Problems in the thermal investigation of the BaF2-YF3 system

Article abstract of DOI:10.1007/s10973-008-9005-3

The BaF₂-YF₃ system was partially investigated, with focus given to the BaY₂F₈ compound and its neighboring phases. In this report, various difficulties that hinder the thermal analysis investigation of this bin

Full text of DOI:10.1007/s10973-008-9005-3

Journal of Thermal Analysis and Calorimetry, Vol. 95 (2009) 1, 43–48
PROBLEMS IN THE THERMAL INVESTIGATION OF THE
BaF2-YF3 SYSTEM
G. H. G. Nakamura1, S. L. Baldochi1,*, V. L. Mazzocchi1, C. B. R. Parente1, M. E. G. Valério2
and D. Klimm3
1IPEN-CNEN/SP, CP 11049, 05422-970, S±o Paulo, SP, Brazil
2Federal University of Sergipe, Physics Department, Campus Universitário, S±o Cristov±o, Brazil
3Institute for Crystal Growth, Max-Born-Str. 2, 12489 Berlin, Germany
The BaF2–YF3 system was partially investigated, with focus given to the BaY2F8 compound and its neighboring phases. In this re-
port, various difficulties that hinder the thermal analysis investigation of this binary fluoride system are described in detail. Samples
of various compositions ranging from 58 to 79% YF3 were prepared and subjected to thermal analysis (DTA, TG and DSC) and
X-ray diffraction. Diffraction patterns were analyzed through the Rietveld method for the calculation of phase concentrations in
samples and determination of the lattice parameters of monoclinic BaY2F8. Thermal results were compared with data from the
literature and discrepancies were found.
Keywords: BaY2F8, DSC, DTA, fluorides, phase diagrams, TG, XRD
Introduction
solutions of up to 1% excess of YF3. The transition
temperature differs in each side of the compound: it is
928°C on the BaF2 rich side and drops to 920°C for
the YF3 rich solid solutions. An YF3-rich eutectic at
940°C, a BaF2-rich eutectic at 946°C and the congruent
melting of the compound at Tf=960°C are also reported.
In the same report, however, the authors remark about
the difficulty of experimentally splitting the broad
thermal event observed in the melting of Ba(Y,Ln)2F8
compounds, which is attributed to the closeness of the
polymorphic transformation to the melting point of
these compounds. They do not discard, therefore, the
possibility that further studies might reach different
interpretations for the phase equilibria in the region in
which these compounds are formed.
The compound BaY2F8 (BAYF) possesses great potential
for use in solid-state lasers. Crystals of this material
have been the focus of numerous studies regarding its
spectroscopy and laser applications when doped with
rare-earth (RE) elements [1–4]. Nevertheless, the
thermal behavior of this compound has yet to be fully
understood, which significantly hinders the process of
growing BAYF single crystals.
In the 1970s and 1980s, the BaF2–YF3 binary
system was investigated by various authors such as
Ippolitov et al. [5], Tkachenko et al. [6], Sobolev et al. [7]
among others. These works resulted in detailed phase
diagrams for the system and the proposal of
Sobolev et al. [7] is shown in Fig. 1.
Thermal investigation of this system is reportedly
non-trivial. Obstacles include huge supercooling
effects and remarkable vulnerability to oxygen
contamination, which is particularly damaging since
it leads to the formation of rare-earth oxifluorides. In
the composition range surrounding BAYF, interpretation
of thermal analysis results is additionally complicated
by the overlapping of thermal events, all of which
occur at similar temperatures, making the identification
of individual events difficult.
According to Sobolev and Tkachenko [7], the
thermal character of BAYF is marked by a polymorphic
transition from the monoclinic ‘m-phase’ to possibly
an orthorhombic ‘o-phase’ and the formation of solid
Fig. 1 Reproduction of the BaF2–YF3 phase diagram [7]
*
Author for correspondence: baldochi@ipen.br
1388–6150/$20.00
Akadémiai Kiadó, Budapest, Hungary
Springer, Dordrecht, The Netherlands
© 2008 Akadémiai Kiadó, Budapest

NAKAMURA et al.
The disputed existence of the BaYF5 compound
Differential scanning calorimetry (DSC) and
thermogravimetry experiments were simultaneously
performed on samples on a Netzsch STA 449C/Jupiter
system. A standard DSC/TG sample carrier was used.
Sample powders were placed in lidded Pt DSC cruci-
bles. Two sets of experiments were performed; the
first using heating rates of 10 and the second with
5 K min–1; 50 and 10 mg of the samples were employed,
respectively, in each set. Maximum temperature was
around 1100–1150°C. Prior to each experiment the
furnace was evacuated to 6·10–2 Pa. A first heating
run above the melting point was always performed for
sample homogenization and then followed by a second
consecutive heating run using the same leftover mate-
rial. The first and second runs were always identical
in terms of maximum temperature and thermal rates.
Other consecutive runs were performed for samples
with the (1:2) composition (related to the stoichio-
metric BAYF phase), using 2 and 1 K min–1 heating
rates. All runs were performed under a constant flow
of argon (99.999% purity). The DSC apparatus was
previously calibrated for temperature and sensitivity
at the melting points of Zn, Au, Ni and at the phase
transformation point of BaCO3.
is another feature of the BaF2–YF3 system that is not
completely conclusive. Though none of the studies
regarding phase equilibria describe this compound,
growth of BaYF5 crystals has been reported in the lit-
erature [8, 9].
With our present work we hope to clarify some
discrepancies about this system, as well as contribute
to the study of fluoride compounds through thermal
analysis methods. This first report stands as a preliminary
step in that direction. The BaF2–YF3 system was par-
tially investigated, with focus given to the BaY2F8
compound and its neighboring phases; we relate the
difficulties that arose during our thermal investigation
of this system.
Experimental
Several samples were prepared for this study. Compo-
sitions vary within the range of 58–79 mol% YF3, which
includes the stoichiometric composition for BAYF
and the two eutectics (60% YF3 and 77% YF3) de-
scribed by the phase diagram from Fig. 1.
The starting fluoride materials for the samples
were obtained through fluorination under HF flow of
high-purity BaCO3 (>99.99%) and Y2O3 (>99.99%),
both obtained commercially. This procedure is de-
scribed in detail in [10]. Samples were initially pre-
pared by weighing out appropriate quantities of YF3
and BaF2 and mixing them mechanically. In order to
enhance homogeneity, the samples were individually
melted (heated to the temperature of 985°C) in Pt
boats under a constant flow of HF and the obtained in-
gots were subsequently pulverized in a mortar. Re-
sulting powders were used for thermal analysis and
X-ray diffraction (XRD) experiments.
Results and discussion
X-ray diffraction
Calculated phase concentrations of several samples
are presented in Table 1. Observed concentrations
show some deviation; for example, according to the
lever principle [15], the sample with starting compo-
sition 22% BaF2:78% YF3 should contain approxi-
mately 66% of BAYF and 34% of YF3, but a greater
quantity of the latter compound is observed. Nonethe-
less, the observed phases in each sample mostly agree
with what should be expected after the synthesis pro-
cess (which involves its melting and subsequent so-
lidification), in accordance with the phase diagram
(Fig. 1). The sample with the (1:2) composition, how-
ever, should be composed solely by the BaY2F8
phase, though a measurable quantity of YF3 was de-
tected. It should be noted that three different samples
were synthesized from a mixture of 33.33% molar
fraction of BaF2 plus a molar fraction of 66.66% YF3
(1:2 composition) and all presented a small amount of
YF3 in the X-ray diffraction data. Only crystalline
samples obtained from crystal growth processes pre-
sented pure BaY2F8 in the X-ray diffraction pattern.
Two possible explanations may be either the exis-
tence of a non-stoichiometric range of BaY2F8 that
would accept BaF2, resulting in an YF3 excess in the
synthesized samples, or that the excess is the result of
incomplete mixing during the synthesis process.
XRD measurements were performed at room
temperature in a Rigaku DMAX 2000/PC diffractometer.
The diffraction patterns were taken in the 2q range
from 14 to 80° in step scan mode, with steps of 0.02°,
2 s per point, using CuKa radiation. Resulting patterns
were used for quantitative calculation of phase
concentrations through the Rietveld method; the analyses
were performed using program DBWS-9807a [11].
Thermogravimetry (TG) and differential thermal
analysis (DTA) measurements were simultaneously
performed on each sample using a TA Instru-
ments SDT 2960, under constant helium flow
(99.999% purity, 100 cc min–1). Unlidded Pt–Au cru-
cibles containing approximately 50 mg of the samples
were used. Two sets of experiments were performed
using heating and cooling rates of 10 and 40 K min–1
up to the temperature of 1200°C. The DTA apparatus
was previously calibrated for temperature at the melt-
ing points of LiF, NaI, NaF, MgF2 and KCl.
44
J. Therm. Anal. Cal., 95, 2009

THERMAL INVESTIGATION OF THE BaF2-YF3 SYSTEM
Table 1 Phase concentrations calculated through the Rietveld method
BaY2F8 [12]/
mol%
YF3 [13]/
mol%
Ba4±xY3±xF17±x [14]/
mol%
Sample YF3
RP
RWP
Re
S
78
36.5
41.2
81.2
87.6
98.3
83.0
82.0
70.6
63.5
58.8
18.8
12.4
9.11
8.72
12.73
11.38
10.58
12.42
18.77
9.85
5.21
4.93
5.24
5.38
9.29
5.68
5.87
6.27
2.44
2.30
2.02
2.30
2.02
1.73
1.98
2.41
77
67.7
66.7
65.7
61
8.03
9.65
1.7
17.0
18.0
29.4
14.43
7.38
60
8.59
11.65
15.15
11.40
59
RP=‘R-pattern’, RWP=‘R-weighed pattern’, Re=‘R-expected’, S=‘Goodness-of-fit’ [11]
Table 2 BaY2F8 lattice parameters
Sample
a/
b/
c/
b/°
67.7% YF3
66.7% YF3
65.7% YF3
[15]
6.97257(13)
6.97347(21)
6.97334(22)
6.9781(5)
10.50262(19)
10.50479(32)
10.50453(34)
10.5099(7)
4.257180(79)
4.25780(13)
4.25779(14)
4.2603(3)
99.6696(4)
99.6699(5)
99.6719(6)
99.682(8)
99.678(8)
[16]
6.9829(5)
10.519(1)
4.2644(4)
Thermal analysis
DTA results obtained in the first set of experiments,
performed with 10 K min–1 rates, were found to be un-
reliable; even though thermal events were observed
during heating, upon cooling none of the curves pre-
sented any peaks whatsoever. An example of this can
be seen in Fig. 3.
Experiments performed with the higher heating/
cooling rate (40 K min–1) produced DTA curves which
did present exothermic peaks upon cooling, as
expected for this material. The remarkably different results
were a serious indication that, during the experiments
Fig. 2 Diffraction pattern and its corresponding fitted curve;
sample 66.7% YF3
One of the diffraction patterns obtained is presented
in Fig. 2, along with the Rietveld fitted theoretical curve
and the difference pattern. It should be noted that the
compound Ba4±xY3±xF17±x was detected in the samples
with excess BaF2, whereas BaYF5 was detected in
none; however, none of the prepared samples has the
exact stoichiometric composition of this compound.
Diffraction patterns of three samples were used
for obtaining the lattice parameters of monoclinic
BAYF; results are presented in Table 2, along with
two values found in the literature [12, 16].
Fig. 3 DTA curves from experiments performed with differing
rates; sample composition 78% YF3
J. Therm. Anal. Cal., 95, 2009
45

NAKAMURA et al.
nation by oxygen, as suggested by the partial
YF3–Y2O3 phase diagram from [17]. The presence of a
hint of contamination as early as the first heating run
demonstrates how really vulnerable to oxygen the
compound YF3 is. The curve obtained from a third
consecutive heating, vastly different from the previ-
ous curve, shows that the sample has changed almost
completely. The transformation of YF3 into one or
more oxyfluorides throughout of the experiment is
evidenced by the thermal event at 568°C (onset). This
event likely corresponds to the rhombohedral®cubic
polymorphic transition of YOF, reported to occur at
571°C by [18] and at 560°C by [19].
Hence, the absence of thermal events upon
cooling in the experiments performed with 10 K min–1
may be explained by the formation of oxyfluoride
compounds that do not present transitions in the
temperature range explored. It is also possible that
composition variations caused by differential evaporation
of compounds within the samples coupled with
massive oxygen contamination may have caused the
samples to undergo a glass transition.
Fig. 4 TG curves obtained simultaneously with the DTA
curves from Fig. 3
The use of a much faster rate of heating and
cooling certainly reduced, but did not eliminate the
contamination problem in the second set of DTA
experiments. Moreover, peak overlap was unduly
exacerbated. Results were found to be entirely un-
reliable for accurate identification of phase transitions
and precise determination of the temperatures in
which they occur.
The DSC system used in this work allowed for a
much better controlled atmosphere. Vacuum treatment
of the furnace and samples prior to the measurement
led to much smaller loss of mass during experiments,
despite the use of lower thermal rates: typically less
than 1% per heating/cooling run. This indicates that
oxyfluoride formation was negligible. Evaporation
accounts for the remaining loss and was also small
since crucibles lids were used.
Fig. 5 DTA curves of the first and third heating runs of pure YF3
performed with the 10 K min–1 rate, samples underwent
some process which led to a complete change in
character. In the experiments performed with the
40 K min–1 rate, four times faster than the previous
ones, samples were clearly less affected.
Indeed, TG curves from the first set of experi-
ments show a steep loss of mass: total loss varied
from 8 to 14% of initial masses. In the second set of
experiments, total mass loss was smaller, though still
significant: 2–6% of initial masses. A comparative
example is shown in Fig. 4.
The diminutive loss of mass made possible a
second consecutive heating/cooling run during each
experiment. This is desirable, since the sample homo-
geneity problem detected through XRD is reduced, in
the second run.
Evaporation and oxygen contamination account
for the observed loss of mass. Formation of oxy-
fluoride phases is likely the cause of the anomalous
results from the first set of DTA experiments. It is
well known that heating rare-earth fluorides under a
non-fluorinating atmosphere may produce oxy-
fluoride phases due to reaction with even trace
amounts of oxygen and moisture [17]. Figure 5 shows
the DTA curves of two heating runs of pure YF3. The
curve obtained from the first heating of the compound
shows two large thermal events corresponding to the
polymorphic transition and melting of YF3 and a
smaller event in-between probably related to contami-
Overlapping of peaks remained a problem in all
thermal analysis curves; even for those DSC curves
obtained with 5 K min–1 rates. Curiously, very little
variation was observed between the onset temperatures
of the very first event in each thermal analysis curve. In
all DSC curves, the first thermal event always occurred
at the temperature of 954°C (onset); the same was
observed in all DTA curves as well, with slightly larger
temperature variation. This can be observed in the
example given in Fig. 6.
46
J. Therm. Anal. Cal., 95, 2009

THERMAL INVESTIGATION OF THE BaF2-YF3 SYSTEM
BAYF may be actually very slightly incongruent: it
may decompose before reaching complete melt.
Conclusions
Samples with varying concentrations of BaF2 and YF3
were synthesized under reactive atmosphere. X-ray
diffraction patterns of the samples were obtained and
quantitative phase calculations of their final composition
and the determination of the lattice parameters of BAYF
phase were performed through the Rietveld method.
Thermal analysis was performed under different
conditions in order to study the influence of experi-
mental settings in thermal measurements of the binary
fluoride system BaF2–YF3. Contamination of the
samples with traces of oxygen and moisture during
thermal analysis experiments was determined to be a
very relevant hindrance to the thermal investigation
of the system. It must be mitigated through the use of a
well controlled atmosphere, preferably with vacuum
treatment of the furnace and the sample previously to
the each experiment. Overlapping of thermal peaks
remains a problem even if low thermal rates are used.
Interpretation of resulting thermal curves was found
to be further complicated by intense supercooling
effects, as described in the literature. Cooling curves
are unreliable for determination of phase transition
temperatures.
Fig. 6 DSC and DTA heating curves of sample 77% YF3
Experimental data mostly corroborate the phase
diagram from the literature at low temperatures, but
some deviation was observed at higher temperatures,
at least in the composition range surrounding the
BAYF compound. Thermal analysis curves show no
variation in the polymorphic transition temperature of
BAYF in each side of the compound. Conclusive
evidence of the presence of the solid solution range of
BAYF mentioned in previous literature has not been
found. All thermal peaks were observed to occur at
temperatures much higher than described by the
diagram. DSC results suggest the possibility that
BAYF may actually be incongruent. The observed
discrepancies are under further investigation and will
be the subject of a next publication.
Fig. 7 Two DSC curves obtained consecutively for samples
with 1:2 compositions (BAYF phase)
Partial separation of peaks was achieved in the
DSC curve of the 1 K min–1 heating run of sample
66.7% YF3. The relevant portions of this curve and
the one obtained with 10 K min–1 can be seen in Fig. 7.
The first peak was identified as corresponding to the
polymorphic phase transition of BAYF reported
by [6, 7]. This transition must be the one responsible
for the invariant first onset (954°C) observed in all
samples. Results disagree with the phase diagram in
Fig. 1, since no difference in the temperature of this
transition in each side of the stoichiometric BAYF
line in the diagram was observed. Moreover, peaks in
all thermal curves occur at temperatures systematically
higher than what the phase diagram would allow:
according to the diagram BAYF melts at 960°C and
its phase transition occur at 920°C (eutectoid phase
transition at 928°C).
Acknowledgements
The authors acknowledge the financial support given by CNPq,
Conselho Nacional de Desenvolvimento Científico e Tecnológico
(project number 480365/2004-0) and CAPES/PROBRAL,
Coordena¸±o de Aperfei¸oamento de Pessoal de Nível Superior
(project number 249/06).
Remaining thermal events in the curve from Fig. 7
(1 K min–1) were attributed to melting. The elongated
profile of the melting suggests the possibility that
J. Therm. Anal. Cal., 95, 2009
47

NAKAMURA et al.
12 L. H. Guilbert, J. Y. Gesland, A. Bulou and R. Retoux,
References
Mater. Res. Bull., 28 (1993) 923.
13 A. K. Cheetham and N. Norman, Acta Chem. Scand.,
28 (1974) 55.
1 A. Agnesi, G. Carraro, A. Guandalini, G. Reali, E. Sani,
A.Toncelli and M. Tonelli, Opt. Express, 12 (2004) 3765.
2 A. Agnesi, A. Guandalini, A. Lucca, E. Sani, A. Toncelli
and M. Tonelli, Opt. Express, 11 (2003) 1149.
3 E. Sani, A. Toncelli, L. M. Tonelli, A. Agnesi,
A. Guandalini and G. Reali, IEEE J. Quant. Elec.,
39 (2003), 971.
14 B. A. Maksimov, Kh. Solans, A. P. Dudka, E. A. Genkina,
M. Font-Badria, I. I. Buchinskaya, A. A. Loshmanov,
A. M. Golubev, V. I. Simonov, M. Font-Altaba and
B. P. Sobolev, Cristallogr. Rep., 1 (1996) 56.
15 G. B. Clifton and H. R. Subhash, Introduction to Phase
Equilibria in Ceramics, The American Ceramic Society
Inc., Columbus 1997.
4 A. F. H. Librantz, L. Gomes, S. L. Baldochi, I. M. Ranieri
and G. E. Brito, J. Lumin., 121 (2006) 137.
5 E. G. Ippolitov and A. G. Maklachko, Inorg. Mater.
(USSR) (Engl. Transl.), 6 (1970) 124.
16 O. Greis, P. Stede and M. Kiesser, Anorg. Allg. Chem.,
477 (1981) 133.
17 B. P. Sobolev, P. P. Fedorov, D. B. Shteynberg, B. V. Sinitsyn
and G. S. Shakhkalamian, J. Solid State Chem., 17 (1976) 191.
18 D. B. Shinn and H. A. Eick, Inorg. Chem., 8 (1968) 232.
19 K. Niihara and S.Yajima, Bull. Chem. Soc. Jpn.,
45 (1972) 20.
6 N. L. Tkachenko, M. Svantner and B. P. Sobolev,
Inorg. Mater. (USSR) (Engl. Transl.), 13 (1977) 693.
7 B. P. Sobolev and N. L. Tkachenko, J. Less Common Met.,
85 (1982) 155.
8 H. J. Guggenheim and L. F. Johnson, Appl. Phys. Lett.,
15 (1969), 51.
9 L. Xingren, X. Gang and R. C. Powell, J. Solid State Chem.,
62 (1986) 83.
Received: January 16, 2008
Accepted: April 15, 2008
OnlineFirst: August 15, 2008
10 P. Hagenmuller, Inorganic Solid Fluorides. Chemistry and
Physics, Academic Press Inc., New York 1985.
11 R. A. Young, A. C. Larson and C. O. Paiva-Santos, User´s
Guide to Program DBWS-9807a for Rietveld Analysis of
X-ray and Neutron Powder Diffraction Patterns, Institute
of Technology, Atlanta 1999.
DOI: 10.1007/s10973-008-9005-3
48
J. Therm. Anal. Cal., 95, 2009