Synergistically enhanced energy storage performance of Bi_0.47Na_0.47Ba_0.06TiO₃-based relaxor ferroelectrics via dual engineering of dynamic nanodomains and defect regulation

Jiangping Huang, Liang Deng*, Yu Zhang, Yue Pan, Xiuli Chen*, Xu Li* and Huanfu Zhou
Key Laboratory of New Processing Technology for Nonferrous Metal and Materials, Ministry of Education, Guangxi Key Laboratory of Optical and Electronic Materials and Devices, College of Materials Science and Engineering, Guilin University of Technology, Guilin, 541004, China. E-mail: 809188659@qq.com; cxlnwpu@163.com; lx100527@163.com

First published on 19th August 2025

---

---

1. Introduction

Over the past few decades, with the massive consumption of fossil fuels and other non-renewable resources, as well as the ever-increasing environmental awareness, the development and use of renewable and environmentally friendly resources has become an urgent priority.^1–3 Ceramic capacitors, known for their high-power density, fast charging and discharging capabilities, and excellent high-temperature stability, have attracted widespread attention and research in the field of energy storage.⁴ The availability of high-performance energy storage materials is key to the practical application of pulse power devices. Lead-based materials, known for their superior piezoelectric and dielectric properties, have been the most extensively studied ferroelectric materials.^5,6 However, concerns about the environmental degradation and human health risks associated with lead-based materials, coupled with recent environmental legislation, have spurred significant research efforts to develop lead-free alternatives to replace lead-based materials. To meet the urgent demand for high-performance lead-free energy storage ceramics, which are essential components in advanced pulse power devices and high-power equipment, there is an urgent need to develop new types of energy storage ceramic capacitors with high energy storage density and high efficiency.^7,8 Typically, the energy storage performance of ceramic capacitors can be described by the following formula:^9–12

	(1)

	(2)

	(3)

where P_max is the maximum polarization, P_r is the remanent polarization, W_total represents the total energy storage density, W_rec denotes the recoverable energy storage density, η signifies the energy storage efficiency, and E denotes the applied electric field. Therefore, achieving a high polarization difference (ΔP) and a high breakdown electric field (E_b) are key factors in enhancing the energy storage performance.

Relaxor ferroelectrics, characterized by their slim hysteresis loops and high P_max, have garnered significant attention in the field of energy storage materials.¹³ Bismuth sodium titanate (Bi_0.5Na_0.5TiO₃, BNT), a lead-free ferroelectric ceramic with an ABO₃ perovskite structure, exhibits superior ferroelectric properties and an easily manipulated phase structure.¹⁴ In particular, compared to other lead-free ceramics such as BaTiO₃ (BT), AgNbO₃ (AN), (K_0.5Na_0.5)NbO₃ (KNN), NaNbO₃ (NN), and SrTiO₃ (ST), BNT with its large high polarization capacity positions it as one of the most promising lead-free candidate materials.^15,16 However, pure BNT ceramic, because of its significant residual polarization and low breakdown strength, hinders its direct application in energy storage.^17,18 Research indicates that BNT-based relaxor ferroelectric ceramics, modified by doping, exhibit high P_max, low P_r, and high breakdown strength, making them suitable for advanced pulsed power applications. As a result, many optimization strategies have been proposed to improve the energy storage performance. For example, Long et al. conducted doping modifications on BNT, significantly reducing energy losses and enhancing polarization response through optimization of the grain size and domain structure, leading to a marked improvement in energy storage density. Ultimately, a high energy storage density of 8.63 J cm⁻³ was achieved, along with a high efficiency of 89.6%.¹⁹ Furthermore, the solid solution of Bi_0.5Na_0.5TiO₃ with BaTiO₃, which establishes a morphotropic phase boundary (MPB) where rhombohedral (space group R3c) and tetragonal (space group P4bm) phases coexist, can significantly enhance the polarization response and piezoelectric properties.^20–22 Due to the possibility of extensive compositional tuning and rare earth doping modification, various superior BNT-BT-based ceramics can be obtained.^23–25 Consequently, the (1 − x)BNT-xBT system has long been considered one of the important materials for the modification of energy storage ceramics. For instance, Cao et al. achieved an ultra-high energy storage density of 12.2 J cm⁻³ and a high η of 88.8% by incorporating Ca_0.7Bi_0.2(Sn_0.5Ti_0.5)O₃ into the Bi_0.5Na_0.47Ba_0.06O₃ (BNBT) ceramic and optimizing the E_b through an interfacial polarization strategy.²⁶ Research shows that doping modifications can regulate domain sizes, forming nanodomains and even refined polar nanoregions, which play a crucial role in improving the energy storage efficiency and achieving stability over a wide temperature range.^19,27,28 For example, Dong et al. modulated the domain configuration by introducing Sr(Ni_1/3Nb_2/3)O₃ as the second component in BNBT, ultimately achieving a W_rec of 7.0 J cm⁻³ and a η of 81.5%.²⁹ Breakdown electric strength is a key factor in achieving high energy storage performance. In general, grain refinement, reduction of dielectric losses, reduction of volatilization of low melting point elements and optimization of the preparation process can significantly enhance the breakdown strength of ceramics.¹⁶ For example, Liu et al. obtained a high W_rec of 12.2 J cm⁻³ at 950 kV cm⁻¹ through a grain boundary optimization strategy.³⁰ Deng obtained an ultra-high W_rec of 12.39 J cm⁻³ and a η of 87.1% at 710 kV cm⁻¹ via grain engineering in BNT-based relaxor ferroelectrics. It is shown that doping rare earth elements in BNT can enhance local heterogeneity, achieve domain structure modulation, and improve sintering, which in turn improves the energy storage performance.^31–33 For instance, Tang et al. introduced Sm₂O₃ into BNBT, which significantly reduced the grain size and enhanced the relaxor characteristics, resulting in a W_rec of 4.41 J cm⁻³ and a η of 83.96%.³⁴ Li et al. enhanced the energy storage performance of Bi_0.5Na_0.5TiO₃ ceramic through SmFeO₃ doping, achieving an ultrahigh W_rec of 9.05 J cm⁻³ by improving relaxor behavior and delaying polarization saturation.³³ Although many studies have achieved breakthroughs, realizing ultra-high performance (W_rec >10 J cm⁻³) remains rare, and there is an urgent need to develop new high-performance lead-free energy storage materials. Furthermore, while many researchers have explored the evolution of domain structures in relaxor ferroelectric materials, a deeper investigation of the intrinsic relationship between micro-domain structures and macroscopic energy storage performance is still required.

To solve the problem of low energy storage density, we have made a breakthrough in achieving high energy storage performance in a novel (1 − x)(Bi_0.5Na_0.47Ba_0.06O₃)-xSm(Mg_2/3Sb_1/3)O₃ (abbreviated as BNBT-xSMS) lead-free relaxor ferroelectric ceramics via the regulation of highly dynamic polar nanoregions and defects. Firstly, we selected Bi_0.5Na_0.47Ba_0.06O₃-based ceramics with high polarization as the matrix material for our study.^35,36 The coexistence of the R3c (rhombohedral) and P4bm (tetragonal) phases at the morphotropic phase boundary (MPB) provides a foundation for constructing the polymorphic relaxor phase.^37,38 However, the high P_r and low breakdown strength remain the primary limiting factors for improving the energy storage density in BNBT ceramics, necessitating further modification. Secondly, we introduced the rare earth-based complex Sm(Mg_2/3Sb_1/3)O₃ into BNBT to form a solid solution and optimize the energy storage performance. On the one hand, Sm₂O₃ can optimize sintering, reduce the grain size and decrease the defect concentration, which improve the breakdown electric strength.^33,39–41 On the other hand, the introduction of the B-site trivalent complex cation (Mg_2/3Sb_1/3)³⁺ effectively disrupts the long-range ferroelectric order of the matrix and induces more weakly polar P4bm phases, thereby constructing polar nanoregions (PNRs), which improve the domain response to the external electric field and consequently enhance the energy storage efficiency.^42–44 Additionally, the high bandgap oxides of MgO (7.03 eV) and Sm₂O₃ (4.4 eV) further enhance the breakdown resistance of the ceramics.³³ We have demonstrated through theoretical modeling and experimental characterization that the construction of highly dynamic polar nanodomains in BNBT-based ceramics provides the conditions for achieving high energy storage efficiency through rapid reversible switching under an external electric field. As expected, we achieve an ultrahigh W_rec of 13.6 J cm⁻³ with a high η of 84.9% in the BNBT-0.16SMS ceramic. In addition, the ceramic exhibits excellent temperature/frequency stability and superior charge–discharge performance. These findings provide new insights for the development of ceramics with high energy storage performance.

2. Experimental procedures

The relaxor ferroelectric ceramics (1 − x)Bi_0.47Na_0.47Ba_0.06TiO₃-xSm(Mg_2/3Sb_1/3)O₃ (abbreviated as BNBT-xSMS with x = 0.08, 0.12, 0.16, 0.20) were prepared using traditional solid-state sintering reaction technology. The preparation methods and characterization studies of BNBT-xSMS bulk ceramics are available in the SI.

3. Results and discussion

All ceramic compositions were fabricated via conventional solid-state sintering. X-ray diffraction (XRD) analysis was conducted to characterize the crystal structure (Fig. 1a). As the doping content increased from 0.08 to 0.16, the ceramics exhibited pure perovskite structures. At x = 0.20, secondary phase peaks were detected, indicating that excessive SMS doping exceeded the solid solubility limit of the BNBT-based ceramics, leading to secondary phase formation. To further investigate the phase content of ceramics, Rietveld structural refinement was performed on representative compositions (x = 0.08 and x = 0.16) using the R3c (R) and P4bm (T) phases.²⁹ As shown in Fig. 1b and c, the experimental data exhibited excellent agreement with the refined patterns, with low reliability factors (including the weighted profile residual factor (R_wp), pattern residual factor (R_p), and chi-square value (χ²) below 5%), confirming the robustness of the refinement results. At x = 0.08, the ceramic predominantly exhibited the R phase, while x = 0.16 showed a dominant T phase. This phase transition, driven by composition tuning, disrupts long-range ordered structures and induces a disordered local field, promoting the formation of polar nanoregions (PNRs).⁴⁵ The resulting weakly polar T phase combined with a minor polar R phase not only maintains high polarization strength but also improves polarization response, thereby improving the energy storage efficiency through reduced hysteresis losses. To investigate the local structural evolution in BNBT-xSMS ceramics, room-temperature Raman spectroscopy was performed (Fig. 1d). The raw spectra exhibited four primary vibrational modes: A-site cation vibrations (100–200 cm⁻¹), B–O stretching vibrations (180–420 cm⁻¹), BO₆ octahedral bending/tilting modes (420–670 cm⁻¹), and A₁ + E (longitudinal optical) symmetric modes (670–850 cm⁻¹).^46–49 After Gaussian–Lorentzian deconvolution (Fig. 1e and f), seven resolved peaks emerged. Progressive intensity attenuation and red shifts of ν₄ and ν₅ modes were observed, indicating B-site displacement induced by SMS doping. This structural disorder, arising from lattice distortion and phase fraction variation, facilitates the formation of nanodomains, thereby enhancing the energy storage performance.⁵⁰ To elucidate the local structural evolution of BNBT-0.16SMS ceramic, transmission electron microscopy (TEM) was employed to analyze domain morphology, high-resolution lattice imaging, and selected area electron diffraction (SAED) patterns (Fig. 1). TEM observations revealed the presence of Moiré fringe-like nano-domain size (∼5 nm in size) with high dynamic activity,^51,52 which facilitates rapid domain switching under electric fields, thereby minimizing energy loss (Fig. 1g). High-resolution TEM (HRTEM) imaging (Fig. 1h) revealed distinct crystalline fringes, confirming the excellent crystallinity of the sample. SAED patterns exhibited two distinct superlattice reflections: 1/2(ooe) along the (001) zone axis and 1/2(ooo) along the (110) zone axis, as shown in Fig. 1i and j, where “o” and “e” denote odd and even Miller indices, respectively.⁵³ The appearance of these superlattice diffraction spots can be attributed to oxygen octahedral tilting and B-site cation displacement.⁵⁴ Therefore, these superlattice spots provide direct evidence for the coexistence of the R-phase (trigonal symmetry) and T-phase (tetragonal symmetry), which further confirms the accuracy of the XRD refinement results.⁵⁴ This local structural heterogeneity contributes to enhanced energy storage performance, as demonstrated by the increased recoverable energy density accompanied by reduced hysteresis loss.

The energy storage properties of BNBT-xSMS ceramics were characterized through unipolar P–E hysteresis measurements at room temperature at 10 Hz and 350 kV cm⁻¹ (Fig. S1a). With increasing SMS content, the loops progressively became slimmer, accompanied by reductions in both P_max and P_r (Fig. S1b), signifying that SMS doping effectively disrupts long-range ferroelectric ordering while promoting the formation of PNRs, thereby enhancing relaxor characteristics.⁵⁵ Although both W_total and W_rec decreased with increasing SMS content, a significant enhancement in η was observed. This apparent performance limitation likely results from the insufficient testing electric fields, as higher electric fields are expected to yield better results. The unipolar P–E hysteresis loops of BNBT-xSMS ceramics tested at E_b is shown in Fig. 2a. Notably, higher E_b and P_max values were achieved with increasing SMS content, clearly demonstrating composition-dependent optimization of energy storage performance. The composition with x = 0.12 exhibited superior P_max compared to that with x = 0.08, owing to its enhanced E_b and improved polarization response. By contrast, the x = 0.20 sample showed degraded performance attributable to reduced P_max and E_b, which originated from impurity phases. According to the calculated energy storage parameters, the W_rec values of BNBT-xSMS ceramics were 7.2, 11.4, 13.6, and 7.6 J cm⁻³ along with corresponding η values of 78.3%, 82.0%, 83.9%, and 86.8%, respectively, for increasing x. The primary factor governing these improvements was the enhancement in E_b. The reliability of E_b was evaluated using Weibull distribution analysis based on the following equation:⁵⁶

Xi = ln(Ei)	(4)

	(5)

where n represents the sample count and E_i denotes individual breakdown fields. The calculated shape parameters (β) were all >20, confirming outstanding E_b stability. This reliability stems from high specimen density and defect reduction, as discussed later in the E_b analysis section. As shown in Fig. 2d, the unipolar P–E loops for BNBT-0.16SMS were measured under different electric fields. When the electric field was increased from 100 kV cm⁻¹ to 760 kV cm⁻¹, P_max values increased significantly from 6.2 μC cm⁻² to 46.1 μC cm⁻², while P_r remained nearly unchanged. Consequently, W_rec improved from 0.3 J cm⁻³ to 13.6 J cm⁻³ with consistently high η >83% (Fig. 2e), demonstrating excellent electric field stability. Fig. 2f provides a performance comparison between BNBT-0.16SMS and other state-of-the-art lead-free systems, revealing that it achieves competitive performance and showcasing significant application potential for high-performance energy storage capacitors.

In contrast to conventional ferroelectric BNBT ceramics, the modified relaxor ferroelectric BNBT-xSMS system demonstrates significantly improved dielectric temperature and frequency stability. Fig. 3a–c present the temperature- and frequency-dependent relative permittivity (ε_r) and dielectric loss (tanδ) of BNBT-xSMS ceramics. With increasing x, the ceramics exhibit enhanced frequency insensitivity accompanied by reduced tanδ (Fig. 3a), which can be ascribed to the fast response and reversibility of PNRs.⁵⁷ The gradual decrease in ε_r with SMS doping arises from weakened coupling between PNRs, confirming enhanced structural disorder. The temperature-dependent ε_r behavior is illustrated in Fig. 3b. Pure BNBT ceramics typically show two dielectric anomalies, which provide the basis for broad dielectric plateaus. The BNBT-0.08SMS composition displays characteristic dual peaks: a low-temperature shoulder peak (T_s ∼150 °C) associated with thermal evolution between R3c-phase PNRs and P4bm-phase PNRs, exhibiting pronounced frequency dispersion, and a high-temperature peak (T_m ∼280 °C) corresponding to the transition of R3c-phase PNRs into P4bm-phase PNRs.⁵⁸ With further SMS doping, the T_s peak shifts to lower temperatures, while the T_m peak diminishes, thereby forming a broad dielectric plateau between the two peaks. This behavior results from enhanced A-/B-site ionic disorder and the disruption of long-range ferroelectric domains. This diffuse phase transition characteristic significantly improves the temperature stability of ε_r. The observed strong frequency dispersion and diffuse phase transition behavior provide clear evidence for the improved relaxor properties in this system.⁵⁹ The enhancement of relaxor behavior can be attributed to the incorporation of heterogeneous ions into the lattice, which disrupts the original long-range ferroelectric ordering and facilitates lattice reorganization into nanodomain structures with smaller dimensions.⁶⁰ This structural modification manifests macroscopically as classic relaxor ferroelectric behavior in the dielectric temperature spectra. Fig. 3c displays the temperature-dependent tanδ of BNBT-xSMS ceramics at 100 kHz. Increasing SMS content progressively suppresses tanδ, thereby mitigating energy accumulation and reducing thermal breakdown risks. The decrease in tanδ with temperature (up to 300 °C) correlates with the increasing activity of PNRs. However, the subsequent tanδ increase at elevated temperatures (>300 °C) likely originates from high-temperature leakage conduction. The characteristic frequency dispersion of dielectric response, induced by PNRs, substantially enhances the relaxor properties of BNBT-xSMS ceramics. Quantitative evaluation of relaxor behavior was conducted using the modified Curie–Weiss law:⁶¹

	(6)

where C is the Curie constant and γ characterizes the degree of relaxor behavior (ranging from 1 to 2). Note that γ = 1 corresponds to normal ferroelectrics, while γ = 2 represents ideal relaxor ferroelectrics; therefore, the larger the γ is, the stronger the relaxation is.⁶² As shown in Fig. 3d, the fitted γ values for the studied compositions range from 1.84 to 1.98, confirming strong relaxor behavior. Additionally, the relaxor characteristics were further characterized by ΔT_relaxor (ΔT_relaxor = T_s (1 MHz) − T_s (5 kHz)).⁵⁵ Fig. 3e demonstrates that BNBT-xSMS ceramics exhibit ΔT_relaxor values of 44–55 °C, indicating enhanced relaxor characteristics due to cation disorder and the coexistence of multiphase PNRs structures. The broadening and flattening of dielectric peaks caused by these collective relaxor effects improve dielectric thermal stability. To quantify this stability, we examined the temperature coefficient of capacitance (TCC) with 150 °C as the reference temperature (ε_r/ε_150°C). As shown in Fig. 3f, the TCC variation remains remarkably stable within ±15% over a broad temperature range at 10 kHz, directly demonstrating that strong relaxation behavior effectively suppresses thermal fluctuations in ε_r.

In relaxor ferroelectrics, the thermal evolution of PNRs is accompanied by changes in their size and dynamics. Based on three characteristic temperatures—the freezing temperature (T_f), phase transition temperature (T_m), and burning temperature (T_B)—the dielectric temperature spectra can be divided into non-ergodic relaxation (NER), ergodic relaxation (ER), and paraelectric (PE) states.⁶³ In the paraelectric state (>T_B), PNRs exhibit smaller sizes. As the temperature decreases in the ER state (T_f–T_B), PNRs nucleate and grow, acquiring high dynamic activity that enables reversible polarization reversal under the electric field. Below T_f, the system transitions to the NER state, where PNRs freeze and evolve into long-range ordered ferroelectric domains.³⁵ By tuning the ER state through composition control, enhanced dynamic responses of PNRs can be achieved, improving energy storage characteristics. To quantify these transitions, we employed Vogel–Fulcher (V–F) fitting for T_f and Curie–Weiss fitting for T_B using the following formulas:^55,64

	(7)

	(8)

where f₀ is the attempt frequency, E_a is the activation energy for PNR dynamics, k_B is the Boltzmann constant, and T_C is the Curie temperature. A larger E_a indicates stronger energy barriers between dipole clusters, suppressing long-range dipole correlations and promoting slim hysteresis loops. Here, T_f and E_a were extracted from the dielectric frequency dispersion peak T_s using V–F fitting (Table S1). As shown in Fig. 3g, T_f decreases from 71.8 °C to −41.2 °C with increasing x, confirming the transition from NER to ER states under SMS doping. Concurrently, E_a increases from 0.096 eV (x = 0.08) to 0.264 eV (x = 0.20) (Fig. 3h), indicating smaller dipole cluster interactions that suppress long-range correlations and mitigate dielectric nonlinearities. Fig. S2a–d present T_B fitting results, showing a reduction from 350 °C (x = 0.08) to 267 °C (x = 0.20). The composition-temperature phase diagram highlights how SMS doping systematically shifts BNBT ceramics into the ER regime (Fig. 3i). By engineering the ER state near room temperature, PNRs achieve weakly coupled relaxation behavior, enhanced large P_max under a high electric field, high energy storage efficiency, and suppressed macroscopic nonlinear polarization responses, which are key factors for optimizing the energy storage performance. These findings establish that controlling PNR thermal evolution through chemical modification is a key strategy for developing high-performance energy storage dielectrics.

The energy storage characteristics of relaxor ferroelectric ceramics are fundamentally governed by the domain structure configuration and its dynamic response behavior. Hence, the domain reversible switching behavior of ceramics during voltage application and removal was characterized by piezoresponse force microscopy (PFM). The opposite voltage (±20 V) was applied to the tip of the scanned sample over a scanning area of 4 × 4 μm². The incorporation of SMS doping into the BNBT system induces local structural disorder in the long-range ferroelectric order of the BNBT matrix, promoting the formation of fine nanodomains or PNRs. Fig. 4a and b present the domain morphology phase images of pure BNBT and BNBT-0.16SMS ceramics, respectively, where distinct color orientations correspond to different domain configurations. Pure BNBT exhibits maze-like domain structures characterized by large-scale (sub-micron) irregular domain distributions, reflecting its conventional long-range ferroelectric ordering. In striking contrast, the SMS-modified counterpart demonstrates a complete transition to nanoscale domain structures (verified by the TEM domain structure), manifested as featureless PFM response patterns due to PFM accuracy limitation. The formation of such nanodomains is mainly due to the local structural disorder and enhanced random fields resulting from the increased ionic disorder by doping SMS with different ionic radii. Detailed domain switching and relaxation property analysis uncovers the fundamentally different polarization behaviors between these systems. The litho-PFM test was used to write domains on the ceramics by applying different voltages, and the applied voltages are shown schematically in Fig. 4c. Pure BNBT shows complete domain reversal under ±20 V bias with negligible relaxation after 10 minutes, characteristic of strong ferroelectric coupling that leads to high remanent polarization and low-field saturated polarization behavior and consequently poor energy storage capability (Fig. 4d₁–d₃). In contrast, the BNBT-0.16SMS composition exhibits entirely distinct switching dynamics. Its displays weak domain-flipping behavior at ±20 V due to the weakly coupled characteristics of PNRs, as shown in Fig. 4e₁. Higher voltages are required to induce the formation of large-size domains and achieve a high polarization response. After 5 and 10 minutes of relaxation, the domain-switching behavior fully recovers to its initial state (Fig. 4e₂ and e₃). This unique behavior stems from the weakly coupled, highly dynamic PNRs that enable dramatically reduced P_r.⁶⁵ Therefore, the engineered PNR configuration in BNBT-0.16SMS yields exceptional energy storage performance metrics: ultrafast polarization response, minimal dielectric loss, outstanding energy storage efficiency, and remarkable field stability.

Additionally, achieving excellent energy storage performance is inseparable from attaining a high E_b. Generally, the E_b value is influenced by multiple comprehensive factors, including grain size, density, electrical insulation properties, and defect states. The surface morphology of the specimens was characterized by scanning electron microscopy (SEM), as shown in Fig. 5a, b, S3a and b. With increasing SMS content, the average grain size (G_a) of BNBT-xSMS ceramics initially decreases and then increases, reaching a minimum value of 2.14 μm for BNBT-0.16SMS ceramic. The reduction in G_a is attributed to the increased lattice strain energy caused by Sm³⁺, Mg²⁺, and Sb⁵⁺ ions with different ionic radii, which hinder grain boundary migration.⁶⁶ However, excessive doping induces abnormal grain growth, likely associated with secondary phase formation. As shown in Fig. 5c, the correlation between various SMS contents and E_b, G_a, and bulk density (ρ) reveals that both the reduced G_a and increased ρ positively contribute to improving E_b. Complex impedance spectroscopy was employed to investigate the roles of grains and grain boundaries in charge transport mechanisms and to elucidate the intrinsic factors enhancing E_b. The impedance spectra of BNBT-xSMS ceramics with x = 0.08 and x = 0.16 measured at high temperatures (520–600 °C) are shown in Fig. 5d and e. Notably, the total resistance of the ceramics significantly increases with higher SMS doping, which favors improved E_b. For the BNBT-0.08SMS ceramic, two distinct response modes can be significantly observed, corresponding to the grain boundary and grain contributions, respectively. In contrast, the BNBT-0.16SMS ceramic exhibits a single response mode predominantly governed by grain boundaries. Further analysis of the impedance imaginary part (Z′′) versus electrical modulus (M′′) reveals that the disappearance of double peaks and peak coalescence confirm enhanced electrical homogeneity (Fig. 5f).²⁶ Therefore, grain and grain boundary responses must be differentiated during impedance fitting. The grain and grain boundary resistances derived from impedance analysis are interpreted as thermally activated processes. Based on the Arrhenius relationship:⁶⁷

	(9)

where R is the fitted equivalent resistance, k_B is the Boltzmann constant, E_a is the activation energy, and T is the temperature. The fitting results yield activation energies of 1.31 eV (grains) and 1.03 (grain boundaries) for x = 0.08, while x = 0.16 shows a unified E_a = 1.36, which indicates that the systems are all caused by the oxygen vacancy conducting mechanism (0.5–2 eV). Thus, optimal SMS doping not only refines the grain size and enhances electrical homogeneity but also elevates the thermal activation energy, thereby significantly improving E_b. The increase in E_a is inversely correlated with the reduction of oxygen vacancy (O_V) content. X-ray photoelectron spectroscopy (XPS) was employed to analyze the O_V concentration as a function of SMS doping levels. The O 1s XPS spectra of the ceramics and the fitting of three Gaussian–Lorentzian peak positions corresponding to lattice oxygen (O_L), absorbed oxygen (O_A) and oxygen vacancies are shown in Fig. 5h, where the O_L, O_A and O_V ratios are summarized in Fig. 5i.³⁵ For x = 0.08, the O_V content reached 16.59%, whereas it decreased to 8.75% at x = 0.16. This variance underscores the positive impact of the varying SMS contents on inhibiting the generation of oxygen vacancies. In conclusion, the superior E_b of the ceramics can be primarily attributed to the decrease in G_a, increase in electrical insulation, increase in E_a and decrease in O_V content.

For energy storage capacitor applications in advanced pulsed power devices, operational stability under varying conditions is paramount. Here, our systematic evaluation of the representative BNBT-0.16SMS ceramic demonstrates exceptional performance across wide temperature and frequency ranges under a high electric field (450 kV cm⁻¹). As shown in Fig. 6a, the P–E loops shape remains stable across the temperature range of 30–140 °C. The W_rec and η exhibit minimal variation, with values of 5.7 J cm⁻³ (84.5%) at 30 °C and 5.99 J cm⁻³ (84.0%) at 120 °C (Fig. 6b). This negligible fluctuation confirms the exceptional thermal stability of BNBT-0.16SMS ceramic. The ceramic exhibits similarly impressive frequency insensitivity from 10 to 100 Hz, maintaining stable P_max values (ΔP_max <1.1 μC cm⁻²) (Fig. 6c). At 10 Hz, W_rec and η are 5.8 J cm⁻³ and 85.4%, respectively. At 100 Hz, these values decrease slightly to 5.6 J cm⁻³ and 84.0%, corresponding to marginal variations in W_rec and η of 3.0% and 1.6%, respectively (Fig. 6d). These results highlight the outstanding frequency insensitivity of the material. The excellent temperature stability originates from the structural design ensuring broad temperature stability. Fig. 6e presents the temperature-dependent Raman spectra of BNBT-0.16SMS ceramic. Within the temperature range of room temperature to 400 °C, no new Raman peaks emerge, and no significant peak shifts are observed, confirming that the temperature insensitivity of the local structure symmetry is a critical factor enabling superior thermal energy storage stability. Deconvolution analysis reveals that with increasing temperature, the Raman shift of the ν₅ vibrational mode associated with the BO₆ octahedra exhibits a red shift (lower wavenumber), accompanied by reduced Raman intensity and increased full width at half-maximum (FWHM). These spectral changes indicate enhanced local structural disorder, which facilitates the maintenance of stable P_r even at elevated temperatures.⁶⁸ Pulse charge–discharge performance is critical for evaluating practical application potential. Fig. 6g presents the overdamped charge–discharge current curves of BNBT-0.16SMS ceramic with optimal energy storage performance. Under test electric fields ranging from 100 kV cm⁻¹ to 720 kV cm⁻¹, the discharge curves remain stable, with discharge currents gradually increasing. The discharged energy density (W_dis) can be calculated using the following formula:⁶⁹

	(10)

where R and V are the overload resistance (10 kΩ) and the volume of the ceramic sample, respectively. Fig. 6h summarizes the relationship between W_dis and t_0.9 (time required to release 90% of stored energy) as a function of electric field. When the voltage increases from 100 kV cm⁻¹ to 720 kV cm⁻¹, W_dis sharply increases from 0.2 J cm⁻³ to 11.8 J cm⁻³, while t_0.9 remains below 2 μs. The slight decrease in W_dis compared to W_rec under the same electric field is attributed to thermal dissipation during practical charge–discharge processes. The abundant PNRs induced by this composition enable rapid domain reorientation, achieving exceptional charge–discharge performance. Fig. 6i compares the W_dis of BNBT-0.16SMS ceramic with other advanced lead-free energy storage bulk ceramics. Most lead-free ceramics exhibit W_dis below 8.0 J cm⁻³, with only a few reaching 10.0 J cm⁻³. In contrast, the BNBT-0.16SMS ceramic achieves an ultrahigh W_dis of 11.8 J cm⁻³, demonstrating its significant potential for advanced pulsed discharge devices. Therefore, the excellent thermal and frequency stability of BNBT-0.16SMS ceramic, combined with its high pulse discharge performance, demonstrate its significant potential for practical applications in high-power devices.

4. Conclusions

In this work, the domain configuration and defect structure were systematically modulated through chemical modification via SMS doping modification. This strategic chemical engineering approach simultaneously promotes highly dynamic polarization switching behavior and improved dielectric breakdown, thereby achieving superior energy storage performance. The optimized BNBT-0.16SMS ceramic demonstrates exceptional energy storage capability, with a W_rec of 13.6 J cm⁻³ and η of 84.9%. It also possessed outstanding temperature/frequency stability and remarkable W_dis = 11.8 J cm⁻³ under high electric fields up to 720 kV cm⁻¹. Microstructural analysis reveals that SMS incorporation effectively reduces grain size and suppresses oxygen vacancy formation, leading to significantly improved E_b. Complementary Vogel–Fulcher modeling and PFM analysis confirm that the engineered fine-scale weakly coupled nanodomains substantially reduce the energy barrier for polarization reversal, which facilitates rapid polarization switching, resulting in the simultaneous achievement of W_rec and η at high electric fields. Our work provides a feasible reference for the development of dielectric materials to meet future high energy storage characteristics and promotes the research of other lead-free energy storage dielectric materials.

Author contributions

Jiangping Huang: writing – review & editing, writing – original draft, data curation, investigation, conceptualization, methodology, visualization. Liang Deng: investigation, conceptualization, visualization. Yu Zhang: formal analysis, visualization. Yue Pan: formal analysis, visualization. Xiuli Chen: supervision, funding acquisition. Xu Li: supervision, writing – review & editing, supervision, funding acquisition. Huanfu Zhou: supervision, funding acquisition, resources.

Conflicts of interest

The authors declare that they have no known competing financial interests.

Data availability

The data collected and produced in this investigation are available upon reasonable request from the corresponding author.

The P–E hysteresis loops of BNBT-xSMS ceramics, and corresponding variations in energy storage parameters; The fitting of TB value; SEM images of BNBT-xSMS (x = 0.12 and 0.20) ceramics; The V–F fitting parameters of BNBT-xSMS ceramics. See DOI: https://doi.org/10.1039/d5ta04848f.

Acknowledgements

This study was supported by the Guangxi Science and Technology Plan Project (GuikeAD25069100), the Guangxi Natural Science Foundation Project (2025GXNSFBA069167), the Seedling Talent Inclusive Policy Scientific Research Fund, and the Guangxi Postdoctoral Innovation Talent Support Program.

References

1. H. Zubairi, Z. Lu, Y. Zhu, I. M. Reaney and G. Wang, Chem. Soc. Rev., 2024, 53, 10761–10790 RSC .
2. F. Yan, J. Qian, S. Wang and J. Zhai, Nano Energy, 2024, 123, 109394 CrossRef CAS .
3. S. Pattipaka, Y. Lim, Y. H. Son, Y. M. Bae, M. Peddigari and G. T. Hwang, Materials, 2024, 17, 2277 CrossRef CAS .
4. Y. Zhang, L. Chen, H. Liu, S. Deng, H. Qi and J. Chen, InfoMat, 2023, 5, 12488 CrossRef .
5. B. Guo, F. Jin, L. Li, Z.-Z. Pan, X.-W. Xu and H. Wang, Rare Met., 2023, 43, 853–878 CrossRef .
6. X. Lv, T. Zheng, C. Zhao, J. Yin, H. Wu and J. Wu, Acc. Mater. Res., 2022, 3, 461–471 CrossRef CAS .
7. F. P. Zhuo, H. Qiao, J. M. Zhu, S. Z. Wang, Y. Bai, X. P. Mao and H. H. Wu, Chin. Chem. Lett., 2021, 32, 2097–2107 CrossRef CAS .
8. Z. Yang, H. Du, L. Jin and D. Poelman, J. Mater. Chem. A, 2021, 9, 18026–18085 RSC .
9. G. Wang, Z. L. Lu, Y. Li, L. H. Li, H. F. Ji, A. Feteira, D. Zhou, D. W. Wang, S. J. Zhang and I. M. Reaney, Chem. Rev., 2021, 121, 6124–6172 CrossRef CAS .
10. D. X. Li, X. J. Zeng, Z. P. Li, Z. Y. Shen, H. Hao, W. Q. Luo, X. C. Wang, F. S. Song, Z. M. Wang and Y. M. Li, J. Adv. Ceram., 2021, 10, 675–703 CrossRef CAS .
11. B. Deka and K.-H. Cho, Materials, 2021, 14, 7188 CrossRef CAS .
12. W. Liu, B. Fu, J. Zhang, X. Ma, Y. Mao, Q. Zong, Z. Zhu, H. Yuan, Y. Zhou and W. Bai, Chem. Eng. J., 2025, 512, 162477 CrossRef CAS .
13. F. Z. Yao, Q. B. Yuan, Q. Wang and H. Wang, Nanoscale, 2020, 12, 17165–17184 RSC .
14. D. Yang, J. Gao, L. Shu, Y.-X. Liu, J. Yu, Y. Zhang, X. Wang, B.-P. Zhang and J.-F. Li, J. Mater. Chem. A, 2020, 8, 23724–23737 RSC .
15. V. Veerapandiyan, F. Benes, T. Gindel and M. Deluca, Materials, 2020, 13, 5742 CrossRef CAS .
16. L. Yang, X. Kong, F. Li, H. Hao, Z. Cheng, H. Liu, J.-F. Li and S. Zhang, Prog. Mater. Sci., 2019, 102, 72–108 CrossRef CAS .
17. R. Kang, Z. Wang, M. Wu, S. Cheng, S. Mi, Y. Hu, L. Zhang, D. Wang and X. Lou, Nano Energy, 2023, 112, 108477 CrossRef CAS .
18. F. Yang, Z. Pan, Z. Ling, D. Hu, J. Ding, P. Li, J. Liu and J. Zhai, J. Eur. Ceram. Soc., 2021, 41, 2548–2558 CrossRef CAS .
19. C. Long, Z. Su, H. Song, A. Xu, L. Liu, Y. Li, K. Zheng, W. Ren, H. Wu and X. Ding, Energy Storage Mater., 2024, 65, 103055 CrossRef .
20. Z. F. Zhang, Y. Zhang, H. R. Bai, P. Li, H. H. Huang, Z. L. Li, M. K. Joshi, W. Li, J. G. Hao, J. Du and P. Fu, J. Energy Storage, 2024, 44, 15160–15166 Search PubMed .
21. C. Long, Z. Su, A. Xu, H. Huang, L. Liu, L. Gu, W. Ren, H. Wu and X. Ding, Nano Energy, 2024, 124, 109493 CrossRef CAS .
22. S. Cheng, K. Zhang, C. Li, B. Zhang, J. Chen, K. Sun, C. zhou, J. Zhao, Q. Lin, G. Rao and S. Shi, J. Mater., 2024, 10, 803–810 Search PubMed .
23. R. Zhao, Y. Zhang, W. Li, X. Tang, K. Wang, J. Hu, Z. Shen, H. Fan, Y. Jiang and X. Guo, ACS Appl. Electron. Mater., 2023, 5, 6104–6113 CrossRef CAS .
24. Q. Lin, L. Li, W. Dou, Y. Wu, Y. Xie and Z. Deng, Ceram. Int., 2023, 49, 16225–16234 CrossRef CAS .
25. D. Wang, Z. Fan, D. Zhou, A. Khesro, S. Murakami, A. Feteira, Q. Zhao, X. Tan and I. M. Reaney, J. Mater. Chem. A, 2018, 6, 4133–4144 RSC .
26. W. Cao, L. Li, K. Chen, X. Huang, F. Li, C. Wang, J. Zheng, X. Hou and Z. Cheng, Adv. Sci., 2024, 11, 2409113 CrossRef CAS .
27. M. Zhang, S. Lan, B. B. Yang and H. Pan, Science, 2024, 384, 185–189 CrossRef CAS .
28. H. Pan, F. Li, Y. Liu, Q. Zhang, M. Wang, S. Lan, Y. Zheng, J. Ma, L. Gu, Y. Shen, P. Yu, S. Zhang, L.-Q. Chen, Y.-H. Lin and C.-W. Nan, Science, 2019, 365, 578–582 CrossRef CAS .
29. X. Dong, X. Wu, X. Lv and J. Wu, J. Mater. Chem. A, 2024, 12, 21772–21781 RSC .
30. J. Liu, Y. Jiang, W. Zhang, X. Cheng, P. Zhao, Y. Zhen, Y. Hao, L. Guo, K. Bi and X. Wang, Nat. Commun., 2024, 15, 8651 CrossRef CAS .
31. X. Qiao, A. Sheng, D. Wu, F. Zhang, B. Chen, P. Liang, J. Wang, X. Chao and Z. Yang, Chem. Eng. J., 2021, 408, 127368 CrossRef CAS .
32. J. Chen, F. Si, P. Zhao, S. Zhang and B. Tang, Ceram. Int., 2021, 47, 26215–26223 CrossRef CAS .
33. Z. Li, B. Xie, Z. Liu, K. Guo, K. Li, H. Zhang and H. Luo, J. Mater. Chem. A, 2025, 13, 9339–9346 RSC .
34. X. Tang, Z. Hu, V. Koval, B. Yang, G. C. Smith and H. Yan, Chem. Eng. J., 2023, 473, 145363 CrossRef CAS .
35. T. Shi, Q. Feng, J. Wu, Z. Cen, X. Chen, N. Luo, Y. Wei, X. Liu, J. Xu and C. Yuan, Chem. Eng. J., 2023, 470, 144205 CrossRef CAS .
36. O. Turki, A. Slimani, Z. Sassi, H. Khemakhem, N. Abdelmoula and L. Lebrun, Appl. Phys. A, 2022, 128, 186 CrossRef CAS .
37. Y. Pan, Q. Dong, J. Huang, X. Chen, X. Li and H. Zhou, Chem. Eng. J., 2024, 497, 154695 CrossRef CAS .
38. X. Dong, T. Hu, X. Wu, J. Yin, Z. Fu and J. Wu, SusMat, 2023, 4, 116–125 CrossRef .
39. Y. Pan, Z. Dai, C. Liu, X. Zhao, S. Yasui, Y. Cong and S. Gu, J. Mater. Sci., 2024, 59, 3284–3296 CrossRef CAS .
40. Q. Liu, J. Liu, D. Lu, W. Zheng and C. Hu, J. Alloys Compd., 2018, 760, 31–41 CrossRef CAS .
41. P. Shi, X. Wang, X. Lou, C. Zhou, Q. Liu, L. He, S. Yang and X. Zhang, J. Alloys Compd., 2021, 877, 160162 CrossRef CAS .
42. Y. Lin, D. Li, M. Zhang and H. Yang, J. Mater. Chem. C, 2020, 8, 2258–2264 RSC .
43. S. M. Wang, F. Yan, J. Qian, G. L. Ge, Z. Q. Fu, Z. B. Pan, F. Q. Zhang, J. F. Lin, K. Zeng, C. K. Chen, B. Shen, Z. F. Liu and J. W. Zhai, Energy Storage Mater., 2024, 66, 103155 CrossRef .
44. H. Yang, H. Qi and R. Zuo, J. Eur. Ceram. Soc., 2019, 39, 2673–2679 CrossRef CAS .
45. C. Long, Z. Su, A. Xu, F. Li, Y. Li, W. Ren, H. Wu, X. Ding and L. Liu, J. Adv. Ceram., 2025, 14, 9221063 CrossRef CAS .
46. M. b. Abdessalem, A. Aydi and N. Abdelmoula, J. Alloys Compd., 2019, 774, 685–693 CrossRef CAS .
47. A. Xie, L. Liu, Y. Zhang, A. Rahman and R. Zuo, J. Eur. Ceram. Soc., 2024, 44, 882–890 CrossRef CAS .
48. M. K. Bilal, R. Bashir, S. U. Asif and X. Jiyang, J. Energy Storage, 2024, 97, 112841 CrossRef .
49. H. Wang, S. Wu, B. Fu, J. Zhang, H. Du, Q. Zong, J. Wang, Z. Pan, W. Bai and P. Zheng, Chem. Eng. J., 2023, 471, 144446 CrossRef CAS .
50. Z. Wang, R. Kang, L. Zhang, X. Lou, Y. Zhao, P. Mao and J. Wang, Chem. Eng. J., 2023, 474, 145506 CrossRef CAS .
51. Q. Dong, D. Zeng, Y. Pan, P. Nong, X. Chen, X. Li and H. Zhou, Chem. Eng. J., 2024, 493, 152786 CrossRef CAS .
52. Y. Pan, Q. Dong, J. Huang, Y. Zhang, X. Chen, X. Li, L. Deng and H. Zhou, J. Mater. Chem. A, 2025, 13, 3749–3764 RSC .
53. Y. Fan, W. Qu, H. Qiu, S. Gao, L. Li, Z. Lin, Y. Yang, J. Yu, L. Wang, S. Luan, H. Li, L. Lei, Y. Zhang, H. Fan, H. Wu, S. Yu and H. Huang, Nat. Commun., 2025, 16, 3818 CrossRef CAS PubMed .
54. J. H. Duan, K. Wei, Q. B. Du, L. Z. Ma, H. F. Yu, H. Qi, Y. C. Tan, G. K. Zhong and H. Li, Nat. Commun., 2024, 15, 6754 CrossRef CAS PubMed .
55. X. Li, Y. Cheng, F. Wang, Q. Xu, Y. Chen, L. Xie, Z. Tan, J. Xing and J. Zhu, Chem. Eng. J., 2022, 431, 133441 CrossRef CAS .
56. J. Chen, P. Zhao, K. Chen, F. Si, Z. Fang, S. Zhang and B. Tang, Chem. Eng. J., 2024, 502, 157866 CrossRef CAS .
57. P. Chen and B. Chu, J. Eur. Ceram. Soc., 2016, 36, 81–88 CrossRef CAS .
58. P. Li, H. Yang, Q. Yuan and Y. Lin, Mater. Today Phys., 2024, 43, 101420 CrossRef CAS .
59. C. Long, Z. Su, H. Song, A. Xu, L. Liu, Y. Li, K. Zheng, W. Ren, H. Wua and X. Ding, Energy Storage Mater., 2023, 65, 103055 CrossRef .
60. H. Xie, H. Du, L. Liu, Q. Kou, J. Xu, Y. Sun, R. Lv, Y. Chang and D. Wang, Chem. Eng. J., 2022, 450, 138432 CrossRef CAS .
61. F. Yan, X. Zhou, X. He, H. Bai, S. Wu, B. Shen and J. Zhai, Nano Energy, 2020, 75, 105012 CrossRef CAS .
62. D. Yang, X. Wu, X. Lv and J. Wu, J. Eur. Ceram. Soc., 2025, 45, 117240 CrossRef CAS .
63. A. Deng and J. Wu, J. Eur. Ceram. Soc., 2021, 41, 5147–5154 CrossRef CAS .
64. X. Dong, X. Li, X. Chen, J. Wu and H. Zhou, Chem. Eng. J., 2021, 409, 128231 CrossRef CAS .
65. Y. Pan, Y. Zhang, Q. Dong, J. Huang, S. Zhao, X. Chen, X. Li and H. Zhou, Chem. Eng. J., 2025, 512, 162551 CrossRef CAS .
66. L. Ma, Z. Che, Y. Luo, C. Xu, Z. Cen, F. Toyohisa, S. Zhang, J.-F. Li and N. Luo, Acta Mater., 2025, 289, 120950 CrossRef CAS .
67. Q. Xu, M. T. Lanagan, W. Luo, L. Zhang, J. Xie, H. Hao, M. Cao, Z. Yao and H. Liu, J. Eur. Ceram. Soc., 2016, 36, 2469–2477 CrossRef CAS .
68. F. Chen, M. Chen, J. Zhang, W. Liu, H. Du, Q. Zong, H. Yu, Y. Zhang, J. Hao, J. Wang and J. Zhai, ACS Nano, 2025, 19, 1809–1818 CrossRef CAS .
69. Q. B. Liao, T. Deng, T. Lu, Z. Liu, N. Narayanan, S. Li, S. G. Yan, Y. Z. Bao, Y. Liu and G. S. Wang, Chem. Eng. J., 2024, 488, 150901 CrossRef CAS .

This journal is © The Royal Society of Chemistry 2025