrate
K2Cu(C2O4)2Á2H2O
This pale-green complex is light sensitive and was
stored in a blackened container , but no colour change
was observed when a sample was illuminated in the
apparatus for 15 min at room temperature (Fig. 6). On
heating, dehydration occurred almost immediately to
produce light yellow crystals which became opaque at
just above 1008C. These reactions corresponded to
two strong endothermic peaks on the DSC trace. The
This blue, crystalline compound gives the thermal
analysis curves shown in Fig. 7. The dehydration gives
a single endotherm peaking at about 1208C and
accompanied by a colour change from blue crystals
to an opaque light-blue mass, resulting in an increase
in RLI. The mass loss for this stage (10.1%) con®rms
the presence of two molecules of water of crystal-

P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
291
single DTA exotherm, the two stages of the decom-
position, of the complex and of the potassium oxalate
were well separated on the DSC and the complemen-
tary TG runs, provided small samples were used.
However, since the decompositions are highly ex-
othermic larger samples cause them to overlap con-
siderably.
4. Conclusion
The advice to employ every technique to separate
and identify the changes which occur on heating has
been heeded here. Direct observation of the changes,
their physical or chemical nature and the alteration of
colour provides a useful additional method of con-
®rming thermal events in a range of inorganic com-
pounds.
The study of thermochromic changes has several
potential applications. While some changes occur over
a suf®ciently narrow temperature range to act as
temperature sensors or calibration materials, others
may be useful as thermal indicators. Work on thermo-
chromic changes in solution [35] and in solids [36±38]
has demonstrated some of these applications and
coating such temperature-sensitive materials onto
®bre±optic probes may offer advantages in tempera-
ture control.
Fig. 7. DSC±RLI of potassium dioxalato cuprate (II) dihydrate.
lisation. The anhydrous complex oxalate is stable up to
about 2708C when decomposition occurs in two
exothermic stages. During the ®rst stage up to
3208C, the sample turns black with a large decrease
in RLI, whereas in the second, from 3208C to around
4008C, little change in the intensity trace occurs. This
strongly suggests that the ®rst exotherm is due to the
decomposition of the complex to copper (II) oxide,
while the second stage corresponds to the loss of CO
and the formation of potassium carbonate. The total
mass loss of 38.9% agrees with the reactions:
Other inorganic materials which have been studied
include additives which impart ®re-retardancy in poly-
mer systems, organometallic catalysts and antiperspir-
ants.
Acknowledgements
The collaboration of the School of Applied Chem-
istry at Kingston University is most gratefully
acknowledged. Dr. Alan Vincent kindly provided
several samples and con®rmatory analyses and Dr.
Simon de Mars carried out TG runs and ICP analysis.
Their help was most valuable.
K2CuꢀC2O4†2 Á 2H2O ˆ K2CuꢀC2O4† ‡ 2H2O;
2
loss 10:2% ꢀcalc:†
K2CuꢀC2O4† ˆ K2C2O4 ‡ CuO ‡ CO ‡ CO2;
2
loss 20:4% ꢀcalc:†
Much of this work would not have been undertaken
without the earlier, painstaking and well-planned
investigations of Prof. David Dollimore, and without
his encouragement through lectures, papers and con-
versations. In congratulating him on reaching a most
signi®cant birthday, it is a great pleasure to pay tribute
to his untiring and stimulating help.
K2C2O4 ˆ K2CO3 ‡ CO; loss 7:9% ꢀcalc:†
In the presence of air, the CO is oxidised to CO2 and
thus little reduction of the copper oxide occurs. No
evidence of metallic copper was found in these experi-
ments. Although Dollimore and co-workers [34]
found the decomposition reactions in air to yield a

292
P.J. Haines / Thermochimica Acta 340±341 (1999) 285±292
[21] J. Lamure, G. Colin, Compt. Rend. 258 (1964) 6433.
References
[22] J.W. Mellor, Comprehensive Treatise on Inorganic Chemistry,
vol. IX, Longman, London, 1935, p. 738.
[1] D. Dollimore, Symposium on Thermal Analysis, Battersea,
London, 1964.
[23] T. Sata, E. Komada, Y. Ito, Kogyo Kagaku Zasshi 71 (1968)
643.
[2] M. Harmelin, C. Duval, Mikrochim. Acta 1 (1962) 275.
[3] D. Dollimore, D.L. Griffiths, D. Nicholson, J. Chem. Soc.
(1963) 2617.
[24] G.A. El-Shobaky, K.A. El-Barawy, F.H.A. Abdalla, Thermo-
chim. Acta 96 (1985) 129.
[25] D. Dollimore, T.J. Taylor, Thermal Analysis, Proceedings of
the Sevent ICTA, Wiley±Heyden, New York, 1982, p. 636.
[26] H. Henmi, T. Hirayama, S. Shanmugarajah, N. Mitzutani, M.
Kato, Thermochim. Acta 96 (1985) 145.
[4] D. Broadbent, D. Dollimore, J. Dollimore, J. Chem. Soc. A
(1966) 278.
[5] D. Broadbent, D. Dollimore, J. Dollimore, J. Chem. Soc. A
(1967) 451.
[27] H. Henmi, T. Hirayama, S. Shanmugarajah, N. Mitzutani, M.
Kato, Thermochim. Acta 106 (1986) 263.
[6] D. Dollimore, Thermochim. Acta 117 (1987) 331.
[7] D. Dollimore, T.A. Evans, Thermochim. Acta 178 (1991)
263.
[28] A.H. Verdonk, Thermochim. Acta 4 (1972) 85.
[29] K.A. Foster, J.E. House, Thermochim. Acta 60 (1983) 389.
[30] D. Broadbent, D. Dollimore, J. Dollimore, The Analyst 94
(1969) 543.
[8] D. Dollimore, T.A. Evans, Thermochim. Acta 179 (1991)
149.
[9] D. Dollimore, J.A. Gupta, D.V. Nowell, Proceedings of the
First ESTA, Heyden, London, 1976, p. 233.
[10] D. Dollimore, G.R. Heal, R.Sh. Mikhail, N.R. El Nazer,
Proceedings of the First ESTA, Heyden, London, 1976, p. 51.
[11] K.A. Broadbent, J. Dollimore, D. Dollimore, Thermochim.
Acta 133 (1988) 131.
[31] W.W. Wendlandt, T.D. George, K.V. Krishnamurty, J. Inorg.
Nucl. Chem. 21 (1961) 69.
[32] J.E. House, T.G. Blumthal, Thermochim. Acta 36 (1980)
79.
[33] A. Horvath, J. Kristof, J. Thermal Anal. 36 (1990) 1471.
[34] D. Broadbent, D. Dollimore, J. Dollimore, in: R.F. Schwen-
ker, P.D. Garn (Eds.), Thermal Analysis, Proceedings of the
Second ICTA, Academic Press, New York, 1969, p. 739.
[35] D. Lavabre, J.C. Micheau, G. Levy, J. Chem. Educ. 65 (1988)
274.
[12] P.J. Haines, G.A. Skinner, Thermochim. Acta 59 (1982) 343.
[13] H.G. Wiedemann, J. Thermal Anal. 40 (1993) 1031.
[14] J. Fenerty, P.G. Humphries, J. Pearce, Thermochim. Acta 61
(1983) 319.
[15] P.J. Haines, G.A. Skinner, Thermochim. Acta 134 (1988) 201.
[16] P.J. Haines, TAC'97, Oxford, Unpublished work, 1997.
[17] A. Vincent, Educ. Chem. 18 (1981) 38.
[18] I. Dellien, Thermochim. Acta 55 (1982) 181.
[19] Z.G. Szabo, I.K. Thege, E.E. Zapp, Proceedings of the First
ESTA, Heyden, London, 1976, p. 268.
[36] F.A. El-Saied, A.M. Donia, S.M. Hamza, Thermochim. Acta
165 (1990) 183.
[37] C. Paria, A. Ghosh, N.R. Chaudhuri, Thermochim. Acta 268
(1995) 153.
[38] C. Paria, A. Ghosh, N.R. Chaudhuri, Thermochim. Acta 273
(1996) 185.
[20] M. Taniguchi, T.R. Ingraham, Can. J. Chem. 42 (1964) 2467.