The use of Hi-Res TGA, TG-FTIR, HT-DRIFT and HT-XRD in the study of the decomposition of La2(C2O4)3·10H2O

Article abstract of DOI:10.1016/S0040-6031(00)00469-X

Using a combination of various techniques, more intermediate products could be detected during the thermal decomposition of La₂(C₂O₄)₃·10H₂O than from conventional TGA measurements. In this way a more complete and detailed decomposition mechanism is proposed. Direct identification of the intermediate phases with HT-DRIFT revealed the presence of another oxycarbonate La₂O(CO₃)₂, while with Hi-Res TGA a thermally unstable intermediate La₂(CO₃)₃ was detected. All this information is consistent with the data obtained from TG-FTIR spectroscopic measurements. (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0040-6031(00)00469-X

Thermochimica Acta 354 (2000) 145±151
The use of Hi-Res TGA, TG-FTIR, HT-DRIFT and HT-XRD in the
study of the decomposition of La2(C2O4)3Á10H2O
G. Vanhoyland, R. Nouwen, M.K. Van Bael, J. Yperman, J. Mullens*, L.C. Van Poucke
Laboratory of Inorganic and Physical Chemistry, IMO, Limburgs Universitair Centrum, B-3590 Diepenbeek, Belgium
Received 20 September 1999; accepted 16 February 2000
Abstract
Using a combination of various techniques, more intermediate products could be detected during the thermal decomposition of
La2(C2O4)3Á10H2O than from conventional TGA measurements. In this way a more complete and detailed decomposition
mechanism is proposed. Direct identi®cation of the intermediate phases with HT-DRIFT revealed the presence of another
oxycarbonateLa2O(CO3)2, whilewithHi-ResTGA athermallyunstable intermediateLa2(CO3)3 was detected. Allthisinformation
is consistent with the data obtained from TG-FTIR spectroscopic measurements. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Lanthanum oxalate; Thermal decomposition; HT-DRIFT; Hi-Res TGA; HT-XRD; TG-FTIR
1. Introduction
temperature plots, this indicates that overlapping
decomposition reactions are involved. In order to
resolve these reactions as much as possible a high-
resolution variant of the normal TG was used.
La2(C2O4)3Á10H2O was synthesised in a study to
obtain a homogeneous LSCO (C=Co) precursor by the
oxalate co-precipitation method. Although the decom-
position of the oxalate has already been studied by
conventional thermal analysis techniques [1±3], TG
analysis suggested that more intermediate products
were involved besides those reported La2(C2O4)3,
La2O2(CO3) and La2O3. In this study, the TGA appa-
ratus was coupled to FTIR spectrometer in order to
identify the evolved gases. However, in some cases it
was favourable for also the intermediate solids to be
identi®ed directly. This was achieved by HT-DRIFT
and HT-XRD. Moreover, since no well-de®ned
plateaux were observed in the TG weight loss±
2. Experimental
2.1. Methods and apparatus
Thermogravimetric measurements (in dry air and
argon) were performed with TGA Model 951-2000
and Hi-Res TGA 2950, both from TA Instruments.
The ®rst unit was on-line coupled to a Bruker FTIR
IFS 48 spectrometer in order to identify the evolved
gases [4]. In all coupled experiments, a resolution of
1
8 cm was selected for the MCT detector. The inert
working conditions of all the TGA equipment were
checked using copper oxalate as a standard [5]. DTA
measurements were carried out by heating at
* Corresponding author. Tel.: ‡32-11-26-83-08;
fax: ‡32-11-26-83-01.
E-mail address: jules.mullens@luc.ac.be (J. Mullens)
0040-6031/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 0 0 ) 0 0 4 6 9 - X

146
G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
1
108C min in dry air using a DTA 1600-2000 (TA
Instruments).
temperature. In the region of interest (290±6808C)
every 108C a FTIR spectrum was recorded. At lower
temperatures, the stepsize was 508C.
X-ray diffraction experiments were performed
using a Siemens D-5000 with 2y values from 8 to 708.
The con®guration for the high temperature experi-
È
3. Preparation of lanthanum oxalate [6]
ments consisted of a Gobel mirror (Ka1‡2, Huber),
a high temperature device with a Pt rod (Anton Paar,
HTK 10) and a position sensitive detector (Braun).
The Pt rod also contribute to the diffraction pattern
(peaks at 39.76, 46.24 and 67.458 2y) and could,
therefore, be used as an internal standard in order to
determine shifts due to temperature differences. Each
measurement took about 5 min but was preceded by a
1 min delay. In the range from 320 to 6508C, a
diffraction pattern was measured at 58C intervals from
8 to 708 2y (0.01848 2y stepsize). Outside this range,
temperatures were selected from the TG weight loss
responce. For indexing purposes, a Ge-monochroma-
tor (Ka1), standard sample holder and scintillation
detector were used.
A 0.1 M aqueous solution of La(NO3)3Á6H2O
(Acros) was mixed with a 1.0 M aqueous solution
of H2C2O4Á2H2O (Merck). The water contents of both
starting materials were checked in advance with TGA.
The mixing of both solutions was carried out in a
computer-controlled way with two motor-driven bur-
È
3
ettes (Schott Gerate T100). In each step, 0.2 cm of
each solution was added with continuous stirring, up
to a ®nal volume of 40 cm3. The white precipitate was
®ltered through a 0.45 mm Millipore ®lter and washed
with de-ionised water. Finally, the powder was dried
in air.
The quality of the powder was checked by X-ray
diffraction. The observed peak positions were
corrected by applying a correction curve for SRM
675 (NIST), used as an external standard. Since
the diffraction pattern could be indexed with
Treor97 [7] and Dicvol91 [8] as a monoclinic unit
HT-DRIFT measurements were performed on a
Bruker IFS 66 spectrometer equipped with a high
temperature Ð high pressure chamber (Spectratech)
with parabolic ZnSe windows. A DTGS detector was
1
used, for which a resolution of 4 cm was selected.
Ê
Ê
Ê
cell (aˆ11.351 A, bˆ9.589 A, cˆ10.454 A and
bˆ114.4988) with reasonable accuracy (M20ˆ22 and
F20ˆ49 h0.012, 35i), the powder was considered to be
as a single phase.
About 2% of La2(C2O4)3Á10H2O was mixed with KBr.
This sample was continuously ¯ushed with N2. In
accordance with the X-ray diffraction experiments,
there was a 1 min delay after arriving at the desired
Fig. 1. DTA, normal and Hi-Res TGA of La2(C2O4)3Á10H2O in dry air.

G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
147
Table 1
Comparison between calculated and observed weights from the Hi-Res TGA curve in dry aira
Phase
Temperature (8C)
Observed weight (%)
Calculated weight (%)
La2(C2O4)3Á3H2O
La2(C2O4)3
La2(CO3)3
72
347
390
556
601
82.1
74.4
62.6
50.7
44.8
82.0
74.5
63.0
50.9
44.8
La2O2(CO3)
La2O3
a All weights were calculated supposing that the last plateau corresponds with pure La2O3.
4. Results
low temperature region, at least one different hydrated
oxalate can be identi®ed from a DTG analysis. Only at
the temperature of about 3508C was the oxalate
completely water free. From the observed decomposi-
tion curves in dry air and argon, it was apparent that
the decomposition mechanism was not sensitive to the
atmosphere. The rate of weight loss±temperature have
the same shape, except for the fact that in argon the
®nal decomposition reaction occurs at a somewhat
elevated temperature. From DTA, it is clear that the
decomposition of the `dry' oxalate is highly exother-
mic. The consequence of this will be discussed later.
4.1. TGA and DTA
1
In the normal mode, heating rate at 108C min
,
only the dehydration of the oxalate and the ®nal
decomposition from La2O2(CO3) to La2O3 could be
identi®ed from the observed weight losses. The weight
loss in the temperature range from about 300 to 7008C
is clearly the result of overlapping decomposition
reactions. Hi-Res TGA revealed which reactions were
occurring (Fig. 1). The principle used was that when a
reaction started, the heating rate was decrease in
accordance to the mass loss. As a consequence, the
temperatures at which the transitions occurred were
better de®ned. Moreover, the experimental weight
(62.6%) at the sharp curve transition around 3908C
was in good correspondence with the calculated
weight (63.0%) of La2(CO3)3 (Table 1). Also in the
4.2. TG-FTIR
Experiments were performed both in dry air and
argon atmosphere. The evolved gases were followed
as a function of time (or temperature), and in both
cases, four distinct regions could be distinguished
Fig. 2. Evolved CO and CO2 as a function of time.

148
G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
Fig. 3. Detailed IR spectra of: (a) region II; (b) region III; (c) region IV.

G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
149
4.3. HT-XRD
Up to about 608C, there are only some minor
modi®cations in the high angle region. Once this
temperature is exceeded, the crystallinity decreased
very fast and at about 1008C, the material was amor-
phous. No diffraction peaks were observed, but the
background absorption level decreased, especially
at low angles, until a temperature of about 4258C
was reached. At this temperature, the ®rst diffraction
peaks of the La2O2(CO3) appeared and increased
rapidly, but in the temperature range from 470 to
5708C, the intensity was almost constant. There was
a further increase in peak intensity in combination
with a small decrease in background level (Fig. 4).
Finally, at 7008C, the diffraction peaks of La2O3 are
observed.
4.4. HT-DRIFT
From these experiments, it was clear that the
hydrated oxalate lost H2O (3600±3000 cm 1) up to
about 2708C and in accordance with the TGA results,
there was a large decrease in water around 1008C.
Dehydration also changed the oxalate bands. Up to
Fig. 4. Variation of the (1 0 3) diffraction peak of La2O2(CO3): (A)
425±4708C; (B) 470±5708C; (C) 570±6308C.
were gases were detected. In the ®rst region (I), only
water from the dehydration was detected. In all other
regions (II, III, IV), CO2 (2400±2300 cm 1) absor-
bance could be detected. Furthermore, in the ®rst part
of region II, CO (2200±2100 cm 1) was undoubtedly
detected (Figs. 2 and 3).
1
1008C, they were situated at 1617, 1318 and 797 cm
[9] but from 1508C to the decomposition of the
1
oxalate, they shift to 1638, 1307 and 784 cm
,
respectively. Moreover, at 1508C, a small new band
Fig. 5. HT-DRIFT spectra as a function of temperature.

150
G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
Fig. 6. Detailed IR of La2O(CO3)2 (470±5008C) and La2O2(CO3) (530±6008C).
appeared at about 1460 cm 1 which increased slightly
on heating. At 4208C, this band rapidly increased in
absorbance and became broad at the expense of the
oxalate bands, which have completely disappeared by
4808C. At the same time, new carbonate peaks
appeared at 1074 cm (weak) and 843 cm (med-
ium) [10]. Above 5008C, this phase decomposed into
another, which remained stable up to 530±6008C. The
bands of this new phase are at almost in the same
the H2O molecules are present in different ways in the
crystal structure.
The
decomposition
products
La2(C2O4)3,
La2O2(CO3) and La2O3 can be identi®ed by mass
calculation from the TGA and are con®rmed by at
least one other technique. For example, since
La2(C2O4)3 is amorphous, XRD is not suitable but
from the infrared absorption bands, the oxalate groups
are clearly present.
1
1
positions as that of the previous phase before, except
1
As mentioned before, a high resolution TGA reveals
subtle differences in the weight loss±temperature
curves than in normal mode, and it was apparent that
the decomposition of the La2(C2O4)3 and La2O2(CO3)
occurred in a two-step process. The real decomposi-
tion steps are both times preceded by a small decrease
in mass that is observed over a restricted temperature
area. The same effect is also seen with DTA and HT-
DRIFT. Concerning the decomposition of La2(C2O4)3
(Æ3908C), for example, the oxalate band around
for broad band at 1460 cm
which shifts to
1
1379 cm with a clear shoulder at 1460 cm 1. Inte-
grating from 1860 to 1175 cm 1 showed that the peak
area was somewhat less than half that of the previous
peak, suggesting that there are only half as much
2
CO3 groups present compared to the ®rst phase.
2
At 6008C, all peaks due to the CO3 started to
decrease and completely vanished at about 6508C
(Figs. 5 and 6).
1
1638 cm is nearly constant while the stretching
vibration of the carbonate group (1460 cm 1) slightly
increases in the temperature range 350±4108C. Once
this last temperature is exceeded, there is a fast
decrease in intensity of the oxalate bands, while for
the carbonate bands the reverse is true.
Of all the techniques used, only Hi-Res TGA was
capable, up to a certain level, to separate overlapping
reactions. In this way, a thermally unstable intermedi-
5. Discussion
From coupled TG-FTIR and HT-DRIFT spectro-
scopies, it is clear that water is set free over quite a
large temperature range. The highest mass reduction
however, occurred below 1008C, consistent with
7H2O molecules being removed. This suggests that

G. Vanhoyland et al. / Thermochimica Acta 354 (2000) 145±151
151
ate phase such as La2(CO3)3 could be identi®ed. This
6. Conclusions
was de®nitely not possible with HT-DRIFT and HT-
XRD due to the discontinuous measuring regime.
However, the existence of the La2(CO3)3 as an inter-
mediate phase is also suggested by the coupled TG-
FTIR experiment, since in the whole experiment only
once CO is evolved. This means that all the oxalate
groups have to decompose to carbonate groups in one
single step. In the coupled TG-FTIR experiment (con-
tinuous heating rate) CO2 is detected at the same time
due to the decomposition of the carbonate to an
oxycarbonate, since both decomposition reactions
are not enough resolved in time.
From the HT-DRIFT experiments, it's evident that
in fact two different oxycarbonates must exist. First of
all La2O2(CO3), which was also identi®ed from HT-
XRD, in the temperature range 530±6008C. However,
from the curve integration, it follows that in the region
470±5008C, the La2O(CO3)2 is uniquely present. In
between, it's probably a mixture which exists.
These remarks in mind, we can summarise the
whole decomposition mechanism:
The decomposition mechanism of La2(C2O4)3Á10-
H2O is rather complicated by the fact that some
decomposition reactions overlap with each other, so
not all the intermediate products can be identi®ed from
the conventional TGA curve. Hi-Res TGA, however,
has a much better resolving power. Moreover, the
intermediate products can be identi®ed in other ways:
indirectly from the evolved gas analysis (TG-FTIR) or
directly by other techniques as HT-DRIFT and HT-
XRD. As has been showed in this case, combining all
the information obtained from the different techniques
is an ef®cient way in solving the puzzle.
Acknowledgements
G. Vanhoyland is research assistant of the Fund for
Scienti®c Research, Flanders (Belgium) (F.W.O).
M.K. Van Bael is postdoctoral fellow of the
Fund for Scienti®c Research, Flanders (Belgium)
(F.W.O.).
La2ꢀC2O4† Á 10H2O ! La2ꢀC2O4†
3
3 Á 3H2O ‡ 7H2O
(1a)
References
La2ꢀC2O4†3 Á 3H2O ! La2ꢀC2O4† ‡ 3H2O
Â
[1] A. Gorski, A.D. Krasnicka, J. Therm. Anal. 32 (1987) 1229±
3
1241.
[2] W. Ma-Shine, F. Tsang-Tse, Mater. Chem. Phys. 37 (1994)
278±283.
(1b)
La2ꢀC2O4† ! La2ꢀCO3† ‡ 3CO
(2a)
(2b)
(3)
3
3
[3] R.L.N. Sastry, S.R. Yoganarasimhan, P.N. Mehrotra, C.N.R.
Rao, J. Inorg. Nucl. Chem. 28 (1966) 1165.
[4] J. Mullens, in: M. Brown, P. Gallagher (Ed.), Handbook of
Thermal Analysis and Calorimetry, Vol. 1, Elsevier, Am-
sterdam, 1998 p. 509.
La2ꢀCO3† ! La2OꢀCO3† ‡ CO2
3
2
La2OꢀCO3† ! La2O2ꢀCO3† ‡ CO2
2
La2O2ꢀCO3† ! La2O3 ‡ CO2
(4)
[5] J. Mullens, A. Vos, R. Carleer, J. Yperman, L.C. Van Poucke,
Thermochim. Acta 207 (1992) 337.
However, since reactions 2a and 2b are highly
exothermic, due to local heat build-up also reaction
3 can occur at the same time. In fact, this is clearly a
mass effect and depending on the technique used. HT-
XRD samples for example are about 50 times larger
than those for HT-DRIFTand Hi-Res TGA. Moreover,
in the HT-DRIFT set-up the sample is embedded in an
inert material (KBr). This yields a more ef®cient heat
removal compared to HT-XRD. For this reason
La2O2(CO3) is partly formed when the oxalate decom-
poses in the HT-XRD experiment.
[6] A. Vos, R. Carleer, J. Mullens, J. Yperman, L.C. Van Poucke,
Eur. J. Solid State Inorg. Chem. 28 (1991) 657±662.
[7] P.E. Werner, L. Eriksson, M. Westdahl, J. Appl. Cryst. 18
(1985) 367±370, Treor97 is a further evolution of the well-
known Treor90 indexing program.
È
[8] A. Boultif, D. Louer, J. Appl. Cryst. 24 (1991) 987±993.
[9] N.B. Colthup, L.H. Daly, S.E. Wiberley, in: Introduction to
Infrared and Raman Spectroscopy, 3rd Edition, Academic
Press, New York, 1990, p. 318.
[10] N.B. Colthup, L.H. Daly, S.E. Wiberley, in: Introduction to
Infrared and Raman Spectroscopy, 3rd Edition, Academic
Press, New York, 1990, p. 310.