Cold sintered LiMgPO 4 based composites for low temperature co-fired ceramic (LTCC) applications

Cold sintered, Li 2 MoO 4 -based ceramics have recently been touted as candidates for electronic packaging and low temperature co-fired ceramic (LTCC) technology but MoO 3 is an expensive and endangered raw material, not suited for large scale commercialization. Here, we present cold sintered temperature-stable composites based on LiMgPO 4 (LMP) in which the Mo (and Li) concentration has been reduced, thereby significantly decreasing raw material costs. Optimum compositions, 0.5LMP-0.1CaTiO 3 -0.4K 2 MoO 4 (LMP-CTO-KMO), achieved 97% density at <300°C and 600 MPa for 60 minutes. Raman spectroscopy, X-ray diffraction, scanning electron microscopy, and energy dispersive X-ray mapping confirmed the coexistence of end-members, LMP, CTO, and KMO, with no interdiffusion and parasitic phases. Composites exhibited temperature coefficient of resonant frequency ~ –6 ppm/°C, relative permittivity ~9.1, and Q × f values ~8500 GHz, properties suitable for LTCC technology and competitive with commercial incumbents.

Traditional MW ceramics are sintered at high temperature (T > 1000°C) to impart strength, integrity, and to optimize the required physical properties, 6−9 but energy consumption and associated carbon emissions are substantial and high T limits the integration of low cost metal electrodes (Ag) and polymers.−43 The energy consumed by cold sintering is <50% of that used by its conventional counterpart, 6 facilitating integration with polymers and base metals and liberating RF design space for ceramics, previously forbidden by their high sintering temperatures.To date, there have been numerous cold sintered materials touted to show promise as potential replacements of LTCC and ULTCC but almost all are based on MoO 4 2− , eg refs 25, 26.Mo is an expensive and endangered raw material 44,45 not suited for mass production in the electronics sector, and new ceramics and composites are required that either eradicate or reduce the concentration of Mo if cold sintering is to compete for applications in MW technology.
The geometric method was used to calculate the bulk density (ρ) of samples, as reported elsewhere. 25,26The ceramic microstructure and grain morphology were examined by scanning electron microscopy (SEM, FEI Inspect F-50).The crystal structure and phase assemblage were determined by X-ray powder diffraction (XRD, Bruker D2 Phaser).Raman spectra were obtained using a Renishaw inVia Raman spectroscopy.The measurement of MW properties (ε r , TCF, tanδ and Q × f) was conducted by the TE 01δ mode with a vector network analyser (Advantest R3767CH), following previously defined protocols. 25,26The cavity was heated by a Peltier device and the resonant frequency (f) was measured from 25°C to 85°C.The corresponding TCF values were obtained using the formula: where the f T and f T 0 are the TE 01δ resonant frequencies at tem- peratures, T and T 0 respectively. (1)

Composition
Sintering condition  The values of density (ρ) and relative density (ρ r ) for coldsintered LMP ceramics are plotted in Figure 1A,B and listed in Table 1.As sintering temperature increases, the values of ρ and ρ r both increase from 2.42 g/cm 3 (82%) at 150°C to 2.66 g/cm 3 (90%) at 250°C, followed by a decrease for higher sintering temperatures (Figure 1A).The optimum cold sintering temperature is therefore, 250°C.Sintering time at 250°C is subsequently increased and ρ and ρ r is further optimized to 2.73 g/cm 3 (93%) for samples cold-sintered 60 minutes (Figure 1B).The maximum values of ρ ~ 2.73 g/cm 3 and ρ r ~ 93% for cold-sintered LMP are higher than those reported for conventionally sintered LMP, (2.72 g/cm 3 and 92%) (Figure 1B), illustrating the great promise of cold sintered LMP as a base to develop new materials.Room-temperature XRD patterns of LMP calcined powders, conventionally sintered ceramics, and cold-sintered samples are shown in Figure 1D,E.LMP has a Mg 2 SiO 4 -type olivine structure (space group: Pnma, ISCD collection code: 201 138) and is composed of PO 4 tetrahedra and Li/MgO 6 octahedra (Figure 1C).Only diffraction peaks associated with olivine-structured LMP are detected in the XRD patterns (Figure 1D,E) with no impurity phases.Full-pattern Rietveld refinement of XRD data for LMP ceramics cold-sintered 60 minutes at 250°C was conducted using a Topas 5 software.The calculated pattern is in good agreement, with low values of R exp = 8.75% and R wp = 11.21%, Figure 1F.The calculated lattice parameters are a = 10.7418Å, b = 5.9070 Å, c = 4.6909 Å for LMP, which agree with those previously reported. 20he data presented above confirms that LMP may be cold sintered to moderately high density (93%) but further improvements were not obtained in the current study.Moreover, LMP has a large negative TCF (-55 ppm/°C) that falls outside excepted values for LTCC and ULTCC (+/-30 ppm/°C) applications.To adjust TCF closer to zero and to improve density, a bespoke cold sintering flux, 0.8KMO-0.2CTO(TCF = +70 ppm/°C) was developed based on KMO (TCF = -70 ppm/ o C) but with a large positive TCF, adjusted through adding CTO (+850 ppm/ o C). 25 Forming composites in this manner is typically used in commercial LTCC which are composed of a low melting temperature glass matrix with negative TCF and high ε r , positive TCF phases such as TiO 2 . 49 and ρ r for cold-sintered LMP-CTO-KMO composites as a function of sintering temperature are plotted in Figure 2A and listed in Table 1.As sintering temperature increases to 200°C and 250°C, ρ and ρ r increase to 2.84 g/cm 3 (94.5%)and 2.92 g/cm 3 (97%), respectively, followed by a slight decrease at higher sintering temperatures.
Room-temperature XRD patterns of cold-sintered LMP-CTO-KMO composites are shown in Figure 2B,D which reveal little change in phase assemblage as a function of sintering temperature.Only peaks from each end member are present: CTO exhibits an orthorhombic, perovskite structure (Pbnm, ISCD collection code: 62149); KMO is monoclinic (C12/m1, ISCD collection code: 16154) and LMP is as described in Figure 1.There is no evidence of secondary phases.Rietveld refinement was performed using a three-phase mix of LMP (Pnma), KMO (C12/m1), and CTO (Pbnm).The calculated pattern matches well with experimental data (R p = 9.65% and R wp = 12.92%), where the weight fractions (LMP ~ 49.9%, CTO ~ 10.6%, KMO ~ 39.5%) are close to the nominal compositions, as shown in Figure 2E.
To confirm further the coexistence of three phases in composites, Raman spectra of cold-sintered LMP-CTO-KMO samples are shown in Figure 2F.As reported previously, 36, 10, and 39 Raman bands are commonly observed in LMP, CTO, and KMO respectively. 25,50,51For LMP, there are 18 external modes (<400 cm −1 , 12 translations of Li, Ni, PO 4 tetrahedra, six vibrations of PO 4 tetrahedra), and 18 internal modes of PO 4 tetrahedra (>400 cm −1 , P-O stretching: ν 1 = 974 cm −1 , ν 3 = 1020-1080 cm −1 , O-P-O bending and P vibration: ν 2 = 416-468 cm −1 , ν 4 = 590-650 cm −1 ), Figure S1.For CTO, Raman bands of 630 and 678 cm −1 are related to symmetric stretching of Ti-O.The 463 and 493 cm −1 bands are the torsional modes of Ti-O.The bands of 171, 214, 234, 275, and 323 cm −1 are related to the bending of O-Ti-O.The 139 cm −1 band belongs to the Ca ions motion.For KMO, Raman bands in the range of 100 ~ 160 cm −1 correspond to a combination of the translations and vibrations of MoO 4 tetrahedra and translations of K ions.The 310 ~ 370 cm −1 bands are related to bending modes of MoO 4 tetrahedra.The 820 ~ 890 cm −1 bands are related to stretching modes of MoO 4 tetrahedra.Raman spectra of LMP-CTO-KMO composites therefore, represent an overlay of Raman bands from individual phases, confirming the existence of LMP, KMO and CTO but without a significant volume fraction of interaction (Figure 2F).
SEM and BSE images of cold-sintered LMP ceramics and cold-sintered LMP-CTO-KMO composites are shown in Figure 3A,B and Figure 3C,D respectively.A denser microstructure in LMP-CTO-KMO composites than in LMP ceramics is evident, coincident with the higher density listed in Table 1.The variations in contrast in BSE images of LMP-CTO-KMO suggest that there are three chemically discrete CTO, KMO, and LMP rich phases, confirmed by the EDS mapping (Figure 3E-L).
The MW dielectric properties of LMP ceramics and LMP-CTO-KMO composites as a function of sintering temperature and time are presented in Figure 4 and listed in Table 1.As sintering temperature and time increase, ε r and Q × f of LMP increases initially before decreasing.The highest values of ε r ~ 6. (Figure 4A,B).The same conditions resulted in optimized values of cold sintered LMP-CTO-KMO composites with ε r ~ 9.1, Q × f ~ 8500 and a near zero TCF ~ -6 ppm/°C (Figure 4C).
The comparison of sintering temperature and MW properties for recently reported and commercial LTCCs are listed in Table 2. LMP-CTO-KMO composites exhibit the lowest sintering temperature (250°C), reducing energy costs in manufacture and ensuring compatibility with all low-cost electrode systems.Q × f of cold-sintered LMP-CTO-KMO composites are is superior to commercial LTCCs, ε r is ideal and they are temperature stable (<+/-30 ppm/°C). 49Cold sintering does not result in lateral shrinkage and hence issues relating to dissimilar shrinkage and thermal expansion between electrode and substrate are alleviated.Many materials are reported with higher Q × f and ε r but these have not been commercialized due to either TCF> +/-30 ppm/°C, sintering temperature >900°C, high cost and environmental issues (Mo and V based systems) or they are over designed for the application; metallized LTCCs do not require ultra-high Q × f as losses are dominated by the metal/ceramic interface.Cold sintered LMP-CTO-KMO, therefore, satisfies the criteria for LTCC applications but we note that the production of cold sintered LTCC requires a radical rethink of ceramic processing and scale-up.

Compounds
Bulk and relative density of cold sintered LMP-CTO-KMO composites as a function of sintering temperature.B, XRD patterns of cold sintered LMP-CTO-KMO composites at different sintering temperatures.C, Schematic of the crystal structures of KMO and CTO.D, XRD patterns of cold-sintered LMP-CTO-KMO, LMP, and KMO and commercial CTO powder.E, Rietveld refinement of cold sintered LMP-CTO-KMO composites.F, Raman spectra of cold-sintered LMP-CTO-KMO, LMP, and KMO and commercial CTO powder [Color figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 3 A and B, SEM images of LMP ceramics cold-sintered 60min at 250°C/600MPa.C and D, BSE images of LMP-CTO-KMO composites cold-sintered 60min at 250°C//600MPa.EDS mapping of LMP-CTO-KMO: (E) layered image, (F) P, (g) Mg, (H) O, (I) Ca, (J) T, (K) K, (L) Mo [Color figure can be viewed at wileyonlinelibrary.com] 5 and Q × f ~ 16 000 are achieved for LMP cold sintered 60 min at 250°C under a uniaxial pressure of 600 MPa F I G U R E 4 The microwave dielectric properties of LMP as a function of (A) sintering temperature (B) sintering time; (C) The microwave dielectric properties of LMP-CTO-KMO as a function of sintering temperature [Color figure can be viewed at wileyonlinelibrary.com] 49 Comparison of sintering temperature (ST) and MV properties of LMP ceramics and LMP-CTO-KMO composites with other LTCCs