вход по аккаунту


Electrochemical Metallization of Solid Terbium Oxide.

код для вставкиСкачать
DOI: 10.1002/ange.200503571
Electrochemical Metallization of Solid Terbium
Dihua Wang, Guohong Qiu, Xianbo Jin, Xiaohong Hu,
and George Z. Chen*
Heavy rare-earth (HRE) metals function very efficiently in
advanced materials for magnetooptical-information storage,
giant magnetostrictive energy conversion, and high-strength
permanent magnets.[1] A 7 % annual growth in global demand
for rare-earth (RE) metals was recently predicted.[2]
Although the separation of mixed RE compounds has been
greatly refined,[3] extraction of pure RE metals, especially the
heavy ones, remains a historical challenge. The difficulty is
multifold but particularly related to the high oxygen affinity
of HRE metals. Current industrial methods convert HRE
oxides into fluorides (e.g. by reaction with HF or NH4F), from
which the metal is extracted by calciothermic reduction at
1500 8C in argon.[4] The obtained HRE metal needs further
separation from excess calcium under vacuum and deoxygenation by special methods.[5] HRE metals have high melting
points (> 1350 8C), and hence cannot be extracted as a liquid
in the same way as aluminum and light RE metals.[6]
Electrodeposited solid reactive metals from molten salts are
always dendritic and absorb oxygen easily when exposed to
air. Dendrite formation may be avoided by deposition onto a
suitable transition-metal substrate to form a liquid alloy,[7] but
the alloy composition is difficult to control precisely.
Recently, solid TiO2,[8] SiO2,[9] and some mixed oxides[10]
were electrochemically reduced to the respective solid metals
or alloys in molten salts. The new process requires that 1) the
metal oxides to be reduced are thermodynamically less stable
than the oxide of the metal element of the molten salt
used,[8–13] and 2) electrodeposition of the metal of the molten
salt should be avoided.[8] CaO is extremely stable, and hence
molten CaCl2 is the preferred choice for the new process.
However, HRE oxides are almost as stable as CaO, and their
reduction may not occur at potentials more positive than that
[*] Dr. D. H. Wang, G. H. Qiu, Dr. X. B. Jin, Dr. X. H. Hu,
Prof. Dr. G. Z. Chen
College of Chemistry and Molecular Sciences
Wuhan University
Wuhan, 430072 (P.R. China)
Fax: (+ 86) 276-875-6319
Prof. Dr. G. Z. Chen
School of Chemical, Environmental and Mining Engineering
University of Nottingham
Nottingham, NG7 2RD (UK)
Fax: (+ 44) 115-951-4115
[**] The authors thank the National Natural Science Foundation of
China for financial support (Grant Numbers: 50 374 052 and
20 125 308).
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 2444 –2448
of Ca deposition. Furthermore, in previous demonstrations,
the metal produced was separated from the solidified salt by
leaching in water, a method not applicable to the active HRE
metals.[8–10] Most importantly, for both academic and commercial interests, there are continual debates on the electroreduction mechanism of solid metal oxides (MOx) in molten
CaCl2. Electrodeposition of Ca was claimed to be necessary
for removal of oxygen from the oxide (Equations (1) and
(2);[12, 13] electrocalciothermic reduction).
Ca2þ þ 2 e Ð Ca
x Ca þ MOx Ð M þ x Ca2þ þ x O2 ðor x CaOÞ
Figure 1 a). In the CV of the oxide, in addition to the c1/a1
couple that is also present in that of the bare MCE, the
appearance of the c2/a2 couple demonstrates that the
electroreduction of Tb4O7 (actually Tb2O3 as discussed
later) continues up to potentials sufficiently negative for Ca
deposition. Similar CVs were recorded for Dy2O3.
If Ca atoms are formed through [Eq. (1)] and dissolve in
molten CaCl2, the two reduction mechanisms may be
thermodynamically indistinguishable. Therefore, they have
to be differentiated in other ways. For HRE oxides, the
standard potential difference between [Eq. (1)] and [Eq. (3)]
is very small (DEo = 21 mV and 34 mV for Tb2O3 and Dy2O3,
respectively, at 927 8C), which agrees with the CVs in
Figure 1 a. If the Ca atoms produced in [Eq. (1)] were partly
This statement differs from the original
patent claim of the oxygen-ionization mechanism (Equation (3);[8] direct electroreduction), and also contradicts later cyclic voltammetric findings for TiO2.[11]
MOx þ 2x e Ð M þ x O2
Herein we demonstrate, for the first time,
that the electroreduction of solid terbium
oxide to solid terbium metal can only proceed
under the conditions for Ca deposition and
predominantly through [Eq. (3)], with little
contribution, if at all, from [Eqs. (1) and (2)].
This electrochemical-metallization approach
differs in principle from all previous ones, and
can be applied to the extraction of other RE
metals, such as dysprosium and yttrium,
directly from their solid oxides.
Figure 1 a shows the cyclic voltammograms (CVs) of Tb4O7 powder (> 99.9 %
purity, particle size 2–3 mm) recorded in
molten CaCl2 at 850 8C by using a molybdenum cavity electrode (MCE).[14a] The molten
salt ( 50 g) was prepared from the
Figure 1. Electrochemistry in molten CaCl2. a) CVs of the MCE (inset) without (dashed line) and
CaCl2·2 H2O granules by thermal drying in
with (solid line) Tb4O7 powder. b) Current versus time plots of constant-voltage electrolysis of
air ( 300 8C), then in argon ( 600 8C). In
Tb4O7 pellets ( 2 g at 3.1–3.6 V; 4 g at 3.8–4.0 V; 6 g at 4.2 V). c) Correlations between
reduction depth and electrolysis time at indicated voltages for electrolysis of Tb4O7 and of mixed
both cases, the heating time was at least 4 h.
4O7 and NiO (Ni/Tb = 5). d) Correlations between reduction depth and cell voltage for 6 and
The thermally dried salt was heated to 850 8C
12 h of electrolysis. The inset shows the cross-section of a Tb4O7 pellet after electrolysis at 4.2 V for
for melting and underwent preelectrolysis at
6 h. The pellet was enclosed by a layer of Ca and Tb, but was only partly reduced inside.
about 2.6 V for more than 4 h to remove trace
moisture and other redox-active impurities
that might be present in the salt.[14] The MCE was constructed
or fully consumed in [Eq. (2)], one would expect a small or no
Ca-reoxidation current represented by a1. However, the total
by drilling two 0.5-mm-diameter through-holes (Figure 1 a
charge under a1 in the two CVs are approximately equal,
inset) in a 0.5-mm-thick molybdenum foil with a laser
which means that the Ca atoms produced by (1) must have
beam.[14a] It was filled with the oxide powder by repeated
contributed very little, if at all, to the reduction of the oxide
manual pressing, then inserted into the molten salt contained
through [Eq. (2)] on the voltammetric timescale. Therefore,
in a small graphite crucible (25-mm internal diameter, 250the oxide reduction, represented by c2, must have proceeded
mm height, 3-mm wall thickness), which also functioned as
predominantly through [Eq. (3)].
the counterelectrode.[14a] In the work reported herein, CVs
Constant-voltage electrolysis (two electrodes) of pellets of
were recorded by using either a Pt wire as a pseudo-reference
Tb4O7 powder ( 2.0 g, pressed at 4 MPa, and sintered at
electrode or a quartz-sealed Ag/AgCl reference electrode.[14a–c] Potential data were referenced to the current-onset
1000 8C for 2 h in air, 40–45 % porosity, 19.0-mm diameter,
2.0-mm thickness) was performed in 500 g or more of
potential for the reduction of Ca2+ ions (denoted Ca/Ca2+) in
molten CaCl2 in a large graphite crucible (90-mm internal
the CV of a bare MCE in the same molten salt (dashed line in
Angew. Chem. 2006, 118, 2444 –2448
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
diameter, 235-mm height, 10-mm wall thickness). The
molten salt was prepared as above. A graphite rod was used
as the anode (20-mm diameter, 30–50-mm immersion depth in
the molten salt during electrolysis).[8, 9, 14] The pellets were
sandwiched between two pieces of Mo mesh to form an
assembled cathode (Figure 2 a). XRD analysis of the sintered
Figure 2. Optical photographs. a)–d) Assembled cathode of a Tb4O7
pellet a) enclosed by Mo mesh/wires, b) after electrolysis, and crosssection after electrolysis c) at 3.1 V and 900 8C for 23 h and d) at 3.2 V
and 850 8C for 16 h; e)–g) Tb foil (2.5 mm thick) with six drilled holes
e) before oxidation, f) after oxidation in air and immersion in molten
CaCl2, showing all six holes filled with solidified salt, and g) after
oxidation in air and electrodeoxidation in molten CaCl2, showing five
of the six holes free of solidified salt.
pellet with CuKa1 radiation (l = 1.5406 A) at 40 kV, 40 mA,
and a 2 q scan rate of 4 deg min1 showed similar features to
the Tb4O7 standard (Figure 3 a). However, heating in argon
converted the pellet into Tb2O3 through the thermal decomposition reaction Tb4O7 Q 2 Tb2O3 + 1=2 O2. This conversion
was also evident in electrolysis experiments at voltages below
2.8 V, as confirmed by XRD analysis (Figure 3 b).
Figure 3. X-ray diffraction spectra. a) Sintered Tb4O7 pellet (1000 8C,
2 h); b) its electrolysis product at low voltage (2.5 V, 900 8C, 15 h,
molten CaCl2); c) yellow phase between the central Tb2O3 phase and
the surface metallic phase; d) surface metallic phase shown in
Figure 2 c and d; e)–h) HRE intermetallic compounds prepared by
electroreduction of the respective oxides: e) TbFe2 (3.1 V, 900 8C, 10 h),
f) Tb0.27Dy0.73Fe2 (3.1 V, 900 8C, 10 h), g) TbNi5 (3.1 V, 850 8C, 10 h,
washed in water), h) TbNi4Al (3.1 V, 850 8C, 10 h, washed in water).
All current–time plots of electrolysis at voltages greater
than 3.1 V exhibited an initial decline, followed by a plateau
(Figure 1 b). The current was larger at higher voltage, temperature, and/or oxide mass. Such behavior is similar to previous
findings in the potentiostatic electrolysis (three electrodes) of
transition-metal oxides, for example, Cr2O3, with the
MCE,[14a] and is believed to have resulted mainly from the
reaction kinetics and the state change on the oxide cathode.
Furthermore, because CaCl2 and CaO decompose at 3.2
and 2.7 V, respectively, at the working temperature (and
1.7 V for CaO if CO/CO2 forms at the graphite anode), the
observed current must have resulted from both electroreduction of the oxide and Ca deposition. The dependence of
the current plateau on voltage also suggests electronic
conduction through the molten salt.[15] Upon electrolysis, the
Tb4O7 pellet quickly turned metallic on the surface (Figure 2 b), and the interior changed from dark brown to white
(Figure 2 c). In some cases, a yellowish interlayer was seen
(Figure 2 c), which was confirmed by XRD analysis as TbOCl
(Figure 3 c) produced by the reaction Tb2O3 + CaCl2 Q
2 TbOCl + CaO. Surface metallization is consistent with the
reduction taking place at the metal j oxide j electrolyte threephase interline, which first propagates along the surface, then
invades the oxide phase[9, 16] (Figure 2 c and d).
The inward progression of metallization can be represented by a plot of reduction depth against electrolysis time
(Figure 1 c). When the cell voltage was increased from 3.1–
3.5 V, the reduction became faster. This observation cannot
be explained by Equations (1) and (2). The Ca activity varies
with electrode potential, but is forced to unity (i.e. the
deposition of a distinct solid or liquid Ca phase) at the Eo
value of [Eq. (1)] or at more negative potentials. Consequently, when the potential is made more negative, [Eq. (2)]
should proceed at a constant rate, but [Eq. (3)] would become
faster. Thus, Figure 1 c is strong evidence for [Eq. (3)] despite
the occurrence of Ca deposition.
Plots in Figure 1 c were analyzed for mass transfer. In
diffusion-controlled processes, the parabolic law, depth =
(mDt)1/2, in which m is proportional to the diffusion coefficient
D and may be approximated by 2 D for simple 1D diffusion,
applies.[17] However, in more-complicated cases, mass transfer
can result from, and be affected by, both concentration
gradient (diffusion) and other factors such as electric fields on
charged species (electromigration). In previous work on
electroreduction of solid oxides, in particular, the logarithmic
law, depth = log(nDt+1), in which n is a constant related to the
mass-transfer rate, was observed.[9b, 14d] Such behavior may be
associated with the fact that 1) it is the O2 ions that are
removed from the pellet cathode under an electric field (i.e.
the electrode potential) and 2) the pore geometry in the pellet
cathode is complex.
Interestingly, the data presented in Figure 1 c satisfied
approximately both parabolic and logarithmic laws with m
and n increasing with the applied voltage. Diffusion coefficients estimated from m varied from 6.0 D 108 to 1.5 D
107 cm2 s1. These values are greater than those for oxygenatom diffusion in some solid metals at similar temperatures
(e.g. 109 cm2 s1 for the Y–Co (15 wt %) alloy,[18a]
1010 cm2 s1 for Ti,[18b,c] and 1012 cm2 s1 for the Ti–Nb
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 2444 –2448
(4 wt %) alloy[18b]). Therefore, oxygen diffusion in the solid Tb
metal formed is not likely to be responsible for Figure 1 c.
Because the estimated diffusion coefficients are also smaller
than those in a liquid, Figure 1 c probably reflects the transfer
of O2 ions in the molten salt contained in the non-linear
pores of the pellet.[17] This understanding is supported by the
fact that both m and n depend on the applied cell voltage,
which is expected for O2 ions in the electrolyte contained in
the pores of the pellet, but not for the oxygen atoms in the
solid metal.
Tb4O7 pellets with over 70 % porosity were prepared by
mixing the Tb4O7 powder with NH4HCO3 ( 10 wt %),
pelletizing the mixture by die-pressing at 4 MPa, heating to
and at 300 8C for 2 h to remove NH4HCO3 by decomposition,
and then sintering the remaining mass at 1000 8C for 2 h.
Electrolysis at 3.2 V and 850 8C for 4 h resulted in reduction
depths of 0.28 and 0.43 mm for the denser and more-porous
pellets, respectively. This difference implies that mass transfer
of O2 or CaO is, indeed, the rate-control step. However,
prolonged electrolysis did not lead to a greater difference in
reduction depth. Tb melts at 1356 8C and can sinter quickly at
850–950 8C.[19] Because the sintering speed is inversely
proportional to the particle size,[19] it is expected that,
regardless of the initial porosity of the pellet, the Tb metal
particles formed will grow quickly to similar sizes and
interconnect into less-porous structures that slow mass transfer.
The formation of liquid Ca on the pellet may also
influence mass transfer between the pellet and the molten
salt. The reduction depths measured at different voltages for 6
and 12 h of electrolysis are compared in Figure 1 d. The two
plots show an inflexion point at about 3.5 V beyond which the
reduction depth became either smaller or unchanged, considering experimental errors. A Ca layer was often observed
on electrolyzed pellets beyond 3.5 V but the interior was only
partially reduced. An example is shown in Figure 1 d (inset) in
which the dark interior contained Tb2O3, CaO, Tb, and Ca,
but the bright surface layer was Ca metal mixed with Tb, as
revealed by XRD and energy-dispersive X-ray (EDX)
analyses. These findings suggest that the reduction was
effectively retarded when Ca was overdeposited on the pellet.
The particle sizes in the Tb4O7 pellet before and after
electrolysis (Figure 4 a and b) confirmed that the nodular Tb
particles were well-sintered together.[8] The same was
observed when electrolyzing Dy2O3. The Tb image was
taken from the cross-section of a pellet sample washed in
DMSO. Surprisingly, the sample was very clean even though
CaCl2 is poorly soluble in DMSO. Scanning electron microscopy analysis of the interior of the pellet (Figure 4 c) also
revealed very little salt-induced charging. EDX analyses at
different locations in the pellet confirmed that the porous
metal contained very little solidified CaCl2 (Table 1). As it is
unlikely that the porous structure is not filled by molten salt
during electrolysis, wetting could be poor between the molten
salt and the internal surface of the porous metal.[10] Consequently, the molten salt might have been removed by
gravity and capillary force when the reduced pellet was lifted
from it, which is consistent with the observation of more salt
near the surface of the electrolyzed pellet (Figure 4 c and
Angew. Chem. 2006, 118, 2444 –2448
Figure 4. Scanning electron micrographs. a) Sintered Tb4O7 powder;
b) Tb powder from electro-reduction; c) cross-section of an electroreduced Tb4O7 pellet, showing solidified salt near the surface (bright
region A due to charging), but no salt in the interior (regions B–E) of
the fully metallized pellet.
Table 1: EDX analyses of marked locations in Figure 4 c.
Tb [%]
Ca [%]
Cl [%]
Table 1). This nonwetting hypothesis was further confirmed in
the following experiment.
Six holes were drilled in a small Tb sheet (Figure 2 e). It
was then attached to a Mo wire and heated briefly ( 30 s)
above the molten CaCl2 (900 8C) in the open reactor. This
treatment led to the formation of an observable surface oxide
layer on the sample. The oxidized sample was then fully
immersed vertically in molten CaCl2 for about 15 s before
lifting into air to cool. Removal of the solidified CaCl2 on the
surface of the sheet with a knife revealed that the six holes
were fully filled with solidified CaCl2 (Figure 2 f). The sample
was then suspended in the molten salt again. The reactor was
sealed and purged with argon. At the same time, electrolysis
was carried out between the sample and a graphite anode at
3.2 V for 30 min. A longer electrolysis time was used to ensure
that sufficient air was purged by argon, as a low oxygen partial
pressure in the reactor is necessary for satisfactory electroreduction or deoxidation of the sample. After electrolysis, the
sample was lifted into air and cooled. On removing the top
salt layer, it was found that five of the six holes were empty
(Figure 2 g). This observation demonstrates qualitatively but
conclusively that the electrodeoxidized Tb metal surface was
either poorly or not wetted by molten CaCl2.
Electrolysis at 3.1–3.5 V and 850–900 8C is fairly common
in industry. In the work described herein, up to 30 kWh
(kg Tb)1 of energy was used to attain low oxygen levels
( 4200 ppm) for the 2-mm-thick pellet. Energy consumption
for industrial Tb production is unavailable in the literature,
but it should be comparable with that for Ti production, that
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
is, 50 kWh (kg Ti)1. Furthermore, according to Figure 1 c, if
the pellet thickness were decreased to 1.5 mm, the energy
consumption could be nearly halved. Notably, both Dy2O3
and Y2O3 were electroreduced to the respective metals in the
way described above. The latter, in particular, is more stable
than CaO, so that [Eq. (2)] would not be feasible in that case.
Currently, HRE metals are mostly used to make alloys or
intermetallic compounds by melting. Because this process
involves metal extraction followed by combination, the
energy released by the formation of the intermetallic
compound cannot be recovered and hence used to offset the
energy required for extraction of the individual metals, which
results in elevated costs. The former can, however, partially
compensate the latter if the intermetallic compounds are
formed directly from the minerals by electrolysis. This
method is, therefore, preferred. Indeed, we have prepared
many HRE intermetallic compounds with well-controlled
compositions directly from mixed-oxide powders (Figure 3 e–
h) and at lower voltages and faster rates than for the
individual metals (Figure 1 c). Energy consumption ranged
from 6.5 kWh kg1 for TbNi5 to 11.1 kWh kg1 for TbFe2.
In conclusion, we have shown that porous Tb4O7 pellets
can be electroreduced to Tb metal under Ca deposition in
molten CaCl2. However, overdeposited Ca blocked ionic
exchange between the pellet and the molten salt and
effectively retarded the reduction process. The metal produced was porous but was not wetted by molten CaCl2, thus
allowing cleaning or leaching in DMSO. This process also
suits the extraction of Dy and Y metals from their oxides. In
all cases, electrolysis energy consumption was satisfactorily
Received: October 8, 2005
Revised: December 21, 2005
Published online: March 8, 2006
[9] a) T. Nohira, K. Yasuda, Y. Ito, Nat. Mater. 2003, 2, 397; b) X. B.
Jin, P. Gao, D. H. Wang, X. H. Hu, G. Z. Chen, Angew. Chem.
2004, 116, 751; Angew. Chem. Int. Ed. 2004, 43, 733.
[10] a) A. J. Muir Wood, R. C. Copcutt, G. Z. Chen, D. J. Fray, Adv.
Eng. Mater. 2003, 5, 650; b) G. Z. Chen, D. J. Fray, Light Met.
2004, 881.
[11] a) K. Dring, R. Dashwood, D. Inman, J. Electrochem. Soc. 2005,
152, E104; b) G. Z. Chen, D. J. Fray, J. Electrochem. Soc. 2002,
149, E455.
[12] a) T. H. Okabe, M. Nakamura, T. Oishi, K. Ono, Metall. Trans. B
1993, 24, 449; b) K. Hirota, T. H. Okabe, F. Saito, Y. Waseda,
K. T. Jacob, J. Alloys Compd. 1999, 282, 101.
[13] a) K. Ono, R. O. Suzuki, JOM-J. The Min. Met. Mater. Soc. 2002,
54, 59; b) R. O. Suzuki, J. Phys. Chem. Solids 2005, 66, 461.
[14] a) G. H. Qiu, M. Ma, D. H. Wang, X. B. Jin, X. H. Hu, G. Z.
Chen, J. Electrochem. Soc. 2005, 152, E328; b) P. Gao, X. B. Jin,
D. H. Wang, X. H. Hu, G. Z. Chen, J. Electroanal. Chem. 2005,
579, 321; c) K. Jiang, X. H. Hu, M. Ma, D. H. Wang, G. H. Qiu,
X. B. Jin, G. Z. Chen, Angew. Chem. 2006, 118, 442 – 446; Angew.
Chem. Int. Ed. 2006, 45, 428 – 432; d) M. Ma, D. H. Wang, W. G.
Wang, X. H. Hu, X. B. Jin, G. Z. Chen, J. Alloy. Compd. 2006,
DOI:10.1016/j.jallcom.2005.10.048, published online: December
5, 2005.
[15] a) U. Stohr, W. Freyland, Phys. Chem. Chem. Phys. 1999, 1, 4383;
b) G. Z. Chen, D. J. Fray, J. Appl. Electrochem. 2001, 31, 155.
[16] a) G. Z. Chen, D. J. Fray in Proc. 6th Int. Symp. Molten Salt
Chem. Technol. (Eds: N. Y. Chen, Z. Y. Qiao), Shanghai
University, Shanghai, China, 2001, p. 79; b) Y. Deng, D. H.
Wang, W. Xiao, X. B. Jin, X. H. Hu, G. Z. Chen, J. Phys. Chem. B
2005, 109, 14 043.
[17] a) U. R. Evans, The Corrosion and Oxidation of Metals:
Scientific Principles and Practical Applications, Edward
Arnold, London, 1960, p. 810; b) C. S. Cha, Electrode Kinetics,
3rd ed., Science Press of China, Beijing, 2002, p. 345.
[18] a) Y. Niu, Y. S. Li, F. Gesmundo, F. Viani, Intermetallics 2000, 8,
293; b) M. Gobel, V. A. C. Haanappel, M. F. Stroosnijder, Oxid.
Met. 2001, 55, 137; c) D. David, G. Beranger, E. A. Garcia, J.
Electrochem. Soc. 1983, 130, 1423.
[19] L. D. Zhang, Preparation and Applications of Ultra-Fine Powders, China Petrochemical Press, Beijing, 2001, p. 61.
Keywords: electrochemistry · metal extraction ·
rare-earth metals · terbium
[1] a) M. Kaneko, J. Magn. Magn. Mater. 1995, 148, 351; b) E.
Quandt, B. Gedach, K. Seemann, J. Appl. Phys. 1994, 76, 7000.
[2] a) J. N. Xu, China Rare Earth Information 2004, 9, 14; b) 2004
Roskill Report, Rare Earths and Yttrium:
[3] a) M. Lederer, Nature 1955, 176, 462; b) T. Uda, K. T. Jacob, M.
Hirasawa, Science 2000, 289, 2326.
[4] a) J. L. Moriarty, Jr., J. Met. 1968, 20, 41; b) V. T. Volkov, A. M.
Ionov, T. V. Nikiforova, Vacuum 1999, 53, 105.
[5] C. K. Lee, J. S. Park, S. H. Yeon, K. I. Rhee, Met. Mater. Int.
2001, 7, 343.
[6] a) K. Grojtheim, C. Krohn, M. Malinovsky, J. Thonstad,
Aluminium Electrolysis, Aluminium-Verlag, DMsseldorf, 1997;
b) G. X. Xu, Rare Earths (II), 2nd ed., Metallurgical Industry
Press of China, Beijing, 1995, pp. 97 – 180.
[7] B. Q. Wu, Metallurgy of Rare Earths, Metallurgical Industry
Press of China, Beijing, 1995.
[8] a) D. J. Fray, T. W. Farthing, Z. Chen, 1998, PCT Patent,
WO9964638; b) G. Z. Chen, D. J. Fray, T. W. Farthing, Nature
2000, 407, 361.
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 2444 –2448
Без категории
Размер файла
442 Кб
oxide, solis, electrochemically, terbiums, metallization
Пожаловаться на содержимое документа