Effect of calcination temperature and duration on structural and dielectric properties of CaFeO_3-δ

2.1 Materials and methods

The synthesis of CaFeO_3-δ perovskite was achieved using a solid-state method, as depicted in Fig. 1. In this method, the diffusion capacity of the cations and anions in the solid state plays a critical role, requiring high temperatures and long reaction times to increase the diffusion rate and reaction rate for obtaining a stable phase. To achieve this, stoichiometric amounts of high-purity dried powders of CaCO3 (99.9 atom %, sigma–aldrich) and Fe2O3 (99.9 atom %, sigma–aldrich) were ground in an agate mortar for 60 min, following the general chemical equation (1). After grinding, the product was homogenized in a container containing acetone for 4 h.

Close

Fig. 1

The different stages of CaFeO_3-δ preparation.

Export to PPT

The ceramics were heated at different temperatures (600 °C, 700 °C, 800 °C, 900 °C, 1000 °C, and 1100 °C) for two different durations (4 h and 10 h), as shown in Fig. 2. After heating, the powders were ground with an agate mortar for 30 min and mixed with a 3 % solution of polyvinyl alcohol (PVA) binder. The resulting mixture was pressed under 8 tons (685 MPa) of pressure using a hydraulic press to make pellets with dimensions of 12 mm in diameter and 1–2 mm thickness for dielectric measurement. These pellets were then sintered at 1100 °C for four hours in the air.

Close

Fig. 2

Clacination cycle of CaFeO_3-δ.

Export to PPT

Using X-ray diffraction (RX) (XPERT-PRO with CuK radiation where (λ = 1.5406) across the range 2θ ∼10-80° with a step size of 0.02°, the effect of calcination temperature and duration on the proprieties structural of the CFO samples was examined. The Rietveld refinement was carried out using FULLPROF suite software and was extended to refine the crystal phases, also for calculating the lattice parameters. The Raman spectroscopy was successfully used to study the structural evolution of ceramics, and measurements were performed at room temperature in the region of 100–1000 cm⁻¹, This spectrometer is equipped with a monochromatic excitation source with a laser power of ≈10 mW and a wavelength of 410 nm. In addition, the microstructures of the pellets are examined using a scanning electron microscope (SEM, model JEOL JSM5410). An energy-dispersive X-ray analyzer (EDX, model Kevex, Sigma KS3) is used for the elemental analysis of sintered samples.

The dielectric measurements are performed on disk-shaped pellets. After sintering, both sides of the pellets are polished with an abrasive paper. The sides of the pellets are then metalized by a thin layer of silver lacquer, to form a plane capacitor. The capacitance and corresponding dielectric losses (Tanδ) as a function of temperature from ambient to 500 °C and frequency from 1000 Hz to 2 MHz, of the capacitor thus formed, are measured using an independence meter operating under a low excitation level of 1 V inducing an electric field. These dielectric measurements are performed with an automatic impedance bridge HP4284A. The degree of charge mobility within the substance is measured by its electrical conductivity. The movement of the charge can be used to determine the product's qualities (insulator, semiconductor, and conductor). The obtained dielectric measurements were used to determine the AC conductivity (σ_ac), which was assessed using the relationship shown below (2):

The Nyquist plot is a crucial experimental piece of information used in impedance spectroscopy to evaluate the circuit. The parallel RC circuit produces an arc on the Z“-Z' plane; and an additional capacitive element connected in parallel with a resistor adds a semicircular arc, while the resistive element produces a displacement on the real axis. The impedance Z is calculated using the formula (3):

To trace the Nyquist diagram, we separate Z _Reel (Z') and Z _imaginary (Z’’) following the Eq. (4):

Furthermore, ω represents the angular frequency (2πf), C for the sample's geometric capacitance, and ε₀ for the permittivity of space (8.854*10^-12F/m). The contribution of the ceramics grain and grain boundary impacts is studied using complex impedance and Cole–Cole plots which were fitted using the Z-view.

3.1 XRD results

In order to identify the crystallographic structure of CaFeO_3-δ samples, calcined at different temperatures (600 °C to 1100 °C) and durations (4 h and 10 h), the X-ray diffraction was investigated as presented in Fig. 3. The crystallographic structures were refined using the Rietveld technique while varying temperature and calcination time. This analysis was facilitated by FullProf software, in conjunction with WinPlotr. To model the peak profiles, a pseudo-Voigt function was employed, and a 12-coefficient polynomial function defined the background level. The refinement process included various parameters such as fractional atomic coordinates, scale coefficient, atomic occupancy, preferential orientation coefficient, isotropic displacement coefficient (in Å^2), zero-point detector, global isotropic displacement parameters (temperature), cell coefficients, background refinement (via a 12-coefficient polynomial function), half-width coefficients, asymmetry coefficients, and profile shape coefficient (η). The results show that all the powders crystallize in the pure perovskite phase without the presence of a secondary phase. Above 700 °C of calcination temperature, the characteristic peaks of the pure orthorhombic perovskite-like phase appear and remain the same with the change of temperature and duration of calcination. The effect of calcination temperature and duration on crystallinity phase parameters of CaFeO_3-δ powders we carried out by the Rietveld refinement method using Full-Prof software. The fitted results are shown in Fig. 4 and confirmed a single orthorhombic crystallographic phase formation of CaFeO_3-δ powders with a space group of Pcmn. The structure obtained was in accordance with the previous work reported by Zafar, N. et al (Zafar and Khan, 2015). The lattice parameters (a, b, and c), cell volume (V), B position, site occupancy, and reliability factors obtained from the Rietveld fitting are grouped in Table 1. As expected, the fitted cell parameters, lattice coordinate, and B position are nearly in a zigzag by varying the time and temperature treatment, This might be caused by substitution disorder and compositional changes in the cation arrangement at one or more Ca and Fe crystallographic positions (Gouitaa Najwa et al., 2022), in addition, the change of the calcination conditions can affect the thermal vibrations of the atoms in the crystal structure. But the cell volume increases with the increase of temperature (for both 4 h and 10 h), due probably to many factors such as internal stress, surface energy, and surface stress. The reliability factor χ 2; which represents the ratio between the theoretical and experimental fit; is between 1 and 2 for all the powders and close to 1 for the ceramic calcined at 1000 °C/4H, which confirms the good crystallinity of this powder. A good agreement between the observed and calculated patterns was obtained with Rp ≤ 5 %, and Rwp ≤ 7 % for all ceramics. The broadness of the diffraction peaks of the CaFeO_3-δ phase (Table 1) confirms the nano-crystalline nature of the prepared materials, which indicates that the diffraction peak half-width FWHM and average crystallite size depends clearly on the temperature and duration treatment. The average crystallite size was calculated from Scherrer's formula (Akl and Hassanien, 2021; Hassanien and Akl, 2018) as follows (5):

Close

Fig. 3

XRD spectra of CaFeO_3-δ powders calcined at different temperatures and durations.

Export to PPT

Close

Fig. 4

Refinement of CaFeO_3-δ compounds calcined at different temperatures and durations using the Rietveld method by Full-prof program.

Export to PPT

where D is the average grain size, β is full width at half maximum (FWHM), and γ is the wavelength of the copper target (0.15405 nm). The Bragg’s diffraction angle is denoted by θ. The values of FWHM and average crystallite size found are reported in Table 1 and show that the half-height width was narrower and the average crystallite size increased with increasing calcination temperature for the two durations 4 h and 10 h. But with changing the calcination time, an inverse phenomenon is observed which consists of an increase in FWHM and a reduction of crystallite size (Basak et al., Feb. 2022). This indicates that the crystallinity of the samples was improved with temperature increasing at low time duration. The same increase in crystallite size was found by Barbara Kucharczyk et al (Kucharczyk et al., 2019) for LaFeO₃ compounds with a change in calcination temperature. Also, This result was consistent with that of Mahmoud Lebid (Lebid and Omari, 2014) and Qing Lin (Lin et al., 2018). The influence of calcination temperature and duration on the crystallite size of CaFeO_3-δ has not been reported in the previous works.

3.2 SEM results

The morphology of the grains of CaFeO_3-δ ceramics at different temperatures of calcination (900 °C-1100 °C) and durations (4 h and 10 h) and sintered at 1100 °C/4h is shown in Fig. 5. SEM analysis revealed that the layers of perovskites are independent of the calcination temperature and duration, and they are not homogeneous with the presence of a mixture of small and large grains in each sample. The grains formed in the pellets are porous in their bulk and surrounded by a number of small particles, whose average diameter is 0.5 µm. In addition, we can observe an aggregation phenomenon in the sample (especially in samples calcined at 1000 °C) and a decrease of porosity, this indicates that the chemical stress between the crystal increases with the increase in this calcination temperature (Chetan et al., 2020), leading to the shrinkage and aggregation of crystals. The same phenomenon was reported by Barbara Kucharczyk et al. for LaFeO₃ ceramics synthesis using the sol–gel method (Kucharczyk et al., 2019). We can also notice that the ceramics calcined for 10 h are more porous than those calcined for 4 h. The calculated average grain size obtained from the SEM image (Table 2) confirmed that the samples calcined at high temperature ∼ 1100 °C, exhibit a large particle size in both duration (4 h and 10 h). These results agree with the crystallite size found in DRX results, and they are well matched with the values of LaFeO₃ synthesized by Barbara Kucharczyk (Kucharczyk et al., 2019). In addition, Fig. 6 presents an Energy Dispersive Spectroscopy micrograph of the CaFeO_3-δ perovskite sample calcined at different temperatures and durations and sintered at 1100 °C/4H, and Table 3 records the relative abundance of the elements (%) for all samples. The EDS analyses indicate the presence of Ca, Fe, and O elements in the CaFeO_3-δ sample with some impurities of C since it is an outcome of the initial reagent; the percentage of C decreases with increasing temperature and calcination duration. according to the EDS studies, the Ca/Fe ratio is close to 1:1, as Table 3 demonstrates, which is consistent with the stoichiometry expected in the range of experimental errors, It can also be noticed that the atomic percentage of oxygen increases with the increase of the temperature and the duration of calcination which can mean the decrease of the vaccination sites of oxygen. These results highlight the importance of temperature and calcination conditions in influencing the elemental composition and properties of CaFeO3-δ perovskite.

Close

Fig. 5

Micrographs SEM of CaFeO_3-δ powders calcined at different temperatures and durations and sintered at 1100 °C/4H.

Export to PPT

Close

Fig. 5

Micrographs SEM of CaFeO_3-δ powders calcined at different temperatures and durations and sintered at 1100 °C/4H.

Export to PPT

Close

Fig. 6

The EDS micrograph of the CaFeO_3-δ perovskite samples calcined at different temperatures and durations and sintered at 1100 °C/4H.

Export to PPT

3.3 FT-IR results

The infrared spectra of CFO samples calcined at various temperatures and durations are depicted in Fig. 7, and Table 4 summarizes the vibrations and assignment. The antisymmetric stretching vibration of the Fe-O bond in the regular octahedron FeO₆ (Swatsitang et al., 2017; Lin et al., 2018; Saikia et al., 2022) is shown to represent the main band between 569 and 600 cm⁻¹. This result is in line with those of a prior study by Qing Lin et al (Lin et al., 2018). The Fe-O band is present in the synthetic perovskite oxides La_0.85Mg_0.15FeO₃ with a wavenumber of about 550.16 cm⁻¹ (Lin et al., 2018). Additionally, Qing Lin et al. (Lin et al., 2018) noted that the LaFeO₃ samples' spectral absorption pattern revealed a major band of Fe-O stretching vibrations at 574 cm⁻¹ and 400 cm⁻¹. Furthermore, a triplet of 3 bands consisting of a broad, intense absorption band at 1459 cm⁻¹, a thin, intense band at 874 cm⁻¹, and a thin, weak band at 712 cm-¹ appears, These are the three modes of C-O elongation of the carbonate group that characterizes CaCO₃ (Lin et al., 2018). The latter shows that the samples calcined at 700 °C/4H and 800 °C/4H have a certain level of impurity since it is an initial reagent that disappears with an increase in calcination temperature. The symmetrical stretching vibration band of the carboxyl root appears at around 2359 cm⁻¹ (Lin et al., 2018), and the existence of a peak at 3642 cm⁻¹ indicates the presence of calcium hydroxide Ca(OH)₂ according to the work reported in the literature (Lin et al., 2018). On the other hand, with the increase in calcination temperature, the intensities of the bands related to carbon bonds and OH bonds decrease gradually. These OH and C are probably related to small amounts of acetone and precursor residues. It has been observed that certain bonds are not detectable by XRD, but can be identified through Infrared analysis. These bonds are formed mainly at the surface due to exposure to ambient air. Additionally, the appearance of CO2 can be attributed to the release of these bonds on the surface of the powders. These results confirm the purity of the synthesized CFO and are consistent with the literature spectra (Kozakov, 2016).

Close

Fig. 7

FT-IR spectra of CaFeO_3-δ powders calcined at different temperatures and durations.

Export to PPT

3.4 RAMAN results

The Raman spectroscopy analysis of all the CFO powders calcined at different temperatures and durations of calcination is shown in Fig. 8. Many vibrational modes of the orthorhombic CaFeO_3-δ structure in the Pcmn space group are undetectable because of their low polarizabilities. For CFO calcinated at 700 °C/4H and 800 °C/4H, four modes 218 cm⁻¹, 287 cm⁻¹, 403 cm⁻¹ and 602 cm⁻¹ are registered that refers to a spectrum of Fe₂O₃ (Kozakov, 2016). these peaks disappeared with increasing temperature and calcination time for the other CFO; which means that the 900 °C temperature is the beginning of the disappearance of the precursors. For the other powders calcined at 900 °C, 1000 °C, and 1100 °C for 4 h and 10 h, the Raman spectra contain the same modes. Indeed, the symmetric-stretching modes at 250, 305, and 707 cm⁻¹, are corresponded to the breathing-type distortion of the FeO₆ octahedra (Deka et al., 2018). In addition; the peak 375 cm^−1, can be attributed to the mixed motion of Ca and O ions, as Ghosh et al (Kucharczyk et al., 2019) illustrated. The mode close to 667 cm⁻¹ is assigned to the maghemite γ-Fe₂O₃; while the observation of maghemite iron implies a reduction in the atmosphere (Gouitaa et al., 2023). Less study of lattice dynamical calculations and assignment of the Raman modes for CaFeO_3-δ materials are reported. Moreover; there are some controversial points in the assignment of the mode. Further details will be studied in future work. According to the spectra found in the literature, the CFO therefore had a good structural organization with an orthorhombic perovskite phase (Kozakov, 2016).

Close

Fig. 8

Raman spectra of CaFeO_3-δ powders calcined at different temperatures and durations.

Export to PPT

3.5 Dielectric studies

The temperature dependence on dielectric permittivity of CaFeO_3-δ ceramics at different calcination temperatures (900 °C-1100 °C) and durations (4 h and 10 h) is shown in Fig. 9. In both calcination durations (4 h and 10 h), we can notice the presence of two anomalies T₁ and T₂ for calcination temperatures of 900 °C and 1000 °C. While at 1100 °C, the two anomalies are merged into one broad phase transition. The samples calcined at 1000 °C showed a good enhancement of dielectric properties. However, the dielectric permittivity reaches colossal values for both duration times which are about 10⁵. This colossal dielectric response in CaFeO_3-δ is related to the high capacitance of the grain boundaries which is an extrinsic effect. The understanding of the origin of this response is recorded according to the IBLC model described by I. Jalaf et al. (Jalafi, 2023), and by Y. Zhang et all (Zhang, 2019) in which the ε′ is given by the following Eq. (6):

Close

Fig. 9

The dielectric permittivity as a function of temperature rise at 1000 Hz frequency for samples CaFeO_3-δ calcinated at 900––1000-1100 °C for two durations a) 4 Hours, b) 10 Hours, and sintered at 1100 °C/4H.

Export to PPT

Close

Fig. 10

The dielectric permittivity as a function of frequency rise at ambient temperature for samples CaFeO_3-δ calcinated at 900––1000-1100 °C for two durations a) 4 Hours and b) 10 Hours and sintered at 1100 °C/4H.

Export to PPT

3.6 Electrical properties

The effect of calcination temperature and duration time on the conductivity of CaFeO_3-δ ceramics can be investigated by studying the alternative conductivity. Thus, the evolution of Ac-conductivity as a function of frequency, at 340 °C test temperature, for all samples, is represented in Fig. 11. It is clear to see that the conductivity is frequency-independent at low-frequency regions which is called continuous conductivity σ_dc, which means that the conductivity follows a direct current due to the presence of free charge carriers and the weak electron hopping (Singh Yadav et al., 2020.; Narsimulu et al., 2015; Hajlaoui et al., 2019). At high-frequency regions, the conductivity increases with a frequency that is attributed to the hopping mechanism of the small polaron as described by N. Gouitaa et al (Gouitaa et al., 2021)and by Kyeong-Han Na et al (Na, 2018). It is confirmed that for solid materials the small polaron concept is present which is responsible for the enhancement of Ac-conductivity. According to Maxwell–Wagner theory and Koop’s model, ferrite grains that conduct electricity are separated by non-conductive, oxygen-rich grain boundaries. These non-conductive grain boundaries are more significant in low-frequency regions. However, at higher frequency regions, conductive grains are more prominent, thus the frequency of skipping between Fe²⁺ and Fe³⁺ ions of calcium ferrite increases (Hakeem, 2021; Murugesan and Chandrasekaran, 2015; Unal, 2022). With changing the calcination conditions, a noticeable increase in σ_ac values for the sample calcined at 1000 °C/10 h (The maximum conductivity is 0.22 S/m at 1 MHz) more than the sample calcined for 4 h, which is probably due to the grain growth showed in Fig. 5.

Close

Fig. 11

Electrical conductivity curves as a function of frequency rise at 340 °C for all samples CaFeO3-δ calcined for 4H (a) and calcined for 10H (b) and sintered at 1100 °C/4H.

Export to PPT

The frequency dependence of the conductivity, often described as the Universal Dynamic Response (UDR), can be succinctly expressed using Jonsher's law (Singh Yadav et al., 2020.; Jafarpour et al., 2022) by the following Eq. (7):

Fig. 12 displays the variation of ln(σ_DC) as a function of the inverse of temperature (1000/T), which provides the activation energy value for electrical conduction of CFO ceramics at 100 kHz where the temperature dependence of the conductivity, determined from Arrhenius law (8).

Close

Fig. 12

. Conductivity ln(σDC) as a function of 1000/T at 100 kHz for samples CaFeO_3-δ calcinated at 900––1000-1100 °C for two durations a) 4 Hours and b) 10 Hours, and sintered at 1100 °C/4H.

Export to PPT

It is known that complex impedance spectroscopy is a powerful tool to separate the effect of grain and grain boundaries, it is used to study the electrical relaxation behavior in these materials (Singh Yadav et al., 2020.). For this purpose, the complex impedance using the Cole-Cole plot is studied. Fig. 13.a shows the fitted plots using a series circuit of two parallel combinations [(R1, C1), (R2, CPE1)] Fig. 13.c, thus indicating the contribution of the sample grains in the high-frequency region and the grain boundaries in the low-frequency part. From this fit, we were able to extract the resistance values R1 and R2 attributed to the grain resistance (R_G) and grain boundaries resistance (R_GB) respectively. these circuits are parallel and connected in series for the sample calcined at 1000 °C/4h (all the other samples exhibit the same behavior and are not represented here). Two semi-arcs are obtained in the tested samples; one is related to the grain contribution, and the other is related to the grain boundaries contribution. It is known that when the center of the semicircle is on the real axis, the relaxation follows the ideal Debye behavior.

Close

Fig. 13

A. Nyquist plots of CaFeO_3-δ ceramics calcined at 1000 °C/4h and sintered at 1100 °C/4H. b. Nyquist plots at 220 °C of CaFeO_3-δ ceramics calcined at different temperatures and durations and sintered at 1100 °C/4H. c. equivalent circuits.

Export to PPT

In the present study, the center of semicircles is below the z’ axis; confirming the poly-dispersive nature of the dielectric response in this material (Giannaccini et al., 2018). The radius of these semicircles decreases when the temperature increases which indicates a negative thermal resistivity coefficient (NTCR) behavior of the materials (Gouitaa et al., 2018). This variation in radii characterizes semiconductor materials indicating a non-ideal Debye relaxation behavior (Khelifi et al., 2015) which could be caused by many factors such as stress phenomena, deformation and distribution of atomic defects as well as the orientation of the grain and the grain boundaries. In addition, we notice that the semicircles partially overlap as the temperature increases. According to the IBLC model, these phenomena are typically related to the existence of a relaxation time distribution; the various components or elements within the system have different time scales for returning to equilibrium or responding to an external perturbation; in which the resulting complex impedance is composed of two overlapping half circles (Mallmann, 2015). Upon analyzing the curve of Nyquist plots of CaFeO_3-δ ceramics calcined at different temperatures and durations at 220 °C in Fig. 13-b, it was observed that the CFO calcined at 1100 °C/10H and 900 °C/10H had the largest circle radii. This phenomenon is believed to be one of the reasons for the decrease in conductivity values (Hakeem, 2021). Likewise, CFO calcined at 1000 °C/10H, which has a lower conductivity value, was found to have a smaller radius and sample resistance (R) as compared to the others, as shown in Table 7, which groups the values of the adjusted parameters (R_B, R_GB, C_B, and C_GB) at T = 220 °C. Based on the same table, all samples had Rg values that were significantly lower than Rgb values. And this demonstrates the capacity for the creation of the internal barrier layer (IBLC). The IBLC model predicts that the behavior of Rg < Rgb can result in a significant amount of polarization at the grain boundaries and interface (Singh Yadav et al., 2020.).