Pyrolytic synthesis and Eu3+→Eu2+ reduction process of blue-emitting perovskite-type BaLiF3:Eu thin films

Article abstract of DOI:10.1016/j.jssc.2003.10.008

Perovskite-type barium lithium fluoride (BaLiF₃) was synthesized by pyrolysis of metal trifluoroacetates. The reaction temperature necessary for producing a single-phase material was found to be 600°C, which was lower than that for a conventional solid-state reaction or a melting method. Eu-doped BaLiF₃ was also prepared and characterized to examine the suitability of trifluoroacetates for precursors in synthesizing homogeneous complex metal fluoride materials. It was demonstrated that trivalent Eu³⁺, which was used as acetate for a starting material, was reduced to divalent Eu²⁺ in the pyrolysis process of BaLiF ₃, as indicated by a broad blue emission due to an allowed 4f ⁶5d→4f⁷ transition at 408nm with a ultraviolet excitation at 254nm. The concentration quenching of the blue emission occurred at 5at% of Eu in BaLiF₃, indicating that Eu was homogeneously dispersed in the BaLiF₃ host lattice. Mechanisms of the formation and reduction process of BaLiF₃ were discussed based on pertinent chemical reactions.

Full text of DOI:10.1016/j.jssc.2003.10.008

ARTICLE IN PRESS
Journal of Solid State Chemistry 177 (2004) 1032–1036
Pyrolytic synthesis and Eu3+-Eu2+ reduction process
of blue-emitting perovskite-type BaLiF3:Eu thin films
Shinobu Fujihara,ꢀ Yoko Kishiki, and Toshio Kimura
Department of Applied Chemistry, Faculty of Science and Technology, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan
Received 11 July 2003; received in revised form 6 October 2003; accepted 13 October 2003
Abstract
Perovskite-type barium lithium fluoride (BaLiF3) was synthesized by pyrolysis of metal trifluoroacetates. The reaction
temperature necessary for producing a single-phase material was found to be 600ꢁC, which was lower than that for a conventional
solid-state reaction or a melting method. Eu-doped BaLiF3 was also prepared and characterized to examine the suitability of
trifluoroacetates for precursors in synthesizing homogeneous complex metal fluoride materials. It was demonstrated that trivalent
Eu3+, which was used as acetate for a starting material, was reduced to divalent Eu2+ in the pyrolysis process of BaLiF3, as
indicated by a broad blue emission due to an allowed 4f 65d-4f 7 transition at 408 nm with a ultraviolet excitation at 254 nm. The
concentration quenching of the blue emission occurred at 5 at% of Eu in BaLiF3, indicating that Eu was homogeneously dispersed
in the BaLiF3 host lattice. Mechanisms of the formation and reduction process of BaLiF3 were discussed based on pertinent
chemical reactions.
r 2003 Elsevier Inc. All rights reserved.
Keywords: Barium lithium fluoride; Trifluoroacetates; Decomposition; Reduction; Photoluminescence
1. Introduction
an atomic level can be achieved in precursor solutions
and the resultant samples are expected to have high
homogeneity. In the present work, BaLiF3 films were
successfully prepared from mixed metal trifluoroacetates
that were dissolved in an organic solvent. Eu was doped
in BaLiF3 to examine the homogeneity of samples as
well as the potential use as phosphors. It was found that
Eu3+ ions, which were used as acetate for a raw
material, were reduced to Eu2+ in the pyrolysis process
at temperatures as low as 600ꢁC. This contrasts with the
high-temperature Eu3+-Eu2+ reduction which has
been accomplished in oxide and fluoride materials so
far [9,10]. The Eu-doped BaLiF3 films exhibited efficient
blue photoluminescence (PL), which could be used as a
spectroscopic probe for the sample homogeneity.
Mechanisms of the formation and reduction of
BaLiF3: Eu prepared from the trifluoroacetates were
discussed based on pertinent chemical reactions.
Barium lithium fluoride (BaLiF3) has an inverted
perovskite structure where monovalent Li+ ions occupy
the center of fluorine octahedra [1]. Doped, as well as
undoped, BaLiF3 materials have been investigated
aiming at practical applications to X-ray storage
phosphors [2], vacuum ultraviolet (UV) optical materi-
als [3], lasers [4], and ion conductors [5]. Because
electrical and optical properties of BaLiF3 are greatly
dependent on defect structures as well as dopants, a
number of researches have been reported focusing on
the development of synthetic methods by which high-
quality samples can be produced.
While melting of a powder mixture of BaF2, LiF, and
dopant fluorides at high temperatures in a HF atmo-
sphere has been generally utilized to produce doped
BaLiF3 single crystals [6,7], the pyrolytic synthesis
starting from mixed metal-organofluorine compounds
is expected as an attractive alternative for preparing fine
powders and thin films [8]. This synthetic method is
advantageous in that the mixing of metal constituents at
2. Experimental procedure
Metal acetates, Ba(CH3COO)2, LiCH3COO, and
Eu(CH3COO)3 ꢂ 4H2O, were dissolved in isopropanol
to which trifluoroacetic acid (CF3COOH) and water
ꢀCorresponding author. Fax: +81455661551.
E-mail address: shinobu@applc.keio.ac.jp (S. Fujihara).
0022-4596/$ - see front matter r 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2003.10.008

ARTICLE IN PRESS
S. Fujihara et al. / Journal of Solid State Chemistry 177 (2004) 1032–1036
1033
were added. Dopant concentrations of Eu were varied
between 0, 1, 2, 3, 5, 10, and 15 at%. After stirring for
2 h, the resultant solutions were spin-coated on quartz
glass substrates at 2000 rpm. The coated substrates were
then immediately placed in a furnace kept at 400–700ꢁC
and heated for 10 min in a flowing nitrogen atmosphere,
followed by quenching. This coating/heating procedure
was repeated two additional times to increase the film
thickness up to B200 nm. This increase was necessary to
analyze the films accurately.
Phase identification of the films was performed with
an X-ray diffractometer equipped with a thin-film
attachment using CuKa radiation (Rigaku). Thermal
analysis of the solution that was dried at 90ꢁC for 4 days
was done by thermogravimetry-differential thermal
analysis (TG-DTA) (Mac Science, type TG-DTA2020S
thermal analyzer). A heating rate of 5ꢁC/min was
adopted. PL spectra were measured at room tempera-
ture with a spectrofluorophotometer (Shimadzu, type
RF-5300PC) using a Xe lamp (150 W) as a light source.
Emission scans were performed with 5 nm bandpass
emission slits. A filter was used to remove a second-
order peak of the excitation light.
Fig. 1. XRD patterns of thin films obtained by the pyrolysis of
trifluoroacetates at 400–700ꢁC for 10 min in the flowing nitrogen;
BaLiF3, BaF2, and unknown peaks are labelled (K), (.) and ( ꢃ ),
respectively.
100
90
exo.
endo.
80
70
3. Results and discussion
DTA
3.1. Pyrolytic synthesis of BaLiF3
60
The thermal decomposition of metal trifluoroacetates
generally occurs at temperatures higher than 300ꢁC,
which results in the formation of metal fluorides [8].
Thus alkaline earth (MgF2, CaF2, SrF2, and BaF2) and
rare-earth fluorides (LaF3, NdF3, EuF3, etc.) can be
prepared by the trifluoroacetate method. In case of
complex metal fluorides containing two or more metal
constituents, the trifluoroacetates should be pyrolysed at
relatively higher temperatures because of differences in
decomposition temperatures of each individual metal
trifluoroacetate [11]. Fig. 1 shows the X-ray diffraction
(XRD) patterns of the BaLiF3 films obtained by heating
the coated solution at 400–700ꢁC for 10 min in the
flowing nitrogen. Sharp diffraction peaks appearing at
2y ¼ 22:2ꢁ; 31.7ꢁ, 39.0ꢁ, 45.4ꢁ, 51.1ꢁ, and 56.4ꢁ are
assigned to (100), (110), (111), (200), (210), and (211)
planes, respectively, of BaLiF3 with a cubic structure
and a lattice constant of a ¼ 3:9950 A according to
JCPDS No. 18-715. While the film that was heated at
600ꢁC is a single-phase BaLiF3, the films heated at the
lower temperatures of 400ꢁC or 500ꢁC contain BaF2 as
an impurity phase. A possible formation of LiF was not
confirmed by XRD because of its lower detectability for
light elements. The film heated at 700ꢁC is contaminated
with unknown crystalline phases, which are supposed to
be decomposition products of BaLiF3.
50
TG
40
0
200
400
600
800 1000
Temperature/˚C
Fig. 2. TG-DTA curves of the dried trifluoroacetate precursor
solution.
Fig. 2 shows the TG-DTA curves of the dried
trifluoroacetate solution. A weight loss and an en-
dotherm around 200ꢁC are due to dehydration of the
trifluoroacetates and vaporization of organic com-
pounds such as isopropanol, trifluoroacetic acid, and
acetic acid that were derived from the solvent and
starting materials. The DTA curve exhibits exothermic
peaks that are characteristic of the decomposition of the
trifluoroacetates. The three peaks at 299ꢁC, 320ꢁC, and
352ꢁC are ascribed to the formation of LiF, BaF2, and
BaLiF3, respectively. These exotherms accompany a
large weight loss. A sharp endothermic peak at 850ꢁC is
due to melting of BaLiF3. The formation reaction of
BaLiF3 from the trifluoroacetates is nominally expressed
by
BaðCF3COOÞ2 þ LiCF3COO-BaF2 þ LiF-BaLiF3
ð1Þ

ARTICLE IN PRESS
S. Fujihara et al. / Journal of Solid State Chemistry 177 (2004) 1032–1036
1034
referring to the suggestion by Rillings and Roberts [12].
The weight loss observed around 300ꢁC is 51.1%, which
agrees roughly with a calculated weight loss, 58.3%, for
the above reaction. The results of the thermal analysis
explain the crystallization behavior of the films shown in
Fig. 1. The heating temperature should be enough high
to avoid the separate formation of each constituent
fluoride (BaF2 and LiF) and to produce the single-phase
BaLiF3 samples.
samples. Very recently, Hua et al. [18] have reported
the synthesis of the Eu-doped BaLiF3 through a
solvothermal process and have found the similar lattice
constants between the doped and the undoped sample.
The case is such that Eu2+-doped and undoped
BaMgF4 have similar lattice constants and the same
space group [19].
8
The divalent europium ion, which has a 4f 7 S7=2
electronic configuration in the ground state, shows a
broad absorption band in a long-wavelength UV region
due to an allowed 4f 7-4f 65d transition. Emission
wavelengths for transitions from the excited to the
ground state (4f 65d-4f 7) can vary from long-wave-
length UV to yellow depending on crystal fields of host
lattices [20]. Fig. 4 shows the PL spectra of the
BaLiF3:Eu (3 at%) film that was heated at 600ꢁC in
the flowing nitrogen atmosphere. The excitation wave-
length used was 254 nm, which would provide efficient
blue luminescence in BaLiF3:Eu2+ [10]. The spectra in
Fig. 4 are typical of the Eu2+ ions doped in the BaLiF3
crystal of which the Ba2+ site symmetry is Oh: A broad
blue emission with a peak wavelength of 408 nm is due
to the 4f 65d-4f 7 transition. A narrow peak at 360 nm
3.2. Eu-doped BaLiF3
Fig. 3 shows the XRD patterns of the Eu-doped
BaLiF3 films that were heated at 600ꢁC. Almost single-
phase samples are obtained with the Eu concentration
up to 5 at%. At 10 and 15 at%, peaks due to the
perovskite phase become weaker, indicative of lower
crystallinity. Thus the solubility limit of Eu in BaLiF3 is
suggested to be 5 at%. The valence state of Eu in BaLiF3
cannot be estimated from the XRD data. In oxide
perovskite materials, doped Eu ions generally exist in
the trivalent state [13–16]. In contrast, Couto dos Santos
et al. [17] have suggested that the solution energy for the
Eu3+ ions in the BaLiF3 matrix is highly positive,
implying that the incorporation of Eu3+ into the Ba2+
sites has a low probability. Because Eu2+ and Ba2+
have similar ionic radii as well as the same configuration
of outermost electrons (5s25p6), it is reasonable to think
that the Eu ions are doped in BaLiF3 in the divalent
state. Even so, the XRD analysis seems to be incapable
of examining the structural difference induced by the
Eu-doping because no appreciable change in lattice
constants was observed for the Eu-doped BaLiF3
6
is ascribed to a P1=2-8S7=2 transition. Thus the
presence of the optically active Eu2+ ions in BaLiF3
has been demonstrated.
The Eu3+-Eu2+ reduction is generally achieved by
firing materials at high temperatures in a reducing
atmosphere or air. When a reducing H2 gas is used, a
reduction mechanism can be easily explained as follows:
2Eu3þ þ H2-2Eu2þ þ 2Hþ:
ð2Þ
The Eu3+ ions are efficiently reduced by reacting with
H2. In contrast, the reduction in air should be caused by
chemical interactions between Eu3+ and constituents of
host lattices. For example, Peng et al. [9] have explained
350 400 450 500 550 600 650 700
Wavelength/nm
Fig. 3. XRD patterns of Eu-doped BaLiF3 thin films obtained by the
pyrolysis of trifluoroacetates at 600ꢁC for 10 min in the flowing
nitrogen; BaLiF3 and unknown peaks are labelled (K) and ( ꢃ ),
respectively. The Eu concentration was varied between 0 and 15 at%.
Fig. 4. PL spectra of the BaLiF3:Eu (3 at%) film. The excitation
wavelength used was 254 nm.

ARTICLE IN PRESS
S. Fujihara et al. / Journal of Solid State Chemistry 177 (2004) 1032–1036
1035
the reduction of Eu in Sr4Al14O25 based on a charge
compensation model where strontium vacancies play a
key role. The Eu reduction seems to be more feasible in
fluoride lattices due to chemical interactions between
Eu3+ and F–. When Eu3+ is doped in the BaLiF3 lattice
through a solid-state reaction between BaF2, LiF, and
EuF3, it is reduced to Eu2+ by a following reaction:
intensity cannot be ascribed only to the higher Eu
concentrations. The concentration quenching, however,
is clearly demonstrated for the film with 5 at% of Eu
which has well crystallinity as judged from Fig. 3.
Therefore the critical concentration, xc; is suggested to
be equal to x ¼ 0:03 in Ba1ꢄxEuxLiF3.
The concentration quenching of the luminescent
Eu2+ ions doped in host crystals occurs by energy
transfer between them [21]. According to Blasse [22], the
critical transfer distance, Rc; can be estimated by
Eu3þ þ Fꢄ-Eu2þ þ F:
ð3Þ
This reduction was actually achieved at 750ꢁC in the
flowing nitrogen atmosphere [10]. Electrons of the F–
ions are trapped by the Eu3+ ions forming Eu2+. The F–
ions, which lose electrons, form interstitial F atoms. The
inverse reaction of (3) can naturally occur because
fluorine is a strongly oxidizing species. Therefore, it is
thought that reaction (3) is possible only in a special case
where Eu2+ can be stabilized in the solid state of
BaLiF3. In the present case, the Eu-doped BaLiF3 has
been prepared from the trifluoroacetates. Because
reaction (1) actually involves the generation of gaseous
phases such as (CF3CO)2O, CF3COF, and COF2 [12], it
seems that F– can be supplied from these organofluorine
species. Reminding that the film formation was achieved
by placing the coated substrates immediately in the
furnace kept at the heat-treatment temperature, it is
highly possible that several reaction steps proceed
almost simultaneously such as the dehydration, the
vaporization of the organic components, the decom-
position of the trifluoroacetate, the generation of the
gases, and the formation of the Eu-doped BaLiF3. In the
presence of water, F– ions can be supplied by hydrolysis
of CF3COF and COF2 [12]. Thus the Eu3+-Eu2+
reduction process is regarded as a solid–gas reaction.
Fig. 5 shows the dependence of the PL intensity (the
408 nm emission) on the Eu concentration in the
BaLiF3:Eu films. The intensity increases with increasing
concentration up to 3 at% and then decreases. Because
the crystallinity of BaLiF3:Eu (10 and 15 at%) is
relatively low as indicated in Fig. 3, the decreased
ꢀ
ꢁ
1=3
3V
RcE2
;
ð4Þ
4pxcN
where N is the number of ions, which can be replaced by
the luminescent ions, in the unit cell and V is the unit
cell volume. For BaLiF3, N is the number of the Ba2+
ion and equals to 1. V is calculated to be 63.76 A3. By
taking xc of 0.03, Rc is found to be 16 A. This value is
close to those of other Eu2+-activated phosphors
[21,23], implying that the Eu2+ ions are well dispersed
in the BaLiF3 host. If Eu-aggregations, clusters, or EuF2
were formed, the xc value would be much smaller and
the Rc value would not agree with the appropriate one.
Thus the present experiments demonstrate that the high-
quality, blue-emitting Eu-doped BaLiF3 can be prepared
from the metal trifluoroacetates by pyrolysis.
4. Conclusions
The undoped and Eu-doped BaLiF3 films were
prepared from the metal trifluoroacetates. The crystal-
lization of the metal fluorides from the trifluoroacetate
was influenced by the decomposition temperature, and
the single-phase BaLiF3 film was obtained by pyrolysis
at 600ꢁC. The Eu-doped BaLiF3 films exhibited blue PL,
which indicated the reduction of Eu3+ to Eu2+ during
pyrolysis and BaLiF3-formation process. Such the
reduction was explained by the generation of F– species,
the reaction between Eu3+ and F–, and stabilization in
the BaLiF3 lattice. The concentration quenching of the
408 nm PL emission was observed with the Eu
concentrations over 5 at%, which demonstrated the well
dispersion of the Eu2+ ions in BaLiF3.
Acknowledgments
This work was supported by Grant-in-Aid for
Scientific Research (No. 15760506) and for the 21st
century COE program ‘‘KEIO Life Conjugate Chem-
istry’’ from the Ministry of Education, Culture, Sports,
Science, and Technology, Japan.
0
5
10
15
Eu concentration/at.%
Fig. 5. Dependence of the PL intensity (the 408 nm emission) on the
Eu concentration in the BaLiF3:Eu thin films.

ARTICLE IN PRESS
S. Fujihara et al. / Journal of Solid State Chemistry 177 (2004) 1032–1036
1036
References
[11] S. Fujihara, S. Ono, Y. Kishiki, M. Tada, T. Kimura, J. Fluorine
Chem. 105 (2000) 65.
[12] K.W. Rillings, J.E. Roberts, Thermochim. Acta 10 (1974) 285.
[13] Y. Tadokoro, Y.J. Shan, T. Nakamura, S. Nakamura, Solid State
Ionics 108 (1998) 261.
[1] A. Boumriche, J.Y. Gesland, A. Bulou, M. Rousseau,
J.L. Fourquet, B. Hennion, Solid State Commun. 91 (1994) 125.
[2] C.T. Xia, C.S. Shi, Mater. Res. Bull. 32 (1997) 107.
[3] A. Bensalah, K. Shimamura, T. Fujita, H. Sato, M. Nikl,
T. Fukuda, J. Alloys Compd. 348 (2003) 258.
[14] T. Tsuji, T. Nagano, Solid State Ionics 136 (2000) 179.
[15] B. Nalini, T. Takeuchi, H. Kageyama, Solid State Ionics 154
(2002) 629.
[4] M. Mortier, J.Y. Gesland, M. Rousseau, F. Auzel, D. Meichenin,
Radiat. Eff. Defects Solids 136 (1995) 975.
[16] P.J. Deren, J.C. Krupa, J. Lumin. 102 (2003) 386.
[17] M.A. Couto dos Santos, M.E.G. Valerio, R.A. Jackson, J.F. de
Lima, Chem. Phys. Lett. 369 (2003) 90.
[5] M. Kamata, A. Matsumoto, T. Esaka, Denki Kagaku 66 (1998)
443.
[18] R. Hua, B. Lei, D. Xie, C. Shi, J. Solid State Chem. 175 (2003)
284.
[6] A.M.E. Santo, S.P. Morato, S.L. Baldochi, J. Cryst. Growth 203
(1999) 156.
[19] F. Kubel, H. Hagemann, H. Bill, Z. Kristallogr. 214 (1999) 139.
[20] G. Blasse, B.C. Grabmaier, Luminescent Materials, Springer,
Berlin, 1994.
[7] S.L. Baldochi, K. Shimamura, K. Nakano, N. Mujilatu,
T. Fukuda, J. Cryst. Growth 200 (1999) 521.
[8] S. Fujihara, in: S.G. Pandalai (Ed.), Recent Research Develop-
ments in Materials Science, Vol. 3, Research Signpost, Kerala,
India, 2002, p. 619.
[21] D. Wang, Q. Yin, Y. Li, M. Wang, J. Lumin. 97 (2002) 1.
[22] G. Blasse, Philips Res. Rep. 24 (1969) 131.
[23] S.H.M. Poort, W.P. Blokpoel, G. Blasse, Chem. Mater. 7 (1995)
1547.
[9] M. Peng, Z. Pei, G. Hong, Q. Su, Chem. Phys. Lett. 371 (2003) 1.
[10] Y. Tan, C. Shi, J. Phys. Chem. Solids 60 (1999) 1805.