3.1. Structural investigation:
The preliminary room temperature structural investigation was carried out in the diffraction angle (2θ) ranging from 200- 800 using X-ray diffraction (XRD) technique and depicted in Fig. 1. The observed sharp and unique peaks in the XRD profile differ from raw chemical suggest new structure has formed. The structural identification was done by utilizing standard programming software “POWD”. In this technique, all the XRD peaks were refined in the framework of seven crystal system. The deviation between experimental and computational value of interplanar spacing (shown in Table 1) was found minimum (σ = 0.024) for tetragonal structure, which suggest the investigated sample crystallize in tetragonal structure. The lattice parameters obtain are a = b = 7.804 Ȧ, c = 20.751 Ȧ, c/a = 2.6591 and V = 1263.81 (Ȧ)3. As all the XRD peaks were index well for tetragonal structure with above unit cell parameters, it suggests formation of single phase compound. The appearance of strong intensity XRD peak with Miller index (2 2 0) along 2θ = 32.30 affirms majority of the crystal is growing along [2 2 0]. The observed slight broadening in the XRD profile was expected due to titling of the octahedral FeO6 and TiO6 from their ideal position. Further investigation on structural parameters, atomic positions, reliability factors were done by reitveld refinement technique using MATCH software (version 2.O). The refinement profile and detail atomic position obtained through reitveld technique is illustrated in Fig. 2 and Table 2 respectively. The observed least value of reduce (chi)2 (χ2 = 2.3) and average Bragg R-factor (RB = 12.8 %) for tetragonal system (space group = P4bm) once again clarify that the investigated sample crystallize in tetragonal structure.
Table 1
Comparison of experimental and calculated d-spacing for different Bragg’s planes.
Serial No. | Bragg’s angle (2θ) | d-spacing (experimental) | d-spacing (calculated) | h k l |
1 | 22.77 | 3.9020 | 3.9020 | 2 0 0 |
2 | 26.93 | 3.3079 | 3.3079 | 2 1 2 |
3 | 32.37 | 2.7633 | 2.7591 | 2 2 0 |
4 | 39.89 | 2.2580 | 2.2594 | 2 1 7 |
5 | 43.57 | 2.0755 | 2.0752 | 0 0 10 |
6 | 46.45 | 1.9533 | 1.9553 | 3 0 7 |
7 | 52.37 | 1.7455 | 1.7450 | 4 2 0 |
8 | 57.81 | 1.5935 | 1.5953 | 4 1 7 |
9 | 63.41 | 1.4656 | 1.4653 | 2 2 12 |
10 | 67.89 | 1.3794 | 1.3796 | 4 4 0 |
11 | 72.53 | 1.3022 | 1.3019 | 5 2 7 |
12 | 77.17 | 1.2350 | 1.2354 | 4 0 13 |
The surface morphological investigation of the prepared sample Li2LaFeTiO6 was carried out on pellet sample at room temperature using scanning electron microscope (SEM) is shown in Fig. 3. The polycrystalline nature was reveal from the micrograph as it consists of grains of varying size and shape distributed non-uniformly throughout the pellet sample. Certain sort of clustering of grains observed in the micrograph may be due to high temperature sintering of the pellet sample. Even after sintering at high temperature, certain amounts of voids are noticed on close observation which can trigger the hopping conduction mechanism of charges [14]. Relatively high concentration of small grains in the micrograph was expected due to collapsing of large grains during the transport of oxygen ion vacancies through the grain boundaries. The average grain size was found 3.82 µm through intercept technique.
3.2. FTIR spectroscopy:
The investigation of different modes of vibration and functional groups present in the sample was done by FTIR spectroscopic technique as shown in Fig. 4. The FTIR spectra was found to consists of some characteristic bands located at around 453, 567, 876, 927, 1000, 1134, 1433, 1505 and 3405 cm− 1. The perovskite based materials generally found to possess some characteristic band linked with lattice vibration in the range 850 − 400 cm− 1 [15]. The bands at 453 and 567 cm− 1 assigned to stretching of Ti-O and Fe-O bonds. The observed bands in the range 876–1000 cm− 1 is ascribing to Fe-O-Fe asymmetric vibration. The vibration of Gd-O bonds attribute to the band at 1134 cm− 1[16]. The observed weak intensity bands at 1433 and 1505 cm− 1 may be attributing to vibration of carbonate ion (CO32−) [17]. The absorption of OH− radicals attribute to the band at 3405 cm− 1 in the investigated sample [17].
3.3. UV-Visible spectroscopy:
The opto-electric response of the prepared sample was investigated through UV-Visible spectroscopic technique. The absorbance and reflectance spectra of the investigated sample were depicted in Fig. 5. It was noticed from Fig. 5 (a) and Fig. 5 (c) that the response of the sample was relatively weak in the UV range whereas reverse trend was observed in Visible range. The absorbance cut-off of the sample was found nearly 650 nm which suggest the sample absorb visible radiation in the range 380–650 nm quite strongly. In order to calculate the energy bandgap (Eg) we have utilized the relation (αhν)1/n = A (hν - Eg) [18] proposed by Tauc, Mott and Davis. Where α = absorbance coefficient, ν = frequency of incident radiation, A = proportionality constant & n = index ascribing different electronic transition. For n = 1/2, 3/2, 2 and 3, the transition is direct allowed, direct forbidden, indirect allowed and indirect forbidden respectively. In the present investigation we take n = 1/2 as perovskite materials generally undergoes direct allowed transition. The bandgap (Eg = 1.73 eV) of the material was extracted from the Tauc plot (Fig. 5 (b)) by extrapolating the tangent line on hν axis. Besides Tauc function, we have also utilized Kubelka-Munk function [19] F (R) = (1 - R)/2R (where R = reflectance coefficient) to calculate the bandgap from reflectance data. The estimated band gap (Eg = 1.87 eV) from the reflectance data was found comparable with the value obtained from absorbance data.
3.4. Dielectric spectroscopy:
The frequency dependent variation of relative permittivity/dielectric constant (εr) and loss tangent (tanδ) is depicted in Fig. 6 (a) and Fig. 6 (b) respectively. It was found that both the parameters εr and tanδ decreases with enhancement of frequency for all the selected temperatures. As per literature, the dielectric property of ceramic is very much influence by four kinds of polarizations, namely, electronic, ionic, dipolar/orientation and space charge polarization [20]. At low frequency, all the polarizations contribute significantly whereas at high frequency only electronic polarization attribute towards the total dielectric of the ceramic. The gradual disappearance of different polarizations with enhancement of frequency is the significant cause of dielectric dispersion in the investigated sample. The Maxwell-Wagner and Koop’s phenomenological theory [21] was found quite consistent with the observed dielectric anomaly in the investigated sample. As per the theory, the ceramic dielectrics are consists of two types of layers (i) semiconducting grains and (ii) insulating grain boundaries. The grain boundaries are more prominent at low frequency region whereas the grains are more prominent at high frequency region. Due to insulating nature of grain boundaries, the charge carriers (like electrons) get accumulate on it which increases the trapped charge density and hence εr value. To move the charge carriers via grain boundaries, energy required is more and hence high tanδ value. On increasing the frequency, the density of trapped charges on grain boundaries keep on decrease and charges start to slide over the grain by dissipating less energy and hence εr and tanδ is also low.
Also it is noticed from Fig. 7 (a) and Fig. 7 (b) that the value of dielectric parameters εr and tanδ almost attain saturation level below 3500C and above it both of them start to rise sharply without encountering any transition (magnetic and/or ferroelectric) in the specified temperature range 250C-5250C. Such observation suggests the transition may be lie beyond the experimental constraint temperature. The asymptotic variation of εr and tanδ above 3500C may be assign to thermally activated hopping conduction in the system. As B-site of the investigated sample was occupied by Fe and Ti ion, it was expected that the creation of oxygen vacancies due to high temperature sintering can cause valence fluctuation (Fe3+↔Fe2+) and (Ti4+↔ Ti3+) and enhancing the hopping conduction of charge carriers in the material. The other possibilities of high value of dielectric parameters at elevated temperature are (i) electron-phonon interaction at the lattice site (ii) presence of some imperfection/dislocation in the crystal and (iii) domination of conductivity.
3.5. Impedance study:
Non-destructive impedance spectroscopy method was adopted to study the influence of grains, grain boundaries and electrodes in the characterization of electrical properties. In this method, an electrical perturbation was induced in the dielectric system and its output response was recorded over varying range of frequency and temperature. In order to evaluating the resistive (Z') and reactive (Z'') components of complex impedance (Z*), the following equations were utilized
Z* = Z' + i Z''
Z' = R/(1 + ω2τ2)
Z'' = - ωR2C/(1 + ω2τ2)
Where τ is the relaxation time and is define as τ = RC.
The frequency-temperature dependent variation of resistive (Z') and reactive (Z'') is depicted in Fig. 8 (a) and Fig. 8 (b) respectively. From Z' versus frequency spectra, it was noticed that Z' value fall in a sigmoidal manner below 10 kHz which suggests the lowering in density of trapped charges at grain boundaries [22]. The overlapping of all the curves above 10 kHz irrespective of temperature signifies the release of space charges and domination of ac conductivity in the material at high frequency [23]. It is observed from Z'' versus frequency spectra, the value of Z'' raise proportionally with frequency and attain maxima at a characteristics frequency (commonly known as relaxation frequency) and then gradually fall and attain saturation in high frequency. The shifting of maxima peak in the direction of increasing frequency with enhancement of temperature suggests the relaxation in the material is temperature dependent and is non-Debye type [24]. The broadening of maxima peak in Z'' spectra with enhancement of temperature suggests spread of relaxation and fall of relaxation time in the investigated material. The low temperature relaxation in the material is ascribing to immobile charges and that at high temperature is due to oxygen vacancies and/or some intrinsic defect. The NTCR behavior in the material is also reflected from both Z' and Z'' spectra.
The change of Nyquist plots (i.e., Z'' vs Z') at various temperatures was depicted in Fig. 8 (c). The observed single semicircular arc in the Nyquist plot suggests the resistive and capacitive property was highly influence by the grains in the material. As per Debye hypothesis, an ideal semicircular arc with center lie on the Z' axis corresponds to single relaxation process and homogeneity in the material. The depressed semicircle with center lies below the Z' axis suggests the non-Debye type of relaxation and inhomogeneity in the investigated sample [25]. To correlate the electrical and microstructural properties, an equivalent circuit model consists of capacitor (C) and resistor (R) was incorporated. The experimental data were fitted with the theoretical ones using software ZSIMWIN (version 2.O). In order to counter the observed depression in the semicircular arc, a new element, namely constant phase element (CPE) Q was introduce in the RC network. It was noticed that the experimental data were fitted well for the equivalent circuit (C || Q || R) (insert in Fig. 8 (c)). Such observation suggests significant contribution of grains in the electrical and transport mechanism [26]. The slight deviation observed in the semicircular arc from the model generated curve at low frequency was might be due to weak contribution of grain boundaries.
3.6. Modulus study:
The complex modulus spectroscopic method is also incorporated in addition of impedance technique to have better understand of different processes such as (i) relaxation (ii) conduction (iii) transport etc. occurring in the investigated sample. The advantage with modulus technique is that it can extract the contribution of small capacitance in the material which sometimes gets suppressed in impedance technique due to inhomogeneous distribution of grains [27]. The real (M') and imaginary (M'') components of complex modulus (M*) was extracted using the following equations
M* = M' + i M''
M' = C0/C [(ω τ)2/{1 + (ω τ)2}]
M'' = C0/C [(ω τ)/{1 + (ω τ)2}]
Where C0 is the geometrical capacitance and is define as C0 = ε0A/d (ε0 = permittivity of free space, A and d are respectively area and thickness of the circular pellet)
It is observed from the variation of M' versus frequency spectra (Fig. 9 (a)) that M' has negligibly small value in the low frequency region which suggest insignificant contribution of electrode polarization. The monotonic dispersion observed in M' spectra at intermediate frequency region attributes the short range mobility of charge carriers influence the conduction process. The mobility is due to the lack of restoring force of charge carriers in external electric perturbation. The merging of all curves irrespective of temperature ascribe to lack of space charges in high frequency region [28, 29].
From M'' versus frequency spectra (Fig. 9 (b)), it is found that M'' value increases with frequency and reach a maxima. The relaxation in the material is once again clarify from M'' spectra. The temperature dependent relaxation is observed from the shifting of relaxation peak towards high frequency which is due to activation of hopping conduction through thermal agitation. The asymmetric broadening of M'' spectra signifies non-Debye type of relaxation [30, 31].
In order to illustrate, the response of small capacitance and large resistance in the material, the relative variation of Z'' and M'' with frequency (Fig. 10 (a)) is studied at a particular temperature. It also enables to distinguish whether the relaxation in the material is dominated by short range or long range charge carriers. For long range process, Z'' and M'' relaxation peaks appear at nearly common frequency whereas for short range process, they appear at different frequencies. The observed mismatching of Z'' and M'' peaks in the investigated sample suggests short range charge carriers are dominating the relaxation mechanism [32, 33].
The relaxation time (τ) obtained from Z'' and M'' spectra is studied as a function of temperature as depicted in Fig. 10 (b). In order to estimate the value of τ we have utilize the relation τ = 1/2πfr, where fr corresponds to frequency of relaxation. The observed values of τ are found to fitted well with Arrhenius relation [34] τ = τ0 exp (-Ea/KBT), where Ea = activation energy. The Ea = 1.109 eV value estimated from Z'' spectrum correspond to localized conduction while Ea = 1.087 eV obtained from M'' spectrum correspond to delocalized conduction [33, 34]. The observed closeness of Ea value illustrates that similar sort of charge carriers are participated in both the conduction process.
3.7. AC conductivity study:
The frequency dependent variation of ac conductivity (σac) at selected temperatures is depicted in Fig. 11 (a). It is noticed that the variation of σac is relatively small in low frequency range whereas at high frequency it rises sharply. The observed dispersion in σac spectra can be ascribe to the presence of space charges as well as disorderness of the cations at A-site and B-site [35]. The conductivity spectra was found very much consistent with Jonscher’s power law σac = σ0 + Aωn [36]. As per the law, the origin of frequency dependent conductivity is related to the relaxation phenomenon of mobile charges. Also it is noticed that the slope of the ac conductivity spectra is relatively high at high frequency than that at low frequency region. Such observation suggests the hopping mechanism of charges is quit dominating in the conduction process. The presence of multi-valent ions Fe and Ti at B-site can initiate the hopping conduction of electron between (Fe3+ ↔ Fe2+) and (Ti4+ ↔ Ti3+) in the investigated sample.
The temperature dependent variation of σac at selected frequencies is shown in Fig. 11 (b). The NTCR behaviour of the material was confirm as σac increases with temperature. The high value of σac at high temperature region attributes to higher mobility of charges due to hopping between localised sites [37]. The activation energy (Ea) was calculated in high temperature region by fitting the Arrhenius equation σac = σ0 exp (-Ea/KBT) in the experimental data. The obtained values of Ea were 1.027 eV, 0.791 eV, 0.475 eV, 0.329 eV and 0.295 eV at frequencies 1 kHz, 10 kHz, 100 kHz, 1 MHz and 5 MHz respectively. The Ea value was found high at low frequency than that of high frequency. With increase in frequency, the transportation of charge carriers between localized sites enhances and hence activation energy decreases.
Table 2: Detail atomic position of various elements in Li2GdFeTiO6 obtained through Rietveld refinement using MATCH software
Atom site
|
Atom type
|
X
|
Y
|
Z
|
Li 1
|
Li+
|
0.5
|
0
|
0.5615
|
Gd 1
|
Gd3+
|
0.5
|
0
|
0.5615
|
Fe 1
|
Fe3+
|
0
|
0
|
0.0843
|
Ti 1
|
Ti4+
|
0
|
0
|
0.0843
|
O 1
|
O2-
|
0
|
0
|
0.4539
|
O 2
|
O2-
|
0.2642
|
0.2357
|
- 0.006
|