Thin films of cubic pyrochlore bismuth zinc niobate, bismuth zinc tantalate, and bismuth zinc niobate tantalate were fabricated using chemical solution deposition. This family of materials exhibited moderate relative permittivities between 55 ± 2 and 145 ± 5 for bismuth zinc tantalate and bismuth zinc niobate, respectively, and low loss tangents on the order of 0.0008 ± 0.0001. Increases in the concentration of the tantalum end member increased the dielectric breakdown strength. For example, at 10 kHz, the room temperature breakdown strength of bismuth zinc niobate was 5.1 MV/cm, while that of bismuth zinc tantalate was 6.1 MV/cm. This combination of a high breakdown strength and a moderate permittivity led to a high discharged energy storage density for all film compositions. For example, at a measurement frequency of 10 kHz, bismuth zinc niobate exhibited a maximum recoverable energy storage density of 60.8 ± 2.0 J/cm3, while bismuth zinc tantalate exhibited a recoverable energy storage density of 60.7 ± 2.0 J/cm3. Intermediate compositions of bismuth zinc niobate tantalate offered higher energy storage densities; at 10 mol. % tantalum, the maximum recoverable energy storage density was ∼66.9 ± 2.4 J/cm3.

Due to their fast discharge (less than one second), solid state capacitors typically supply high power densities and small energy storage densities compared to batteries or supercapacitors.1,2 Pulsed power capacitors that discharge on the order of several kilohertz are used in a variety of applications,3 ranging from industrial lasers to implantable medical devices, such as heart defibrillators.4,5 Implantable heart defibrillators provide approximately 30 joules of energy to the human heart;6,7 improved energy storage densities of the capacitors in defibrillator pulse generators will allow for the further miniaturization of these devices, which are currently the size of a pocket watch.8 This defibrillator example requires the capacitor to remain charged to allow for immediate, high-frequency discharge upon detection of a heart event, although other power electronics applications, such as pulsed lasers, flash tubes, and military equipment, require constant cycling of the capacitors.9,10

Figure 1 shows the correlation between breakdown strength and relative permittivity for several materials reported to have a high energy storage density.9,11–26 As seen in Figure 1, many materials fall above the historical “best-fit” line,27 primarily due to increases in the breakdown strengths associated with improved processing and/or reduction in dielectric thicknesses. The relative permittivity and, more importantly, breakdown strength are critical materials properties for an energy storage dielectric, as the maximum energy that can be stored by a linear dielectric is calculated using Eq. (1),11 

(1)

in which J is the energy density, E is the electric field sustained by the dielectric, P is the induced polarization of the dielectric, Pmax is the maximum induced polarization, Po is the polarization at zero electric field, εr is the relative permittivity of the material, and εo is the permittivity of free space. It has previously been demonstrated that modest values for the relative permittivity of a material may be compensated for by an enhanced breakdown strength, improving the overall energy storage density of the material.12,13,28,29

FIG. 1.

The maximum breakdown strength as a function of relative permittivity for materials reported to exhibit a high energy storage density.9,11–25 Improvements in the energy storage density are primarily due to increases in the breakdown strength above the historical “best-fit” line.27 The abbreviations are: PLZT, lead lanthanum zirconate titanate; PLZST, lanthanum-doped lead zirconate stannate titanate; PVDF, polyvinylidene fluoride; HFP, hexafluoropropene; CTFE, chlorotrifluoroethylene; PP, polypropylene; BST, barium strontium titanate; PZN-PMN-PT, lead zinc niobate-lead magnesium niobate-lead titanate; BT-BMT-PZ, barium titanate-bismuth magnesium titanate-lead zirconate; BNZ-PT, bismuth nickelate zirconate-lead titanate.

FIG. 1.

The maximum breakdown strength as a function of relative permittivity for materials reported to exhibit a high energy storage density.9,11–25 Improvements in the energy storage density are primarily due to increases in the breakdown strength above the historical “best-fit” line.27 The abbreviations are: PLZT, lead lanthanum zirconate titanate; PLZST, lanthanum-doped lead zirconate stannate titanate; PVDF, polyvinylidene fluoride; HFP, hexafluoropropene; CTFE, chlorotrifluoroethylene; PP, polypropylene; BST, barium strontium titanate; PZN-PMN-PT, lead zinc niobate-lead magnesium niobate-lead titanate; BT-BMT-PZ, barium titanate-bismuth magnesium titanate-lead zirconate; BNZ-PT, bismuth nickelate zirconate-lead titanate.

Close modal

The bismuth pyrochlores are a compositionally tunable family of materials that possess high breakdown strengths and moderate permittivities, in addition to exhibiting low loss tangents, making them candidates for energy storage dielectrics. For example, bismuth zinc niobate, a cubic pyrochlore, has a relative permittivity of 150–200 (Refs. 30–32) and dielectric loss tangents between 0.0005 and 0.005.30,31,33,34 Breakdown strengths for Bi1.5Zn0.9Nb1.5O6.9 have been reported to exceed 5.2 MV/cm.26 It has been demonstrated26 that bismuth zinc niobate thin films prepared using a Pechini-based solution chemistry possess an energy storage density exceeding 60 J/cm3. Further improvement in the energy storage density of Bi1.5Zn0.9Nb1.5O6.9 would be expected if the breakdown strength of the material could be increased.

The pyrochlore crystal structure accommodates a wide variety of substituents, allowing material properties to be varied with composition.35 To enhance the breakdown strength, and, hence, the energy storage capacity of bismuth zinc niobate, it is useful to increase the band gap of the material.36 Tantalum oxide exhibits a band gap of approximately 4 eV,37 while niobium oxide has a band gap of approximately 3.4 eV.38 Above its dielectric relaxation, bismuth zinc tantalate has a relative permittivity of approximately 70 and dielectric losses of approximately 0.005.39–41 Thus, the permittivity of bismuth zinc tantalate is about half that of bismuth zinc niobate, while the losses for the two systems are comparable. It is anticipated that the lower polarizability of the tantalate may be compensated by a higher breakdown strength, enabling improvements in the energy storage density of the deposited thin film. The objective of this paper is to characterize Bi1.5Zn0.9Nb(1.5−x)Ta(x)O6.9 thin films prepared by chemical solution deposition and evaluate their compositional tunability with respect to permittivity and breakdown strength, properties that are critical for their use in dielectric energy storage applications.

Cubic pyrochlore Bi1.5Zn0.9Nb(1.5−x)Ta(x)O6.9 thin films were deposited via spin-coating of chemical solutions. Precursor solutions were synthesized using a modified Pechini method. The cation sources used were bismuth nitrate pentahydrate, zinc acetate dihydrate, niobium ethoxide, and tantalum ethoxide (Sigma-Aldrich). The mole percent of tantalum ethoxide was varied from 0% to 100%. Details of the solution preparation can be found elsewhere; the solutions used to deposit these films used a 3:1 molar ratio of citric acid to metal cation precursor.26 The 0.15 M solution was spin coated at 4000 rpm for 40 s on a platinum-coated silicon wafer (Pt (100 nm)/Ti (20 nm)/SiO2 (500 μm)/Si, NOVA Electronic Materials, Flower Mound, TX). The substrate was dried on a hotplate at 250 °C for 3 min, pre-pyrolyzed on a second hotplate at 350 °C for 10 min, then pyrolyzed in air for 10 min at 400 °C to remove organic species and densify the film. The film was crystallized in a rapid thermal anneal system (RTP-600S, Modular Process Technology Corp., San Jose, CA) for 2 min at 600 °C. This process was repeated several times to increase film thickness. After the deposition of four layers, the film thickness was typically 150 nm.

The phase content of the film was analyzed with an x-ray diffractometer (Empyrean, PANalytical, Almelo, The Netherlands) configured in focusing geometry using Cu Kα radiation. Patterns were collected over a 2θ range from 20° to 73°. The instrument step size was 0.02°, and the scan rate was 2° 2θ per minute.

The band gaps of Bi1.5Zn0.9Nb(1.5−x)Ta(x)O6.9 films on magnesium oxide substrates were analyzed using UV-Vis spectroscopy (Lambda 950, Perkin-Elmer, Waltham, MA) in transmission mode over a wavelength range from 220 to 800 nm. The step size was 1 nm, and the transmission measurements were referenced to a clean, blank MgO single crystal.

To measure the dielectric properties of the films, circular electrodes with diameters ranging from 200 μm to 1 mm were patterned on the film using a double layer lithography process. After patterning, a 500 Å thick layer of platinum was sputtered (CMS-18 Sputter System, Kurt J. Lesker Company, Pittsburgh, PA) onto the film; lift-off processing patterned the top electrodes. The bottom platinum electrode was exposed using a 30% aqueous hydrofluoric acid solution to remove the dielectric film. The film was annealed a final time at 600 °C for 2 min in a rapid thermal annealing system to improve the film-top electrode interface. Film thickness was measured using a profilometer (Alpha-Step 500 Surface Profilometer, Tencor, Portsmouth, NH).

The relative permittivity and loss tangent of the films were measured with an LCR meter (Hewlett-Packard 4284 A Precision, Agilent Technologies, Inc., Palo Alto, CA) at an AC oscillation voltage of 0.03 V over a frequency range from 100 Hz to 1 MHz. Polarization-electric field behavior was measured using a multiferroic analyzer (Precision Multiferroic, Radiant Technologies, Inc., Albuquerque, NM).

Impedance measurements were performed using an impedance analyzer (ModuLab 2100 A, Solartron Analytical, Hampshire, UK) connected to a probe station with a temperature-controllable stage under an oscillation voltage of 1 V. Measurements of capacitance as a function of voltage prior to impedance spectroscopy confirmed that this field was within the range of the linear response for the films. A frequency range of 0.001 Hz–1 MHz was probed on 1 mm electrodes, and spectra were analyzed using Z-View software (Scribner Associates, Southern Pines, NC). The data were fit to an equivalent circuit consisting of a capacitor and resistor in parallel for the grains and a second capacitor and resistor in parallel for the grain boundaries.

Crystalline Bi1.5Zn0.9Nb1.5−xTaxO6.9 films were prepared using the method described above. The diffraction patterns of Bi1.5Zn0.9Nb1.5O6.9 (BZN), Bi1.5Zn0.9Nb1.425Ta0.075O6.9 (BZNT-5), Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (BZNT-10), Bi1.5Zn0.9Nb1.275Ta0.225O6.9 (BZNT-15), and Bi1.5Zn0.9Ta1.5O6.9 (BZT) are shown in Figure 2. The diffraction patterns indicate that the films have a cubic pyrochlore structure (PDF # 04-009-5437, 2002); there are no peaks that correspond to the formation of a secondary phase. Bismuth zinc niobate and bismuth zinc tantalate should produce a substitutional solid solution, in accordance with the Hume-Rothery rules.42 Tantalum and niobium both have an ionic radius of 64 pm,43 an electronegativity of 1.5 and 1.6,44 respectively, and an oxidation state of +5. In this case, given the similarity between the tantalate and niobate end members, it was not possible to use lattice parameters to assess Vegard's law.45–47 

FIG. 2.

X-ray diffraction patterns of Bi1.5Zn0.9Nb1.5O6.9 (BZN), Bi1.5Zn0.9Nb1.425Ta0.075O6.9 (BZNT-5), Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (BZNT-10), Bi1.5Zn0.9Nb1.275Ta0.225O6.9 (BZNT-15), and Bi1.5Zn0.9Ta1.5O6.9 (BZT) films on platinum-coated silicon. Peaks labelled with the abbreviation C.P. correspond to the cubic pyrochlore structure, while peaks labelled with an asterisk (*) are due to the substrate or to diffraction from wavelengths other than Cu Kα.

FIG. 2.

X-ray diffraction patterns of Bi1.5Zn0.9Nb1.5O6.9 (BZN), Bi1.5Zn0.9Nb1.425Ta0.075O6.9 (BZNT-5), Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (BZNT-10), Bi1.5Zn0.9Nb1.275Ta0.225O6.9 (BZNT-15), and Bi1.5Zn0.9Ta1.5O6.9 (BZT) films on platinum-coated silicon. Peaks labelled with the abbreviation C.P. correspond to the cubic pyrochlore structure, while peaks labelled with an asterisk (*) are due to the substrate or to diffraction from wavelengths other than Cu Kα.

Close modal

Figure 3 shows the direct band gap Tauc plot48 of Bi1.5Zn0.9Nb1.5O6.9 (dashed line), Bi1.5Zn0.9Ta1.5O6.9 (dotted line), and Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (solid line) films deposited on magnesium oxide substrates. To ensure that small errors (in the background subtraction or absorbance) due to small differences in film thickness do not perturb the analysis, the maximum absorbance numbers were normalized to the same value. Therefore, the observed shift in absorbance onset is due to a change in band gap. As seen in Figure 3, the band gap of bismuth zinc niobate was 3.72 ± 0.06 eV, while that of bismuth zinc tantalate was 3.88 ± 0.04 eV. The band gap for the solid solution sample shown fell between the magnitudes of the band gaps of the end members. All three compositions exhibited direct band gaps, which is consistent with the type of transition reported for other bismuth-based pyrochlores.49,50 The observed shift in band gap was reproducible across the surface of the film and across multiple films of these compositions.

FIG. 3.

The normalized Tauc plots of Bi1.5Zn0.9Nb1.5O6.9 (dashed line), Bi1.5Zn0.9Ta1.5O6.9 (dotted line), and Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (solid line) on a magnesium oxide substrate. The shift in the onset of absorbance indicates that modification with tantalum increased the band gap of the material. The abbreviations on the y-axis are α, absorption coefficient; h, Planck's constant; ν, frequency.

FIG. 3.

The normalized Tauc plots of Bi1.5Zn0.9Nb1.5O6.9 (dashed line), Bi1.5Zn0.9Ta1.5O6.9 (dotted line), and Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (solid line) on a magnesium oxide substrate. The shift in the onset of absorbance indicates that modification with tantalum increased the band gap of the material. The abbreviations on the y-axis are α, absorption coefficient; h, Planck's constant; ν, frequency.

Close modal

To determine whether the increase in band gap due to substitution of the niobium end member for tantalum correlates to an increase in the breakdown strength of the material, a film of bismuth zinc tantalate was electrically characterized. This film exhibited a room-temperature relative permittivity of 55 ± 2 and a loss tangent of 0.0004 ± 0.0001. The measured permittivity value is lower than previously reported values for bulk ceramics, and the loss tangent is approximately an order of magnitude lower than previous reports in the literature.39–41 The tunability of bismuth zinc tantalate was approximately an order of magnitude lower than that of bismuth zinc niobate. At DC fields of 1.5 MV/cm and an AC oscillation voltage of 0.03 V, bismuth zinc tantalate exhibited a tunability of 4.5%; under identical field and measurement conditions, bismuth zinc niobate exhibited a tunability of 24%. As anticipated, the breakdown strength of the tantalate films was higher than that of the bismuth zinc niobate films, and the high field loss tangent remained low. Weibull plots were used to determine the breakdown strength of these films using bipolar polarization-electric field hysteresis loops. The values of the Weibull parameters are given in Table I. The Weibull parameter for each condition was calculated using a total of ten electrodes. The large Weibull parameter values indicate that the breakdown strengths exhibited little variation from electrode to electrode. Figure 4 shows the maximum 10 kHz electric field sustained by the film prior to breakdown. The 10 kHz breakdown field of this film, which is listed in Table I, was 6.1 MV/cm. At 1 kHz, the breakdown strength was 5.5 MV/cm. The combination of an improved breakdown strength and lower high field loss correlates well with an increase in the band gap of the material.

TABLE I.

Weibull parameters for the breakdown strength of Bi1.5Zn0.9Ta1.5O6.9.

1 kHz10 kHz
Breakdown strength (MV/cm) 5.5 ± 0.1 6.1 ± 0.1 
Weibull parameter 34.0 ± 2.4 95.1 ± 6.0 
1 kHz10 kHz
Breakdown strength (MV/cm) 5.5 ± 0.1 6.1 ± 0.1 
Weibull parameter 34.0 ± 2.4 95.1 ± 6.0 
FIG. 4.

The polarization-electric field behavior at 10 kHz for films of Bi1.5Zn0.9Ta1.5O6.9.

FIG. 4.

The polarization-electric field behavior at 10 kHz for films of Bi1.5Zn0.9Ta1.5O6.9.

Close modal

The maximum discharged energy storage density for the bismuth zinc tantalate films was 60.7 ± 2.0 J/cm3 at 10 kHz, which is comparable with the maximum discharged energy storage density of bismuth zinc niobate films.26 This is the discharged energy storage density calculated using the integral in Eq. (1) (i.e., losses were excluded and tunability was fully accounted for). For this case, any improvement in the energy storage density of the tantalate composition due to the enhancement in breakdown strength, reduced tunability, and lower high field loss is offset by the reduction in relative permittivity of the bismuth zinc tantalate. However, some applications require that the loss of the material remain below 2%;51 in many materials targeting these applications, the maximum energy storage density is limited by the loss of the film, rather than the breakdown strength. Given this constraint, bismuth zinc tantalate exhibited low high field losses across a range of temperatures and frequencies. As seen in Figure 5, for most temperatures and frequencies, the films can be subjected to their maximum field prior to breakdown without the loss tangent exceeding 2%, resulting in high efficiencies for the material.

FIG. 5.

The energy storage density and efficiency of Bi1.5Zn0.9Ta1.5O6.9. Energy storage density data points connected by solid lines represent the maximum energy storage density of the material, while data points connected by a dashed line represent the energy storage density of Bi1.5Zn0.2Ta1.5O6.9 when loss values are required to remain below 2%.

FIG. 5.

The energy storage density and efficiency of Bi1.5Zn0.9Ta1.5O6.9. Energy storage density data points connected by solid lines represent the maximum energy storage density of the material, while data points connected by a dashed line represent the energy storage density of Bi1.5Zn0.2Ta1.5O6.9 when loss values are required to remain below 2%.

Close modal

To determine if any intermediate tantalum concentrations will enhance the energy storage density of the films, the tantalum concentration was varied from 0% to 15%. The energy storage density as a function of tantalum concentration was investigated, and the results are summarized in Table II. As the concentration of tantalum increased, the permittivity of the films decreased, as expected. Further, as the concentration of tantalum was increased, the maximum field sustained by the films increased. A concentration of 10% tantalum was determined to be optimal, as the energy storage density was maximized at this tantalum concentration.

TABLE II.

Energy storage density of Bi1.5Zn0.9Nb(1.5−x)Ta(x)O6.9 as a function of tantalum concentration.

BZNBZN with 5% TaBZN with 10% TaBZN with 15% TaBZT
Permittivity 145 ± 5 130 ± 5 122 ± 4 108 ± 4 55 ± 2 
1 kHz maximum field (MV/cm) 4.7 4.9 5.0 5.1 5.5 
10 kHz maximum field (MV/cm) 5.1 5.4 5.5 5.7 6.1 
1 kHz energy storage (J/cm347.1 ± 1.7 46.7 ± 1.6 54.1 ± 1.9 49.6 ± 1.8 49.6 ± 1.7 
10 kHz energy storage (J/cm360.8 ± 2.0 61.7 ± 2.2 66.9 ± 2.4 65.0 ± 2.3 60.7 ± 2.0 
BZNBZN with 5% TaBZN with 10% TaBZN with 15% TaBZT
Permittivity 145 ± 5 130 ± 5 122 ± 4 108 ± 4 55 ± 2 
1 kHz maximum field (MV/cm) 4.7 4.9 5.0 5.1 5.5 
10 kHz maximum field (MV/cm) 5.1 5.4 5.5 5.7 6.1 
1 kHz energy storage (J/cm347.1 ± 1.7 46.7 ± 1.6 54.1 ± 1.9 49.6 ± 1.8 49.6 ± 1.7 
10 kHz energy storage (J/cm360.8 ± 2.0 61.7 ± 2.2 66.9 ± 2.4 65.0 ± 2.3 60.7 ± 2.0 

As an example of tantalum modification leading to an enhancement in the energy storage density, Figure 6 shows the high-field polarization versus electric field behavior for bismuth zinc niobate with 10% tantalum at a measurement frequency of 10 kHz; the area used to calculate the energy storage density is shaded. The low field relative permittivity of the films was 122 and the loss tangent was 0.0008. The permittivity values of the tantalum-modified films are somewhat lower than those predicted by dielectric mixing laws, which may be due, in part, to modest amounts of porosity in the deposited films. The loss tangents of the solid solutions are of the same order of magnitude as the end member compositions. For the optimized composition, Bi1.5Zn0.9Nb1.35Ta0.15O6.9, and a measurement frequency of 10 kHz, the material sustained an electric field of 5.5 MV/cm. At a lower measurement frequency of 1 kHz, the films sustained fields of 5.0 MV/cm. The films reached catastrophic electrical breakdown at applied fields of 5.5 MV/cm at 10 Hz. This breakdown strength exceeds those reported in the literature for Bi1.5Zn0.9Nb1.5O6.9 by approximately 0.3 ± 0.1 MV/cm.26 

FIG. 6.

The polarization-electric field behavior for films of Bi1.5Zn0.9Nb1.35Ta0.15O6.9. The shaded area was used to calculate the energy storage density.

FIG. 6.

The polarization-electric field behavior for films of Bi1.5Zn0.9Nb1.35Ta0.15O6.9. The shaded area was used to calculate the energy storage density.

Close modal

The high breakdown strengths for these films were attributable, at least in part, to a fine-grained microstructure. As seen in Figure 7(a), the films exhibited small grains, on the order of several tens of nanometers in lateral size and thickness. In dielectric energy storage applications, smaller grains may be advantageous. Grain boundaries in oxide thin films are often resistive;52–54 thus, the presence of many grain boundaries likely impedes conduction through the film, increasing breakdown strength. Impedance spectroscopy revealed that the grain boundaries had a much higher resistivity than the interior of the grain. Figure 7(b) shows the impedance spectra of a Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film measured from 180 °C to 200 °C; the fits to the data are indicated by black lines. The spectra show two overlapping semicircles indicative of grain and grain boundary resistances.55,56 The modeled resistivity values for the grain boundary and grain interior are given in Table III. The corresponding conductivities exhibited Arrhenius behavior, as shown in Figure 7(c), allowing for the activation energies to be calculated. The activation energy for the interior of the grain was 0.20 ± 0.002 eV, while the activation energy for grain boundary conduction was 1.12 ± 0.003 eV. The resistivity values for the grain boundary are two orders of magnitude higher than the resistivity values for the grain interior, confirming that resistive grain boundaries likely impede conduction through the film, contributing to the high breakdown strengths observed here.

FIG. 7.

(a) SEM cross section of the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film used for impedance spectroscopy. (b) The impedance spectra of the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film measured from 180 °C to 200 °C. The inset high-frequency data are not shown on equivalent axes to capture the relatively smaller grain interior contribution. (c) The Arrhenius plot for the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film used to extract activation energies for conduction.

FIG. 7.

(a) SEM cross section of the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film used for impedance spectroscopy. (b) The impedance spectra of the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film measured from 180 °C to 200 °C. The inset high-frequency data are not shown on equivalent axes to capture the relatively smaller grain interior contribution. (c) The Arrhenius plot for the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 film used to extract activation energies for conduction.

Close modal
TABLE III.

Resistivity values of Bi1.5Zn0.9Nb1.35Ta0.15O6.9 films.

Temperature (°C)Resistivity of the grain boundaries (Ω m)Resistivity of the grain interior (Ω m)
180 2.90 × 1010 ± 1.6 × 108 1.83 × 108 ± 6.9 × 106 
190 1.56 × 1010 ± 9.3 × 107 1.63 × 108 ± 1.7 × 107 
200 8.6 × 109 ± 5.2 × 107 1.40 × 108 ± 2.6 × 107 
Temperature (°C)Resistivity of the grain boundaries (Ω m)Resistivity of the grain interior (Ω m)
180 2.90 × 1010 ± 1.6 × 108 1.83 × 108 ± 6.9 × 106 
190 1.56 × 1010 ± 9.3 × 107 1.63 × 108 ± 1.7 × 107 
200 8.6 × 109 ± 5.2 × 107 1.40 × 108 ± 2.6 × 107 

The high breakdown strength of tantalum-modified bismuth zinc niobate contributes to a high energy storage density. Figure 8 shows the maximum storage density achieved for the bismuth zinc niobate tantalate films as a function of temperature and measurement frequency; these data points are connected by a solid line. These energy storage densities exceed those of bismuth zinc niobate and bismuth zinc tantalate, indicating that bismuth zinc niobate tantalate may be suitable for dielectric energy storage, particularly in pulsed power applications. The maximum energy storage density was ∼66.9 ± 2.4 J/cm3 for measurement frequencies of 10 kHz, while for a measurement frequency of 1 kHz, the maximum recoverable energy density was ∼54.1 ± 1.9 J/cm3.

FIG. 8.

The energy storage density of Bi1.5Zn0.9Nb1.35Ta0.15O6.9. Data points connected by solid lines represent the maximum energy storage density of the material, while data points connected by a dashed line show the energy storage density of Bi1.5Zn0.2Nb1.35Ta0.15O6.9 when loss values are required to remain below 2%.

FIG. 8.

The energy storage density of Bi1.5Zn0.9Nb1.35Ta0.15O6.9. Data points connected by solid lines represent the maximum energy storage density of the material, while data points connected by a dashed line show the energy storage density of Bi1.5Zn0.2Nb1.35Ta0.15O6.9 when loss values are required to remain below 2%.

Close modal

The dashed lines in Figure 8 show the energy storage density of the bismuth zinc niobate tantalate films recalculated with the maximum fields reduced to keep the high field loss tangent below 0.02. This loss metric may be important for power electronics applications that require constant cycling of the capacitors.9,10 Although the energy storage densities under these loss conditions are reduced, the energy storage values are still high for all measurement frequencies. The larger high field losses of the BZNT (manifested by the larger area enclosed by the polarization-electric field hysteresis loop) reduced the energy storage density that could be obtained under 2% loss restrictions relative to the energy storage density for BZT films under identical conditions. Therefore, in applications requiring 2% loss tangent restrictions, BZT may be better suited than BZNT, despite the larger maximum discharged energy density of the BZNT composition. The electric field at which the loss tangent for each composition surpasses 2% is given in Table IV. As can be seen in Table IV, the BZT films could be subjected to higher electric fields than the BZNT films prior to the loss tangent exceeding 0.02.

TABLE IV.

10 kHz electric field at which loss tangent exceeds 0.02.

Temperature (°C)Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (MV/cm)Bi1.5Zn0.9Ta1.5O6.9 (MV/cm)
25 3.8 5.8 
50 3.6 5.6 
100 3.2 5.3 
150 2.7 4.5 
200 2.5 3.9 
Temperature (°C)Bi1.5Zn0.9Nb1.35Ta0.15O6.9 (MV/cm)Bi1.5Zn0.9Ta1.5O6.9 (MV/cm)
25 3.8 5.8 
50 3.6 5.6 
100 3.2 5.3 
150 2.7 4.5 
200 2.5 3.9 

As the temperature is increased, the maximum discharged energy storage density of the films remains high, and exceeds 20 J/cm3 at 200 °C. This is comparable to the maximum energy storage density reported for BZN films and borosilicate glass at 200 °C.26,57 The reduction in the energy density of bismuth zinc niobate tantalate at elevated temperatures is due to a decrease in the breakdown strength, as well as an increase in the loss tangent, of the material at higher temperatures. As an example, at 200 °C and a measurement frequency of 10 kHz, the breakdown field of the material is reduced to 4.0 MV/cm.

The maximum delivered power density for a linear dielectric is related to the energy storage density using Eq. (2),57 

(2)

in which P is the maximum power density, f is the measurement frequency, and tanδ is the loss tangent for the electric field at which J, the energy storage density, is calculated. The high field loss was calculated using Eq. (3),57 

(3)

where we is the hysteresis, which was defined as the area enclosed by the positive field polarization-electric field loop. The power densities for bismuth zinc niobate tantalate are listed in Table V.

TABLE V.

Power storage density of Bi1.5Zn0.9Nb1.35Ta0.15O6.9.

Temperature (°C)Power storage density at 10 kHz (MW/cm3)Power storage density at 1 kHz (MW/cm3)
25 850 ± 35 45 ± 2 
50 720 ± 30 40 ± 2 
100 710 ± 30 25 ± 1 
150 700 ± 30 8 ± 0.5 
200 530 ± 20 3 ± 0.5 
Temperature (°C)Power storage density at 10 kHz (MW/cm3)Power storage density at 1 kHz (MW/cm3)
25 850 ± 35 45 ± 2 
50 720 ± 30 40 ± 2 
100 710 ± 30 25 ± 1 
150 700 ± 30 8 ± 0.5 
200 530 ± 20 3 ± 0.5 

Despite the decrease in the relative permittivity of bismuth zinc niobate upon modification with tantalum, the energy storage density of the material is improved due to the increase in the breakdown strength achieved by compositionally tuning the band gap. The reproducible energy storage density of Bi1.5Zn0.9Nb1.35Ta0.15O6.9 exceeds that of Bi1.5Zn0.9Nb1.5O6.9.26 High energy storage density values are maintained at elevated temperatures up to 200 °C.

Compositional tuning of bismuth zinc niobate with tantalum was pursued as a method to increase the band gap, and, consequently the breakdown strength, of this substitutional solid solution for energy storage applications. This family of materials exhibited moderate relative permittivities of between 55 ± 2 and 145 ± 5, for bismuth zinc tantalate and bismuth zinc niobate, respectively, and low loss tangents on the order of 0.0008. Increases in the tantalum concentration of the solid solution increased the dielectric breakdown strength. At an optimal tantalum concentration of ten mole percent, the maximum field sustained by the films at 10 kHz was 5.5 MV/cm, and the permittivity was 122. This improvement in the breakdown strength compensated for the reduction in the relative permittivity of material, which led to a high discharged energy storage density. At a measurement frequency of 10 kHz, the Bi1.5Zn0.9Nb1.35Ta0.15O6.9 films exhibited a maximum recoverable energy storage density of 66.9 ± 2.4 J/cm3, and at 1 kHz, the recoverable energy storage density was 54.1 ± 1.9 J/cm3.

The authors would like to thank The Dow Chemical Company for funding this research. This material is also based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE1255832.

1.
J.
Sowash
,
Am. Ceram. Soc. Bull.
88
,
19
(
2009
).
2.
M.
Winter
and
R. J.
Brodd
,
Chem. Rev.
104
,
4245
(
2004
).
3.
J. D.
van Wyk
, in
Proceedings of the Power Electronics and Motion Control Conference
(
2000
), Vol. 1, p.
9
.
4.
M. V.
Fazio
and
H. C.
Kirbie
,
Proc. IEEE
92
,
1197
(
2004
).
5.
M.
Glikson
and
P. A.
Friedman
,
Lancet
357
,
1107
(
2001
).
6.
D.
Graham-Rowe
, MIT Technology Review, 26 November
2008
.
7.
D. M.
Zogbi
, Passive Component Magazine, 30 October
2001
.
8.
American Heart Association
, see http://www.heart.org/HEARTORG/.
9.
H.
Tang
and
H. A.
Sodano
,
Nano Lett.
13
,
1373
(
2013
).
10.
F.
MacDougall
,
J.
Ennis
,
X. H.
Yang
,
K.
Seal
,
S.
Phatak
,
B.
Spinks
,
N.
Keller
,
C.
Naruo
, and
T. R.
Jow
,
IEEE International Pulsed Power Conference
(
2005
), p.
1215
.
11.
N. H.
Fletcher
,
A. D.
Hilton
, and
B. W.
Ricketts
,
J. Phys. D: Appl. Phys.
29
,
253
(
1996
).
12.
E. K.
Michael
and
S.
Trolier-McKinstry
,
J. Ceram. Soc. Jpn.
122
,
250
(
2014
).
13.
L. A.
Fredin
,
Z.
Li
,
M. T.
Lanagan
,
M. A.
Ratner
, and
T. J.
Marks
,
Adv. Funct. Mater.
23
,
3560
(
2013
).
14.
X.
Zhou
,
X.
Zhao
,
Z.
Suo
,
C.
Zou
,
J.
Runt
,
S.
Liu
,
S.
Zhang
, and
Q. M.
Zhang
,
Appl. Phys. Lett.
94
,
162901
(
2009
).
15.
A. V.
Wagner
,
G. W.
Johnson
, and
T. J. W.
Barbee
(
Mater. Res. Soc. Symp. Proc.
,
1999
), Vol.
574
, p.
219
.
16.
Z.
Xie
,
Z.
Yue
,
G.
Ruehl
,
B.
Peng
,
J.
Zhang
,
Q.
Yu
,
X.
Zhang
, and
L.
Li
,
Appl. Phys. Lett.
104
,
243902
(
2014
).
17.
T.
Chen
,
J.
Wang
,
X.
Zhong
,
F.
Wang
,
B.
Li
, and
Y.
Zhou
,
Ceram. Int.
40
,
5327
(
2014
).
18.
J.
Luo
,
J.
Du
,
Q.
Tang
, and
C.
Mao
,
IEEE Trans. Electron Devices
55
,
3549
(
2008
).
19.
M. S.
Mirshekarloo
,
K.
Yao
, and
T.
Sritharan
,
Appl. Phys. Lett.
97
,
142902
(
2010
).
20.
N. J.
Smith
,
B.
Rangarajan
,
M. T.
Lanagan
, and
C. G.
Pantano
,
Mater. Lett.
63
,
1245
(
2009
).
21.
U.
Balachandran
,
M.
Narayanan
,
S.
Liu
, and
B.
Ma
,
Jpn. J. Appl. Phys., Part 1
52
,
05DA10
(
2013
).
22.
H.
Ogihara
,
C. A.
Randall
, and
S.
Trolier-McKinstry
,
J. Am. Ceram. Soc.
92
,
1719
(
2009
).
23.
G.
Sethi
,
M.
Olszta
,
J.
Li
,
J.
Sloppy
,
M. W.
Horn
,
E. C.
Dickey
, and
M. T.
Lanagan
,
2007 Annual Report Conference on Electrical Insulation and Dielectric Phenomena
(
2007
), p.
815
.
24.
P.
Khanchaitit
,
K.
Han
,
M. R.
Gadinski
, and
Q.
Wang
,
Nat. Commun.
4
,
2845
(
2013
).
25.
S.
Chao
and
F.
Dogan
,
Int. J. Appl. Ceram. Technol.
8
,
1363
(
2011
).
26.
E. K.
Michael
and
S.
Trolier-McKinstry
,
J. Am. Ceram. Soc.
98
,
1223
(
2015
).
27.
J. W.
McPherson
,
J.
Kim
,
A.
Shanware
,
H.
Mogul
, and
J.
Rodriguez
,
IEEE Trans. Electron Devices
50
,
1771
(
2003
).
28.
V.
Tomer
,
C. A.
Randall
,
G.
Polizos
,
J.
Kostelnick
, and
E.
Manias
,
J. Appl. Phys.
103
,
034115
(
2008
).
29.
P.
Barber
,
S.
Balasubramanian
,
Y.
Anguchamy
,
S.
Gong
,
A.
Wibowo
,
H.
Gao
,
H. J.
Ploehn
, and
H.-C.
zur Loye
,
Materials
2
,
1697
(
2009
).
30.
W.
Ren
,
S.
Trolier-McKinstry
,
C.
Randall
, and
T. R.
Shrout
,
J. Appl. Phys.
89
,
767
(
2001
).
31.
J.
Lu
and
S.
Stemmer
,
Appl. Phys. Lett.
83
,
2411
(
2003
).
32.
X.
Zhang
,
W.
Ren
,
P.
Shi
,
X.
Wu
,
X.
Chen
, and
X.
Yao
,
J. Alloys Compd.
553
,
8
(
2013
).
33.
J.
Park
,
J.
Lu
,
S.
Stemmer
, and
R. A.
York
,
J. Appl. Phys.
97
,
084110
(
2005
).
34.
R. L.
Thayer
,
C. A.
Randall
, and
S.
Trolier-McKinstry
,
J. Appl. Phys.
94
,
1941
(
2003
).
35.
M. A.
Subramanian
,
G.
Aravamudan
, and
G. V.
Subba Rao
,
Prog. Solid State Chem.
15
,
55
(
1983
).
36.
A.
Yoshikawa
,
H.
Matsunami
, and
Y.
Nanishi
,
Wide Bandgap Semiconductors: Fundamental Properties and Modern Photonic and Electronic Devices
(
Springer
,
New York
,
2007
), p.
1
.
37.
E.
Franke
,
C. L.
Trimble
,
M. J.
DeVries
,
J. A.
Woollam
,
M.
Schubert
, and
F.
Frost
,
J. Appl. Phys.
88
,
5166
(
2000
).
38.
K.
Yoshimura
,
T.
Miki
,
S.
Iwama
, and
S.
Tanemura
,
Thin Solid Films
281–282
,
235
(
1996
).
39.
D. P.
Cann
,
C. A.
Randall
, and
T. R.
Shrout
,
Solid State Commun.
100
,
529
(
1996
).
40.
H.-J.
Youn
,
C.
Randall
,
A.
Chen
,
T.
Shrout
, and
M. T.
Lanagan
,
J. Mater. Res.
17
,
1502
(
2002
).
41.
H.-J.
Youn
,
T.
Sogabe
,
C. A.
Randall
,
T. R.
Shrout
, and
M. T.
Lanagan
,
J. Am. Ceram. Soc.
84
,
2557
(
2001
).
42.
W. D.
Callister
and
D. G.
Rethwisch
,
Materials Science and Engineering
(
Wiley
,
2011
).
43.
R. D.
Shannon
,
Acta Crystallogr., Sect. A
32
,
751
(
1976
).
44.
L.
Pauling
,
The Nature of the Chemical Bond
(
Cornell University Press
,
1960
).
45.
I.
Levin
,
T. G.
Amos
,
J. C.
Nino
,
T. A.
Vanderah
,
I. M.
Reaney
,
C. A.
Randall
, and
M. T.
Lanagan
,
J. Mater. Res.
17
,
1406
(
2002
).
46.
C. C.
Khaw
,
K. B.
Tan
,
C. K.
Lee
, and
A. R.
West
,
J. Eur. Ceram. Soc.
32
,
671
(
2012
).
47.
I.
Levin
,
T. G.
Amos
,
J. C.
Nino
,
T. A.
Vanderah
,
C. A.
Randall
, and
M. T.
Lanagan
,
J. Solid State Chem.
168
,
69
(
2002
).
48.
J.
Pankove
,
Optical Processes in Semiconductors
(
Cornell University Press
,
1971
).
49.
S. M.
Zanetti
,
S. A.
da Silva
, and
G. P.
Thim
,
J. Solid State Chem.
177
,
4546
(
2004
).
50.
G.
Perenlei
,
P. C.
Talbot
, and
W. N.
Martens
,
J. Nanomater.
2014
,
695973
.
51.
S.
Islam
, “
Polymer nano-dielectrics for high density energy storage
,” Ph.D. thesis (
University of South Carolina
,
Columbia
,
2014
).
52.
R. E.
Newnham
,
Rep. Prog. Phys.
52
,
123
(
1989
).
53.
54.
J.
Liebault
,
J.
Vallayer
,
D.
Goeuriot
,
D.
Treheux
, and
F.
Thevenot
,
J. Eur. Ceram. Soc.
21
,
389
(
2001
).
55.
E.
Barsoukov
and
J. R.
Macdonald
,
Impedance Spectroscopy: Theory, Experiment, and Applications
(
Wiley
,
2005
).
56.
J. T.
Irvine
,
D. T.
Sinclair
, and
A. R.
West
,
Adv. Mater.
2
,
132
(
1990
).
57.
M. P.
Manoharan
,
C.
Zou
,
E.
Furman
,
N.
Zhang
,
D. I.
Kushner
,
S.
Zhang
,
T.
Murata
, and
M. T.
Lanagan
,
Energy Technol.
1
,
313
(
2013
).