Photostriction effect and electric properties of La-doped PMN-PT transparent ferroelectric ceramics

ABSTRACT

Ferroelectric ceramics display a remarkable photostriction effect, which could facilitate the prevention of electromagnetic interference and promote the integration and miniaturization of actuator systems, indicating potential applications in glimmer sensors. In this work, highly transparent La-doped 0.9Pb(Mg_1/3Nb_2/3)O₃-0.1PbTiO₃ (0.9PMN–0.1PT) ceramics were synthesized using a solid-state reaction method in an oxygenated atmosphere. XRD and SEM were used to analyze the phase structure of the dense samples, which displayed a pure perovskite structure, with no trace of a pyrochlore phase. With La doping, the phase transition temperature shows a decreasing trend, and its dielectric constant could reach up to 20,000. The P-E loop of the samples was gradually refined into a straight line with an increase in the doping amount. After La doping, the transparency of the 0.9PMN–0.1PT ceramics was significantly enhanced and reached up to 61% at the near-infrared band. Finally, the photogenerated voltage and deformation of the La-doped 0.9PMN–0.1PT ceramics were investigated under ultraviolet light of various intensities. After 20 s of irradiation at an intensity of 800 mW/cm², the maximum voltage and deformation, with values of 40 V and 250 nm, respectively, were both found in the 1%La-doped 0.9PMN–0.1PT ceramics.

Graphical abstract

Photostrictive effect in 1%La-doped 0.9PMN-0.1PT and pure 0.9PMN-0.1PT transparent ceramics.

KEYWORDS:

• Photostriction
• ferroelectric
• Raman
• piezoelectric

1. Introduction

In ferroelectrics, materials produce a photogenerated voltage under irradiation with high-energy ultraviolet light and then undergo mechanical deformation through the action of an inverse piezoelectric effect. This cross-coupling between the anomalous photovoltaic effect and inverse piezoelectric effect results in a photo-deformation phenomenon called the photostriction effect [1,2]. Based on the electro-optical effects of ferroelectric materials, many devices such as electro-optical modulators, fiber grating sensors, optical switches, and electro-optical attenuators have been widely used in the field of optical communication due to the advantages of their small size, controllable voltage, and fast response [3–15].

The photoelectric effect can be external or internal and refers to the phenomenon in which substances release electrons under the action of light. In general, the photovoltaic voltage will not exceed the electronic bandgap width (E_g) of a material. Transparent ferroelectric ceramics characterized by high optical transmittance in visible and near-infrared and by considerable electro-optic (EO) coefficients are promising materials for optoelectronic devices. In ferroelectric materials, there is an anomalous photovoltaic effect; the photogenerated voltage is not limited by the electronic bandgap width of the crystal material, and its value can reach up to 10³ ~ 10⁵ V/cm [16–19]. In 1985, Uchino et al. investigated the effect of the incident light wavelength on the photo-generated current of lead lanthanum zirconate titanate (PLZT) ceramics and found that photo-deformation can reach 10⁻⁶ orders of magnitude when exposed to light with a wavelength near 365 nm [20]. Visible light-driven photostriction of 10⁻³ orders of magnitude was found in a Bi(Ni_2/3Nb_1/3)O₃-PbTiO₃ material [21]. In practical applications, the ferroelectric properties of rare-earth doped materials should be maintained in order to obtain greater photostriction effect.

PMN-PT transparent materials are considered to be the most promising nonlinear optical material due to their high electro-optical coefficient and strong optical refraction effect. At room temperature, its electro-optical coefficient is 100 times higher than that of LiNbO₃ crystals and 2 ~ 5 times higher than that of PLZT ceramics. With excellent transparency over a wide wavelength range of 400–7000 nm, it is suitable for almost all optical applications across the visible to mid-infrared light spectrum [22,23]. PMN-PT transparent ferroelectric ceramics can be prepared using sintering technology combined with an oxygenated atmosphere and hot pressing, the products of which can be applied in many devices, such as variable optical attenuators, polarization controllers, variable gain tilt filters, and tunable optical filters [22,24]. Recently, rare earth doped PMN-PT materials have shown ultra-high piezoelectric properties and novel optical properties [25–30]. For La-doped 0.75PMN–0.25PT transparent ceramics synthesized in a two-stage sintering process in an oxygenated atmosphere and with hot pressing, the secondary electro-optical coefficient can reach up to 66 × 10⁻¹⁶ m²/V², and, under these conditions, the transmittance has been reported as 58% at 633 nm, and 65% at the near-infrared band [31]. For Sm-doped PMN-PT transparent ceramics prepared using the same conditions, the transmittance is 67% at 633 nm and 69% at the near-infrared band [9,32–34].

Until now, there are few reports on direct measurement of photostriction effect in PMN-PT ceramics. In this work, the photostriction effects of rare earth doped PMN-PT ceramics are investigated in detail, including simplified transparent ceramic synthesis, novel ferroelectric domain structures, and direct photostriction deformation measurement. First, transparent La-doped 0.9PMN–0.1PT ceramics were prepared using a sintering process within an oxygenated atmosphere. Second, the structure, dielectric, ferroelectric, and transmittance properties of pure and La-doped 0.9PMN–0.1PT ceramics were measured, respectively. Before directly measuring the degree of photostriction, the Raman spectrum of the samples was characterized, and the photo-induced strain was determined under ultraviolet light. Finally, the photogenerated voltage and deformation were measured and analyzed.

2. Materials and methods

La-0.9Pb(Mg_1/3Nb_2/3)O₃-0.1PbTiO₃ (La-0.9PMN–0.1PT) ferroelectric ceramics were prepared using a B-site cation precursor method. The MgNb₂O₆ precursor materials were first fabricated at 1200°C for 4 h using Nb₂O₅ (99.99%) and MgO (99.9%) powders. The Pb₃O₄ (99%), La₂O₃ (99.99%), MgNb₂O₆, and TiO₂ (99.9%) powders were wet-mixed using zirconium ball milling with alcohol for 10 h. Second, the mixed powders were calcined at 850°C for 2 h and subjected to vibratory milling in alcohol with a binder for 10 h. After drying at 80°C for 10 h, the powders with poly(vinyl alcohol) as binder were pressed into pellets of 5 mm thickness and 20 mm diameter under a uniaxial pressure of 500 MPa. At 550°C, the binder was burned out over 2 h, and the samples were sintered in an oxygenated atmosphere at 1250°C for 4 h [35,36]. The average relative density of the samples was 96% calculated by applying the Archimedes’ principle.

The crystal structure of the samples was determined using data collected via X-ray diffraction (Bruker D8 Advance) operating with a Cu Kα radiation source under 40.0kV and 30.0 mA. The morphology of the samples was measured using a scanning electron microscope (SU1510). For further electric measurement, silver paste was fired on both sides of the samples at 600°C for 10 min to form the electrodes. The samples were poled in silicone oil at room temperature for 30 min using a DC electric field with a strength of 1 kV/mm. The temperature-dependent dielectric properties were determined using an LCR meter (E4980A) connected to a computer-controlled cooling – heating stage. The P-E loop was measured with precision premier II model standard ferroelectric devices. The measuring principle was based on an improved Sawyer–Tower measuring circuit. The transmittance was measured with a spectrophotometer (Lambda 950). For the photostriction effect measurement, the variable power UV light provided by an LED-UV surface light source (ULAS30–250) with a wavelength of 365 nm was utilized. The deformation of the PMN-PT ceramic specimens was measured using a dispersion confocal displacement meter (Keyence). The voltage sensor used was a high impedance voltmeter (Keithley 6517A).

For the 0.9PMN–0.1PT transparent ceramics, the theoretical transmittance T can be calculated using the following formula:

(1) $\mathit{R}=(\mathit{n}-1{)}^2/{\left(\mathit{n}+1\right)}^2$(1)

(2) $\mathit{T}={\left(1-R\right)}^2\exp \left(-\mathit{\beta t}\right)$(2)

where n is the refractive index, R is the reflectivity, t is the thickness of the sample, and β is the effective scattering constant, which is related to the absorption and scattering of the incident light in the material.

The refractive index can be considered to be fixed in 0.9PMN–0.1PT transparent ferroelectric ceramics [37]. At 633 nm, the refractive index, n, is about 2.44. In the near-infrared band, its refractive index n is about 2.3 [31]. Therefore, the theoretical transmittance of the 0.9PMN–0.1PT transparent ceramics can be calculated using Formula (1) and (2), which results in values of 68% at 633 nm of visible light and 71% at the near-infrared band.

3. Results

3.1. Microstructure

Figure 1 shows the XRD patterns of the 0.9PMN–0.1PT ceramics doped with different amounts of the La element (0, 1%, 2%, and 4%). By comparing the diffraction pattern with the standard PDF card (No.01-076-9083), all the diffraction peaks can be calibrated in the range of 20°~80°, and no miscellaneous peaks appear. This indicates that the components with different doping amounts of La can form a binary solid solution phase, and their structure is a pure perovskite phase with no trace of a pyrochlore phase.

Figure 1. X-ray pattern 0.9PMN–0.1PT transparent ceramics doped with various amounts of La. The embedded image is a photo of La-doped 0.9PMN–0.1PT transparent ceramics.

High-transparency ceramics require the absence of residual impurities and pores. The density of such ceramics is evaluated using scanning electron microscopy (SEM). Figure 2(a–) presents the SEM images of the untreated natural surface of the 0.9PMN–0.1PT transparent ceramic samples doped with the La element. It can be observed that there is almost no porosity in all of the 0.9PMN–0.1PT ceramics, resulting in a high density. This is mainly due to the reduction of oxygen vacancies in ceramics in the sintering process of oxygen-containing atmosphere. According to our observations, the average grain size of all the ceramic samples is less than 10 μm. The average grain size of the La-doped samples is slightly smaller than that of the undoped ceramics, and decreases with an increase in the doping amount. The average grain size is of the order of 4.40, 4.29, 4.27, and 4.10 μm for pure, 1%La-, 2%La-, and 4%La-doped 0.9PMN–0.1PT ceramics, as shown in Figure 2(a–). The grain size statistics show that the grain distribution of 1%La doped 0.9PMN–0.1PT ceramics is more homogeneous. It means that the grain boundaries of ceramic grains are consistent and the mechanical properties are more uniform. In order to investigate the domain structure of La doped 0.9PMN–0.1PT ceramics, the samples were first thermal etched, and a non-domain structure was observed in Figure 2(e). Then, the samples were etched by mixed hydrofluoric acid and hydrochloric acid, and a clear domain structure with width on the order of 150–200 nm were obtained in pure and La doped 0.9PMN–0.1PT ceramics, as shown in Figure 2(f–). Finally, the EDS mapping was employed to investigate the La element dispersion in 0.9PMN–0.1PT ceramics, as shown in Figure 3(a–).

Figure 2. The SEM images of natural growth surfaces of La-doped 0.9PMN–0.1PT transparent ceramics: (a) pure 0.9PMN–0.1PT, (b) 1%La-doped 0.9PMN–0.1PT, (c) 2%La-doped 0.9PMN–0.1PT, (d) 4%La-doped 0.9PMN–0.1PT, (e) thermal-etched surface of 1%La-doped 0.9PMN–0.1PT, (f)-(g) hydrofluoric acid and hydrochloric acid etched surfaces of 1%La-doped 0.9PMN–0.1PT, and (h) hydrofluoric acid and hydrochloric acid etched surfaces of pure 0.9PMN–0.1PT.

Figure 3. EDS mapping of 1%La-doped 0.9PMN–0.1PT ceramics.

3.2. Dielectric properties

The temperature-dependent dielectric constant and loss of the 0.9PMN–0.1PT transparent ceramics with different La dopant amounts are shown in Figure 4. It can be seen from the figure that all the samples display frequency dispersion and relaxor phase transitions. The dielectric peak is not sharp, but wide, forming an envelope peak [38–40]. The peak value of the 0.9PMN–0.1PT transparent ceramics decreases with an increase in the frequency and shifts toward a high temperature, which indicates that the prepared 0.9PMN–0.1PT transparent ceramics have typical relaxor ferroelectric characteristics. As La is a trivalent cation, its use as a dopant will destroy the ferroelectric oxygen’s octahedral structure, resulting in the formation of polar nanoregions. In ABO₃ structures, the B site is occupied by a variety of ions in a disorderly manner, which also leads to the formation of nano-polar regions. This means that the normal ferroelectric properties of a material are affected and it will exhibit relaxation. Therefore, nano-polar regions are considered to be an important cause of the relaxation of relaxor ferroelectrics. As the La dopant amount increases, the maximum relative dielectric constant decreases gradually, and the Curie temperature (T_c) decreases continuously. When the La dopant amount exceeds 1 mol%, the T_c will be near 0°C. However, an accurate value of T_c cannot be obtained from the dielectric spectrum due to frequency dispersion. The T_c trend is consistent with previous literature report, in which the T_c will decrease by about 25°C with La dopant amount increases by 1% [41].

Figure 4. The temperature-dependent dielectric constant and dielectric loss in 0.9PMN–0.1PT transparent ceramics with various amounts of La dopant.

3.3. Ferroelectric properties

Figure 5 shows the P-E loops of the 0.9PMN–0.1PT transparent ceramics with different amounts of La dopant at a test frequency of 0.2 Hz. In contrast to the almost square P-E loop seen in the normal ferroelectric materials, the area of the P-E loop of the samples decreases and its inclination degree increases, indicating that the ferroelectric properties are weakened, and displaying the ferroelectric relaxation of the La-doped 0.9PMN–0.1PT transparent ceramics. With an increasing amount of the La dopant, the coercive field and remnant polarization decrease gradually, as shown in Figure 5. The small coercive field and remnant polarization equate to a small corresponding P-E loop area. In practical applications, this kind of material only needs to consume less energy to follow the changes in the applied external electric field. With rare earth element doping, the La ion will occupy the original Pb ion at the A site, which will inevitably destroy the long-range order of the original oxygen octahedron, and then reduce the stability of the ferroelectricity, affecting the inherent ferroelectric property of the materials.

Figure 5. The P-E loop in 0.9PMN–0.1PT transparent ceramics with various amounts of La dopant under different external electric fields.

3.4. Transmittance properties

For polycrystalline materials, there will be many pores, grain boundaries, and other uneven microstructures, which will cause the reflection, scattering, and refraction of the incident light, thus affecting the transmittance of the samples. The different refractive indices of the grains and grain boundaries inevitably cause the continuous scattering and refraction of light, thus affecting the transmittance of ceramics. Therefore, in cases of smaller grain size, no second phase and pores, and dislocation defects, the material will have a higher transmittance. In order to obtain significant electric properties, PMN-PT ceramics need to obtain appropriate domain sizes and domain wall densities in a uniform grain distribution.

In order to measure the effective transmittance using a spectrophotometer, the sample needs to be polished to reduce diffuse reflection. Figure 6(a) shows the transmittance curves of the La-doped 0.9PMN–0.1PT ceramics, in which the thickness of the sample is 0.5 mm and the test wavelength ranges from 380 to 1200 nm. The transmittance of the 0.9PMN–0.1PT ceramics is 21% at 633 nm and is 31% at the near-infrared band. The maximum transmittance of the La-doped 0.9PMN–0.1PT ceramics is more than 54% at 633 nm, and the maximum transmittance is more than 61% at the near-infrared band. As shown in Figure 6(a), the transmittance rate shows an increasing trend with an increase in the La dopant amount in the 0.9PMN–0.1PT ceramics.

Figure 6. The transmittance and (αhν)²-hν relationship in 0.9PMN–0.1PT transparent ceramics with various La dopant amounts.

In direct transition absorption, the electron directly transitions from the valence band to the conduction band under the action of incident photons, in which the absorption coefficient and the incident photon energy follow the Tauc equation [42,43]:

(3) ${\left(\mathit{\alpha }\mathrm{h\nu }\right)}^2=\mathrm{A}\left(\mathrm{h}\mathit{\nu }-{\mathrm{E}}_{\mathrm{g}}\right)$(3)

(4) $\mathit{\alpha }=\frac{1}{\mathrm{t}}\ln (\frac{1}{\mathrm{T}})$(4)

(5) $\mathit{v}=\frac{C}{\lambda }$(5)

where α is the absorption rate (unit: cm⁻¹), h is the Planck constant (its value is 4.1357 × 10–15 eV), A is the transition coefficient as a constant, hν is the incident photon energy (unit: eV), t is the sample thickness (unit: cm), T is the transmittance, and C is the speed of light (its value is: 3 × 10⁸m/s), λ is the wavelength, and E_g is the optical band gap.

As shown in Figure 6(b), the (αhν)²-hν relationship diagram of the PMN-PT transparent ceramics can be obtained in accordance with formulas (3–5). Through the linear fitting extension of the linear part of the curve, its intersection point with the horizontal coordinate provides the optical band gap (E_g). The E_g of La-doped 0.9PMN–0.1PT transparent ceramics can be obtained from Figure 6(b), and these values fall in the range of 3.15–3.20 eV. This value does not markedly differ from that of Pb(In_1/2Nb_1/2)O₃-Pb(Mg_1/3Nb_2/3)O₃-PbTiO₃ (PIN-PMN-PT) crystals (3.18 eV) [44], and lower than that of PLZT ceramics (3.35–3.67 eV) [45]. Meanwhile, these optical band gaps indicate that the 0.9PMN–0.1PT transparent ceramic samples doped with La can be transparent in the visible light range.

3.5. Raman spectrum properties

Raman spectroscopy has gradually become an effective method to reveal the photostriction effect of ABO₃ perovskite phase-structured ferroelectric materials [21,46–48]. The frequency difference between the Raman scattering and Rayleigh scattering is called the Raman shift, which is the frequency of the vibration or rotation of the molecule. The Raman shift is only related to the molecular structure of the material being measured. The Raman spectra of the PMN-PT samples were measured using a Raman spectrometer at room temperature. The wavelength of the light source was 532 nm, and the different laser powers used were 2 mW, 5 mW, 10 mW, and 20 mW, as shown in Figure 7(a–). In Figure 7(a,b), the peaks in the range between 560 and 590 cm⁻¹ display the redshift phenomenon with an increase in the laser power, indicating that there is a significant correlation between the excitation power intensity and the lattice strain, that is, the samples demonstrate the photostriction effect. The maximum redshift in the samples was 5.3 cm⁻¹. It is worth mentioning that there is no correlation between the photostriction effect of the ceramic sample and its components.

Figure 7. Raman spectra in PMN-0.1PT transparent ceramics with various La dopant amounts: (a) pure 0.9PMN–0.1PT, (b) 1%La-doped 0.9PMN–0.1PT, (c) 2%La-doped 0.9PMN–0.1PT, and (d) 4%La-doped 0.9PMN–0.1PT.

In the Raman spectra of ABO₃ perovskite materials, there are three major band regions: a low-frequency region (below 180 cm⁻¹) attributed to the Pb-BO₆ stretching mode, an intermediate-frequency region (from 180 to 400 cm⁻¹) assigned as B-O-B bending and O-B-O stretching, and a high-frequency region (from 400 to 1000 cm⁻¹) classified as B-O-B stretching. Spectra below 400 cm⁻¹ are mainly generated by the vibration of the cations at the A and B sites, and spectral bands in the 400 cm⁻¹ to 1000 cm⁻¹ region are mainly attributed to the bending and stretching between different ion chains in the BO₆ oxygen octahedron [49–51]. With the increase in the laser source power, some peaks in the Raman spectrum can be observed to be redshifted. This redshift represents a change in the physical size of the tensile strain produced by the ferroelectric material under light induction, that is, the photo-deformation caused by the photostriction effect of the material.

3.6. Photostriction effect

Figure 8 shows the variation curve of the end voltage between the electrodes of the 1%La-, 2%La-, and 4%La-doped 0.9PMN–0.1PT and pure 0.9PMN–0.1PT ceramics under various UV light intensities. The response speed of voltage at both ends of 0.9PMN–0.1PT ceramics was almost the same as the increase in light intensity and saturation value of the increasing voltage with light intensities of 100, 200, 400, 600, and 800 mW/cm². After 20 s of illumination, the photogenerated voltage reached 4, 8, 20, 28, and 40 V, respectively, and the values remained unchanged for 80 s. As shown in Figure 8(a–), after 20 s of light irradiation with an intensity of 200 mW/cm², the voltage rises to 8 V, and the expansion capacity of the 0.9PMN–0.1PT ceramic is 90 nm. When the light intensity increases to 800 mW/cm², the voltage rises to 40 V, and the expansion capacity of the ceramic is 250 nm. After the light source is turned off, the voltage drops and the amount of expansion is also reduced. The La-doped 0.9PMN–0.1PT ceramics demonstrate an obvious photovoltaic effect. The values of the photogenerated voltage and the amount of expansion and the light intensity are positively correlated. The voltage and expansion reached a stable value in 20 s. If the intensity of the light is increased, the response speed can also be improved, and the size of its expansion can also be controlled according to light intensity. As La dopant increases, ferroelectricity decreases at room temperature, as shown in Figure 5, indicating that photostriction is positively correlated with ferroelectricity. In practical applications, the ferroelectric properties of rare-earth doped materials should be maintained in order to obtain greater photostriction effect.

Figure 8. Photogenerated voltage and displacement in (a–b)1%La-doped 0.9PMN–0.1PT, (c–d) pure 0.9PMN–0.1PT transparent ceramics, and (e–f) comparison to 2%La- and 4%La-doped 0.9PMN–0.1PT transparent ceramics.

4. Conclusions

Highly transparent La-doped 0.9PMN–0.1PT ceramics with good electrical properties were prepared using an optimized process. First, the La-doped 0.9PMN–0.1PT-based transparent ceramics were prepared using a solid-state reaction and a sintering process in an oxygenated atmosphere. With an increase in amount of dopant, the remnant polarization of ceramics decreased, displaying a thin and inclined P-E loop. With a La dopant amount of 4%, the maximum value of transmittance in ceramics was 54% and 61% at 633 nm and near-infrared band, respectively. When irradiated to ultraviolet light with an intensity of 800 mW/cm², the photogenerated voltage and deformation reached up to 40 V and 250 nm, respectively, in 1% mol La-doped 0.9PMN–0.1PT ceramics. Due to their high photostriction effect, the transparent 0.9PMN–0.1PT ceramics can directly convert light energy into mechanical energy, achieving mechanical control without an external electric field. By using high-energy ultraviolet light as a driving source, wireless transmission and remote control can be achieved. This could allow for a reduction in actuator system’s size, and also for the removal of wire from devices, effectively diminishing the adverse effects of electromagnetic interference.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was partially funded by the start-up foundation for introduced talents of Wuxi University, grant number [No. 2021r001].