Comparative study for stepwise chlorination-chemical vapor transport characteristics of pure rare earth oxides from Sc2O3 to Lu2O3 mediated by the vapor complexes LnAlnCl3n+3

Article abstract of DOI:10.1016/S0925-8388(99)00042-0

A systematic study has been carried out for the stepwise chlorination-chemical vapor transport (SC-CVT) characteristics of pure rare earth oxides from Sc₂O₃ to Lu₂O₃ mediated by the vapor complexes LnAl_nCl_3n+3 (where Ln = rare earth elements) under identical conditions. The results show that the total transported amount of the produced chlorides is the highest for YCl₃; it is also high for ScCl₃, but low for LaCl₃, and then increases systematically from the early lanthanide chlorides to the end lanthanide chlorides except CeCl₃, EuCl₃ and GdCl₃. The results also show that the main deposition temperature of the chlorides is in the increasing order ScCl₃ < YCl₃ < LaCl₃ and then with a systematically decreasing trend from the early lanthanide chlorides to the end lanthanide chlorides. Based on the literature data for the solid complexes LnAl₃Cl₁₂ and LnAl₃Br₁₂, a similar coordination structure assumption is introduced for the vapor complexes LnAl₃Cl_3n+3 from Ln = Sc to Ln = Lu with the same stoichiometry to explain the SC-CVT characteristics of the pure rare earth oxides from Sc₂O₃ to Lu₂O₃ using a 4f electron hybridization assumption.

Full text of DOI:10.1016/S0925-8388(99)00042-0

Journal of Alloys and Compounds 287 (1999) 109–113
L
A comparative study for stepwise chlorinatxion–chemical vapor transport
characteristics of pure rare earth oxides from Sc2O3 to Lu2O3 mediated by
the vapor complexes LnAlnCl3n13
Zhi-Chang Wanga, ,1, Yan-Hui Suna, Lei Guob
*
aDepartment of Chemistry, Northeastern University, Shenyang 110006, Liaoning, People’s Republic of China
bInstitute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110015, Liaoning, People’s Republic of China
Received 10 January 1999; received in revised form 24 January 1999
Abstract
A systematic study has been carried out for the stepwise chlorination–chemical vapor transport (SC–CVT) characteristics of pure rare
earth oxides from Sc2O3 to Lu2O3 mediated by the vapor complexes LnAlnCl3n13 (where Ln5rare earth elements) under identical
conditions. The results show that the total transported amount of the produced chlorides is the highest for YCl3; it is also high for ScCl3,
but low for LaCl3, and then increases systematically from the early lanthanide chlorides to the end lanthanide chlorides except CeCl3,
EuCl3 and GdCl3. The results also show that the main deposition temperature of the chlorides is in the increasing order ScCl3,YCl3,
LaCl3 and then with a systematically decreasing trend from the early lanthanide chlorides to the end lanthanide chlorides. Based on the
literature data for the solid complexes LnAl3Cl12 and LnAl3Br12, a similar coordination structure assumption is introduced for the vapor
complexes LnAl3Cl3n13 from Ln5Sc to Ln5Lu with the same stoichiometry to explain the SC–CVT characteristics of the pure rare earth
oxides from Sc2O3 to Lu2O3 using a 4f electron hybridization assumption.
1999 Elsevier Science S.A. All rights reserved.
Keywords: Stepwise chlorination–chemical vapor transport reaction; Pure rare earth oxides from Sc2O3 to Lu2O3; Rare earth vapor complexes;
LnAlnCl3n13
1. Introduction
dynamic measurements for the vapor complexes
LnAlnCl3n13 from Ln5Sc to Ln5Lu [7–10].
Systematic analysis has been carefully carried out for
physical and chemical properties of rare earth elements,
compounds and complexes in the solid and liquid states
since the early 1950s. In particular, Gschneidner [1,2]
recently represented them as functions of metallic or ionic
radium from Sc to Lu and introduced a 4f electron
hybridization assumption to explain the systematics and
anomalies. Until now, however, a similar systematic
analysis has not been carried out for rare earth vapor
compounds or complexes [3–5] due to the lack of a
complete set of experimental data for all of the rare earth
elements, except Adachi’s chemical vapor transport (CVT)
investigations for pure rare earth chlorides from LaCl3 to
LuCl3 mediated by the vapor complexes LnAlnCl3n13
(where Ln5rare earth elements) [6] and Wang’s thermo-
Rare earth vapor complexes LnAlnCl3n13 have been
used as intermediate materials for the preparation of
anhydrous rare earth chlorides in high purities [7–16], for
mutual separation of rare earth elements [6,17–26], and for
recovery of rare earth metals [27,28] by using the CVT
reactions. Adachi’s CVT investigations [6] have showed a
systematic increase in the total transported amount of the
reproduced chlorides from LaCl3 to LuCl3 except EuCl3
within 1 h. Our thermodynamic measurements [7–10] have
suggested the predominant vapor complexes to be
LnAl2Cl9 for Ln5Sc and Ln5Y but LnAl3Cl12 from
Ln5La to Ln5Lu together with a Gd divergence for both
standard enthalpy and standard entropy, roughly in the
temperature range 500–800 K and pressure range 0.01–
0.22 MPa.
On the other hand, both Wang’s [22,23] and Adachi’s
[3,20,21] groups have carried out comparative studies on
the CVT characteristics between rare earth oxide mixtures
and rare earth chloride mixtures with AlCl3 under identical
conditions and observed more efficient CVT results from
*Corresponding author.
Present address: Max-Planck-Institut fur Metallforschung, Pulver-
metallurgisches Laboratorium, Heisenbergstrasse 5, D-70569 Stuttgart,
Germany.
1
¨
0925-8388/99/$ – see front matter
PII: S0925-8388(99)00042-0
1999 Elsevier Science S.A. All rights reserved.

110
Z. Wang et al. / Journal of Alloys and Compounds 287 (1999) 109–113
the oxide mixtures than from the chloride mixtures, where,
in the former case, the additional carbothermic reduction–
chlorination subreaction of the oxides and the CVT
subreaction of the chlorides produced are realized almost
at the same time and under the same conditions. Very
recently, we [24,25] have further developed a stepwise
chlorination–chemical vapor transport (SC–CVT) reaction
for rare earth oxide mixtures, which allows both chlorina-
tion and CVT subreactions to be realized under different
conditions, resulting in a further improvement for the
separation efficiency between neighboring rare earth ele-
ments due to an additional selective chlorination of the
oxide mixtures. In this study, we tried to make a compara-
tive study for the SC–CVT characteristics of pure rare
earth oxides from Sc2O3 to Lu2O3 mediated by the vapor
complexes LnAlnCl3n13 under identical conditions.
2. Experimental details
The chemicals used in this study were of 98.0% purity
for anhydrous AlCl3 and $99.9% purity for all the rare
earth oxides. Anhydrous AlCl3 was further purified by
careful sublimation under vacuum. A raw mixture was
formed by mixing active carbon with a pure rare earth
oxide at an atomic ratio of C–Ln53:1.
The SC–CVT reaction was carried out in a cylindrical
alumina reactor tube, 25 mm inner diameter and 1000 mm
length, with a given temperature and pressure gradient as
described in the previous papers [24–26]. Let T denote the
highest temperature in the tube reactor, where the raw
material placed. The raw material was chlorinated by dry
Cl2 gas with a flow-rate of 20 cm3 min21 at T5800 K for
2 h. Within the temperature range of T5800–1300 K, the
Cl2 gas was replaced by a dry CO–HCl mixed gas with the
flow-rates of 40 and 20 cm3 min21, respectively. The rare
earth chlorides produced were then transported with AlCl3
at T51300 K for 6 h, using a dry CO gas with a flow-rate
of 40 cm3 min21 as carried gas. During the whole reaction,
a given pressure gradient was maintained by a subatmos-
pheric pressure of 2.7 kPa at the outlet of the tube reactor.
At the end of each run, the amounts of the rare earth
chloride produced were determined from the peak intensity
Fig. 1. Distribution of rare earth chlorides deposits in a SC–CVT
reaction formed from their pure oxides mediated by the vapor complexes
LnAlnCl3n13: (A) Sc2O3, (B) Y2O3, (C) La2O3, (D) CeO2, (E) Pr6O11
,
(F) Nd2O3, (G) Sm2O3, (H) Eu2O3, (I) Gd2O3, (J) Tb4O7, (K) Dy2O3,
(L) Ho2O3, (M) Er2O3, (N) Tm2O3, (O) Yb2O3, and (P) Lu2O3.
deposition profiles for the rare earth chlorides produced vs.
fraction numbers (FN) of the receptors, where the tempera-
ture was 1300 K at FN51 and 750 K at FN515,
respectively, at T51300 K. Fig. 2 shows the total trans-
ported amount of LnCl3 vs. atomic number from Ln5La to
of the characteristic bands [29]: 361.395 nm for Sc31
324.228 nm for Y31, 408.672 nm for La31, 418.660 nm
for Ce31, 414.311 nm for Pr31, 430.358 nm for Nd31
442.434 nm for Sm31,412.970 nm for Eu31, 342.247 nm
,
,
for Gd31, 350.917 nm for Tb31, 353.170 nm for Dy31
,
339.898 nm for Ho31, 337.271 nm for Er31, 313.126 nm
for Tm31, 328.937 nm for Yb31 and 261.542 nm for Lu31
on an inductively coupled plasma atomic emission spec-
trometry (Perkin Elmer, Optima 3000).
3. Results and discussion
Fig. 2. Relationship between the total transported amount of rare earth
chlorides in a SC–CVT reaction, formed from their pure oxides mediated
by the vapor complexes LnAlnCl3n13, and atomic number from Ln5La to
Ln5Lu.
Fig. 1 (A–P) show the SC–CVT reaction results of the
pure rare earth oxides from Sc2O3 to Lu2O3 in the form of

Z. Wang et al. / Journal of Alloys and Compounds 287 (1999) 109–113
111
CeO2(s) 1 (3/2)Cl2(g) 1 2C(s) 5 CeCl3(s) 1 2CO(g) (2)
(1/6)Pr6O11(s) 1 (3/2)Cl2(g) 1 (11/6)C(s) 5 PrCl3(s)
1 (11/6)CO(g)
(3)
(1/4)Tb4O7(s) 1 (3/2)Cl2(g) 1 (7/4)C(s) 5 TbCl3(s)
1 (7/4)CO(g)
(4)
and by the complexation subreaction of rare earth chlorides
at T51300 K:
LnCl3(l) 1 (n/2)Al2Cl6(g) 5 LnAlnCl3n13(g)
(Ln 5 Sc, Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Th, Dy, Ho, Er,
Tm, Yb and Lu)
(5)
Here, the very low total transported amount of CeCl3 and
EuCl3 might be described by the use of CeO2 as starting
material, whose chlorination ability is very low at T5800
K [30], and by the lower stability of EuCl3 than EuCl2 at
1300 K and the very low complexation ability of EuCl2
with AlCl3 [31], respectively. While the relatively lower
total transported amount of ScCl3 than YCl3 might be
explained by the very low chlorination ability of Sc2O3 as
compared to Y2O3 at T5800 K [30]. On the other hand,
the high total transported amount of ScCl3 and YCl3 might
be explained by the higher thermodynamic stability of the
vapor complexes ScAlnCl3n13 and YAlnCl3n13 [7], the
slightly low total transported amount of GdCl3 might be
explained by the Gd divergence shown in the thermo-
dynamic behavior of the vapor complexes LnAlnCl3n13
[9], and the systematic increase in the total transported
amount from the early lanthanide chlorides to the end
lanthanide chlorides would indicate the systematic increase
in the thermodynamic stability of the predominant vapor
complexes LnAlnCl3n13 from early lanthanides to end
lanthanides at 1300 K. The distribution characteristics of
LnCl3 in the reactor tube would mainly be determined by
the total transported amount of LnCl3 at 1300 K and by the
complicated CVT subreaction along the temperature gra-
dient, where the predominant vapor complexes may be
LnAlCl6 and LnAl2Cl9 in the higher temperature range but
LnAl2Cl9, LnAl3Cl12 and LnAl4Cl15 in the lower tempera-
ture range, while the produced chlorides are in the liquid
state at the higher temperatures but in the solid state at the
lower temperatures. The large deposition in the lower
temperature range for ScCl3 and in the middle temperature
range for YCl3 would indicate the high stability in the
higher and middle temperature ranges for the vapor
complexes ScAlnCl3n13 and in the higher temperature
range for the vapor complexes YAlnCl3n13, respectively.
The large deposition of early lanthanide chlorides in the
higher temperature range would indicate the low stability
of the early lanthanide vapor complexes LnAlnCl3n13 even
at the higher temperature. While the large deposition of end
lanthanide chlorides in the lower temperature range would
indicate the higher stability of the end lanthanide vapor
Fig. 3. Relationship between the total transported amount of rare earth
chlorides in a SC–CVT reaction, formed from their pure oxides mediated
by the vapor complexes LnAlnCl3n13, and ionic radius of the trivalent
rare earth ion Ln31 with a six-fold coordination number from Ln5Sc to
Ln5Lu.
Ln5Lu. In order to represent the systematics and
anomalies in the SC–CVT characteristics of pure rare earth
oxides from Sc2O3 to Lu2O3 more clearly, Fig. 3 shows
the total transported amount of LnCl3 vs. ionic radius of
Ln31 with a six-fold coordination number from Ln5Sc to
Ln5Lu though the actual structural information is lacking
for the predominant vapor complexes LnAlnCl3n13 at 1300
K. It can be seen from Fig. 1 that the produced ScCl3
concentrates in the lower temperature range, YCl3 mainly
deposits in the middle temperature range, while LaCl3
enriches in the higher temperature range, and then the
deposition temperature decreases systematically from the
early lanthanides to the end lanthanides. It can also be seen
from Figs. 2 and 3 that the total transported amount is the
largest for YCl3, also high for ScCl3, but low for LaCl3,
and then increases systematically from the early lanthanide
chlorides to the end lanthanide chlorides except CeCl3,
EuCl3 and GdCl3. It would be interesting to point out that
the plot for the total transported amount of LnCl3 vs. ionic
radium in Fig. 3 is similar to those for various physical
arid chemical properties of rare earth elements, compounds
and complexes in solid and liquid states vs. metallic or
ionic radium shown in [1].
The total transported amount of LnCl3 would mainly be
determined by the chlorination subreaction of rare earth
oxides at T5800 K:
(1/2)Ln2O3(s) 1 (3/2)Cl2(g) 1 (3/2)C(s) 5 LnCl3(s)
1 (3/2)CO(g) (Ln 5 Sc, Y, La,
Nd, Sm, Eu, Gd, Dy, Ho, Er, Tm, Yb and Lu)
(1)

112
Z. Wang et al. / Journal of Alloys and Compounds 287 (1999) 109–113
complexes LnAlnCl3n13 in the higher and middle tempera-
ture ranges. These arguments would be supported by the
limited literature thermodynamic data [4,5,7–13,32–35].
It would be interesting to point that the plot of the total
transported amount of LnCl3 vs. atomic number from
Ln5La to Ln5Lu in Fig. 2 is very similar to that reported
by Adachi’s group [5,6] only except for Ln5Ce and Gd.
Here, the small difference in CVT characteristics for Ln5
Gd might be related to the difference in starting materials
and reaction conditions, while the large difference for
Ln5Ce, by the very low chlorination ability of CeO2 at
T5800 K in this study. Moreover, the deposition profiles
of ScCl3, YCl3 and LaCl3 vs. fraction number of the
receptors in Fig. 1 (A–C) are also somewhat similar to
those reported by Adachi’s group [20], who determined the
CVT characteristics of pure rare earth oxides Sc2O3, Y2O3
experiments for various physical and chemical properties
of rare earth elements, compounds and complexes in solid
and liquid states by using his 4f electron hybridization
assumption. This assumption is also supported by the
recent theoretical calculation of Temmerman et al. [38],
who found the unoccupied 4f bands in Pr metal to
hybridize strongly with the conduction s, p and d bands.
However, it would be required to keep the same coordina-
tion structure for one particular kind of rare earths from
Ln5Sc to Ln5Lu when using the 4f electron hybridiza-
tion assumption [1,2]. It would seem that a rare earth solid
complex may reasonably be assumed to have the same
coordination structure as its corresponding vapor complex.
Therefore, the rare earth vapor complexes LnAl3Cl12 from
Ln5Sc to Ln5Lu might be assumed to have the same
coordination structure if further experiments of the solid
complexes LnAl3Cl12 (or LnAl3Br12) for the end lanth-
anides show the same microstructure as those for the early
lanthanides. If the same coordination structure assumption
for LnAl3Cl12 from Ln5Sc to Ln5Lu might further be
extended to the other LnAlnCl3n13 (or LnAlnBr3n13) with
the same stoichiometry, then the 4f electron hybridization
assumption [1,2] might more reasonably be used to explain
the systematics and anomalies in the rare earth vapor
complexes LnAlnCl3n13 from Ln5Sc to Ln5Lu. We [9]
have tried to use the 4f electron hybridization assumption
to explain both the difference in the stoichiometry of the
predominant vapor complexes LnAl2Cl9 for Ln5Sc and
Ln5Y and LnAl3Cl12 from Ln5La to Ln5Lu and the Gd
divergence from the behavior of the vapor complexes
LnAl3Cl12 from Ln5La to Ln5Lu in roughly the same
temperature and pressure ranges (i.e. 500–800 K and
0.01–0.22 MPa). Adachi et al. [20] have used the 4f
electron hybridization assumption to explain CVT sepa-
and La2O3 mediated by the vapor complexes LnAlnCl3n13
,
except the total transported amount order: ScCl3 .YCl3 .
LaCl3 in [20] and ScCl3 ,YCl3 .LaCl3 in this study. That
might be described by the great difference in the chlorina-
tion temperatures: 1300 K in [20] and 800 K in this study.
Let K1(Ln) denote the equilibrium constant of reaction (1)
for the rare earth element Ln. The value of equilibrium
constant ratio hK1(Sc)/K1(Y)j is 1.431024 at 1300 K but
2.131026 at 800 K [31]. Besides, the total transported
amount order of YCl3 .ErCl3 .HoCl3 .DyCl3 shown in
Fig. 3 is somewhat different from ErCl3 .YCl3 .HoCl3 .
DyCl3 reported by Adachi’s group [21], who determined
the CVT characteristics for pure rare earth chlorides YCl3,
DyCl3, HoCl3 and ErCl3 mediated by the vapor complexes
LnAlnCl3n13. This would further indicate the great effect
of the starting materials and reaction conditions on the
CVT efficiency, which has been shown in the literature
[5,6,17–26].
Sc31 is the smallest rare earth trivalent ion but Y31 is at
least larger than Tm31, Yb31, and Lu31. However, the
SC–CVT determinations reported in this study show both
the difference between Sc2O3 and Y2O3 on the one hand
and the lanthanide oxides on the other hand, and a
systematic trend from La2O3 to Lu2O3. Therefore, the rare
earth ionic radii are not the decisive factors for the SC–
CVT characteristics of the pure rare earth oxides from
Sc2O3 to Lu2O3. Moreover, scientists have had different
opinions on the nature of the rare earth complexes even in
the condensed states. The Gd divergence effect in the rare
earth liquid complexes, for example, has been explained by
a change either in the 4f electron configuration or in the
coordination number [36]. Similar to the second explana-
tion, different coordination structures have been assumed
for early and end lanthanides in the vapor complexes
LnAl3Cl12 [3,4,11]. However, recent experiments found
the same microstructure for the solid complexes DyAl3Cl12
[16], HoAl3Cl12 [15], LaAl3Br12 [37], PrAl3Br12 [37] and
NdAl3Br12 [37], which were produced from the corre-
sponding vapor complexes.
ration results for the binary oxide mixtures Ln2O3 –Ln2O3
9
(Ln±Ln95Sc, Y, La). Similarly, we may also use the 4f
electron hybridization assumption to explain the difference
in the SC–CVT behavior of the pure rare earth oxides
Sc2O3 and Y2O3 on the one hand and from La2O3 to
Lu2O3 on the other hand, as well as the systematics from
La2O3 to Lu2O3 together with anomalies for CeO2, Eu2O3
and Gd2O3 reported in this study. Therefore, further
experimental and theoretical studies will be very interest-
ing on the microstructures of the rare earth solid and vapor
complexes LnAlnCl3n13 (or LnAlnBr3n13) to have a deeper
understanding of the mechanisms of the systematics and
anomalies both in the thermodynamic properties of the
vapor complexes LnAlnCl3n13 and in the SC–CVT prop-
erties of the rare earth oxides mediated by the vapor
complexes LnAlnCl3n13 from Ln5Sc to Ln5Lu.
4. Conclusions
This paper presents a complete set of experimental data
for SC–CVT characteristics of pure rare earth oxides from
On the other hand, Gschneidner [1,2] has explained the

Z. Wang et al. / Journal of Alloys and Compounds 287 (1999) 109–113
113
¨
[13] G. Steidl, F. Dienstbach, K. Bachmann, Polyhedron 2 (1983) 727.
[14] H. Gunsilius, W. Urland, R. Kremer, Z. Anorg. Allg. Chem. 550
(1987) 35.
Sc2O3 to Lu2O3 mediated by the vapor complexes
LnAlnCl3n13. The results show the difference in Ln5Sc
and Ln5Y on the one hand and from Ln5La to Ln5Lu
on the other hand as well as the systematics from Ln5La
to Ln5Lu together with anomalies for Ln5Ce, Eu and
Gd.
[15] D. Hake, W. Urland, Angew. Chem. 101 (1989) 1416.
[16] D. Hake, W. Urland, Z. Anorg. Allg. Chem. 586 (1990) 99.
[17] G. Adachi, S. Shinozaki, Y. Hirashima, K. Machida, J. Less-Com-
mon Met. 169 (1991) L1.
[18] G. Adachi, K. Murase, S. Shinozaki, K. Machida, Chem. Lett.
(1992) 511.
[19] K. Murase, S. Shinozaki, K. Machida, G. Adachi, Bull. Chem. Soc.
Jpn. 65 (1992) 2724.
Acknowledgements
[20] J. Jiang, T. Ozaki, K. Machida, G. Adachi, J. Alloys Comp. 264
(1998) 157.
[21] T. Ozaki, J. Jiang, K. Murase, K. Machida, G. Adachi, J. Alloys
Comp. 265 (1998) 125.
[22] Z.-C. Wang, J. Yu, Y.-L. Yu, Bull. Chem. Soc. Jpn. 69 (1996) 2369.
[23] Z.-C. Wang, J. Yu, Y.-L. Yu, Y.-H. Sun, J. Alloys Comp. 264 (1998)
147.
This work was supported by the National Natural
Science Foundation of China (Grant No. 59574020) and
the Ministry of Metallurgical Industry.
References
[24] Z.-C. Wang, Y.-H. Sun, Chem. Lett. (1997) 1113.
[25] Y.-H. Sun, Z.-C. Wang, L. Guo, J. Alloys Comp. 269 (1998) 88.
[26] Y.-H. Sun, Z.-C. Wang, L. Guo, J. Alloys Comp., in press.
[27] K. Murase, K. Machida, G. Adachi, Chem. Lett. (1994) 1555.
[28] K. Murase, K. Machida, G. Adachi, J. Alloys Comp. 217 (1995)
218.
[29] R.K. Winge, V.J. Peterson, V.A. Fassel, Appl. Spectr. 33 (1979) 206.
[30] O. Knacke, O. Kubaschewski, K. Hesselmann, in: Thermochemical
Properties of Inorganic Substances, 2nd ed., Springer-Verlag, Berlin,
1991.
[1] K.A. Gschneidner Jr., J. Alloys Comp. 192 (1993) 1.
[2] K.A. Gschneidner Jr., J. Alloys Comp. 223 (1995) 165.
[3] G.N. Papatheodorou, in: E. Kaldis (Ed.), Current Topics in Materials
Science, Vol. 10, North Holland Publishing Co, New York, 1982, p.
249.
[4] S. Boghosian, G.N. Papatheodorou, in: K.A. GschneidnerJr., L.
Eyring (Eds.), Handbook On the Physics and Chemistry of Rare
Earths, Vol. 23, Elsevier Science B.V, Amsterdam, 1996, p. 435.
[5] J. Jiang, T. Ozaki, K. Machida, G. Adachi, J. Alloys Comp. 260
(1997) 222.
[31] M. Sørlie, H.A. Øye, J. Inorg. Nucl. Chem. 91 (1978) 493.
[32] H.A. Øye, M. Gruen, J. Am. Chem. Soc. 91 (1969) 2229.
[33] M. Cosandey, F.P. Emmenegger, J. Electrochem. Soc. 126 (1979)
1601.
[6] K. Murase, S. Shinozaki, Y. Hirashima, K. Machida, G. Adachi, J.
Alloys Comp. 198 (1993) 31.
¨
¨
[34] H. Schafer, U. Florke, Z. Anorg. Allg. Chem. 479 (1981) 89.
[35] H. Opperrmann, D.Q. Huong, Z. Anorg. Allg. Chem. 621 (1995)
659.
[7] Z.-C. Wang, L.-S. Wang, R.-J. Gao, Y. Su, J. Chem. Soc. Faraday
Trans. 92 (1996) 1887.
[8] L.-S. Wang, R.-J. Gao, Y. Su, Z.-C. Wang, J. Chem. Thermodyn. 28
(1996) 1093.
[9] Z.-C. Wang, L.-S. Wang, Inorg. Chem. 36 (1997) 1536.
[10] Z.-C. Wang, L.-S. Wang, J. Alloys Comp. 265 (1998) 153.
[11] G.N. Papatheodorou, G.H. Kucera, Inorg. Chem. 18 (1979) 386.
[36] H. Kanno, J. Alloys Comp. 192 (1993) 271.
[37] D. Hake, W. Urland, Z. Anorg. Allg. Chem. 613 (1992) 45.
[38] W.M. Temmerman, Z. Szotek, W. Winter, Phys. Rev. B 47 (1993)
1184.
¨
[12] G. Steidl, K. Bachmann, F. Dienstbach, J. Phys. Chem. 87 (1983)
5010.