Effect of TiO2 doping on crystallization, microstructure and dielectric properties of CBS glass-ceramics

Article abstract of DOI:10.1016/j.jallcom.2019.05.152

In this study, calcium borosilicate (CaO-B₂O₃-SiO₂, CBS) glass-ceramics were fabricated using chemically pure CaO, SiO₂, and B₂O₃ as raw materials. Differential scanning calorimetry, X-ray diffraction, scanning electron microscopy, and electrical measurements were conducted to explore the effect of titanium dioxide (TiO₂) doping on crystallization, microstructure, and dielectric properties of CBS glass-ceramics. Furthermore, the influence of sintering temperature and sintering schemes on crystalline phases of CBS glass-ceramics was systematically investigated. Results showed that the increase of TiO₂ content led to the reduction in sintering temperature of CBS glass-ceramics and promoted the precipitation of wollastonite crystal phase. For instance, with the increase of content of TiO₂ from 0 to 3 wt%, transition temperature and crystallization temperature were reduced by 12.9 and 27.6 °C, respectively. However, excessive TiO₂ affected the precipitation of wollastonite crystal phase, destroyed crystal structure, and damaged close arrangement of crystal grains. Moreover, higher TiO₂ content was found to weaken dielectric properties of CBS glass-ceramics. In this study, the best molar ratio of ingredients, meeting that of ideal target material, is n(Ca): n(Si): n(B)= 1:1:0.6, with doping mass fraction of 2 wt% of TiO₂. After optimal sintering procedure, dielectric constant of the best sample was found to be 1.4 (1 MHz), 1.3 (10 MHz), and dielectric loss was 3.9 × 10^?3 (1 MHz) and 3.8 × 10^?3 (10 MHz).

Full text of DOI:10.1016/j.jallcom.2019.05.152

Accepted Manuscript
Effect of TiO doping on crystallization, microstructure and dielectric properties of
2
CBS glass-ceramics
Dongfeng He, Hao Zhong, Chong Gao
PII:
S0925-8388(19)31832-8
DOI:
https://doi.org/10.1016/j.jallcom.2019.05.152
JALCOM 50684
Reference:
To appear in: Journal of Alloys and Compounds
Received Date: 2 April 2019
Revised Date: 11 May 2019
Accepted Date: 13 May 2019
Please cite this article as: D. He, H. Zhong, C. Gao, Effect of TiO doping on crystallization,
2
microstructure and dielectric properties of CBS glass-ceramics, Journal of Alloys and Compounds
(2019), doi: https://doi.org/10.1016/j.jallcom.2019.05.152.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Effect of TiO doping on crystallization, microstructure and dielectric
2
properties of CBS glass-ceramics
Dongfeng He*, Hao Zhong, Chong Gao
Department of Ferrous Metallurgy, School of Metallurgical and Ecological Engineering, University of Science and
Technology Beijing, Beijing 100083, China
*Corresponding author.
E-mail address: hdfcn@163.com (D. He)
Abstract: In this study, calcium borosilicate (CaO-B O -SiO , CBS) glass-ceramics
2
3
2
were fabricated using chemically pure CaO, SiO , and B O as raw materials.
2
2
3
Differential scanning calorimetry, X-ray diffraction, scanning electron microscopy,
and electrical measurements were conducted to explore the effect of titanium dioxide
(
TiO ) doping on crystallization, microstructure, and dielectric properties of CBS
2
glass-ceramics. Furthermore, the influence of sintering temperature and sintering
schemes on crystalline phases of CBS glass-ceramics was systematically investigated.
Results showed that the increase of TiO content led to the reduction in sintering
2
temperature of CBS glass-ceramics and promoted the precipitation of wollastonite
crystal phase. For instance, with the increase of content of TiO from 0 to 3 wt.%,
2
transition temperature and crystallization temperature were reduced by 12.9 and
2
7.6 °C, respectively. However, excessive TiO affected the precipitation of
2
wollastonite crystal phase, destroyed crystal structure, and damaged close
arrangement of crystal grains. Moreover, higher TiO content was found to weaken
2
dielectric properties of CBS glass-ceramics. In this study, the best molar ratio of
ingredients, meeting that of ideal target material, is n(Ca): n(Si): n(B) = 1:1:0.6, with
doping mass fraction of 2 wt.% of TiO . After optimal sintering procedure, dielectric
2
constant of the best sample was found to be 1.4 (1 MHz), 1.3 (10 MHz), and dielectric
−3
−3
loss was 3.9 × 10 (1 MHz) and 3.8 × 10 (10 MHz).
Keywords: Calcium borosilicate, Glass-ceramics, Dielectric properties, Low sintering
temperature, Titanium dioxide

ACCEPTED MANUSCRIPT
1
. Introduction
With the rapid development of modern information technology, large-scale and
ultra-large-scale integrated circuits have become an important direction of research
and development. Electronic packaging technology has evolved from traditional
high-density single-layer packaging to stacked three-dimensional packaging, thus
multi-layer wiring packaging technology is widely used nowadays [1–3]. Low
temperature co-fired ceramic (LTCC) materials, multi-chip modules, chip size
package, etc., have emerged in recent years as well-known multidisciplinary
components for integration technology [4–6]. Among them, LTCC has become the
mainstream of electronic component integration design with its excellent integration
density and high frequency characteristics, such as electrical, mechanical, thermal,
and process characteristics [7–12].
Calcium borosilicate (CaO-B O -SiO , CBS) glass-ceramics have low dielectric
2
3
2
constant, low dielectric loss, and thermal expansion coefficient similar to that of
silicon chips. Moreover, these glass-ceramic materials have an excellent match with
metallic pastes and ability to reduce high-frequency signal relaxation. Therefore, CBS
glass-ceramic materials are considered as one of the most promising LTCC substrate
materials to create robust electronics [13–15]. Significant research efforts have been
devoted to the study of the component ratios of CBS glass-ceramics, the heat
treatment system, and the influence of nucleating agents on their microstructure and
properties. As a result, some excellent CBS glass-ceramic materials have also been
obtained [16–18].
Zhou et al. [19] studied the effects of different Ca/Si ratios on the microstructure
and properties of CBS glass-ceramics. Zhu et al. [20] used CBS glass-ceramic as raw
material and added 0.5%P O and 0.5%ZnO to fabricate LTCC. Chiang et al. [21]
2
5
systematically studied the densification behavior and properties of CBS
glass-ceramics with different compositions, and analyzed the effects of temperature,
heating rate, and holding time on the properties of glass-ceramics. Shao et al. [22]
prepared CBS glass-ceramic by low temperature water quenching method, and
studied the effect of composition on the melting temperature and the influence of
added titanium dioxide (TiO ) on the dielectric properties of CBS glass-ceramics.
2
Chang et al. [23] studied the crystallization kinetics and mechanism of low dielectric,

ACCEPTED MANUSCRIPT
low temperature and co-fired CBS glass-ceramics.
He and Gao [24] investigated the effect of added boron on the crystallization,
microstructure, and dielectric properties of CBS glass-ceramics. Zhu et al. [25] used
P O and ZnO as additives to optimize the synthesis conditions of CBS glass-ceramics
2
5
for LTCC. Hsiang et al. [26] studied the effects of ZrO addition on the crystallization
2
behavior, sintering characteristics, and dielectric properties of CaO-MgO-Al O -SiO
2
2
3
glass-ceramics. Zhou et al. [27] investigated the effects of La O -B O as a sintering
2
3
2
3
aid on the flexural strength and microwave dielectric properties of CBS
glass-ceramics. Furthermore, Wang et al. [28] studied the effects of the addition of
Li CO , TiO , and Al O on the densification and microwave dielectric properties of
2
3
2
2
3
CBS glass-ceramics.
As a nucleating agent, TiO has always been the focus of research. For instance,
2
G. A. Khater [29] explored the effect of TiO nucleating agent on the crystallization
2
behavior and microstructure of glass-ceramics based on blast furnace slag. Yang et al.
[
30] studied the effect of TiO addition on the crystallization, microstructure, and
2
mechanical properties of stainless steel slag glass-ceramics. Rezvani et al. [31]
investigated the effects of Cr O , Fe O , and TiO as nucleating agents on the
2
3
2
3
2
crystallization behavior of SiO -Al O -CaO-MgO(R O) glass-ceramics. Mukherjee et
2
2
3
2
al. [32] analyzed the influence of TiO content on the crystallization and
2
microstructure of machinable glass ceramics. Ebadzadeh and Ghasemi [33] studied
the effect of adding TiO on the stability of t-ZrO in mullite-ZrO . Dugué et al. [34]
2
2
2
studied the effects of structural transformation and spectral properties of Ni-doped
magnesium aluminosilicate glass-ceramics doped with TiO and ZrO nucleating
2
2
agents on broadband near-infrared light. The above mentioned studies clearly show
that TiO as a nucleating agent can significantly promote the crystallization of
2
glass-ceramics and improve their performance. However, studies focusing on the
content of TiO are rare and the influence of TiO as nucleating agent on the crystal
2
2
phases and microstructure of CBS glass-ceramics has not been explored yet.
Therefore, the role of TiO in CBS glass-ceramics should be further investigated.
2
In this study, CBS glass-ceramics were prepared from chemically pure materials.
The effects of different contents of TiO and sintering process on the crystal phase and
2
microstructure of CBS glass-ceramics were studied. Moreover, the effect of TiO on
2
the dielectric properties of glass-ceramics was analyzed from the perspective of
crystal structure and microstructure evolution. Finally, the optimum doping content of

ACCEPTED MANUSCRIPT
TiO and heat treatment system were determined to obtain CBS glass-ceramics with
2
excellent properties.
2
Experimental
2
.1 Raw materials and design
CBS glass-ceramics are wollastonite-based crystalline materials. Wollastonite
−4
has low dielectric constant (ε = 5) and dielectric loss (tanδ = 1–3 × 10 ). In the
r
preparation of CBS glass-ceramics, different crystalline phases such as CaB O ,
2
4
CaB SiO , CaSiO , and CaSiO precipitate due to different sintering systems and
2
8
3
4
compositions. Different types and quantities of crystalline phases can be obtained by
selecting suitable sintering system to obtain materials with different properties [35–
3
7].
Chemically pure raw materials were used in this study. Distribution ratio of the
raw materials was Ca Si B On. According to the preliminary experimental results [24],
x
y z
we configured the raw materials according to the molar ratio of the elements in the
composition X:Y:Z = 1:1:0.6. The molar fraction of the raw materials was converted
into mass fraction, and then the compounding scheme was set according to the
different doping amount of TiO , the detailed composition of formulation I, II, III, and
2
IV is listed in Table 1.
Table 1. The detailed composition of raw materials according to different
formulations (g).
Formulation
Formulation I
Formulation II
Formulation III
Formulation IV
CaO
40.88
40.47
40.06
39.65
SiO2
43.79
43.36
42.92
42.48
B O
TiO2
2
3
15.33
15.18
15.02
14.87
0
1
2
3
2
.2 Experiment process
The ingredients were prepared according to the designed ingredient list (Table 1)

ACCEPTED MANUSCRIPT
and operated according to the experimental procedure, schematically illustrated in Fig.
1
[38–40].
Fig. 1. Flow chart for the glass melting process and characterization
2
.3 Characterization
Differential scanning calorimetry (DSC) was performed (STA409C integrated
thermal analyzer, NETZSCH Co.) to determine the heat treatment system of
glass-ceramics. Briefly, the sample for DSC was crushed into powder using a DF-4
electromagnetic type ore grinder and sieved through a 200 mesh. The powder sample
was weighed using an electronic balance and 20 mg of the sample was used for DSC
experiments. α-Al O was used as the standard reference, argon gas was used as the
2
3
−1
shielding gas, and the heating rate was 10 °C min .
The phase composition was determined by X-ray diffraction (XRD) using an
M21X-SRA X-ray diffractometer (MAC Science Co., Ltd, Japan) equipped with
graphite crystal diffracted-beam monochromator. The working voltage and the working
current of the tube were 40 KV and 300 mA, respectively. CuKα radiations with a

ACCEPTED MANUSCRIPT
wavelength of 1.5406 Å were used as the X-ray source and XRD patterns were
recorded in the 2θ range of 20–60° at a scanning speed of 0.044°/sec.
Microstructure of the prepared samples was characterized by scanning electron
microscopy (SEM, ZEISS, Germany). Before microstructural analysis, the sample
was etched for 15 s in 2 wt.% hydrofluoric acid solution, ultrasonically cleaned for 3
min, dried, and then sprayed with gold. The working voltage of 20 kV and working
current of 4–20000 pA were used. The maximum resolution of SEM images was 1
nm.
The dielectric properties of the samples were measured using an impedance
analyzer (E4990A, Keysight, USA). Sample was first sprayed with gold on both sides
to maintain electrical conductivity. The frequency range of the E4990A impedance
analyzer was from 20 Hz to 120 MHz, providing a basic accuracy of 0.045% over a
wide impedance range.
The density of the glass-ceramics was measured using the Archimedes method.
The dry weight m and hanging weight m of the sample were determined by using an
1
2
electronic balance. Sample density
ρ
was calculated as follows:
m1
ρ
=
×
ρH
(2-1)
m − m2
1
3
ρH is the density of the plasma water, 1 g/cm .
3
Results and discussion
Uniformly mixed powder was poured into a platinum crucible and heated in a
muffle furnace to 1360 °C. The melted glass liquid was obtained by maintaining the
temperature for 2 h at 1360 C. The molten glass was then quickly poured into
°
deionized water at room temperature to obtain water quenched sample. The
water-quenched samples of CBS glass-ceramics based on the formulations I, II, III,
and IV were crushed and sieved through a 200-mesh sieve before DSC measurements.
The results of the test are shown in Fig. 2.
The endothermic peak of DSC curve corresponds to the transition temperature
(
T ) and softening temperature (T ) of glass, and the exothermic peak corresponds to
g
f
the crystallization temperature of glass. The results of the DSC detection show that

ACCEPTED MANUSCRIPT
the transition temperatures of formulation I, II, III, and IV are 683.5, 676.2, 672.3,
and 670.6
and 857.7
℃
, respectively, and the crystallization temperatures are 885.3, 872.6, 861.9,
℃
, respectively. This indicates that the T and crystallization temperatures
g
gradually decrease with the increase of TiO content. When the mass percentage of
2
TiO increases from 0 to 3%, the transition temperatures decrease by 12.9
℃
, and the
2
crystallization temperature decreases by 27.6 °C. This can be attributed to the
4
+
existence of Ti in tetracoordinated state in the molten glass and the base glass
4
+
4+
formed by extreme cold. Radius of Ti is larger than that of Si , and the bond energy
4
+
4+
of the Ti–O bond is lower than that of Si–O; therefore, Ti partially replaces Si in
the silicon–oxygen tetrahedron structure, thus destroying the complex silicon
tetrahedral system, reducing the viscosity of glass, increasing its fluidity, facilitating
the diffusion of ions, and finally reducing the crystallization temperature of glass [29].
The peak value of exothermic peak represents the crystallization strength, and the
exothermic peak corresponding to formula IV is the sharpest, which indicates that the
crystallization of glass-ceramics is enhanced with the increase of TiO content.
2
Fig. 2. The DSC curves of CBS glass-ceramics with different compositions
Samples with formulation I were molded and placed in a muffle furnace for
heating and sintering. According to the DSC image of formulation I, different
sintering systems were designed to prepare samples 1–5, and the specific scheme is

ACCEPTED MANUSCRIPT
summarized in Table 2. In general, glass powder was heated to the vicinity of the
glass softening temperature, maintained for one hour, heated to near crystallization
temperature, and then followed by two hours insulation. This heat treatment system is
referred to as two-step sintering.
Table 2. Different sintering regimes for formulation I.
Sample no.
Sintering regimes
−1
Heating up to 800 °C at a heating rate of 3 °C min . After 2
1
2
3
#
#
#
h of holding time, cooling with a muffle furnace
−1
Heating up to 900 °C at a heating rate of 3 °C min . After 2
h of holding time, cooling with a muffle furnace
−1
Heating up to 900 °C at a heating rate of 5 °C min . After 2
h of holding time, cooling with a muffle furnace
−1
Heating up to 700 °C at a heating rate of 3 °C min . After 1
h of holding time, the temperature was increased to 850 °C
4
5
#
#
−1
(
3 °C min ). After 2 h of holding time, the samples were
cooled down with muffle furnace.
−1
Heating up to 700 °C at a heating rate of 3 °C min . After 1
h of holding time, the temperature was increased to 900 °C
−1
(
3 °C min ). After 2 h of holding time, the samples were
cooled down with muffle furnace.
XRD measurements were used to explore the effects of different composition
and sintering system on the precipitation, phase composition, and crystal phase
transformation of CBS glass-ceramics. SEM was utilized to study the effects of
different composition and sintering system on the microstructure and densification
sintering of CBS glass-ceramics. The impedance analyzer was used to study the
dielectric properties of CBS glass-ceramics and determine the optimal sintering
system.
Fig. 3 shows the XRD patterns of samples 1–5.

ACCEPTED MANUSCRIPT
Fig. 3. The XRD patterns of samples 1–5.
XRD patterns of samples 1–5 reveal that the main crystalline phase of each
sample is wollastonite, with a small amount of quartz and dicalcium silicate. The ratio
of CaO to SiO in the batch is 1:1, which conforms to the composition design of
2
wollastonite. However, some Ca and wollastonite combine to form Ca SiO ; therefore,
2
4
the remaining (SiO ) is self-associated to form cristobalite phase. Sample 1 was
4
sintered at 800 °C, which is lower than its crystallization temperature of 885.3 °C,
thus resulting in a relatively low content of wollastonite crystal phase. The crystal
phases of samples 2–5 are roughly similar. Compared to the crystal phase of sample 1,
the amount of wollastonite crystal phases in the samples 2–5 gradually increases, and
the quartz crystal phase content relatively decreases. This shows that increasing
sintering temperature and adopting two-step sintering method can make the samples
precipitate more wollastonite crystalline phase. Two-step sintering can lead to the
growth and nucleation of nucleus at low nucleation rate. When numerous nuclei are
uniformly formed in the glass, the temperature increases to the temperature near the
maximum value of crystal growth rate for heat preservation treatment. Thus, a large
number of uniform microcrystals can be grown in the glass, so that the samples can
acquire excellent properties.
Fig. 4 shows the SEM images of samples 1–5. In general, the microstructures of
samples 2 and 3 and 4 and 5 are, respectively, similar; therefore, the SEM images of
samples 1, 3, and 5 are presented herein. Sample 1 shows a large amount of glass

ACCEPTED MANUSCRIPT
liquid phases, few crystal grains, and a large number of voids, which can be attributed
to the incomplete sintering. Compared to sample 1, sample 3 shows a relative
decrease in the liquid phase and increase in crystalline grains, and most grains are
closely connected with each other, which indicates significant improvement in
crystallization. However, as-formed sample 5 is dense with a large number of acicular
wollastonite grains, and the crystal grains are full and closely arranged, with average
particle size of about 5 µm.

ACCEPTED MANUSCRIPT
Fig. 4. SEM images of samples 1, 3, and 5 (See Table 2 for sintering regimes).
The dielectric properties of samples 1–5 were analyzed using an impedance
analyzer and the density of the samples was measured using the Archimedes method.
The results are presented in Table 3 (Fig. 5). The dielectric constants of samples 1–5
are generally lower than 7, which meet the expected target. Furthermore, the density
of samples 1–5 increases gradually, which indicates that the densification of structure
was enhanced. It should be noted that the density of sample 3 is slightly larger than
sample 2 and sample 5 is slightly larger than sample 4, which is why the dielectric
constant of sample 3 is lower than sample 2, and the dielectric constant of sample 5 is
lower than sample 4. In addition, the dielectric constant of sample 3 is comparatively
lower, which is 5 at 1 MHz and 10 MHz, however, its dielectric loss is higher,
−2
reaching 1.27
×
10 . The dielectric constant of sample 5 is 6, and the dielectric loss
−3
−3
is 2.3 ×ꢀ10 at 1 MHz and 3.37 ×ꢀ10 at 10 MHz, which can be attributed to good
crystallization and dense structure. Moreover, the dielectric properties of sample 5
meet the expected requirements of dielectric constant and dielectric loss of less than 7
−3
and 5 × 10 , respectively. Considering dielectric constant and dielectric loss, sample
5
is the optimal sample and meets the market requirements.
Table 3. The dielectric constant, dielectric loss and density of samples 1–5.
Density
Frequency
(MHz)
Dielectric
constant ( r
Dielectric
Sample no.
1
3
–3
(
g/cm )
ε
)
loss(10 )
1
8
7
6
6
5
5
7
8
6
3.9
1.97
9
.8
1
0
0
1
3.2
1
2
3
2.26
2.28
1
1
18.7
6.0
1
1
2.7
.9
0
5
1
4
5
2.34
2.37
1
0
18.7
2.3
1

ACCEPTED MANUSCRIPT
1
0
6
3.4
Fig. 5. The dielectric constant and dielectric loss of samples 1–5.
The results of XRD, SEM, and the dielectric properties of samples 1–5 were
comprehensively analyzed. The results indicated that the increase of the sintering
temperature or the use of two-step sintering could precipitate more wollastonite
crystal phase and enhance sample performance. Therefore, among formulations I, II,

ACCEPTED MANUSCRIPT
III, and IV, the optimal sintering system was selected to prepare the samples. The
details are presented in Table 4.
Table 4. Sintering regimes for formulations I, II, III, and IV
Sample
no.
Sintering regimes
−
1
Heating up to 700 °C at a heating rate of 3 °C min .
After holding for 1 h, the temperature was increased
Formulation I
Formulation II
Formulation III
Formulation IV
5
6
7
8
−
1
to 900 °C (3 °C min ). After holding for 2 h,
samples were cooled down in a muffle furnace
−
1
Heating up to 700 °C at a heating rate of 3 °C min .
After holding for 1 h, the temperature was increased
−
1
to 892 °C (3 °C min ). After holding for 2 h,
samples were cooled down in a muffle furnace
−
1
Heating up to 700 °C at a heating rate of 3 °C min .
After holding for 1 h, the temperature was increased
−
1
to 882 °C (3 °C min ). After holding for 2 h,
samples were cooled down in a muffle furnace
−
1
Heating up to 700 °C at a heating rate of 3 °C min .
After holding for 1 h, the temperature was increased
−
1
to 872 °C (3 °C min ). After holding for 2 h,
samples were cooled down in a muffle furnace
XRD measurements were carried out to analyze the crystal structure of samples
–8 and results are shown in Fig. 6.
5
The main crystalline phase of each sample is wollastonite. With the increase of
the content of TiO , the diffraction peak of the main crystalline phase becomes
2
sharper and the peak value is higher, which indicates the presence of more
wollastonite phase in samples and also the samples undergo better crystallization.
However, in sample 5 without TiO doping, only a small amount of crystal grains is
2
formed due to the difficulty of nucleation and crystal growth. Few wollastonite phase
and more danburite and quartz phases are observed in SEM image of sample 5. These
results indicate that TiO doping contributes to the precipitation of wollastonite
2
crystalline phase. Moreover, noteworthy, a small amount of anatase (TiO ) and
2

ACCEPTED MANUSCRIPT
titanium silicide (TiSi ) were also detected for sample 7, and these impurities
2
crystalline phases increased significantly in sample 8, which led to the weakening of
the properties of the sample. This result indicates that an optimum amount of TiO2
doping is required, and if the amount exceeds the optimum value, new crystalline
phases are introduced.
Therefore, the effect of TiO doping on CBS glass-ceramics was further
2
investigated. According to the analysis, in the glass network structure, Ti ions act as
network intermediates, and elemental Ti exists in the tetracoordinated and
hexacoordinated states in the glass. In the molten glass and the extremely cold formed
base glass, Ti ions exist in a tetracoordinate state. However, Ti ions exist in a
hexacoordinated form at the nucleation temperature, and this structure is greatly
different from the network structure formed by glass. When the content exceeds the
solubility, it precipitates out from the glass matrix to form a Ti-rich phase, which
causes the glass to be phase-separated, thereby promoting glass crystallization.
Fig. 6. XRD patterns of samples 5–8.
The factsage software was used to simulate the change of the crystal phase of the
CBS-based glass-ceramic system in the temperature range of 500–1000 °C and results
are shown in Fig. 7.
According to the software calculation results, the wollastonite phase content of
the samples with 1 and 2 wt.% (mass fraction) of TiO increased gradually, while the
2

ACCEPTED MANUSCRIPT
content of quartz and cyanite phases decreased slightly, compared to the samples
without added TiO . Moreover, the content of wollastonite crystal phase containing 3
2
wt.% (mass fraction) TiO was slightly less than that of 2% TiO sample. This is
2
2
consistent with the experimental results. It is thus confirmed that TiO doping can
2
promote the precipitation of wollastonite crystal phase; however, excessive TiO2
doping introduces a new crystal phase and reduces the content of wollastonite crystal
phase in sample.
Fig. 7. Thermodynamic calculation of CBS glass-ceramic system
Fig. 8 presents SEM images of samples 5–8. Sample 5 corresponding to
formulation I is densely formed by a large number of acicular wollastonite crystal
grains, and the crystal grains are full and closely arranged, and the average particle
diameter is about 5
crystal grains, and the crystal grains are evenly and completely filled, and the average
particle diameter is in the range of 3–5 m. Sample 7 of formulation III consists of
closely arranged crystals, the grains are well developed, the structure is dense, and the
ꢀm. Sample 6 of formulation II is made up of a large number of
ꢀ

ACCEPTED MANUSCRIPT
average particle size is in the range of 3–5
ꢀ
m, showing a good microscopic
arrangement. SEM image of sample 8 of formulation IV exhibits that the crystal
grains are loosely arranged, the particle sizes are different, and the degree of
densification is inferior to that of samples 5–7. This is attributed to the fact that the
higher TiO content produces more crystal nuclei, which interfere with the crystal
2
growth, affect the dense arrangement of the wollastonite crystal phase, and produce a
small amount of Ti-containing crystal phase. This further affects the compactness of
the microstructure. This is consistent with the XRD detection results, which indicates
that certain amount of TiO doping promotes the crystallization and enhances the
2
densification. However, excessive TiO leads to the precipitation of the Ti-containing
2
crystal phase and affects the microstructure of sample.
Fig. 8. SEM images of samples 5–8 (See Table 4 for sintering regimes).
The results of the dielectric properties of samples 5–8 and the density of the
samples are presented in Table 5 and Fig. 9. The dielectric constant of samples 5–7
gradually decreases with the increase of TiO content, and maintains within a
2

ACCEPTED MANUSCRIPT
reasonable range, which indicates that the increase of TiO content can enhance the
2
dielectric properties of samples. However, the performance of sample 8 was slightly
lower than that of sample 7, indicating that excessive TiO would impair the dielectric
2
properties of sample. Excellent dielectric property of sample 7 is attributed to the
large content of wollastonite crystal phase, and uniform grain development, dense
structure, thereby reducing the number of polarizable atoms per unit volume, and
greatly reducing the structural loss of the samples. Moreover, this makes the samples
obtain lower dielectric constant and dielectric loss, which is consistent with the results
of XRD and SEM.
Comparative analysis indicates that sample 7 of formulation III is the optimal
sample and meets the market requirements. Its dielectric constant is 1.4 (1 MHz), 1.3
−3
−3
(
10 MHz), and the dielectric loss is 3.9 ×ꢀ10 (1 MHz), 3.8 ×ꢀ10 (10 MHz).
Table 5. Dielectric constant, dielectric loss and density of samples 5–8.
Density
Frequency
(MHz)
1
Dielectric constant
Dielectric loss
Sample no.
3
−3
(
g/cm )
(
εr
6
6
)
(10 )
2.3
3.4
6.8
7.9
3.9
3.8
4.8
5.4
5
6
7
8
2.37
1
0
1
2.1
2.2
1.4
1.3
2.3
2.4
2.58
2.70
2.46
1
1
0
1
0
1
1
0

ACCEPTED MANUSCRIPT
Fig. 9. Dielectric constant and dielectric loss of samples 5–8.

ACCEPTED MANUSCRIPT
4
. Conclusions
In this study, CBS glass-ceramics with four formulations were obtained to
investigate the effect of TiO on crystal phase, microstructure, and dielectric
2
properties. In summary, following conclusions can be drawn:
ꢀ
1
ꢁ
TiO doping reduced the transition temperature and crystallization
2
temperature of CBS glass-ceramics and promoted the precipitation of
wollastonite crystal phase. With the increase in the mass fraction of TiO2
from 0 to 3 wt.%, the transition temperature and crystallization
temperature of CBS glass-ceramics decreased by 12.9 and 27.6 °C,
respectively. However, excessive TiO caused the CBS glass-ceramics to
2
precipitate a hetero-crystalline phase, which led to the destruction of
close arrangement of the crystal grains in the microstructure. Furthermore,
two-step sintering promoted crystallization, enhanced densification, and
reduced dielectric constant and dielectric loss. The two-step sintering was
carried out at a temperature higher than the crystallization temperature of
2
0 °C, and the performance of the sample was found to be superior.
ꢀ2ꢁ Herein, sample 7, with the composition of CaO: SiO : B O = 1:1:0.6 and
2
2
3
doping mass fraction of 2 wt.% of TiO ., exhibited the best performance,
2
which meets the requirements for the target materials. The sintering
−1
system was heated to 700 °C at a rate of 3 °C min , held for 1 h, and
heated to 882 °C at the same rate. After 2 h of heat preservation, samples
were cooled down in a muffle furnace. After two-step sintering process,
the sample 7 acquired dielectric constant of 1.4 (1 MHz), 1.3 (10 MHz),
−3
−3
and a dielectric loss of 3.9 ×ꢀ10 (1 MHz), 3.8 ×ꢀ10 (10 MHz).
Acknowledgements
This study was funded by the National Natural Science Foundation of China
(Grant No. 51534001).

ACCEPTED MANUSCRIPT
References
[
[
[
[
[
[
[
[
1] O. Dernovsek, M. Eberstein, W.A. Schiller, et al. LTCC glass-ceramic
composites for microwave application[J]. Journal of the European Ceramic
Society, 2001, 21(10):1693-1697.
2] H. Wang, Z. He, D. Li, et al. Low temperature sintering and microwave
dielectric properties of CaSiO –Al O ceramics for LTCC applications[J].
3
2
3
Ceramics International, 2014, 40(3):3895-3902.
3] X. Chen, W. Zhang, S. Bai, et al. Densification and characterization of
SiO -B O -CaO-MgO glass/Al O composites for LTCC application[J].
2
2
3
2
3
Ceramics International, 2013, 39(6): 6355-6361.
4] J. Rodel, A. B. N. Kounga, M. Weissenberger-Eibl, et al. Development of a
roadmap for advanced ceramics: 2010-2025 [J]. Journal of the European
Ceramic Society, 2009, 29(9):1549-1560.
5] S. Arcaro, F.R. Cesconeto, F. Raupp-Pereira, et al. Synthesis and
characterization of LZS/
α-Al
2
3
O glass ceramic composites for applications in the
LTCC technology[J]. Ceramics International, 2014, 40(4):5269-5274.
6] L. J. Golonka. Technology and application of Low Temperature Cofired
Ceramic
ꢀLTCCꢁbased sensors and Microsystems[J]. Bulletin of the Polish
Academy of Sciences-Technical Sciences, 2006, 54(2):221-231.
7] L. Yuan, B. Liu, N. Shen, et al. Synthesis and properties of borosilicate/AlN
composite for low temperature co-fired ceramics application[J]. Journal of
Alloys and Compounds, 2014, 593:34-40.
8] N. Mori, Y. Sugimoto, J. Harada, et al. Dielectric properties of new
glass-ceramics for LTCC applied to microwave or millimeter-wave
frequencies[J]. Journal of the European Ceramic Society, 2006,
2
6(10-11):1925-1928.
[
[
9] H. Dannheim, A. Roosen. Effect of metallization on the lifetime prediction of
mechanically stressed low-temperature co-fired ceramic multilayers[J]. Journal
of the American Ceramic Society, 2005, 88(8):2188-2194.
10] A. Karamanov, M. Pelino. Induced crystallization porosity and properties of
sintereds diopside and wollastonite glass-ceramics[J]. Journal of the European
Ceramic Society, 2008, 28(3):555-562.

ACCEPTED MANUSCRIPT
[
[
11] K. Singh, I. Bala, V. Kumar. Structural, optical and bioactive properties of
calcium borosilicate glasses[J]. Ceramics International, 2009, 35(8):3401-3406.
12] M. He, M.Q. Wu, Z.R. Zhang, et al. Quantitative analysis of crystalline and
remaining glass phases in CaO–B O –SiO ternary system glass ceramics[J].
2
3
2
Journal of Alloys and Compounds, 2010, 506(2):757-760.
[
[
[
[
[
[
13] G.H. Chen, L.J. Tang, et, al. Synthesis and characterization of CBS
glass/ceramic composites for LTCC application[J]. Journal of Alloys and
Compounds, 2008, 478(1-2):858-862.
14] C.S. Cheng, C.C. Chou, C.S. Chen, et al. Microwave dielectric properties of
glass-MCT low temperature co-firable ceramics[J]. Journal of the European
Ceramic Society, 2004, 24(6):1795-1798.
15] H.K. Zhu, M. Liu. H. G. Zhou, et al. Preparation and properties of
low-temperature co-fired ceramic of CaO–SiO –B O system[J]. Journal of
2
2
3
Materials Science: Materials in Electronics, 2006 17(8):637-641.
16] J.H. Jean, C.R. Chang, C.D. Lei, Sintering of a Crystallizable CaO-B O -SiO
2
2
3
Glass with Silver[J]. Journal of the American Ceramic Society, 2004,
7(7):1244-1249.
8
17] S. Kemethmüller, M. Hagymasi, A. Stiegelschmitt, et al. Viscous Flow as the
Driving Force for the Densification of Low-Temperature Co-Fired Ceramics[J].
Journal of the American Ceramic Society, 2007, 90(1):64-70.
18] D.U. Tulyaganov, M.J. Ribeiro, J.A. Labrincha. Development of glass-ceramics
by
calcium-magnesium-aluminosilicate glass[J]. Ceramics International, 2002,
8(5):515-520.
sintering
and
crystallization
of
fine
powders
of
2
[
[
[
[
19] X.H. Zhou, B. Li, S.R. Zhang, et al. Effect of Ca/Si ratio on the microstructures
and properties of CaO–B O –SiO glass-ceramics[J]. Journal of Materials
2
3
2
Science: Materials in Electronics, 2009 20(3):262-266.
20] H. Zhu, M. Liu, H. Zhou, et al. Preparation and properties of low-temperature
co-fired ceramic of CaO–SiO –B O system[J]. Journal of Materials Science
2
2
3
Materials in Electronics, 2006, 17(8):637-641.
21] C.C. Chiang, S.F. Wang, Y.R. Wang, et al. Densification and microwave
dielectric properties of CaO-B O -SiO system glass–ceramics[J]. Ceramics
2
3
2
International, 2008, 34(3):599-604.
22] H. Shao, T. Wang, Q. Zhang. Preparation and properties of CaO–SiO –B O
3
2
2

ACCEPTED MANUSCRIPT
glass-ceramic at low temperature[J]. Journal of Alloys Compounds, 2009,
4
84(1):2-5.
23] C.R. Chang, J.H. Jean. Crystallization kinetics and mechanism of low
low temperature, cofirable CaO-B O -SiO glass-ceramics[J]. Journal of the
American Ceramic Society, 1999, 82(7):1725-1732.
[
[
[
[
-dielectric,
-
2
3
2
24] D.F. He. C Gao. Effect of boron on crystallization, microstructure and dielectric
properties of CBS glass-ceramics[J]. Ceramics International, 2018, 44(14):
1
6246-16255.
25] H.K. Zhu, H. Zhou, M. Liu, et a1. Low temperature sintering and properties of
CaO-B O -SiO system glass ceramic for LTCC applications[J]. Journal of
2
3
2
Alloys and Compounds, 2009, 482(1-2): 272-275.
26] H.I. Hsiang, S.W. Yung, C.C. Wang. Effects of the addition of alumina on the
crystallization, densification and dielectric properties of CaO–MgO–Al O –
2
3
SiO , glass in the presence of ZrO [J]. Ceramics International, 2014,
2
2
4
0(10):15807-15813.
[
[
[
[
[
[
[
27] X.H. Zhou, E.Z. Li, S.L. Yang, et al. Effects of La O –B O on the flexural
2 3 2 3
strength and microwave dielectric properties of low temperature co-fired CaO–
B O –SiO glass–ceramic[J]. Ceramics International, 2012, 38(7):5551-5555.
2
3
2
28] S.F. Wang, C.C. Chiang, Y.R. Wang, et al. Effects of additives on the
densification and microwave dielectric properties of binary CaO-B O -SiO
2
2
3
glass[J]. Japanese Journal of Applied Physics, 2010, 49(49):1-5.
29] G.A. Khater. Influence of Cr O , LiF, CaF and TiO nucleants on the
2
3
2
2
crystallization behavior and microstructure of glass-ceramics based on
blast-furnace slag[J]. ceramics International, 2011, 37(7):2193-2199.
30] J. Yang, S.G. Zhang, B. Liu, et al. Effect of TiO on Crystallization,
2
Microstructure and Mechanical Properties of Glass-ceramics[J]. Journal of Iron
and Steel Research, International, 2015, 22(12):1113-1117.
31] M. Rezvani, B. Eftekhari-Yekt, M. Solati-Hashjin, et al. Effect of Cr O , Fe O
3
2
3
2
and TiO nucleants on the crystallization behavior of SiO -Al O -CaO-MgO
2
2
2
3
ꢀ
R Oꢁglass-ceramics[J]. Ceramics International, 2005, 31(1):75-80.
2
32] D.P. Mukherjee, S.K. Das, Influence of TiO content on the crystallization and
2
microstructure of machinable glass-ceramics[J]. Journal of Asian Ceramic
Societies, 2015, 4(1):55-60.
33] T. Ebadzadeh, E. Ghasemi. Effect of TiO addition on the stability of t-ZrO in
2
2

ACCEPTED MANUSCRIPT
mullite-ZrO composites prepared from various starting materials[J]. Ceramics
2
International, 2002, 28(4):447-450.
[
34] A. Dugué, O. Dymshits, et al. Structural transformations and spectroscopic
properties of Ni-doped magnesium aluminosilicate glass-ceramics nucleated by
a mixture of TiO and ZrO for broadband near-IR light emission[J]. Journal of
2
2
Alloys and Compounds, 2019, 780:137-146.
[
[
[
[
[
[
35] H.K. Zhu, H.Q. Zhou, M. Liu, et al. Microstructure and microwave dielectric
characteristics of CaO–B O –SiO glass ceramics[J]. Journal of Materials
2
3
2
Science Materials in Electronics, 2009, 20(11):1135-1139.
36] C.C. Chiang, S.F. Wang, Y.R. Wang, et al. Characterizations of CaO–B O –SiO
2
2
3
glass–ceramics: Thermal and electrical properties[J]. Journal of Alloys and
Compounds, 2008, 461(1-2):612-616.
37] S.F. Wang, Y.R. Wang, Y.F. Hsu, et al. Densification and microwave dielectric
behaviors of CaO–B O –SiO glass-ceramics prepared from a binary glass
2
3
2
composite[J]. Journal of Alloys and Compounds, 2010, 498(2):211-216.
38] P.Y. Li, S.B. Wang, J.G. Liu, et al. Effects of B O content and sintering
2
3
temperature on crystallization and microstructure of CBS glass–ceramic
coatings[J]. Applied Surface Science, 2015, 356:1180-1188.
39] L.C. Ren, X.F. Luo, L.S. Hu, et al. Synthesis and characterization of LTCC
compositions with middle permittivity based on CaO-B O -SiO glass/CaTiO
3
2
3
2
system[J]. Journal of the European Ceramic Society, 2016, 37(2):619-623.
40] H.Y. Yang, E.Z. Li, C.L. Sun, et al. The Influence of Sintering Temperature on
the Microwave Dielectric Properties of Mg SiO Ceramics with CaO-B O -SiO
2
2
4
2
3
Addition[J]. Journal of Electronic Materials, 2017, 46(2):1048-1054.

ACCEPTED MANUSCRIPT
reduced the sintering temperature of CBS glass-ceramics.
promoted the precipitation of wollastonite crystal phase.
TiO
2
2
TiO
2
Excessive TiO destroyed crystal structure, and damaged close arrangement of crystal
grains.