Effect of mono-dopants (Mg2+) and co-dopants (Mg2+, Zr4+) on the dielectric, ferroelectric and optical properties of BaTiO3 ceramics

In this work, BaTiO3, Ba(Mg0.01Ti0.99)O3, Ba(Mg0.015Ti0.985)O3, Ba(Mg0.02Ti0.98)O3 and Ba(Mg0.01Zr0.15Ti0.84)O3 ceramics have been prepared through conventional solid-state route to investigate the effects of Mg2+ and Zr4+ dopants as mono-substitution (only Mg2+) and co-substitution (Mg2+ and Zr4+) of B-site on the structural, electrical and optical properties of BaTiO3 ceramics. Exhibiting perovskite structure, Ba(MgxTi1−x)O3 ceramics revealed a decrement pattern of tetragonality with the increment of the concentration of MgO which was confirmed through Rietveld analysis. Morphological analysis of the sintered samples by scanning electron microscope showed a grain growth retardation phenomenon with Mg2+ addition. Releasing from this retardation process, Ba(Mg0.01Zr0.15Ti0.84)O3 showed a maximum dielectric constant of ∼1269.94 due to the enhanced domain wall motion and the confinement within the solubility limit of Mg2+. The ferroelectric characteristic of Ba(MgxTi1−x)O3 was sluggish due to the effects of grain size and its boundary. The optical band gap for BaTiO3 was found to be decreased from 3.55 eV to 3.06 eV with the addition Mg2+ content but for Ba(Mg0.01Zr0.15Ti0.84)O3, the value increased due to the Burstein-Moss effect. Again the FTIR analysis proved that no impurity phases were formed during the doping phenomenon, but in Ba(MgxTi1-x)O3 ceramics, a significant reduction of Ti-O bond strength was observed. However, BaTiO3, Ba(Mg0.01Ti0.99)O3, Ba(Mg0.015Ti0.985)O3 and Ba(Mg0.02Ti0.98)O3 ceramics had manifested P-E loop having lower remanent polarization and coercive field compared to Ba(Mg0.01Zr0.15Ti0.84)O3 ceramics with moderate electrical and optical properties. So, co-doping with Mg2+ and Zr4+ evidenced a favorable accession for the increment of the properties of BaTiO3 ceramics.

Just after the first advancement of MLCCs, Ag-Pd was used as an internal electrode [3]. However, a few years later, their usages were minimized due to some costing related issues. Afterward, the Ni electrode was thought to be used as a replacement of the previous electrode [4]. But, during the propagation of Ni electrode into MLCCs through the sintering mode under a partial pressure of oxygen gas (P O 2 > 10 −10 MPa), reduction of BaTiO 3  Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. evolution of oxygen vacancies with the n-type carries were resulted in accordance with the following reaction [5]: Where, V o ·· prevails an oxygen vacancy into BaTiO 3 lattice as derived by Krӧger Vink notation [6]. So in order to prevent the afore-mentioned unexpected situation, the structural compositions could be manipulated through the addition of acceptor additives, i.e. Mg 2+ , Al 3+ , Mn 3+ , Fe 3+ and Co 2+ in a radical range of 0.54 Å to 0.83 Å [4]. Amongst the acceptor elements, MgO is considered as the most efficient additive for attaining nondeductible compositions in MLCCs with Ni electrodes [3]. Several research works were already performed to dig up the effects of MgO addition on the microstructural and the dielectric properties of BaTiO 3 ceramics [7,8]. Whither, S H Yoon observed a special feature that the accession of Mg 2+ dominates the rate of grain growth and simultaneously reduces the grain size of BT ceramics [9].
Apart from that, among the doped BaTiO 3 systems, Ba(Zr x Ti 1−x )O 3 ceramics have recently received attention due to their high strain level and piezoelectric effect in both single crystals and polycrystalline ceramics [10]. From the research of Zhi Yu, it is cleared that a distinct phase transition is triggered with Zr enrichment, whereas a merged broad peak corresponds to a three-phase transition [11]. Again, as the replacement of titanium with zirconium takes place, the transition of the polymorphous phase is reconfigured to higher temperatures and also arises the degeneration of the Curie temperature [12]. Besides, regarding higher doping levels (more than 10 mol%), the commencement of the relaxor behavior is obtained which is broadly investigated by C Ciomaga et al [1]. Whereas, Jean Ravez claimed a ferroelectric-relaxor behavior at x>0.25 [13].
However, following the previous researches, we tried to perform further exploration on BT ceramics by adopting variable Mg component (0 mol%, 1 mol%, 1.5 mol%, and 2 mol%) as mono substitution dopants and 15 mol% Zr with 1 mol% Mg as co-dopants, where co-doping (Mg 2+ +Zr 4+ ) supposed to be a new approach which perpetuates the dielectric and ferroelectric behavior of the corresponding ceramics. Moreover, the results were analyzed in terms of crystal anisotropy correlated with the impression of grain size. Awfully sterling raw materials including BaCO 3 (purity>99%, Merck Specialties, India), MgO (purity>99%, Merck Specialties, India), ZrO 2 (purity>99%, SRL, India) and TiO 2 (purity>99%, Merck Specialties, India) were used with appropriate stoichiometric ratios. The weighed powders were ball-milled using ethanol (purity>99%, Merck, Germany) as milling media with yttria-stabilized alumina balls for 20 h. After milling, the slurry was dried in an oven at 100°C for 24 h. Followed by the incineration at 900°C for 2 h, the incinerated powder was re-ground, compounded with a binder (2 wt% polyvinyl alcohol) and compacted into disk pellets under an axial pressure of 2.5 tons. The prepared pellets were then sintered at 1250°C for 4 h. X-ray diffraction (XRD) measurement was carried out at room temperature for phase detection using 40 kV-40 mA (scanning step of 0.01°and counting time of 1 s per step) and Cu-K α radiation of wavelength K α1 =1.54060 Å and K α2 =1.54439 Å (Panalytical Empyrean, Netherlands) in the range of 10-80°. Structural refinement was redacted using a standard refinement program 'Full Prof'. Scanning Electron Microscope (SEM) (JEOL JSM-6510, Netherlands) was devoted to observing the morphologies of the swatches. The average grain size was explored by the linear intercept method and the grain size distribution was attained via ImageJ and OriginPro software. Dielectric measurements were effectuated at 500 mV over the frequency range of 100 Hz to 3.17 MHz through Impedance Analyzer (Wayne Kerr 6500B series, UK). Ferroelectric polarization-electric field (P-E) hysteresis loop was obtained using Multiferroic Tester (Radiant Tech. Inc., USA) retaining the reliability on voltage (200 V-1 kV at 1 Hz). The optical band gap energy was assessed using UV-vis spectroscopy (SHIMADZU UV/Vis-1650 PC, Japan) over the range of 200-800 nm. Fourier Transform Infrared spectroscopic measurements were accomplished by using the FTIR spectrometer (JASCO FTIR 6100, Japan) in the wavenumber range of 400-4000 cm −1 .
Moreover, incorporation of Zr 4+ into BaTiO 3 ceramics usually follows three mechanisms to attain structural stability as revealed through the leading equations [16,17]: Consideration for network modifier: . 4 x O . 5 Consideration for network formers: . 7 x O . 8 ] are featuring both characteristics equally. However, the stabilization of defect mechanisms is prompted by the presence of these clusters. Cavalcante and Gurgel believed that the vacancies could be eliminated by the loss of energy of electrons residing into the conduction band and the re-acquirement of the hole in the valence band [16]. Figure 2 shows the final outputs of the Rietveld refinement programs, confirming that the configurations belong to tetragonal symmetry with space group p4mm which were carried out using 'FullProf' and 'Maud: Materials Analysis Using Diffraction' software by adopting Wyckoff's series [18]. The refined lattice parameters and reliability factors (R factors) derived from Rietveld analysis are enlisted in table 1. In consideration of co-doping, an expanded cell volume results in due to the replacement of lower radii Ti 4+ (0.605 Å) with higher radii Zr 4+ (0.720 Å) and Mg 2+ (0.605 Å) (table 1) [19]. However, a minor variation in tetragonality takes place as the axial ratio (c/a) alters in the compositional sequence [20].
Debye-Scherrer formula [21] was used to measure the crystallite size (table 2) for the most vivid peak (110). The formula can be expressed as: Hither, β is the Full Width at Half Maxima, k being a dimensionless shape factor having a constant value of 0.94 associated with the crystal symmetry, while λ and ϴ are the wavelengths of Cu K α radiation (1.54 Å) and Bragg angle respectively.
Through the x-ray diffraction analysis, the degree of crystallinity (x c ) can be usually derived by two-phase model according to the following formulae: Where, I crystalline and I amorphous belong to the area of the crystalline and the amorphous peaks respectively. However, BMZT1 exhibits lower x c as compared to BT and BMT ceramics which are enlisted into table 2.
The structural transformation could be evaluated by using R D Shannon's ionic radii table [22] and Goldschmidt's rule [23]. Following the formula, the tolerance factor (t) could be computed by, and O [CN=6] ion correspondingly. The decrement pattern of t (table 2) implies that BaTiO 3 doped with MgO and ZrO 2 is attaining structural transformation as well as more stability in the perovskite structure [24]. The value for BMT varies between 1.333495 to 1.332535 inducing the reduction of the tetragonal phase and approaching a cubic phase which will be again evidenced as the discretion of c/a ratio. Besides, the obtained values of the tolerance factor (t>1) reflect the presence of the ferroelectric phase at room temperature [25].
The resulted lattice strain arisen in the sample was obtained using Williamson-Hall equation An enhancement of the tensile strain is emerged with the accession of Mg 2+ and Zr 4+ due to the lattice alteration sourced from the substitution of Ti 4+ ions by larger Mg 2+ and Zr 4+ ions. Figures 4(a)-(e) displays the SEM micrographs of the empirical formulas of BaTiO 3 ceramics doped with Mg and Zr. The average grain size was enumerated by using the linear intersection method [28]. At a lower proportion of MgO (1.0 mol%), inhabiting in the solubility range of MgO in BT, the microstructure shows aesthetic conformation, with an average grain size of 811.37 nm [4]. But a significant reduction in grain size is observed when the MgO content is above 1.0 mol% (table 2) [14]. This result reveals that the grain growth of BT ceramics is suppressed by Mg which designates it as the grain growth inhibitor [29]. It is due to the fact that the incorporation of Mg 2+ ions persuades the formation of oxygen vacancy (V O ·· ). Simultaneously, the generated V O ·· introduces a motion into the crystal lattices which attempts to consume some energy. However, the consumption of the energy is benefited as the segregation of the solutes (Mg 2+ ) is manifested in the grain boundary. Positioning themselves into the grain boundary, Mg 2+ ions counteract the motion of the boundary resulting the suppression of the grain growth. Besides, in consideration of Zr embodiment, a symbolical enhancement in the grain size is observed accompanied by more effectiveness in the densification mechanisms as shown in figure 4(d) [30]. This can be caused by the accelerated matter transport mechanism which is generally materialized at the interim locus of the grains during the densification course [17]. A narrative scatter diagram associated with the distribution contour of the grain size distribution (appraised by exploiting ImageJ software and OriginPro 2018 software) are stereotyped in figures 4(a)-(d) [31]. Furthermore, from the quantitative Energy-dispersive x-ray spectroscopy (EDS), the weight percentage of the elements in the samples were computed (table 3). The curves as stereotyped in the inset of figure 5 reveal the elemental provinces of the perovskite phases of un-doped and doped BT ceramics. However, the tabulated values ensure the absence of impure phases in the experimental specimens and also dispel the burden of losing any ingredients during its formulation [32,33].  [34]. But proceeding at high frequencies, the dipoles show incapability to provide sufficient response to the imposed field resulting in degradable dielectric constant. The observed dielectric behavior at the low frequencies due to the oxygen vacancies can be coordinated with the Maxwell-Wagner interfacial polarization model with Koop's phenomenological theory. According to the hypothesis, the dielectric structure comprises two layers, i.e. layers of immensely conducting grains and layers of flimsy conducting grain boundaries, where the grain boundaries become more enterprise than the grains at a lower frequency and only the grains withhold their distinctiveness at the higher frequencies [35].

Dielectric properties
Besides, the acquired data confirms a strong relationship between κ and dopant %. Enhanced dielectricity is observed at 1 mol% MgO, while further increment of MgO demonstrates a declined κ value [14]. It is backed by the dominance of grain size and grain boundary permittivity. Generally, Grain boundary comprises space charges which exclude polarization charges from the grain surface and creates a depletion layer on it. However, the layer introduces a depolarization field which lowers the polarization value simultaneously. Whereas, in consideration of the ZrO 2 accession into BT, a maximal value of the relative permittivity is demonstrated by the emergence of the largest grain size as well as the dominance of maximum tetragonality [36].   into table 4. Actually, there involves two mechanisms contributing to the loss factor, i.e. resistive loss and relaxation loss. For the consideration of resistive loss, a certain quantity of energy is devoured by the whirling charge bearers, while for later one, the contribution results from the relaxation of electric doublets. However, the dependence of the dielectric loss on frequency turns into nonpartisan at the higher frequency range [37].   Figure 7 represents the characteristic polarization versus electric field (P-E) loops of BT, BMT1, BMT2, BMT3, and BMZT1 ceramics. The arisen of interruption at 0 V coupling with the generation of an unsaturated hysteresis loop express the appearance of leakage current into the samples. Actually, P-E loops describe a total polarization value comprising of remanent and non-remanent polarization. The remanent polarization is concerned with switchable dipoles which retain their polarization in the removal of the applied field. While for the non-remanent polarization, it is mainly influenced by the dielectric linear capacitance and its loss factor accompanied by the non-retainable polarized dipoles. The obtained remanent polarization (P r ) and the coercive field (E c ) are enlisted in table 4. However, the maximal value of remanent polarization (P r ) of 0.21 μC/cm 2 with a field of maximum coercive value (E c ) of 2.10 kV/cm were found for BMZT solid solution. It is caused by the presence of Zr 4+ which influences to generate the space charges and in the meantime preserve the domain configuration resulting in the higher poling efficiency as well as enhanced remanent polarization value. Moreover, the graphs (figures 7(a)-(c)) for BMT exhibit a decreased pattern of P r with the increment of the mol% of Mg. Actually, reduction of poling proficiency can be usually caused by the effects of grain size, grain boundary, internal stress, depolarizing field and also the mutual interaction between the domain walls and their clamping characteristics. Again, the larger area of the grain boundaries results in enhanced electrical insulation associated with a haphazard arrangement of space charges that interrupts and lowers the polarizability.

Optical property
The UV-Visible spectra of BT, BMT1, BMT2, BMT3, and BMZT1 ceramics are interpreted in figure 8(a). The optical band gap energy was computed via allowed direct inter-band transition between valence and conduction bands by adopting Tauc's Law [38]: Where, A is a constant, α being the absorption coefficient, hυ stands for the photon energy and E g represents the optical band gap energy. E g can be attained from the (αhυ) 2 versus hυ plot and the extrapolation of the graphical segment of the trajectory to (αhυ) 2 =0 as shown in the inset of figure 8(b). With proceeding from BMT1 to BMT3, the band gap significantly reduces from 3.18 to 3.06 eV while for BT, the value is 3.55 eV (table 4). Generally, the obtained E g values are strongly dominated by the appearance of ordered or disordered patterns into the lattice structure. The disordered arrangement is mainly caused due to the presence of asymmetry into the O-Ti-O bond or due to the misrepresentation of the TiO 6 clusters. Again, another contributing factor is the  defect mechanisms into the structure [39,40]. However, the inauguration of lattice defects into the crystal structure allows for the establishment of certain intermediate energy levels (comprising of oxygen 2p, titanium 3d, and magnesium 3p orbitals) between the valence band and the conduction band (as shown in figure 9) which reduce the band gap energy significantly. Moreover, as the amount of Mg increases, the corresponding defect mechanism enhances which triggers the shifting of E g to a lower extent for BMT ceramics.
On the other hand, in consideration of Zr 4+ addition, the E g shifts to a higher extent value (3.71 eV) as compared to BMT. Whenever these materials are heavily doped, the electrons seize themselves into the lowest   potency level adjacent to the conduction band which is ordinarily commenced from the defect mechanisms i.e. oxygen vacancy and this phenomenon is considered as Burstein-Moss effect [41]. Another contributing factor known as crystallite size also affects the optical band gap energy significantly. Whereas, reduced band gap results from the cause of increment in the crystallite size of BT ceramics doped with Mg 2+ and Zr 4+ as derived from the XRD analysis [42]. Figure 10 illustrates the room temperature transmittance spectra of BT, BMT1, BMT2, BMT3, and BMZT1 ceramics. All of the samples (except BMT1) exhibit a broad absorption band at 3250 cm −1 due to the stretching vibration of water molecules [43]. However, the characteristic band at 2452 cm −1 is employed for the asymmetric stretching of CO 2 [44]. Moreover, the absorption band at 1630 cm −1 appears due to the bending vibration of H 2 0 molecules [45]. The band at 585 cm −1 is caused due to the stretching vibration of TiO 6 with barium, while for the bending vibration of Ti-O bond in [TiO 6 ] 2octahedron, the band at 475 cm −1 has appeared. In addition, for BMT1, BMT2, BMT3 and BMZT1 ceramics, the maximum absorption occurs at the band of 578 cm −1 , 576 cm −1 , 571 cm −1 , and 583 cm −1 respectively. It is observed that as the incorporation rate of MgO increases, the absorption band of BMT ceramics shifts to a lower extent. It is mainly caused due to the weakening of the Ti-O bond [46] and this weakening behavior is mainly caused by two reasons. The first one involves the replacement of Ti 4+ ions with Mg 2+ and Zr 4+ ions having higher ionic radius than that of the replaced ions and the second one involves the formation of oxygen vacancies (V O ·· ) for retaining the electric charge balance into the crystal lattice [47]. Whereas, for BMZT1, the absorption band increased due to the higher stabilization mechanism of Zr 4+ as well as having a lower concentration of Mg 2+ .

Conclusion
In summary, BT doped with 0, 1, 1.5, 2 mol% Mg and co-doped with 1 mol% Mg and 15 mol% Zr ceramics were successfully fabricated by conventional solid-state reaction method. A single perovskite tetragonal phase was found in all the ceramic specimens without the evidence of any additional phase which was simultaneously proved by XRD and FTIR analysis. An occurrence of retardation in grain growth of BMT ceramics was exhibited whenever the MgO content is >1 mol%. The dielectric behaviors were predominantly attributed to the existence of oxygen vacancies, experiencing the extensive motion of V O ·· accompanied by the pinning phenomena derived by  Mg .
Ti However, due to the effect of larger grain size, BMZT1 exhibits enhanced relative permittivity value along with the appearance of higher remanent polarization as compared with other synthesized samples. However, the obtained optical band gap energy of the samples increases its probability to use in optoelectronic purposes.