Preparation of pure MgTiO3 powders and the effect of the ZnNb2O6-dope onto the property of MgTiO3-based ceramics

Article abstract of DOI:10.1016/j.jallcom.2009.11.140

Magnesium titanate (MgTiO₃) powders have been prepared and characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM). The effects of calcination temperature and MgO: TiO₂ ratios on phase formation, morphology an

Full text of DOI:10.1016/j.jallcom.2009.11.140

Journal of Alloys and Compounds 492 (2010) 461–465
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
Preparation of pure MgTiO3 powders and the effect of the ZnNb2O6-dope onto
the property of MgTiO3-based ceramics
Bin Tang∗, Shuren Zhang, Xiaohua Zhou, Chao Deng, Shengquan Yu
State Key Laboratory of Electronic Thin Films and Integrated Devices, University of Electronic Science and Technology of China, Chengdu 610054, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Magnesium titanate (MgTiO3) powders have been prepared and characterized by X-ray diffraction (XRD)
and scanning electron microscopy (SEM). The effects of calcination temperature and MgO: TiO2 ratios
on phase formation, morphology and particle size distribution of the proposed powders are exam-
ined. Single-phase MgTiO3 can be synthesized by heating the non-stoichiometrically 1.03MgO–1TiO2
at 1100 ◦C for 3 h. The densification, microstructural evolution and microwave dielectric characteristics
of 0.935MgTiO3–0.065CaTiO3 (935MCT) compound are developed. It is found that the content of ZnNb2O6
has an apparent effect on the physical properties and the microwave dielectric characteristics of 935MCT
ceramics. The 935MCT sample with 5 wt% of ZnNb2O6, sintered at 1300 ◦C for 2 h, was measured to show
excellent microwave dielectric properties: εr equal to 22.5, Q × f value equal to 93,561 GHz (at 9 GHz) and
ꢀf equal to 6.2 ppm/◦C.
Received 2 March 2009
Received in revised form
12 November 2009
Accepted 18 November 2009
Available online 24 November 2009
Keywords:
Powders
Solid-state reaction
Composites
© 2009 Elsevier B.V. All rights reserved.
X-ray methods
Electron microscopy
Dielectric properties
1. Introduction
methods. However, it seems highly difficult to synthesize pure
MgTiO3 without additional phases MgTi2O5 or Mg2TiO4 by con-
Dielectric resonators used in microwave frequency have been
widely investigated due to the fast growth of microwave telecom-
munication and satellite broadcasting, which can create, filter and
select frequencies in oscillators, amplifiers and tuners in communi-
cation circuits. Dielectric resonators provide significant advantages
in terms of compactness, light weight, temperature stability and
relatively low cost in the fabrication of high frequency devices.
Currently, materials used in the miniaturization of microwave cir-
cuits have three categories: high relative dielectric constant (εr),
high quality factor (Q) and low temperature coefficient of reso-
nant frequency (ꢀf) [1]. These three parameters are relative to the
size, frequency selectivity and temperature stability of the system,
respectively.
Magnesium titanate (MgTiO3) has been reported to exhibit
excellent microwave dielectric properties (εr ≈ 17, ꢀf ≈ −50 ppm/◦C
and Q × f = 110,000 at 10 GHz) that make it become an ideal res-
onator material for electronics [2]. The synthesis of monophase
MgTiO3 has, therefore, been a subject of several investigations
comprising of solid-state reaction methods [3], co-precipitation
[4], sol–gel method [5,6], hydrothermal [7] and mechano–chemical
complexation [8]. Unfortunately, these processes have little practi-
cability in the ceramic industry, except for the solid-state reaction
ventional solid-state reaction [3], even in mechano–chemical
complexation method a small quantity of MgTi2O5 shows up in
the final product [8]. These parasite phases can induce delay in
the shrinkage processes and modifications in the final dielectric
properties of the ceramic, which is highly detrimental to the indus-
trial use of this material in spite of its high importance for type I
multilayer ceramic capacitors.
Moreover, It must be kept in mind that the ꢀf value of the MgTiO3
ceramics has a big negative value ꢀf ≈ −50 ppm/◦C [2]. In principle,
the shift of ꢀf to a near zero value may be achieved by adding other
compounds having a ꢀf of opposite sign. CaTiO3, which has a high
positive ꢀf value, is generally introduced into the main composi-
tion [9]. With doping perovskite CaTiO3, 0.95MgTiO3–0.05CaTiO3
(defined as 95MCT) ceramics has been found to yield εr ≈ 21,
Q ≈ 8000 at 7 GHz, and a zero ꢀf value [10] due to the temperature
compensating effect of the mixed compositions and widely applied
in communication systems, such as radar and global positioning
systems (GPS) operating at microwave frequencies.
There has been a strong desire for developing a practical method
of preparing pure MgTiO3 powders. Due to the sintering temper-
ature as high as 1400–1450 ◦C of MgTiO3 ceramics, many efforts
had been made to sintering MgTiO3 ceramics at lower temperature.
ZnNb2O6 ceramics possess low sintering temperature (1150 ◦C) and
microwave dielectric properties: εr ≈ 25, a relatively high quality
value (Q × f ≈ 87,300 at 10 GHz), a relatively low temperature coeffi-
cient of resonant frequency (ꢀf ≈ −56 ppm/◦C) [11], and in addition,
∗
Corresponding author. Tel.: +86 28 8320 8048; fax: +86 28 83202139.
E-mail address: tangbin@uestc.edu.cn (B. Tang).
0925-8388/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2009.11.140

462
B. Tang et al. / Journal of Alloys and Compounds 492 (2010) 461–465
the microstructure and dielectric properties of ZnNb2O6-doped
MgTiO3 ceramics have not been reported.
In this study, pure MgTiO3 powders were prepared by the con-
ventional solid phase reaction using Mg(OH)2·4MgCO3·5H2O and
sub-micrometer size TiO2 as starting materials. And the effects
of ZnNb2O6 additive on sintering behavior, microstructure and
microwave dielectric properties of MgTiO3-based ceramics were
investigated.
2. Experimental procedures
CaTiO3 additions were prepared by the conventional solid-state ceramic route.
High purity CaCO3 (>99.9%) and TiO2 (main phase of anatase and additional phase of
rutile, Sub-micrometer size, >99.0%) were used as the starting materials. Stoichio-
metric amounts of the powder samples were mixed and ball milled using zirconia
balls in deionized water for 24 h. The resultant slurry was then dried and calcined
at 1100 ◦C for 4 h.
Using conventional solid-state synthesis, MgTiO3 were prepared from high
purity oxides. The raw materials were Mg(OH)2·4MgCO3·5H2O (>99.9%) and TiO2
(main phase of anatase and additional phase of rutile, Sub-micrometer size, >99.0%).
Mg(OH)2·4MgCO3·5H2O and TiO2 in the molar ratio of 1:1–1.06:1 were mixed by
ball-milling in a polyethylene container using ZrO2 balls in deionized water for 18 h.
After calcined at 700–1100 ◦C for 3 h, the crystal structure and composition were
examined by X-ray diffraction (XRD) using a Philips X’Pert diffractometer within
the range of 2Â = 20–80◦ with Cu K␣ radiation. The calcined pure MgTiO3 and CaTiO3
were weighed according to the 0.935MgTiO3–0.065CaTiO3 composition with La2O3
(0.5 wt%) and different contents of ZnNb2O6 (1–7 wt%). The doping La2O3 into MCT
ceramics has been extensively used to optimize the microwave dielectric proper-
ties [12–14]. Mixtures were re-milled with ZrO2 balls for 12 h in deionized water
and then dried. The dried powders were mixed with 5 wt% polyvinyl alcohol (PVA),
uniaxially pressed under a pressure of 20 MPa to obtain green compacts with diam-
eter of 20 mm and thickness of 2 mm, and then sintered at 1250–1350 ◦C in air for
2 h.
Microstructures of the calcined MgO–TiO2 powders and MgTiO3-based ceram-
ics were observed using a JEOL JSM-6490LV scanning electron microscope (SEM).
The bulk densities of the sintered pellets were measured by the Archimedes meth-
ods. The relative permittivity and quality factors were measured by Hakki-Coleman
dielectric resonator method [15] and an HP83752A network analyzer was employed
in the measurement. Dielectric properties were measured in the microwave fre-
quencies between 6 GHz and 10 GHz. The temperature coefficients of the resonant
frequency (ꢀf) were defined as follows:
f85 − f20
ꢀf
=
(1)
f20 × 65
where f20 and f85 were the resonant frequencies at 20 ◦C and 85 ◦C, respectively.
3. Results and discussion
3.1. The phase and morphology of MgO–TiO2 powders
Fig. 1 shows the X-ray diffraction patterns of 1MgO–1TiO2,
1.03MgO–1TiO2 and 1.06MgO–1TiO2 powders calcined at differ-
ent temperatures. The phases of the calcined powders depend on
the raw materials and sintering conditions. Table 1 summarizes
the phase compositions of MgTiO3 powders calcined at different
temperatures. As shown in Fig. 1 and Table 1, the XRD patterns
of MgO–TiO2 powders calcined at 700 ◦C show that the anatase
TiO2 phase is well preserved after calcination. It is clear that none
anatase phase is detected after calcined at 800 ◦C showing that
phase transition anatase to rutile occurred during the heated pro-
cess [16]. Fig. 1(a) shows the crystalline phase of the 1MgO–1TiO2
powders sintered at 900 ◦C and 1100 ◦C is MgTiO3 accompanied by
Fig. 1. X-ray diffraction patterns of (a) 1MgO–1TiO2, (b) 1.03MgO–1TiO2 and (c)
1.06MgO–1TiO2 powders calcined at different temperatures for 3 h.
the unreacted rutile TiO2 phase. The results indicate that it is dif-
ficult to prepare pure MgTiO3 from the mixtures of 1MgO–1TiO2
with a controlled stoichiometry because the excess TiO2 was pre-
sented in the calcined powders. According to Fig. 1(b) and Table 1,
the phase transition in 1.03MgO–1TiO2 system can be expressed as
follows:
◦
Mg(OH)2 • 4MgCO3 • 5H2O≤−70→0 CMgO
(2)
Table 1
The phase compositions of MgTiO3 powders calcined at different temperatures.
Mg: Ti = 1:1
Mg: Ti = 1.03:1
MgO + TiO2 + TiO2(A)
MgTiO3 + TiO2 + MgO
MgTiO3 + MgTi2O5 + MgO + TiO2
MgTiO3 + MgTi2O5 + MgO
MgTiO3
Mg: Ti = 1.06:1
700 ◦
C
–
–
MgO + TiO2 + TiO2(A)
MgTiO3 + TiO2 + MgO
MgTiO3 + MgTi2O5 + MgO
MgTiO3 + MgO
800 ◦C
900 ◦C
1000 ◦C
1100 ◦C
MgTiO3 + TiO2
–
MgTiO3 + TiO2
MgTiO3 + Mg2TiO4 + MgO
TiO2 (A): TiO2 (Anatase), TiO2: TiO2 (Rutile)

B. Tang et al. / Journal of Alloys and Compounds 492 (2010) 461–465
463
◦
MgO + TiO2(rutile) + TiO2(anatase)8−0→0 CMgTiO3
particle size and degree of agglomeration tended to increase with
calcination temperature.
+MgO + TiO2(rutile)
MgO + TiO2(rutile)800−−→900 CMgTiO3
MgTiO3 + TiO2(rutile)900−−1→000 CMgTi2O5
(3)
(4)
(5)
(6)
3.2. Microstructural evolution and densification of doped MgTiO3
ceramics
◦
◦
The SEM micrographs of the 0.935MgTiO3–0.065CaTiO3–
xZnNb2O6 ceramics revealed in Fig. 3 show the changes in the den-
sification and grain growth. As shown in Fig. 3, with the addition
of ZnNb2O6, few pores were observed for all compositions. SEM
micrographs show that the grain shape of 935MCT ceramics with
1 wt% ZnNb2O6 was equiaxed and uniform. The average grain size
ranged from 3 ␮m to 5 ␮m (Fig. 3(a)). However, the grain shape
changed to long column and the average grain size increased for the
samples doped with more ZnNb2O6 (Fig. 3(b)–(d)). It suggested that
large amount of ZnNb2O6 additives could induce the grain growth
of samples.
Fig. 4 shows the apparent density of 935MCT ceramics doped
with ZnNb2O6 as a function of sintering temperature. The densities
of samples steadily improved with increasing sintering tempera-
ture, and saturated at 1300–1350 ◦C and the addition of ZnNb2O6
to MgTiO3 enlarged densities of the specimens, since the theoret-
ical density of MgTiO3 is 3.86 g/cm3 [19], while that of ZnNb2O6
is 5.65 g/cm3 [20]. It was difficult to densify MgTiO3 ceramics
when the sintering temperature was lower than 1400 ◦C. How-
ever, when the sintering temperature was higher than 1300 ◦C,
97% theoretical density could be obtained for all samples due
to that the densification temperature of MgTiO3-based ceramics
decreased with increasing amount of ZnNb2O6. The results sug-
gested that the sample doped with ZnNb2O6 could significantly
lower the sintering temperature of MgTiO3-based ceramics from
1400 ◦C to 1300 ◦C.
◦
MgTi2O5 + MgO1000−1100 C
−→
2MgTiO3
As shown in Eqs. (2)–(6), the 1.03MgO–1TiO2 powders show
various phases at different calcining temperatures. It was found
that the minor phases of MgO and MgTi2O5 tend to form
together with MgTiO3, depending on calcination conditions. For-
tunately, single-phase MgTiO3 could be synthesized by heating
the 1.03MgO–1TiO2 at 1100 ◦C for 3 h. However, in the case of
1.06MgO–1TiO2 powders, the XRD results shown in Fig. 1(c) and
Table 1 indicated that all the as-produced powders were com-
pletely polyphase, as all the samples have MgO peaks. At 900 ◦C,
both MgTi2O5 and MgO were observed in the powders. At 1000 ◦C,
a small amount of MgO was appeared. MgO remained even at
1100 ◦C, but Bernard et al. [17] found that heating at 1000 ◦C
via mixing and grinding process produced powders consisting
almost entirely of pure stoichiometric MgTiO3. Moreover, a trace
of Mg2TiO4 was observed in the 1.06MgO–1TiO2 powders calcined
at 1100 ◦C because of the reaction between MgTiO3 and MgO at a
high temperature. Sreedhar and Pavaskar [18] observed that excess
MgO (above 5 mol%) results in the formation of Mg2TiO4 phase. The
reaction can be expressed as follows:
◦
MgTiO3 + MgO10−0→0 CMg TiO4
(7)
2
The JCPDS powder diffraction files employed in this study for
anatase, rutile, MgO, MgTiO3, Mg2TiO4 and MgTi2O5 were 73-1764,
86-0147, 77-2179, 79-0831, 73-1723 and 35-0796, respectively.
The morphological changes in the 1.03MgO–1TiO2 powders are
illustrated in Fig. 2 as a function of calcination temperature. In
general, the particles were agglomerated and basically irregular in
shape, with a substantial variation in particle size, particularly in
samples calcined at high temperatures. The range of particle diam-
eter was found to be about 0.1–0.4, 0.2–1.0, 0.3–2.5 and 0.5–3.5 ␮m
for the samples calcined at 800 ◦C, 900 ◦C, 1000 ◦C and 1100 ◦C for
3 h, respectively. The results indicated that both the difference in
3.3. Dielectric properties of doped MgTiO3 ceramics
Fig. 5 shows the dielectric constant (εr) of 935MCT ceram-
ics doped with ZnNb2O6 as a function of sintering temperature.
The relationships between εr values and sintering temperatures
revealed the same trend with those between densities and sintering
temperatures. On the other hand, the εr values of the ceramics also
increased slightly with increasing amount of ZnNb2O6 additions.
Fig. 2. SEM micrographs of the 1.03MgO–1TiO2 powders calcined for 3 h at (a) 800 ◦C, (b) 900 ◦C, (c) 1000 ◦C, (d) 1100 ◦C.

464
B. Tang et al. / Journal of Alloys and Compounds 492 (2010) 461–465
Fig. 3. SEM photographs of 935MCT ceramics with various contents of ZnNb2O6 sintered at 1300 ◦C. (a) 1 wt%; (b) 3 wt%; (c) 5 wt%; (d) 7 wt%.
The εr values of the samples varied from 18 to 24 as the tempera-
ture increased from 1250 ◦C to 1350 ◦C. The improvement of εr was
closely related to the increase in relative density of ceramics at the
same sintering temperature due to the increasing sintering aids.
Fig. 6 shows the quality factor (Q × f) of 935MCT ceramics at
9 GHz with ZnNb2O6 additions as a function of sintering tempera-
ture. The Q × f values of 935MCT ceramics with different ZnNb2O6
contents also increased with increasing sintering temperature, then
decreased after reaching the maximum values. On the other hand,
the Q × f values increased with the increasing amount of ZnNb2O6
and changed from 72,643 GHz (1 wt%) to 93,561 GHz (5 wt%) when
sintering at 1300 ◦C. However, Q × f values of samples with more
ZnNb2O6 addition (7 wt%) were much the same as that of 5 wt%
ZnNb2O6-doped ceramics. The variation tendency of Q × f values
for the ceramics with increasing ZnNb2O6 additions matches well
with the earlier report that the Q × f values increased with the
average grain size [21]. According to research by Djuniadi and
Sagala [22], as the grain size increased, the pores and the grain
boundary area decreased so as to reducing the lattice imperfec-
tions and increasing the Q × f values. In this study, the increases
in Q × f values were attributed to the increase in densities and
grain growth. But in the case of incorporation of more ZnNb2O6
additive, that phase rich in Nb and Zn grew may cause impu-
rity in ceramics, then anharmonicity in phonon vibration was
enhanced so as to cause the decrease of Q × f value since the Q × f
Fig. 4. The densities of 935MCT ceramics with various ZnNb2O6 addition as a func-
tion of sintering temperature.
Fig. 5. The dielectric constant (εr) of 935MCT ceramic with various ZnNb2O6 addi-
tions as a function of sintering temperature.
Fig. 6. The Q × f values of 935MCT ceramic with various ZnNb2O6 additions as a
function of sintering temperature.

B. Tang et al. / Journal of Alloys and Compounds 492 (2010) 461–465
465
The resulting MgO–TiO2 powders consist of agglomerated parti-
cles of 0.1–3.5 ␮m in size, which are rounded in morphology. The
microstructural evolution, physical and microwave dielectric prop-
erties of the 935MCT ceramics by the addition of ZnNb2O6 have
been investigated. It is obtained that the addition of the ZnNb2O6 in
the ceramics can effectively lower the sintering temperature, and
increases the bulk density, dielectric constant and Q × f values of
the sintered ceramics. The temperature coefficient of resonant fre-
quency of 935MCT shifted to more negative values as the amount
of ZnNb2O6 increased. The 935MCT sample with 5 wt% ZnNb2O6
exhibited the best microwave dielectric constant: εr of 22.5, Q × f
value of 93,561 GHz (at 9 GHz) and ꢀf of 6.2 ppm/◦C.
References
[1] S.-F. Wang, Y.-F. Hsu, Y.-R. Wang, L.-T. Cheng, Y.-C. Hsu, J.P. Chu, C.-
Y. Huang, Journal of the European Ceramic Society 26 (2006) 1629–
1635.
[2] J.-H. Sohn, Y. Inaguma, S.-O. Yoon, M. Itoh, T. Nakamura, S.-J. Yoon, H.-J. Kim,
Japanese Journal of Applied Physics 33 (1994) 5466–5470.
[3] W.W. Cho, K.-I. Kakimoto, H. Ohsato, Materials Science and Engineering B 121
(2005) 48–53.
Fig. 7. ꢀf values of 935MCT ceramics doped with various ZnNb2O6 additions sintered
at 1300 ◦C.
value was independent of the density for highly densified ceramics
[23].
[4] A.B. Gaikwad, S.C. Navale, V. Samuel, A.V. Murugan, V. Ravi, Materials Research
Bulletin 41 (2006) 347–353.
Fig. 7 shows the temperature coefficient of resonant frequency
(ꢀf) of 935MCT ceramics with different amounts of ZnNb2O6 addi-
tions sintered at 1300 ◦C. The temperature coefficient of resonant
frequency (ꢀf) is related to the composition and the second phase
of a material. The ꢀf values of samples tended to move toward
negative direction as the amount of ZnNb2O6 addition increased.
Since ZnNb2O6 has a much more negative ꢀf value compared
with 935MCT, ꢀf values of ZnNb2O6-doped samples could have
more negative with increasing ZnNb2O6 addition. It varied from
14.5 ppm/◦C to 2.8 ppm/◦C as the amount of ZnNb2O6 addition
increased from 1 wt% to 7 wt%, and thus, there was a method to
adjust ꢀf value to the nearly 0 ppm/◦C by variating of the ZnNb2O6
addition. In this study, the optimum Q × f = 93,561 GHz (at 9 GHz)
and ꢀf value = 6.2 ppm/◦C were revealed in 5 wt% ZnNb2O6-doped
935MCT ceramics.
[5] Y.-M. Miao, Q.-L. Zhang, H. Yang, H.-P. Wang, Materials Science and Engineer-
ing: B 128 (2006) 103–106.
[6] I.R. Abothu, A.V. Prasada Rao, S. Komarneni, Materials Letters 38 (1999)
186–189.
[7] V. Parvanova, M. Maneva, Thermochimica Acta 279 (1996) 137–141.
[8] J. Liao, M. Senna, Materials Research Bulletin 30 (1995) 385–392.
[9] A. Belous, O. Ovchar, D. Durilin, M.M. Krzmanc, M. Valant, D. Suvorov, Journal
of the American Ceramic Society 89 (2006) 3441–3445.
[10] K. Wakino, Ferroelectrics 91 (1989) 69–86.
[11] H.-J. Lee, I.-T. Kim, K.S. Hong, Japanese Journal of Applied Physics 36 (1997)
L1318.
[12] T.A. Vanderah, V.L. Miller, I. Levin, S.M. Bell, T. Negas, Journal of Solid State
Chemistry 177 (2004) 2023–2038.
[13] Y.-B. Chen, C.-L. Huang, S.-H. Lin, Materials Letters 60 (2006) 3591–3595.
[14] Y.-B. Chen, Journal of Alloys and Compounds 477 (2009) 883–887.
[15] B.W. Hakki, P.D. Coleman, IEEE Transactions on Microwave Theory and Tech-
niques 8 (1960) 402–410.
[16] T. Ohsaka, Solid State Communications 30 (1979) 345–347.
[17] J. Bernard, F. Belnou, D. Houivet, J.-M. Haussonne, Journal of Materials Process-
ing Technology 199 (2008) 150–155.
[18] K. Sreedhar, N.R. Pavaskar, Materials Letters 53 (2002) 452–455.
[19] C.-L. Huang, C.-H. Shen, C.-L. Pan, Materials Science and Engineering: B 145
(2007) 91–96.
4. Conclusion
[20] S.-H. Wee, D.-W. Kim, S.-I. Yoo, Journal of the American Ceramic Society 87
(2004) 871–874.
[21] S. Kucheiko, J.-W. Choi, H.-J. Kim, H.-J. Jung, Journal of the American Ceramic
Society 79 (1996) 2739–2743.
[22] S.N. Djuniadi, A. Sagala, Journal of the American Ceramic Society 75 (1992)
2573–2575.
[23] W.S. Kim, T.H. Kim, E.S. Kim, K.H. Yoon, Japanese Journal of Applied Physics 37
(1998) 5367–5371.
Single-phase of magnesium titanate powders may be pro-
duced by employing
a
solid-state reaction process using
Mg(OH)2·4MgCO3·5H2O and sub-micrometer size TiO2 as raw
materials. The calcination temperature and MgO: TiO2 ratios have
been found to have a pronounced effect on the phase formation
and particle size of the calcined magnesium titanate powders.