Introduction

Nowadays, inorganic oxides with high ionic conductivity are of great interest because of their potential applications as electrolytes in solid oxide fuel cells (SOFCs), gas sensors, gas separation membranes, etc. [1–5]. Recent developments in materials with high oxide ion conductivity have generated enormous interest in the design and development of new perovskite type oxides, because of their unique structural features suitable for aliovalent doping. Such doping creates vacancies in oxygen sublattice, which helps the migration of oxygen ions through the lattice [6, 7]. Among perovskites, LaTiO3 is a promising material in diverse branches of applications such as electro-optical, piezoelectric, dielectric, ferroelectric, electrical conductivity and catalysis due to its unique properties [8]. Earlier, studies have been made to evaluate structural, electrical and morphological features of LaTiO3 synthesized by various methods. It has been reported that doping in the A-site or B-site of this compound improves its overall properties [9–17]. Several authors have doped in the A-site divalent cations and studied its various properties [18–25]. Zang et al. reported the photocatalytic property by doping Cu at B-site of LaTiO3, which was found to be highly effective for Fenton mechanism [26]. Literatures indicate that LaScO3 is both proton and oxide ion conductor. The reason for high conductivity in scandate is a larger B-site ion, which will widen the lattice and create more space for the oxide ion to travel through the lattice [27–30]. Titanium exists in both trivalent and tetravalent oxidation states, and the ionic size of scandium is nearer to that of titanium when compared to other possible dopants such as Cu. Therefore, doping Sc in LaTiO3 + δ would not alter its crystal structure and thus affect its properties. The present work reports the incorporation of Sc3+ into LaTiO3 + δ to progress the total conductivity and structural properties of LaTiO3 + δ to make it more suitable for high-temperature applications. After structural and thermal analysis of the materials, their total conductivity is studied through impedance analysis.

Experimental details

A series of LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3) were synthesized by sol-gel method. High-purity (99.9 %) La(NO3)3·6H2O, Sc(NO3)3·6H2O and C12H28O4Ti (Aldrich chemicals) were used as precursors. The precursor gel was prepared by using the following chemicals: 4.4 mmol La(NO3)3·6H2O, 3 mmol C12H28O4Ti and 0.5 mmol Sc(NO3)3·6H2O (for the perovskite composition with x = 0.2) which were dissolved in 50 mL distilled water. After the formation of clear solution, 26 mmol citric acid and 5 mmol polyethylene glycol were added and the solution was heated to 60 °C with continuous stirring for 8 h to obtain a gel. Similarly for other compositions, calculated amounts of precursors were taken and reactions were carried out. The gel was heated to 120 °C in a hot air oven to get a brown porous solid which was calcined at 800 °C for 2 h to obtain a white powder of LaTi0.8Sc0.2O3 + δ . The calcined powder of each composition was ground into a fine powder in an agate mortar and then uniaxially pressed into cylindrical pellet of 10 mm diameter and 1.6 mm thickness using hydraulic press by applying a pressure of 1,800 kg/cm2. Subsequently, the pellets were sintered at 1,200 °C for 10 h in air atmosphere and further characterizations and conductivity measurements were carried out.

The precursors in the form of dried gel were subjected to thermogravimetry/differential scanning calorimetry (TG/DSC) performed using NETZSCH STA 449F3 in static air atmosphere from room temperature to 1,000 °C at a heating rate of 10 °C min−1 to determine the temperature of formation of perovskite oxide phase. The phase identification and structural characterization of the perovskites was carried out by X-ray diffraction (XRD) using PHILIPS X-ray Diffractometer with CuKα radiation (λ = 1.540598 Å) with 2θ ranging from 10° to 60° at an acquisition rate of 0.02°/s. Unit cell dimensions of the crystal structure were refined by Rietveld (Le-Bail fitting) method using X-pert HighScore Plus program. The microstructure of the sintered pellets was recorded using scanning electron microscopy (SEM) in Carl-Zeiss Auriga SEM in secondary emission mode. Oxidation states of the perovskites were analysed using by X-ray photoelectron spectroscopy (XPS, Specs-phoibos analyser) with a monochromatized Al K X-ray source of 1,486.58 eV. Total conductivity of the perovskites was studied through impedance measurement using a computer controlled impedance analyser HIOKI 3532 LCR HITESTER in the frequency ranging from 50 Hz to 10 MHz and at a temperature range of 30–650 °C. For better ohmic contact, high-temperature curing silver paste was applied on both faces of the pellet.

Results and discussion

Formation process of LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3) using TG/DSC

Thermal analysis results of all the dried gels obtained during the synthesis of undoped and Sc-doped LaTiO3 + δ perovskites are shown in Fig. 1. Figure 1a shows the thermal analysis of undoped LaTiO3 + δ . The curves depict the weight loss profiles that occur in several stages. The weight loss observed in the range between room temperature and 300 °C is due to the removal of adsorbed and structural water from the sample. The second stage between 300 and 500 °C is accompanied by an endothermic peak at the DSC curve with weight loss that is due to the decomposition of nitrates and citrate complexes. The third stage in the range from 500 to 800 °C, and its accompanying weight loss represents the loss of organics and carbon dioxide from metal oxides. This is also represented by an endothermic peak at the DSC curve. The fourth stage in the range of 800–1,000 °C with weight loss is attributed to the oxygen loss and formation of perovskite phase and the structure variation in TiO6 octahedron under higher calcination temperatures [31]. After doping Sc with different concentrations in the B-site as shown in Fig. 1b–d, no major changes were observed in TG/DSC curve. This indicates that all samples can be calcined at 800 °C for 2 h to obtain an oxygen-deficient perovskite phase.

Fig. 1
figure 1

TG/DSC data of LaTi(1 − x)Sc x O3 + δ synthesized by sol-gel method for various compositions: a x = 0, b x = 0.2, c x = 0.25 and d x = 0.3

Full size image

Powder X-ray diffraction and structure

Powder XRD patterns of the sintered perovskites, LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3), are shown in Fig. 2a. Well-defined peaks indicate the crystalline nature of the synthesized compounds without any impure phase. The characteristic peaks observed in these patterns are indexed according to an orthorhombic crystal structure having space group Pnma (62). The refined unit cell parameters are listed in Table 1. The refinement plot is shown in Fig. 2b. Ionic radii of trivalent Ti and Sc for octahedral coordination are 0.67 and 0.745 Å, respectively [32]. As shown in the inset in Fig. 2a, with increase of Sc concentration, slight shifts in the peak positions due to change in cell dimension can be observed. We observe a gradual decrease in lengths of a-axis and increase in b- and c-axes of the orthorhombic unit cell with increase in dopant concentration. We also observe a substantial decrease in peak intensities (as shown in the inset for the 221 plane) for doped perovskites. This may be due to the substitution of dopants among certain preferred crystallographic axes or planes and resulting octahedral tilting/deformation. The overall effect can be observed in the unit cell volume calculated from the experimentally obtained cell dimensions. Maximum cell volume is observed for the composition with x = 0.2. As observed from our results, this may be the solubility limit and doping Sc beyond this level may result in small amount of secondary phases which can be undetectable in XRD.

Fig. 2
figure 2

Powder XRD patterns of sintered LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3)

Full size image
Table 1 Refined lattice parameters, total conductivity and activation energies for the undoped and doped LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3) perovskites
Full size table

Scanning electron microscopy studies

SEM micrographs of the sintered pellets are shown in Fig. 3. Figure 3a shows the morphology of the undoped LaTiO3 + δ perovskite. Figure 3b–d shows micrographs of the doped LaTi(1 − x)Sc x O3 + δ for x = 0.2, 0.25 and 0.3, respectively. Figure 3e represents a representative lower-magnification image (for x = 0.2). A similar morphology was exhibited by all the other compositions. Micrographs reveal that the pellets are well sintered and have defined grains with distinct boundaries but with irregular morphology. The surface morphology of the samples indicates that ordered grains of unequal size are distributed throughout the sample with very small porosity. From the figure, it can be seen that the surface of each particle is composed of smaller particles. Extended contact to high temperature has resulted in agglomeration and fusion of the particles which appear as larger grains in microstructural images. No significant change in the morphology was observed due to different concentrations of doping.

Fig. 3
figure 3

SEM images of the sintered LaTiO3 + δ and LaTi(1 − x)Sc x O3 + δ . a x = 0, b x = 0.2, c x = 0.25, d x = 0.3 e Image corresponding to the composition with x = 0.2 at lower magnification

Full size image

X-ray photoelectron spectroscopy studies

The oxidation states of lanthanum, titanium and scandium in the perovskite LaTi(1 − x)Sc x O3 + δ with x = 0.2 were studied by using XPS. XPS spectra are shown in Fig. 4. Figure 4a represents La 3d state, Fig. 4b shows Ti 2p state, Fig. 4c shows Sc 2p state and Fig. 4d shows the O 1 s state of the perovskite. The binding energy of La 3d is 853 eV. Because of the shake-up effect, we observe two peaks in La 3d3/2 spectra from 853.3 to 849.5 eV. The presence of small satellite peaks in the spectra is due to the suborbitals in 3d orbital. From the data, it is clear that for La, the La 3d peaks that appear at 853.3–849.5 eV (3d3/2) correspond to La3+. It has been observed that the binding energy splitting between the components of both La 3d3/2 and La 3d5/2 lines is ~4.1 eV in La2Ti2O7, LaNbO4 and Ca4LaO(BO3)3 crystals, and this value is higher than ~3.5 eV in La2O3 [33–35]. The presence of satellite peak La 3d is due to the monopole excitation arising from a change in the screening of the valence electrons upon the removal of a core electron. Similar phenomenon was reported earlier by Miao et al. [36]. In our sample, the Ti 2p located at binding energies of 463.2 and 458.2 eV, respectively, indicates the presence of Ti3+. After sintering at high temperature, there is a possible transfomation tendency from Ti4+ to Ti3+ with a peak shift. This suggests that the Ti ions on the surface of the sample are trivalent. Similar studies for Ti 2p spectra of TiO4 and TiO7 were reported by Rao et al. [37]. Figure 4c represents the Sc 2p spectra. In the Sc 2p spectra, the satellite peaks observed on the high-binding energy side of the main peak by about 4.5 eV can be interpreted in terms Sc existing with the oxidation state 3+. The O 1 s spectra (Fig. 4d) shows the maximum binding energy at 529.82 eV. When lanthanum and titanium interact with oxygen, valence electrons are transferred from metals to oxygen atoms with variation in electrical screening of the inner shells. Thus, binding energies of the inner electrons of metal ions increase with synchronous decrease in the binding energy of O 1-level oxygen ions. The shift in binding energies is because of lanthanum- and titanium-bearing oxides. In the O 1 s spectra, the appearance of a low tail at higher binding energy confirms the oxygen deficiency of perovskite [38].

Fig. 4
figure 4

XPS spectra of the surface of sintered pellet LaTi(1 − x)Sc x O3 + δ (x = 0.2), depicting a La 3d, b Ti 2p, c Sc 2p and d O 1 s peaks

Full size image

Total conductivity measurements

The total charge transport properties were measured as a function of frequency and temperature using impedance spectroscopy over a wide range of frequency (50 Hz to 10 MHz) within the temperature range 30–650 °C. Figure 5 shows the Cole-Cole plot of LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3) at 650 °C. The experimental semicircle is related to ionic conduction in both bulk and grain boundary. Contribution from the electrode is not observed at lower frequencies. The value of resistivity can be determined from the intercepts of the semicircle on the x-axis. It can be observed from Fig. 5b that after doping Sc, the resistivity decreases when compared with the undoped perovskite Fig. 5a. The total DC conductivity is calculated using the following equation:

Fig. 5
figure 5

Cole-Cole plot of LaTi(1 − x)Sc x O3 + δ at 650 °C, a x = 0, b x = 0.2, c x = 0.25 and d x = 0.3

Full size image
$$ \sigma =\frac{l}{Rb\ A}\mathrm{S}\ {\mathrm{cm}}^{-1} $$
(1)

where R b is the resistance of the sample, l is the thickness of the sample and A is the area of the sample.

Figure 6 shows the total conductivity (log σ) as a function of inverse of absolute temperature (1,000/T). The nature of variation is almost linear in accordance with Arrhenius equation

Fig. 6
figure 6

Variation of log conductivity with respect to reciprocal temperature for LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3)

Full size image
$$ \sigma ={\sigma}_0 \exp \left(-{E}_{\mathrm{a}}/{K}_{\mathrm{B}} T\right) $$
(2)

where E a is the activation energy, T is the absolute temperature and σ 0 is the pre-exponential factor. It is obvious from the plot that the conductivity increases with the increase in temperature. This indicates negative temperature coefficient of resistance (NTCR) behaviour of the samples. The activation energy of the carriers for conduction process is calculated using Eq. (2). Activation energy and total conductivity at 650 °C for all the perovskites studied in this work are given in Table 1.

The diffusion coefficient of carriers in LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3) is calculated under different temperatures ranging from 520 to 650 °C. Figure 7 shows the variation of diffusion coefficient of carriers as a function of reciprocal temperature. The diffusion coefficient is calculated using the formula:

Fig. 7
figure 7

Variation of log diffusion coefficient (D) with respect to reciprocal temperature for LaTi(1 − x)Sc x O3 + δ (x = 0, 0.2, 0.25, 0.3)

Full size image
$$ D=\frac{\sigma KBT}{N{ e}^2} $$
(3)

where σ is the electrical conductivity, K B is Boltzman constant, T is the temperature, N is the number of atoms in an unit cell and e is the electronic charge [39]. From the figure, it is clear that after doping trivalent cations in the Ti site, there is an increase in diffusion coefficient with the increase in temperature which brings about enhanced mobility of the carriers. Consequently, it is assumed that at higher temperature, ionic mobility is higher. The inherent oxygen vacancies that can exist in unit cells due to various levels of doping and sintering conditions can also contribute to mobility and alter the diffusion coefficient.

Our results indicate that doping Sc3+ at Ti3+ site has altered the total conductivity of LaTiO3 + δ . It can be observed that conductivity attains a maximum for the composition x = 0.2 and, further, a decrease is observed for x = 0.25 and x = 0.3. The calculated unit cell volume from cell dimensions obtained through XRD analysis (Table 1) shows that the composition with x = 0.2 has the maximum value. The perovskite with the same composition also shows the maximum diffusion coefficient and highest total conductivity at all measured temperatures. From Table 1, it is evident that the perovskite with composition x = 0.2 has higher activation energy (0.62 eV) and exhibited the highest conductivity of 4.56 × 10−3 S/cm, which is due to the predominant ionic conductivity at higher temperature. Since the conductivity is temperature dependent, it is assumed that ions are responsible for the conductivity at higher temperature. This type of behaviour was reported by Stevenson et al., [40] in nonstoichiomertric perovskites. Based on XPS results (which conveys that both are trivalent), it is evident that doping Sc in the position of Ti does not create any additional oxygen vacancies, since after sintering at high temperature, Ti will no longer exist as tetravalent. Therefore, the possible reason for enhanced conductivity in the perovskite with composition x = 0.2 might be enhanced mobility of the carriers due to larger cell volume which involves localized hopping with either rotation or translation of the mobile species [7]. When compared with the conductivity data using other dopants than Sc at B-site of LaTiO3 − δ available in literature [41], Sc doping (up to x = 0.2) is more effective in enhancing the total conductivity at high temperature when compared to other dopants studied so far.

Conclusion

LaTi(1 − x)Sc x O3 − δ (x = 0, 0.2, 0.25, 0.3) perovskites were successfully synthesized using sol-gel method. XRD confirmed a single phase showing an orthorhombic structure. A similar microstructure with fine grain size was observed in all the perovskites. XPS data confirmed the presence of Sc in the material and the valence states for La3+, Ti3+ and Sc3+, related to oxygen-deficient perovskite structure. This investigation shows that among the studied compositions, doping Sc on LaTiO3 + δ perovskites can successfully improve the total conductivity only up to a certain dopant concentration (x = 0.2), beyond which a reverse effect occurs.