The Enhanced Thermal Stability of (Mg0.95Ni0.05)2TiO4 Dielectric Ceramics Modified by a Multi-Phase Method

The thermal stability of (Mg0.95Ni0.05)2TiO4 dielectric ceramics has been improved by mixing with CaTiO3 phases owing to higher positive temperature coefficients. The pure (Mg0.95Ni0.05)2TiO4 and the mixture phase systems of CaTiO3-modified (Mg0.95Ni0.05)2TiO4 were verified by XRD diffraction patterns to ensure the crystallite of different phases. The microstructures of the CaTiO3-modified (Mg0.95Ni0.05)2TiO4 were observed by SEM and EDS to investigate the relation between element ratios and grains. As a result, it can be seen that the thermal stability of the CaTiO3-modified (Mg0.95Ni0.05)2TiO4 can be effectively enhanced, compared with the pure (Mg0.95Ni0.05)2TiO4. Moreover, the radio frequency dielectric performances of CaTiO3-modified (Mg0.95Ni0.05)2TiO4 dielectric ceramics are strongly dependent upon the density and the morphology of the specimens. The champion sample with the ratio of (Mg0.95Ni0.05)2TiO4 and CaTiO3 of 0.92:0.08 showed an εr value of 19.2, an Qf value of 108,200 GHz, and a τf value of −4.8 ppm/°C, which may encourage (Mg0.95Ni0.05)2TiO4 ceramics to broaden the range of novel applications and match the requirements of 5G or next-generation communication systems.


Introduction
As a result of the continuous evolution of science and technology, millimeter-wave technology has been extensively developed and considered as a branch of 5G communication or next-generation mobile communication technology. Due to a large amount of data transmission, a multilayer board is required as the main circuit board design, leading to materials which should possess both low dielectric loss and conductor loss characteristics [1][2][3][4][5]. Dielectric materials, to date, are a good choice for creating high-frequency components that have a higher dielectric constant (ε r ), which can contribute to the miniaturization of components; high-quality factors (Qf values) can improve the energy of stored electromagnetic waves and the temperature coefficient of the resonance frequency (τ f ) approaching zero, which can enhance the thermal stability of the components. For example, when a filter possesses the three abovementioned characteristics, the filter results in effective downsizing, lower dielectric loss rates, greater filtering activity, and greater stability, unaffected by the external ambient temperature [6,7].
The ceramic systems based on MgTiO 3 have always received considerable attention and applications in the literature. For example, dielectric passive components of MgTiO 3 in a communication system have been employed to be a resonator, duplex, filter, and antenna [8][9][10][11][12]. Therefore, it is considerably important to enhance the dielectric performance of MgTiO 3 . When the ratio of Mg and Ti is 2:1, the binary titanate ceramic Mg 2 TiO 4 possesses a spinel-type structure that belongs to the cubic phase with Fd-3m space group (2 2 7). It demonstrates an ε r value of 14, a Qf value of 150,000 GHz, and a τ f value of −50 ppm/ • C [13][14][15]. In addition, to further improve the dielectric characteristics, the method of partial substitution that selects Ni 2+ (0.069 nm) with a radius resembling Mg 2+ (0.078 nm) to perform the substitution was presented [16]. When the Mg 2+ ions were replaced by Ni 2+ ions to form (Mg, Ni) 2 TiO 4 compositions, the (Mg 0.95 Ni 0.05 ) 2 TiO 4 compositions possessing a spinel-type structure presented a good dielectric performance with an ε r value of 16.43, a Qf value of 238,000 GHz, and a τ f value of −55 ppm/ • C [17]. Moreover, an inferior (Mg 0.95 Ni 0.05 )TiO 3 phase also appeared during the synthesizing of (Mg 0.95 Ni 0.05 ) 2 TiO 4 , which may be attributed to the effect of the thermal decomposition mechanism [18]. However, the slight (Mg 0.95 Ni 0.05 )TiO 3 -doped (Mg 0.95 Ni 0.05 ) 2 TiO 4 possess a comparable performance, with an ε r value of 17.2, a Qf value of 180,000 GHz, and a τ f value of −45 ppm/ • C [19], compared to those of the pure (Mg 0.95 Ni 0.05 ) 2 TiO 4 composition [17].
On the other hand, a multi-phase method has been employed to modify the dielectric characteristics. In Mg 2 TiO 4 -SrTiO 3 systems [20], Mg 2 TiO 4 and SrTiO 3 have spinal cubic (lattice parameters: a = 8.439 Å, space group Fd3m) (ICDD-PDF#00-003-0858) and cubic perovskite (ICDD-PDF#01-084-0443), respectively. When x increased, the peak intensity of SrTiO 3 increased and the lattice parameters of Mg 2 TiO 4 remained unchanged, which demonstrates that the two-phase system was used to modify relative permittivity in the Mg 2 TiO 4 -SrTiO 3 system. In (Mg 0.95 Zn 0.05 ) 2 TiO 4 -SrTiO 3 systems [21], the lattice parameters had slight influence and remained unchanged after the SrTiO 3 was added into (Mg 0.95 Zn 0.05 ) 2 TiO 4 , which confirms that the presence of a two-phase system could effectively promote densification in the (Mg 0.95 Zn 0.05 ) 2 TiO 4 matrix. In addition, CaTiO 3 -and SrTiO 3 -modified (Mg 0.95 Co 0.05 ) 2 TiO 4 ceramics [22,23] were presented using a multi-phase method for low-loss dielectric properties at microwave frequencies. Furthermore, thermal stability of dielectric ceramics is another important factor in practical applications. However, few studies have been conducted to discuss the thermal stability of (Mg 0.95 Ni 0.05 ) 2 TiO 4 . For this purpose, we have made great efforts to enhance the thermal stability factor of (Mg 0.95 Ni 0.05 ) 2 TiO 4 compositions using a perovskite structured SrTiO 3 additive to form a two-phase system in our previous report [24].
In this study, we demonstrate the microstructure and radio frequency dielectric characteristics of the CaTiO 3 -modified (Mg 0.95 Ni 0.05 ) 2 TiO 4 ceramic system for enhancing its thermal stability factor using the multi-phase method. We attempted to add CaTiO 3 with a high positive temperature coefficient to form a three-phase system and to compensate the negative temperature coefficient of (Mg 0.95 Ni 0.05 ) 2 TiO 4 . X-ray diffraction (XRD), scanning electron microscopy (SEM), and energy-dispersive X-ray spectrometer (EDS) analyses were also employed to study the microstructure grain boundary, and compositions of the ceramic system. The radio frequency dielectric performances of the CaTiO 3 -modified (Mg 0.95 Ni 0.05 ) 2 TiO 4 were quantified and examined by employing the formation of mutiphase coexistence.

Materials and Method
In this study, (Mg 0.95 Ni 0.05 ) 2 TiO 4 and CaTiO 3 were produced by a mixed-oxide solidstate reaction using the following high-purity chemical powders: magnesium oxide (MgO), nickel oxide (NiO), calcium carbonate (CaCO 3 ), and titanium dioxide (TiO 2 ). Because MgO is hygroscopic in nature, it first removes the moisture remaining at 600 • C for 2 h. The mixture of the above oxides was mixed according to the stoichiometries of (Mg 0.95 Ni 0.05 ) 2 TiO 4 and CaTiO 3 . They were then ball ground in distilled water (ball grinding medium) for 24 h. All mixed powders were parched in a kiln and pre-phased (calcine) at 1100 • C for 4 h in a high-temperature furnace. Then, the pre-phased reagents were mixed again according to the chemical molar ratio of (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 , and ball ground into a fine powder for 24 h. Then, polyvinyl alcohol (PVA 500; Showa, Tokyo, Japan) as a binder was added to the calcined powder, uniformly granulated, screened with a 100-mesh screen, and compressed with a pressure of 200 MPa to form a tablet form with a height of 0.5 cm and diameter of 1.1 cm. The sintering temperatures of the tablets were set at 1300~1425 • C for 4 h in air. The risen and dropped temperature rates were set at 10 • C/min for all samples.
The crystallization-phase observation of the pre-phased powder and mixture compositions were analyzed by XRD (Rigaku D/Max III. V., Tx for USA) The lattice constant was calculated using the Rietveld method to fit the XRD patterns. SEM (Philips XL-40FEG, Eindhoven, The Netherlands) was employed to observe the surface morphologies of samples, and EDS was utilized to demonstrate the different phases and compositions. The apparent densities of the samples were measured using the Archimedes method. The ε r and Qf values at radio frequencies were measured by the Hakki-Coleman dielectric resonator method [25][26][27].
The measurement was mainly composed of a vector network analyzer (HP8757D, Agilent Technologies, Taipei, Taiwan) and sweep oscillator connections (HP8350B, Agilent Technologies, Taipei, Taiwan). The thermal stability (τ f values) was evaluated with a temperature range from 20 to 80 • C. The following Formula (1) was utilized to obtain the τ f value (ppm/ • C): where f 1 and f 2 represent the resonant frequencies at T 1 and T 2 , respectively.  Figure 1. In addition, the reason for the presence of the (Mg 0.95 Ni 0.05 )TiO 3 phase is generally considered to be the uneven particle size of the initial materials, which increases the probability of secondary crystallization nucleation. Another reason for the formation of the (Mg 0.95 Ni 0.05 )TiO 3 phase may be the effect of the thermal decomposition mechanism [18]. However, the ratio of (Mg 0.95 Ni 0.05 )TiO 3 phase in the ceramics was decreased due to the increased grain boundary motion with the increasing sintering temperature, as shown in Table 1. Moreover, the (Mg 0.95 Ni 0.05 )TiO 3 had no noticeable impact on the dielectric properties of (Mg 0.95 Ni 0.05 ) 2 TiO 4 in our previous study [19]. Therefore, the influence of the inferior (Mg 0.95 Ni 0.05 )TiO 3 phase can be negligible when the sintering temperature is over 1350 • C.  Table 2. It was also observed that all samples have a spinal cubic structure with the lattice constants (a = b = c) from 0.84005 to 0.83456 nm, indicating that the (Mg 0.95 Ni 0.05 ) 2 TiO 4 is still in the primary phase. When CaTiO 3 was blended with (Mg 0.95 Ni 0.05 ) 2 TiO 4 , no obvious influence on the lattice constants of (Mg 0.95 Ni 0.05 ) 2 TiO 4 could be found (Table 2). Furthermore, the growth of mixed phases in the one ceramic may cause a negative effect due to structural dissimilarities and the larger ionic radii values of Ca 2+ (0.106 nm) compared to those of Mg 2+ (0.078 nm) and Ni 2+ (radii = 0.069 nm) [16]. However, the XRD analysis confirms the coexistence of the multiple phases without the structural dissimilarities in our samples. Figure 3 shows the SEM images of the morphologies of the 0.92·(Mg 0.95 Ni 0.05 ) 2 TiO 4 ·0.08 CaTiO 3 sintered at different temperatures. When the sintering temperature was increased, the grain size increased in Figure 3a-f, which is consistent with the XRD results ( Figure 1). The pores reduced with the sintering temperature increasing from 1300 to 1350 • C. The reason for the increase in the grain size and the decrease in pores is the thermal drive energy, which enables connection and expands the neck between the grains. Therefore, Figure 3a-c exhibits the grain growth with better movement of the grain boundary and a more uniform and dense morphology. However, when the sintering temperature was increased from 1350 to 1425 • C, the grain size still increased, and the pore size became larger. This is attributed to the excessive extrusion between grains and air pressure under too-high sintering temperatures, which is similar with previous reports [20][21][22][23][24]. Furthermore, grains of 0.92·(Mg 0.95 Ni 0.05 ) 2 TiO 4 ·0.08CaTiO 3 can be roughly divided into three shapes, as shown in Figure 3c. The EDS results of each grain are summarized in Table 3. Therefore, the different grains were identified as follows: spot A is (Mg 0.95 Ni 0.05 )TiO 3 ; spot B is (Mg 0.95 Ni 0.05 ) 2 TiO 4 ' and spot C is CaTiO 3 . The EDS results are consistent with the XRD analysis, verifying that the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 is a three-phase coexistence system.       Figure 3 shows the SEM images of the morphologies of the 0.92·(Mg0.95Ni0.05)2TiO4·0.08CaTiO3 sintered at different temperatures. When the sintering temperature was increased, the grain size increased in Figure 3a-f, which is consistent with the XRD results ( Figure 1). The pores reduced with the sintering temperature increasing from 1300 to 1350 °C. The reason for the increase in the grain size and the decrease in pores is the thermal drive energy, which enables connection and expands the neck between the grains. Therefore, Figure 3a-c exhibits the grain growth with better movement of the grain boundary and a more uniform and dense morphology. However, when the sintering temperature was increased from 1350 to 1425 °C, the grain size still increased, and the pore size became larger. This is attributed to the excessive extrusion between grains and air pressure under too-high sintering temperatures, which is similar with previous reports [20][21][22][23][24]. Furthermore, grains of 0.92·(Mg0.95Ni0.05)2TiO4·0.08CaTiO3 can be roughly divided into three shapes, as shown in Figure 3c. The EDS results of each grain are summarized in Table 3. Therefore, the different grains were identified as follows: spot A is (Mg0.95Ni0.05)TiO3; spot B is (Mg0.95Ni0.05)2TiO4' and spot C is CaTiO3. The EDS results are consistent with the XRD analysis, verifying that the (1-x) (Mg0.95Ni0.05)2TiO4·xCaTiO3 is a three-phase coexistence system.      Figure 4 shows the results of the apparent density and dielectric constant (ε r values) of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 with different x values and temperatures. The apparent density was increased with the increasing x value because CaTiO 3 (~4.036 g/cm 3 ) possesses a higher density than that of (Mg 0.95 Ni 0.05 ) 2 TiO 4 (~3.49 g/cm 3 ) [17]. Moreover, the apparent density value increased when the temperature increased from 1300 to 1350 • C, which is due to the grain growth (Figure 3a-c). However, the apparent density was reduced due to the enlarged pore size (Figure 3d-f) when the temperature increased from 1350 to 1425 • C. These results lead to the optimal sintering temperature of 1350 • C being obtained. Due to much higher dielectric constant (ε r ) of (Mg 0.95 Ni 0.05 ) 2 TiO 4 (~16.4) and CaTiO 3 (~170) than that of air (~1), the ε r of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 xCaTiO 3 correlates positively with apparent density. Therefore, the ε r changed with the change in the apparent density and the x value. Therefore, the dielectric constant changes from 18.6 to 20.9 as x increases from 0.06 to 0.12, as shown in Table 4.      Figure 5 shows that the Qf value (quality factor) of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 xCaTiO 3 decreased with the increasing the x value. It is attributed to the fact that the Qf value of CaTiO 3 (~3600 GHz) was relatively smaller than that of (Mg 0.95 Ni 0.05 ) 2 TiO 4 (~238,000 GHz). In addition, the Qf of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 rises at 1300-1350 • C and drops at 1350-1425 • C in Figure 5. Compared with Figure 4, the Qf of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 xCaTiO 3 correlated positively with the ε r and apparent density. Although there are many factors that impact the Qf, such as dielectric loss (which is caused by the second phase), oxygen vacancy, grain size, and porosity [28], the Qf is mainly dependent on the apparent density in our case. Similar with the results of the apparent density at different temperatures (Figure 4), the Qf also reached the optimal value at 1350 • C, as listed in Table 4. The τ f of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 sintered at different temperatures seems to maintain a constant value, which is strongly dependent on the CaTiO 3 contents (x value), as shown in Figure 5 and Table 4. The composition and phase determined the thermal stability of the (1-x) (Mg 0.95 Ni 0.05 ) 2 TiO 4 ·xCaTiO 3 because a high positive temperature coefficient of CaTiO 3 can compensate the negative temperature coefficient of (Mg 0.95 Ni 0.05 ) 2 TiO 4 . Therefore, the τ f was changed from −20.7 to 30.5 (ppm/ • C) when x increased from 0.04 to 0.12, indicating that τ f = 0 is possible after controlling x. It is worth noting that the τ f was only −4.8 ppm/ • C at x = 0.08 but the Qf was kept at 108,200 GHz.

Conclusions
In conclusion, the thermal stability of the CaTiO3-modified (Mg0.95Ni0.05)2TiO4 ceramic system was investigated. The dense morphology without pores can be obtained at 1350 °C, resulting in the optimal sintering temperature of the apparent density, εr, and Qf. A high positive temperature coefficient of CaTiO3 can be used to improve the thermal stability of (Mg0.95Ni0.05)2TiO4. The τf can be adjusted with changing the CaTiO3 contents, even closing to zero. The excellent dielectric characteristics of the 0.92·(Mg0.95Ni0.05)2TiO4·0.08CaTiO3 sintered at 1350 °C was presented with an εr of 19.2, an Qf of 108,200 GHz, and a τf of −4.8 ppm/°C. Therefore, the CaTiO3-modified (Mg0.95Ni0.05)2TiO4 ceramic system showed high thermal stability and performance, suggesting the potential of these ceramics as dielectric substrate materials and radio frequency passive components in the microwave field to miniaturize components and transmit signals neglecting the temperature factor.

Conclusions
In conclusion, the thermal stability of the CaTiO 3 -modified (Mg 0.95 Ni 0.05 ) 2 TiO 4 ceramic system was investigated. The dense morphology without pores can be obtained at 1350 • C, resulting in the optimal sintering temperature of the apparent density, ε r , and Qf. A high positive temperature coefficient of CaTiO 3 can be used to improve the thermal stability of (Mg 0.95 Ni 0.05 ) 2 TiO 4 . The τ f can be adjusted with changing the CaTiO 3 contents, even closing to zero. The excellent dielectric characteristics of the 0.92·(Mg 0.95 Ni 0.05 ) 2 TiO 4 ·0.08CaTiO 3 sintered at 1350 • C was presented with an ε r of 19.2, an Qf of 108,200 GHz, and a τ f of −4.8 ppm/ • C. Therefore, the CaTiO 3 -modified (Mg 0.95 Ni 0.05 ) 2 TiO 4 ceramic system showed high thermal stability and performance, suggesting the potential of these ceramics as dielectric substrate materials and radio frequency passive components in the microwave field to miniaturize components and transmit signals neglecting the temperature factor.