Introduction

The increase of energy demand over the years and its cost coupled with different environmental issues, such as greenhouse effect, global warming, air pollution, and acid rain1,2, drive the world to go for cleaner, more reliable and renewable energy resources (i.e., wind, solar, heat, stress, vibration, and magnetic field). In the United States, around 68% of the primary energy generated is wasted as heat each year3. The waste-heat is ubiquitous, abundant, and freely available4 making energy-harvesting a cost-effective, maintenance-free, and self-sustainable, hence they are suitable for supplying power in remote areas. Waste-heat is classified into three types, depending on the temperature range5: high-grade (≥ 649 °C), medium grade (232—649 °C), and low-grade (≤ 232 °C). Usually medium and high-grade heat can be easily recovered when compared to low-grade waste heat. However, a substantial amount ~ 60% of the wasted heat is low-grade, which is difficult to scavenge efficiently5.

Considerable efforts are currently underway to utilize the low-grade waste-heat with improved efficiency and reduced carbon emission using pyroelectric (ferroelectric) materials which require periodic temperature fluctuation.

Ferroelectric materials demonstrate a spontaneous and switchable polarization, which are strongly electric field and temperature dependent. The change of polarization with the variation of temperature is called the pyroelectric effect (PEE). This occurs in ceramics, single crystals, thin films, and polymers with an acentric crystal structures possessing a unique axis of rotational symmetry, giving the structure an inherent spontaneous polarization6.

The pyroelectric performance of a material is determined by its pyroelectric coefficient (π), which is the rate of change of polarization with temperature and can be computed by probing the pyroelectric current,

$${i}_{p }=\pi A\left(\frac{dT}{dt}\right)$$
(1)

where A is the area of the electrode and dT/dt is the rate of change of the material’s temperature with respect to time. Many techniques have been reported to measure the pyroelectric properties of large samples7,8. However, it is more challenging to estimate π in smaller samples because of poor accuracy in temperature measurement, non-uniform heating, and thermally induced current produced when the trapped electrons are released9,10. Only a few techniques, such as, hot plate heating, micro-fabricated resistive heater-based experiments11 and modulated laser-based measurements12 are reported to estimate π for small-sized samples.

The power density of different thermal energy harvesting techniques is published in the literature. The thermoelectric generators show very high-power density (~ 3.0 kW/cm3) when the difference in temperature is > 600 °C. But their power density significantly decreases with the decrease of temperature difference. Nguyen et al.13 has reported a power density of 10.7 mW/cm3 at a temperature difference of 14.8 °C in P(VDF-TrFE) pyroelectric material whereas Sato et al.14 observed power density of ~ 2.0 mW/cm3 on thermoelectric material (shape memory alloy) at similar temperature difference. Therefore, for the small temperature difference (ΔT < 100 °C), pyroelectric materials are one of the most capable of waste-heat scavenging exhibiting an appreciable power density4,15.

To this date, only a few materials were studied to recover waste heat resulting in some exciting results. Recently, Pandya et al.16 studied the conversion of pyroelectric energy from low-grade waste-heat, which exploits both temperature and field-dependent polarization, on lead-based 0.68Pb(Mg1/3Nb2/3)O3-0.32PbTiO3 relaxor-ferroelectric thin films. Similarly, Bhatia et al.11 demonstrated three frequency domain thermal measurement (a hot plate, microfabricated heater, and modulated laser) techniques and compared their corresponding pyroelectric currents based on PZT epitaxial film from 0.02 Hz to 1.3 MHz. To date, most of the studies have been focused on lead-based ferroelectric materials and reported their energy conversion density. Even though the lead-based ferroelectric materials exhibit large polarization with excellent pyroelectric as well as piezoelectric coefficients, their applications are limited because of their toxicity and environmental concerns.

Regarding the lead-free ferroelectric materials, Bhatia et al.17 systematically studied a BaTiO3 thin film and presented detailed studies, using the pyroelectric Ericsson cycle, to optimize the conversion of thermal energy into electrical energy. While BaTiO3 is one of the most popular lead-free ferroelectric materials, its ferroelectric, dielectric, piezoelectric, and pyroelectric properties are lower than the lead-based ferroelectrics. Previous research showed the enhancement of these properties by suitable site modification at both sites in BTO solid solutions BaZr0.2Ti0.8O3 (BZT), Ba0.7Ca0.3TiO3 (BCT), and BaZr0.2Ti0.8O3–Ba0.7Ca0.3TiO3–(BZT-BCT) for different applications in multifunctional devices4a,b. The images are scanned for a size of 1 × 1 μm2 at multiple locations, which exhibits relatively smooth, well-connected grains with an RMS surface roughness (Rq) of ~ 1 nm. The 3D micrograph, in Fig. 4a, shows a good quality film due to uniform deposition as the upper surface of the film exhibiting a layer that is homogeneous and free of cracks and pores/holes. Such a well-connected, dense, and uniform granular microstructure with low surface roughness is required to enhance the dielectric, the ferroelectric, the pyroelectric, and the electrical properties of the film. Figure 4b depicts the 2D micrograph, which demonstrates that the surface is smooth with densely packed grains of uniform size ~ 40 nm.

Figure 4
figure 4

AFM images of the heterostructures BZT/BCT: (a) 3D and (b) 2D.

Full size image

I-V characteristic of BZT/ BCT heterostructures

The current density as a function of electric field (J-E) of BCT/BZT heterostructures in metal-ferroelectric-metal (MFM) architecture at ambient temperature is shown in Fig. 5a. We noticed a low leakage current of < 10–6 A/cm2 through the film with an applied electric field of 0.1 MV/cm, which is comparable to the literature on similar materials22,23. This low leakage current is originated due to the presence of oxygen (100 mTorr) during deposition and the completeness of perovskite phase formation at higher annealing temperature (800 °C) after the deposition24. At a higher oxygen environment, the leakage current is reduced by suppressing the defects, such as oxygen vacancies, which reduces the ferroelectricity. There is a sharp increase of current, even at a lower electric field, followed by almost saturation at a higher electric field. At a low electric field, the J-E curve obeys Ohm’s law, at a medium electric field, space-charge induced conduction is dominant whereas, at a higher electric field, Schottky-Frankel conduction mechanism dominates25,26.

Figure 5
figure 5

Variation of (a) leakage current density with electric field (J–E) curves and (b) dielectric constant with frequency and electric field at 1 kHz (inset) of BZT/BCT films.

Full size image

Dielectric properties of BZT/BCT heterostructures

Dielectric tunability is one of the important features of the ferroelectrics. The dielectric permittivity of BZT/BCT heterostructures is found to decrease when the frequency increases, Fig. 5b, which indicates a typical characteristic of ferroelectric thin films. The dielectric constant in solids, generally, is due to different types of polarizations such as: (1) space charge/interfacial (2) dipolar/orientational (3) ionic and (4) electronic27. At lower frequencies, the dielectric constant is high, which is due to the presence of most of polarization dynamics. The dielectric permittivity of these nanostructures is found to be ~ 1000 at 1 kHz, which is smaller compared to similar materials in the literature18,28,29. Such a low dielectric constant is crucial for the improvement of the figure of merit of the pyroelectric device. The variation of dielectric constant with electric field at 1 kHz of the ferroelectric materials is demonstrated in the capacitance–voltage (C–V) characteristics, as shown in Fig. 5b (inset). The polarization modulation, via the variation of the dielectric constant as a function of dc electric field, is related to the domain movement. The higher dielectric response, at a low dc field, can be attributed to the huge increase in polarization due to the reversal of domains. At a higher electric field, the maximum number of domains are aligned along the direction of the electric field, resulting in a significantly reduced dielectric response. Such a typical C–V butterfly characteristic, i.e. non-linear dielectric properties with applied field, demonstrates the ferroelectric nature of the materials. The occurrence of smaller asymmetry in this C–V curve is due to dissimilar materials used for the top (Pt) and bottom (SRO) electrodes.

Pyroelectric properties of BZT/BCT heterostructures

As explained in our previous publication21, the BCT/BZT heterostructures demonstrated well-saturated and slim ferroelectric hysteresis loops, at ambient temperature in the applied field from 0.7 to 1.7 MV/cm, confirmed the ferroelectric behavior of the heterostructures even at higher applied voltages. Pyroelectric behavior occurs in most of the ferroelectrics as they exhibit large spontaneous and switchable polarization in a wide range of temperature30. To estimate the energy conversion density of the thin films of ferroelectrics, we recorded the P-E hysteresis curves of these thin films at different temperatures, Fig. 6a. We found that both the saturation and remnant polarization values decrease monotonically with the increase of temperature15,29. The polarization is regulated by the lattice structure of the material and shows a large change when the material undergoes structural transition with the variation of temperature.

Figure 6
figure 6

Variation of polarization with electric field (P–E) hysteresis loops of BZT/BCT films at the frequency of 1 kHz at different temperatures (a). Sketch of the pyroelectric Ericsson cycle in four-step on a ferroelectric P–E loops at two temperatures (b).

Full size image

A ferroelectric loop is usually recorded in an anticlockwise direction and the area enclosed by the loop indicates the losses by which electrical energy is transformed into thermal energy. However, if the loop is measured in a clockwise direction, it might be possible to generate electrical energy from thermal energy31,32. The investigation of pyroelectric energy conversion, with a ferroelectric loop using the Ericsson cycle, is displayed in Fig. 6b and it has the following four steps:

  1. 1.

    Isothermal charging: a high E-field is applied at a constant temperature until it arrives at its saturated polarization where all the dipoles are oriented in the direction of the applied field (represented by step 1–2).

  2. 2.

    Heating at constant E-field: when the material reaches saturated polarization and heats at constant electric field, a large depolarization current is generated due to the change in orientation of the dipoles (represented by step 2–3).

  3. 3.

    Isothermal discharging: once the material achieves thermal equilibrium, the applied E-field is removed slowly, so that all the dipoles are randomly oriented (represented by step 3–4).

  4. 4.

    Cooling at constant E-field: the material at low electric field, removed from the heat source, and then kept at the sink so that all the dipoles restore to their original position giving rise to large polarization current (represented by step 4–1).

The enclosed area 1–2–3–4 by the four steps on a hysteresis loop indicates the energy conversion density (ND) of the thin film, which is computed using the following equation 33,34,

$${N}_{D}={\oint }_{1234}E dP$$
(2)

and found to be 10,970 kJ/m3 per cycle. This energy conversion density value is higher than the one found in similar lead-free materials and even higher than some of the lead-based structures as seen in Table 2.

Table 2 The pyroelectric energy density harvested per cycle of various recently reported pyroelectric materials.
Full size table

There is a great interest in exploring lead-free structures for energy harvesting and recently, it has been found that a maximum energy conversion density of 149 kJ/m3 was attained by Patel et al.15 with a saturated E- field and temperature of 24 °C and 96 °C on a lead-free 50BZT-50BCT bulk system. Similarly, Chauhan et al.39 reported a large energy conversion density of 1,986 kJ/m3 on a lead-free ferroelectric thin film of Bi0.5Na0.44K0.06TiO3 under an operational temperature of 25 °C and 120 °C. The energy conversion density we measured in our lead-free BZT/BCT heterostructures is 10,970 kJ/m3 under an operational temperature of 20 °C and 100 °C, which is showing better performance than the other reported data15,39.

Figure 7 shows the actual temperature fluctuations on the material and the corresponding pyroelectric currents generated from BCT/BZT multilayer structures at different laser illumination repetition rates.

Figure 7
figure 7

Temporal profile for temperature, its time derivative, and the corresponding pyroelectric current generated (a,b) for laser repetition rate 0.1 Hz, (c,d) for 0.5 Hz and (e,f) for 1 Hz, respectively. At a repetition rate of 0.1 Hz (5 s heating and cooling periods), the observed pyroelectric current is symmetric but at 1 Hz repetition rate (0.5 s heating and cooling periods) the positive current is unchanged and the negative current decreases but is non-zero.

Full size image

The heating conditions and the observed pyroelectric current maxima and minima are presented in Table 3.

Table 3 Temperature range and current density for a range of heating repetition rates.
Full size table

In Table 3, the conditions observed due to laser heating at four different repetition rates are presented with the corresponding temperature ranges observed as well as the current observed during each cycle. As shown in the table, as the repetition rate decreases, with a 50% duty cycle, the period of exposure increases, and the temperature swing experienced by the sample increases as well. Due to this lengthened heating period, the maximum temperature of the sample increases since more heat is delivered to the sample. However, as the heating frequency increases, the average temperature increases slightly since cooling time also decreases. The heating/cooling curves shown in Fig. 8a can be approximated as an exponential rise and fall for 0.1 Hz repetition rate but appear much more linear for faster repetition rates. The temperature range for each repetition rate is presented in Fig. 8b. In the figure, the amplitude is plotted versus the phase of the repetition rate to highlight the amplitude of the oscillation. It is clear from the figure that the amplitude of the heating cycle increases with increasing period. For repetition rates below 0.2 Hz, the cooling period and the cooling time are approaching each other, so that the sample has enough time to reach its unperturbed temperature before the next heating cycle. At repetition rates higher than 0.2 Hz, the sample reaches equilibrium at a higher temperature. This increasing temperature is shown in Fig. 8a. This will be discussed in more detail in a later section of this work.

Figure 8
figure 8

Plot of (a) average temperature for different heating frequency, (b) the amplitude of temperature change as a function of the phase cycle at heating frequencies, and (c) pyroelectric current density versus heating frequency for positive-going current only.

Full size image

Figure 7 also contains the waveform of the pyroelectric current generated during each heating and cooling cycle. When the laser repetition rate is 0.1 Hz and a laser incident power of 150 mW and a spot size of ~ 100 microns, the temperature fluctuates with a magnitude of 11.8 ℃ and the corresponding symmetric pyroelectric current is recorded from – 25.8 mA/m2 to + 25 mA/m2, as shown in Fig. 7a,b. Therefore, at a repetition rate below 0.2 Hz, the pyroelectric waveform is symmetric about the x-axis with approximately the same amplitude in both polarities. As the repetition rate increases above 0.2 Hz, the waveform becomes less symmetric with the amplitude of the negative portion decreasing. When the repetition rate of the laser is increased to 0.5 Hz, a smaller temperature fluctuation of 3.4 °C is measured and the corresponding pyroelectric current is asymmetric with a reduction in the negative current peak value, as shown in Table 3 and also in Fig. 7c,d. With a further increase of laser repetition rate to 1 Hz, the amplitude of the temperature fluctuation is further reduced to 1.7 °C (Fig. 7e,f). At 1 Hz, the negative amplitude of the pyroelectric current is also reduced but is not zero. It is worth noting that, both the maximum temperature and positive pyroelectric current only decrease slightly as the incident energy of the laser decreases with the length of the excitation periods. Additionally, the positive current is relatively unchanged with an order of magnitude reduction in the incident energy of the laser excitation.

From the above observations, we can conclude that with increasing laser repetition rates and, thus decreasing heating and cooling periods, the sample cannot cool to its lowest pre-excitation temperature. Therefore, with shorter cooling periods, the temperature to which the sample cools before the next pulse is higher than it would be if left to cool longer. The competition between the heating rate driven by the laser pulse width and cooling rate driven by the thermal conductivity of the film ultimately determines the range of temperatures within the film particularly the lower temperature. When the “laser-off” period is longer than the cooling period, the sample reaches a lower temperature. When the “laser-off” period is shorter than the cooling period, the sample cools less. This is also observed in the pyroelectric current waveform. Since the pyroelectric current is the direct result of the change in temperature, the magnitude of the current will be proportional to the change in temperature. As this range decreases for the cooling part of the cycle as the repetition rate increases, the resulting negative portion of the pyroelectric current will also decrease. At repetition rates where the cooling rate is competitive with the time between pulses, the change in temperature is large and the pyroelectric current is also larger. As the range of temperature changes decreases, the sample develops a steady-state temperature closer to the maximum temperature. For our samples, when excited with repetition rates faster than 2 Hz, the polarity of the bulk of the waveform is positive current, however, a small negative current is also observed. Our material demonstrates the characteristic of symmetric alternating current signature of pyroelectric current for 0.1 Hz repetition rates or slower.

Figure 8c is a plot of the pyroelectric current density for each of the heating frequencies listed above. The current density is calculated based on the positive portion of the waveforms presented in Fig. 7. The mean current density was 23.7 \(\pm \) 1.2 mA/m2. Since the current density has been determined from only the positive-going current, the maximum density may be larger. The pyroelectric coefficient was calculated for positive pyroelectric current density at 0.1 Hz heating frequency using Eq. 1 and found to be 3425 μC/m2K.

To date, few ferroelectric materials are studied for their pyroelectric measurement and reported their current densities. Bhatia et al.11 has reported the current density varies from 10–5 A/m2 to 103 A/m2 over the range 10–2 to 106 K/s on PbZr0.2Ti0.8O3 thin film. They have observed the current density of 10–4, 0.1, and 103 A/m2 with temperature oscillation amplitude of 3, 1, and 4.5 K over the heating frequency 0.02, 100, and 105 Hz, respectively. Similarly, Biancoli et al.40 has demonstrated a pyroelectric current density of 10 pA/cm2 and 3.5 pA/cm2 on Ba0.6Sr0.4TiO3 ceramic and BaTiO3 single crystal, respectively, with the temperature fluctuation of 1 K and heating frequency 0.02 Hz. Recently, You et al.41 demonstrated a pyroelectric current density of 12.5 μA/m2 with a temperature oscillation of 6 K on a PVDF polymer at a heating frequency of 0.125 Hz. However, the BCT/BZT thin films, reported here, show relaxor-ferroelectric nature exhibit excellent pyroelectric performance compared to similar lead-free materials reported in the existing literature. The present results show that, in multilayer heterostructures, as the temperature fluctuates, there is abrupt change in polarization due to structural transformation, which in turn, significantly improves the pyroelectric performance. These relaxor-ferroelectric multilayer structures have the potential to be used for future low-grade waste-heat harvesting applications.

Conclusions

BZT/BCT multilayer nanostructures are successfully grown on SRO coated STO single crystalline substrate using optimized PLD technique. The (l00) reflection peaks from STO substrate, SRO, and multilayer structure indicate a crystalline, phase pure, and highly c-axis orientated of these nanostructures. The AFM images, recorded in tap** mode, exhibit a very smooth surface indicating good quality growth. These BZT/BCT multilayer heterostructures demonstrate excellent pyroelectric performance compared to the lead-based materials reported in the existing literature. The above-mentioned results show that these relaxor-ferroelectric multilayer heterostructures demonstrate excellent pyroelectric energy density of 10,970 kJ/m3 per cycle and a current density of ~ 25 mA/m2 and the corresponding pyroelectric coefficient of 3425 μC/m2K with 11.8 °C temperature fluctuation. These pyroelectric results are better than those reported for other lead-free materials and demonstrate a potential to be used in low-grade thermal energy harvesting applications.