Microwave dielectric properties of Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) for LTCC applications

Two low-firing Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) dielectrics were prepared, and their microwave dielectric properties were investigated. The XRD patterns indicated both samples crystallized in an orthorhombic structure with different space groups. The microstructures, lattice parameters, and Raman spectra of these ceramics were also systematically studied. The microwave dielectric properties were obtained with ε r ~ 9.1, Q × f ~ 34 000 GHz and τ f ~ −72 ppm/°C for Li 2 Co 2 (MoO 4 ) 3 at 840°C, ε r ~ 9.6, Q × f ~ 28 000 GHz and τ f ~ −71 ppm/°C for Li 2 Ni 2 (MoO 4 ) 3 at 660°C. With 0.05 mol% Zn substitution, excellent microwave dielectric properties ( ε r ~ 10.9, Q × f ~ 56 000 GHz and τ f ~ −62 ppm/°C) can be


| INTRODUCTION
Due to the requirement of miniaturization and integration of microwave passive components and devices, low-temperature co-fired ceramics (LTCC) technology has become more important and been widely used in the fabrication of multilayer devices for today's communication systems in the last two decades. LTCC multi-layer devices are normally composed of layers of alternating microwave dielectric ceramics and internal Ag electrodes. To avoid the migration of Ag, sintering temperatures of the devices must be lower than the melting point of Ag (melting point ~ 960°C). 1,2,3,4 In addition, other major requirements for LTCC materials include a low dielectric constant (ε r < 10), a high Q × f value, and a near-zero resonant frequency (τ f ) temperature coefficient. These requirements are also valid for ultra-low-temperature co-fired ceramics (ULTCC) applications (<700°C). 5 Thus, the research to enable co-firing either using additives or developing new materials has been widely conducted. However, low-melting glass additions need to go through a prior glass preparation followed by a synthesizing condition of sintering and normally it also degrades the microwave dielectric properties of the ceramics, in particularly the Q × f value, which would limit its applications at high-frequency regime such as 5G system. As a result, the demand for new materials with excellent microwave dielectric properties and low sintering temperature is desired than ever.
In this paper, the crystal structure and the microwave dielectric properties of Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics were firstly investigated. The XRD patterns, surface morphology, and Raman spectrum were also reported. In addition, small amount Zn was employed to replace Ni to lower the microwave dielectric loss of the specimens.

| EXPERIMENTAL PROCEDURE
The Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics were prepared by the solid-state reaction method. The starting materials of high-purity Li 2 CO 3 , CoO, NiO, ZnO, and MoO 3 oxide powders (>99.9%) were weighed and ball-milled using distilled water and zirconia balls in anion container for 24 hours. The mixtures were dried and calcined at 500°C for 2 hours, mixing with 5 wt% of a 10% PVA solution as a binder, and granulated by sieving through a 200 mesh and pressed into pellets with 5 mm in thickness and 11 mm in diameter using an automatic uniaxial hydraulic press at 2000 kg f / cm 2 . These pellets were sintered at 600-900°C for 4 hours. The finished sample was performed by the following test. Relative density is the ratio of the density of the sample to the density of the Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni). The density of the Li 2 Co 2 (MoO 4 ) 3 and Li 2 Ni 2 (MoO 4 ) 3 is 4.3418 and 4.3797 g/cm 3 , respectively. 11,12 The crystal structures of the sintered ceramics were analyzed with a Siemens D5000 diffractometer with Cu Kα radiation operated at 40 kV and 40 mA. Raman spectra were excited with a 532 nm Raman Spectrometer light source (Jobin Yvon/Labram HR). The microstructures were evaluated using scanning electron microscopy (UHR-SEM; HitachiSU-8000), and the composition of the samples was analyzed with an energy-dispersive X-ray spectrometer (EDS, Philips). The bulk densities of the ceramics were measured using the Archimedes method. The dielectric constant (ε r ) and the quality factor values (Q) at microwave frequencies were measured using the dielectric resonator method suggested by Hakki-Coleman 13 and Courtney. 14 A system combining a HP8757D network analyzer and a HP8350B sweep oscillator was employed in the measurement. An identical technique was applied to the measurement of the resonant frequency (T f ) temperature coefficient. T f (ppm/°C) can be calculated by considering the change in resonant frequency (Δf), for which the resonant frequency (T f ) temperature coefficient can be defined as.
where f 1 and f 2 represent the resonant frequencies at T 1 and T 2 , respectively. The entire test setup was then placed over a thermostat in a temperature range of 25 to 80°C. Rietveld refinement of the powder X-ray diffraction data was performed with the General Structure and Analysis (GSAS) System. The patterns were refined for the lattice parameters, scale factor, background, bond length, bond angle, and atomic coordinates.  3 , respectively. It is observed that there are several XRD peaks slightly shifting toward higher angles. This distortion might be induced from a compressed lattice volume owing to materials sintering process. The specimens exhibited homogeneous orthorhombic phase but belonged to different space groups which are Pnma(62) for Li 2 Co 2 (MoO 4 ) 3 and Pmcn(62) for Li 2 Ni 2 (MoO 4 ) 3 . To further

| RESULTS AND DISCUSSION
The XRD patterns of Li 2 Co 2 (MoO 4 ) 3 ceramics sintered at different temperatures for 4 h clarify the crystal structure of Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics, and demonstrate the lattice size decreased, Rietveld refinements were carried out using GSAS software and the refined lattice parameters, molecular volume, reliability factors, and goodness-of-fit indicators are illustrated in Tables 1 and 2. The variations of lattice parameters and cell volumes were limited and within 0.5%, and the goodness-of-fit indicator values (χ 2 ) were almost in the range 0.83-1.79, indicating that the structural model is valid and the refinement result is reliable. Also, the refined plots of the specimens are shown in Figure 2 and the results are identical to the crystallographic data achieved from Rietveld refinement, as shown in Tables 1 and 2. These results also confirmed the formation of a single orthorhombic structure phase. Figures 3 and 4 demonstrate SEM images of Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) specimens sintered at different temperatures for 4h. As can be seen that the Li 2 Co 2 (MoO 4 ) 3 specimen at 840°C exhibits typical uniform polygonal grains along with some small grains and both are corresponding to a single Li 2 Co 2 (MoO 4 ) 3 phase as further confirmed by the EDS results ( Figure 3f). As the temperature is up to 870°C or higher, the grain starts to melt, and a little secondary phase precipitates at the grain boundary (shown in Figure 3d,e), which is bad for grain densification (which will be shown in Figure 7). For Li 2 Ni 2 (MoO 4 ) 3 specimens sintered from 600°C to 720°C, well-densified microstructures were developed and there was no second phase for the entire temperature range; also, both round and bar grains have same composition which agrees with the EDS results ( Figure 4f). Specimen at 720°C was found to have an obviously larger grain size and a less uniform morphology because of a higher sintering temperature. The influence of grain growth on the relative density is not obvious because of the little grain size change (which will be discussed in Figure 7). As a result, these materials have individual optimum sintering temperature for densification. Figures 5 and 6 illustrate the Raman spectrum of the Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics recorded at room temperature. The Raman bands of Co-contained specimen are apparently distributed in three distinct regions: between 316 and 360 cm −1 corresponding to the MoO 4 bending vibration modes, between 801 and 886 cm −1 representing the MoO 4 anti-symmetric stretching modes, and between 948 cm −1 and 970 cm −1 in response of the MoO 4 symmetric stretching modes 15 . In comparison with that of Co-contained specimen, small pattern shift was observed for Li 2 Ni 2 (MoO 4 ) 3 ceramic due to the radius difference between Co and Ni. 16 The relative densities and ε r values of the orthorhombic Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics as a function of sintering temperature are shown in Figure 7. The relative density was very low at approximately 86% of the theoretical F I G U R E 2 The XRD patterns of density for the Li 2 Co 2 (MoO 4 ) 3 specimen sintered at 780°C most likely due to the porous microstructure as shown in Figure 3. However, it increased with increases in the sintering temperature, reaching a maximum of 93.3% at 840°C, and decreased thereafter due to the secondary phase precipitation. Variation of ε r was consistent with that of the relative  density indicating that density might be a primary factor affecting ε r in this experiment. In addition, the relative density of Li 2 Ni 2 (MoO 4 ) 3 specimen also shows a maximum of 91% at 660°C and slightly decreases at temperatures higher than 660°C corresponding to a less uniform morphology, suggesting that 660°C is the optimal sintering temperature, and a maximum ε r value of ~9.6 was obtained.
The Q × f and T f values of Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics at different sintering temperatures are illustrated in Figure 8. The measured microwave dielectric loss represents the overall loss, including not only intrinsic loss related to the lattice vibration modes, but also extrinsic contributions related to density, second phases, impurities, surface morphology, and the lattice defect. 17,18 The variations of Q × f values for both specimens were consistent with their relative densities suggesting that the dielectric loss was mainly controlled by the densification of the specimens. Maximum Q × f of 34 000 GHz and 28 000 GHz were obtained for Co-contained and Ni-contained specimens, respectively. Moreover, crystal structure may also play an important role in affecting the dielectric properties. Because there is no obvious second-phase precipitation in the pure orthorhombic Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) ceramics, accordingly, the T f values of the specimens showed no significant change and ranged from −72ppm/°C to −94ppm/°C and −61ppm/°C to −71ppm/°C for Co-contained and Ni-contained specimens, respectively, since the variations in the lattice parameters and cell volumes were limited and within 0.5%. Small amount of Zn was employed to partially replace Ni in the Li 2 Ni 2-x Zn x (MoO 4 ) 3 (x = 0-0.1) ceramics to improve its microwave dielectric properties, and the XRD patterns of the specimens sintered at 660°C for 4 hours are shown in Figure 9. All the specimens can be identified as a single Li 2 Ni 2 (MoO 4 ) 3 phase with the orthorhombic structure. However, expecting peak shift due to the ionic radius difference was not observed, which might be because of the substitution amount was small and could not be identified from the XRD patterns. The cell volumes, lattice parameters, and reliability factors of specimens after Rietveld refinements are listed in Table 3, and the variations of lattice parameters and cell volumes were all less F I G U R E 9 The XRD patterns of   than 0.3%. In addition, the goodness-of-fit indicator values (χ 2 ) were in the range 1.01-1.38, suggesting that the structural model is valid and the refinement result is reliable. Figure 10 shows the relative densities and ε r values of the Li 2 Ni 2-x Zn x (MoO 4 ) 3 (x = 0-0.1) ceramics at different sintering temperatures. For all specimens, with increasing sintering temperature the relative density increased to a maximum value at 660-690°C and then slightly decreased and it showed a relative density of 93% for x = 0.05 at 660°C. The ε r value also exhibited a similar variation trend with that of the relative density and a ε r of 10.9 can be obtained for x = 0.05 at 660°C. Instead of intrinsic factor, it suggests that the ε r value was mainly controlled by the extrinsic factor, which is relative density in this case.
The Q × f values and packing fractions of the Li 2 Ni 2-x Zn x-(MoO 4 ) 3 (x = 0-0.1) ceramics are demonstrated in Figure 11. The packing fraction, defined by summing the volume of packed ions over the volume of a primitive unit cell, can be expressed as: 7,19 where Z is the number of formula units per unit cell. Accordingly, the dependence of the Q × f on the packing fraction of the ceramics is weak. The maximum Q × f of 56 000 GHz was obtained at 660°C for x = 0.05. In comparison with that of x = 0, it shows a 100% enhancement in the x Zn x (MoO 4 ) 3 (x = 0-0.1) ceramics Q × f value which would make it a more promising dielectric material for high-frequency applications such as 5G system. The T f value of the specimen as shown in Figure 12 was functions of Zn contain and sintering temperature and it ranged from -54ppm/°C to -84ppm/°C.
The SEM images and related EDS results of Li 2 Ni 2-x Zn x-(MoO 4 ) 3 (x = 0.05) ceramics sintered at 660°C are illustrated in Figure 13. The morphologies of the specimens are similar, and no significant difference was observed. Corresponding to EDS results, large and small grains were confirmed as the same Li 2 Ni 2-x Zn x (MoO 4 ) 3 (x = 0.05) composition ( Figure 13b).

| CONCLUSION
For Li 2 M 2 (MoO 4 ) 3 (M = Co, Ni) specimens sintered at 600-900°C for 4 hours, a single orthorhombic structure with different space groups belonging to Pnma(62) and Pmcn(62), respectively, was detected. Contrast to Li 2 Co 2 (MoO 4 ) 3 having a higher Q × f value, a much lower sintering temperature (660°C) was required to obtain single Li 2 Ni 2 (MoO 4 ) 3 phase, which is more attractive for the ULTCC applications. The dielectric properties are correlated with the temperatures, the densities, and the microstructures of the specimens. Moreover, a small Zn substitution in the Li 2 Ni 2-x Zn x (MoO 4 ) 3 (x = 0.05) composition could effectively increase its Q × f (28 000 GHz for pure one) by 100% to a value of 56 000 GHz that would make it a more promising dielectric material for high-frequency applications such as 5G system.