Crystal Structure and Electrical Properties of Ruthenium-Substituted Calcium Copper Titanate

This paper reports a detailed study of crystal structure and dielectric properties of ruthenium-substituted calcium-copper titanates (CaCu3Ti4−xRuxO12, CCTRO). A series of three samples with different stoichiometry was prepared: CaCu3Ti4−xRuxO12, x = 0, 1 and 4, abbreviated as CCTO, CCT3RO and CCRO, respectively. A detailed structural analysis of CCTRO samples was done by the Rietveld refinement of XRPD data. The results show that, regardless of whether Ti4+ or Ru4+ ions are placed in B crystallographic position in AA’3B4O12 (CaCu3Ti4−xRuxO12) unit cell, the crystal structure remains cubic with Im3¯ symmetry. Slight increases in the unit cell parameters, cell volume and interatomic distances indicate that Ru4+ ions with larger ionic radii (0.62 Å) than Ti4+ (0.605 Å) are incorporated in the CaCu3Ti4−xRuxO12 crystal lattice. The structural investigations were confirmed using TEM, HRTEM and ADF/STEM analyses, including EDXS elemental mapping. The effect of Ru atoms share in CaCu3Ti4−xRuxO12 samples on their electrical properties was determined by impedance and dielectric measurements. Results of dielectric measurements indicate that one atom of ruthenium per CaCu3Ti4−xRuxO12 unit cell transforms dielectric CCTO into conductive CCT3RO while preserving cubic crystal structure. Our findings about CCTO and CCT3RO ceramics promote them as ideal tandem to overcome the problem of stress on dielectric-electrode interfaces in capacitors.


Introduction
The CaCu 3 Ti 4 O 12 (CCTO) materials belong to a large family of the AA' 3 B 4 O 12 A-site ordered perovskites. During the last two decades, CCTO materials have been extensively investigated due to very high dielectric permittivity (up to 10 5 ), which is practically constant in a broad temperature (100 and 600 K) and frequency (1 kHz-1 MHz) range [1][2][3][4]. Due to such characteristics, CCTO ceramics attract the attention of researchers as very promising materials for applications in microelectronics and microwave devices [5]. The giant dielectric permittivity of CCTO materials promotes their utilization for the fabrication of capacitors with improved capacitive performance for portable electronic devices [6]. Though, it was observed that the nature of the ceramic-electrode interface could affect the decrease of the dielectric permittivity. Significant differences in the crystal structure and electrical properties of the CaCu 3 Ti 4 O 12 as a dielectric material and a metallic electrode cause an energy barrier and the occurrence of stress on the ceramic-electrode interface, which reduces the dielectric permittivity. Therefore, for the fabrication of capacitors with good capacitive performance, it is important to combine dielectric ceramic and electrodes Chemical Company Inc.), which was dissolved in 10 mL of acidified water. The stable solutions were mixed, and, in this solution, a RuO 2 (99.9%, Aldrich) was dispersed. The prepared dispersion was evaporated in a silicon oil bath while mixing with a magnetic stirrer. Obtained dry powder was homogenized with an agate mortar and pestle and consecutively calcined at 700 • C. The resulting powder was milled in planetary YTZ ball-mil for 0.5 h in ethanol and fired in the air at 900 • C for 20 h.
The structural investigations were performed by the Rietveld refinement [19] of the XRD data recorded at room temperature on a Philips PW 1050 diffractometer with CuKα1,2 (λ = 1.54178 Å) Ni-filtrated radiation. The diffraction intensity was measured from 10 to 130 • 2θ, using a step size of 0.02 • with a counting time of 12 s per step. The working conditions were 40 kV and 20 mA. The Rietveld refinement of the CaCu 3 Ti 4−x Ru x O 12 crystal structure was performed using the FullProf computing program in the WinPLOTR environment [26][27][28][29]. The unit cell parameters were calculated using the program LSUCRI (least squares unit cell refinement with indexing) [30]. The Rietveld refinement was started from the calculated unit cell parameters and previously reported atomic positions [15]. A TCH-pseudo-Voigt function was chosen for the line shape of the diffraction peaks, whereas a 6-coefficient polynomial function was used for the background description. Transmission electron microscopy (TEM) and ADF/STEM with EDXS elemental mapping were performed using Jeol JEM 210 TEM and Titan G2 80-200 with ChemiSTEM technology. The fast Fourier transformation (FFT) analysis was done in the Digital Micrograph computer program.
The synthesized CCTRO powders were uniaxially pressed in a die (∅ 10 mm) under the pressure of about 400 MPa. Each compact has a thickness of circa 3 mm and 60 ± 2% of theoretical density. Sintering of the green compacts was carried out in a Protherm tube furnace in an air atmosphere, with a heating rate of 5 • /min up to 1050 • C and with a dwell time of 12 h. The electrical measurements were done on the sintered CCTRO pellets. The capacitance and conductivity measurements were performed in function of temperature at a frequency of 1 kHz (internal frequency of the instrument) at the Wayne Kerr Universal Bridge B224. All measurements were done in the cooling mode from 150 to 21.3 • C in an air atmosphere. In order to improve the contact between the silver electrodes and the surface of the samples, both sample bases are coated with a suspension of silver powder and ethyl acetate. Impedance measurements were performed on the Potentiostat-Galvanostat PAR 273A coupled with Dual Phaslock-in Amplifier 5210 brand. The measurements were done in the frequency interval from 0.01 Hz to 100 kHz, and the temperature range from 150 • C to room temperature. Mathematical modeling of experimentally obtained data was done by Z-View2 (version 2.6 demo) computer program.

Results and Discussion
The structural refinement of room temperature XRD data of three investigated powders: CaCu 3 Ti 4 O 12 (CCTO), CaCu 3 Ti 3 RuO 12 (CCT3RO) and CaCu 3 Ru 4 O 12 (CCRO), were performed in the space group Im3. The Rietveld analysis was done by using the FullProf computing program. A pseudo-Voigt function was chosen to generate the line shape of the diffraction maximums. The scale factor, peak-shape parameters, lattice parameters, variable fractional atomic coordinates, and displacement parameters were refined. The unit cell parameters of the CCTO materials are doubled in all three crystallographic axes (V = 2ap·2ap·2ap) compared to the ideal (ABO 3 ) cubic perovskite lattice (ap). The Rietveld refinements were started from the previously reported atomic positions: Ca atoms were situated in 2a Wyckoff sites at the (0 0 0) position, Cu atoms occupied the 6b at (0 1 2 1 2 ), Ti in 8c (0 1 4 1 4 ), and O in 24g (x y 0) sites. The literature data [15] of the positional coordinates were used as starting values for structural refinement. Figure 1 illustrates the good agreement between calculated and observed structural models. The obtained results confirm that the powders are a single-phase perovskite oxide without a secondary phase. the displacement parameters were constrained at the same value. The titanium and ruthenium atoms in the CaCu3Ti4−xRuxO12 crystal structure are situated in the same crystallographic positions, so their fractional atomic coordinates were set to be equal during refinement. Their occupational factors are interrelated, and only the Ti/Ru ratio was allowed to vary, keeping the overall occupancy of the respective sites constant according to the appropriate stoichiometric ratio. The refined values of the unit cell parameters, atomic position parameters, and displacement parameters are listed in Table 1.   The structural refinement was started by allowing varying the values of the zero point, the scale factor, and the background coefficients. Then, lattice parameters, asymmetry, and half-width parameters (U, V, W) were refined. When satisfactory values of profile parameters were obtained, the structural parameters, i.e., atomic positional parameters and isotropic displacement parameters, were allowed to vary. The overall displacement parameter was refined for every atom; the obtained value was set as starting isotropic displacement parameter for each of the atoms. The isotropic displacement parameters for Ca, Cu, Ti and Ru atoms were refined independently, while for the O atoms, the displacement parameters were constrained at the same value. The titanium and ruthenium atoms in the CaCu 3 Ti 4−x Ru x O 12 crystal structure are situated in the same crystallographic positions, so their fractional atomic coordinates were set to be equal during refinement. Their occupational factors are interrelated, and only the Ti/Ru ratio was allowed to vary, keeping the overall occupancy of the respective sites constant according to the appropriate stoichiometric ratio. The refined values of the unit cell parameters, atomic position parameters, and displacement parameters are listed in Table 1.

CaCu 3 Ru 4 O 12 Im3
Ti/Ru The refined values of the unit cell parameters and unit cell volume increase with increases in ruthenium ions content in the crystal structure of the investigated powders (Table 1, Figure 2). The increases in unit cell parameters are provoked by partial (1 of 4) or complete (4 of 4) replacement of Ti 4+ ions (0.605 Å ionic radii for six coordination) by larger Ru 4+ ions (0.62 Å ionic radii for six coordination) in the CaCu 3 Ti 4−x Ru x O 12 crystal structures. The results of the structural investigation were confirmed by TEM and SAED analyses. TEM and SAED analyses of CCTO and CCRO powders are presented in Figure 3. Figure 3a reveals significant agglomeration of CCTO single-crystalline particles whose size is about several tens of nanometers (inset in Figure 3a). SAED pattern, presented in Figure 3b Figure 3c. Their single-crystalline nature is exposed in Figure 3d, which shows the overlapping of crystals with 2 different zone axes  The results of the structural investigation were confirmed by TEM and SAED anal yses. TEM and SAED analyses of CCTO and CCRO powders are presented in Figure 3 Figure 3a reveals significant agglomeration of CCTO single-crystalline particles whose size is about several tens of nanometers (inset in Figure 3a). SAED pattern, presented in Figure 3b, confirms crystallographic information obtained through Rietveld refinement Interplanar distances of 3.701, 2.597, 1.837, 1.516, and 1.253 Å correspond well too (200) (220), (400), (422), and (433) planes of CaCu3Ti4O12 phase, JCPDS 75-2188. Similarly, many randomly oriented CCRO particles are presented in Figure 3c. Their single-crystalline nature is exposed in Figure 3d, which shows the overlapping of crystals with 2 differen zone axes with low mismatching. Observed interplanar distances of 3.740, 1.860, 2.630 and 1.320 Å are assigned to (200)   High-angle annular dark-field STEM and corresponding EDXS elemental mapping presented in Figure 5   High-angle annular dark-field STEM and corresponding EDXS elemental mapping presented in Figure 5 implies the homogeneous distribution of Ca, Cu, Ti, Ru and O in CCT3RO particles, independently of their size. Complex impedance spectroscopy (IS) was used to determine the electrical properties of sintered CCTRO ceramics. Recorded Nyquist plots are presented in Figure 6.  High-angle annular dark-field STEM and corresponding EDXS elemental mapping presented in Figure 5 implies the homogeneous distribution of Ca, Cu, Ti, Ru and O in CCT3RO particles, independently of their size. Complex impedance spectroscopy (IS) was used to determine the electrical properties of sintered CCTRO ceramics. Recorded Nyquist plots are presented in Figure 6. Complex impedance spectroscopy (IS) was used to determine the electrical properties of sintered CCTRO ceramics. Recorded Nyquist plots are presented in Figure 6.
The recorded complex impedance spectra of CCTRO sintered ceramics were fitted using the Z-View2 computing program according to the equivalent circuit. The equivalent circuit consisting of one resistor (R) and one capacitor (constant phase element, CPE) connected in parallel best describes the impedance spectra of CCTRO materials. The calculated resistance values are in the range of 1 to 150 MΩ. Relatively high values of resistance indicate good dielectric properties of CCTO material. On the contrary, the impedance spectra of sintered ceramics with ruthenium content show an accumulation of points for higher and middle-frequency values of about 0.1 Ω and 0.05 Ω for CCT3R and CCRO, respectively. Only at very low-frequency values, from 0.1 to 0.01 Hz, the impedance spectra of these ceramics can be described by an equivalent circuit of one resistor (R) and one capacitor (CPE) connected in parallel. However, the difference in the obtained resistivity values for CCTRO materials with and without ruthenium content is more than eight orders of magnitude in the direct current circuit. Such behavior indicates good conductivity of CCT3RO and CCRO ceramics. The electrical conductivity of investigated ceramics at four characteristic frequency points is presented in Table 3. The recorded complex impedance spectra of CCTRO sintered ceramics were fitted using the Z-View2 computing program according to the equivalent circuit. The equivalent circuit consisting of one resistor (R) and one capacitor (constant phase element, CPE) connected in parallel best describes the impedance spectra of CCTRO materials. The calculated resistance values are in the range of 1 to 150 MΩ. Relatively high values of resistance indicate good dielectric properties of CCTO material. On the contrary, the impedance spectra of sintered ceramics with ruthenium content show an accumulation of points for higher and middle-frequency values of about 0.1 Ω and 0.05 Ω for CCT3R and CCRO, respectively. Only at very low-frequency values, from 0.1 to 0.01 Hz, the impedance spectra of these ceramics can be described by an equivalent circuit of one resistor (R) and one capacitor (CPE) connected in parallel. However, the difference in the obtained resistivity values for CCTRO materials with and without ruthenium content is more than eight orders of magnitude in the direct current circuit. Such behavior indicates good conductivity of CCT3RO and CCRO ceramics. The electrical conductivity of investigated ceramics at four characteristic frequency points is presented in Table 3. The difference in electrical conductivity can be explained by different electron configurations of Ti 4+ and Ru 4+ ions [14,31,32]. Titanium is a 3d element with an electronic configuration: Ti: 1s 2 2s 2 2p 6 3s 2 3p 6 4s 2 3d 2 However, the electronic configuration of Ti 4+ cation in the CCTO crystal structure turns into:  The difference in electrical conductivity can be explained by different electron configurations of Ti 4+ and Ru 4+ ions [14,31,32]. Titanium is a 3d element with an electronic configuration: Ti: 1s 2 2s 2 2p 6 3s 2 3p 6 4s 2 3d 2 However, the electronic configuration of Ti 4+ cation in the CCTO crystal structure turns into: Ti 4+ : 1s 2 2s 2 2p 6 3s 2 3p 6 with closed external 3p 6 shells, which explains the relatively low conductivity of these materials. The isomorphic replacement of the Ti 4+ ions by Ru 4+ leads to different electronic configurations. Explicitly, Ru 4+ has the following electronic configuration: Ru 4+ : 1s 2 2s 2 2p 6 3s 2 3p 6 4s 2 3d 10 4p 6 5s 0 4d 4 , with uncompleted d subshells. These four electrons in the octahedral crystalline environment are situated in t g crystalline orbitals containing two unpaired d electrons, i.e., Ru 4 + represents a magnetic ion in the CCRO crystal. The CCRO conductivity can be explained by the Kondo mechanism. Namely, CCRO materials are d-electron heavy-fermion system, which behaves similarly to an f -electron heavy-fermion system. In CCRO structures, the Cu 3d electrons are localized, while 4d electrons of magnetic Ru 4+ ions contribute to the conductivity [14,32,33]. Figure 7 displays the complex impedance spectra (Nyquist plot) of CCTO ceramic measured in an air atmosphere, in the temperature range from 150 to 23 • C (room temperature) Materials 2022, 15, 8500 9 of 14 and frequency interval from 100 kHz to 0.01 Hz. The obtained impedance diagrams represent the sum of the grain-interior and grain boundary resistance, which is impossible to separate. Thus, the fitting of summary data was done. The fitting was achieved based on the corresponding equivalent circuit (inset in Figure 7) and the Z-View2 computing program.
represents a magnetic ion in the CCRO crystal. The CCRO conductivity can be ex-plained by the Kondo mechanism. Namely, CCRO materials are d-electron heavy-fermion system, which behaves similarly to an f-electron heavy-fermion system. In CCRO structures, the Cu 3d electrons are localized, while 4d electrons of magnetic Ru 4+ ions contribute to the conductivity [14,32,33]. Figure 7 displays the complex impedance spectra (Nyquist plot) of CCTO ceramic measured in an air atmosphere, in the temperature range from 150 to 23 °C (room temperature) and frequency interval from 100 kHz to 0.01 Hz. The obtained impedance diagrams represent the sum of the grain-interior and grain boundary resistance, which is impossible to separate. Thus, the fitting of summary data was done. The fitting was achieved based on the corresponding equivalent circuit (inset in Figure 7) and the Z-View2 computing program.  Table 4. The calculated values of electrical conductivity (Table 4) were used to determine the activation energies Ea based on the Arrhenius equation, Equation (1) [34,35]: (1) where Ea is the activation energy, kB is the Boltzmann's constant (8.8617343 × 10 −5 eV/K), and σo is a constant. The obtained relationship is almost linear (Figure 8). The calculated value of activation energy of 0.49 eV is in accordance with literature data [1][2][3]. The values of electric resistivity determined by fitting with an appropriate equivalent circuit were used to calculate the values of electrical resistivity ρ [Ω·cm] and specific electrical conductivity σ (S·cm −1 ). The calculated values are presented in Table 4. The calculated values of electrical conductivity (Table 4) were used to determine the activation energies E a based on the Arrhenius equation, Equation (1) [34,35]: where E a is the activation energy, k B is the Boltzmann's constant (8.8617343 × 10 −5 eV/K), and σ o is a constant. The obtained relationship is almost linear (Figure 8). The calculated value of activation energy of 0.49 eV is in accordance with literature data [1][2][3].  Figure 9 represents the influence of both the temperature and the frequency on the dielectric properties of the CCTO ceramic. The measurements were done in the frequency interval from 100 kHz to 0.01 Hz and in the temperature range between 177 and 23 • C.
177.0 4.85 × 10 5 0.22 × 10 −3 137. 3 2.17 × 10 4 0.48 × 10 −4 98. 9 9.39 × 10 4 0.11 × 10 −4 67. 6 2.72 × 10 5 0.35 × 10 −5 38. 9 8.96 × 10 5 0.12 × 10 −5 23.0 1.47 × 10 6 0.72 × 10 −6 Figure 8. The Arrhenius plot of conductivity ln(σ) versus 1000/T of CCTO ceramic. Figure 9 represents the influence of both the temperature and the frequency on the dielectric properties of the CCTO ceramic. The measurements were done in the frequency interval from 100 kHz to 0.01 Hz and in the temperature range between 177 and 23 °C. It is obvious that CCTO materials possess a high dielectric permittivity ε' (> 10 3 ) which significantly depends on temperature and frequency (Table 5).    It is obvious that CCTO materials possess a high dielectric permittivity ε' (> 10 3 ) which significantly depends on temperature and frequency (Table 5). It is obvious that CCTO materials possess a high dielectric permittivity ε' (>10 3 ) which significantly depends on temperature and frequency (Table 5). The influence of temperature and frequency on the dielectric loss (tgδ) of the CCTO ceramic is shown in Figure 10. For the three highest temperatures, the dielectric loss monotonically decreases with increasing frequency. For the three lowest temperatures, the dielectric losses slightly increase in the frequency range from 10 to 100 kHz. The influence of temperature and frequency on the dielectric loss (tgδ) of the CCTO ceramic is shown in Figure 10. For the three highest temperatures, the dielectric loss monotonically decreases with increasing frequency. For the three lowest temperatures, the dielectric losses slightly increase in the frequency range from 10 to 100 kHz. The relative dielectric constant εr for CCTO material was measured in cooling mode from 150 °C to room temperature at a constant frequency of 1 kHz. The relative dielectric constant increases with increasing temperature from about 1460 at room temperature to 2460 at 150 °C (Figure 11a). It can be seen that these measurements are in good agreement with the values obtained from the impedance measurements at the same frequency of 1 kHz (Table 5). The relative dielectric constant ε r for CCTO material was measured in cooling mode from 150 • C to room temperature at a constant frequency of 1 kHz. The relative dielectric constant increases with increasing temperature from about 1460 at room temperature to 2460 at 150 • C (Figure 11a). It can be seen that these measurements are in good agreement with the values obtained from the impedance measurements at the same frequency of 1 kHz (Table 5).
Materials 2022, 15, x FOR PEER REVIEW 12 of 15 Figure 11. The temperature dependence of (a) relative dielectric constant was calculated based on impedance measurements (circle) and based on measured capacitance (square), and (b) dielectric loss for CCTO ceramic.
The dependence of dielectric loss (tgδ) on temperature is represented in Figure 11b. It can be seen that the tgδ increases with an increase in the temperature. These values are in the range between 0.13 at room temperature and 0.65 at 150 °C, which is in good agreement with the values obtained at 1 kHz ( Figure 10).
The results of electrical measurements clearly show that the stoichiometry of Ca-Cu3Ti4−xRuxO12 significantly affects their electrical properties. It is known that Ca-Cu3Ti4O12 material exhibits excellent dielectric properties with a dielectric constant in excess of 10 4 . Quite the contrary, the CaCu3Ru4O12 material exhibits conductive behavior [7,[36][37][38]. The results presented in this paper revealed that the substitution of one Ti 4+ ion by Ru 4+ per the CaCu3Ti4O12 unit cell changes the electrical characteristics of these materials from dielectric to conductive. Such electrical properties of these materials are very important for their potential application in the electronics industry for capacitors production. However, in these components, the problem of stress on the dielectric-electrode Figure 11. The temperature dependence of (a) relative dielectric constant was calculated based on impedance measurements (circle) and based on measured capacitance (square), and (b) dielectric loss for CCTO ceramic.
The dependence of dielectric loss (tgδ) on temperature is represented in Figure 11b. It can be seen that the tgδ increases with an increase in the temperature. These values are in the range between 0.13 at room temperature and 0.65 at 150 • C, which is in good agreement with the values obtained at 1 kHz ( Figure 10).
The results of electrical measurements clearly show that the stoichiometry of CaCu 3 Ti 4−x Ru x O 12 significantly affects their electrical properties. It is known that CaCu 3 Ti 4 O 12 material exhibits excellent dielectric properties with a dielectric constant in excess of 10 4 . Quite the contrary, the CaCu 3 Ru 4 O 12 material exhibits conductive behavior [7,[36][37][38]. The results presented in this paper revealed that the substitution of one Ti 4+ ion by Ru 4+ per the CaCu 3 Ti 4 O 12 unit cell changes the electrical characteristics of these materials from dielectric to conductive. Such electrical properties of these materials are very important for their potential application in the electronics industry for capacitors production. However, in these components, the problem of stress on the dielectric-electrode interfaces occurs due to the different structural characteristics of the metal (electrode) and ceramic materials. This problem can be overcome by the use of materials that would have the same or similar structural characteristics both with dielectric and with an electrode, which would be used as an interface. Potential materials that could be used for the production of a contact or interlayer are CCT3RO and CCRO, owing to the same crystal structure and close lattice parameters (Table 1).

Conclusions
In this paper, we presented a detailed analysis of crystal structure and dielectric properties of three CaCu 3 Ti 4−x Ru x O 12 (x = 0, 1 and 4) samples. The detailed analysis of the crystal structure was determined by Rietveld's refinement of XRD data. The results of the structural analysis show that regardless of whether Ti 4+ or Ru 4+ ions are placed in B crystallographic position in the CaCu 3 Ti 4−x Ru x O 12 unit cell, the crystal structure remains cubic with the Im3 space group. Slight increases in the unit cell parameters, cell volume and interatomic distances indicate that Ru 4+ ions with larger ionic radii (0.62 Å) than Ti 4+ (0.605 Å) are incorporated in the CaCu 3 Ti 4−x Ru x O 12 crystal lattice. Results of TEM, HRTEM and ADF/STEM analyses with EDXS elemental mapping confirmed the results of structural investigations. The electrical properties of the CaCu 3 Ti 4−x Ru x O 12 materials were investigated based on impedance spectroscopy and dielectric measurements. Values of electrical resistance measured at room temperature indicate that the ruthenium ions in the CaCu 3 Ti 4 O 12 structure change the electrical properties of these materials from dielectric to conductive. What is more, just one atom of ruthenium per unit cell is enough to drastically improve the conductive properties of these materials. With the addition of ruthenium ions, conductivity increases for more than five orders of magnitude for the frequency of 100 kHz and for almost eight orders of magnitude for the frequency of 100 Hz.
Our findings about CCTO and CCT3RO ceramics, as both an inexpensive and ideal tandem to overcome the problem of stress on dielectric-electrode interfaces in capacitors, promote their commercial application in the electronic industry.