Journal of the Chinese Chemical Society, 2009, 56, 1112-1117
1112
Effect of Li2O Addition on the Preparation of (Y2-yLiy)Ti2O7-y
Wen-Ping Sua (
), Yu-Hsuen Leea (
), Ching-Tien Hsieha (
b
), Jyh-Fu Leeb (
),
Hwo-Shuenn Sheu (
b
) and H.-C. I. Kaoa,* (
)
Yong-Ping Chiang (
),
a
b
Department of Chemistry, Tamkang University, Tamsui 25137, Taiwan, R.O.C.
National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan, R.O.C.
A new series of (Y2-yLiy)Ti2O7-y having an ordered pyrochlore phase was prepared by a solid state reaction method with a solid solution range of 0.05 £ y £ 0.10. Unit cell parameters obtained by the Rietveld
refinement method shows that the a-axis decreases linearly with increasing the amount of Li ion addition,
indicating the successful incorporation of the Li ion into unit cell. The average x-fractional coordinate of
the O(1) site depends on the ionic radius ratio of r(A3+)/r(Ti4+) in the A2Ti2O7 with a pyrochlore phase. The
Ti K-edge XANES spectra of the (Y2-yLiy)Ti2O7-y show that the valence of the Ti ions is slightly less than 4
so that Ti is in the mixed valence state. Average particle size increases with increasing the amount of extra
Li ion addition, which acts as a flux to lower the melting point of the materials.
Keywords: (Y2-yLiy)Ti2O7-y; Flux; Pyrochlore; SEM; Ti K-edge XANES.
INTRODUCTION
Pyrochlore (A2B2O7) has a space group of Fd3m. The
prototype phase that gives its name to the pyrochlore structure is a mineral with a formula as (Ca,Na) 2(Nb,Ta) 2O 6
(O,OH,F).1 A wide variety of additional cations are commonly found to be incorporated in solid solution in specimens that occur in nature. Brixner had prepared several
Ln2Ti2O7, Ln = Sc, Y, La-Lu (all the rare earth cations).
Their unit cell a-axis and resistivity had been reported.2
The pyrochlore phase is a superstructure derivative of the
simple fluorite (AO2) phase. Every four units of AO2 has an
oxide vacancy, as a result, three oxygen sites are identified
in a pyrochlore unit cell. For an ordered pyrochlore phase,
O(1) at the 48f site and O(2) at the 8b site are fully occupied, but O(3) at the 8a site is completely vacant.3-5 As a result, the x-fractional coordinate of the O(1) site is not in the
special position, (3/8, 1/8, 1/8). The amount of this shift depends on the occupancy factor of the O(3) site.
Both Ln2Zr2O7 and Ln2Ti2O7 (Ln = rare earth cation)
are possible candidates as electrolytes in the solid oxide
fuel cell.5-9 Their oxygen ion conductivity reaches to 10-2
S/cm at 700 °C and 10-1 S/cm at 1000 °C.10 In order to measure the ionic conductivity, highly densed materials are required. Due to the high melting point of these zirconates
and titanates, adding a flux to lower the melting point of the
* Corresponding author. E-mail: [email protected]
parent compound is one way to approach.11-13
In this report, a series of samples with the nominal
compositions of (Y2-yLi3y)Ti2O7-y was prepared. One-third
of the Li2O acts as a dopant substituting into the Y site of
Y 2Ti2O 7 and the rest of them play as a flux to lower the
melting point of the titanates. Solid solution range, Rietveld
refinement results and the roles played by the Li atom were
investigated.
The Ti valence and coordination in oxide compounds
have been studied by the Ti K-edge XANES spectra. The
pre-edge position, width and height, which mainly originate from the dipole transition is interpreted in terms of the
overlapping of the Ti 4p with Ti 3d orbitals in different coordination.14-16 The Ti K-edge XANES spectra have been
studied to find the valence and coordination environment
of the Ti atom in the title compounds.
EXPERIMENTAL
Solid state reaction method was employed for preparing (Y2-yLiy)Ti2O7-y. The amount of Li2CO3 addition is 3y
per formula unit. Appropriate amount of the starting materials including 99.9% of Gd2O3 and Y2O3, 99% of Li2CO3
and TiO2 obtained from Riedel-de Haen were weighed and
ground thoroughly, then pressed at ambient temperature
and sintered at 850 °C for 4 h to decompose Li2CO3, fol-
Effect of Lithium Oxide Addition on (Y2-yLiy)Ti2O7-y
lowed by furnace cooling in a box furnace in static air.
Sample was ground, pressed and sintered at 1180 °C for 10
h and repeated 3 times. In order to reduce the amount of the
Li ion evaporation, powder with the same composition was
placed on top of the pellet and the container was covered.
Phase purity of the sample is checked by the XRD patterns
obtained from a Bruker MXP3 Diffractometer equipped
with Cu Ka lines with a weighted average wavelength of
1.5418 Å and a graphite monochromator. The GSAS (General Structure Analysis System) program developed by
Larson and von Dreele17 was employed for the structural
parameters analysis with Rietveld method. In the beginning of the Rietveld refinement, thermal parameter, Uiso,
was set as 0.025 Å2 for all the atomic sites. All three oxygen
sites are constrained to the same Uiso and they are not allowed to vary independently because of the light mass of
the oxygen atom compared with other atoms in the sample.
In the refinement, the formulas of the compounds were assumed as (Y2-yLiy)Ti2O7-y. Morphology of the sample was
observed with a VEGA\\SBH Scanning Electron Microscope. All samples are electrical insulator at room temperature so that their surface should be covered with a thin layer
of gold prior to the examination. X-ray absorption spectra
were conducted on beamline BL17C1 at the NSRRC in Taiwan. All spectra were measured in transmission mode at
room temperature. Powder samples was uniformly spread
onto a Scotch tape and folded to get the desired thickness.
The spectra were collected using a gas ionization detector.
The ion chambers used for measuring the incident (I0) and
transmitted (I) synchrotron beam intensities were filled
with a H 2/N 2 gas mixture and N 2 gas, respectively. Data
were collected from 200 eV below the Ti K-edge (4966 eV)
to 1100 eV above the edge. Ti2O3, TiO2 anatase and rutile
were used as reference standards.
RESULTS AND DISCUSSION
In the preparation of (Y2-yLiy)Ti2O7-y by the addition
of 3y moles of Li2O in the samples at 1180 °C, with y =
0.04, impurity phase, Y2O3, was observed in the XRD pattern; and with y = 0.11, Li2Ti2O4 appeared; so that single
phase materials were obtained in the range of 0.05 £ y £
0.10. XRD of all the single phase samples is shown in Fig.
1. Because of the light weight of the Li atom and the amount
of extra Li2O addition is quite small, no diffraction peaks
associate to the Li containing impurity is observed, sug-
J. Chin. Chem. Soc., Vol. 56, No. 6, 2009
1113
gesting that the extra Li2O is probably evenly distributed in
the grain boundaries.
All the (Y2-yLiy)Ti2O7-y have a space group of Fd3m
with an ordered pyrochlore phase. Rwp for all samples are
refined to 6.7-8.6%, which are reasonable values. One
Rietveld refinement result with y = 0.05 is shown in Fig. 2.
All the refinement results are similar so that only one sam-
Fig. 1. XRD of all (Y2-yLi y)Ti 2 O 7-y samples prepared
with 3y moles of Li2O.
Fig. 2. Rietveld
refinement
results
of
a
(Y2-yLiy)Ti2O7-y with y = 0.05 prepared with 3y
moles of Li2O. The figures include the experimental, calculated, and difference XRD profiles. Crosses are the experimental data; solid
lines are the calculated profile. All the possible
Bragg reflections are indicated with short vertical tics below the calculated profile. The difference between the experimental and calculated
results is plotted below the Bragg refraction
tics.
1114
J. Chin. Chem. Soc., Vol. 56, No. 6, 2009
ple is shown. Fig. 3 plots the unit cell a-axis versus the
amount of Li+ ion substitution into the Y site of Y2Ti2O7-y.
A linear line is fitted and the unit cell volume decreases
with increasing the amount of Li ion doping. Ionic radii of
8-coordination Y3+ and Li+ are 1.109 and 0.92 Å, respectively.18 A linear decrease of the a-axis suggests that Li+
was successfully incorporated into the unit cell. In the
Rietveld refinement process, the formula, (Y2-yLiy)Ti2O7-y,
is assumed, although the nominal composition of the Li
atom is 3y instead of one y, it is not possible to include 3 Li+
ions at one site. Extra Li ions are likely to act as a flux to
lower the melting point of the materials, evidence of this
assumption is shown later.
In order to prove that only one Li ion is substituted
into the pyrochlore unit cell, materials with the nominal
composition of (Y2-yLiy)Ti2O7-y were prepared for y = 0.06
and 0.09. Unit cell parameters obtained from the Rietveld
refinement method are marked in Fig. 1 with (O) symbols.
These two points fixed well into the linear line, indicating
that all the materials prepared have the same formula and
it is (Y 2-yLi y)Ti 2 O 7-y. Therefore, no matter the nominal
composition of the materials is (Y 2-yLi 3y)Ti 2 O 7-y or
(Y 2-yLiy)Ti2O 7-y, the pyrochlore phase obtained has the
same composition.
In (Y 2-yLiy)Ti2O 7-y series having an ordered pyrochlore structure, the O(3) site is completely empty, as a
consequence, the O(1) site shifts from a special position
with x = 0.375 to and average value of 0.4227(8) and the
average shift, Dx, is 0.0477(8). Because of the low diffraction power of the oxygen atoms, refinement results from
Fig. 3. The unit cell a-axis vs. y of (Y 2-yLi y)Ti 2 O 7-y,
(O) prepared with one y mole, and (n) prepared
with 3y moles of Li2O.
Su et al.
XRD data are not as accurate as those obtained from the
neutron diffraction patterns so that average value of the
x-fractional coordinate of the O(1) site is presented. Checking into literature reports, Gd 2Ti2O 7, 19 Y 2Ti2O 7, 20 and
Yb2Ti2O710 have x = 0.4291(6), 0.4247(4) and 0.4201(3)
with Dx = 0.0541(6), 0.0497(4) and 0.0451(3), respectively. The amount of shift seems related to the ionic radius
ratio of the cations, r(A3+)/r(Ti4+), so that the average cation
radius ratio is calculated and plotted in Fig. 4. The r(A3+)/
r(Ti4+) for the (Y2-yLiy)Ti2O7-y is 1.68, which is bigger than
Yb2Ti2O7,10 and smaller than Y2Ti2O720 and Gd2Ti2O7. It is
found that x decreases with decreasing the ionic radius ratio. A linear line is obtained and fitted, it follows x = 0.08(1)
r(A 3+ )/r(Ti4+ ) + 0.29(2). This phenomenon has not been
emphasized yet. Extrapolating the x-fractional coordinate
to x = 0.375, r(A3+)/r(Ti4+) = 1.1. It means that when the
ionic radius ratio is 1.1, the O(1) site is located at the special position so that crystal structure should become fluorite and all of the oxygen atoms are evenly distributed in
all the oxygen sites. Actually, it is not necessary to lower
the cation radius ratio to 1.1 to obtain fluorite phase.
Yamamura et al. found that if the r(A3+)/r(B4+) is smaller
than 1.48, it is possible to prepare A2B2O7 with a fluorite
phase.21
The Ti K-edge XANES spectra for all the samples is
shown in Fig. 5. Looking into the main absorption edge
peak near 4985 eV, it is found that all samples has absorption energy slightly less than the reference, TiO2 rutile. Us-
Fig. 4. The x-fractional coordinate of the O(1) site vs.
ionic radius ratio of r(A 3+ )/r(Ti 4+ ) in the
A 2 Ti 2 O 7 . (o) Yb 2 Ti 2 O 7 , 10 (D) Y 2 Ti 2 O 7 , 20 (O)
Gd 2 Ti 2 O 7 19 and (n) (Y 2-yLi y)Ti 2 O 7-y prepared
with 3y moles of Li2O.
Effect of Lithium Oxide Addition on (Y2-yLiy)Ti2O7-y
Fig. 5. The Ti K-edge XANES spectra of references,
and (Y2-yLiy)Ti2O7-y prepared with 3y moles of
Li2O.
J. Chin. Chem. Soc., Vol. 56, No. 6, 2009
1115
ing Ti2O3 and TiO2 rutile as standards, the valence of the Ti
atom calculated in the samples with is 3.87 for y £ 0.07 and
3.75 for y > 0.07. All of them have the Ti valence less than
4+. The amount of the Ti atom reduction slightly increases
with increasing the amount of Li ion doping. It means that
during the preparation at 1180 °C for 30 h in the static air
atmosphere, Ti ions were partially reduced. Ti atom is in
the mixed valence state. As a result, samples become pale
blue.
Looking into the pre-edge peaks at 4971.5 eV, which
is a transition of Ti 1s to 3d, a forbidden transition happened when the Ti 3d is hybridized with the 4p orbitals so
that this transition is coordination dependent.14-16 For TiO2
rutile, it has tetragonal structure with 6 Ti-O bonds, two of
them have a bond lengths of 1.9803 Å and four others of
Fig. 6. SEM images (2000x) of (Y2-yLiy)Ti2O7-y with (a) 0.05, (b) 0.06, (c) 0.09 and (d) 0.10 prepared with 3y moles of Li2O;
(e) 0.06 and (f) 0.09 are samples prepared with one y mole of Li2O.
1116
J. Chin. Chem. Soc., Vol. 56, No. 6, 2009
1.9487 Å.15 Our samples have a cubic ordered pyrochlore
phase. The Ti atom is octahedrally coordinated with six oxygen atoms and there is only one Ti-O bond distance for
each sample. Variation of these Ti-O bond length is quite
small among different samples so that average distance is
calculated, it is 1.947(3) Å for all the (Y2-yLiy)Ti2O7-y samples. In an ordered pyrochlore phase, the Ti atom is central
symmetrically surrounded by 6 oxygen atoms but the Ti-O
bonding in rutile is not, so that the intensity of this peak of
rutile is stronger than that of the title compounds.
Morphology of four samples, from (a) to (d) prepared
by 3y moles of Li2O, observed under SEM with 2000 times
magnification is shown in Fig. 6. All samples are quite homogenous with no second phase observed. The extra Li2O
probably exists in the grain boundaries. In addition, increasing the amount of y, more extra Li ions were added
into the samples and the materials gradually melt. Melting
point of these materials decreases with increasing the
amount of Li2O. The extra Li ions act like a flux to lower
the melting point of the samples. On the other hand, Fig.
6(e) and 6(f) are two samples prepared by the nominal composition of (Y2-yLiy)Ti2O7-y, only one y mole of Li2O, with y
= 0.06 and 0.09, respectively. Particle size in Fig. 6(e) and
6(f) is smaller than their counterparts in Fig. 6(b) and 6(c),
respectively. In addition, samples 6(b) and 6(c) have grains
growing together. Samples are about to melt, but for the
6(e) and 6(f), grains were separated with each other. Difference between them is obvious. The extra Li2O does help to
lower the melting point of the materials.
Particle sizes of the materials with extra Li2O esti-
Fig. 7. Average particle size vs. y of (Y2-yLiy)Ti2O7-y
prepared with 3y moles of Li2O.
Su et al.
mated from the SEM images are plotted in Fig. 7. They are
in the range of 3 to 6 mm for those prepared by 3 y moles of
Li2O. On the other hand, most of the particles in the samples prepared by one y mole of Li 2 O (see Fig. 6(e) and
6(f)) are quite small. Extra Li2O does enhance the particle
growth.
CONCLUSION
Li ion is successfully brought into the Y site of
Y 2Ti2O 7. Extra Li2O acts as a flux to lower the melting
point and increase the particle growth of the materials. The
pyrochlore phase prepared in this report has a formula of
(Y2-yLiy)Ti2O7-y, which is a new series. The valence of the
Ti ion in the (Y2-yLiy)Ti2O7-y is 3.8, slightly lowered than 4
and the Ti atom is in a mixed valence state. The x-fractional
coordinate of the O(1) site in an ordered pyrochlore unit
cell depends on the ionic radius ratio of the A and Ti cations
in the A2Ti2O7. The relationship has not been emphasized
before.
ACKNOWLEDGMENTS
This work was supported by the National Science
Council of Taiwan under contract No. NSC 96-2113-M032-003 and NSC 97-2113-M-032-005.
Received June 5, 2009.
REFERENCES
1. Subramanian, M. A.; Aravamudan, G.; Subba Rao, G. V.
Prog. Solid State Chem. 1983, 15, 55.
2. Brixner, L. H. Inorg. Chem. 1964, 3, 1065.
3. Heremans, C.; Wuensch, B. J. J. Solid State Chem. 1995,
117, 108.
4. Wuensch, B. J.; Eberman, K. W. J. Miner. Met. Mater. Soc.
2000, 52, 19.
5. Wuensch, B. J.; Eberman, K. W.; Heremans, C.; Ku, E. M.;
Onnerud, P.; Yeo, E. M. E.; Haile, S. M.; Stlick, J. K.;
Jorgensen, J. D. Solid State Ionics 2000, 129, 111.
6. Abrantes, J. C. C.; Levchenko, C. A.; Shlyakhtina, A. V.;
Shcherbakova, L. G.; Horovistiz, A. L.; Fagg, D. P.; Frade, J.
R. Solid State Ionics 2006, 177, 1785.
7. Lian, J.; Wang, L. M.; Wang, S. X.; Chen, J.; Boatner, L. A.;
Ewing, R. C. Phys. Rev. Lett. 2001, 14, 14591-01-1.
8. Kramer, S. A.; Tuller, H. L. Solid State Ionics 1995, 82, 15.
9. Kharton,V. V.; Mzrques, F. M. B.; Atkinson, A. Solid State
Ionics 2004, 174, 135.
10. Shlyakhtina, A. V.; Fedtke, P.; Busch, A.; Kolbanev, I. V.;
Barfels, T.; Wienecke, M.; Sokolov, A. E.; Ulianov, V. A.;
Effect of Lithium Oxide Addition on (Y2-yLiy)Ti2O7-y
11.
12.
13.
14.
15.
16.
Trounov, V. A.; Shcherbakova, L. G. Solid State Ionics 2008,
179, 1004.
Gao, H.; Liu, J.; Chen, H.; Li, S.; He, T.; Ji, Y.; Zhang, J.
Solid State Ionics 2008, 179, 1620.
Gil, V.; Moure, C.; Durán, P.; Tartaj, J. Solid State Ionics
2007, 178, 359.
Gorbova, E.; Maragou, V.; Medvedev, D.; Demin, A.;
Tsiakaras, P. Solid State Ionics 2008, 179, 887.
Farges, F.; Brown, Jr., G. E.; Rehr, J. J. Phys. Rev. B 1997,
56, 1809.
Wu, Z. Y.; Ouvrard, G.; Gressier, P.; Natoli, C. R. Phys. Rev.
B 1997, 55, 10382.
Luca, V.; Djajanti, S.; Howe, R. F. J. Phys. Chem. B 1998,
J. Chin. Chem. Soc., Vol. 56, No. 6, 2009
1117
102, 10650.
17. Larson, A. C.; von Dreele, R. B. General Structure Analysis
System; Report La-UR-86-748; Los Alamos National Laboratory: Los Alamos, 2003.
18. Shanon, R. D. Acta Crystallogr. A 1976, 32, 751.
19. Hess, N. J.; Begg, H. D.; Conradson, S. D.; McCready, D. E.;
Gassman, P. L.; Weber, W. J. J. Phys. Chem. B 2002, 106,
4663.
20. Chtoun, E.; Hanebali, L.; Garnier, P.; Kiat, J. M. Eur. J. Solid
State Inorg. Chem. 1997, 34, 553.
21. Yamamura, H.; Nishino, H.; Kakinuma, K. Solid State
Ionics 2003, 158, 359.