The Copper Ampoule: A New Reactor for the Solid-State Synthesis of Complex Lanthanide Fluorides

Article abstract of DOI:10.1006/jssc.1998.8008

For the first time, copper ampoules have been used as reactors for the solid-state synthesis of complex lanthanide fluorides under vacuum. The high ductility of copper allows for high-vacuum tight sealing of the ampoules by cold welding. Furthermore, no chemical reactivity of copper toward solid fluorides has been observed. Thus, for an examination in view of optical upconversion properties, the superstructure phases Ba_4-xY_3+xF_17+x(x≈0.08) and Pb_4?xY_3±xF_17±x(x≤0.2) have been synthesized by annealing of mixtures of binary fluorides in evacuated copper ampoules.

Full text of DOI:10.1006/jssc.1998.8008

Journal of Solid State Chemistry 142, 152—155 (1999)
Article ID jssc.1998.8008, available online at http://www.idealibrary.com on
The Copper Ampoule: A New Reactor for the Solid-State Synthesis
of Complex Lanthanide Fluorides
Philipp Egger, Nils Peters, and Ju¨ rg Hulligerꢀ
Department of Chemistry and Biochemistry, University of Berne, Freiestrasse 3, CH-3012 Berne, Switzerland
Received April 30, 1998; in revised form August 12, 1998; accepted August 17, 1998
(
(
oxidation) and water (hydrolysis) at elevated temperatures
formation of oxyfluorides and oxides with release of fluor-
For the first time, copper ampoules have been used as reactors
for the solid-state synthesis of complex lanthanide fluorides ine or hydrogen fluoride), synthesis has to be performed
under vacuum. The high ductility of copper allows for high- under an inert atmosphere (inert gas or vacuum) in an
vacuum tight sealing of the ampoules by cold welding. Further-
more, no chemical reactivity of copper toward solid fluorides has
been observed. Thus, for an examination in view of optical upcon-
appropriate reactor (12, 13). As described in the literature,
fluoride powders have mainly been obtained by annealing in
argon-purged and sealed ampoules of tantalum, gold, plati-
num, or nickel (14—16). However, for a screening procedure,
requiring the synthesis of a large number of compounds,
a reactor material has to be found which is inexpensive and
easy to handle. Furthermore, it has to be inert toward both
reactants and products and resist possible corrosive gases.
version properties, the superstructure phases Ba4؊x
Y
F
3؉x 17؉x
(
x+0.08) and Pb4Gx
Y
3؎x
F
17؎x (x4 0.2) have been synthesized
by annealing of mixtures of binary fluorides in evacuated copper
ampoules. ( 1999 Academic Press
Key Words: evacuated copper ampoules; solid state synthesis
in copper ampoules; solid state synthesis of complex lanthanide
fluorides; synthesis of barium yttrium fluorides; synthesis of lead D e´ ne` s introduced bent copper tubes as new inexpensive and
yttrium fluorides.
convenient reactors for fluorides of metals in suboxidation
states (17). According to Gmelin (18), copper is not—or just
slightly—reactive toward fluoride compounds and hydro-
gen fluoride. Fluorine, however, will attack copper at elev-
INTRODUCTION
ated temperatures with formation of CuF . For the
ꢁ
synthesis of SnF -based fluorides, De
´
ne
`
s used copper tubes
Lanthanide (¸n) compounds exhibit many interesting
solid-state properties, including so-called upconversion
UC). UC describes the conversion of infrared to visible or
even ultraviolet light (1). UC is attractive in view of the
realization of compact visible UC lasers (2, 3) or UC phos-
phors (4—7), which explains the interest for new and highly
efficient UC materials. High UC efficiency has been found in
n halides (8), showing lower phonon energies and, conse-
quently, reduced radiationless multiphonon relaxation
9, 10). However, among ¸n halides, only fluorides are air
stable (not hygroscopic) and are therefore of interest in view
of real world applications.
For the search for new and efficient fluoride UC com-
pounds, including new as well as known phases, we have
developed a screening procedure, synthesizing a broad
range of compounds by annealing and assessing their UC
properties using a specially designed powder test setup (11).
Since ¸n fluoride compounds are sensitive to oxygen
ꢁ
which were sealed under inert gas by flattening and bending
the end around itself three or four times. This method is
efficient for leak-free reactions under moderate heating (e.g.,
(
1
h at 400—500°C). However, leakage occurred at higher
temperatures. Furthermore, copper tubes sealed by this
technique cannot be evacuated. Altogether, this reduces the
value of the bent copper tube technique for the solid-state
synthesis of complex ¸n fluorides.
In this paper, we present the use of evacuated copper
ampoules as reactors for the solid-state synthesis of ternary
¸
(
¸
n fluorides. Copper ampoules, being tight up to 900°C and
8 h, are obtained from tubes which are cold welded using
4
a special pair of pliers. This represents a simple sealing pro-
cedure, which can be applied on a vacuum line system or in
an inert gas purged glovebox. We have developed the new
ampoule technique for the synthesis of the complex ¸n
fluorides. Our interest is directed toward yttrium-contain-
ing ternary fluoride host compounds, which may be doped
with, erbium as the optically active ion. By introducing
a second heavy cation (barium or lead), our aim is to further
enhance UC efficiency. BaY F has already proved its
ꢀ
To whom correspondence should be addressed. E-mail: juerg.hulliger@
iac.unibe.ch.
ꢁ
ꢂ
1
52
0
022-4596/99 $30.00
Copyright ( 1999 by Academic Press
All rights of reproduction in any form reserved.

REACTIONS IN EVACUATED COPPER AMPOULES
153
potential as an UC laser host material (19) and established
our interest for further intermediate compounds of the
phase systems BaF —YF and PbF —YF , respectively. We
ꢁ
ꢃ
ꢁ
ꢃ
have found Ba
Y
F
(x40.05) (20, 21) and
G
$
$
ꢁ
V ꢀ V ꢄ V
Pb
Y
F
(x40.2) (14, 22) to be attractive
G
$
$
ꢅ V ꢃ V ꢀꢄ V
candidates for an assessment of their potential as an UC
host. This is mainly based on interesting compositional
(phase width) and structural (superstructure) features of
these phases. Furthermore, they exhibit compositional anal-
ogy or proximity to highly efficient chloride UC compounds
[
Ba ErCl and Ba YCl :Erꢃ>, respectively (23, 24)]. By
ꢁ
ꢄ
ꢁ
ꢄ
annealing corresponding mixtures of binary fluorides in
evacuated copper ampoules, we have obtained Pb G
ꢅ V
Y
F
(x40.2) as a pure phase. However, annealing
$
$
ꢃ V ꢀꢄ V
of the (BaF :YF ) sample yielded a mixture of a solid
solution (Ba, Y)F
(
ꢁ
ꢃ
ꢁ
(x+0.28) and Ba
Y
F
>V
ꢅ\V ꢃ>V ꢀꢄ>V
x+0.08), a further superstructure phase. This is due to
Ba
Y
F
disproportionating into the two above-
G
$
$
ꢁ
V ꢀ V ꢄ V
mentioned phases below 945°C.
FIG. 1. Welding pliers (a) and copper ampoule (b) as used for the
solid-state synthesis of ternary lanthanide fluorides.
EXPERIMENTAL AND RESULTS
Yttrium fluoride was prepared from the corresponding
oxide (Aldrich, 99.9%) by dissolving in hot hydrochloric vacuum tightness. The obtained ternary ¸n fluorides were
acid and subsequent precipitation of the hydrated fluoride ground, yielding crystallites 50—100 lm in diameter.
with hydrofluoric acid. Drying, dehydration, and recrystall-
The powders were cast into an epoxy resin and analyzed
ization were performed by slow heating in a horizontal tube by an electron microprobe (Cameca SX50, 20 kV, 20 nA),
furnace at 800°C for 12 h under a gently flowing mixture of determining the composition of individual crystallites. The
N and HF. The YF obtained in this way could be handled results showed strong scattering due to the dependence of
ꢁ
ꢃ
in a normal atmosphere without showing any sign of hygro- the analysis on the surface roughness of the crystallites.
scopic degradation for hours. However, they indicated the presence of a single phase in the
The ternary ¸n fluorides were obtained by annealing (PbF :YF ) product and of two phases in the (BaF :YF )
ꢁ
ꢃ
ꢁ
ꢃ
anhydrous powders of YF , BaF (Fluka, 597%), and sample. In the latter case, the elemental compositions were
ꢃ
ꢁ
PbF (Fluka, 599%). The mixture ratios were 4:3 for additionally determined as &16 at.% Ba/&12 at.%
ꢁ
PbF :YF and 2:1 for BaF :YF . The loose powder mix- Y [Ba Y F
ꢅ ꢃ ꢀꢄ
(25, 26): 16.7 at.% Ba/12.5 at.% Y] and
(25, 27): 22 at.%
ꢁ
ꢃ
ꢁ
ꢃ
ture or a pressed pellet was placed into a glassy carbon &20 at.% Ba/&8 at.% Y [(Ba, Y)F
ꢁ
ꢆꢁꢂ
crucible (0.3 cmꢃ) and closed with graphite wool to minim- Ba/8.5 at.% Y], respectively. For all powder samples, no
ize transport by sublimation. The crucible was put into contamination with copper (detection limit"100 ppm)
a copper tube (commercially available oxygen-free copper; could be found.
inner diameter: 8 mm; wall thickness: 1 mm) which was
previously sealed on one side by squeezing using a special der diffraction camera (Enraf-Nonius, Guinier camera,
Analysis of the (PbF :YF ) product with an X-ray pow-
ꢁ
ꢃ
pair of pliers (rounded jaws, Fig. 1a). Thereafter, the loaded CuKa , 2 h) allowed unambiguous identification of the
ꢀ
tube was connected to a high-vacuum line and evacuated for superstructure phase Pb
2
Y
F
(x40.2) on the
G
$
$
ꢅ V ꢃ V ꢀꢄ V
h at 100°C and &10\ꢅ mbar. Finally, the copper tube basis of X-ray reflections corresponding to both the basis
was sealed a second time by squeezing. This yielded an structure and the superstructure (14). The (BaF :YF ) prod-
ampoule around 5 cm in length (Fig. 1b). Reaction was uct was analyzed with an X-ray powder diffractometer (Phi-
ꢁ
ꢃ
performed in a resistively heated furnace (nearly isother- lips PW1800, CuKa , int. stand.: LiF). Using data available
ꢀ
mal temperature distribution), with an argon-purged silica on the solid solution (Ba, ¸n)F
(27), Ba
Y
F
G
$
$
ꢁ
>V
ꢅ V ꢃ V ꢀꢄ V
tube protecting the copper against oxidation by atmo- (26), and Ba
Y
F
(20), we could assign the reflec-
G
$
$
ꢁ
V ꢀ V ꢄ V
spheric oxygen. The ampoules were kept for 48 h at 600 tions to (Ba, Y)F
(
(x+0.28) and Ba
Y
F
(x+
ꢁ
>V
ꢅ\V ꢃ>V ꢀꢄ>V
PbF :YF ) and 900°C (BaF :YF ), respectively. After 0.08) (Table 1). The latter exhibited weak superstructure and
ꢁ
ꢃ
ꢁ
ꢃ
cooling to room temperature, they were opened using a tube split basis-structure reflections, both being characteristic for
cutter. The ampoules did not show any sign of collapsing or the rhombohedral distorted fluorite-type superstructure.
corrosion. Additionally, they were successfully checked for Ba
Y
F
, however, was not obtained.
G
$
$
ꢁ
V ꢀ V ꢄ V

1
54
EGGER, PETERS, AND HULLIGER
TABLE 1
Ba
Y
F
(x"0.01) and (Ba, Y)F
(x"0.25) if
ꢅ\V ꢃ>V ꢀꢄ>V
ꢁ>V
Measured X-Ray Powder Diffraction Pattern of the cooled down rather than quenched below 945°C, Rao et al.
BaF :YF ) Product and Calculated Diffraction Patterns for
Ba4؊xY3؉xF17؉x [x؍0.08, Hexagonal, a؍11.055 A
26)] and (Ba, Y)F2؉x [x؍0.28, Cubic, c؍6.0086 A> (27)]
(
2 3
(21) obtained Ba YF by solid-state reaction of BaF and
YF at 900°C and subsequent furnace cooling to room
temperature. Our observations correspond to the former.
ꢁ
ꢄ
ꢁ
>
, c؍20.346 A
>
a
ꢃ
(
Powders of complex lanthanide fluorides can be obtained
in a solid-state reaction using evacuated and sealed copper
ampoules as reactors. We have prepared vacuum-tight am-
poules providing up to 4-cmꢃ reaction volume. The high
ductility of the copper enables cold welding, leading to the
formation of a joint which is high-vacuum tight (10\ꢇ mbar)
up to 900°C at least. Tightness is also maintained when an
internal pressure of 5 bar is applied at 500°C for several
hours. This time might be extended by applying a counter-
pressure to prevent swelling of the joint. Furthermore, some
problems can arise from copper recrystallization, leading to
considerable grain growth and the formation of leaks in the
joint region. This limits the use of evacuated copper am-
poules at elevated temperatures to around 48 h, which,
however, is adequate for most solid-state reactions. As for
the rest, the joint is found to be mechanically robust (shock
resistant) and remains tight even after repeated quenching in
liquid nitrogen or water.
(
BaF :YF ) product
Ba
Y
F
(26) (Ba, Y)F
(27)
ꢁ
ꢃ
ꢅ\V ꢃ>V ꢀꢄ>V
ꢁ>V
dꢀ"ꢁ (A
>
)
Intensity (%)
d#!ꢂ# (A
>
)
hkl
d#!ꢂ# (A
>
)
hkl
4
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
.2875
2
3
7
100
75
32
10
21
29
1
1
1
2
3
2
6
3
4
37
33
29
10
1
2
1
21
12
15
23
(s) 4.2847
(s) 3.7450
(s) 3.5627
113
015
211
.7434
.5625
.4696
.4114
.3865
.1647
.0036
.9488
.7602
.7030
.6328
.5613
.4844
.3784
.2663
.2250
.1772
.1249
.0868
.0806
.0070
.9576
.9324
.9075
.8132
.8026
.7840
.7789
3.4691
3.0043
111
200
(b) 3.4093
(b) 3.3910
*
122
006
(b) 2.9486
(s) 2.7638
(s) 2.7041
(s) 2.6330
(s) 2.5594
(s) 2.4845
(s) 2.3771
(s) 2.2661
(s) 2.2238
(s) 2.1837
214
220
125
131
223
027
401
217
315
321
Focusing on the chemical properties of copper, this turns
out to be both stable and inert during synthesis. Though
D e´ ne` s (17) does not report any contamination of the prod-
uct with copper, it appears advisable to prevent direct
contact between the solid fluorides and the copper by the
use of platinum or glassy carbon crucibles. The latter have
proved their inertness in single-crystal melt-growth experi-
2.1244
1.8117
220
311
(b) 2.0892
(b) 2.0807
(s) 1.9966
(s) 1.9604
(s) 1.9328
(s) 1.9064
410
128
413
137
235
051
ments of LiYF (28). Furthermore, emphasis has to be laid
(s) 1.8022
(b) 1.7813
(b) 1.7787
241
422
416
ꢅ
on the elimination of traces of oxygen and water, which can
attack the copper as well as reactants and products. This
calls for the use of perfectly dry reactants, with heating of the
loaded and evacuated tubes being necessary prior to sealing.
In conclusion, simple cold welding with a special pair of
pliers yields tightly sealed copper ampoules and can be
?
(s), superstructure reflection; (b), basis-structure reflection; *, not identi-
fied.
With the synthetic procedure mentioned above, Erꢃ>- applied on a vacuum line to evacuate and/or fill ampoules
doped samples of (Ba, Y)F
(x+0.28)/Ba
Y
F
with a gaseous reactant (halogenization) prior to sealing.
ꢁ
>V
ꢅ\V ꢃ>V ꢀꢄ>V
(
x+0.08) and Pb G
Y
F
(x40.2), together with Alternatively, sealing may be done in a glove box under an
$
$
ꢅ V ꢃ V ꢀꢄ V
BaY F for comparison, were prepared and their UC effi- inert gas atmosphere. The ampoules can be used as reactors
ꢁ
ꢂ
ciency determined on an UC powder test setup (11). for the solid-state synthesis of oxygen-sensitive and/or
BaY F :10% Erꢃ>, excited with laser light of &980-nm moisture-sensitive compounds. Furthermore, they allow
ꢁ
ꢂ
wavelength, turned out to be the strongest UC system by up quenching after reaction, which might be necessary to ob-
to one order of magnitude. As for the rest, no significant tain metastable phases. In summary, this new ampoule
influence of the host structure type or the composition could technique represents a valuable tool for solid-state chemists.
be found. However, considerable impact could be attributed
to the Erꢃ> concentration (optimum at &10 mol%) and the
ACKNOWLEDGMENTS
excitation wavelength (being best at 980 nm).
We thank G. Rizzoli and T. Peters for the measurement of the X-ray
powder diffractograms as well as L. Diamond and M. Engi for the use of
the electron microprobe. This work was supported by the Priority Pro-
DISCUSSION AND CONCLUSIONS
Whereas Kieser and Greis (20) found Ba YF to be meta-
gram ‘‘Optics II’’ of the Swiss Board of the Federal Institutes of Technology
ꢁ
ꢄ
stable at room temperature and to disproportionate into (Project No. 232).

REACTIONS IN EVACUATED COPPER AMPOULES
155
REFERENCES
14. A. Dib, S. Ale
1984).
5. D. Kissel and R. Hoppe, Z. Anorg. Allg. Chem. 557, 161 (1988).
16. J. Scheffler and R. Hoppe, Z. Anorg. Allg. Chem. 521, 79 (1985).
. M. C. Brierley, J. F. Massicott, T. J. Whitley, C. A. Millar, R. Wyatt, 17. G. Denes, J. Solid State Chem. 77, 54 (1988).
18. ‘‘Gmelin Handbuch der Anorganischen Chemie,’’ 8th ed., Vol. 60,
´ onard, and M. T. Roux, J. Solid State Chem. 52, 292
(
1
1
2
. F. Auzel, J. ¸umin. 45, 341 (1990).
´
`
S. T. Davey, and D. Szebesta, B¹ ¹echnol. J. 11, 128 (1993).
. R. M. Macfarlane, J. Phys. I» (Suppl.) 4, C4-289 (1994).
. J. Ohwaki and M. Otsuka, Electron. ¸ett. 31, 752 (1995).
. J. Ohwaki and Y. Wang, Jpn. J. Appl. Phys. 33, L334 (1994).
. J. Ohwaki and Y. Wang, Appl. Phys. ¸ett. 65, 129 (1994).
. J. Ohwaki and Y. Wang, Electron. ¸ett. 29, 351 (1993).
3
4
5
6
7
8
Part A, p. 1230/4. VCH, Weinheim/New York, 1955.
19. L. F. Johnson and H. J. Guggenheim, Appl. Phys. ¸ett. 19, 44 (1971).
20. M. Kieser and O. Greis, J. ¸ess-Common Met. 71, 63 (1980).
21. U. R. K. Rao, A. K. Tyagi, K. V. Muralidharan, and R. M. Iyer,
J. Mater. Sci. ¸ett. 11, 435 (1992).
. M. P. Hehlen, K. Kra
¨
mer, H. U. Gu
¨
del, R. A. McFarlane, and R. N. 22. J. M. Re
´
au, P. P. Fedorov, L. Rabardel, S. F. Matar, and P. Hagenmul-
Schwartz, Phys. Rev. B 49, 12475 (1994).
ler, Mater. Res. Bull. 18, 1235 (1983).
9
. L. A. Riseberg and H. W. Moos, Phys. Rev. 174, 429 (1968).
23. P. Egger, P. Rogin, T. Riedener, H. U. Gu
¨
del, M. S. Wickleder, and
1
1
1
1
0. J. M. F. v. Dijk and M. F. H. Schuurmans, J. Chem. Phys. 78, 5317
1983).
1. P. Egger, P. Rogin, H. U. Gu
Sol. (b) (submitted).
2. ‘‘Gmelin Handbuch der Anorganischen Chemie,’’ 8th ed., Vol. 39, Part
C3, p. 109. Springer-Verlag, Berlin/New York, 1976.
J. Hulliger, Adv. Mater. 8, 668 (1996).
24. T. Riedener, P. Egger, J. Hulliger, and H. U. Gu
(
¨
del, Phys. Rev. B 56,
¨
del, N. Peters, and J. Hulliger, Phys. Stat.
1800 (1997).
25. N. L. Tkachenko, L. S. Garashina, O. E. Izotova, V. B. Aleksandrov,
and B. P. Sobolev, J. Solid State Chem. 8, 213 (1973).
26. M. Kieser and O. Greis, Z. Anorg. Allg. Chem. 469, 164 (1980).
3. J. Grannec and L. Lozano, ‘‘Inorganic Solid Fluorides’’ (P. Hagenmul- 27. B. P. Sobolev and N. L. Tkachenko, J. ¸ess-Common Met. 85, 155 (1982).
ler, Ed.), p. 43. Academic Press, San Diego, 1985. 28. P. Rogin and J. Hulliger, J. Cryst. Growth 172, 200 (1997).