Giant electrostrain coefficient under low driving electric field in sodium potassium niobate piezoelectric ceramics with symmetrical bipolar strain

Abstract

Currently, achieving highly symmetrical bipolar strain and high electrostrain under low driving electric field remains challenging in piezoelectric materials. The designed potassium sodium niobate-based ceramics exhibit highly symmetrical bipolar strain and ultrahigh electrostrain coefficient (~2000 pm/V) under a low driving electric field of 8.4 kV/cm through A-site defect engineering and charge compensation. The highly symmetrical bipolar strain is related strongly to the lowly aligned defect dipoles by partially substituting A-site (Na⁺/K⁺) ions with Mn²⁺. The eye-catching performance is ascribed to the unique microstructure of atomic-scale polar nanoregions embedded in nano-domains (~34 nm) by tuning Na⁺/K⁺ ions deficiency and coexistence of multiple phases. Phase-field simulations reveal that flattened energy barrier and multiphase nanodomains interplay to boost electrostrain at low driving fields. This work provides an innovative way of designing lead-free piezoelectric materials with highly symmetrical bipolar strain and giant electrostrain coefficient at low driving electric field, promising for high-precision actuators applications.

The authors design potassium sodium niobate-based ceramics exhibiting highly symmetrical bipolar strain and high electrostrain coefficient (~2000 pm/V) under a low driving electric field of 8.4 kV/cm through A-site defect engineering and charge compensation.

Introduction

Piezoelectric actuators are widely used in various fields that require high precision and fast response, such as precise optical instruments, microelectromechanical systems, and robotics, etc.^1,2. The miniaturization and integration of electronic components have increased the performance requirements for piezoelectric actuators, particularly the need for a large electrostrain coefficient ($d_{33}^{*}$) under a low driving electric field (E_D)^3,4 ($d_{33}^{*}=\frac{S}{E_{\mathrm{D}}}$, S is electrostrain at E_D, E_D refers to the electric field under the highest $d_{33}^{*}$) and high precision with low displacement accumulated error (DAE) during the application process⁵. DAE can be mitigated by constructing stable defect dipoles, which facilitate the design of highly symmetrical bipolar strain^5–8. As a core element for piezoelectric actuators, the piezoelectric material with large $d_{33}^{*}$, low E_D, and highly symmetrical bipolar strain is urgently expected.

To achieve a large electrostrain coefficient, extensive research efforts have been devoted in recent decades. However, the manufacturing and applications of lead-based materials have been limited due to the raised concerns from the health and environmental issues^9,10. Therefore, lead-free piezoelectric material systems have garnered significant attention. Among the lead-free piezoelectric systems, the (Na_1/2Bi_1/2)TiO₃ ceramics exhibited highly asymmetrical bipolar strain and large $d_{33}^{*}~1500\mathrm{pm}/\mathrm{V}$ at fairly high E_D of 100 kV/cm¹¹. However, the high E_D may restrict its commercial applications due to reasons such as non-linearity and potential electrical breakdown³. In 2022, Huangfu et al. proposed an innovative strategy that is the synergistic effects (Fig. 1a1) of defects engineering and domains engineering, enabling (K,Na)NbO₃-based ceramics to achieve excellent $d_{33}^{*}$ (2100 pm/V) at E_D (50 kV/cm), which has reignited researchers’ attention to (K,Na)NbO₃-based systems¹². Subsequently, based on this strategy, Li/Sr-doped (K,Na)NbO₃ ceramics with a high $d_{33}^{*}$ (~3080 pm/V) at 50 kV/cm¹³ and (K_0.48Na_0.52)_0.99NbO_2.995 ceramics with a giant $d_{33}^{*}$ (~3500 pm/V) at 20 kV/cm¹⁴ are obtained, attributed to the interaction between the nano-domains and highly aligned defect dipoles ($V_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{O}}^{\bullet \bullet }$). Nano-domains constructed by multiphase coexistence can effectively decrease the energy barrier¹⁵. However, the introduction of $V_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{O}}^{\bullet \bullet }$, the energy profile in these nano-domains promotes defect dipoles alignment along the external electric fields, resulting in the global system falling into a deep potential well. Consequently, a significantly asymmetric Landau energy profile with slightly high energy barrier is formed as depicted in Fig. 1a2, resulting in asymmetric large bipolar strain and a slightly high E_D as shown in Fig. 1a3^12,13. Thus, a unique microstructure is urgently designed to significantly flatten and symmetrize the Landau free energy plot, which dramatically increases electrostrain under low E_D and improves symmetry of bipolar strain.

Fig. 1. Design schematic diagram of (K,Na)NbO₃-based ceramics with low E_D induced large electrostrain and high bipolar strain symmetry.

The relationship between structural design (a1 and b1), Landau energy (a2 and b2), and bipolar strain (a3 and b3). (LSH refers to local structural heterogeneity. The region marked in the red oval frame is enlarged as the smaller-scale octahedral structure on the right).

A flattening of the Landau free energy plot can decrease the energy barrier and lead to the system anomalously soft, i.e., a small external electric field can result in comparatively large piezoelectric response¹⁶. Nano-size domains formed in ceramic can effectively reduce the energy barrier during polarization switching¹⁷. However, it is difficult to further flatten and symmetrize the Landau free energy plot solely by reducing the domain size via establishing multiphase coexistence. Since polar nanoregions (PNRs) embedded in the nano-domains matrix are considered to act as accelerators during domain switching, flatten the Landau energy profile and then benefit the electrostrain and decrease E_D^18–23. We consider that constructing smaller-scale PNRs would significantly increase the interfacial energy, thereby further flattening of Landau energy profile, ultimately achieving significantly improved piezoelectric performance and reduced E_D¹⁹. It can be designed that the scale of PNRs is further reduced by constructing a higher level of local structural heterogeneity (LSH), which can be achieved by the ions (such as Mn²⁺ ions, 1.27 Å: 12 coordination number) with smaller ionic radius occupying A-site of perovskite structure^{11,20,24–26}, as shown in Fig. 1b1. Additionally, Mn²⁺ ions diffusion path increases due to the introduction of a reducing atmosphere during sintering and more spaces in the A-site of the (K, Na)NbO₃ lattice formed by K/Na ions deficiency, which promotes more Mn²⁺ ions to entry into A-site to form more atomic-scale PNRs (Fig. 1b1). Besides, the incorporation of Mn²⁺ ions into the A-site (K/Na) vacancies leads to the defects ${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }$ with positive charge, which inhibits the formation of oxygen vacancies^27,28. A decrease of $V_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{O}}^{\bullet \bullet }$ concentration results in low defect dipoles alignment along the external electric fields, decreasing the potential well. Thus, Landau free energy plots are symmetrized, thereby enhancing the bipolar strain symmetry¹⁵. Finally, we believe that low defect dipoles alignment and a unique microstructure that is atomic-scale PNRs embedded in the nano-domains will make a great contribution to induce flattened and symmetrized Landau free energy plot (Fig. 1b2) and then increase electrostrain, decrease E_D and improve symmetry of bipolar strain (Fig. 1b3).

Herein, we provide a readily applicable A-site defect engineering strategy to achieve ultra-high $d_{33}^{*}$ up to 2000 pm/V under a low driving electric field of 8.4 kV/cm, associated with highly symmetrical bipolar strain (β = 15.08% at 40 kV/cm). Our work provides an innovative way of designing new high-performance lead-free piezoelectric ceramics that are promising for high-precision piezoelectric actuator applications.

Results

Electrostrain and temperature stability

Figure S1 in Supplementary Information shows the microstructure of (KN)_xNT-BZ+MnO ceramics. All compositions exhibit closely compacted grains with a slight increase in size by K/Na deficiency, revealing high quality, which is also confirmed by the high density of over 95% in Table S1. Figure 2a presents the bipolar strain of all ceramics. The maximum strain (S_max) obviously increases with the decrease of K/Na content and reaches a maximum value of 0.41% at x = 0.98, while it decreases with further reduction of K/Na content. Interestingly, all samples exhibit relatively symmetrical bipolar strain curves, demonstrating small variation for S_max in the first and second quadrants. This good symmetry is further improved by reducing K/Na content, as quantitatively revealed by the decreasing β from 37.87% to 10.96% (Fig. S2). This good symmetry of the bipolar strain can also be reflected by the relatively symmetrical P-E loops without obvious translational shift, as shown in Fig. S3. Of particular importance is that the E_i decreases obviously from ~0.72 kV/cm to ~0.06 kV/cm with decreasing K/Na content (Fig. S4), which is much smaller in comparison with other systems that can generate giant asymmetrical bipolar strain^15,29–31.

Fig. 2. Electrostrain and its temperature stability of (KN)_xNT-BZ + MnO ceramics.

a Bipolar strain curves of (KN)_xNT-BZ + MnO ceramics at a fixed electric field of 40 kV/cm and 1 Hz; b Electric field dependence of unipolar strain curves at 1 Hz for x = 0.98; c Electric field dependence of $d_{33}^{*}$ for all ceramics; d A compared study of $d_{33}^{*}$ and applied electric field between the materials in this work and the state-of-the-art piezoelectric materials in the literatures^{8,12,32,52–61}; e Temperature dependence of unipolar strain curves for x = 0.98 at 8.4 kV/cm and 1 Hz; f $d_{33}^{*}$ and its stable temperature span ($T_{90}^{*}$) in various piezoelectric materials^32–39: Parameter $T_{90}^{*}$ where the $d_{33}^{*}$ deteriorates to 90% of its value at room temperature is used to judge the temperature stability⁶².

To get the electrostrain, unipolar strain curves of x = 0.98 ceramics are measured at different applied electric fields and 1 Hz, as depicted in Fig. 2b. The strain obviously increases with the enhancement of the applied electric field, reaching a maximum value up to 0.47% at 50 kV/cm. Furthermore, the maximum strain also exhibits strongly compositional dependent behavior, which increases first and then decreases with the decrease of K/Na content, for example, from 0.23% at x = 1, to 0.47% at x = 0.98 and 0.40% at x = 0.97 (Fig. S5). As for each independent component, the as-calculated $d_{33}^{*}$ increases first and then decreases with increasing applied electric field, reaching a maximum value at E_D, as shown in Fig. 2c. It should be noticed that the x = 0.98 ceramic demonstrates ultra-high $d_{33}^{*}$ over 2000 pm/V at a considerably low E_D of 8.4 kV/cm. In comparison with the state-of-the-art piezoelectric bulks, the (KN)_xNT-BZ+MnO ceramics showcase more outstanding electrostrain coefficient, as displayed in Fig. 2d. It is worth noting that the unipolar strain hysteresis of x = 0.98 ceramic is only 26.35% at E_D and 14.33% at 20 kV/cm (Tables S2 and S3 and Fig. S6), which is significantly reduced than those of commercially used PZT-5H ceramics (~23% at 20 kV)¹², demonstrating better reversibility.

The temperature stability is also a very important merit for piezoelectric actuators in practical applications. Figure 2e gives the temperature dependence of unipolar strain and the corresponding $d_{33}^{*}$ of x = 0.98 ceramic at its E_D of 8.4 kV/cm. The unipolar strain increases from ~0.169 % (~2000 pm/V) to a maximum value of ~0.226% (~2690 pm/V) at 100 °C, and then decreases to ~0.152% (~1809 pm/V) at 160 °C, demonstrating that the x = 0.98 ceramics possess a high temperature stability ($T_{90}^{*}=160^{\circ }\mathrm{C}$). Of particular importance is that, in comparison with the state-of-the-art piezoelectric bulks^32–39, the x = 0.98 ceramic not only demonstrates superior $d_{33}^{*}$, but also exhibits larger $T_{90}^{*}$ with better temperature stability. Besides, a good fatigue unipolar strain of the x = 0.98 ceramics, at which only degrades 7.9% after 10⁶ cycles under unipolar E_D = 8.4 kV/cm (Fig. S7). These distinguished merits make it great potential for piezoelectric micro-actuators applications.

Discussion

Microstructure

The design of A-site ions deficiency should cause an increase in oxygen vacancy according to Eq. (1). However, it is surprising that the concentration of oxygen vacancies decreases continuously with reducing K/Na content, e.g. from 24.66% for x = 1 to 15.67% for x = 0.97, as depicted in the refinement of O1s curves of X-ray photoelectron spectroscopy in Fig. S8 and the results presented in Fig. 3a. In addition, the refinement of O1s curves of X-ray photoelectron spectroscopy in ceramics sintered in air (Fig. S8) shows the concentration of oxygen vacancies increases continuously from 22.31 to 31.55%. According to Figs. S8 and S9, this may be ascribed to the Mn²⁺ ions partially substituting for A-site ions (K⁺ and Na⁺), as hinted in Eqs. (2) and (3). To determine the valence state of the Mn element, Mn 2p curves of X-ray photoelectron spectroscopy for (KN)_xNT-BZ+MnO ceramics are measured and refined, as shown in Fig. S10. The majority of Mn²⁺ (75–90%) with minor Mn⁴⁺ (10–25%) is determined in all samples sintered in reducing atmosphere, as shown in Fig. 3b. Interestingly, the amount of Mn²⁺ ions increases from 76.68% for x = 0 to 89.94% for x = 0.97 with more serious K/Na deficiency. However, the amount of Mn⁴⁺ ions in ceramics sintered in air increases from 27.14 to 36.87% in Fig. S11, which means more Mn⁴⁺ ions enter the B-site and then increase oxygen vacancy concentration. It indicates that the reducing atmosphere hinders the oxidation of Mn²⁺ ions that prefer to locate at the B-site, which, on the contrary, promotes more Mn²⁺ ions to occupy the A-site space left by the K/Na deficiency. A higher K/Na content produces more A-site vacancies (${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }$), which also leads to more Mn²⁺ ions locating at ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }$ and the formation of ${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\cdot }$ defects to keep valence balance. This is also the reason for the lower concentration of oxygen vacancy for the compositions with a high level of K/Na deficiency.

$$2{\mathrm{Na}}_{\mathrm{Na}}+{2\mathrm{K}}_{\mathrm{K}}+2{\mathrm{O}}_{\mathrm{o}}^{*}\to {\mathrm{K}}_2\mathrm{O}+{\mathrm{Na}}_2\mathrm{O}+2{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }+{2\mathrm{V}}_{\mathrm{Na}}^{\prime }+2{\mathrm{V}}_{\mathrm{K}}^{\prime }$$ 1

$$2\mathrm{MnO}{-\to }^{{\mathrm{K}}_2\mathrm{O}}{2\mathrm{Mn}}_{\mathrm{K}}^{\cdot }+{\mathrm{O}}_2+2{\mathrm{e}}^{\prime }+{\mathrm{O}}_{\mathrm{o}}$$ 2

$$2\mathrm{MnO}{-\to }^{{\mathrm{Na}}_2\mathrm{O}}{2\mathrm{Mn}}_{\mathrm{Na}}^{\cdot }+{\mathrm{O}}_2+2{\mathrm{e}}^{\prime }+{\mathrm{O}}_{\mathrm{o}}$$ 3

Fig. 3. Microstructure of (KN)_xNT-BZ + MnO ceramics.

Relative amount of a lattice oxygen and oxygen vacancies, b Mn²⁺ and Mn⁴⁺ ions, and c rhombohedral (R) phase and orthorhombic (O) phase in (KN)_xNT-BZ+MnO ceramics. The domain structure in d x = 1 and e x = 0.98 ceramics. The statistic polar vector angles image in f x = 1 and g x = 0.98 ceramics, the blue rectangle highlights the atomic-scale PNRs. h A statistical percentage of polar vector angles (φ) that are off the [111] direction. Strain map on an experimental HAADF STEM image of i x = 1 and j x = 0.98 ceramics: ε_xx refers to lattice strain.

The change of various defects when increasing the K/Na deficiency certainly will lead to the increase of ${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }-V_{\mathrm{Na}/\mathrm{K}}^{\prime }$ defect dipoles and decrease of ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ defect dipoles in (KN)_xNT-BZ+MnO ceramics. Theoretically, the highly stable ${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }-{\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }$ defect dipoles make a very low contribution to E_i due to their high migration activation energy⁴⁰. On the contrary, the ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ defect dipoles are more easily reoriented along the spontaneous polarization direction to form P_d and consequently lead to E_i, as a result of oxygen vacancy motion^15,41. In the newly designed compositions, the Mn²⁺ ions will lead to the decreasing concentration of oxygen vacancy to keep charge balance, which remarkably hinders the formation of highly oriented ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ defect dipoles, thus resulting in smaller E_i. In this case, the Landau free energy plot is more symmetrical due to the lack of contribution from ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ defect dipoles, which improves the symmetry of electrostrain.

Theoretically, a system with a single phase typically has a higher energy barrier for polarization switching. Since the reversibility of the rhombohedral (R)-monoclinic (M) phase transition in R-orthorhombic (O) phase coexistence tends to optimize the path of polarization rotation and result in nano-domains, which leads to the Landau energy profile more shallower and flatter in comparison with other phases^19,42,43. As a result, an R-O phase boundary in KNN-based ceramics is designed via K/Na ions deficiency, which may be easier to induce nano-domains and then flatten the Landau free energy profile. The XRD patterns and the corresponding Rietveld-refinement plots of (KN)_xNT-BZ+MnO ceramics are shown in Fig. S12. After Rietveld refinement, it is found that all ceramics can be well-fitted by two mixed phases, namely, a dominant R phase (space group R3m) and a small amount of O phase (space group Amm2), as shown in Fig. 3c. Interestingly, the fraction of R phase decreases from 95.25 to 84.77% with decreasing K/Na content, whereas the mount of O phase increases from 4.75 to 15.23%. This indicates that an R-rich multiple-phase boundary is successfully constructed through A-site deficiency engineering.

Figure 3d, e presents the TEM images of (KN)_xNT-BZ+MnO ceramics with representative compositions of x = 1 and 0.98. Typical stripe-like domains are clearly observed in both compositions, but with a remarkable decrease in width from 160 nm to 34 nm. The formation of nano-domains is mainly associated with the following two reasons. (1) The instability of multiphase with optimal phase ratios decreases polarization anisotropy so which reduces the domain size^19,44. (2) The A-site point defects cause a stronger random field and then result in a higher level of LSH to make additionally interfacial energies (such as electrostatic and elastic energies associated with the heterogeneous interfaces), which results in nano-domains. These nano-domains will be more susceptible to reorientation when subjected to an electric field, which plays a crucial role in achieving a high $d_{33}^{*}$ under low E_D^45,46.

To further investigate the intrinsic mechanism of domain response, aberration-corrected STEM is employed to perform atomic mapping of domains (Red box) (Fig. 3d, e) along the [001]_pc zone axis. The polar vectors of the ceramics are quantitatively calculated based on the relative displacements of the Nb⁵⁺ cations concerning the center of the nearest A-site cations. Figure 3f, g respectively shows the statistical polar vector angles image in x = 1 and x = 0.98 ceramics. When x = 1, the vast majority of polar vectors share a common polarization direction along the [111]_pc axis, signifying the predominant existence of the R phase. In contrast, polar vectors are observed in all directions for x = 0.98 ceramic, indicating a significant anisotropy of the ferroelectric polarization. Meanwhile, polarity vectors with opposite directions in local structure and atomic scale PNRs (Blue box) are also observed in x = 0.98 ceramic. This unique microstructure (Fig. S13) can also be found in other locations of the TEM (x = 0.98) sample. This means the A-site defect engineering indeed induces a higher level of LSH and results in the formation of many atomic-scale PNRs embedding in the nano-domains, which is verified by the enhancement of dielectric dispersion (Figs. S4 and S15).

Figure 3h presents a systematic statistical analysis of the polar vector angles and their proportions in x = 1 and x = 0.98 ceramics. It is worth noting that the x = 1 ceramic exhibits a very concentrated polar vector distribution with the majority of polar vectors (~54.25%) locating at low polar vector angles (0–20°), while the percentage of high-angle polar vectors (>80°) is <8 %, in consistence with the high percentage of R phase. On the contrary, the x = 0.98 ceramic displays a highly diffused polar vector with a wide angle (φ) distributed from 0° to 180°. We consider that the polarization vectors of a large number of atomic-scale PNRs deviate from their original <111> directions of R phase and randomly distribute in all directions, because a higher level of LSH results in the enhancement of the random field to disrupt the long-range order of the submicron domains^18,47. The diffused polarization vectors distribution at wide angles from 0° to 180° in x = 0.98 ceramic illustrates polarization switching path is more diverse, which promotes energy barrier of polarization switching to decrease. Thus, both intrinsic and extrinsic contributions to piezoelectricity can be enhanced due to easier polarization rotation and domain switching under an electric field^48,49.

The LSH can also be revealed by the fluctuation of lattice strain (ε_xx). Figure 3i, j presents the strain map on experimental HAADF STEM images for x = 1 and 0.98 ceramics. The ε_xx of x = 1 maintains a small variation in the range of ±0.005, whereas the ε_xx fluctuation is much larger with a value up to ±0.06 for x = 0.98, demonstrating more noticeable strain heterogeneity among the B-site cations after A-site engineering. Furthermore, more nanoscale regions with larger ε_xx fluctuation are also observed in x = 0.98 ceramic, as shown in Fig. 3j. There should be a significant lattice mismatch between these nanoscale regions due to the introduction of various defects, resulting in larger lattice strain in x = 0.98 ceramic. This is also strong evidence of the high level of LSH.

Figure 4a, b displays vertical piezoresponse force microscopy (PFM) images of x = 1 and x = 0.98 ceramics. Before measurement, a very high positive DC voltage (30 V) was first applied to the tip during scanning of a 2 × 2 μm² area, to enable complete polarization reorientation. Subsequently, a negative DC voltage with various magnitudes was applied to the tip during scanning of a 0.5 × 0.5 μm² area to investigate the reversibility of the poled domains. A high voltage as high as 6 V is required to completely induce a reversal of domains in x = 1 ceramic, where the square exhibits inverse phase contrast compared to the surrounding part. Interestingly, the domains begin to switch at a very small voltage of 1 V and complete under 2 V in x = 0.98 ceramic, which is much lower than that observed in x = 1 ceramic. This indicates a considerably stronger domain switching response for the x = 0.98 ceramic, due to its lower energy barrier attributed to the unique nano-domains embedded by smaller PNRs⁵⁰. These atomic-scale PNRs induced by A-site defects (${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }$ and ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }$) are embedded in nano-domains and then result in more easily domain switching as accelerators^24,51. According to the Rayleigh analysis (in Fig. S17), the ceramic (x = 0.98) has significantly higher values of piezoelectric Rayleigh coefficient (α = 11 × 10⁻¹⁶ m²/V²) than that α (3.2 × 10⁻¹⁶ m²/V²) of ceramic (x = 1), also confirming that extrinsic contribution from these unique nano-domains response under small external electric field is enhanced significantly via the design of A-site deficiency. These are also the reasons that accounts for the much lower driving electric field and the highest electrostrain coefficient (d₃₃) (shown in Fig. S17) in x = 0.98 ceramic.

Fig. 4. Phase-field simulation results and domain response of (KN)_xNT-BZ + MnO ceramics.

PFM images of a x = 1 and b x = 0.98 ceramics. Three-dimensional Landau free-energy profiles of c x = 1 and d x = 0.98 ceramics. Two-dimensional Landau free energy profile at different phase transition path of e x = 1 (with bias field 0.7 kV/cm) and f x = 0.98 (with bias field 0.15 kV/cm) ceramics.

To uncover the origin of the large electrostrain in (KN)_xNT–BZ + MnO ceramics at low electric fields, we combined Landau thermodynamics with phase-field simulations. An eighth-order Landau potential was fitted to dielectric-temperature spectra, yielding 3D free-energy surfaces at 300 K for x = 1 and x = 0.98 (Fig. 4c, d). Both compositions exhibit an R-phase ground state (O and T tetragonal phases metastable), yet the R → O and R → T barriers are lower for x = 0.98 (Fig. 4e, f). The asymmetric profile at x = 1 stems from a defect-dipole induced bias field (0.7 kV cm⁻¹), whereas the nearly symmetric curve at x = 0.98 reflects the small experimental bias field (0.15 kV cm⁻¹). Phase-field simulations based on the fitted potential reveal that increased Mn doping enhances local random electric fields, transforming the domain structure from monolithic R domains (x = 1) to nanoscale multiphase coexistence (x = 0.98) with markedly reduced domain size (Fig. S18). The bipolar strain loop calculated from the phase-field simulation. Loops show a pronounced electrostrain of 0.35 % for x = 0.98. With unchanged electrostrictive coefficients, the enhanced $d_{33}^{*}$ correlates directly with (1) flattened free-energy barriers and (2) the multiphase nanodomain state. The reduced barriers enable R → T transitions at lower fields, and the larger c/a ratio of the T phase yields greater strain. Moreover, the multiphase configuration coupled with local random fields promotes reversible domain reconfiguration upon field removal. Thermodynamic and phase-field analyses thus demonstrate that flattening of the energy landscape and multiphase nanodomains collectively drive the high electrostrain in x = 0.98 under modest electric fields.

In summary, ultra-high electrostrain coefficient ($d_{33}^{*}~2000\mathrm{pm}/\mathrm{V}$) at a low driving electric field of 8.4 kV/cm and high symmetry bipolar strain curves (15.08% at 40 kV/cm) were achieved in (KN)_0.98NT-BZ+MnO lead-free ceramics by synergy effect from A-site defect engineering and charge compensation. Furthermore, the (KN)_0.98NT-BZ+MnO ceramics also possessed good temperature stability $\left(T_{90}^{*}~160^{\circ }C\right)$ and relatively low unipolar strain hysteresis (26.35%) under 8.4 kV/cm. The partial Mn²⁺ ions as donor dopants in A-site of KNN perovskite form point defects (${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }$) and reduce ${\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ concentration, which weakens the contribution from ${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }-{\mathrm{V}}_{\mathrm{o}}^{\bullet \bullet }$ and then improves bipolar strain symmetry. The coexistence of R/O multiple phases induces nano-domains that help to flatten the Landau energy profile to a certain degree in x = 0.98 ceramic. The atomic-scale PNRs induced by point defects (${\mathrm{V}}_{\mathrm{Na}/\mathrm{K}}^{\prime }$, ${\mathrm{Mn}}_{\mathrm{Na}/\mathrm{K}}^{\bullet }$) embedding in nano-domains further flatten the Landau energy profile by increasing interfacial energies. The flattened and symmetrized Landau free energy plot and multiphase nanodomains results in ultra-high electrostrain, low E_D and highly symmetrical bipolar strain in (KN)_0.98NT-BZ+MnO ceramics.

Outlook

This work successfully addressed the long-standing challenge in the development of lead-free piezoelectric ceramics by achieving high electrostrain under low driving electric field and high bipolar strain symmetry. In this work, we proposed a strategy in potassium sodium niobate (KNN) based ceramics for designing a unique microstructure of atomic-scale polar nanoregions (PNRs) embedded in nano-domains by tuning the A-site ions deficiency and coexistence of multiple phases. The designed KNN-based ceramics exhibited highly symmetrical bipolar strain and ultrahigh electrostrain coefficient ($d_{33}^{*}\approx 2000\mathrm{pm}/\mathrm{V}$) under a low E_D (8.4 kV/cm). The concept presented here provides a valuable way of developing other lead-free piezoelectric ceramics with good symmetrical bipolar strain, and ultra-high electrostrain coefficient under low E_D. However, it should be pointed out that the hysteresis of electrostrain curves at low E_D for KNN-based ceramics still needs to be further improved. More work should be done to achieve low-hysteresis electrostrain curves, which are of great significance for the practical use of lead-free piezoelectric ceramics.

Methods

Sample preparation

The KNN-based ceramics with composition of 0.94(K_0.48Na_0.52)_xNb_0.96Ta_0.04O₃-0.06BaZrO₃ + 0.08MnO (abbreviated as (KN)_xNT-BZ+8MnO, x is 1, 0.99, 0.98, and 0.97) were prepared through a conventional solid-state method in a reducing atmosphere. 3 mol% ZrO₂ was added to each sample to improve the anti-reduction property. Analytically pure raw materials, including BaCO₃, Nb₂O₅, Na₂CO₃, K₂CO₃, ZrO₂, MnO, and Ta₂O₅, were weighed according to the stoichiometric formula and then mixed via planetary ball-milling with zirconia balls in ethanol for 24 h. The ball-milled flurries were dried at 70 °C through rotary evaporation and subsequently were calcined at 850 °C for 2 h in air. The calcined powders were ball-milled again and mixed with a PVB binder before being compacted into pellets under uniaxial pressure in a stainless-steel die. The green pellets were 8 mm in diameter and 0.4 mm in thickness. The green pellets were first heated to 600 °C for 2 h in air to remove the binder, followed by sintering at ~1120 °C for 2 h in a reducing atmosphere consisting of H₂ and N₂ (H₂/N₂ ~ 0.6%) to hinder the oxidation of Mn²⁺ ions during high-temperature sintering. The as-sintered samples were then cooled to ~850 °C and held for 9 ~ 10 h in high-purity nitrogen before being naturally cooled. For comparison, part of the green pellets were sintered at ~1120 °C for 2 h in air. All sintered ceramics were ~6.8 mm in diameter and ~0.4 mm in thickness.

Structure characterization

Phase purity and crystal structure of the as-prepared samples were characterized by using an X-ray powder diffractometer (XRD, SmartLab-3kW, Rigaku Ltd., Tokyo, Japan) with Cu Kα radiation. The microstructure was observed by using a scanning electron microscope (SEM, FE-SEM Sigma 300, ZEISS Corp., Germany). Gold electrodes were sputtered on the samples using ion sputtering equipment (SBC-12, KYKY Technology Corp., Beijing, China). The amount and valence of Mn and O ions were obtained using an X-ray photoelectron microprobe (ESCALAB 250Xi, Thermo Fisher, Britain) equipped with a standard monochromatic AlKa excitation source (hυ = 1361 eV). The piezoelectric response was characterized using an atomic force microscope (AFM, MFP-3D, Asylum Research, USA) with a functionality of a piezoresponse force microscope (PFM) for the local poling experiments and switching spectroscopy. A negative DC voltage large enough (30 V) to induce complete polarization orientation was first applied to the tip during the scanning of a 2 × 2 μm² area. Then, a positive DC voltage with different values, depending on the composition, was applied to the tip during the scanning of a 0.5 × 0.5 μm² area. Microstructure observation by transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM). Samples for electron microscopy were prepared by an FEI Helios NanoLab focused ion beam (FIB). Domain morphology was observed using a TEM (Talos F200X G2, FEI, America) operated at 200 kV equipped with a charge-coupled device camera. Aberration-corrected scanning transmission electron microscope was performed on a probe corrected FEI Spectra 300/TEM (Thermo Fisher Scientific, Eindhoven, Netherlands) equipped with a Schottky field-emission gun and operated at an accelerating voltage of 300 kV. The beam current was 40 pA and had a convergence semi-angle of 17.8 mrad. The STEM images were collected by a high-angle annular dark-field detector with inner- and outer-collection semi-angles of 73 and 200 mrad, respectively. The image analysis and quantification were conducted using custom MATLAB scripts. The polar vectors were calculated by the projected B-site (of high contrast) cation displacement from the center of the nearest four A-site (of low contrast) cations. The position of cations was specified by 2D Gaussian fitting on HAADF-STEM images using the Atomap module. The polar vector map was visualized by the Temul toolkits. For each atomic column (A or B site) found in the STEM images, we recorded its coordinates, intensity, and polar vector for further analyses, including drawing the intensity map, polar vector map, and polar vector angle map.

Measurement of piezoelectric properties

After confirming the sample condition, poling of the samples was conducted in a silicone oil bath under T = 80 °C and a DC electric field of 30 kV/cm for 30 min. The poled sample is left at room temperature for one week for aging. The bipolar/unipolar electric-field-induced strain curves including temperature dependence of the strain curves under 1 Hz in the temperature range from 25 to 220 °C and polarization-electric field (P-E) loops were measured under 1 Hz using a TF ANALYZER 3000 ferroelectric measuring system (AixACCT Systems GmbH, Aachen, Germany). The unipolar fatigue cycles at room temperature were carried out using 50 Hz, E_D unipolar triangular waves, and then tested at 1 Hz and E_D.

Phase-field simulation

The simulation details are provided in the Support materials.

Author contributions

Z.C. and N.L. conceived the original idea and directed the project. Z.C. wrote the manuscript. Z.C. and Z.X. revised the manuscript. F.C. prepared the sample and processed the experimental data. Z.C. and F.C. performed and analyzed the aberration-corrected STEM experiments and X-ray photoelectron spectroscopy experiments. Z.D. performed SEM experiments. Z.X., X.C., and D.Z. carried out the PFM measurements. X.S., H.H., and Z.L. performed the phase-field simulations. F.C., Y.-X. L., and X. Z. conducted the TEM observations. All authors discussed and edited the paper. K.W. and C.W. provided helpful advice and critical discussions. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Maxim Morozov (eRef) who co-reviewed with Mariia Mikhailova (ECR); and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information files and from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-65521-5.