The phase diagram of the system LiF-GdF3

Article abstract of DOI:10.1016/j.jallcom.2004.02.010

The phase diagram of the system LiF-GdF₃ has been revised, using differential thermal analysis (DTA). We observed a eutectic reaction at 25mol% of GdF₃ and 698°C and a peritectic reaction at 34mol% of GdF ₃ and 755°C. We found indications for a GdF₃ phase transformation from hexagonal to orthorhombic at 900°C. An identification of the formed phases was made by X-ray diffraction and SEM.

Full text of DOI:10.1016/j.jallcom.2004.02.010

Journal of Alloys and Compounds 379 (2004) 95–98
The phase diagram of the system LiF–GdF3
I.M. Ranieria,∗, A.H.A. Bressianib, S.P. Moratoa, S.L. Baldochia
a
Center for Lasers and Applications, Inst. Pesquisas Energeticas & Nucl., CP 11049, Butantã 05422-970, São Paulo, SP, Brazil
b
Center for Material Science and Technology, IPEN, CP 11049, Butantã 05422-970, São Paulo, SP, Brazil
Received 15 December 2003; received in revised form 10 February 2004; accepted 10 February 2004
Abstract
The phase diagram of the system LiF–GdF3 has been revised, using differential thermal analysis (DTA). We observed a eutectic reaction at
25 mol% of GdF3 and 698 ◦C and a peritectic reaction at 34 mol% of GdF3 and 755 ◦C. We found indications for a GdF3 phase transformation
from hexagonal to orthorhombic at 900 ◦C. An identification of the formed phases was made by X-ray diffraction and SEM.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Phase equilibrium; Binary systems; Rare earth fluorides; DTA
1. Introduction
2. Experimental procedures
The phase diagram of the system LiF–GdF3 was stud-
ied by Thoma et al. in 1960s [1,2], and it was found that it
presents two invariant points: a eutectic at 26 mol% GdF3
and 700 ◦C, a peritectic at 39 mol% GdF3 and 755 ◦C and
a GdF3 phase transformation from hexagonal to orthorhom-
bic at 875 ◦C. LiGdF4 is the only intermediary compound,
showing an incongruent melting behavior. Pham et al. [3]
revised superficially this diagram determining the following
values: a eutectic at 20 mol% GdF3 and 627 ◦C and a peri-
tectic at 32 mol% GdF3 and 727 ◦C.
LiGdF4 crystals are suitable for doping with light rare
earth ions with a particular interest in neodymium ions
to develop new laser media [4,5]. To grow good quality
crystals from the melt the knowledge of the peritectic com-
position is important in order to minimize the excess of
one of the components, because this excess acts as im-
purity causing defects (as inclusions) in the crystals. For
this reason we aimed at examining this diagram more
carefully, using differential thermal analysis (DTA) to con-
struct the phase diagram. The formed phases have been
identified by X-ray diffraction and the microstructure of
some samples has been examined using scanning electron
microscopy.
The samples utilized to construct the LiF–GdF3 phase dia-
grams were prepared from LiF (99.9%, Aldrich), previously
purified by zone refining under a flux of hydrofluoric acid
(HF) and argon (Ar) gases. GdF3 was obtained by Gd2O3
(99.99%, Alfa Aesar) hydrofluorination at high temperature
in HF atmosphere. Samples weighting around 5 g with dif-
ferent compositions were melted in the same atmosphere
and then pulverized for homogenization.
DTA curves were obtained in a TGA–DTA equipment,
model 2960, TA Instruments. The samples weighing around
50 mg were placed in open platinum crucibles without a ref-
erence material. The measurements were performed under
a flux of purified helium, with a heating rate of 10 ◦C/min.
These conditions allowed that the broad melting curves
could be resolved. The measurements were performed for
temperatures up to 900 ◦C, because above this temperature
there was a sensible loss of mass causing data scattering
under these conditions the maximum loss of mass was
about 1%.
The reasons for this mass loss are the evaporation of LiF
and the high reactivity of the rare earth fluorides with resid-
ual air or moisture present during the experiment at high
temperatures [6]. The evidence of some contamination was
confirmed during the determination of the rare earth fluoride
melting points, they could be obtained only with heating
rates of 40 ◦C/min; otherwise, the samples were completely
converted into the oxide above 900 ◦C. Therefore, for the
∗
Corresponding author. Tel.: +55-11-3816-9306;
fax: +55-11-3816-9315.
E-mail address: iranieri@net.ipen.br (I.M. Ranieri).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.02.010

96
I.M. Ranieri et al. / Journal of Alloys and Compounds 379 (2004) 95–98
construction of the phase diagram we considered only the
range of compositions up to 60 mol% GdF3, this limit being
set by the peak temperatures of the transformations in the
heating mode. Sealed crucibles or evacuation of the system
prior the measurements would minimize this effect.
To determine the phases present in each of the two-phase
region of the phase diagrams, two samples of each com-
position were melted under a flux of hydrogen fluoride gas
and cooled at a rate of 15 ◦C. One of the samples was ex-
amined using a scanning electron microscope model XL30,
from Philips, and the phase compositions were estimated
with an energy-dispersive spectrometer model EDXAUTO,
from EDAX. The other sample was pulverized for analysis
by powder X-ray diffraction, using a Bruker AXS diffrac-
tometer, model D8 Advance, operated at 40 kV and 30 mA,
in the 2θ range of 18–66◦. The lattice constants of the iden-
tified phases were obtained by least squares fittings.
Fig. 2. X-ray curves of the samples: (a) 76 mol% LiF:24 mol% GdF3, (b)
68 mol% LiF:32 mol% GdF3, (c) 50 mol% LiF:50 mol% GdF3.
3. Results and discussion
are present in the samples Gd24 and Gd32 (Fig. 2a and b),
and LiF, LiGdF4, and GdF3 in sample Gd50 (Fig. 2c).
The lattice parameters of LiGdF4 and GdF3 obtained by
XRD are presented in Table 1. They are in accordance with
those cited in the literature [7,8].
The microstructures were observed by SEM and the
phases present were identified also by EDS for the fol-
lowing samples: (a) 76 mol% LiF:24 mol% GdF3 (Gd24),
(b) 74 mol% LiF:26 mol% GdF3 (Gd26), (c) 68 mol%
LiF:32 mol% GdF3 (Gd32), and (d) 50 mol% LiF:
50 mol%GdF3 (Gd50).
The sample Gd24 presented a microstructure (Fig. 3)
formed by LiF dendrites (dark regions) and a fine structured
eutectic formed by LiF/LiGdF4 mixture (lighter regions). In
Fig. 3, the LiF primary phase is present in many locations
of the analyzed area but the eutectic microstructure is dom-
inant. It can be seen that during the solidification various
crystallization fronts occurred simultaneously, generating a
colony structure. The eutectic microstructure changes from
a disordered arrangement of the LiF/LiGdF4 phases near the
LiF dendrites to a regular one with a fine fibrous morphol-
ogy. Therefore, the sample composition was very close to
the eutectic composition, though hypoeutectic in character.
The microstructures of the sample G26 consist almost en-
tirely of eutectic colonies with the same fibrous morphology
(Fig. 4a) and in a small area crystallization of the LiGdF4
primary phase took place (gray part of Fig. 4b). This is due
to the sample composition being slightly richer in GdF3
than the eutectic composition. In fact, considering the eutec-
tic composition determined in the phase diagram (Fig. 1),
one expects the solidification of 4% of the LiGdF4, thus,
confirming that the eutectic composition is around 25 mol%
GdF3.
In Fig. 1, we present the phase diagram obtained for the
LiF–GdF3 system. We found a eutectic reaction at 25 mol%
of GdF3 and 698 ◦C and a peritectic reaction at 34 mol% of
GdF3 and 755 ◦C. The eutectic composition is in agreement
with that obtained by Thoma et al. [2], while the peritectic
composition agrees with that obtained by Pham et al. [3].
The reaction temperatures are in accordance with those ob-
served by Thoma et al. despite the measurements were done
during cooling of the samples. These authors used a method
specified as a fast cooling with higher cooling rates than
that used in conventional DTA equipments. It can then be
concluded that the samples were not supercooled.
The identification of the phases present in each two phases
region was performed by X-ray diffraction for the follow-
ing compositions: (a) 76 mol% LiF:24 mol% GdF3 (Gd24),
(b) 68 mol% LiF:32 mol% GdF3 (Gd32), and (c) 50 mol%
LiF:50 mol% GdF3 (Gd50). The X-ray curves of these sam-
ples shown in Fig. 2 confirm that the LiF and LiGdF4 phases
The intermediary composition of the sample Gd32, be-
tween the eutectic and peritectic invariant points, exhib-
ited the LiGdF4 primary phase. This phase was formed on
Fig. 1. The phase diagram determined for the LiF–GdF3 system.

I.M. Ranieri et al. / Journal of Alloys and Compounds 379 (2004) 95–98
97
Table 1
Values obtained for the lattice parameters of the phases present in the LiF–GdF3 system
Lattice parameters
Composition (1 − x)LiF–xGdF3 (mol%)
LiGdF4 (x = 24)
LiGdF4 (x = 32)
LiGdF4 (x = 50)
LiGdF4 [7]
5.219
GdF3 (x = 50)
GdF3 [8]
a (Å)
b (Å)
5.222 (1)
5.223 (1)
5.225 (1)
6.571 (3)
6.995 (4)
6.570
6.984
c (Å)
Volume (Å3)
10.985 (4)
299.56 (21)
10.986 (1)
299.67(70)
10.987 (3)
299.95 (16)
10.97
4.393 (2)
201.92 (28)
4.393
201.63
Fig. 3. Gd24 SEM photographs, in two different regions of the sample: (a) dendrites of the primary phase (dark region), (b) region where the eutectic is
dominant (light gray regions).
the bottom of the ingot that was in contact with the Pt
crucible and the first one to solidify. In Fig. 5, one can
see that the fibrous eutectic colonies surround the primary
phase.
Finally the solidification of the LiGdF4 stoichiometric
composition begins with the crystallization of the GdF3 pri-
mary phase (gray regions in Fig. 6). When the sample tem-
perature reaches the peritectic temperature, formation of the
stoichiometric phase takes place (dark gray regions) ensuing
from the reaction of the primary phase with the liquid. As
the peritectic reaction is time-consuming, the LiGdF4 phase
is formed only on the GdF3 grain boundary. In consequence
of the high cooling rate there was not enough time for dis-
solution of the primary phase generating a non-equilibrium
condition and then the simultaneous crystallization of the
eutectic mixture LiGdF4/LiF.
This high incongruent melting behavior of LiGdF4 makes
the crystal growth from the melt difficult, as observed by
Ranieri et al. [9], when the peritectic composition was uti-
lized. The crystals were very sensitive to variations in di-
Fig. 4. Gd26 SEM photographs, in two different regions of the sample: (a) dominant fibrous eutectic microstructure (light gray regions), (b) LiGdF4
primary phase (gray region).

98
I.M. Ranieri et al. / Journal of Alloys and Compounds 379 (2004) 95–98
Fig. 5. Gd32 SEM photographs: (a) LiGdF4 primary phase (dark gray region), (b) detail of the eutectic microstructure (light gray region).
Fig. 6. Gd50 SEM photographs, in two different regions of the sample with the LiGdF4 (dark gray regions), GdF3 (gray regions), and the eutectic (light
gray regions).
ameter and presented inclusions, also using raw materials
with high purity and a reactive CF4 atmosphere. The liq-
uid presented high viscosity, which interferes in the con-
vective flow of the melt and inhibits mixing of the melting
components.
References
[1] R.E. Thoma, in: L. Eyring (Ed.), Progress in the Science and Tech-
nology of the Rare Earths, vol. 2, Pergamon Press, Oxford, 1966,
pp. 90–122.
[2] R.E. Thoma, G.D. Brunton, R.A. Penneman, T.K. Keenan, Inorg.
Chem. 9 (1970) 1096.
[3] A. Pham, J. Lefaucheur, G. Lutts, B. Chai, J. Nicholls, in: A.A.
Pinto, T.Y. Fan (Eds.), OSA Proceedings on Advanced Solid State
Lasers, vol. 15, Optical Society of America, Washington, DC, 1993,
pp. 178–184.
[4] X.X. Zhang, M. Bass, A.B. Villaverde, J. Lefaucheur, A. Pham, B.H.T.
Chai, Appl. Phys. Lett. 62 (11) (1993) 1197–1199.
[5] L.C. Courrol, E.P. Maldonado, L. Gomes, N.D.I.M. Ranieri, S.P.
Morato, N.D. Vieira Jr., Opt. Mater. 14 (2000) 81–90.
[6] R.C. Pastor, M. Robinson, Mater. Res. Bull. 9 (1974) 569–578.
[7] G. Keller, H. Schmutz, J. Inorg. Chem. 27 (1965) 900–901.
[8] A. Zalkin, D.H. Templeton, J. Am. Soc. 75 (1953) 2453–2458.
[9] I.M. Ranieri, K. Shimamura, K. Nakano, T. Fujita, L.C. Courrol, S.P.
Morato, T. Fukuda, J. Crystal Growth 217 (1–2) (2000) 145–150.
4. Conclusions
An accurate phase diagram of the system LiF–GdF3 was
determined, showing that this system presents a eutectic at
25 mol% GdF3 and 700 ◦C and peritectic at 34 mol% GdF3
and 753 ◦C. The diagram obtained in this work shows that
the range of compositions in which one can grow LiGdF4
crystals is narrower than that expected initially, limiting the
amount of mass to be extracted during the crystal growth
process.