1 Introduction

In recent years, the Information Communications Technology (ICT) has progressed in areas like Internet of Things (IoT) technology, microwave telecommunications, Direct-broadcast satellite television (DBS TV), satellite broadcasting, Intelligent Transport Systems (ITS) and Industry 4.0 which promoted scientific research of new materials [1,2,3,4,5]. Microwave dielectric materials applied in dielectric resonators, filters, substrates, etc. play a key role in communication system [6,7,8,9,10]. Li2TiO3 ceramic is one of the members of the ternary rock salt oxide ceramic systems that have been reported to have excellent microwave dielectric properties [11,12,13,14,15,16]. Li2TiO3 possesses high quality factor (Q × f > 15,000 GHz for pure Li2TiO3) [17,18,19] and moderate dielectric constant (22) [15]. The synthesized powders are usually characterized by a large grain size and inferior sinterability [18]. In addition, porous microstructure due to sublimation of lithium [19, 20] and crack formation due to phase transition under high sintering temperatures (above 1150 °C) [19, 21] adversely affect the densification of Li2TiO3 and therefore reducing the sintering temperature is very important to control the above-mentioned defects. Several approaches have been generally applied to enhance the sintering characteristics of ceramics such as addition of dopants [22,23,24,25], addition of sintering aids with low melting points [26,27,28,29], using wet chemistry synthesis methods (hydrothermal, sol–gel,…) [30,31,32,33] and modified solid state reaction methods like high-energy ball mill [34, 35]. Two last approaches have been attractive as a consequence of nanopowders production capability which provide powders with high surface area and stored energy that enhance solid state densification at lower temperatures [36,37,38,39]. Because sol–gel routes can be tedious and often require toxic and/or expensive reactants, it seems that the high-energy ball mill is an enhanced approach to synthesis of materials [40]. For practical applications, high-energy ball mill is the most widely used since it is easy to scale-up, and the precursors are cheap and abundant [41].

In the present work, the synthesis of Li2TiO3 fine powders through a solid state reaction from high-energy ball milled Li2CO3 and TiO2 mixtures has been reported. Furthermore, the characteristics of synthesized powder and sintered parts were examined using XRD, FE-SEM, Raman spectroscopy, BET, PSA, and TG–DTA. The effect of high-energy milling on the microstructures, densification behavior and microwave dielectric properties of Li2TiO3 ceramics was also investigated.

2 Materials and methods

Li2CO3 (Sigma Aldrich, 13,010, 99% purity) and TiO2 (Merck, 1.00808.1000, > 99% purity) were used as starting materials. High-energy ball milling (HEBM) was performed in a PM 400 Retsch planetary ball mill using 250 mL tungsten carbide vials with diameter of 7.5 cm and height of 7.2 cm and tungsten carbide balls with 10–20 mm diameter at 300 rpm. The ball-to-powder ratio of 10:1 was chosen for different milling times of 1, 5, 10 and 15 h denoted as HEBM-1, HEBM-5, HEBM-10 and HEBM-15, respectively. The relative speed of vials and disk was − 2.5. Low-energy ball milling (LEBM) was also carried out in a planetary ball mill using a polyethylene vial and zirconia balls in 99.97% pure ethanol and the sample were denoted as LEBM.

The particle size of as-received powders after high-energy mill was measured by Fritsch Particle Sizer analysette. Milled powders were calcined in an electrical box furnace at different temperatures of 500, 550, 600, 700 and 800 °C with 4 h holding time at maximum temperature and heating rate of 5 °C/min.

The surface area of prepared powders was determined by means of nitrogen adsorption at − 196 °C using an automated gas adsorption analyser (GEMINI 2375, Micromeritics, USA). The mean particle size (D) was estimated using the following equation (Eq. 1)

$${\text{D}}={\text{6}}/{\text{s}}\; \times \;\rho$$
(1)

In which ‘s’ is the BET surface area (m2/g) and ‘ρ’ is the powder density (kg/m3) [42]. The density of synthesized powder was calculated by helium-gas pycnometry (AccuPyc 1330, GA, Micromeritics, USA). Thermal gravimetry-differential thermal analysis (TG–DTA, STA) was performed using Simultaneous Thermal Analyser (BÄHR-STA 503) in air. Differential thermal gravimetry (DTG) curves were also demonstrated.

The synthesized Li2TiO3 powders were subject to grinding for 4 h at 180 rpm in ethanol using zirconia balls and a planetary mill in order to break hard agglomerates. The slurries were then dried, granulated using 5 Wt% of PVA and uniaxially pressed into pellets under a pressure of 200 MPa. The samples were sintered in an electrical muffle furnace at ambient atmosphere between 1000 and 1300 °C with heating rate of 10 °/min and for 3 h holding time at maximum temperature. The densities were measured by Archimedes method (ASTM C373 standard test method). The relative density was calculated by dividing the Archimedes density over theoretical density. The theoretical density of samples was achieved after crushing the pellets and then measuring the powder density by helium-gas pycnometry. Both LEBM and HEBM-5 powder densities were measured 3.40 g/cm3. The X-Ray Diffraction (XRD) patterns of powders and pellets were obtained by Philips PW 3710 with Cu-kα line of 1.5406 Å (40 kV, 30 mA). The data were collected in the 2θ range of 20–80°. The morphology and microstructure of samples were investigated by means of field emission scanning electron microscope (FE-SEM) (TESCAN MIRA 3 LMU). The microstructural observations were performed on polished and thermally etched samples. Thermal etching was performed for each sintered sample at 50 °C below the sintering temperature for 30 min dwell time. Raman spectra have been recorded by DXR-Smart Raman (Thermo-scientific) instrument with Smart Excitation Laser (Thermo Scientific) with laser wavelength at 780 nm. Resolution, number of scans and integration time were 2 cm−1, 5 and 1500 s, respectively. Microwave dielectric properties were measured in the frequency range of 8–10 GHz with TE01δ resonant mode using HP8719C vector network analyzer. The temperature coefficient of resonant frequency τ f was also measured in the temperature range from 25 to 75 °C.

3 Results and discussion

Figure 1 shows the structural examination of Li2CO3 and TiO2 powder mixtures subject to different high-energy milling times. All peaks are attributed to Li2CO3 (JCPDS card number 721216) and TiO2 (JCPDS card number 211272), while no Li2TiO3 or any other phases were identified. The most prominent effect of using high-energy ball milling is peak broadening and intensity reduction which can be attributed to the reduction of crystallite size, loss of long-range periodicity and accumulation of strains in lattice which is accompanied by an increase in the background.

Fig. 1
figure 1

XRD patterns of HEBM-1, HEBM-5, HEBM-10 and HEBM-15 (L: Li2CO3, T: TiO2)

The particle size (D90) of as-received powders are presented in Table 1. As it is obvious, the size of powders has reduced with milling time.

Table 1 Particle size of powders milled for different durations

In order to find out the influence of milling on reaction between precursors, they were analyzed by STA (DTA/TG) and the results of LEBM and HEBM-5 powders are presented in Fig. 2.

Fig. 2
figure 2

a TG and DTG analysis of LEBM, b DTA of LEBM, c TG and DTG analysis of HEBM-5, d DTA of HEBM-5 with heating rate of 10 °C/min

The DTA curve of LEBM powder with the heating rate of 10 °C/min in Fig. 2b exhibits two overlapping endothermic peaks positioned at around 641 and 706 °C. There exists only one endothermic peak in DTA curve of HEBM-5 powder shown in Fig. 2d at about 592 °C, which is about 50 °C lower than the firs peak in DTA curve of LEBM powder. The first derivative of TG (DTG) graphs presented in Fig. 2a shows that the starting and ending points for the decomposition of Li2CO3 in LEBM powder happens at around 511 and 768 °C, respectively, and in HEBM-5 powder reduce to around 354 and 709 °C, respectively. The reduction of decomposition temperature in HEBM-5 powder reveals that the CO2 gas escapes at lower temperature by milling process followed by increasing specific surface area of starting powders. Since as reported in [39] the reaction between Li2CO3 and TiO2 starts from the surface of particles and with increase in powder surface area more active sites for reaction develop. The results of estimated area under DTA peaks of LEBM and HEBM-5 powders (Fig. 2b, d) are 2623 and 1921 respectively which shows that the milling process decreases the heat of formation of Li2TiO3 (about 26%) as well as the reaction temperature between starting precursors.

The mixture of starting powders (LEBM) was annealed at 600 °C without any holding time and the XRD patterns of the powder as well as XRD patterns of starting materials are shown in Fig. 3. Patterns show the existence of all three phases of Li2CO3, TiO2 and Li2TiO3 that determines the broad endothermic peak to be related to the reaction between starting powders as explained by Eq. (2) and the release of CO2 happens simultaneously along with Li2TiO3 formation:

Fig. 3
figure 3

XRD patterns of LEBM powders annealed at 600 °C without holding time along with starting materials

$${\text{L}}{{\text{i}}_{\text{2}}}{\text{C}}{{\text{O}}_{\text{3}}}+{\text{Ti}}{{\text{O}}_{\text{2}}}~ \to {\text{L}}{{\text{i}}_{\text{2}}}{\text{Ti}}{{\text{O}}_{\text{3}}}+{\text{C}}{{\text{O}}_{\text{2}}}$$
(2)

The second peak in Fig. 2b may be related to the melting of remaining Li2CO3 as the transformation temperature is close to melting point of Li2CO3 (Tm = 740 °C [43,44,45]).

Figure 4 illustrates the room temperature XRD patterns of LEBM and HEBM-5 powders annealed at different temperatures of 500, 550, 600, 700 and 800 °C for 4 h. All existing peaks in samples annealed at 700 and 800 °C are related to monoclinic β-Li2TiO3 (JCPDS card: 330,831) without any additional phases. Increasing the temperature causes the increase of peak intensities of superstructure due to higher order of crystallinity, while the width of peaks become narrower due to crystallite growth. Based on the results of phase analysis taken from synthesized powders presented in Fig. 4 the calcination temperature of 700 °C were adopted for the rest of research as no additional phases were observed.

Fig. 4
figure 4

XRD patterns of a LEBM and b HEBM-5 powders annealed at 500, 550, 600, 700 and 800 °C with a holding time of 4 h (LT: Li2TiO3, L: Li2CO3, T: TiO2)

SEM micrographs of Li2TiO3 powders synthesized via two different processes (LEBM and HEBM-5) are demonstrated in Fig. 5. Figure 5a reveals the presence of hard agglomerates with non-spherical morphology related to LEBM powder particles, while Fig. 5b shows different morphology and particle size of HEBM-5 powders with spherical morphology and measured particle size of 147 ± 45 nm. The surface area of this powder was measured to be 15.75 m2/g and based on Eq. (1) the particle size was obtained to be 115 nm in fair agreement with SEM observations.

Fig. 5
figure 5

Morphology of a LEBM and b HEBM-5 powders calcined at 700 °C

Green density of LEBM pellets was calculated to be 60% of the theoretical density which increased to 65% after high-energy mill. Figure 6a shows the results of density measurements of HEBM-1, HEBM-5, HEBM-10 and HEBM-15 sintered at 1100 °C. Increasing the milling time from 1 to 5 h enhances the density from 3.11 to 3.19 g/cm3 which remains almost unchanged adopting higher milling durations. Figure 6b illustrates the density variation of HEBM-5 sample with respect to temperature as well as LEBM. HEBM-5 sample reaches the maximum density of 3.19 g/cm3 at sintering temperature of 1100 °C that remains almost unchanged at higher sintering temperature of 1200 °C and then decreases at 1300 °C. A similar density (3.18 g/cm3) is reported in [18] after sintering the nano-size Li2TiO3 ceramics prepared via sol–gel method with average particle size of 6–12 nm. The density of LEBM samples enhances with increasing the sintering temperature to 1200 °C (ρmax = 3.08 g/cm3) and reduces with further increasing the temperature. Decreasing the density of both samples at 1300 °C could be due to lithium sublimation. Figure 6b also demonstrates the superiority of sintering behavior of high-energy ball milled parts in comparison with low-energy ball milled parts. Improvement of density can be explained by high surface area of fine synthesized powders which increases the driving force of sintering.

Fig. 6
figure 6

a Density of high-energy ball milled samples for different milling times of 1, 5, 10, 15 (HEBM-1, HEBM-5, HEBM-10, HEBM-15) sintered at 1100 °C and b density of LEBM and HEBM-5 specimens sintered at different temperatures for 3 h

Figure 7 shows the room temperature XRD patterns of LEBM and HEBM-5 pellets sintered at 1200 °C for 3 h. Both patterns include peaks related to the monoclinic phase of Li2TiO3.

Fig. 7
figure 7

XRD patterns of LEBM and HEBM-5 samples sintered at 1200 °C for 3 h (all peaks attributed to monoclinic β-Li2TiO3 (JCPDS card: 330831))

Raman spectra of sintered pellets are presented in Fig. 8 and the spectra are very similar and verify the phase purity of samples. In Li2TiO3, the frequencies within 700–550, 400–550, and 250–400 cm−1 region are assigned to Ti–O stretching in TiO6 octahedral, Li–O stretching in tetrahedral, and Li–O stretching in octahedral coordinations respectively. In the Li2TiO3 structure, the lithium can occupy both octahedral and tetrahedral positions [46, 47]. Lorentzian functions were employed to deconvolute the spectra into different modes of vibration (Fig. 9) and the FWHM of different peaks are presented in Table 2.

Fig. 8
figure 8

Raman spectra of LEBM and HEBM sintered at 1100 and 1200 °C

Fig. 9
figure 9

Deconvoluted Raman spectra of a LEBM and b HEBM-5 sintered at 1200 °C into different modes of vibration

Table 2 FWHM of Raman peaks located at 658, 419 and 348 cm–1

Comparing the FWHM bands located at 658, 419 and 348 cm−1 in two samples reveals the higher value in HEBM rather than LEBM which can be due to lower crystallinity of the HEBM-5 sample.

The SEM micrographs of thermal-etched LEBM and HEBM-5 samples sintered at different temperatures are illustrated in Fig. 10. SEM images of LEBM samples sintered at all temperatures exhibit a finer microstructure containing grains within a narrow range of size in comparison with HEBM-5 pellets sintered at the same temperatures albeit HEBM-5 powders possess fine particle size due to mechanochemical effect. The grain size of LEBM and HEBM-5 ceramics sintered at 1100 °C were calculated 4.16 ± 0.38 and 9.16 ± 1.89 µm, respectively. SEM micrographs of the HEBM-5 samples indicate the occurrence of abnormal grain growth. The same result is reported in [18] after sintering of Li2TiO3 powders with particle size in the range of 6–12 nm and is believed to be responsible for deteriorated quality factor of ceramic. Preferential orientation of grain growth is observed in all microstructures after sintering at higher temperatures of 1200 and 1300 °C. The presence of micro-cracks due to order–disorder transition and micro-pores due to lithium sublimation are also obvious at higher temperatures.

Fig. 10
figure 10

SEM micrographs of LEBM ceramic sintered at a 1100 °C, b 1200 °C, c 1300 °C and HEBM-5 ceramic sintered at d 1100 °C, e 1200 °C, f 1300 °C

Figure 11a, b show the microwave dielectric properties of the LEBM and HEBM-5 ceramics sintered at different temperatures. HEBM parts possess higher ε r value at all temperatures with maximum value of 21.8 obtained at 1100 °C. The superiority of ε r in HEBM-5 over LEBM is in agreement with the superiority of its bulk density as shown in Fig. 6 and can be explained by Clausius–Mossotti equation. Based on this equation the dielectric constant of ceramic is controlled by the dipoles present in unit cell volume and dielectric polarizabilities of ions [19, 48]. The higher the density of ceramic, the higher the dipoles in a unit cell volume and therefore the higher the value of ε r [19, 20, 49]. The relation can also be explained because of the unity of ε r of air [18].

Fig. 11
figure 11

a Dielectric constant and b Q × f of LEBM and HEBM-5 versus the sintering temperature

From Fig. 11b it can be observed that quality factor tends to decreases using high-energy mill. The reason can be attributed to the microstructure showing abnormal grain growth (Fig. 10) or lower crystallinity of samples proved by Raman technique (Table 1). The Q × f value of HEBM ceramic increases with increasing temperature up to 1100 °C and then decreases at higher temperatures but the maximum quality factor of LEBM parts was measured at 1200 °C. The temperature variation of resonant frequency τ f of HEBM-5 parts sintered at 1100 and 1200 °C were measured + 26 ppm/°C.

4 Conclusion

Reaction mechanism for the synthesis of Li2TiO3 from the mixture of Li2CO3 and TiO2 were investigated by means of DTA/TG and XRD procedures which proved milling process decreases both the reaction temperature between starting precursors (about 60 °C) and the heat of formation of Li2TiO3 (about 26%). Fine Li2TiO3 powders (< 200 nm) were synthesized through high-energy ball mill for 5 h and these powders had higher sinterability because of the resultant high surface area. The maximum value of ε r  = 21.85 is obtained through density increase resulting from fine particle size. The SEM investigations of sintered pellets showed that abnormal grain growth occurred in high-energy milled sample that could be the reason for deteriorated Q × f of the ceramic. The temperature variations of resonant frequency τ f of ceramic sintered at 1100 °C was measured + 26 ppm/°C.