Microwave dielectric properties and sintering behavior of nano-scaled (α + θ)-Al2O3 ceramics

Article abstract of DOI:10.1016/j.materresbull.2007.06.032

The dielectric properties at microwave frequencies and the microstructures of nano (α + θ)-Al₂O₃ ceramics were investigated. Using the high-purity nano (α + θ)-Al₂O₃ powders can effectively increase the value of the quality factor and lower the sintering temperature of the ceramic samples. Grain growth can be limited with θ-phase Al₂O₃ addition and high-density alumina ceramics can be obtained with smaller grain size comparing to pure α-Al₂O₃. Relative density of sintered samples can be as high as 99.49% at 1400 °C for 8 h. The unloaded quality factors Q × f are strongly dependent on the sintering time. Further improvement of the Q × f value can be achieved by extending the sintering time to 8 h. A dielectric constant (ε_r) of 10, a high Q × f value of 634,000 GHz (measured at 14 GHz) and a temperature coefficient of resonant frequency (τ_f) of -39.88 ppm/°C were obtained for specimen sintered at 1400 °C for 8 h. Sintered ceramic samples were also characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM).

Full text of DOI:10.1016/j.materresbull.2007.06.032

Materials Research Bulletin 43 (2008) 1463–1471
www.elsevier.com/locate/matresbu
Microwave dielectric properties and sintering behavior
of nano-scaled (a + u)-Al2O3 ceramics
a
b
b
Cheng-Liang Huang a, , Jun-Jie Wang , Fu-Su Yen , Chi-Yuen Huang
*
a Department of Electrical Engineering, National Cheng Kung University, Tainan, Taiwan
b Department of Resource Engineering, National Cheng Kung University, Tainan, Taiwan
Received 15 July 2006; received in revised form 14 March 2007; accepted 12 June 2007
Available online 16 June 2007
Abstract
The dielectric properties at microwave frequencies and the microstructures of nano (a + u)-Al2O3 ceramics were investigated.
Using the high-purity nano (a + u)-Al2O3 powders can effectively increase the value of the quality factor and lower the sintering
temperature of the ceramic samples. Grain growth can be limited with u-phase Al2O3 addition and high-density alumina ceramics
can be obtained with smaller grain size comparing to pure a-Al2O3. Relative density of sintered samples can be as high as 99.49% at
1400 8C for 8 h. The unloaded quality factors Q ꢀ f are strongly dependent on the sintering time. Further improvement of the Q ꢀ f
value can be achieved by extending the sintering time to 8 h. A dielectric constant (er) of 10, a high Q ꢀ f value of 634,000 GHz
(measured at 14 GHz) and a temperature coefficient of resonant frequency (tf) of ꢁ39.88 ppm/8C were obtained for specimen
sintered at 1400 8C for 8 h. Sintered ceramic samples were also characterized by X-ray diffraction (XRD) and scanning electron
microscopy (SEM).
# 2007 Elsevier Ltd. All rights reserved.
Keywords: A. Ceramics; A. Oxides; B. Crystal growth; B. Dielectric properties
1. Introduction
The development of microwave dielectric resonators for applications in communication systems, such as cellular
phones, wireless local area networks (WLAN), direct broadcasting satellite (DBS) and global positioning systems, has
been rapidly progressing in the past decades [1,2]. The unique electrical properties of ceramic dielectric resonators
have revolutionized the microwave-based wireless communications industry by reducing the size and cost of filter and
oscillator components in circuit systems. The advantage of using dielectric resonators is that it makes the size
reduction of microwave components possible. Requirements for these dielectric resonators must be the combined
dielectric properties of a high dielectric constant (er > 25), a low dielectric loss (Q > 5000, where Q = 1/tan d) and a
near-zero temperature coefficient of resonant frequency (tf) [3]. High dielectric constant material can effectively
pffiffiffi
reduce the size of resonators since the wavelength (l) in dielectrics is inversely proportional to er of the wavelength
pffiffiffi
(l0) in vacuum (l ¼ l0= er). However, due to the shift of the frequency of interest from 900 MHz to 2.4, 5.2 or even
5.8 GHz, their dielectric constants need not to be as high as required in the past. In terms, high quality factor (inverse of
* Corresponding author.
E-mail address: huangcl@mail.ncku.edu.tw (C.-L. Huang).
0025-5408/$ – see front matter # 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2007.06.032

1464
C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
the dielectric loss, Q = 1/tan d) becomes more important since Q ꢀ f is almost constant in the microwave region. In
general, high dielectric constant materials exhibit high dielectric loss (low Q ꢀ f value), while low loss ceramics are
usually accompanied by low er value. Low loss dielectrics, however, normally require high sintering temperatures.
Low melting glass addition, chemical processing and small particle sizes of the starting materials are generally
advantageous to reduce the sintering temperature of dielectric materials [4–7]. Using these methods, it is possible to
reduce the firing temperatures. Dielectric materials with lower dielectric constants often exhibit higher unloaded
quality factors.
Alumina ceramics is well known as the material for substrates, dielectric resonators and patch antennas.
Commercially available alumina normally has a Q ꢀ f value around 10,000–20,000 GHz. With TiO2 addition,
alumina ceramics possesses a quality factor (Q ꢀ f) of 333,000 GHz at 1500 8C for 5 h [8]. Using nano-particle-sized
starting material, a-Al2O3 ceramics can be sintered at 1500 8C and promotes its Q ꢀ f value to ꢂ500,000 GHz [9]. In
this paper, according to the Furnas packing model [10], the nano (a + u)-Al2O3 ceramic powder is composed of
70 wt.% a-Al2O3 and 30 wt.% u-Al2O3. In this ratio, the mixed powder can be easily sintered at temperatures as low as
1400 8C, preparing high density with nano-scaled grain size. The 30/70 u/a powder mixture not only lowered its
sintering temperature but also decreased microwave dielectric loss. The dielectric properties at microwave frequencies
of the sintered ceramics were characterized. The relation between crystalline phases and microstructures of the
samples were also investigated.
2. Experimental procedure
The starting materials were high-purity oxide powders (>99.99%): 70 wt.% a-Al2O3 and 30 wt.% u-Al2O3. The
particle sizes of a-Al2O3 and u-Al2O3 were 50 and 25 nm, respectively. The primary particle size of the starting
dried powder is 150 nm. Table 1 lists some basic physical properties of the powder. Particle size distribution of the
powder in slurry state was measured using laser light scattering methods (Zetasizer 1000, Malvern). Then, the
synthesized method was described below. The 30/70 u/a powder mixture were weighed, dispersed in reagent-grade
ethanol. The solid content of the powder suspension was 45 wt.%. Phosphate ester (Re 610, Toho Japan) was used
as the dispersant with 1.5 wt.% of the powder weight. The powder suspension was ball mixed in polyethylene bottle
for a period of 15 h with yttrium-stabilized-zirconia (YTZ) beads (Ø = 2 mm) as the grinding media. After ball
mixing, the powder suspension was dried overnight. The dried powder was milled for 5 h with 5 wt.% of a 10%
solution of PVA as a binder. Pellets with 11 mm in diameter and 5 mm in thickness were pressed using uniaxial
pressing. A pressing pressure of 400 MPa was used for all samples. A dilatometer (DIL420 C, Natzsch) was
employed to examine the thermal shrinkage behavior before the compacts were thermal treated at 1400 8C for
2–8 h.
The crystalline phases of the sintered ceramics were identified by XRD using Cu Ka (l = 0.15406 nm) radiation
with a Seimens D5000 diffractometer operated at 40 kVand 40 mA. Before SEM analysis, the samples were put into
the ultrasonic machine by acetone for 15 min. Therefore, the impurities on the surface of the samples were removed.
Then, the surfaces of the samples were coated by Pt for the increase of the electric conduction, so that the clear and
small grain size could be observed. The apparent densities of the sintered pellets were measured by the Archimedes
method. The dielectric constant (er) and the quality factor values (Q) at microwave frequencies were measured using
the Hakki–Coleman dielectric resonator method [11,12]. A system combining a HP8757D network analyzer and a
HP8350B sweep oscillator was employed in the measurement. For temperature coefficient of resonant frequency (tf),
Table 1
Basic physical properties of starting powders
Sample name
A70
Xtallite sizea (nm)
a-Phase
u-Phase
50
25
BET (m2/g)
26
a
XRD-Scherrer formula.

C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
1465
Fig. 1. (a) The particle size distribution of the starting powders. (b) The SEM micrograph (ꢀ30 k) of starting powder of the nano (a + u)-Al2O3
ceramics. (c) XRD phase identification of starting powders.

1466
C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
Fig. 2. (a) The DTA profile of nano-Al2O3 green samples. (b) The thermal shrinkage curve of the green compact prepared by starting powders.
the technique is the same as that of quality factor measurement. The temperature coefficient of resonant frequency (tf)
at microwave frequencies was measured in the temperature range from 25 to 80 8C and is defined by
D f0
tf ¼
ðppm=ꢃCÞ
(1)
f0DT
where Df0 is the shift in the center frequency introduced by temperature change DT.
Fig. 3. X-ray diffraction patterns of the nano (a + u)-Al2O3 ceramics sintered at 1400 8C for (a) 2 h, (b) 4 h and (c) 8 h.

C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
1467
3. Results and discussion
Fig. 1(a) shows that 90% of the powder is finer than 300 nm in particle sizes and the primary particle size of the
starting dried powder is 150 nm. The nanometer-scaled particles were also observed through the SEM micrograph as
demonstrated in Fig. 1(b). Fig. 1(c) shows XRD phase identification of starting powders. The result illustrates the
powder is composed of u- and a-Al2O3 phases. Therefore, the mixture is bimodal. The relative density of the green
sample of nano (a + u)-Al2O3 ceramics reached 55% at a pressing pressure of 400 Mpa. The reason was according to
the pressing pressure for nano-scaled alpha alumina. The pressure was suitable to form green sample of nano-scaled
alpha alumina with high relative density. This was because appropriate pressing pressure was helpful to get dense
samples easily. Table 2 shows that the relative density of green sample of nano-scaled alpha alumina reached 53.1% at
a pressing pressure of 400 MPa. The appropriate pressing pressure could destroy the agglomeration in the nano-scaled
alpha alumina powder and gain high relative density of green sample. As shown in the Table 2, more than the pressing
Fig. 4. SEM micrographs (ꢀ30 k) of the specimens sintered at 1400 8C using nano (a + u)-Al2O3 for (a) 2 h, (b) 4 h and (c) 8 h, and using nano a-
Al2O3 for (d) 2 h, (e) 4 h and (f) 8 h.

1468
C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
Table 2
Relations of relative density of green samples and pressing pressure for nano-scaled alpha alumina
Relative density of green samples (%)
Pressing pressure (MPa)
45.7
49.5
53.1
53.4
200
300
400
500
pressure of 400 MPa was not necessary for getting excellent relative density of green samples. Therefore, we adopted
the pressing pressure of 400 MPa for nano-scaled high-purity (a + u)-Al2O3 powders in agreement with that for pure
nano-scaled alpha Al2O3 in Ref. [9]. Therefore, removing the agglomeration in the green body by appropriate pressing
pressure would obtain a sintering temperature lower than 1400 8C.
The DTA profile of sample is shown in Fig. 2(a). Phase transformation from u to a for u-Al2O3 occurred at 1200 8C.
However, it was increased to 1270 8C as the sample is a mixture of (a + u)-Al2O3. The dilatometric curve shows that
the thermal shrinkage started at the temperature of 1100 8C with a maximum occurring at the temperature of 1350 8C,
as shown in Fig. 2(b). The lower starting shrinkage temperature may result from u- to a-phase transformation.
Fig. 3 shows the XRD patterns of nano (a + u)-Al2O3 ceramics sintered at 1400 8C for different sintering time. At
1400 8C, the crystalline phases of the nano (a + u)-Al2O3 ceramics have become a-alumina ceramics completely.
Phase transformation occurred at 1270 8C as illustrated in Fig. 2. It is understood that the crystal system of the
Fig. 5. (a) The relative densities of the nano (a + u)-Al2O3 ceramics as a function of its sintering time at 1400 8C. (b) The relative pore volume of the
nano (a + u)-Al2O3 ceramics as a function of its sintering time at 1400 8C.

C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
1469
Fig. 6. The dielectric constant of the nano (a + u)-Al2O3 ceramics as a function of its sintering time at 1400 8C.
a-alumina ceramics is hexagonal (JCPDS #82-1468). Similar X-ray diffraction patterns were detected for the
specimens at different sintering time. No second phase was observed throughout the entire experiment.
The SEM micrographs of the nano-Al2O3 ceramics sintered at 1400 8C for different sintering time are illustrated in
Fig. 4. For specimen with starting material nano a-Al2O3, its average grain size increased from 610 to 980 nm as the
sintering time increased from 2 to 8 h. However, the variation was from 280 to 540 nm while nano (a + u)-Al2O3 was
selected as the starting powders. It indicates that grain growth can be limited with appropriate u-phase addition.
Comparing to uniform morphology from pure a-Al2O3, matrix of (a + u)-Al2O3 obviously consisted of two groups of
grains. Large grains were attributed to pure a-Al2O3 through grain growth. However, small ones were results from u-
Al2O3 since u-Al2O3 possesses relatively slow grain growth even after its transformation to a-phase. Moreover, small
grains would not be combined by large ones during grain growth. The porosity of sample decreased as the sintering
time was extended. Significant grain growth was observed and the pores almost disappeared at 1400 8C for 8 h. These
may directly affect the microwave dielectric properties of the nano (a + u)-Al2O3 ceramic samples.
The relative densities of the nano (a + u)-Al2O3 ceramics (apparent density/theoretical density; JCPDS #82-1468)
as a function of its sintering time at 1400 8C are indicated in Fig. 5(a). It shows that 97.12–99.49% relative densities
were obtained for the nano (a + u)-Al2O3 ceramics sintered at 1400 8C from 2 to 8 h. Comparing to pure a-Al2O3 [9],
it possesses a higher relative density with appropriate u-Al2O3 addition. As shown in Fig. 4, SEM morphology was a
combination of large and small grains. The small grains could occupy pores from large ones and squeezes each other
resulted in an increase in the density. With increasing sintering temperature, the relative density could increase to
Fig. 7. The Q ꢀ f value of the nano (a + u)-Al2O3 ceramics as a function of its relative density.

1470
C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
Fig. 8. The temperature coefficients of resonant frequency (tf) of the nano (a + u)-Al2O3 ceramics as a function of its sintering time at 1400 8C.
99.49% of its theoretical density at 1400 8C for 8 h. Fig. 5(b) shows the relative pore volume of the nano (a + u)-Al2O3
ceramics as a function of its sintering time at 1400 8C. The pores in the sample decrease with the increase of sintering
time. The increase in relative density with increasing sintering time was due to the decrease in the number of pores as
well as grain growth as observed in Fig. 4.
Fig. 6 shows the dielectric constant of the nano (a + u)-Al2O3 ceramics as a function of its sintering time at
1400 8C. The variation of er value was consistent with that of relative density since higher relative density represents
lower porosity. When samples were sintered at 1400 8C from 2 to 8 h, the er values slightly increased and saturated at
10 for the well-sintered ceramics.
The quality factor value (Q ꢀ f) of the nano (a + u)-Al2O3 ceramics as a function of its relative density was
demonstrated in Fig. 7. With increasing relative density to maximum value, the Q ꢀ f value increased and reached a
value of 634,000 GHz when the sample of the nano (a + u)-Al2O3 ceramics was sintered at 1400 8C for 8 h. Relative
density plays an important role in controlling the dielectric loss, and has been shown for other microwave dielectric
materials. Since the Q ꢀ f value of the nano (a + u)-Al2O3 ceramics was consistent with the variation of relative
density, it suggested the dielectric loss of the nano (a + u)-Al2O3 ceramics was mainly controlled by relative density.
Fig. 8 illustrates the temperature coefficients of resonant frequency (tf) of the nano (a + u)-Al2O3 ceramics as a
function of its sintering time at 1400 8C. In general, the temperature coefficient of resonant frequency (tf) was related
to the composition and the phases that existed in the ceramics. Since there is no second phase, significant change was
not observed in the tf value at different sintering time. It possesses a temperature coefficient of resonant frequency (tf)
of ꢁ39.88 ppm/8C for specimen sintered at 1400 8C for 8 h.
4. Conclusion
The microstructures and the microwave dielectric properties of the nano (a + u)-Al2O3 ceramics were investigated.
The microwave dielectric properties are strongly related to relative density. High-density alumina ceramics can be
achieved since small grains (from u-phase) could occupy pores from large ones (from a-phase) and squeezes each
other to eliminate pores. The tf value was not significant changed at different sintering time. In this work, a dielectric
constant (er) of 10, a high Q ꢀ f value of 634,000 GHz (measured at 14 GHz) and a tf value of ꢁ39.88 ppm/ 8C were
obtained for the nano (a + u)-Al2O3 ceramics sintered at 1400 8C for 8 h. Specimen can be sintered to a relative
density higher than 90% at low temperature and still possesses excellent microwave dielectric properties. The nano
(a + u)-Al2O3 ceramics is a very promising material for extremely low loss microwave dielectric components.
Acknowledgements
This work was supported by Ministry of Economic Affairs, Taiwan (92-EC-17-A-08-S1-023) through Particulate
Materials Research Center at National Cheng Kung University.

C.-L. Huang et al. / Materials Research Bulletin 43 (2008) 1463–1471
1471
References
[1] S. Nishigaki, H. Kato, S. Yano, R. Kamimure, Am. Ceram. Soc. Bull. 66 (1987) 1405.
[2] K. Wakino, K. Minai, H. Tamura, J. Am. Ceram. Soc. 67 (1984) 278.
[3] T. Takada, S.F. Wang, S. Yoshikawa, S.J. Yang, R.E. Newnham, J. Am. Ceram. Soc. 77 (1994) 1909.
[4] T. Kakada, S.F. Wang, S. Yoshikawa, S.J. Jang, R.E. Newnham, J. Am. Ceram. Soc. 77 (1994) 1909.
[5] T. Kakada, S.F. Wang, S. Yoshikawa, S.J. Jang, R.E. Newnham, J. Am. Ceram. Soc. 77 (1994) 2485.
[6] S.I. Hirno, Takashi, Hayashi, A. Hattori, J. Am. Ceram. Soc. 74 (1991) 1320.
[7] V. Tolmer, G. Desqardin, J. Am. Ceram. Soc. 80 (1997) 1981.
[8] N.M. Alford, S.J. Penn, J. Appl. Phys. 80 (1996) 5895.
[9] C.-L. Huang, J.-J. Wang, C.-Y. Huang, Mater. Lett. 59 (2005) 3746.
[10] J.S. Reed, Principles of Ceramics Processing, 2nd., Wiley, 1995, pp. 219–222.
[11] B.W. Hakki, P.D. Coleman, IEEE Trans. Microw. Theory Tech. 8 (1960) 402.
[12] W.E. Courtney, IEEE Trans. Microw. Theory Tech. 18 (1970) 476.