Dielectric thin film capacitors, storing large charge density, are useful in electric energy storing devices. Highly oriented lead-free BaZr0.20Ti0.80O3 (BZT20) thin films were grown on a conducting bottom layer La0.7Sr0.3MnO3 deposited on a MgO (100) substrate under an oxygen atmosphere using a pulsed laser deposition technique. X-ray diffraction studies indicate that BZT20 films were stabilized in a (100) oriented tetragonal phase. Microstructural studies on thin films indicate a smooth film (a roughness of ∼1.25 nm) with a thickness of around 320 nm. The structural sensitive A1(TO2) Raman band exhibits a discontinuous change across the tetragonal-cubic phase transition temperature Tc ∼ 275 K. The appearance of the broad Raman band in the cubic (Pm−3m) phase at an elevated temperature suggests the activation of symmetry forbidden Raman active bands. The temperature dependent band frequency and integrated intensity of the structural sensitive A1(TO2) band show anomaly across Tc. Temperature dependent dielectric studies (100–650 K) carried out in a wide range of frequencies 102–106 Hz on a fabricated Pt/BZT20/LSMO metal-insulator-metal capacitor suggest a broad dispersive peak of around 290 K. The polarization relaxation follows the Vogel-Fulcher relation with an activation energy of Ea = 0.047 eV and a freezing temperature of Tf = 246 K. The slim polarization P-E loops with a remanent polarization of ∼89.6 μC/cm2 and an EC value of ∼0.29 MV/cm were observed, suggesting its local ferroelectric ordering in corroboration with Raman and dielectric findings. From the P-E loop analysis, a large energy storage density of 31.9 J/cm3 and an energy storage efficiency of 56% were obtained. Our experimental results revealed that the BZT20 thin film capacitors have potential for energy storage device applications.

Perovskite-type (ABO3) relaxor ferroelectric materials are of current research interest because of their potential applications in a high energy storage capacitor and power generation electronic devices.1–4 Relaxors are characterized by their complex behavior with high dielectric constant, frequency and temperature dependent diffuse dielectric peaks, and high breakdown electric field.5 The central feature for the complex behavior in relaxors is based on the nucleation and dynamics of polar nanoregions (PNRs) dispersed in a nonpolar matrix.6–9 In general, compositional fluctuation or chemical strain in mixed-oxide perovskites give rise to the formation of PNRs and their dynamical evolution governs their physical behavior. Relaxors are manifested by a slim-electric polarization loop, frequency dispersive broad dielectric peaks, and broad phonon spectra.6,10,11 The slim polarization hysteresis loop of relaxors results in a larger energy storage area and higher recoverable energy.12,13 The relaxor in thin film form, in contrast to the bulk, has one dimension being at a nanoscale and can store maximum energy density Ust = ɛoɛrE2BD/2. Since the material in thin film form has a high dielectric constant and a larger breakdown electric field EBD,12–14 it is favorable toward achieving high energy storage capacity. Lead based relaxor ferroelectric thin films have been extensively studied to examine their relaxor behavior and robust charge storage density capacity. Highly oriented (100) La3+ and Sc3+ doped Pb(Zr0.53Ti0.47)O3 thin films, grown by the pulsed laser deposition (PLD) method, exhibit local polar ordering and can have a better recoverable energy density of ∼19 J/cm3 with 66% of efficiency.15 Only Sc3+ doped Pb(Zr0.53Ti0.47)O3 relaxor thin films were reported to store an excellent energy density of ∼54 J/cm3 with a storage efficiency of ∼70%.16 However, from the point of view of environmental friendly lead-free compounds, as an alternative to replacing Pb-based ferroelectric capacitors, a lot of attention has been devoted toward BaTiO3(BTO)-based relaxors.3,13,17 Isovalent zirconium (ionic radius 0.84 Å) substituted in place of a Ti-cation (an ionic radius of 0.605 Å) of comparatively smaller size in BTO form a single phase compound.18 BaTi1−xZrxO3 relaxor ferroelectrics, belonging to the perovskite structure, are known to possess interesting dielectric and ferroelectric behavior.3,11–18 The A-site of the unit cell is occupied by a Ba2+ cation, and the B-site is randomly occupied by Ti4+ and Zr4+ cations. In regard to the individual components of the mixed solid solutions, BaTiO3 is a classical tetragonal ferroelectric, whereas BaZrO3 is a nonpolar cubic dielectric. However, their mixed solid solution stabilizes in the tetragonal relaxor phase at room temperature. Several works on BaTi1−xZrxO3 have mainly concentrated on dielectric tunability,19,20 compositional, and temperature dependent phase diagram.11,18 Pulsed laser deposited BaTi1−xZrxO3 thin films, for x = 0.3–0.5,13 of thickness around 400 nm, grown on the MgO substrate were found to exhibit relaxor behavior with less dielectric loss and high dielectric constant. Among these films, x = 0.3, the lowest dopant Zr films exhibit ultrahigh recoverable energy density of ∼156 J/cm3 with an efficiency of 72%. In regard to BaZr0.2Ti0.8O3 thin films fabricated using a radio frequency magnetic sputtering process on the Nb,SrTiO3 (001) substrate achieved a large energy storage density of 30.4 J/cm3 with a robust thermal stability.3 Ferroelectric-dielectric multilayered thin film capacitors fabricated through the insertion of a dielectric HfO2:Al2O3 layer in the Pt/0.5Ba(Zr0.2Ti0.8)O3-05(Ba0.7Ca0.3)TiO3/Au structure showed an impressive energy storage density of 99.8 J/cm3 and an efficiency of 71% at an applied field of 750 kV/cm.21 Several studies22 on linear dielectrics, ferroelectrics, and relaxor ferroelectrics ceramic films were carried out to examine their high power density and ultrashort discharge time for possible application in pulsed power electronics. (Ba0.955Ca0.045) (Zr0.17Ti0.83)O3 films grown on a conducting buffer layer deposited onto the MgO(100) substrate using the PLD process is reported to exhibit a high recoverable energy density of 39.11 J/cm3 with a poor thermal stability.17 Precisely, little research has been reported on the energy storage performance of BaTi1−xZrxO3 capacitors and remains to be explored.

Herein, we report the fabrication of BaZr0.20Ti0.80O3 (BZT20) thin film capacitors using the pulsed laser deposition technique and studied their structural, vibrational, dielectric, and ferroelectric ordering with temperature. Temperature dependent Raman spectroscopy, as a local probe sensitive to structural transition, is employed to study the phonons’ behavior across tetragonal-cubic phase transition. The energy storage density and charge-discharge efficiency of the thin films’ capacitors have been estimated from the measured polarization hysteresis loop and compared with other reported relaxor systems. A large energy storage density of 31.9 J/cm3 and an energy storage efficiency of 56% were obtained at 1.3 MV/cm electric field and demonstrated BZT20 as an energy storage material for scalable electronic devices.

BaZr0.20Ti0.80O3 (BZT20) thin films of about 320 nm thickness were prepared by using the pulsed laser deposition (PLD) technique. High-purity precursors such as barium carbonate (BaCO3) (99%), zirconium (IV) oxide (ZrO2) (99.7%), and titanium (IV) oxide (TiO2) (99.8%) were used to synthesize PLD targets by a solid state reaction method. A KrF excimer laser (λ = 248 nm, f = 10 Hz) operating at a pulse frequency of 10 Hz was used for BaZr0.2Ti0.8O3/La0.7Sr0.3MnO3 (LSMO) heterostructure thin films fabricated on MgO (100) substrates at 923 K. The LSMO layer was crystallized in situ at 998 K for 60 min (dwelling time) in 300 Torr of oxygen pressure and cooled down slowly to room temperature to form a highly oriented bottom electrode. During the laser ablation of BZT on buffer LSMO coated on MgO at 923 K, a laser energy of 250 mJ was used and an oxygen process pressure of ∼150 m Torr was maintained inside the PLD chamber. The BZT thin films were crystallized by in situ post deposition annealing at 923 K for 60 min in 300 Torr of oxygen and allowed to cool down slowly to room temperature to crystallize preferentially (100)-oriented BZT/LSMO heterostructures. The phase purity and crystal structure of the fabricated samples were tested using the x-ray diffraction technique. Cu-Kα radiation (λ = 1.5404 Å) was used to record the diffraction pattern using a Rigaku Ultima III x-ray diffractometer operating in the Bragg-Brentano geometry. A Nanoscope III atomic force microscope (AFM) (Digital Instrument, USA) was used to obtain the surface morphology and roughness of the thin films employing tapping mode amplitude modulation. For electrical measurements, platinum (Pt) top electrodes of ∼80 μm diameter were dc sputtered through a shadow platinum mask onto the BZT thin films to form Pt/BZT/LSMO metal-insulator-metal (MIM) capacitors. To overcome any defect during the process of sputtering, we annealed the electrode thin films at 650 °C for 1 h in an oxygen pressure of 300 Torr. The dielectric constant of the samples was measured using an HP4294 inductance capacitance resistance (LCR) meter from 100 Hz to 1 MHz in the temperature range of 100–450 K at a 5 K/min cooling/heating rate. The remanent polarization and coercive field of the ferroelectric capacitors were measured using a Sawyer Tower test configuration (Radiant Technologies) up to the applied electric field of 1.5 MV/cm. The Raman spectra were recorded using an ISA T64000 spectrometer equipped with an Olympus microscope and a 50× long-working distance objective. The measurements were carried out in backscattering geometry employing the 514.5 nm excitation line of an Ar-ion laser. The signals were analyzed using a triple monochromator and detected by a CCD detector. The spectrometer resolution for 1800 l/mm grating was ∼1 cm−1. In situ temperature dependent measurements were carried out using a Linkam heating and cooling stage ensuring the temperature stability of ±0.1 K. To obtain the better spectra, the signal-to-noise ratio was improved by adjusting the laser power and data acquisition time. The spectra were analyzed using a damped harmonic oscillator model to obtain the Raman band position, linewidth, and band intensity, employing a PeakFit program.

The x-ray diffraction pattern measured from BZT20 thin films deposited on the LSMO coated MgO substrate is shown in Fig. 1. Figure 1 shows the multiple of (100) diffraction peaks located at 2θ positions 21.9°, 44.7°, and 69.4°, suggesting a highly textured thin film oriented along the (100) direction. One can notice strong (100) reflection peaks from the BZT20, LSMO, and MgO substrates without any impurity peaks from any secondary phases, suggesting that the BZT20 thin films have a single perovskite phase and are highly oriented along the in-plane. The thin films stabilized at the tetragonal structure (P4mm) and the tetragonality (c/a = 1.0004) is evident from the distinct splitting of peaks located at 2θ ∼ 44° into (002) at 44.59° and (200) at 44.7° peaks. The lattice mismatch between the BZT20 thin films and the MgO substrate is expected to induce lattice strain on the thin films. The lattice strain η can be calculated using the relation:17 η = (asubstrate − afilm/afilm) × 100. The lattice parameter for the MgO substrate is asubstrate = 4.213 Å and that for our BZT20 thin film is estimated as afilm = 4.061 Å, a smaller value as compared to the substrate. Therefore, a positive lattice mismatch is evident indicating a tensile strain in the BZT20 thin film grown on the MgO substrate. The tensile strain is found to be around 4%. Upon increasing the Zr concentration (x) from x = 0.3 to 0.5, the tetragonality distortion is reported to decrease13 and disappeared at x = 0.5, and the system transforms to a cubic phase for x ≥ 0.5.11 

FIG. 1.

Room temperature XRD Bragg's peaks of BZT20 thin films. The reflection peaks are labeled using JCPDF file No. 05-0626.

FIG. 1.

Room temperature XRD Bragg's peaks of BZT20 thin films. The reflection peaks are labeled using JCPDF file No. 05-0626.

Close modal

To study the surface morphology of thin films, in-plane and cross-sectional scanning electron microscopy (SEM) of the BZT20 films were carried out, as shown in Fig. 2. From the SEM image, one can see that the BZT20 films are smooth and have a dense surface. From the SEM cross-sectional image, the thickness of the grown thin films was estimated as ∼320 nm, similar to that obtained using a XP-200 Profilometer profile [Fig. 2(c)]. Energy dispersive x-ray spectroscopic measurement on thin films, excited by an electron beam of energy 20 kV, identifies the presence of elemental components Ba, Zr, Ti, and O in the films along with their characteristic x-ray emission lines. Using SEM elemental mapping, scanning over a large area of the thin film surface and a homogeneous distribution of these elements was observed [Fig. 2(d)]. The AFM micrograph of BZT20 thin films scanned in a contact mode in an area of 3 × 3 μm2 is shown in Fig. 3. The thin films are found to be smooth and they possess dense granular of nearly homogeneous distribution of grains. The grain sizes were found to be in the range of 26–42 nm. The average grain size was estimated as ∼32 nm from the statistical analysis involving a sample size of 100 grains. The average surface roughness RQ of the surface is ∼1.25 nm and their root mean square value is found to be Ra ∼ 0.95 nm.

FIG. 2.

(a) Planar SEM micrograph, (b) cross-sectional SEM image of BZT20 thin films grown on a MgO substrate, (c) thickness of the BZT20 thin films is around 320 nm using a profilometer, (d) elemental mapping of BZT20 thin films. The area in blue is “Ba,” that in green is “Zr,” that in yellow is “Ti,” that in reddish indicates “O,” and overall elements distribution is in the last one.

FIG. 2.

(a) Planar SEM micrograph, (b) cross-sectional SEM image of BZT20 thin films grown on a MgO substrate, (c) thickness of the BZT20 thin films is around 320 nm using a profilometer, (d) elemental mapping of BZT20 thin films. The area in blue is “Ba,” that in green is “Zr,” that in yellow is “Ti,” that in reddish indicates “O,” and overall elements distribution is in the last one.

Close modal
FIG. 3.

AFM morphology of MgO/LSMO/BZT20 thin films. Inset: three-dimensional AFM image.

FIG. 3.

AFM morphology of MgO/LSMO/BZT20 thin films. Inset: three-dimensional AFM image.

Close modal

The frequency dependence of dielectric constant ɛr at several temperatures for Pt/BZT20/LSMO thin film capacitors is shown in Fig. 4(a). The changes in dielectric constant are observed to be small in the low frequencies (<104 Hz). However, a highly dispersive in ɛr with a substantial decrease in its values is observed above this frequency.

FIG. 4.

Frequency dependencies of (a) dielectric constant (ɛ) of BZT20 thin films measured at several temperatures and (b) dissipation factor (tan δ) with frequencies.

FIG. 4.

Frequency dependencies of (a) dielectric constant (ɛ) of BZT20 thin films measured at several temperatures and (b) dissipation factor (tan δ) with frequencies.

Close modal

The possible reason for the observed decrease in the dielectric response in high frequency is outlined: as expected, the dielectric constant of BZT20 thin films has contribution from intrinsic (lattice contribution) and extrinsic (grain boundary and interface) factors. The intrinsic contribution depends on grain size, film orientation, and strain on the film. The interfacial layer between the BZT20 film and the bottom electrode La0.7Sr0.3MnO3 (LSMO) has the extrinsic contribution to the dielectric constant,23 and it behaves as a pinning center which essentially perturbs the domain wall motion and consequently affect the dielectric behavior. It is seen that in pure LSMO thin films, the extrinsic contribution to the dielectric constant is pronounced at higher frequencies.14 A comparison of the reported frequency dependent dielectric constant behavior of pure LSMO14 and our BZT20/LSMO thin films shows a similar decreasing trend of dielectric constant at higher frequencies, suggesting that the observed dielectric behavior in our thin films could be due to the dielectric response from the LSMO bottom layer. Furthermore, additional capacitance arising from the interface due to the contribution from two dissimilar materials BZT20 and LSMO is expected to be responsible for the present dielectric response at larger frequencies. Such a dispersive and decrease in the dielectric constant at higher frequencies was observed in PbZr0.52Ti0.48O3/LSMO thin films23 and is argued to be from the extrinsic interfacial contribution. The decrease in the dielectric constant with a rise in the temperature is expected due to the thermal fluctuation of the intrinsic polarization and from space charge polarization and interfacial polarization across the Pt/BZT20/LSMO interface. Another possibility can be attributed to the role of oxygen vacancies present in the thin film since at elevated temperatures oxygen vacancies become mobile and ionic conductivity becomes appreciable, and hence, a reduction in the dielectric constant with temperature is expected. Low dielectric losses are observed (<0.1) for almost all temperatures up to 103 Hz, and consequently, it increases above this frequency [Fig. 4(b)], attributed to oxygen vacancies and other disorders present in the material.24,25Figure 5 shows the temperature dependent dielectric behavior of dielectric constant (ɛr) and loss for the BZT20 thin films measured at 100 Hz–106 Hz. Only one peak with a dielectric maximum of ɛm ∼ 6200 at 1 kHz is observed at a temperature of Tm = 290 K.

FIG. 5.

Dielectric constant (ɛ) and dissipation factor (tan δ) of BZT20 thin films in the temperature range of 85–650 K.

FIG. 5.

Dielectric constant (ɛ) and dissipation factor (tan δ) of BZT20 thin films in the temperature range of 85–650 K.

Close modal

The dielectric peak broadens with temperature and a frequency dispersion was observed. Upon increasing the frequency, the dielectric maximum ɛm value decreases and a slight shift of dielectric peaks toward a higher temperature region is seen. The observed diffuse peak is expected due to the statistical distribution of Curie temperatures originated due to B-site compositional fluctuation.18 The loss (tan δ) variation with temperature increases upon increasing the frequency and is frequency dispersive (Fig. 5). A significant temperature difference between dielectric ɛ and loss tan δ peak positions, suggesting its relaxor behavior.18 The dielectric constant in the high temperature paraelectric phase obeys the Curie-Weiss law,

1ε=TTmC(T>Tm),
(1)

where ɛ is the dielectric constant, Tm is the Curie temperature, and C is the Curie-Weiss constant. The temperature dependence of the inverse of dielectric constant ɛr at 1 kHz is analyzed using the above expression (1) and is shown in Fig. 6(a). The fitting yields the parameters C = 11.5 × 105 K and Tm = 290 K, which are listed in Table I.

FIG. 6.

(a) Curie-Weiss plot for the inverse of the relative dielectric permittivity-temperature (1/ɛ vs T). Inset: linear fit of ln (1/ɛ − 1/ɛm) as a function of the ln (T − Tm) plot for BZT20 thin films and (b) variation of the reciprocal of Tm of Ba(Zr0.2Ti0.8)O3 thin films with frequency. The solid line is fitted using the Vogel-Fulcher relation.

FIG. 6.

(a) Curie-Weiss plot for the inverse of the relative dielectric permittivity-temperature (1/ɛ vs T). Inset: linear fit of ln (1/ɛ − 1/ɛm) as a function of the ln (T − Tm) plot for BZT20 thin films and (b) variation of the reciprocal of Tm of Ba(Zr0.2Ti0.8)O3 thin films with frequency. The solid line is fitted using the Vogel-Fulcher relation.

Close modal
TABLE I.

Dielectric maximum (ɛmax), temperature maximum Tm, Curie-Weiss constant (C), and dielectric constant follows the Curie-Weiss law above Tdev, Tdev − Tm, and the diffuse constant γ for the BZT20 thin film at 1 kHz.

ParameterɛmaxTm (K)C (×105 K)Tdev (K)Tdev − Tm (K)γ
Values 6190 290 13.5 340 50 1.64 
ParameterɛmaxTm (K)C (×105 K)Tdev (K)Tdev − Tm (K)γ
Values 6190 290 13.5 340 50 1.64 

The dielectric constant ɛr starts to deviate from linear behavior at a temperature of Tdev = 340 K, which lies above the Tm. The degree of deviation from the Curie-Weiss behavior is represented by the difference in temperature ΔT(Tdev − Tm) = 50 K. Earlier studies on BZT thin films with a larger Zr dopant (x ≥ 0.3)13,18 reported to have larger Tdev values and a lower dielectric maximum temperature Tm. The smaller magnitude of random field generated by the interaction of polar and nonpolar regions is argued18 to be the reason for the observed larger Tm in our BZT20. Therefore, the diffuseness of phase transition increases upon Zr concentration is apparent. In order to describe the diffusiveness of the phase transition, a modified Curie-Weiss law, expressed as

1ε1εm=(TTm)γC1,
(2)

is useful, where C1 is a constant and γ represents the degree of diffuseness of the phase transition.

The value of γ lies in the range of 1 ≤ γ ≤ 2. For γ = 1, it corresponds to a normal ferroelectric phase transition, and for γ = 2, the quadratic dependence suggests a diffuse phase transition for ideal relaxor ferroelectrics.26,27 However, the value of γ that lies in between 1 and 2 indicates an incomplete diffuse phase transition in relaxor ferroelectrics governed by the correlation of polar nanodomains dispersed in a nonpolar matrix. The slope of ln(1ε1εm) as a function of ln(TTm) essentially represents the γ value. The fitting yields the γ value of 1.64 and the linear fitting curve is shown in Fig. 6(a) (inset), implying a typical relaxor-ferroelectric behavior in our BZT20 thin films. Upon increasing the Zr dopant on the BZT (x ≥ 0.3) thin films’ relaxor, the γ values are estimated as ∼1.8, close to our present estimated value. As mentioned earlier, since the dielectric maximum peak shifts toward a higher temperature, one can also analyze the dielectric relaxation in BZT20 using the known Vogel-Fulcher (V-F) relation,16f=f0exp[EaKB(TmTf)], where f is the input signal frequency in which dielectric constants are measured, f0 is a pre-exponential factor related to the size and nature of the polar nanoregions, Ea is the activation energy of polar fluctuation, KB is the Boltzmann constant, Tm is the dielectric maximum temperature, and Tf is the freezing temperature of the relaxor. A close inspection of the above expression indicates that when the dielectric maximum temperature Tm approaches the freezing temperature Tf, f approaches toward zero; in other words, the kinetics of polarization fluctuation of PNRs tends to arrest as Tm shifted toward Tf and, consequently, freezes at Tf. The nonlinear curve fitting of the ln f ∼ 1000/Tm [Fig. 6(b)] plot using the V-F relation yields fitted parameters f0 = 1012 Hz, Ea = 0.047 eV, and Tf = 246 K. These fitted values are of the same order of magnitude and comparable with similar reported systems.28 The freezing temperature Tf is found to be larger as compared to those of higher Zr-dopant BZT.18 This is due to a large cooperative dipolar interaction possibly due to less density of nonpolar clusters in BZT20 thin film relaxor and, hence, the occurrence of freezing of polarization fluctuations at comparatively higher thermal energy, as compared to that for x > 0.2, is understandable.

To study the ferroelectricity at a nanoscale, ferroelectricity/piezoelectricity in BZT20 thin films was studied using piezoforce microscopy (PFM). A local electric field bias across the conductive AFM tip as the top electrode and LSMO as the ground bottom electrode of the BZT20 thin film was applied to obtain the PFM images. The switching of polarization was carried out by polling two square areas: a larger area 6 × 6 μm2 and the other central area of 4 × 4 μm2 of the thin film surface with +12 V and −12 V bias dc voltage, respectively (Fig. 7). One can observe a contrast behavior in switching of phase and amplitude of piezoelectric response upon the change in the bias voltage, suggesting that the BZT20 thin films exhibit ferroelectricity as inferred from our dielectric and vibrational spectroscopy. The observed asymmetric switching behavior can be attributed to the internal built-in electric field associated with the bottom interface29 and could also be due to mechanical stress from the AFM tip, as the tip/film/electrode configuration results in the shear stress deformation of the grain beneath the tip.17,30 Such ferroelectric imprint behavior was reported in other ferroelectric thin films such as (Ba0.955Ca0.045)(Zr0.17Ti0.83)O3,17 Pb(Zr0.2Ti0.8)O3,29 and Pb(ZrxTi1−x)O3.30 

FIG. 7.

Piezo force microscopy: (a) phase and (b) amplitude images of PZTS/LSMO/MgO thin films of thickness 320 nm.

FIG. 7.

Piezo force microscopy: (a) phase and (b) amplitude images of PZTS/LSMO/MgO thin films of thickness 320 nm.

Close modal

As discussed earlier, BZT20 stabilizes at the tetragonal phase (P4mm) similar to that of the pristine BaTiO3; therefore, the irreducible representations of phonons are the same as the parent compound. These are Γopt = 3A1 + 4E + B1, where the Raman and infrared active modes are A1 and E, whereas only the Raman active mode is B1. The Raman spectrum measured at ambient temperature is found to be broad (Fig. 8).

FIG. 8.

Raman spectrum measured at 298 K, analyzed using the sum of three Lorentzian peaks and an exponentially decay function as a background for BZT20. Individual Raman bands are also shown.

FIG. 8.

Raman spectrum measured at 298 K, analyzed using the sum of three Lorentzian peaks and an exponentially decay function as a background for BZT20. Individual Raman bands are also shown.

Close modal

The Raman bands are located with the center at 305, 517, and 738 cm−1 and can be assigned as A1(TO1), A1(TO2), and A1(LO), respectively.31,32 The A1(TO2) band located at 517 cm−1 is sensitive31 to temperature induced structural changes. A discussion on this band will be presented later. At low temperatures, one can expect a sharpening of Raman bands and overlapping nearby peaks can be resolved due to the decrease in phonon scattering processes. To confirm whether the observed broad nature of the Raman bands is due to the extrinsic inhomogeneous broadening related to the mixed solid solution system or due to the intrinsic anharmonic effect related to the finite temperature (ambient), the Raman spectrum was also measured at the lowest temperature of 82 K [Fig. 9(a)]. It is noticed that the Raman bands do not appear to narrow down as compared to the room temperature one, and thereby, the dominating extrinsic broadening effect is evident as seen in several relaxor-ferroelectric materials.6,33 The fluctuation of bond strengths of BO6 octahedra is apparent due to the substitution of Zr at the B-site, and a statistical distribution of all possible frequencies about a mean value results in the observed broadening of Raman bands. Figure 9(a) shows the Raman spectra of BZT20 measured at an elevated temperature in the T range of 82–420 K. Upon increasing the temperature, the spectra further broaden, and the intensity is found to decrease. At 420 K, the highest temperature of the present study, one can notice those three major Raman bands but with significant broadening and weaker intensity. To identify the temperature induced structural phase transition in BZT20, a study on the thermal evolution of the Raman band position, linewidth, and the intensity of the structural sensitive A1(TO2) band at 517 cm−1 is useful.31 

FIG. 9.

(a) Raman spectra recorded as a function of temperature in the T range of 82–440 K. The Raman bands are assigned as compared to those for tetragonal ferroelectric BaTiO3. Temperature dependencies for the 517 cm−1 A1(TO), (b) Raman band frequency, (c) linewidth, and (d) integrated intensity. The mode frequency and intensity of the band shows an anomaly of around 275 K at the tetragonal to cubic phase transition.

FIG. 9.

(a) Raman spectra recorded as a function of temperature in the T range of 82–440 K. The Raman bands are assigned as compared to those for tetragonal ferroelectric BaTiO3. Temperature dependencies for the 517 cm−1 A1(TO), (b) Raman band frequency, (c) linewidth, and (d) integrated intensity. The mode frequency and intensity of the band shows an anomaly of around 275 K at the tetragonal to cubic phase transition.

Close modal

The band position and the linewidth of the A1(TO2) band located at 517 cm−1 are obtained by analyzing the temperature dependent reduced Raman spectra, eliminated from the thermal population factor, using the damped harmonic oscillator model. The temperature dependent behavior of the Raman band positions and the linewidths are shown in Figs. 9(b) and 9(c). Upon heating, the band position is found to soften as normally expected and shows an anomaly around ∼275 K, the phase transition temperature. The observed decreasing trend of band frequencies with temperature is due to the lattice anharmonicity involved in the respective atomic vibration.34 The linewidth increases almost linearly with the temperature due to the decrease in the phonon lifetime associated with the enhanced multiphonons scattering processes. Since the intensity (I) of a Raman band is directly proportional to the square of the derivative of the mode polarizability (χ) with respect to the normal co-ordinate (q), the analysis of the temperature dependence of the square root of spectral intensity (I) is often useful for the identification of structural phase transition.35 To investigate the manner in which the integrated intensity of the structural sensitive A1(TO2) band located at 517 cm−1 belongs to the low temperature ferroelectric phase, evolved upon the increasing temperature, the spectral intensity was analyzed quantitatively. To repress the effect of extrinsic factors such as laser line focusing, laser power, and data acquisition time on spectral intensity, it was normalized with respect to the total integrated intensity integrated from 60 to 1000 cm−1. The integrated intensities (area) of the Raman band of the reduced Raman spectra, eliminated from the effect of the thermal population factor, were obtained at different temperatures and normalized with that of the total integrated intensity obtained from their corresponding spectrum in the same spectral range. The thermal evolution of intensities of A1(TO2) band is shown in Fig. 9(d). The intensity monotonically decreases up to 275 K, exhibits an anomalous behavior around 275 K, and decreases thereafter, corroborating the cubic phase transition temperature. Since the tetragonal anisotropy (c/a) of the BZT20 is only 1.0004, one can expect a phase transition to a high symmetry cubic phase at an elevated temperature. This is what we obtained the phase transition from our dielectric studies (discussed above). The slight difference in transition temperature obtained from our Raman spectroscopy and dielectric studies is due to the different length scales of the sensitivity of these used probes,6,7 viz., Raman spectroscopy is a local probe (length scale within a few unit cells),7 whereas that for dielectric, as a bulk technique lie in a few micrometers to millimeters (size of the sample).6 

P-E hysteresis loops of the BZT20 thin films’ capacitor measured at various applied electric fields at 10 kHz frequency is shown in Fig. 10. The hysteresis loop is found to be slimmer indicating its relaxor behavior as inferred from our dielectric and Raman studies. The incorporation of Zr4+ in the Ti4+ site randomly lead to a distribution of nonferroelectric ZrO6 and ferroelectric TiO6 octahedra in the crystal. Therefore, one can expect a slimmer P-E loop due to the breaking of a long-range polarization ordering (Ti-O dipole-dipole interaction) in BTO due to the substitution of the Zr4+ cation. Upon increasing applied electric fields, the remanent polarization Pr and the coercive field Ec are found to increase. The possible reason for the nonsaturation of the P-E loop could be due to the conductivity originated from oxygen vacancies, interface limited conduction, and other ionic conduction processes.36 The study on the current-voltage (I-V) characteristic curve on the thin film oxide supports the significant leakage current (conductivity) through the film with bias voltage. At the highest applied electric field of 1.3 MV/cm, close to the AC breakdown field, the measured Pr value of ∼89.6 μC/cm2 and the EC value of ∼0.29 MV/cm were obtained at 10 kHz. These obtained Pr and Ec values from our BZT20 thin films are larger as compared to reported higher Zr4+ dopant BZT thin films.13,18 Since the polar nanoregion number density decreases upon nonpolar Zr4+ cation doping in the BTO matrix, a reduction in the dipolar cooperative interactions is expected, and hence, the less Pr value with larger dopant concentration is expected. In thin films, the presence of defects such as oxygen vacancies, several other impurities, and layer interface often acting as pinning center generate internal electric field37,38 is argued to be responsible for the polarization relaxation, resulting in the discontinuity of the hysteresis loop. The present P-E loop experiments were carried out using the radiant ferroelectric test system in the virtual ground mode; therefore, polarization relaxation is evident in hysteresis loops. Similar behavior of discontinuity in hysteresis curves were reported in other nanoscale ferroelectric thin films such as (Ba0.955Ca0.045)(Zr0.17Ti0.83)O317 and Pb(Zr,Ti)O3 based ferroelectric thin films37,38 In fact, the gap is associated with the temporary memory decay with a transient time scale and can be attributed to polarization relaxation in the films.17 The asymmetric feature of the P-E loop could be due to the oxygen impurity related to the layer interface24,25 and/or different work functions39 of the top (Pt) and bottom (LSMO) electrodes.

FIG. 10.

P-E hysteresis loops of BZT20 thin films measured at several applied electric fields at a frequency of 10 kHz.

FIG. 10.

P-E hysteresis loops of BZT20 thin films measured at several applied electric fields at a frequency of 10 kHz.

Close modal

To study the leakage current conduction behavior of the BZT20 thin films, the current-voltage (I-V) characteristic curve was measured with a voltage step of 1 V and an elapsed time of 0.5 s for each voltage as shown in Fig. 11. It can be noticed that there are two regions from the plot: a lower electric field region, below 0.5 MV/cm, the current density increases linearly with the applied electric field indicating an Ohmic conduction behavior and other (above 0.5 MV/cm) upon increasing electric field, the current density increases exponentially pointing toward the involvement of Schottky or Poole-Frankel emission type conduction processes.17 One can notice a significant leakage current (>0.01 A/cm2) through the film with increasing bias electric field of above 0.5 MV/cm attributed to conductivity originated from oxygen vacancies and interface limited conduction.36 Similar conduction behavior was observed in (Ba0.955Ca0.045)(Zr0.17Ti0.83)O3 thin film capacitors, prepared by the PLD method, exhibiting a larger leakage current density (>0.5 A/cm2) of above 0.7 MV/cm.17 The DC breakdown electric field of BZT thin films infers from a separate I-V leakage current studies that is found to be around 1.23 MV/cm.

FIG. 11.

Variation of leakage current with the applied electric field for the BZT20 film.

FIG. 11.

Variation of leakage current with the applied electric field for the BZT20 film.

Close modal

The energy storage Ust and recoverable energy density Ure per unit volume of our Pt/BZT20/LSMO thin film capacitors can be estimated from the analysis of the experimental P-E loop data. The integration of area under the unipolar P-E loop at 10 kHz, as shown in Fig. 12, essentially provides the Ust, Ure, and loss energy Uloss values. The filled color areas under the curve: brown and yellow areas represent Ure and Uloss, respectively, and the total sum of these areas (Ure + Uloss) is the energy storage density Ust of the capacitor. In other words, one can estimate the Ure using the following expression:12 

Ure=PmaxPrEdP,0EEmax,
(3)

where E is the applied electric field, Pr and Pmax are the remanent and maximum polarization of hysteresis loop, respectively. Similarly, the store energy density of the capacitor can be obtained3 from

Ust=0PmaxEdP;0EEmax.
(4)
FIG. 12.

P-E loop and energy density calculation at a higher electric field for LSMO/BZT20/Pt thin film capacitors. The high energy density of BZT thin films suggests its superb potential to store electrical energy.

FIG. 12.

P-E loop and energy density calculation at a higher electric field for LSMO/BZT20/Pt thin film capacitors. The high energy density of BZT thin films suggests its superb potential to store electrical energy.

Close modal

The loss energy density Uloss is represented by the closed curve area (Fig. 12), obtained from the numerical integration of close area, and can be represented by Uloss = Ust − Ure. It can be mentioned that both recoverable energy storage density Ure and energy storage efficiency η are important parameters for charge storage application of thin film capacitors. The larger their values, the better their charge-discharge recoverable capacity. The energy storage efficiency η is defined as12 

η=Ure/Ust×100.
(5)

Using these formalisms, a recoverable energy storage density Ure of 31.9 J/cm3 and an energy storage efficiency η of 54.6 at 1.3 MV/cm are estimated for 10 kHz frequency. These energy storage parameters estimated for other applied electric fields are shown in Fig. 13. The storage energy density and efficiency are found to increase with the increase in the applied field since the maximum polarization value increases close to the breakdown field. These obtained values are found to be less as compared to reported BZT relaxors for 0.3 ≤ x ≤ 0.513 and is close to the reported40 Ure and energy efficiency η values for A- and B-site doped (Bi0.5Na0.5)TiO3-Bi(Ni0.5Zr0.5)O3 relaxors. Bi3.25La0.75Ti3O12 thin film capacitors prepared by the chemical solution deposition method show a recoverable energy density of 44.7 J/cm3 and an energy efficiency of ∼78%.41 Ba0.85Ca0.15Ti0.9Zr0.1O3 thin films prepared by the chemical solution deposition method in an oxygen enrich environment exhibits relaxor behavior and showed an energy storage density of 64.8 J/cm3 with an energy storage efficiency of 73% at an applied field of 2000 kV/cm.42 Studies on the laser ablated multilayer superlattice of BaTiO3/Ba0.3Sr0.7TiO3 relaxor ferroelectric are reported to have an energy density of 12.24 J/cm3.12 The BZT20 film fabricated on a Nb:SrTiO3 substrate using the radio frequency magnetron sputtering technique achieved an energy storage density of 30.4 J/cm3 with an efficiency of nearly 82%.3 BaTiO3 nanostructure based polymer composite thin films showed the energy density of 12.37 J/cm3 at 4.5 MV/cm.43 A relatively less energy density of ∼2.1 MV/cm and an efficiency of η < 50 were reported in Ba0.955Ca0.045Zr0.17Ti0.83O3 relaxor thin films fabricated using a RF sputtering method.44 The recoverable energy storage density Ure reported for BZT30 is 156 J/cm3 and that for BZT 50 is 144 J/cm3 estimated at a sustained high field of 3 MV/cm with their charge-discharge energy efficiencies η lying between 70% and 80%.13 These reported larger recoverable energy Ure is expected due to the increase in the degree of relaxion upon Zr (≤50%). Since our BZT20 thin film capacitors have less degree of relaxation, as compared to those reported BZT thin films; a comparatively smaller energy storage density is expected. On the other hand, another (Ba0.955Ca0.045)(Zr0.17Ti0.83)O3 thin film capacitors have shown discharge energy density Ure of ∼39.11 J/cm3,17 a value comparable to our present finding.

FIG. 13.

Energy densities and efficiencies of BZT20 thin films estimated for various electric fields.

FIG. 13.

Energy densities and efficiencies of BZT20 thin films estimated for various electric fields.

Close modal

Highly (100) oriented BZT20 thin films were grown on MgO single crystal substrates using the pulse laser deposition technique. X-ray diffraction results suggest that grown thin films were stabilized in the tetragonal phase. Raman spectroscopic studies identified A1(TO1) ∼ 305 cm−1, A2(TO2) ∼ 517 cm−1, and A1(LO) ∼ 738 cm−1 Raman bands corresponding to the tetragonal P4mm phase. The broad Raman spectral features are attributed to substitutional disorder, leading to relaxor behavior in BZT 20 thin films. The analysis of the spectral parameters such as band position, linewidth, and intensity of the structural sensitive A1(TO2) band located at 517 cm−1 identifies the tetragonal-cubic phase transition to be around 275 K. The temperature dependent dielectric studies on the BZT20 metal-insulator-metal capacitor showed a diffused peak of around 290 K. The analysis of dielectric data in the paraelectric phase was used to estimate the degree of diffuseness of the phase transition parameter γ of 1.64. The slim P-E hysteresis loop of thin film capacitors measured at several applied fields corroborates its relaxor behavior. Using the experimental P-E loop curve, a recoverable energy storage of Ure = 31.9 J/cm3 with an efficiency η of 54.6% at an applied electric field of 1.3 MV/cm was estimated. Our experimental results revealed that the BZT20 thin films’ capacitors have large potential for electrical charge storage applications.

The authors acknowledge the financial support from the Department of Defense, USA (DoD, Grant No. FA9550-16-1-0295).

1.
X.
Hao
,
Y.
Wang
,
L.
Zhang
,
L.
Zhang
, and
S.
An
,
Appl. Phys. Lett.
102
,
163903
(
2013
).
2.
Z.
Xie
,
Z.
Yue
,
B.
Peng
,
J.
Zhang
,
C.
Zhao
,
X.
Zhang
,
G.
Ruehl
, and
L.
Li
,
Appl. Phys. Lett.
106
,
202901
(
2015
).
3.
Z.
Sun
,
C.
Ma
,
X.
Wang
,
M.
Liu
,
L.
Lu
,
M.
Wu
,
X.
Lou
,
H.
Wang
, and
C.-L.
Jia
,
ACS Appl. Mater. Interfaces
9
,
17096
(
2017
).
4.
B.
Peng
,
Q.
Zhang
,
X.
Li
,
T.
Sun
,
H.
Fan
,
S.
Ke
,
M.
Ye
,
Y.
Wang
,
W.
Lu
,
H.
Niu
,
X.
Zeng
, and
H.
Huang
,
ACS Appl. Mater. Interfaces
7
,
13512
(
2015
).
5.
J. F.
Scott
,
Integr. Ferroelectr.
19
,
85
(
1998
).
6.
K. K.
Mishra
,
A. K.
Arora
,
S. N.
Tripathy
, and
D.
Pradhan
,
J. Appl. Phys.
112
,
073521
(
2012
).
7.
N.
Waeselmann
,
B.
Mihailova
,
B. J.
Maier
,
C.
Paulmann
,
M.
Gospodinov
,
V.
Marinova
, and
U.
Bismayer
,
Phys. Rev. B
83
,
214104
(
2011
).
8.
G. A.
Samara
,
Phys. Rev. B
71
,
224108
(
2005
).
9.
V.
Sivasubramanian
and
S.
Kojima
,
Phys. Rev. B
85
,
54104
(
2012
).
10.
D. A.
Sanchez
,
A.
Kumar
,
N.
Ortega
,
R. S.
Katiyar
, and
J. F.
Scott
,
Appl. Phys. Lett.
97
,
202910
(
2010
).
11.
N. K.
Karan
,
R. S.
Katiyar
,
T.
Maiti
,
R.
Guob
, and
A. S.
Bhallab
,
J. Raman Spectrosc.
40
,
370
(
2009
).
12.
N.
Ortega
,
A.
Kumar
,
J. F.
Scott
,
D. B.
Chrisey
,
M.
Tomazawa
,
S.
Kumari
,
D. G. B.
Diestra
, and
R. S.
Katiyar
,
J. Phys. Condens. Matter
24
,
445901
(
2012
).
13.
A. A.
Instan
,
S. P.
Pavunny
,
M. K.
Bhattarai
, and
R. S.
Katiyar
,
Appl. Phys. Lett.
111
,
142903
(
2017
).
14.
S.
Majumdar
,
H.
Huhtinen
,
P.
Paturi
, and
H. S.
Majumdar
,
J. Mater Sci.
48
,
2115
(
2013
).
15.
M. K.
Bhattarai
,
K. K.
Mishra
,
S.
Dugu
,
A. A.
Instan
, and
R. S.
Katiyar
,
Appl. Phys. Lett.
114
,
223902
(
2019
).
16.
M. K.
Bhattarai
,
K. K.
Mishra
,
A. A.
Instan
,
B. P.
Bastakoti
, and
R. S.
Katiyar
,
Appl. Surf. Sci.
490
,
451
(
2019
).
17.
V. S.
Puli
,
D. K.
Pradhan
,
S.
Adireddy
,
R.
Martínez
,
P.
Silwal
,
J. F.
Scott
,
C. V.
Ramana
,
D. B.
Chrisey
, and
R. S.
Katiyar
,
J. Phys. D Appl. Phys.
48
,
355502
(
2015
).
18.
A.
Dixit
,
S. B.
Majumder
,
R. S.
Katiyar
, and
A. S.
Bhalla
,
J. Mater. Sci.
41
,
87
(
2006
).
19.
W.
Jie
,
J.
Zhu
,
W.
Qin
,
X.
Wei
,
J.
Xiong
,
Y.
Zhang
,
A.
Bhalla
, and
Y.
Li
,
J. Phys. D Appl. Phys.
40
,
2854
(
2007
).
20.
T.
Wu
,
C.
Wu
, and
M.
Chen
,
Appl. Phys. Lett.
69
,
2659
(
1996
).
21.
J. P. B.
Silva
,
J. M. B.
Silva
,
M. J. S.
Oliveira
,
T.
Weingärtner
,
K. C.
Sekhar
,
M.
Pereira
, and
M. J. M.
Gomes
,
Adv. Funct. Mater.
29
,
1807196
(
2018
).
22.
H.
Palneedi
,
M.
Peddigari
,
G.-T.
Hwang
,
D.-Y.
Jeong
, and
J.
Ryu
,
Adv. Funct. Mater.
28
,
1803665
(
2018
).
23.
D.
Barrionuevo
,
N.
Ortega
,
A.
Kumar
,
R.
Chatterjee
,
J. F.
Scott
, and
R. S.
Katiyar
,
J. Appl. Phys.
114
,
234103
(
2013
).
24.
G.
Singh
,
V. S.
Tiwari
, and
P. K.
Gupta
,
J. Appl. Phys.
107
,
064103
(
2010
).
25.
R. E.
Newnham
,
Properties of Materials Anisotropy, Symmetry, Structure
(
Oxford University Press
,
2005
).
26.
A.
Dixit
,
S. B.
Majumder
,
P. S.
Dobal
,
R. S.
Katiyar
, and
A. S.
Bhalla
,
Thin Solid Films
447–448
,
284
(
2004
).
27.
X.-G.
Tang
and
H. L.-W.
Chan
,
J. Appl. Phys.
97
,
034109
(
2005
).
28.
V.
Reymond
,
O.
Bidault
,
D.
Michau
,
M.
Maglione
, and
S.
Payan
,
J. Phys. D Appl. Phys.
39
,
1204
(
2006
).
29.
A.
Gruvermana
,
A.
Kholkin
,
A.
Kingon
, and
H.
Tokumoto
,
Appl. Phys. Lett.
78
,
2751
(
2001
).
30.
A.
Wu
,
P. M.
Vilarinho
,
D.
Wu
, and
A.
Gruverman
,
Appl. Phys. Lett.
93
,
262906
(
2008
).
31.
V. S.
Puli
,
D. K.
Pradhan
,
W.
Perez
, and
R. S.
Katiyar
,
J. Phys. Chem. Solids
74
,
466
(
2013
).
32.
M.
Didomfnico
, Jr.
,
S. H.
Wemple
, and
S. P. S.
Porto
,
Phys. Rev.
174
,
522
(
1968
).
33.
K. K.
Mishra
,
A. T.
Satya
,
A.
Bharathi
,
V.
Sivasubramanian
,
V. R. K.
Murthy
, and
A. K.
Arora
,
J. Appl. Phys.
110
,
123529
(
2011
).
34.
K. K.
Mishra
,
S.
Chandra
,
N. P.
Salke
,
S. N.
Achary
,
A. K.
Tyagi
, and
R.
Rao
,
Phys. Rev. B
92
,
134112
(
2015
).
35.
J.
Riuz-Fuertes
,
D.
Errandonea
,
O.
Gomis
,
A.
Friedrich
, and
F. J.
Manjon
,
J. Appl. Phys.
115
,
043510
(
2014
).
36.
E.
Gradauskaite
,
J.
Gardner
,
R. M.
Smith
,
F. D.
Morrison
,
S. L.
Lee
,
R. S.
Katiyar
, and
J. F.
Scott
,
Phys. Rev. B
96
,
104104
(
2017
).
37.
R.
Thomas
,
S.
Mochizuki
,
T.
Mihara
, and
T.
Ishida
,
Thin Solid Films
413
,
65
(
2002
).
38.
J.
Lee
and
R.
Ramesh
,
Appl. Phys. Lett.
68
,
484
(
1996
).
39.
M. M.
Toda
,
Int. J. Appl. Phys. Math.
1
,
3
(
2011
).
40.
N.
Sun
,
Y.
Li
,
Q.
Zhang
, and
X.
Hao
,
J. Mater. Chem. C
6
,
10693
(
2018
).
41.
B. B.
Yang
,
M. Y.
Guo
,
D. P.
Song
,
X. W.
Tang
,
R. H.
Wei
,
L.
Hu
,
J.
Yang
,
W. H.
Song
,
J. M.
Dai
,
X. J.
Lou
,
X. B.
Zhu
, and
Y. P.
Sun
,
Appl. Phys. Lett.
111
,
183903
(
2017
).
42.
S. R.
Reddy
,
V. V. B.
Prasad
,
S.
Bysakh
,
V.
Shanker
,
N.
Hebalkar
, and
S. K.
Roy
,
J. Mater. Chem. C
7
,
7073
(
2019
).
43.
Z.
Pan
,
L.
Yao
,
J.
Zhai
,
H.
Wang
, and
B.
Shen
,
ACS Appl. Mater. Interfaces
9
,
14337
(
2017
).
44.
T.
Chen
,
J. B.
Wang
,
X. L.
Zhong
,
Y. K.
Zeng
,
F.
Wang
, and
Y. C.
Zhou
,
Appl. Surf. Sci.
285
,
744
(
2013
).