Understanding the structural, optical, and dielectric characteristics of SrLaLiTe1−xMnxO6 perovskites

In electronic applications, good dielectric permittivity material has huge potential in the capacitive energy storage devices. Herein, in the present work the dielectric study of SrLaLiTe1−xMnxO6 (x = 0.02, 0.04, 0.06, 0.08, and 0.10) double perovskites has been studied and discussed. These compounds were prepared through solid-state reaction method. All of the prepared compounds were confirmed to crystallized in monoclinic structure of P21/n space symmetry with better crystallization when dopant concentrations increased until x = 0.08. The formation of Li–O–Te/Mn bonds in octahedral structures in all compounds were confirmed in this study. The existence of peaks at specific wavenumbers indicated vibrations of B–site cations’ bonds. When dopant amounts were increased from x = 0.02 to x = 0.08, there was an increasing trend of grains sizes formation in the compounds. The discussions on effects of grain sizes towards dielectric properties were included in this paper. Other important results and discussions comprised of the significant effects of dopant on the optical band gap (Eopt) and absorption frequencies of the compounds. The decreasing trend of Eopt towards semiconductor range indicated the compounds’ promising potentials for optoelectronic device application.


Materials and methods
The SrLaLiTe 1−x Mn x O 6 (x = 0.02, 0.04, 0.06, 0.08, and 0.10) compounds were prepared by solid-state reaction method 21 . High purity chemicals (≥ 99.99%) of strontium carbonate (SrCO 3 ), lanthanum oxide (La 2 O 3 ), lithium carbonate (Li 2 CO 3 ) , tellurium dioxide (TeO 2 ), and manganese (III) oxide (Mn 2 O 3 ) powders were purchased from Sigma-Aldrich. The chemicals were mixed at required stoichiometric ratios before grinded with pestle and agate mortar in 1 h duration to attain homogeneity. The mixed powders were calcinated in air using a CWF 11/5 furnace (Carbolite Gero, UK) with heating rate of 15 °C/min until max. temperature of 850 °C in 10 h period. Then, slow cooling rate of 1 °C/min was done in order to ensure the formed stoichiometry match with the desired oxygen stoichiometry 27 . The samples then were pressed to form pellet at pressure of 5 T by using Atlas 15 T hydraulic press (Specac, UK) before sintered in air at max. temperature of 850 °C in 10 h duration. The phase(s) and purity of the samples were investigated by the x-ray diffraction (XRD) characterization within the angle range of 10° to 80° using Xpert PRO MPD diffractometer (PANanalytical, Netherland). The instrument equipped with a Cu K α radiation which has 1.5418 Å of wavelength. The graphical user interface (EXPGUI) and visualisation for electronic structural analysis (VESTA) software were deployed for structural refinement and to visualize the refined structure of compounds, respectively 18,[28][29][30] . The infrared absorption spectra of the samples were collected by Fourier transform infrared (FTIR) characterization within the wavenumber range of 400 to 1500 cm −1 using Drift Nicolet 6700 spectrometer (Thermo Fisher, USA). The morphologies and constituents in the samples were determined by field emission scanning electron microscope (FESEM) and energy dispersive x-ray (EDX) characterizations, which were conducted using SU 8000 FESEM (Hitachi, Japan). The dielectric characteristics of the samples were obtained by electrochemical impedance spectroscopy (EIS) characterization within the frequency range of 50 Hz to 1 MHz and temperature range of 298 K to 343 K using LCR 3532-50 HiTester analyzer (Hioki, Japan) while keeping the compounds and electrodes in sandwich geometry. The optical spectra of the samples were investigated by ultraviolet-visible light (UV-vis) characterization within the wavelength range of 200 to 1000 nm using Lambda 750 spectrometer (Perkin Elmer, USA). The densities of each samples were calculated by applying the Archimedes' formula. Figure 1 shows the refined XRD data of SrLaLiTe 1−x Mn x O 6 (x = 0.02 (Mn 0.02), x = 0.04 (Mn 0.04), x = 0.06 (Mn 0.06), x = 0.08 (Mn 0.08), and x = 0.10 (Mn 0.10)) by the Rietveld refinement method. From these plots, all compounds were formed in double phase, with minor peaks present corresponded to Sr 7 Mn 4 O 12 with percentage of less than 10%. The obtained reliabilities (χ 2 ) were 2.178, 1.903, 1.647, 2.419, and 1.948 for compounds with x = 0.02, 0.04, 0.06, 0.08, and 0.10, respectively which showed good reliability of results. All prepared compounds were crystallized in monoclinic structure of P2 1 /n symmetry. The refined lattice parameters of a, b, and c in all compounds were in 5.57-5.63 Ǻ, 5.58-5.61 Ǻ, and 7.1-7.92 Ǻ ranges, respectively. All compounds possessed α = 90° and γ = 90°, whereas ß = 90.37°, 90.25°, 89.93°, 90.05°, and 90.10° for x = 0.02, 0.04, 0.06, 0.08, and 0.10, respectively. The obtained unit cell volumes (V) for all compounds were 248.7, 247.7, 248.5, 249.5, and 249.9 Å 3 , respectively. The trend was not in agreement with the doping of smaller size of Mn 6+ into larger Te 6+ cations. Table 1 shows the complete parameters that were obtained from the refinement. Figure 2 shows refined structure of SrLaLiTe 1−x Mn x O 6 from bc plane. This figure revealed presence of doped Mn 6+ into Te 6+ alongside Li + at B-site octahedral structure. Mn 6+ /Te 6+ and Li + alternately positioned between each other and being surrounded by six O 2− atoms in each octahedron. A-site cations (Sr 2+ and La 3+ ) placed between octahedral to fill up spaces in the structure. In all compounds, Li + was located at (0.5, 0, 0) and Te 6+ was detected at (0, 0.5, 0) coordinates. The tolerance factor (τ) of the compounds was calculated by the equation below 21 :   19,[35][36][37][38] . It is apparent that as dopant concentration increased, there were more quenches of peaks. This is understandable since the Te-O-Li should be reduced when the doping Mn 6+ into Te 6+ took place to form Mn-O-Li bonds. Nonetheless, peak at 718 cm −1 in Mn 0.02 can be suggested due to introduction of Mn 6+ to form Mn-O bonds with none of peak present at the same wavenumber for pristine SrLaLiTeO 6 39 . Besides, this peak started to shift to lower wavenumbers indicates the increment of abundance of Mn-O bonds as dopant increased with Mn 0.10 depicts the Mn-O bonds vibration peak at 685 cm −1 . Reports have claimed that the Mn-O bonds with inclusion of Mn 3+ / Mn 4+ showed vibration peaks at 600 cm −1 or above 40,41 . Since Mn 6+ has smaller ionic radius compared to Te 6+ and Mn 3+ / Mn 4+ , hence, it is possible for Mn 6+ that form Mn-O bonds to exhibit vibration peaks at higher wavenumbers. Meanwhile, most of the existing peaks redshifted as dopant content increases implying the increase of bond length. These results are in accordance with the Rietveld refinement and tolerance factor. Summary of the peaks obtained were tabulated in Table 2. Table 1. Obtained space symmetry, parameters of lattice, V, angles of bonds, lengths of bond, and fit goodness from Rietveld refinement with calculated ϕ, τ, and D in SrLaLiTe 1−x Mn x O 6 .  (4) 1.935 (7) Mn-O 1 (× 2) 1.934 (5) 1.933 (4) 1.935(4) 1.937 (4) 1.940 (7) Mn-O 2 (× 2) 1.936 (5) 1.935 (4) 1.934 (3) 1.937 (3) 1.941 (7) Mn-O 3 (× 2) 1.924(10) 1.920 (8) 1.921 (6) 1.924 (5) 1.925 (8) Aver. < Mn-O > 1.931 (7) 1.929 (5) 1.930(4) 1.933 (4) 1.935 (7) Aver. < Sr-Sr > 3.962 (19)  www.nature.com/scientificreports/   www.nature.com/scientificreports/ when Mn contents increased and lead to grain size enhancement. EDX graph in Figure 4f exhibits constituted elements in Mn 0.08. This graph confirms that this compounds contains elements of the prepared raw material composition, except Au which originate from coating in sample preparation for FESEM characterization. Other compounds show almost same trend of EDX graph as in Figure 4f. As calculated by Archimedes method, the density of all compounds had difference of less than 4%. The densest compound was Mn 0.06 with 5.450 g cm −3 despite Mn 0.08 having the largest grain sizes. This occurrence most likely because of less porosity in Mn 0.06 compared to all samples. Figure 5a exhibits the UV-vis reflectance spectra versus wavelength. Compared to pristine SrLaLiTeO 6 43 , there are strong absorption bands that were observed at the whole 300-800 nm range. It was probably because of the charge transfer between 3d orbital of Mn 6+ and 2p orbital of O 2− . As Mn 6+ dopant increases, the reduction of reflectance value of compound indicates the increase of the absorbance character. Spectra for the compounds show the changes in shoulder in the spectrum just below 800 nm in comparison to pristine SrLaLiTeO 6 43 , suggesting the E opt of the doped compounds are dramatically altered. When Mn 6+ doped into Te 6+ site, Mn band could be formed within the gap. At low concentration of dopant, the band gap width is not hugely affected. With increasing Mn 6+ concentration, the widths of Mn bands could increase and the overlapping of the bands could happen. Hence, implies the changes in the reflectance spectra 44 . Figure 5b shows the absorption UV-vis of SrLaLiTe 1−x Mn x O 6 plotted from Kubelka-Munk equation 21 ,  where λ is the absorption edge's wavelength. Figure 5c shows the graph plotted by using the Tauc equation:

Results and discussions
where hv is photon energy, A is proportional constant, and E opt is gap energy. Meanwhile, the values of n can vary based on the of transition type in a material where n = 1/2 indicate the direct with allowed transition, n = 2 indicate the indirect with allowed transition, n = 3/2 indicate the direct with forbidden transition, and n = 3 indicate the indirect with forbidden transition 21 . Figure 5b shows the absorbance cut-off wavelength increases with increasing Mn 6+ content. Meanwhile, Fig. 5c shows fitting of n = 1/2 is the most fitted to the graph and hence, implies that only allowed and direct shift of electrons took place from highest occupied molecular orbital (HOMO) to lowest unoccupied molecular orbitals (LUMO) in these compounds where only photon absorption occurred. The comparable values of E opt obtained from both graphs are tabulated in Table 3.   44,45 has reported that E opt is correlated to the existence of energy levels of impurities within the E opt of materials. As wider Mn energy bands (impurity bands) formed with higher dopant concentration, this change reduced the distance between bands and consequently reduced the E opt . On the other hand, reduction in structural distortion when Mn 6+ doped into pristine compound is the other reason to reduce E opt . Opposite explanation regarding effect of distortion onto band gap has been discussed in other report 21 . Besides, oxygen vacancies were the other possible factor which could promote the formation of energy levels of impurities within the E opt 45 in all compounds. It is accepted that oxygen vacancies can affect the optical result. Nonetheless, lattice parameters as well as monoclinic structure in every compounds were close to the ones in SrLaLiTeO 6 on previous report 21 . It is suggested that there were minor differences of oxygen content among the compounds. Therefore, oxygen vacancies should not show major effects towards the optical properties. There is other report that revealed the same trend of E opt in other compound with doping of Mn 2+ realized 46 . The absorbance spectra in Fig. 5a shows that all compounds' optical band gaps are in visible light range, makes SrLaLiTe 1−x Mn x O 6 possible for photovoltaic applications with further study to increase the electrons or holes mobility. Figure 6a exhibits the plot of ε ′ with respect to frequencies from 50 Hz to 1 MHz in SrLaLiTe 1−x Mn x O 6 (x = 0.02, 0.04, 0.06, 0.08, and 0.10) compounds in room temperature. The variation of ε ′ with frequency showed almost the same trend for each compound, with respect to frequencies below or above 100 Hz. For frequencies below 100 Hz, all compounds showed almost similar variation of with frequency where ε ′ dropped instantaneously when frequency increased. At frequencies above 100 Hz, the decrease of ε ′ was moderately and steadily to reach the minimum value at 1 MHz. Compound Mn 0.06 showed the highest ε ′ values from 50 Hz to 700 kHz. However, started from 300 kHz, the gradient of decrease of ε ′ in Mn 0.08 was the most minimum among all compounds before showed the highest value of ε ′ at 1 MHz. Figure 6a (inset) shows changes of at 250 kHz, 500 kHz, and 750 kHz frequencies together with average size of grains for all compounds. It is clear that aside of ε ′ in Mn 0.02, the increasing trend of ε ′ observed with the peak value shown by Mn 0.08 before decrease significantly at higher doping concentrations. This ε ′ peak most probably interrelated with peak of grain size of Mn 0.08. Figure 6b exhibits the dielectric loss, Tan δ variation in SrLaLiTe 1−x Mn x O 6 (x = 0.02, 0.04, 0.06, 0.08, and 0.10) with respect of frequencies in room temperature. This figure shows drop of Tan δ in all compounds at frequencies less than 500 Hz before relatively flat pattern at frequencies above 500 Hz. Compounds with different dopant concentrations exhibit almost the same increasing pattern from 10 kHz to 1 MHz. Figure 6b (inset) shows Tan δ at 300 kHz and above, where all compounds registered increasing Tan δ values at this range of frequencies.
The real part of dielectric constant, ε ′ illustrates the electrical dipoles' aligning ability in compounds with external electric field. The high values of ε ′ at minimum frequencies in all compounds from Fig. 6a were suggested because of the space charge polarization that primarily consists of the accumulated heavy electrical dipoles at grain/grain boundaries interfaces when electrical field alternation existed. Besides, the minor possibility for oxygen vacancies existence could reduce the phonon modes in the compounds' structure, hence, elicited the space charge polarization 47 . The decrease of ε ′ at low frequencies in all compounds was because of these heavy dipoles were impotent to move with the alternation of external field when frequencies were increased. ε ′ drop in low frequencies was in alignment with high losses in Tan δ at the same frequencies range as in Fig. 6b.
Tan δ illustrates the energy loss in the compounds in the midst of electric field alternation. The presence of peak at minimum frequency in Fig. 6b could be due to the impotency of heavy electrical dipoles to move with the external electric field or DC conduction loss. The primary reason for the loss could be the inability of electrical dipoles to move with the field changes as frequency increased. Another plausible reason is the charge carriers' motions across the grain boundaries. Any minor possible formation of singly/doubly ionised oxygen vacancies that were due to the evaporation of lithium in the midst of sintering could produce charge carriers 48 by the following reaction: www.nature.com/scientificreports/ where V ′′ o is double ionised oxygen vacancies. Hence, DC conduction loss might be another factor which assisted the drop of the ε ′ . However, the gradient in plot of ln ε ′ against ln ω (not shown) did not show magnitude close to (-1) to prove the presence of this factor. Meanwhile, the increase of Tan δ starting from 300 kHz could indicate the presence of relaxation peak at higher frequencies than 1 MHz.
Apparently, there are clear effect of doping towards ε ′ values from low to higher frequencies with heavy, medium and light-sized electrical dipoles contributed in ε ′ response at low, medium, and high frequencies, respectively. The ε ′ values keep decreased as frequencies increased, and ε ′ values between all compounds are slightly differ from each other, most likely due to small difference of density between the compounds. The highest density of Mn 0.06 could be the reason of its highest value of ε ′ from 50 Hz to 650 kHz compared to the other compounds 19 . Nonetheless, the lowest decrease of ε ′ variation in Mn 0.08 in medium frequencies range and its highest ε ′ value at 1 MHz compared to other compounds can be pertained for having larger medium-sized dipoles with higher abundance, which is related to smallest Tan δ in the compound at frequencies above 400 kHz as in inset of Fig. 6b. The variation of ε ′ response can be explained in terms of its interrelation with average grain size in inset of Fig. 6a. Larger grains size can form larger medium-sized dipoles and probability of higher amount of medium-sized dipoles. Thus, production of larger electrical dipole moments with larger polarization effect could took place. This suit to explain the relationship between variation in ε ′ in medium frequencies and average grain size. This factor might be the reason of Mn 0.08 possessed highest ε ′ at 750 kHz and above. Other reports that were related to this interrelation between ε ′ and grain size has been recorded before 49 . In comparison to the pristine SrLaLiTeO 6 43 , there is enhancement of ε ′ values in all Mn 6+ doped SrLaLiTeO 6 compounds. From the Eq. (8) below: where ε o is the vacuum permittivity, A is the area of contact between compounds and electrode, and d is thickness of compounds, the calculated values of capacitance at 1 MHz in room temperature for Mn 0.02, Mn 0.04, Mn 0.06, Mn 0.08, and Mn 0.10 were 7.77 nF, 9.4 nF, 9.15 nF, 11.2 nF, and 9.7 nF, respectively. These results clearly showed the highest capacitance obtained by Mn 0.08 at 1 MHz, in accordance with the ε ′ trend. www.nature.com/scientificreports/ Figure 7a illustrates variation of ε ′ value as function of temperature in the range of 298 K to 343 K of Mn 0.08 sample. All other samples exhibit almost identical trend of plot as in Mn 0.08 (not shown). At minimum frequency, the ε ′ values for each temperature enhanced as temperature elevated from 298 to 313 K. In higher temperatures from 323 to 343 K, the values of ε ′ at minimum frequency dropped. Meanwhile, ε ′ variation with respect to frequency in 298 K exhibit steady decreasing trend at medium and high frequencies. Generally, values of ε ′ enhanced in these frequency ranges as temperatures elevated. Figure 7b illustrates the variation of ε ′ against temperature for sample of Mn 0.08 at selected frequencies. Apparently, there were more dispersion of ε ′ with temperature as temperature elevated. All frequencies, except at 1 MHz, exhibited non-linearly dependence of ε ′ response against temperature. For each frequency, the decrease of ε ′ value occurred in different temperatures. Each decrease shifted towards the higher temperature as frequencies increased. On the other hand, the graph increased almost linearly with temperatures at 1 MHz.
The small ε ′ values in low frequencies range in 323-343 K in Fig. 7a can be interpretated as inability of electrical dipoles to follow the external field efficiently which most probably due to the scattering of the dipoles when external heat energy added. Thermal activation started to show effect at 1 kHz and increase the ε ′ values where the medium and light dipoles were able to cope with external field ascribed to external heat energy which aided the electrical dipoles to follow external field. At frequencies higher than 100 kHz, these polarizations failed to effectively follow the external electrical field. The increasing trends and enhanced values of ε ′ from 100 Hz to 100 kHz, in temperatures of 323-343 K as in Fig. 7a,b could indicated the dielectric resonant phenomenon in the compounds. Reports have discussed the resonance occurrence and related to the damped harmonic oscillator model [50][51][52] . However, there are no ε ′′ peaks in Fig. 7c were available at the same frequencies range to support this argument. Thus, the probability of resonance occurrence can be dismissed. Figure 8 exhibits changes of Tan δ variation in Mn 0.08 against various temperature from 298 to 343 K. All other samples exhibited almost same pattern of Tan δ as in Mn 0.08 (not shown). The presence of two peaks at minimum and maximum frequencies were apparent in almost all temperatures. As temperature elevated until 313 K, the minimum frequency peaks are enhanced meanwhile maximum frequency peaks are reduced. In temperature of 323-343 K, there were absence of the minimum frequency peaks. Meanwhile, there is a possibility of maximum frequency peaks to shift towards higher frequencies as temperature elevated.
The presence of peaks at minimum frequency in temperature of 298-313 K in Fig. 8 could be because of heavy electrical dipoles cannot effectively follow the alternation of electric field. In higher temperatures from 323 K until 343 K, smaller values of Tan δ at minimum frequency corresponded to the small value of ε ' at the same frequency. On the other hand, the presence of peaks at 1 MHz frequency and the possibilities of the peaks to shift towards the higher frequencies as temperature elevated can be interpreted as the possibility of relaxation peaks presence 53,54 at frequencies higher than 1 MHz since the electrical dipoles unable to match the alternation of higher external field frequencies.
The obtained results from our study showed that small B-site doping sample preparation can significantly alter the dielectric properties of tellurium based double perovskite. In addition, the ε ′ values at 1 MHz and in room temperature are comparable or better results compared to some reports with full or partial A-site doping samples preparations. Table 4 summarized the comparison of our results with the results from other reports.

Conclusion
SrLaLiTe 1−x Mn x O 6 double perovskites was successfully synthesized by applying solid state reaction method. All samples were confirmed to crystallized in monoclinic symmetry with P2 1 /n space group. The existence of maxima at specific wavenumbers confirms the formation of Te 6+ /Mn 6+ -O-Li + octahedral structure. The grain size of the compounds gets larger towards higher concentration doping until x = 0.08. The reduction of optical band gap can be deduced because of the doping of Mn 6+ cations that capable to reduce the gap distance between highest www.nature.com/scientificreports/ occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) in the compound. Nevertheless, all compounds showed absorption within visible light range which make the application for compounds in optoelectronic device such as solar cell can be realized with modifications on its electrical conductivity or tuning its E opt . Highest dielectric real permittivity and lowest dielectric losses at 1 MHz was recorded by Mn 0.08. This behaviour can be elucidated on the basis of grain size. This study shows that the dielectric and optical characteristics of tellurium based double perovskite can be tailored through small B-doped doping method.
Further studies related to annealing temperature, full doped sample preparation or other means can be done to increase the value of dielectric real permittivity in the future. The enhancement in dielectric and optical characteristics of the studied compounds revealed good potential for the materials to be used in electronic applications.