Introduction

Recently, ferroelectric materials have attracted significant attention due to their excellent physical properties and performance; these include photoelectricity1, pyroelectricity2, piezoelectricity3, and flexoelectricity4. Based on these characteristics, ferroelectrics can be used for harvesting a variety of energy sources from the environment, such as light5,6, thermal7,8, and mechanical energy9,10. To scavenge energy more efficiently, researchers have taken advantage of coupling effects, which includes the coupling of light and thermal energy sources11,12,13,14, the coupling of light and mechanical energy15,16,17,18, and the simultaneous coupling of light, thermal, and mechanical sources19,20,21. These studies have demonstrated that the coupling of different mechanisms can lead to an enhancing effect and thus increase the energy harvesting capability.

Currently, the coupling factor (KC) is commonly used as a metric to evaluate the performance of a coupled generator22,23,24, which represents the ratio of change in the performance of generator after coupling with respect to its performance prior to coupling, and the value of KC therefore represents a relative indicator. The magnitude of the coupling factor can be used to visualize the effect of enhancement, or weakening, of the current, charge, and energy of the device after coupling. However, a variety of ferroelectric materials can generate the coupling enhancement effect, and determining which material has the most significant enhancement remains unclear. Previous reports have demonstrated a quantified series of the triboelectric performance of various materials by triboelectric charge density25,26, which provides a good reference for the selection of high-quality triboelectric nanogenerators. Inspired by this approach, the KC of various ferroelectric films can also be quantitatively ranked, which can be used as a reference for the selection of materials for coupled generators. In addition, since the KC is primarily used to indicate the relative enhancement before and after coupling, it cannot be used to visualize the magnitude of the output of the generator. Therefore, a new metric that combines both the absolute performance and the relative enhancement of the coupled generator is proposed, which is more significant for the evaluation of the coupling performance. Finally, since energy is representative as an important indicator of the performance of electronic devices, is should also be considered as an indicator.

In this work, we therefore developed a standardized experimental platform dedicated to testing the performance of various ferroelectric films in terms of its coupling of pyroelectric and photovoltaic effects, and quantitatively rank the coupling performance of a range of ferroelectric materials. The charge coupling factor (KC,Q) of different materials were first ranked as an evaluation metric, which represents the relative change of charge before and after coupling. Moreover, a new indicator, termed the Yang’s charge (QY) is proposed, which takes into account both the relative enhancement upon coupling and its absolute ability to generate charge, which outlines the overall performance of the material upon coupling, and represents a more rigorous and comprehensive approach compared to traditional metrics.

Furthermore, the output energy of each material was also quantitatively ranked to visualize their ability for energy scavenging applications. In addition, the underlying reasons leading to the enhancement of charge coupling are analyzed theoretically and the calculation of KC,Q is simplified. This work leads to an approach for ranking of the coupling performance of ferroelectric films due to pyroelectric and photovoltaic effects by three evaluation metrics, which provides a unique reference source in terms of material selection for high-performance coupled generators based on ferroelectric materials.

Results

Experimental platform for testing pyroelectric and photovoltaic effect

To quantify the coupling performance of the pyroelectric effect (PyE) and photovoltaic effect (PvE) of different ferroelectric films, an experimental platform dedicated to testing PyE and PvE effects was constructed, as shown in Fig. 1a. The left side of the experimental platform primarily consists of a 3D displacement stage that is fixed on a load platform, on which the ferroelectric devices are located, and the displacement stage is used to adjust the position of the samples to be tested. The right side of the experimental platform consists of two laser transmitters with two different emission wavelengths of 1064 nm and 405 nm, where a long wavelength 1064 nm laser is used to generate a pyroelectric effect and a low wavelength 405 nm laser is used to generate a photovoltaic effect. Changing the wavelength generates different effects related to the balance of photothermal and photoexcitation, as explained in detail in the screening strategies of the power sources section, outlined below. The ports of both laser emitters are fitted with 1.2 mm diameter apertures to ensure a uniform spot size for both lasers. The transmitter is fixed to a rotating platform and a vertical lifting platform in sequence to control the angle and height of the laser beam, respectively. Both the load platform and the lift platform are screwed and mounted to the lower substrate of the setup to ensure the components of the system remain in place during testing. The whole testing process was performed at 27 °C of room temperature. In order to avoid the interference of external light sources, the experimental platform was completely covered by a rectangular light shield, and the photos of both are presented in Supplementary Fig. 1.

Fig. 1: Experimental setup and evaluation metrics of coupling performance.
Fig. 1: Experimental setup and evaluation metrics of coupling performance.
Full size image

a Experimental setup for measuring the coupling of photovoltaic and pyroelectricity. The experiment was measured in a light shield. b Layered structure of the devices based on ferroelectric films. c SEM image of the cross-section of PST ferroelectric film. Scale bar in (c), 200 nm. d Schematic of the current signal upon coupling of the pyroelectric effect (PyE) and photovoltaic effect (PvE). e Schematic of the change in charge and energy before and after coupling of PyE and PvE, and metrics for testing the sequence of coupling performance.

Figure 1b shows the structure of the ferroelectric device, in which mica was chosen as the substrate of the device. LaNiO3 (LNO) and ferroelectric solutions were spin-coated onto the mica in turn to fabricate films, and the 1.2 mm-diameter indium tin oxide (ITO) electrodes were then prepared on the films by magnetron sputtering; finally, the whole device was adhered to a 1-mm-thick acrylic plate. The LNO and ferroelectric films were fabricated by the sol-gel method, and the full description of the various ferroelectric materials and specific fabrication methods is described in detail in Methods and Supplementary Note 1. To maintain consistent test conditions throughout the experiments, all samples had the same dimensions. The process began with sequentially spin-coating LNO and a ferroelectric sol onto a 3 cm × 3 cm mica substrate, ensuring complete coverage with mica for both layers. Subsequently, an array of multiple ITO electrodes, each with a diameter of 1.2 mm, was sputtered onto the ferroelectric film. In the case of a single device, the laser spots for both 1064 nm and 405 nm wavelengths were of the same diameter as the ITO electrodes, and during operation, these spots were well aligned with the electrodes. This alignment ensured that the effective operational area of the entire device was confined to a circular area with a diameter of 1.2 mm. Figure 1c shows an SEM image of a cross-section of ferroelectric film, taking lead strontium titanate (PST) film as an example, the ITO/PST/LNO layered structure can be clearly seen and the thicknesses of each layer are 207 nm, 91 nm, and 223 nm, respectively. Supplementary Fig. 1c shows a photographic image of the PST film. Furthermore, the thicknesses of other films are shown in Supplementary Fig. 2, and explanations for the slight variations in film thicknesses are described in Supplementary Note 2. While the 405 nm or 1064 nm lasers are operating, the spots of the two lasers are focused on the same position on ITO, and Fig. 1d exhibits the currents excited by the lasers. In addition, a coupling of the PyE and PvE occurs when both lasers are operating simultaneously, as demonstrated on the right side of Fig. 1d.

Evaluation metrics of coupling performance

Our previous work has demonstrated that the performance of BTO film is enhanced by the coupling of PyE and PvE24. Based on this, we systematically tested the coupling performance of different ferroelectric films and sequenced the coupling performance of all materials. It is worth noting that the volume is important for the performance of the device, and the comparison of energy versus energy density is equivalent when each device has the same volume. Therefore, we have adopted a rigorous approach to controlling the volume of each sample so that is as consistent as possible to compare the output performance, and specific measures are described in Methods. From Fig. 1e, it can be seen that the charge and energy generated by the coupling of PyE and PvE are larger than the values when working individually, presenting a clear coupling enhancement. In order to visualize the degree of coupling enhancement, the coupling factor is often used as an evaluation metric. Due to the irregularity of the coupled current signal, the charge before and after coupling can be obtained by integrating the current-time, and accordingly the comparison of KC,Q enables a more detailed evaluation of the coupling performance, as illustrated in Fig. 1e, which shows the calculation KC,Q. Significantly, the KC,Q is a dimensionless metric that represents the magnitude of the change in the post-coupling charge relative to the pre-coupling charge as a relative value. To obtain electronic devices with a high charge density, the charge is of critical importance; therefore, ferroelectric films with optimal coupling performance are characterized by both a large KC,Q and a high charge output. Based on this approach, a quantitative metric to characterize the charge generation capability due to coupling is proposed, termed the Yang’s charge (QY), which is the multiplication of KC,Q and the level of charge after coupling. In addition, the output energy (E) is also considered as an important quantitative indicator here to evaluate the practical application capability of the electronic device. In general, this work mainly focuses on KC,Q, QY and E as evaluation indexes to assess in detail the coupling performance of various ferroelectric materials, and thus provides an unique reference for selecting ferroelectric films with optimal coupling performance.

Screening strategies of the power source and their light intensities for photovoltaic and pyroelectric effects

Before testing the PyE and PvE properties and performance of various ferroelectric films, it is important to select the optimal power source for both effects. Typically, the PvE is activated by short-wavelength lasers with high energy, whereas the PyE requires a rapid change in temperature using long-wavelength laser. For bulk ferroelectric materials, heating-cooling plates are commonly employed as the driving force for pyroelectric nanogenerators, which are capable of inducing temperature variations. However, these plates exhibit relatively heating and cooling rates. Furthermore, the extended heating process leads to substantial heat loss into the surrounding air, with minimal heat actually being absorbed by the film. In contrast, when a 1064 nm laser is used, a significant amount of heat is generated upon irradiation. During the instantaneous activation or deactivation of the laser, the surface temperature of the device rapidly rises or falls, resulting in a significant reduction in response time compared to using heating-cooling plates. In this work, a laser capable of producing a large temperature difference which can be rapidly switched on/off is utilized as the driving source for PyE. Notably, the power source of PyE should avoid generating a photovoltaic response, hence lasers with low energy and long wavelength should be selected. Supplementary Fig. 3 illustrates the selection strategy of the laser power source and light intensity for PyE and PvE. The source of PvE is selected to obtain a large photovoltaic response while avoiding a temperature change and a pyroelectric response. In addition, the selection criteria for the light intensity of each source were aimed at obtaining a large KC,Q.

Supplementary Fig. 4 demonstrates the short-circuit current of BTO film when irradiated by various lasers with wavelengths in the range of 405 nm–1064 nm, and the results show that a maximum and stable photocurrent is generated when the material is irradiated by low wavelength 405 nm laser. According to the formula Ephot = hc/λ, the shorter the wavelength, the larger the photon energy (Ephot), thereby generating a larger photocurrent and voltage. To ensure that a range of ferroelectric film materials can generate a PvE, the photon energy should be as large as possible. When the laser wavelength is set to 405 nm, the photon energy can be computed to be 3.06 eV, and it is worth noting that all ferroelectric films exhibit a photovoltaic effect when exposed to 405 nm laser radiation. Importantly, the open-circuit voltage is not strictly constrained by the bandgap of the material due to the presence of a spontaneous polarization within ferroelectric materials. This spontaneous polarization facilitates the separation of photogenerated electron-hole pairs. Consequently, the photovoltaic effect can continue to occur even when the photon energy is lower than the bandgap of the ferroelectric material, since the spontaneous polarization plays a significant role. Supplementary Fig. 5 illustrates the band gaps of various ferroelectric films determined using the Tauc plot method, highlighting that some materials possess a band gap wider than the photon energy of a 405 nm laser. This indicates that the photovoltaic effect in ferroelectric thin films is not solely reliant on the bandgap. Nevertheless, the relatively high energy of photons associated with the shorter wavelength of 405 nm light was chosen for irradiating various ferroelectric materials. At longer laser wavelengths, the photon energy is relatively low and is insufficient to generate a significant photocurrent. However, laser irradiation causes the sample surface to heat up rapidly, thereby generating a pyroelectric signal. The amount of heat generated by different lasers varies, resulting in different currents and voltages. As the laser wavelength increases, the photocurrent gradually decreases, and a spike signal begins to appear at 650 nm. As the wavelength increases to 785 nm, the spike signal dominates and the photocurrent almost disappears. To verify that the current spike is generated by a pyroelectric response due to a change in temperature of the material, the intensity of 785 nm laser was continuously varied in the form of a sine wave, and the temperature of the sample was then also changed in a sinusoidal form, as shown in Supplementary Fig. 6. As the rate of temperature change reaches the maximum, the corresponding current also increases to a maximum, which is in accordance with pyroelectric law, where I = pAdT/dt27. As the wavelength is further increased, the photocurrent completely disappears and only a pyroelectric is generated, with a large current produced for a long wavelength of 1064 nm. Although lasers in the range of 785-1064 nm are capable of generating a pyroelectric response, the 1064 nm laser generates a larger pyroelectric signal and the photoelectric response completely disappears at 785 nm, which avoids the interference of PvE on PyE, and thus each contribution can be assess individually. The 1064 nm laser was selected due to the relatively low energy of its photons, which prevents the induction of a photovoltaic effect.

To verify the universality of this selection method, BSZT and PZT films were irradiated with various lasers, and the change in current and voltage with wavelength is consistent with that of BTO, as presented in Supplementary Fig. 7–Fig. 8, respectively. In addition, Supplementary Fig. 9 shows that the surface temperature of BTO, BSZT and PZT films is changing rapidly on switching the 1064 nm laser on or off, which satisfies the requirement for pyroelectric generation.

Determinants of K C,Q

While the PyE or PvE operate independently, a pulsed alternating current and square-wave direct current are generated, respectively. When the PyE and PvE work together, the coupled current is composed of the direct current due to the photoelectric effect and the alternating current and spike of the pyroelectric effect, as shown in Fig. 2a. For a pyroelectric generator, the charges are \({{{{\rm{Q}}}}}_{1}\) and \({{{{\rm{Q}}}}}_{1}^{{\prime} }\) as the laser is switched on and off, respectively, and the total charge in one cycle is \({Q}_{{PyE}}={Q}_{1}+{Q}_{1}^{{\prime} }\). The charge generated by photovoltaic generator is \({Q}_{{PvE}}={Q}_{2}\). For the coupled generator, the charge includes the PvE portion, denoted as \({{{{\rm{Q}}}}}_{3}\), and the PyE portion denoted as \({{{{\rm{Q}}}}}_{4}\) and \({{{{\rm{Q}}}}}_{4}^{{\prime} }\), implying that the total charge is \({Q}_{{Coup}}={Q}_{3}+{Q}_{4}+{Q}_{4}^{{\prime} }\). Previous work have reported that KC,Q is calculated by the following equation24:

$${K}_{C,Q}=\frac{{Q}_{{Coup}}}{{Q}_{{PyE}}+{Q}_{{PvE}}}$$
(1)
Fig. 2: Determinants of KC,Q and reasons for coupling enhancement.
Fig. 2: Determinants of KC,Q and reasons for coupling enhancement.
Full size image

a Current and integrated charge generated from the ferroelectric film in the presence of only PyE, PvE and the coupling of both. b KC,Q of PZT, BTO, BSZT as a function of the light intensity of driving source of PvE. (Intensity of PyE is 7.413 W/cm2). c KC,Q of PZT, BTO, BSZT as a function of the light intensity of driving source of PyE. (Intensity of PvE is fixed at 0.002 W/cm2 for BTO and BSZT and 0.172 W/cm2 for PZT). d Variation law of KC,Q of PZT when the light intensity of driving source of PyE and PvE is changed simultaneously. e KC,Q versus ICoup-PvE/IPvE for various materials. (Light intensity of driving source is 0.172 W/cm2 for PvE and 7.413 W/cm2 for PyE). f Temperature of the BTO surface when only PvE is present and when PyE and PvE are coupled. g I-V curves of BTO under ambient environment and coupling condition. h–j Energy band schematic of the ITO/BTO/LNO device for (h) PvE, (i) PyE, (j) and the coupling of PyE and PvE.

Since KC,Q is related to the light intensity and the resulting PvE response, also the temperature change and PyE response24, a reasonable regulation of the above parameters is essential to obtain a large coupling factor.

As a result, the effect of the light intensity and the PvE response on KC,Q is firstly investigated, where the light intensity of the 1064 nm laser is maintained at a maximum value of 7.413 W/cm2. Supplementary Fig. 10–Fig. 12 show the change in current before and after coupling of BTO, BSZT and PZT when the intensity and the PvE response is gradually increased, respectively. As the intensity of the 405 nm laser increases, the current due to the PvE response as well as the coupling increase significantly, and the post-coupling charge is slightly larger than the pre-coupling charge, as shown in Supplementary Fig. 13. The KC,Q of PvE at different intensities can be obtained by Eq. (1), as shown in Fig. 2b. As the light intensity and PvE response grows, the KC,Q of PZT, BTO and BSZT gradually decrease and all values of KC,Q are larger than 1, meaning that the PvE can generate a coupling enhancement effect at various light intensities, and KC,Q values are larger at lower light intensities. This is due to the fact that the temperature of the sample is close to room temperature and the photocurrent is small for a low intensity driving source of PvE, whereas the introduction of a PyE contribution increases the temperature of the sample significantly and the photocurrent. At a high intensity of the driving source of PvE, the temperature of the sample is also very high, and the introduction of PyE has little effect on temperature, thereby resulting in a small change in the generated photocurrent. Supplementary Fig. 14–Fig. 15 demonstrate that the introduction of the 1064 nm laser leads to higher temperature rise and a greater percentage increase in photocurrent when the driving source of PvE is at a low light intensity.

Similarly, the effect of the light intensity of the PyE driving source on KC,Q was further investigated in which the temperature of PyE was controlled by modulating the intensity of 1064 nm laser, while keeping the intensity of driving source of PvE constant, which was 2.379 mW/cm2 for BTO and BSZT, and 172.417 mW/cm2 for PZT (the signal could not be detected at a lower intensity for PZT). Figure 2c demonstrates a gradual increase in KC,Q of the three materials as the intensity of driving source of PyE increases, which is due to a more pronounced temperature rise that occurs with a higher intensity of the 1064 nm laser, resulting in more photogenerated carriers. Supplementary Fig. 16 demonstrates the temperature difference of the surface when PZT is irradiated by 1064 nm laser at different intensities, along with a 405 nm laser at a fixed intensity. In addition, the KC,Q at different light intensities is always greater than 1, implying that the PyE also generates a coupling enhancement.

Supplementary Fig. 17–Fig. 19 and Supplementary Fig. 20 show the improvement of current and charge, respectively, of BTO, BSZT and PZT after coupling when the intensity of driving source of PyE is gradually increased. To confirm the reliability of the conclusion that KC,Q decreases with the enhancement of the driving source of PvE and increases with the enhancement of the driving source of PyE, Fig. 2d shows the change in KC,Q of PZT when the intensity of driving source of both PvE and PyE are changed, and the results illustrate that the KC,Q reaches a maximum of 2.72 when the intensity of driving source of PvE is at a minimum and PyE at a maximum, which fully confirms the reliability of the conclusion.

The KC,Q of a variety of ferroelectric films were systematically evaluated and sequenced on the basis of the determination of the light intensities required for PvE and PyE to obtain the optimum KC,Q. In order maintain same testing conditions for all materials, an intensity of 7.413 W/cm2 was chosen for PyE and 0.172 W/cm2 for PvE. In addition, the experimental conditions and surroundings were kept the same for all materials. To minimize effects due to variations in the fabrication of the materials as well as in the testing process, three samples of each material were prepared. All ferroelectric materials were prepared by the sol-gel method, which is described in detail Methods. Figure 2e outlines the KC,Q of all materials, which shows that PZT has the largest KC,Q of 2.72 and the smallest KC,Q of 1.06 for BSZT, and KC,Q of other ferroelectric films is between these extremes. Moreover, KC,Q is larger than 1 for all materials, which implies a coupling enhancement for all materials tested.

To clarify the reason for the different materials having different charge coupling factors, the charges generated by various materials before and after coupling are first analyzed. A theoretical analysis combined with Fig. 2a reveals that the pyroelectric generator produces charge \({{{{\rm{Q}}}}}_{1}\) and \({{{{\rm{Q}}}}}_{1}^{{\prime} }\) when the laser is turned on and off, respectively, and \({Q}_{1}+{Q}_{1}^{{\prime} }=0\) by the principle of charge conservation, so that the total charge of \({{{{\rm{Q}}}}}_{{{{\rm{P}}}}{{{\rm{y}}}}{{{\rm{E}}}}}\) should be equal to zero. The photovoltaic generator generates a direct current and the charge is \({{{{\rm{Q}}}}}_{{{{\rm{PvE}}}}}={{{{\rm{Q}}}}}_{2}\). For the coupled generator, the charge includes \({{{{\rm{Q}}}}}_{3}\) for a PvE and \({Q}_{4}+{Q}_{4}^{{\prime} }\) for a PyE, the same can be analyzed so that \({Q}_{4}+{Q}_{4}^{{\prime} }=0\) by the principle of charge conservation. Therefore, KC,Q can also be expressed as:

$${K}_{C,Q}=\frac{{Q}_{{Coup}}}{{Q}_{{PyE}}+{Q}_{{PvE}}}=\frac{{Q}_{3}+{Q}_{4}+{Q}_{4}^{{\prime} }}{({Q}_{1}+{Q}_{1}^{{\prime} })+{Q}_{2}}=\frac{{Q}_{3}}{{Q}_{2}}$$
(2)

Theoretically, since the direct current generated by PvE is a constant value and the charge is obtained by integration over current-time, the above equation can be further expressed as:

$${K}_{C,Q}=\frac{{Q}_{3}}{{Q}_{2}}=\frac{{\int }_{0}^{t}{I}_{3}{dt}}{{\int }_{0}^{t}{I}_{2}{dt}}=\frac{{I}_{3}\cdot t}{{I}_{2}\cdot t}=\frac{{I}_{3}}{{I}_{2}}=\frac{{I}_{{Coup}-{PvE}}}{{I}_{{PvE}}}$$
(3)

where I2 is the current of PvE when 405 nm laser works individually, which can be denoted as IPvE, and I3 is the current of PvE during the coupling process, which can be denoted as ICoup-PvE. Therefore, the KC,Q can be simplified as the ratio of the current generated by PvE before and after coupling. Figure 2e demonstrates that the ICoup-PvE/IPvE obtained by Eq. (3) for different materials is almost identical to the KC,Q calculated by Eq. (1), illustrating the correctness of the simplified version of KC,Q. In addition, Supplementary Fig. 21 also shows the results of the comparison of the coupling factor values of BTO, BSZT and PZT calculated by Eq. (3) and Eq. (1) for PvE at different light intensities, which clearly shows that the values of the coupling factors calculated by the two methods are similar, which also demonstrates the reliability of the simplified method. The simplified KC,Q method significantly reduces the calculation time and labor costs. Furthermore, examining the factors that influence the KC,Q parameter offers clear insights and valuable direction for enhancing performance and optimization.

Principle of coupling enhancement

The above analysis illustrates that KC,Q is mainly determined by the ratio of the current generated by the PvE effect before and after coupling. The PvE, the current can be expressed as:

$${I}_{{PvE}}={I}_{s}\cdot \exp \left(\frac{{qV}}{{nkT}}\right)-1$$
(4)

where Is, q, n, k, T, and T are the saturation current, charge, ideality factor, Boltzmann constant, temperature, and bias voltage, respectively. The Is can be expressed as:

$${I}_{s}={A}^{*}\cdot S\cdot {T}^{2}\cdot \exp (-\frac{q{\varphi }_{b}}{{kT}})$$
(5)

where A* is Richardson’s constant, S is the electrode area, and qφb is the height of the potential barrier on the metal side at zero bias. Therefore, the photocurrent equation can be systematically expressed as:

$${I}_{{PvE}}={A}^{*}\cdot S\cdot {T}^{2}\cdot \exp (-\frac{q{\varphi }_{b}}{{kT}})\cdot \left[\exp \left(\frac{{qV}}{{nkT}}\right)-1\right]$$
(6)

Equation (5) illustrates that Is is positively correlated with T. Therefore, the change of temperature after coupling is an important reason for the rise of photocurrent with temperature. Figure 2f compares the surface temperature of BTO with only PvE and a coupled response, which shows that the heat generated by the 405 nm laser is low and the heating of the sample surface is minimal, thereby contributing only a little to the pyroelectric response. In addition, it can be seen that the surface temperature of the coupled device rises substantially. Moreover, the current after coupling also increases significantly, as shown in Supplementary Fig. 22, which is in accordance with the theoretical analysis. In addition, Supplementary Fig. 23 also demonstrates a pronounced temperature rise for a range of materials after coupling, meaning that the increase in temperature has a decisive effect on improvement of the coupling performance. According to Eq. (6), it can be analyzed that the electrode area is constant before and after coupling, the Richardson constant can be considered unchanged, and the ideality factor, n, is related to barrier inhomogeneities and interface effects28, but the temperature affects the potential barrier on the interface. Therefore, the generated photocurrent is also related to the potential barrier at the interface. Based on this knowledge, Fig. 2g shows the I-V curves of BTO at an ambient environment and the coupling case, and the inset is a local magnification view. It can be seen that the device exhibits good unidirectional conductivity at room temperature, indicating the presence of a Schottky junction at the interface. When the two lasers are irradiated simultaneously, the device still has a unidirectional conduction characteristic, however the currents at the high and low resistance ends produce a change. The potential barriers at room temperature and coupling can be calculated by the formulae 29:

$${{\mathrm{ln}}}\left({I}_{{HRS}}/{I}_{{LRS}}\right) \sim -\Delta \varphi /{kT}$$
(7)

where IHRS and ILRS correspond to the currents at the high-resistance and low-resistance ends, and ∆φ is the Schottky barrier height. The surface temperature of the device is 27 °C at room temperature and is 48.8 °C when the two lasers are acting simultaneously, which can be substituted into Eq. (7) and indicate that the value of the barriers after coupling are smaller than the value before coupling. This result indicates that the Schottky barrier at the interface is reduced in the coupled case, which is another important reason for the significant enhancement of the photocurrent.

In addition, the temperature changes result in variations of carrier concentration in the ferroelectric films, which also affects the photocurrent, therefore a comprehensive analysis of the coupling enhancement due to the temperature rise is also carried out in terms of the energy band. Taking BTO as an example, the work functions of LNO, BTO and ITO are 4.36 ~ 4.42 eV, 4.9 eV and 4.2 eV, respectively24, and electrons are the main carriers in BTO film24. Thus, the Schottky junction exists mainly at the LNO/BTO interface and forms a built-in electric field directed from BTO to LNO. Since the difference in work function between BTO and ITO is minimal24, an ohmic contact is generated at BTO/ITO interface. In addition, Supplementary Fig. 24 demonstrates that the I-V curves of the materials are essentially the same as that of BTO, with the circuits conducting when a negative bias is applied and cut off when a positive bias is applied. As a result, the devices fabricated from all materials have unidirectional conductive properties and the interface positions of the Schottky junctions are consistent with that of BTO.

For the device based on LNO/BTO/ITO, the electron-hole pairs are generated when irradiated with 405 nm laser. The separation of electron-hole pairs occurs driven by the built-in electric field, and the electrons flow from BTO through ITO and then to the external circuit to produce a photocurrent, which is the principle of PvE, as shown in Fig. 2h. When irradiated with a 1064 nm laser, the charge at the interface is reduced due to heating of the material, thereby allowing electrons in the LNO to be transferred to ITO via an external circuit, resulting in an interfacial pyroelectric response27,30, as shown in Fig. 2i. Interfacial pyroelectrics break through the traditional Curie temperature limitation and are able to generate electrical signals when the temperature is higher than the Curie temperature. When PyE and PvE act simultaneously, the coupling effect not only lowers the potential barrier at the interface but also triggers the formation of additional electron-hole pairs in the BTO materials, which results in an increase in the carrier concentration. More electrons are transferred driven by the built-in electric field, which leads to an enhancement of the photocurrent significantly, resulting in a coupling enhancement effect, as shown in Fig. 2j. Notably, as the temperature decreases, the interfacial barrier rises and the carrier concentration decreases, potentially resulting in a reduced photocurrent, which is unfavorable in terms of a coupling enhancement. In addition, lower temperatures lead to an accumulation of charge at the interface, intensifying the built-in electric field. This phenomenon can, on one hand, enhance the pyroelectric signal with, on the other hand, it can prompt more carriers to migrate. The interplay of these various mechanisms may lead to intriguing alterations in the post-coupling charge dynamics.

Testing for repeatability, stability and durability

In order to verify the stability and reproducibility of the test method for obtaining the optimal KC,Q by modulating the light intensity of driving sources for photovoltaic and pyroelectric effects, repeatability experiments were carried out with BTO films as experimental samples. To obtain accurate results, at least three samples of each material were prepared to avoid variation in material fabrication or experimental operation, and the three samples were tested at different time of the day. Figure 3a shows that the current of the three samples of BTO are steady with only minor differences when only PyE, PvE and coupled effects are generated. It is unavoidable that the currents of the three samples are not identical due to the differences in film uniformity and resistivity of the different samples prepared by the sol-gel method; nevertheless, the current produced by the three samples are similar. Figure 3b shows the charges and corresponding KC,Q of the three samples in response to PyE, PvE and coupled effects. The results indicate that the KC,Q is similar despite the slight difference in charge, which illustrates that KC,Q is not significantly affected by the material fabrication and indicated that the test method is reproducible. Figure 3c demonstrates that KC,Q of BTO film has long-term stability after 20 and 40 days. Four BTO samples were tested each time, and the results indicated that the KC,Q of the four samples were uniformly distributed, fluctuating in the range of 1.1–1.2, and no decay of KC,Q was observed after 40 days. Due to the impact of environmental factors such as oxygen, temperature, and humidity, the performance of materials undergoes some changes over time, with varying degrees of influence on the different materials. Furthermore, the longer the level of exposure, the more significant the changes in properties may become, resulting in some fluctuations in the KC,Q values of different samples after 40 days. In addition, since the materials are prepared using the sol-gel method, achieving complete homogeneity among the various films is not possible, leading to an inevitable variation in the prepared devices. Consequently, this leads to small differences in material resistance, which in turn contributes to the variations in KC,Q among various samples. In addition, Fig. 3d shows the excellent durability of the current of BTO after 60 cycles while only a PvE response is present, which indicates that the test technique will not cause any damage to the sample, and has good stability and reliability. Moreover, Supplementary Fig. 25 shows the current of the device after 60 cycles of continuous irradiation with only 1064 nm laser and coupling conditions, and the results also demonstrate good stability.

Fig. 3: Testing for repeatability, stability and durability.
Fig. 3: Testing for repeatability, stability and durability.
Full size image

a Repeatability test of current before and after coupling of three BTO samples. b Charge and KC,Q before and after coupling of three BTO samples. c Stability of KC,Q for four BTO samples after 20 and 40 days. d Durability of the current of BTO sample after 60 cycles while only PvE is present.

Quantitative series and specific values of coupling performances

Based on the above verification of stability and reproducibility, the coupling performance of various typical ferroelectric films was tested and sequenced, and the results are shown in Fig. 4. All the experiment conditions were controlled to be consistent during the testing process. In order to avoid errors in the fabrication of the materials as well as in the testing process, four samples were tested for each material, and the average of the test results was taken for comparison. Firstly, the KC,Q of all materials are greater than 1, implying that all films exhibit a coupling enhancement, with BSZT being the smallest at 1.06 and PZT being the largest at 2.72. Notably, despite the fact that the KC,Q of PZT is the largest, the amount of charge generated by PZT while generating PyE, PvE, and coupling effects is extremely rare relative to other materials, as demonstrated in Supplementary Fig. 26. Therefore, KC,Q is mainly used to reflect the relative enhancement effect of the material after coupling compared to before coupling, which has a clear advantage when studying the proportion of change in charge after coupling, but it is not sufficient to indicate the total charge generation capability of the material. For a high-performance coupled generator, it is necessary to ensure that the actual output is large, in addition to an enhancement effect. Based on this, a new parameter index termed Yang’s charge is proposed, which is expressed as QY = KC,Q*QCoup, where QCoup refers to the output charge after coupling. For different devices, when the preparation method of the device, the test environment, and other pertinent factors are identical, comparisons of QCoup among different materials indeed hold practical significance. A larger QCoup indicates a stronger charge generation capability of the device, which is informative for selecting high-performance ferroelectric materials. For these comparisons to be valid, it is important to ensure that the structure of various devices and the experimental environment are rigorously controlled to be consistent. This aspect has been considered in detail in Methods. The significance of QY is that it combines the relative enhancement ability of the material after coupling and the absolute ability of the material to generate charge, which is more representative of the coupling performance. The larger the value of QY, the more charge will be produced, and the charge generated by the material increases significantly when heated. This demonstrates that the coupling effect plays a crucial role in enhancing performance, making it suitable for the preparation of high-performance photovoltaic-pyroelectric coupling devices. The middle part of Fig. 4 compares the QY of different materials, and shows that the QY of PZT is extremely small, while that of PST, PTO, and BNT is clearly larger than other materials. Correspondingly, Supplementary Fig. 26 illustrates that the output charge of these materials is significantly better than that of other materials in the occurrence of PvE and coupling effect, which indicates that these materials have both larger KC,Q and larger output charge, which provides a good reference for the selection of materials for coupling generator with excellent overall performance.

Fig. 4: Quantitative series of coupling performances of various ferroelectric films.
Fig. 4: Quantitative series of coupling performances of various ferroelectric films.
Full size image

Quantitative series of KC,Q, QY, and Energy of various ferroelectric films under the same conditions when PyE and PvE are coupled. (All materials were tested at a light intensity of 0.172 W/cm2 for PvE and 7.413 W/cm2 for PyE. Charge in QY was generated in three cycles, each cycle being 40 s.).

In addition, the output energy is of representative significance as an important indicator of the performance of electronic devices. Therefore, the output energy of the above materials after coupling is also compared, which is calculated by the following equation:

$$E=\bar{P}\cdot t=\bar{I}\cdot \bar{U}\cdot t$$
(8)

Where \(\bar{P}\), \(\bar{I}\) and \(\bar{U}\) are the average power, current and voltage respectively, and t is working time. Since the coupled current is irregular, the average current is calculated by integration as follows:

$$\bar{I}=\frac{{\int }_{0}^{t}I{dt}}{t}=\frac{{Q}_{3}+{Q}_{4}+{Q}_{4}^{{\prime} }}{t}=\frac{{Q}_{3}}{t}={I}_{{Coup}-{PvE}}$$
(9)

Since a generator based on ferroelectric film can be regarded as a capacitor31, the average voltage can be calculated from the ratio of the total charge to the capacitance of the sample, which is shown in Supplementary Fig. 27 for a variety of materials, measured at 1 kHz working frequency. The capacitance of the device can be expressed as C = (ε33 ∙ S)/h, where ε33 is the dielectric constant of the device, S is the effective area, in this case the area of the ITO electrode, and h is the thickness of the device. Combined with this equation the average voltage can be calculated as:

$$\bar{U}=\frac{{Q}_{{Total}}}{C}=\frac{{Q}_{3}+{Q}_{4}+{Q}_{4}^{{\prime} }}{C}=\frac{{Q}_{3}\cdot h}{{\varepsilon }_{33}\cdot S}=\frac{{I}_{{Coup}-{PvE}}\cdot \,t\,\cdot \,h}{{\varepsilon }_{33}\cdot S}$$
(10)

Substituting Eqs. (9) and (10) into (8), the energy can be expressed as:

$$E={I}_{{Coup}-{PvE}}\cdot \frac{{I}_{{Coup}-{PvE}}\cdot t\cdot h}{{\varepsilon }_{33}\cdot S}\cdot t=\frac{{{I}_{{Coup}-{PvE}}}^{2}\,\cdot \,{t}^{2}\,\cdot \,h}{{\varepsilon }_{33}\cdot S}=\frac{{Q}_{3}^{2}\,\cdot \,h}{{\varepsilon }_{33}\cdot S}=\frac{{Q}_{3}^{2}\,}{C}$$
(11)

The total energy generated by various materials in one cycle (40 s) was calculated by Eq. (11) and is shown at the bottom of Fig. 4. The results show that the actual output energy of PZT is the lowest, which further illustrates that KC,Q is not sufficient to characterize the coupling performance. Moreover, PST, BNT and PTO exhibit the best output energies of all the materials evaluate, which is generally consistent with the quantization sequence of QY. By comparing the sequences of QY and energy, it can be found that while the position of some materials with similar performance are swapped, such as BST and BTO, the sequence of materials between the two coupling performances similar overall, which also demonstrates the desirability of QY as an indicator for the evaluation of the coupling performance. QY and E have been selected as the primary quantitative indicators for performance evaluation based on their distinct advantages. Firstly, both indicators effectively circumvent the challenge posed by irregular signals, such as coupled currents and voltages, by utilizing reliable integration methods for their calculation. Secondly, QY not only encapsulates the absolute output performance of the device but also highlights the impact of the coupling effect on its performance, thereby offering a more comprehensive and rigorous assessment compared to the traditional KC,Q parameter. On the other hand, E provides a more intuitive measure of the device’s energy harvesting capacity, which is pivotal for determining its practical applications. Despite the similarity between E and QY, the E parameter does not capture the contribution of the coupling effect to the output performance. Based on the analysis presented above, QY is a more representative metric for evaluating pyroelectric and photovoltaic coupling performance.

In addition, the relationship between QY and E is analyzed theoretically in the following: Eq. (11) demonstrates that E = Q32/C, and QY can be expressed as follows combined with Eq. (2):

$${Q}_{Y}={K}_{C,Q}*{Q}_{{Coup}}=\frac{{Q}_{3}}{{Q}_{2}}\cdot \left({Q}_{3}+{Q}_{4}+{Q}_{4}^{{\prime} }\right)=\frac{{Q}_{3}^{2}\,}{{Q}_{2}}$$
(12)

Combining Eqs. (11) and (12), the relationship between QY and E can be expressed as:

$$E=\frac{{Q}_{2}}{C}\cdot {Q}_{Y}$$
(13)

It is apparent that there exists a linear relationship between QY and E, revealing that the QY parameter is not merely an indicator tied to the coupling factor and output charge; rather, it plays a pivotal role in determining the device performance. A higher QY signifies a greater output energy of the device, thus validating the reliability of QY as a metric for assessing the energy harvesting capability of a device. Furthermore, it is worth noting that the sequence of QY and E values may not align perfectly across different devices due to variations in capacitance C and charge \({{{{\rm{Q}}}}}_{2}\). To illustrate the relationship between QY and E, Supplementary Fig. 28 outlines the positions of various materials in serialized tables. The results indicate that both curves exhibit similar trends, and the ranking of materials based on their energy output closely mirrors their ranking in terms of QY. This experimental evidence further underscores the strong correlation between QY and E.

In summary, the above quantitative series can be used to visualize the magnitude of the coupling performance and actual output performance of various ferroelectric materials, which provides a reference for material screening and performance optimization of ferroelectric films.

Table 1 shows the specific values related to the evaluation index of coupling performance. A KC,Q of less than 1 indicates that coupling will lead to a weakening of charge, and a KC,Q larger than 1 indicates an enhanced effect, and the larger KC,Q represents the larger proportion of the increase. A larger QY means the percentage increase in charge after coupling as well as the actual output charge are larger. The larger energy means the higher efficiency of converting light and heat energy into electrical energy. In addition, the temperature change of the sample surface before and after coupling of various ferroelectric films as well as the capacitance of the device are shown in the table. The different temperatures in Table 1 are due to the fact that the absorption of the 1064 nm laser, the transmittance of the device, and the heat dissipation rate of various materials are different. By increasing the temperature change, the current and charge after coupling can be improved and the output energy can be enhanced by decreasing the capacitance, which is of significant value in optimizing the performance of ferroelectric devices.

Table 1 Series table of the coupling performance of ferroelectric films
Full size table

Discussion

In summary, we have developed a methodology for evaluating the performance based on a coupling pyroelectric and photovoltaic effects of different ferroelectric films and have quantitatively ranked the degree of coupling and performance. The 405 nm and 1064 nm laser activated ferroelectric films generated photovoltaic and pyroelectric effects under a controlled environment, and coupling enhancement effects were observed for all films. The quantitative series include KC,Q, QY and energy, in which KC,Q is primarily employed to reflect the change of charge generated after coupling relative to that prior to coupling, QY takes into account both the relative enhancement of charge and the absolute ability to generate charge after coupling, and the output energy demonstrates the performance of the coupling device in terms of energy harvesting visually. In addition, the root causes leading to coupling enhancement are analyzed, which simplifies the calculation of KC,Q and facilitates the directional optimization of the coupling performance. The coupled enhancement of photovoltaic and pyroelectric effects enables ferroelectric devices to harvest more energy, and the above quantified sequences provide a good reference for coupled generators based on ferroelectric films in terms of selection of materials, optimization of properties, and energy harvesting, which may also drive the development of self-driven electronics and photodetectors.

Methods

Preparation of ferroelectric solutions

Solutions of various ferroelectric materials were first prepared by the sol-gel method, and then the solutions were heated and dried to obtain ferroelectric films. For BSZT, 1.409 g of tetrabutyl titanate solution, 0.432 g of zirconium n-butoxide, and 1.1 ml of glacial acetic acid were firstly mixed, then 1.5 ml of acetylacetone was added and the solution was stirred thoroughly. Next, 0.387 g of barium acetate and 0.727 g of strontium acetate powder were fully dissolved in glacial acetic acid. Finally, the above solutions were stirred, and 25 ml of ethylene glycol methyl ether and 5 ml of acetic acid were added, and the solution was stirred thoroughly until clear to obtain a 0.1 mol/L solution of BSZT (Ba0.3Sr0.7Z0.18Ti0.82O3). The preparation methods for other materials are described in Supplementary Note 1.

The LNO solution was also prepared by the sol-gel method. 2.989 g of nickel acetate tetrahydrate was added to 24 ml of glacial acetic acid and stirred until fully dissolved, then 5.167 g of lanthanum nitrate hexahydrate and 60 ml of anhydrous ethanol were added to the above solution and stirred until fully dissolved to obtain the LNO solution.

Fabrication of the device of LNO/ferroelectric film/ITO

Firstly, the LNO solution was spin-coated on a mica substrate by a spin coater with a rotational speed of 500 rpm and a working time of 30 s. Then, the samples were heated for 3 min each through three heating tables at 180, 400, and 680 °C to obtain a monolayered LNO film. The aforementioned steps were replicated six times to build up the desired thickness. Subsequently, the ferroelectric solution was spread onto the prepared samples using a spin coater with a rotational speed of 4000 rpm and a working time of 30 s. The samples were then heated sequentially for 5 min each using the three heating stations described above. Similarly, the above steps were repeated six times. Finally, the samples were heat-treated, with annealing conditions of 680 °C for PTO, PST, and PZT for 5 min, and 750 °C for 5 min for all other materials.

In order to make the thickness of various materials consistent, the concentration, the amount of spinning solution, and the rotational speed and acceleration of the homogenizer are maintained to same during the experiment. Moreover, the temperature of the heating table, the heating time of each film, and the number of layers also kept the same. The above parameters are controlled to ensure that the thicknesses of the films are as similar as possible, thereby reducing the effect of the thickness of the material on the experimental results. Furthermore, to ensure uniformity across all devices in terms of test conditions, we ensured that the spot area size, test environment, temperature, and light intensity remain consistent. These measures maximize the reliability of performance testing and comparison among various thin films under identical conditions.

Electrical measurement and characterization

The current, voltage and I-V curves were measured by a Keithley 2611B programmable electrostatic meter, and the cross section of PST samples were characterized by a scanning electron microscope (SU8020 cold field SEM). The capacitance of various materials was measured by LCR tester (11022). The temperature of the samples was measured by infrared thermography (Optris Xi 400), and the light intensity was measured by photometer (OPHIR).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.