SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

The structural and dielectric properties of (1−x)BaTiO3xBiScO3 (x=0–0.5) ceramics were investigated to acquire a better understanding of the binary system, including determination of the symmetry of the phases, the associated dielectric properties, and the differences in the roles of Bi2O3 and BiScO3 substitutions in a BaTiO3 solid solution. The solubility limit for BiScO3 into the BaTiO3 perovskite structure was determined to be about x=0.4. A systematic structural change from the ferroelectric tetragonal phase to a pseudo-cubic one was observed at about x=0.05–0.075 at room temperature. Dielectric measurements revealed a gradual change from proper ferroelectric behavior in pure BaTiO3 to highly diffusive and dispersive relaxor-like characteristics from 10 to 40 mol% BiScO3. Several of the compositions showed high relative permittivities with low-temperature coefficients of capacitance over a wide range of temperature. Quantification of the relaxation behavior was obtained through the Vogel–Fulcher model, which yielded an activation energy of 0.2–0.3 eV. The attempt characteristic frequency was 1013 Hz and the freezing temperature, Tf, ranged from −177° to −93°C as a function of composition. The high coercive fields, low remanent polarization, and high activation energies suggest that in the BiScO3–BaTiO3 solid solutions, the polarization in nanopolar regions is weakly coupled from region to region, limiting the ability to obtain long-range dipole ordering in these relaxors under field-cooled conditions.


I. Introduction

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

The BaTiO3-based solid solutions have been extensively studied as means of tailoring the ferroelectric transition temperatures (i.e., manipulating the transitions by compositional routes to engineer the temperature dependence of the dielectric constant), and controlling the defect chemistry or defect mobilities as a function of applied dc fields.1–7 The resulting materials have been widely used in capacitors, thermistors, dielectric bolometers, pyroelectrics, and piezoelectrics.8,9 As new perovskite end-member materials are developed, it is thus interesting to study how the structure and functional properties of BaTiO3-based ceramics are modified, both from the standpoint of developing new materials and in developing a fundamental understanding of ferroelectrics.

Bi-doped BaTiO3 has been studied by several authors.10–13 The motivation for Bi doping in BaTiO3 includes the use of Bi as a sintering aid14 or using Bi as a donor dopant on the Ba site.15 The solubility limit of Bi in BaTiO3 has been reported to be 5–10 mol% on the A site of the perovskite lattice.12,13 The dielectric behavior is characterized by a transition from a normal ferroelectric to a relaxor as the Bi concentration increases.

Recently, Tinberg and Trolier-McKinstry16 reported on (1−x)BaTiO3xBiScO3 thin films, following the work of Eitel et al.17,18 on BiScO3–PbTiO3 piezoelectrics. A pure perovskite phase was achieved in the films with x=0.2–0.45, even without epitaxy, which suggests that a stable perovskite phase could be prepared in bulk material as well. The resulting films showed highly dispersive dielectric peaks with modest remanent polarization values at room temperature.

In order to determine whether the dispersion is a consequence of defects or stresses associated with the thin film preparation or an intrinsic material property, in this work the properties of bulk ceramics in this system were studied. In particular, bulk (1−x)BaTiO3xBiScO3 (x=0–0.5) ceramics were investigated to acquire a better understanding of this system, including determination of the phase diagram, the low and high field dielectric properties, and the differences in the roles of Bi2O3 and BiScO3 doping on BaTiO3.

II. Experimental Procedure

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

Bulk (1−x)BaTiO3xBiScO3 (x=0–0.5) ceramic samples were prepared by conventional solid state ceramic processing from reagent grade powders of barium carbonate (BaCO3, 99.0%, Alfa Aesar, Ward Hill, MA), titanium oxide (TiO2, 99.97%, 0.25 μm, Ishihara, Corporation, San Fransisco, CA), bismuth oxide (Bi2O3, 99.9%, MCP, Fairfield, CT), and scandium oxide (Sc2O3, 99.0%, The Low Hanging Fruit Company, CA). Raw materials were batched for the desired composition stoichiometrically, and mixed in an aqueous solution with 0.2 vol% ammonium hydroxide (J. T. Baker, Phillipsburg, NJ) and 0.5 vol% dispersant (Darvan 821A, R. T. Vanderbilt, Norwalk, Connecticut) to reduce the flocculation of the powders. The solution was then ball milled with stabilized zirconia media (cylindrical with ∼9 mm height and 9 mm diameter; TOSOH Ceramics, Tokyo, Japan) for 24 h. After drying the suspension at 130°C, the powder mixture was calcined in air for 4 h at 700°–1000°C as shown in Table I. For better homogeneity, the powder was mixed in the same aqueous solution mentioned above with the media, vibratory milled for 18 h, and dried at 130°C before a second calcination in air for 4 h. Vibratory milling was used for the second milling to crush agglomerates into finer powders. After checking for the formation of the perovskite phase by using X-ray diffraction (XRD), the calcined powder was again vibratory milled with the media and dried as described before preparing ceramic compacts.

Table I.    Calcination and Sintering Temperatures for Each Composition
SampleCompositionTemperature (°C)
(1−x)BaTiO3xBiScO31st calcination2nd calcinationSintering
  1. Note that two temperatures in one calcination step mean 2 h at each temperature (total 4 h).

11080010001320
20.9990.001700, 850750, 9001300
30.9980.002700, 850750, 9001300
40.9970.003700, 850750, 9001300
50.9960.004700, 850750, 9001300
60.9950.005700, 850750, 9001300
70.990.01700, 850750, 9001300
80.980.02700, 850750, 9001300
90.970.03700, 850750, 9001300
100.960.04700, 850750, 9001300
110.950.05800700, 8501300
120.9250.075800700, 8501300
130.90.1800700, 8501300
140.850.15800700, 8501300
150.80.2800700, 8501300
160.70.3800700, 8501250
170.60.4800700, 8501150
180.50.5800700, 8501050

The calcined powder was mixed with 3–4 wt% acrylic binder (Acrylic Resin, Rohm and Haas, Philadelphia, PA) to increase the green strength of the compacts, pulverized, and sieved through a mesh screen with 180 μm openings. The powder was uniaxially pressed at 70 MPa to form disks 13 mm in diameter and about 1 mm in thickness. The binder was burned out at 325°C for 90 min and at 550°C for 60 min. Samples were then cold isostatic pressed at 200 MPa to increase the green density. The pellets were subsequently sintered at 1050°–1320°C, as shown in Table I, for 1–20 h in a closed crucible with a source powder of the same composition to minimize bismuth loss due to its volatility. The exception was the samples with x=0, which were sintered in a clean furnace with an open, O2-rich environment. The weight loss during sintering was confirmed to be <1% for all samples.

In order to achieve densities higher than 95% theoretical density for the electrical measurements, an additional 3 mol% Bi2O3 was added as a sintering aid at the same time as the binder for samples with x=0.001–0.15. Bi2O3 has a melting temperature of about 825°C,19 which is lower than the sintering temperatures used in this study, and so the excess Bi2O3 was believed to improve the density through liquid phase sintering. Weight loss measurement suggested that more than two-thirds of the excess Bi2O3 added evaporated during sintering.

XRD was performed at room temperature for structural analysis and lattice parameter determination using a Scintag PADV diffractometer with CuKα radiation (35 kV, 30 mA, 0.02°/step, 3 s/step conditions were used for lattice parameter calculation and 0.6 s/step for the phase determination). Sintered pellets were crushed and ground into powder, followed by annealing at 300°C to reduce the stress from the grinding. A silicon internal standard with a certified lattice constant of 5.4301±0.0001 Å (Gem Dugout, State College, PA) was used for precise lattice parameter determinations. XRD patterns were calibrated with the internal standard, and peaks were obtained using Jade software (Materials Data Inc., Livermore, CA). Lattice parameters were calculated from the peaks using a least-squares minimization of errors.

Microstructural analysis was performed on some of the sintered samples with a transmission electron microscope (TEM, Philips 420, bright-field imaging at an operating voltage of 120 kV, Eindhoven, Netherlands). Samples for TEM were mechanically polished down to about 30 μm, and then ion milled (Fischione Instruments model 1010, Export, PA or Gatan model 691, Pleasonton, CA) to make the samples electron transparent. The possibility of compositional gradients within samples was analyzed with an energy-dispersive spectrometer (EDS, e2v Scientific Instruments, Buckinghamshire, U.K.) on a Philips 420 TEM.

For electrical measurements on pellet samples, the sintered pellets were polished to achieve parallel, smooth faces, and gold electrodes were sputtered on both faces of the pellets. Dielectric measurements were performed from −120° to 400°C in custom-designed furnaces with a LF impedance analyzer (HP4192A, Agilent, Santa Clara, CA) with computerized control and data collection systems using a heating or cooling rate of 0.5°–2°C/min. The frequencies used for the measurement were from 100 Hz to 1MHz, with an oscillating voltage of 1.0 V. To minimize the effect of moisture on capacitance measurements, data obtained from a cooling cycle were used for measurements <∼100°C. A modified Sawyer-Tower circuit was used for ferroelectric switching measurements in custom-designed furnaces covering the temperature range from −188° to 400°C. PUND measurements were also performed to monitor switchable polarization, Psw, as a function of temperature using an AixACCT TF Analyzer 2000 (Aachen, Germany) in the same temperature chamber as the ferroelectric switching measurements.

III. Results and Discussion

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

(1) Structural Characterization

Figure 1 shows the XRD data for samples with x=0–0.5 after sintering; it can be seen that all of the compositions with x=0–0.4 could be prepared in the desired perovskite phase with little or no second phases. Second phase peaks were clearly observed for x=0.5. The second phases matched the powder diffraction peaks for Bi20TiO32 and Bi2O3 (ICDD PDF #00-042-0202 and #00-027-0050, respectively). The solubility limit for BiScO3 in BaTiO3 is much larger than Bi-doped BaTiO3, where the limit has variously been reported to range from x=0.05 to 0.10 for Ba1−3/2xBixTiO3.12,13 This can be explained by the fact that the simultaneous addition of Sc with Bi should maintain the charge balance without the formation of additional metal and oxygen vacancies. In the BaTiO3–BiScO3 system, the following substitution is expected:

  • image(1)
image

Figure 1.  X-ray diffraction pattern on (1−x)BaTiO3xBiScO3 ceramics. The indexing is based on either tetragonal or pseudocubic perovskite unit cells.

Download figure to PowerPoint

It is confirmed from the XRD patterns that the perovskite structure is tetragonally distorted up to x=0.05 and switches to a pseudocubic phase with x=0.075–0.1; the 220 and 202 peaks were easily resolved up to x=0.05 and then merged to a single peak for x=0.075—within the resolution of our instrument. Clear splitting of the {hhh} family of peaks, which is characteristic of rhombohedral structures, was not observed for any of the samples. Similar results have been recently reported by Datta and Thomas.20

The lattice parameters calculated from the XRD data assuming a tetragonal phase (x=0–0.05) or a pseudocubic phase (x=0.075–0.40) are plotted in Fig. 2. The lattice parameters were determined using a least-squares fit to at least six and four indexed diffraction peaks for tetragonal and cubic cells, respectively. The linear increase in the lattice parameter with BiScO3 concentration in the pseudocubic phase up to x∼0.4 is consistent with Vegard's law,21 confirming solid solution. This limit agrees well with the data on thin films, which reported that a pure perovskite phase was achieved with x=0.2–0.45 without epitaxy.16 The observed decrease in tetragonality with BiScO3 concentration is consistent with the decreasing tolerance factor, t (t=1.062 and 0.907 for pure BaTiO3 and BiScO3, respectively). A relationship between tolerance factor and crystal structure has been well established in perovskites, and a lower tolerance factor generally favors rhombohedral, orthorhombic, or monoclinic phases.17,22 However, the composition for the transition disagrees with the data on films, which showed clear tetragonality even with x=0.20 and 0.30.16 It should be noted that the film data show the symmetry change near a tolerance factor of 1, while the ceramics investigated here become pseudocubic at a tolerance factor of about 1.05, an unusually high value. The origin of this discrepancy is not known, but may be a function of strains associated with the film deposition, or may reflect differences between powders and continuous samples.23

image

Figure 2.  Lattice parameters as a function of BiScO3 concentration (measured on powdered samples after sintering).

Download figure to PowerPoint

The effect of the excess Bi2O3 added as a sintering aid on the crystal structure was investigated using the XRD data of samples with x=0.001 with and without the excess Bi. XRD data on the sintered samples with extra Bi2O3 showed less well-defined tetragonality than the ones without extra Bi2O3. Apparently, the defects associated with the excess Bi2O3 may contribute to a decreased tetragonality of the perovskite. This suggests that some amount of the Bi2O3 added as a sintering aid diffused into the structure during sintering. The impact of this fact needs to be considered carefully in the dielectric property analysis.

(2) Microstructure

The process used in this work resulted in fine-grained material, especially near the BaTiO3 end of the solid solution. Figure 3 shows TEM micrographs of sintered samples for the compositions x=0, 0.02, 0.05, and 0.1 with 1-h sintering. Pure BaTiO3 had grains larger than 5 μm, as shown in Fig. 3(a), but the composition with a few % BiScO3 showed about a 0.3 μm grain size (Fig. 3(b)). The grain size increased slowly with further increase in the BiScO3 concentration; an average grain size of 1 μm was observed for 10% BiScO3 (Fig. 3(d)). This result is counterintuitive because extra Bi2O3, which would presumably act as a sintering aid, was added to the samples with x=0.001–0.15. The mechanism that is responsible for this observation should be studied in future work.

image

Figure 3.  Transmission electron micrographs of (1−x)BaTiO3xBiScO3 ceramics for (a) x=0, (b) x=0.02, (c) x=0.05, and (d) x=0.1.

Download figure to PowerPoint

An interesting feature in the microstructure was also observed for x=0.03–0.1. The samples with these compositions were hard to homogenize during sintering, resulting in core–shell structures observed with TEM, as shown in Fig. 3(c) for x=0.05.2,3 It was found that with prolonged sintering times there are fewer core–shell grains and a reduction in the size of cores. However, the inhomogeneity could not be completely eliminated even with 20 h sintering. It was roughly estimated that more than a half of the grains exhibited core–shell features in 1-h-sintered samples, whereas <10% of the grains did in 20 h sintered samples. The consequence of this core–shell structure on dielectric properties is discussed in the following section. EDS results indicated that shell was enriched in BiScO3, while the cores were depleted.

(3) Dielectric Properties

Figure 4 shows the dielectric constant and loss tangent data of ceramics with x=0, 0.005, 0.02, 0.05, 0.2, and 0.4 measured at frequencies from 100 Hz to 1 MHz over the temperature range from −120° to 400°C. It is clear that as BiScO3 is added to BaTiO3, the materials quickly lose two of the characteristic sharp dielectric peaks and the maximum dielectric constant decreases significantly. That is, even for a small amount of BiScO3, i.e. x=0.005, the peaks associated with the rhombohedral to orthorhombic and orthorhombic to tetragonal phase transitions are difficult to distinguish.

image

Figure 4.  Dielectric constant and dielectric loss tangent data as a function of temperature measured at frequencies from 100 Hz to 1 MHz for (1−x)BaTiO3xBiScO3 ceramics (a) x=0, (b) x=0.005, (c) x=0.02, (d) x=0.05, (e) x=0.05 with 20 h sintering, (f) x=0.2, and (g) x=0.4. Note that all samples except for (e) were sintered for 1 h.

Download figure to PowerPoint

Samples with core–shell microstructures had two permittivity peaks at about 50° and 140°C over the measured temperature range, as shown in Fig. 4(d) for x=0.05. With longer sintering, the lower temperature dielectric peak grows and the high-temperature peak becomes a shoulder, as shown in Fig. 4(e). This is consistent with the observation that samples sintered for 20 h became more compositionally homogeneous. This also suggests that the cores of the grains observed in the TEM micrographs are rich in BaTiO3. These would then be responsible for the higher temperature peak. In contrast, the shell (rich in BiScO3 or the Bi2O3 added as a sintering aid) would then account for the lower temperature peak. The EDS results, which indicated BiScO3-rich shells, support this hypothesis.

For x=0.15 or more, a single diffuse dielectric peak with a dielectric constant of about 800–1000 over a wide range of temperature was observed. These data are very similar to those shown in thin films.16 Thus, the highly dispersive and diffusive dielectric behavior observed in thin films was confirmed in bulk materials as well. However, there is no strong increase in the room-temperature permittivity suggestive of a morphotropic phase boundary near the transition from tetragonal to pseudocubic structures. This contrasts with the observed increase in permittivity in thin films at x=0.4.16 The origin of this discrepancy is not known.

The temperature of the maximum dielectric constant (Tmax) was determined to be a function of composition, as shown in Fig. 5(a) for the BaTiO3–BiScO3 system. Figure 5(b) illustrates comparable data for the Bi-doped BaTiO3 system.13 In BaTiO3–BiScO3 ceramics, for x=0.03–0.1, the dielectric data from the samples sintered for 20 h are used for Tmax as they are assumed to be closer to their equilibrium state than the ones with shorter sintering periods. It was found that Tmax decreases monotonically with BiScO3 concentration for x=0–0.1. From x=0.1, Tmax increases roughly linearly.

image

Figure 5.  (a) Tmax for 100 kHz as a function of composition for (1−x)BaTiO3xBiScO3 ceramics, and (b) Tmax or Tc variation with x for Ba1−3/2xBixTiO3 ceramics (rescaled after Bahri et al.).13

Download figure to PowerPoint

It is important to note that there may be two effects convoluted for the low BiScO3 concentrations. As was described above, in order to densify these samples, excess Bi had to be added, and some of this excess Bi was incorporated into the BaTiO3. It has previously been demonstrated that Bi doping of BaTiO3 induces a slight decrease in Tc, as shown in Fig. 5(b).13 However, it is believed that the BiScO3 concentration (rather than the Bi concentration) dominates the drop in Tmax from x=0 to 0.02, because the drop in Tmax observed here is larger than the changes reported for Bi-doped BaTiO3 with a few % of Bi.

(4) Dielectric Relaxation

As shown in Fig. 4, the dielectric behavior gradually becomes frequency dependent between x=0.02 and 0.05, and strongly dispersive peaks in the real and imaginary permittivities are observed with an increasing concentration of BiScO3. It is thus interesting to consider for what compositions (1−x)BaTiO3xBiScO3 ceramics behave as proper ferroelectrics, relaxor ferroelectrics, or dielectric relaxors. The inverse of the dielectric constant at 1 MHz as a function of temperature was investigated to confirm whether or not the ceramics follows the Curie–Weiss law.24 The 1 MHz data were used here to minimize any space charge contribution to the dielectric constant. Higher temperature measurements (>400°C) were attempted to acquire better fits to the Curie–Weiss law, but could not be used because of increases in the loss tangent around 400°C even at 1 MHz. It was observed that the Curie–Weiss law is obeyed over broad temperature ranges up to x=0.05, and over a more limited range for x=0.1 and 0.2. The Curie constant varies from 1–3 × 105°C for ceramics with x=0–0.2, which is comparable with other ferroelectric materials with displacive phase transitions.24 The limitations at high temperatures prevented determining the deviation temperature in the materials. Nevertheless, the observed agreement with the Curie–Weiss law in the system is characteristic of a proper ferroelectric. Thus, the system should be distinct from the dispersive dielectric behavior in barrier layer capacitor-type materials such as CaCu3Ti4O12.25

Because both dielectric dispersion and deviation from the Curie–Weiss law are characteristic of relaxor ferroelectric materials, this result suggests that the solid solution behaves like a normal ferroelectric from x=0 to about x=0.02 and shows more relaxor-like behavior with a further increase of BiScO3. The trend agrees with the increase in dispersion reported for thin films with increasing BiScO3 concentration.17

Both Arrhenius and Vogel–Fulcher26,27 fits to the frequency dispersion were considered for compositions x=0.1–0.5. The Arrhenius model assumes that a thermally activated process governs the dielectric relaxation. This yields a temperature-dependent relaxation expressed by

  • image(2)

where ω is the measured frequency of the loss tangent maximum, ω0 is related to the attempt jump frequency, Ea is the activation energy, and kB is Boltzmann's constant.

Figure 6(a) shows an Arrhenius plot for x=0.3. The maxima in the measured loss tangent were fitted to an asymmetric sigmoidal function to obtain Tmax, as shown in inset of Fig. 6(a), because it was difficult to determine Tmax values from the real part of dielectric constant. The data from the dielectric loss fit an Arrhenius expression reasonably well over the measured frequency range, although some deviations are apparent (R2∼0.995). However, the attempt jump frequencies obtained for compositions from 0.1 to 0.5 are unrealistic (on the order of 1018–1020 Hz), well beyond the electronic resonance frequency.28

image

Figure 6.  (a) The frequency of the loss tangent maximum as a function of inverse Tmax for a 0.7BaTiO3–0.3BiScO3 sample. The solid line is fitted to an Arrhenius model. Inset is tan δ versus temperature for frequencies from 100 Hz to 1 MHz measured on a 0.7BaTiO3–0.3BiScO3 sample. The symbols are the experimental data; the solid lines are fits using an asymmetric sigmoidal function to obtain Tmax. (b) The inverse frequency of the loss tangent maximum as a function of Tmax for (1−x)BaTiO3xBiScO3 samples with x=0.1–0.5. Solid lines are fits to the Vogel–Fulcher model. (c) Freezing temperature obtained from the fit to the Vogel–Fulcher model for (1−x)BaTiO3xBiScO3 ceramics.

Download figure to PowerPoint

The unphysical results are a consequence of the assumption implicit in the Arrhenius plot that the dipoles become dynamic for all temperatures above absolute zero temperature (0 K). If instead, it is assumed that the dynamics are initiated above a finite freezing temperature, Tf, this gives a dielectric analog of a spin or a dipolar glass. In this case, the dielectric relaxation should follow a Vogel–Fulcher model.29,30 This is expressed by the equation

  • image(3)

where Tf is the freezing temperature. Tf can be thought of as the temperature below which thermal energy is no longer sufficient to permit dipolar cluster dynamics in relaxor ferroelectrics.

Figure 6(b) shows the Vogel–Fulcher plot for compositions x=0.1–0.5. All of the data show excellent fittings (R2>0.999 for all compositions). The key parameters, namely, the activation energy, the attempt frequency, and freezing temperature values obtained from the Vogel–Fulcher fit are listed in Table II. The error bars are plotted using the 95% confidence intervals obtained in the fitting. It is difficult to find a composition trend in either the activation energy or the attempt frequency, but it is clear that the freezing temperature increases as the BiScO3 concentration increases, as shown in Fig. 6(c). Thus, as shown in Fig. 7, both Tf and Tmax increase as a function of the BiScO3 concentration in the pseudocubic structure range.

Table II.    Activation Energy, Attempt Jump Frequency, and Freezing Temperature Obtained from the Fit to the Vogel–Fulcher Model for Various Perovskite Materials
SystemxEa (eV)ωo (Hz)Tf (K)Reference
(1−x)BaTiO3xBiScO3 (this study)0.10.24 ± 0.025.5 × 1013100 ± 6
0.20.26 ± 0.013.2 × 101396 ± 3 
0.30.24 ± 0.019.4 × 1012121 ± 1 
0.40.25 ± 0.011.1 × 1013147 ± 3 
0.50.26 ± 0.012.0 × 1013160 ± 2 
(1−x)PMN–xPT00.0761.0 × 1013220Viehland31
0.070.0613.7 × 1014284 
0.10.0462.4 × 1012296 
0.20.0194.1 × 1012350 
0.250.0173.4 × 1012384 
xBS–(1−x)(PMN−0.53PT)0.220.1041.3 × 1013430Stringer et al.32
(1 mol% excess MgNb2O6)0.320.1111.3 × 1012432 
(Sr1−1.5xBix)TiO30.01330.0361.00 × 101030.7Chen and Yu33
0.05330.0401.73 × 10970.3 
0.1330.0471.05 × 109104.6 
0.20.0253.08 × 108131.9 
Ba(Ti0.7Zr0.3)O30.211.54 × 1010199.6Yu et al.34
Ba(Ti1−xCex)O30.060.00391.45 × 1011374.1Chen et al.35
0.10.00761.32 × 1011320.1 
0.20.0273.21 × 109138.7 
(1−x)BaTiO3xBiScO3 (thin film)0.20.055.3 × 107147Tinberg36
0.40.062.6 × 108175 
0.60.123.4 × 109147 
image

Figure 7.   Tf and Tmax as a function of BiScO3 concentration for (1−x)BaTiO3xBiScO3 ceramics.

Download figure to PowerPoint

Table II also lists the Vogel–Fulcher fitting parameters for other perovskites and BaTiO3–BiScO3 thin films. The attempt frequencies obtained for the BaTiO3–BiScO3 system, ωo∼1013 Hz, compare well with other relaxors. However, the calculated activation energies, Ea∼0.25 eV, are unusually high. Moreover, the values determined here do not correspond to the values obtained in thin films with the same compositions, which are 0.05–0.12 eV.36 The origin of this discrepancy is not known. It should be noted, however, that ωo in films (∼108 Hz) is also anomalously low, possibly as an artifact of the limited frequency range available.

It was previously observed by Stringer et al.32 that Pb(Mg1/3Nb2/3)O3–PbTiO3 with BiScO3 addition (BS–PMN–PT ternary system) shows only a small change in Tf with increasing BiScO3 concentration (0.22–0.32 BiScO3). In that case, however, the PMN concentration was changed at the same time (from 0.25 to 0.15 PMN). Viehland31 demonstrated that Tf decreases with PMN concentration in PMN–PT (see PMN–PT data in Table II). Thus, to account for the small change in Tf reported in the Stringer data,32 BiScO3 would have to increase Tf, which is consistent with the trend observed in BiScO3–BaTiO3 system.

To access the high field dielectric response, the polarization was measured as a function of applied electric field, as shown in Fig. 8. Compositions with x≤0.05 show a ferroelectric hysteresis loop, whereas compositions with x≥0.1 do not show signs of polarization saturation when measured with a maximum applied electric field of 50 kV/cm at room temperature (Fig. 8(a)). Measurements at higher field were performed for a thinner sample with x=0.3 at various temperatures with a maximum applied electric field of about 550 kV/cm. As shown in Fig. 8(b), the polarizations start to saturate at high electric field strengths. There is no clear indication for the existence of a ferroelectric hysteresis loop at 0°C, but the loops open at lower temperatures, suggesting long-range ferroelectricity and relaxor behavior. It should be noted, however, that there is also dielectric loss that contributes to the opening of the loops, which makes it difficult to accurately define the ferroelectricity in the system. It would be useful in the future to monitor the temperature dependence of a property that depends on the remanent polarization.

image

Figure 8.  Polarization versus electric field data for (1−x)BaTiO3xBiScO3 ceramics for (a) x=0.02, 0.05, 0.1, and 0.2 measured with a maximum field of 50 kV/cm and (b) x=0.3 measured at various temperatures with a maximum field of ∼550 kV/cm.

Download figure to PowerPoint

Figure 9 shows the temperature dependence of the remanent polarization, PR, or switchable polarization, Psw, and coercive field, EC, for the 0.7BaTiO3–0.3BiScO3 ceramics obtained from the hysteresis loop in Fig. 8(b) and PUND measurements. The diffuse transition to PR or Psw=0 and EC=0 would be consistent with relaxor ferroelectricity. However, Tf obtained from Vogel–Fulcher analysis for this composition was −150°C, which is about 100°C lower than the Tf estimated from the polarization or coercive field behavior (∼−50°C). The origin for the difference may be, in part, the very different field levels used to probe the sample in the two families of measurements. In thin films, a clearer ferroelectric hysteresis loop was obtained with x=0.4 at lower temperatures, suggesting that these materials are really relaxor ferroelectrics.16 This difference may arise from the fact that much higher electric fields (on the order of MV/cm) can be applied to thin films due to their superior dielectric breakdown strength. Another important finding is that PR, Psw, and EC all decrease at temperatures lower than about −125° to −150°C. Similar behavior was observed in thin films as well.36

image

Figure 9.  (a) Remanent polarization or switchable polarization and (b) coercive field versus temperature data for 0.7BaTiO3–0.3BiScO3 ceramics measured with an applied field of ∼550 kV/cm.

Download figure to PowerPoint

The high activation energies, lower permittivities, and high coercivity implies that it is difficult to obtain long-range dipole alignment in these materials relative to typical complex lead perovskite relaxors such as Pb(Mg1/3Nb2/3)O3 under field-cooled conditions. Therefore it is anticipated that the polar clusters are isolated and frustrated in a matrix that provides only weak coupling between the neighboring clusters. Coupling occurs at low temperatures with very high fields, enabling macroscopic switching of the polarization. It is these relative differences to the classical relaxor that leads us to term these dielectrics “weakly coupled relaxors.” The differences between the classical relaxors, dipolar glasses, and weakly coupled relaxors will be discussed in a future paper. The role of the defect chemistry of the system on the weak coupling should be studied in the future work, particularly given the volatility of the Bi2O3. There could be a wide variety of Bi(Me′, Me″)O3-BaTiO3 dielectrics that show weakly coupled relaxor behavior.37

IV. Conclusions

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

The structural and dielectric properties of bulk (1−x)BaTiO3xBiScO3 (x=0–0.5) ceramics were investigated. The main purposes of the work were to confirm the result from thin films and to acquire a better understanding of the binary system, including determination of the symmetry of the phases, the associated dielectric properties, and the differences in the roles of Bi2O3 and BiScO3 doping in BaTiO3.

The solubility was investigated through the systematic trends in the dielectric data, lattice parameter, and microstructure. The solubility limit for BiScO3 into the BaTiO3 perovskite structure was determined to be about x=0.4, which is much higher than the solubility of Bi alone and in a good agreement with the thin film data (which suggested a limit of x=0.45 without epitaxy). A structural change from tetragonal to pseudocubic was observed at about x=0.05–0.075 at room temperature. The composition for the transition disagrees with the data on films, which showed clear tetragonality even with x=0.20 and 0.30. The origin of this discrepancy is not known, but may be a function of strains associated with the film deposition.

An interesting core–shell feature in microstructure was observed for x=0.03–0.1. It was found the prolonged sintering reduced the number of core–shell grains and the size of cores. However, the inhomogeneity could not be completely eliminated even with 20-h sintering. The EDS results revealed that BiScO3 is enriched in the shell. The process used in this work also resulted in fine-grained material, especially near the BaTiO3 end of the solid solution. Pure BaTiO3 had grains larger than 5 μm (for the sintering condition: 1320°C, 1 h, without Bi2O3 addition), but the composition with a few % BiScO3 showed about a 0.3 μm grain size (1300°C, 1 h, with Bi2O3 addition). The grain size increased slowly with further increase in the BiScO3 concentration; an average grain size of 1 μm (1300°C, 1 h, with Bi2O3 addition) was observed for 10% BiScO3.

Dielectric measurements revealed a gradual change from normal ferroelectric behavior in pure BaTiO3 to a highly diffusive and dispersive relaxor behavior from 10 to 40 mol% BiScO3. Several of the compositions showed high permittivities (approximately 1000) with low-temperature coefficients of capacitance over a wide range of temperature. This result confirmed that the broad, frequency-dependent permittivity maximum reported for thin films is an intrinsic material property, neither a consequence of defects or stresses associated with the thin film preparation, nor a space charge contribution. The Vogel–Fulcher model gave an activation energy of 0.2–0.3 eV for the relaxor behavior range, which was high compared with other relaxors. The attempt frequency was on the order of 1013 Hz and the freezing temperature ranged from −177° to −93°C (96–160 K) with the BiScO3 concentration. It was also found that Tmax increases with increasing BiScO3 concentration in the pseudocubic structure range.

The high coercive fields, low permittivity and remanent polarization, and high activation energies suggest that in the BiScO3–BaTiO3 solid solution, the polarization in nanopolar regions is weakly coupled from region to region, limiting the ability to obtain the metastable long-range dipole ordering in these relaxors under field-cooled conditions.

Acknowledgments

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References

Thanks to the Center for Dielectric Studies for support of this program, and to the National Science Foundation I/UCRC program 0628817. H. O. also wishes to thank Beth Jones and Chris Brink for initial help in the ceramic processing.

References

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Experimental Procedure
  5. III. Results and Discussion
  6. IV. Conclusions
  7. Acknowledgments
  8. References
  • 1
    B. Jaffe, W. R. Cook, and H. Jaffe, Piezoelectric Ceramics. Academic Press, New York, 1971.
  • 2
    D. Hennings, A. Schnell, and G. Simon, “Diffuse Ferroelectric Phase Transitions in Ba(Ti1−yZry)O3 Ceramics,” J. Am. Ceram. Soc., 65 [11] 539544 (1982).
  • 3
    C. A. Randall, S. F. Wang, D. Laubscher, J. P. Dougherty, and W. Huebner, “Structure Property Relationships in Core–Shell BaTiO3–LiF Ceramics,” J. Mater. Res., 8 [4] 8719 (1993).
  • 4
    H. Kishi, N. Kohzu, J. Sugino, H. Ohsato, Y. Iguchi, and T. Okuda, “The Effect of Rare-Earth (La, Sm, Dy, Ho and Er) and Mg on the Microstructure in BaTiO3,” J. Eur. Ceram. Soc., 19 [6–7] 10436 (1999).
  • 5
    H. Kishi, Y. Mizuno, and H. Chazono, “Base-Metal Electrode-Multilayer Ceramic Capacitors: Past, Present and Future Perspectives,” Jpn. J. Appl. Phys., 42 [1] 115 (2003).
  • 6
    S. Lee, Z. Liu, M. Kim, and C. A. Randall, “Influence of Nonstoichiometry on Ferroelectric Phase Transition in BaTiO3,” J. Appl. Phys., 101, 054119 (2007).
  • 7
    C. C. Huang and D. P. Cann, “Phase Transitions and Dielectric Properties in Bi(Zn1/2Ti1/2)O3–BaTiO3 Perovskite Solid Solutions,” J. Appl. Phys., 104, 024117 (2008).
  • 8
    M. E. Lines and A. M. Glass, Principles and Applications of Ferroelectrics and Related Materials. Claredon, Oxford, 1977.
  • 9
    A. J. Moulson and J. M. Herbert, Electroceramics: Materials, Properties, Applications, 2nd edition, Wiley, New York, 2003.
  • 10
    G. I. Skanavi, I. M. Ksendzov, V. A. Trigubenko, and V. G. Prokhvatilov, “Relaxation Polarization and Losses in Nonferroelectric Dielectrics with High Dielectric Constants,” Soviet Phys. JETP, 6 [33] 2509 (1958).
  • 11
    S. V. Bogdanov, K. V. Kiseleva, and V. A. Rassushin, “Effect of Bismuth Impurity on Certain Physical Properties of Barium Titanate Single Crystals,” Soviet Phys.—Crystallogr., 10 [1] 5862 (1965).
  • 12
    L. Zhou, P. M. Vilarinho, and J. L. Baptista, “Solubility of Bismuth Oxide in Barium Titanate,” J. Am. Ceram. Soc., 82 [4] 10646 (1999).
  • 13
    F. Bahri, A. Simon, H. Khemakhem, and J. Ravez, “Classical or Relaxor Ferroelectric Behaviour of Ceramics with Composition Ba1−xBi2/3xTiO3,” Phys. Status Solidi A, 184 [2] 45964 (2001).
  • 14
    I. Burn, “Flux-Sintered BaTiO3 Dielectrics,” J. Mater. Sci., 17 [5] 1398408 (1982).
  • 15
    O. Saburi, “Properties of Semiconductive Barium Titanates,” J. Phys. Soc. Jpn., 14 [9] 115974 (1959).
  • 16
    D. S. Tinberg and S. Trolier-McKinstry, “Structural and Electrical Characterization of xBiScO3–(1−x)BaTiO3,” J. Appl. Phys., 101, 024112 (2007).
  • 17
    R. E. Eitel, C. A. Randall, T. R. Shrout, P. W. Rehrig, W. Hackenberger, and S. Park, “New High Temperature Morphotropic Phase Boundary Piezoelectrics Based on Bi(Me)O3–PbTiO3 Ceramics,” Jpn. J. Appl. Phys., 40 [10] 59996002 (2001).
  • 18
    R. E. Eitel, S. J. Zhang, T. R. Shrout, C. A. Randall, and I. Levin, “Phase Diagram of the Perovskite System (1−x)BiScO3xPbTiO3,” J. Appl. Phys., 96 [5] 282831 (2004).
  • 19
    R. C. Weast, Handbook of Chemistry and Physics, A Ready-Reference Book of Chemical and Physical Data. CRC Press, Boca Raton, FL, 1976.
  • 20
    K. Datta and P. A. Thomas. “In search of Potential Lead Free Materials”, Presentation at 9th European Conference on Applications of Polar Dielectrics, 2008.
  • 21
    L. Vegard, “Die Konstitution der Mischkristalle und die Raumfullung der Atome,” Z. Phys., 5 [1] 1726 (1921).
  • 22
    S. C. Sasaki, T. Prewitt, and R. C. Liebermann, “The Crystal Structure of CaGeO3 Perovskite and the Crystal Chemistry of the GdFeO3-Type Perovskites,” Am. Mineral., 68, 118998 (1983).
  • 23
    D. I. Woodward, I. M. Reaney, R. E. Eitel, and C. A. Randall, “Crystal and Domain Structure of the BiFeO3–PbTiO3 Solid Solution,” J. Appl. Phys., 94 [5] 33138 (2003).
  • 24
    F. Jona and G. Shirane, Ferroelectric Crystals. Dover Publications, New York, 1993.
  • 25
    C. C. Homes, T. Vogt, S. M. Shapiro, S. Wakimoto, and A. P. Ramirez, “Optical Response of High-Dielectric-Constant Perovskite-Related Oxide,” Science, 293 [5530] 6736 (2001).
  • 26
    H. Vogel, “Das Temperature- abhangigkeitsgesetz der Viskositat von Flussigkeiten (The Law of Viscosity Change with Temperature),” Phys. Z., 22, 6456 (1921).
  • 27
    G. Fulcher, “Analysis of Recent Measurements of the Viscosity of Glasses,” Am. Ceram. Soc. J., 8 [6] 33955 (1925).
  • 28
    E. V. Colla, E. Y. Koroleva, N. M. Okuneva, and S. B. Vakhrushev, “Low-Frequency Dielectric Response of PbMg1/3Nb2/3O3,” J. Phys. Condens. Matter, 4, 36717 (1992).
  • 29
    D. Viehland, S. J. Jang, and L. E. Cross, “Freezing of the Polarization Fluctuations in Lead Magnesium Niobate Relaxors,” J. Appl. Phys., 68 [8] 291621 (1990).
  • 30
    A. Chen, Z. Yu, and Z. Jing, “Impurity-Induced Ferroelectric Relaxor Behavior in Quantum Paraelectric SrTiO3 and Ferroelectric BaTiO3,” Phys. Rev. B, 61 [2] 95761 (2000).
  • 31
    D. Viehland, “The Glassy Behavior of Relaxor Ferroelectrics”; Ph.D. Thesis, The Pennsylvania State University, 1992.
  • 32
    C. J. Stringer, T. R. Shrout, and C. A. Randall, “High-Temperature Perovskite Relaxor Ferroelectrics: A Comparative Study,” J. Appl. Phys., 101, 054107 (2007).
  • 33
    A. Chen and Z. Yu, “Dielectric Relaxor and Ferroelectric Relaxor: Bi-Doped Paraelectric SrTiO3,” J. Appl. Phys., 91 [3] 148794 (2002).
  • 34
    Z. Yu, A. Chen, R. Guo, and A. S. Bhalla, “Ferroelectric-Relaxor Behavior of Ba(Ti0.7Zr0.3)O3,” J. Appl. Phys., 92 [5] 26557 (2002).
  • 35
    A. Chen, Z. Jing, and Z Yu, “Ferroelectric Relaxor Ba(Ti,Ce)O3,” J. Phys. Condens. Matter, 14, 890112 (2002).
  • 36
    D. S. Tinberg, “Pulsed Laser Deposition of xBiScO3–(1−x) BaTiO3 Thin Films for Lead Free Dielectric Applications”; M.S. Thesis, The Pennsylvania State University, 2006.
  • 37
    C. C. Huang, D. P. Cann, X. Tan, and N. Vittayakorn, “Phase Transitions and Ferroelectric Properties in BiScO3-Bi(Zn1//2Ti1/2)O3-BaTiO3 Solid Solutions,” J. Appl. Phys., 102 [4] 044103 (2007).