Multiscale microstructure design for high-performance dielectric energy storage materials

Glass ceramics

Glass ceramics and dipole glasses are typically regarded as linear dielectric materials, characterized by their non-switchable polarization behavior and linear dielectric response under applied electric fields^[8]. Nevertheless, a fundamental drawback of these materials lies in their relatively low ε_r, which, despite their high E_b, limits the P_max and ultimately constrains their W_rec. To overcome these intrinsic limitations, microstructural engineering strategies have been increasingly explored.

Among these, compositional tailoring and process optimization remain the most effective and widely adopted approaches^[7], both aimed at introducing or forming a strong polar phase. Controlled heat treatment and crystallization protocols play a crucial role in modulating the microstructure. Compared with amorphous phases, crystalline domains typically exhibit higher dielectric permittivity and polarization capacity. Facilitating the controlled formation of such crystalline phases within a glassy matrix can effectively enhance the overall dielectric performance while retaining the thermal and chemical stability advantages of the glass-ceramic system.

The glass materials intended for dielectric ES are often referred to as dielectric glass ceramics or microcrystalline glasses, emphasizing the coexistence of crystalline and glassy phases^[22]. Liu et al. comprehensively summarized recent advances in high-W_rec FE glass-ceramics, analyzing the critical factors influencing their dielectric performance and outlining prospective directions^[89]. A direct microstructural design strategy for improving both the dielectric permittivity and inducible polarization in such materials lies in promoting controlled crystallization. One of the most effective routes to enhance crystallization and, consequently, the dielectric constant is to increase the crystallization temperature^[90-93]. For example, in the BaO-Na₂O-Nb₂O₅-SiO₂-TiO₂-ZrO₂ glass-ceramic system, the dielectric constant increased markedly from 36.8 at a crystallization temperature of 750 °C to 103.5 at 1,050 °C, primarily due to the formation of NN and Ba₂NaNb₅O₁₅ crystalline phases^[92]. In addition to thermal treatments, compositional engineering via the introduction of suitable dopants has proven highly effective in promoting crystallization and enhancing the dielectric response.

Rare-earth oxides or other functional additives are commonly employed for this purpose^[94-97]. In Gd₂O₃-doped BaO-K₂O-Nb₂O₅-SiO₂ glass ceramics, the increasing Gd₂O₃ content leads to the gradual emergence of crystalline phases such as Ba₂KNb₅O₁₅, BaNb₂O₆, and K₃NbO₄, which contribute to both higher dielectric permittivity and enhanced P_max under the same electric field^[98]. While tungsten bronze and other structures are effective, perovskite-type FE crystals typically offer even higher dielectric constants due to their intrinsic lattice polarizability. Thus, compositional control strategies aimed at promoting perovskite-phase crystallization represent a promising avenue for further dielectric enhancement^[99].

However, it is undeniable that microstructural strategies designed to promote crystallization, while improving the dielectric constant of glass ceramics, inevitably sacrifice other critical advantages inherent to linear dielectrics, such as dielectric stability, polarization linearity, low dielectric loss, and high W_rec^[100]. For linear dielectric materials, a clear trade-off exists between polarization capability and BDS. When employing this structural tuning strategy, careful consideration must be given to the trade-offs between various performance parameters based on the specific application requirements. Overall, linear dielectric ceramics, with their superior thermal/frequency stability, remain highly promising for applications in ultra-high-voltage pulsed circuits. For example, Shang et al. designed the “amorphous-disordered-ordered” microstructure in BT-based glass ceramics by the ingenious strategy of defect formation modulation, as shown in Figure 10, resulting in a high W_rec of 12.04 J/cm³ and a superb P_D of 973 MW/cm^3[101]. This representative work demonstrates a feasible route to obtain glass ceramics with an outstanding ESP and proves the enormous potential of glass ceramics in high and pulsed power applications.

Figure 10. Enhanced ESP in BT-based glass ceramics via defect formation modulation leading to an “amorphous-disordered-ordered” microstructure. (A) Schematic illustrating defect formation during crystallization without and with electric field assistance; (B) High-resolution TEM image of BTAS glass ceramic showing morphology and lattice fringes; insets display the selected area electron diffraction pattern and grain size distribution histogram; (C) P-E loops of BTAS glass ceramics measured at 10 Hz. The BTAS composition corresponds to glass ceramics crystallized under a 3 kV/cm electric field. This figure is quoted with permission from Shang et al.^[101]. ESP: Energy storage performance; TEM: transmission electron microscopy; BT: BaTiO₃; BTAS: BaO-TiO₂-Al₂O₃-SiO₂.

Grain refinement and density enhancement

The grain size, porosity, and phase uniformity of dielectric ceramic materials are key factors influencing their dielectric breakdown field; then, the dielectric and polarization responses are also affected based on size effects and clamping effects. A large porosity is highly detrimental to the breakdown field. Under the requirement of high W_rec, refining the grain size has been widely proven to improve BDS and is one of the most common strategies for grain structure design, which has been extensively validated across various systems^[102-106] as shown in Figure 11.

Figure 11. Grain refinement and densification enhance dielectric BDS. (A) Evolution of electric field and development of breakdown paths under a constant field of 320 kV/cm for 0.1-CS and 0.25-HPS ceramics. The 0.1-CS sample is a 0.9Bi_0.5Na_0.5TiO₃-0.1K_0.5Na_0.5NbO₃ ceramic prepared by conventional sintering, while the 0.25-HPS sample is a 0.75Bi_0.5Na_0.5TiO₃-0.25K_0.5Na_0.5NbO₃ ceramic prepared by hot-press sintering. This figure is quoted with permission from Deng et al.^[103]; (B) Simulated distributions of dielectric constant, electric potential, and local electric field in ceramic macrostructures before and after amorphous phase filling. This figure is quoted with permission from Zhang et al.^[106]. BDS: Dielectric breakdown strength.

Refining the grain structure can reduce defects and pores, thereby mitigating localized electric potential concentration. The increased grain boundary density and number act as insulating barriers, hindering the development of electrical treeing and reducing the probability of dielectric breakdown^[106]. Additionally, the fine-grained structure, through high-density grain boundaries, suppresses the migration of thermally activated charge carriers, reducing leakage current and Joule heat accumulation. This not only enhances the material’s thermal stability under high electric fields but also prevents thermal breakdown. The grain size directly influences the size and distribution of FE domains, with smaller grains significantly reducing domain size. Due to size effects, the FE macrodomains transition into nanosized domains or polar microregions. As a result, the ferroelectricity weakens or may even disappear, significantly lowering the coercive field, remnant polarization, hysteresis, and energy loss.

Moreover, grain refinement may affect polarization behavior. For example, in RFEs, small grains can enhance the local random field, suppressing remnant polarization, and improving η. Wang et al. investigated the impact of grain size and grain boundaries on the ESP of polycrystalline FEs using a phase-field model^[105]. As shown in Figure 12, their findings indicate that the depolarization field at the grain boundary induces vortex domains when the grain size is reduced or when the grain boundary thickness increases to a certain extent, leading to narrow P-E loops. The influence of these grain microstructural features on breakdown characteristics and polarization behavior can be utilized to improve dielectric ESP.

Figure 12. Influence of grain size on polarization behavior and ESP in polycrystalline FEs. (A) Domain structures of polycrystalline models at remnant polarization state with varying grain sizes; (B) P-E loops, (C) ESP, and (D) P_r as a function of average grain size. This figure is quoted with permission from Wang et al.^[105]. ESP: Energy storage performance; FE: ferroelectric.

However, it is important to note that smaller grain sizes are not always better; there may be an optimal size range. Excessively small grains may lead to an excessive number of grain boundaries, which can reduce the dielectric constant or maximum induced polarization, or cause difficulties in preparation and densification. According to Wang’s research, further decreasing the grain size or increasing the grain boundary thickness results in a decrease in W_rec due to a concurrent reduction in both remnant and saturation polarization^[105]. Therefore, it is crucial to select the optimal size range based on the material system^[107].

Grain refinement and densification are fundamental to improving the dielectric ESP of ceramic materials. The underlying mechanism involves enhancing BDS and optimizing polarization characteristics by fine-tuning the microstructure. This approach has proven to be a generally applicable strategy for boosting dielectric ES capacity. Several material preparation methods and chemical processing techniques can effectively achieve this goal. These processes typically encompass ceramic powder fabrication, sintering techniques, and the use of chemical additives.

Reducing the initial powder size, enhancing sintering activity, and inhibiting grain boundary migration during sintering are all effective strategies for controlling grain growth. Applying physical pressure during sintering facilitates particle rearrangement, mass transfer, and plastic flow, thereby promoting densification and lowering sintering barriers. A common approach is hot pressing, in which uniaxial pressure (typically 10-50 MPa) is applied during high-temperature sintering, accelerating densification and limiting grain growth. Rapid sintering techniques also contribute to both densification and grain refinement. One such method is spark plasma sintering (SPS), which generates high temperatures almost instantaneously through pulsed direct current while applying pressure, achieving rapid sintering and densification and suppressing excessive grain growth. Similar techniques include microwave sintering and flash sintering. In addition, chemical methods such as sol-gel processing, hydrothermal or solvothermal synthesis, and co-precipitation enable the preparation of homogeneous, nanometer-sized precursor powders, which, after sintering, yield fine-grained ceramics.

Sintering can also be optimized through two-step processes^[108] or the addition of grain growth inhibitors to enhance density and reduce grain size. For instance, introducing second-phase materials like MgO, SiO₂, or Y₂O₃, or doping with elements such as La³⁺ and Nb⁵⁺, can pin grain boundaries or reduce their migration, effectively suppressing grain growth. By combining physical and chemical methods and optimizing sintering parameters, it is possible to achieve coordinated control over both grain refinement (< 1 μm) and high density (> 98%), significantly improving the dielectric ESP.

When designing high-performance dielectric ceramics for ES, grain refinement is often coupled with other modification techniques. For example, chemical composition adjustments, doping, and multiscale structural designs (such as multilayer structures, core-shell configurations, or composite ceramics) are employed to further enhance performance. Liu et al. demonstrated exceptionally high BDS and ESP in lead-free BT-based RFE using a combined strategy of composition modification, viscous polymer processing, and liquid-phase sintering^[109]. Composition modification disrupted long-range FE order, while liquid-phase sintering enabled low-temperature densification and grain refinement, as shown in Figure 13A. Viscous polymer processing (VPP) allowed for thinner single-layer ceramic samples, illustrating an effective collaborative optimization strategy that could serve as a paradigm for the development of new dielectric ES ceramics.

Figure 13. Grain refinement and densification significantly enhance ESP in dielectric ceramics. (A) Ultrahigh dielectric BDS and superior ESP achieved in BT-based RFE ceramics through combined composition modification, viscous polymer processing, and liquid-phase sintering. This figure is quoted with permission from Liu et al.^[109]; (B) Schematic illustrating strategies to realize outstanding ESP in Bi_0.5Na_0.5TiO₃-based RFE ceramics. This figure is quoted with permission from Zhu et al.^[110]; (C) Grain refinement and pore reduction via two-step firing in 0.6BNT-0.21BT-0.19BiScO₃ ceramics, resulting in doubled BDS and enhanced ESP. This figure is quoted with permission from Xu et al.^[112]. ESP: Energy storage performance; RFE: relaxor ferroelectric; BDS: dielectric breakdown strength; BNT: (Bi_0.5Na_0.5)TiO₃; BT: BaTiO₃

For RFE materials, particularly those with rapidly saturating polarization, simply refining the grain structure to improve the breakdown field provides limited enhancement in ES. It is more effective to combine this with strategies designed to delay saturation polarization for significant performance improvements^[110-112], as illustrated in Figure 13B and C. Li et al. developed a composite dielectric with an intragranular segregation structure, achieving excellent ESP by delaying saturation polarization^[113]. This approach demonstrates how strategic modifications to polarization dynamics can significantly boost the ES capabilities of FE materials.

The FE properties of FE materials are highly dependent on grain size, with various dielectric properties exhibiting a pronounced size effect. Numerous studies have explored these effects and have established that the dielectric response, as well as other electrical characteristics, can be significantly influenced by grain size^[114-118]. Tan et al. have found that the grain size effect is universal and nearly independent of the sintering method and the starting powder^[117]. Zhang et al. studied the grain size dependence of domain wall activity in a lead-free (Ba,Ca)(Zr,Ti)O₃ system and demonstrated that the dielectric properties of polycrystalline FEs are influenced by the grain size, particularly through the dynamics of domain walls^[119]. Their research found that the highest dielectric permittivity occurs at intermediate grain sizes, which facilitate easier and faster domain wall dynamics, coupled with moderate lattice distortion.

Interestingly, the size effect can also be exploited as a strategy for suppressing ferroelectricity. By reducing polarization hysteresis, this approach can enhance dielectric ESP^[120]. It is widely applied in the production of MLCCs. Typically, BT is a FE material, exhibiting high polarization hysteresis and energy loss. However, low-voltage, high-dielectric MLCCs commonly employ nanoscale BT grains. These ceramics exhibit relaxation-like behavior rather than the typical FE response. Additionally, the incorporation of rare-earth ions into grain boundaries further improves performance by promoting the formation of multiphase ceramics and core-shell structures. This strategy extends the temperature range and stability of the dielectric constant, enabling the production of MLCC materials with standardized temperature stability parameters such as X7R, X8R, and X9R^[22,121-124].

Grain size and defect engineering have a significant impact on the dielectric properties of BT-based ceramics and the temperature coefficient of capacitance in MLCCs^[125-127]. For BT ceramics, grain refinement helps suppress tetragonal lattice distortion and ferroelectricity, thereby reducing polarization hysteresis and broadening the dielectric peak. Although the grain size effect is widely applied in low-field applications of BT-based MLCC devices, where a high dielectric constant is leveraged for functions, such as charge storage and filtering, the primary demand for dielectric ES materials focuses on exploiting high polarization responses for power amplification and pulsed discharge. In these materials, the most crucial structural requirement for industrialization remains densification and grain refinement. For dielectric ES ceramics, ensuring finely tuned grains is essential to achieve optimal performance, particularly in applications that rely on high polarization for ES and rapid energy release.

Composite ceramics

The design of composite FE ceramics at the grain scale aims not merely to alter the intrinsic physical or chemical properties of the matrix phase through dopant modification, but rather to harness the distinct advantages of multiple end-member phases through engineered composite microstructures. Broadly defined, a variety of structural configurations fall under the category of composite ceramics. These include multilayer architectures, ceramics with heterogeneous, additive-induced secondary phase precipitations, co-sintered systems comprising mixed compositional constituents, and materials exhibiting core-shell architectures formed via the diffusion of additives into grain lattices during processing. This section specifically focuses on 0-0 type composite ceramics, wherein compositional heterogeneity is introduced at the grain level.

The 0-0 composite paradigm refers to systems in which discrete grains of two different compositions are distributed throughout the ceramic matrix^[66]. These grains may share the same crystal structure but differ in chemical composition, or they may possess fundamentally distinct crystallographic symmetries. For instance, Hu et al. developed a 0-0 type composite by co-sintering two Pb-based RFE compounds with complementary functional characteristics. The high-breakdown-field Pb_0.94La_0.04(Zr_0.99-xSn_xTi_0.01)O₃ was combined with the high-efficiency Pb_0.8925Ba_0.04La_0.045(Zr_0.65Sn_0.3Ti_0.05)O₃ to construct a composite with markedly improved ESP. The resulting ceramic exhibited a high W_rec of approximately 7.15 J/cm³ and an η of about 90.5% under an electric field of 345 kV/cm^[66].

More commonly, high-performance composite systems are achieved by integrating materials with inherently distinct crystal structures and polarization mechanisms. These include combinations of perovskite phases with bismuth-layered, tungsten-bronze, or pyrochlore-type phases. For example, Wu et al. developed a ST-based diphase composite by introducing the bismuth-layered BaBi₂Nb₂O₉ into the perovskite relaxor matrix of Sr_0.6(Na_0.5Bi_0.5)_0.4TiO₃^[128], as shown in Figure 14A. This strategy yielded a ceramic capable of delivering a W_rec of approximately 3.6 J/cm³ and an outstanding η of 94.3% under moderate electric fields. Similarly, Wang et al. proposed a heterogeneous combination strategy involving embedding a high Eb plate-like pyrochlore phase in a high-polarization perovskite phase^[129], as shown in Figure 14B. The embedded plate-like pyrochlore enhances the breakdown field strength and facilitates the dynamic polarization response. Meanwhile, the strong spin-orbit coupling effect of the 5d electrons helps maintain the high polarization value of the perovskite. Other notable examples include pseudo-cubic BT matrix ceramics with Ba₄MgTi₁₁O₂₇ as the secondary phase^[130] (as shown in Figure 14C), as well as composites of BNT (a classic perovskite FE) with the pyrochlore-structured Sm₂Ti₂O₇^[131]. These systems collectively demonstrate that the deliberate design of grain-scale phase coexistence, especially between end-members exhibiting contrasting dielectric and polarization characteristics, provides an effective strategy for enhancing both W_rec and operational η in lead-based and lead-free dielectric ceramics.

Figure 14. Enhanced ESP through grain-scale composite ceramic design. (A) Incorporation of bismuth layer-structured BaBi₂Nb₂O₂ into perovskite Sr_0.6(Na_0.5Bi_0.5)_0.4TiO₃ ceramic via traditional solid-state synthesis to improve overall ESP. This figure is quoted with permission from Wu et al.^[128]; (B) Embedding high-breakdown-strength plate-like pyrochlore phase within high-polarization perovskite matrix to boost ES in BNT-based ceramics. This figure is quoted with permission from Wang et al.^[129]; (C) Superior ESP in BT-based RFE ceramics with reconstructed microstructure featuring elongated rod-shaped and polygonal perovskite grains. This figure is quoted with permission from Yin et al.^[130]. ESP: Energy storage performance; BNT: (Bi_0.5Na_0.5)TiO₃; ES: energy storage; RFE: relaxor ferroelectric; BT: BaTiO₃; BMT: Bi(Mg_1/2Ti_1/2)O₃.

In addition to conventional FE or relaxor components, high-melting-point oxides capable of forming distinct secondary phases, such as HfO₂^[132,133] and ZrO₂^[134,135], have emerged as promising candidates for engineering advanced composite FE ceramics. These oxides can be selectively introduced to create heterostructures with tailored local electric fields and optimized dielectric behaviors. For instance, Lu et al. developed a self-assembled metadielectric nanostructure by embedding HfO₂ nanophases within a BaHf_0.17Ti_0.83O₃ RFE matrix, as shown in Figure 15A. The inclusion of HfO₂ contributed to a significant enhancement in dielectric BDS, improved RFE characteristics, and a reduction in conduction losses. As a result, the fabricated thin-film capacitors achieved an exceptional W_rec of up to 85 J/cm³, with an η surpassing 81%^[132]. In a similar approach, Shen et al. incorporated HfO₂ into 0.75BNT-0.24NN-0.01ST RFE ceramics to construct heterogeneous composites. Here, HfO₂ particles were predominantly located along grain boundaries, acting as physical barriers that effectively suppressed the formation of localized electric branches while preserving high saturation polarization. This structural design strategy ensured a marked improvement in ESP and reliability under high fields^[133]. As shown in Figure 15B, isolated ZrO₂ secondary phases embedded in the perovskite matrix not only enhance the fracture toughness of the ceramics but also improve electrical insulation and broaden the bandgap, resulting in excellent stability, reliability, and superior charge-discharge performance.

Figure 15. Enhanced ESP via self-assembled metadielectric nanostructures and phase toughening in RFE ceramics. (A) Self-assembled BaHf_0.17Ti_0.83O₃-xHfO₂ metadielectric nanostructure: schematic of structural evolution, breakdown phase transitions, and corresponding P-E hysteresis loops under breakdown conditions. This figure is quoted with permission from Lu et al.^[132]; (B) Schematic illustrating improved ESP in ZrO₂-toughened Ba_0.55Sr_0.45TiO₃-Bi_0.5Na_0.5TiO₃-SrZrO₃ RFE ceramics due to enhanced BDS following ZrO₂ phase incorporation. The studied composition is 0.4 Ba_0.55Sr_0.45TiO₃-0.4 Bi_0.5Na_0.5TiO₃-0.2SrZrO₃. This figure is quoted with permission from Yang et al.^[134]. ESP: Energy storage performance; RFE: relaxor ferroelectric; BDS: dielectric breakdown strength.

The impact of composite architectures on polarization response is inherently multifaceted. A key advantage of such composites lies in their ability to leverage and synergize the functional merits of each constituent phase. Accordingly, the selection of end-member phases is typically guided by their complementary properties. One common strategy involves pairing a component possessing intrinsically high polarization capability with another phase, either crystalline or amorphous, that provides superior dielectric BDS. Alternatively, a high-polarization material may be coupled with a secondary phase known for its superior η, thereby balancing W_rec and η within a single composite system. Nonetheless, it is crucial to emphasize that the realization of these synergistic effects critically depends on the reproducibility and precision of the composite fabrication process. Achieving optimal performance requires the formation of dense microstructures with well-defined and stable interfaces. Care must be taken to prevent undesirable interdiffusion between phases, which can otherwise compromise the intrinsic functionalities of the individual components. Controlled phase distribution, grain boundary engineering, and sintering protocols, therefore, play an essential role in translating microstructural advantages into macroscopic dielectric performance.

Core-shell structures

In ceramic materials, the term “core-shell structure” commonly refers to grains exhibiting radial compositional gradients or inhomogeneities. It is important to clarify that the core-shell strategy discussed in this section pertains exclusively to microstructural design at the grain level within sintered ceramics, rather than to particle-level core-shell configurations formed during powder synthesis. For example, BT particles coated with SiO₂ through chemical modification are frequently referred to as core-shell powders. However, upon sintering, the SiO₂ coating typically transforms into an amorphous intergranular network. As such, this structural motif should be more appropriately classified under the domain of grain boundary engineering, rather than as a genuine core-shell structure within the ceramic grains. The distinction between these two strategies is fundamental and warrants careful attention by the reader.

Analogous to the 0-0 type composite ceramics described earlier, core-shell structured grains can be regarded as a particular form of compositional or structural heterogeneity, and thus constitute a subclass of composite ceramics. In practice, this design approach is often coupled with grain refinement and has found widespread application in contemporary BT-based MLCC^[136], where both dielectric performance and reliability critically depend on the intricate interplay between grain size and microstructure. The core-shell model has been extensively employed to rationalize the influence of grain size on the FE and dielectric properties of ceramics^[137,138].

In addition to their well-established role in tailoring dielectric and FE properties, core-shell structural strategies have emerged as a highly effective approach for enhancing ESP in dielectric ceramics. Fundamentally, the rational design of core-shell architectures enables the concurrent optimization of dielectric permittivity and BDS, two parameters that often exhibit an inverse relationship in homogeneous materials. By integrating a high-permittivity core material with a shell that exhibits superior dielectric strength, these heterogeneous structures overcome the performance limitations inherent to conventional single-phase systems. The core supports robust polarization, while the shell acts as a protective barrier that sustains high electric fields without premature failure. Moreover, such structural motifs can be exploited to modulate interfacial polarization and suppress dielectric losses. The introduction of dielectric gradient layers or space charge regions within the shell offers a mechanism to fine-tune the interfacial polarization response, thereby mitigating energy dissipation caused by mobile charge carriers. In addition, the high-resistivity shell layer blocks direct charge injection from the electrodes, significantly reducing conduction-related losses and further enhancing the overall η.

The presence of a shell also plays a crucial role in passivating surface defects at the grain boundaries of the core. By smoothing local potential fluctuations and reducing electric field concentration, the shell effectively delays the onset of dielectric breakdown. Through precise control over the relative dimensions of the core and shell, as well as the compositional gradient across the interface, it becomes possible to engineer a favorable balance between dielectric properties and insulation characteristics. This structural tunability offers a promising route toward simultaneous improvements in BDS, W_rec, and cycling stability. Typically, the core region retains a FE domain structure, contributing substantially to the net polarization, while the shell exhibits either relaxor-like behavior or linear dielectric characteristics with minimal hysteresis and enhanced electrical robustness.

For instance, Xue et al. developed BNT-based RFE ceramics featuring core-shell grains induced by the controlled diffusion of Ta⁵⁺ ions in the BNT-KTaO₃ system^[139]. As shown in Figure 16A and B, this diffusion-limited mechanism yields a chemical heterogeneity in which the Ta⁵⁺-depleted cores contain nanodomains approximately 10 nm in size, while the Ta⁵⁺-rich shells harbor PNRs on the order of 1-2 nm. Such a configuration enhances dielectric relaxation behavior, suppresses grain growth, and collectively contributes to improved temperature stability, elevated BDS, and favorable ESP, as shown in Figure 16C.

Figure 16. Core-shell grain structures enhance dielectric and ESP in lead-free RFE ceramics. (A-C) Core-shell microstructure and properties of BNT-KTaO₃-0.02wt% Li₂CO₃ ceramics: (A) grain morphology and schematic illustration, (B) high-angle annular dark-field (HAADF) image with corresponding EDS elemental mapping, and (C) P-E loops alongside ESP. This figure is quoted with permission from Xue et al.^[139]; (D-F) Ultrahigh ESP in KNN-based RFE ceramics with core-shell grains: (D) grain morphology, (E) EDS mapping, and (F) P-E loops of 0.85KNN-0.15Sr_0.7Nd_0.2ZrO₃-0.03Ag(Nb_0.45Ta_0.55)O₃ ceramics. This figure is quoted with permission from Chai et al.^[141]. ESP: Energy storage performance; RFE: relaxor ferroelectric; BDS: dielectric breakdown strength; BNT: (Bi_0.5Na_0.5)TiO₃; KNN: (K_0.5Na_0.5)NbO₃.

In a related study, Dong et al. engineered core-shell structured bulk ceramics in a (1-x)[0.90NN-0.10Bi(Mg_2/3Nb_1/3)O₃]-x(Bi_0.5Na_0.5)_0.7Sr_0.3TiO₃ system^[140]. Their work demonstrated that the presence of a polar core effectively suppressed local electric field concentrations in the surrounding weakly polar shell, significantly elevating the BDS and enabling a remarkable balance between high polarization and dielectric robustness.

Recently, Chai et al. reported KNN-based bulk ceramics featuring a grain-level core-shell structure combined with polymorphic nanodomains^[141], as shown in Figure 16D and E. Guided by compositional engineering, this microstructural design produced a synergistic enhancement of both polarization and BDS. As a result, the ceramics achieved an outstanding W_rec of approximately 20.4 J/cm³ and an η near 90% under an applied field of around 1,020 kV/cm, as shown in Figure 16F.

Beyond the widely explored FE-RFE and FE-LD composite frameworks, the combination of RAFE and RFE components has also demonstrated significant promise in the construction of core-shell grain structures tailored for dielectric ES applications. A representative example is provided by Xie et al., who developed a core-shell microstructure in a BT-modified NN-BiFeO₃ system^[142]. Their design capitalized on spontaneous compositional segregation, which induced a unique structural evolution from a RAFE to a RFE state. This transition manifested as a progressive transformation of hierarchical AFE nanodomains into nanoscale FE polar regions, enabling the core and shell to respectively inherit the high W_rec potential of the RAFE phase and the high η typical of RFE behavior. The resulting core-shell architecture exhibited an outstanding combination of W_rec and η, highlighting the potential of this strategy for advanced capacitor applications.

The core-shell concept has also been successfully extended to AFE-based bulk dielectrics, where grain-scale structural modulation plays a crucial role in modulating polarization dynamics and mitigating energy loss. Based on phase-field simulations, Wu et al. proposed a model wherein the core adopts AFE characteristics while the shell behaves as a LD^[143]. This heterostructure configuration effectively delayed the onset of saturation polarization and reduced hysteresis-related losses, thereby offering an improved balance between ES capability and η.

Although substantial evidence supports the effectiveness of core-shell grain architectures in enhancing dielectric ESP, several key challenges remain unresolved. In particular, ensuring interfacial chemical compatibility between core and shell phases is critical to maintaining structural coherence and polarization continuity. Moreover, the scalable fabrication of such microstructures remains technically demanding, especially under constraints of industrial viability. Structural stability under high electric fields and elevated temperatures also poses significant limitations for long-term operational reliability. Additionally, the development of precision coating techniques, such as atomic layer deposition, may enable unprecedented control over shell thickness and compositional gradients at the atomic scale. These emerging strategies could unlock the full potential of core-shell microstructures, positioning them as a cornerstone in the next generation of high-performance dielectric ES materials.

Grain boundary engineering

Grain boundary engineering (GBE) has emerged as a powerful microstructural design strategy at the grain scale for enhancing dielectric ESP. By systematically tailoring the structural, chemical, and electrical properties of grain boundaries, which are defined as the interfacial or amorphous regions between adjacent grains, GBE provides a versatile approach to simultaneously enhance critical functional metrics, including dielectric BDS, polarization behavior, and energy loss characteristics. The primary mechanisms underlying these improvements include increased material densification, reduced defect concentrations, enhanced grain boundary resistance, and the suppression of space charge accumulation. These effects collectively contribute to improved insulation and elevated η.

GBE strategies are typically implemented through the introduction of wide-bandgap dopants, intergranular phase modifications, or chemical coating techniques. One of the most representative approaches involves constructing a pseudo 0-3 type composite structure, wherein an amorphous phase forms a percolating three-dimensional network that hosts high-dielectric crystalline grains. This architecture is commonly realized using a two-phase system in which an amorphous matrix, generated from pre-synthesized glass additives or low-melting-point oxides, encapsulates discrete crystalline components during high-temperature sintering. The crystalline grains provide strong polarization capability, while the surrounding amorphous network contributes to high electrical insulation, thereby enabling a synergy between high dielectric permittivity and breakdown endurance^[71,144-147].

A typical example of such a design involves the incorporation of oxides with high bandgap energies and excellent dielectric strength into the composite system. Silica is a common choice for this strategy. As shown in Figure 17A, a core-shell structure of Sr_0.985Ce_0.01TiO₃ (SCT)@ SiO₂ combines a high dielectric permittivity core with an insulating shell material. After sintering, the SiO₂ enriches grain boundaries, enhancing dielectric breakdown^[144]. Silica-coated pre-synthesized powders are suitable for many material systems^[71,148]. For instance, Huan et al. employed a chemical coating technique to deposit a thin SiO₂ shell around KNN-BZT powders^[71]. As shown in Figure 17B, the resulting core-shell composites exhibited significantly improved dielectric BDS and fatigue endurance, attributed to the suppression of electrical conduction pathways and the stabilization of interfacial regions.

Figure 17. Grain boundary engineering enhances BDS to optimize ESP. (A) Experimental processing route and field-emission TEM images of SCT@7.0 wt% SiO₂; inset shows FFT of the selected area. Step 1: Ce-doped SrTiO₃ powders synthesized via the solid-state method. Step 2: Pre-doped powders coated with varying amounts of tetraethoxysilane (TEOS) using the Stöber process. This figure is quoted with permission from Qi et al.^[144]; (B) Schematic of grain configurations in KNN and KNN@SiO₂ FE ceramics, alongside microstructure topography and elemental distributions of KNN-BZT@SiO₂ ceramics. This figure is quoted with permission from Huan et al.^[71]; (C) Schematic illustrating multiscale optimization strategies to achieve ultrahigh ESP in lead-free MLCC based on RFE 0.87BT-0.13Bi(Zn_2/3(Nb_0.85Ta_0.15)_1/3O₃@SiO₂ MLCCs. This figure is quoted with permission from Zhao et al.^[149]. BDS: Dielectric breakdown strength; ESP: energy storage performance; TEM: transmission electron microscopy; KNN: (K_0.5Na_0.5)NbO₃; FE: ferroelectric; MLCCs: multilayer ceramic capacitors; BZT: Ba(Zr_0.4Ti_0.6)O₃; RFE: relaxor ferroelectric; FFT: fast Fourier transform.

In practical device architectures, GBE is often implemented in conjunction with other multiscale microstructural optimization strategies to realize superior overall performance. A particularly noteworthy example is the work of Zhao et al., who integrated GBE with the design of PNRs and multilayer device configurations, as shown in Figure 17C. At the atomic scale, compositional modulation was employed to induce relaxor-type nanodomains; at the grain scale, SiO₂ coatings formed dense intergranular barriers that reduced leakage current and mitigated local electric field concentrations; at the device level, MLCC structures were fabricated to leverage these microscale enhancements^[149]. This holistic design paradigm yielded remarkable results, with the MLCC devices achieving a W_rec of 18.24 J/cm³, an η exceeding 94.5%, and excellent thermal and cycling stability.

Recent advances in GBE have highlighted the transformative potential of multilayer chemical coating techniques applied to pre-synthesized ceramic powders. This strategy enables the construction of hierarchically tailored grain boundary architectures that can simultaneously modulate multiple interfacial properties, such as electric field distribution, defect density, and intergranular polarization dynamics. Among the pioneers in this area, Zhang et al. have conducted a series of influential studies that establish multilayer chemical coating as a highly effective approach for achieving complex microstructural control in dielectric ceramics^[146,147]. In a representative study, a ternary Bi₂O₃-B₂O₃-SiO₂ (BBS) glass was introduced to facilitate the densification of sintered ceramics. Simultaneously, La₂O₃, a wide-bandgap oxide with excellent insulating properties, was employed as a buffer layer at the interface between the glass phase and the ceramic particles, as shown in Figure 18A. This La₂O₃ interfacial coating enhanced uniformity and compactness and mitigated local electric field distortion, which is often a limiting factor for dielectric breakdown. As a result, the coated system demonstrated superior wide-temperature capacitive ES characteristics, with improved stability and η across operating conditions^[146]. Building upon this strategy, a more intricate multilayer configuration was developed by sequentially coating SrZrO₃-BiMg_0.5Sn_0.5O₃ (SZ-BMS) and a SiO₂-based glass onto Na_0.4K_0.1Bi_0.5TiO₃ (NKBT) particles, as shown in Figure 18B. This synergistic design enabled the realization of a ceramic system with outstanding electrical performance^[147].

Figure 18. Core-shell powder design and multiscale optimization in lead-free RFE ceramics. (A) Schematic of core-shell powder architecture for (Na_0.5Bi_0.5TiO₃@La₂O₃)-(SrSn_0.2Ti_0.8O₃@La₂O₃)-Bi₂O₃-B₂O₃-SiO₂ system and the corresponding sintered composite ceramics. This figure is quoted with permission from Zhang et al.^[146]; (B) Multiscale synergistic optimization in Na_0.4K_0.1Bi_0.5TiO₃@(SrZrO₃-BiMg_0.5Sn_0.5O₃)@SiO₂ composite ceramics, featuring pre-synthesized core-shell powders and bright-field TEM imaging. This figure is quoted with permission from Zhang et al.^[147]. TEM: Transmission electron microscopy; NBT: Na_0.5Bi_0.5TiO₃; SST: SrSn_0.2Ti_0.8O₃; NKBT: Na_0.4K_0.1Bi_0.5TiO₃; SZ-BMS: SrZrO₃-BiMg_0.5Sn_0.5O₃; BDS: breakdown strength; NWS: Maxwell-Wagner-Sillars; SST: SrSn_0.2Ti_0.8O₃.

Texturing engineering

An advanced strategy in microstructural engineering at the grain scale involves the deliberate alignment of crystallographic orientations, a technique commonly referred to as texturing. In conventional processing routes, ceramics are typically polycrystalline with randomly oriented grains, resulting in macroscopically isotropic properties. However, through techniques such as templated grain growth or electric-field-assisted sintering, it is possible to fabricate ceramics with controlled anisotropic grain orientations, known as textured ceramics. The principal advantage of texturing lies in its ability to exploit the intrinsic anisotropy of crystal structures to enhance targeted physical properties or mitigate adverse behaviors. This approach has been extensively adopted in piezoelectric and electrostrictive materials to boost their electromechanical response^[150,151].

In the context of dielectric ES ceramics, crystallographic orientation control provides a powerful tool to modulate dielectric response and polarization saturation trends, leveraging the anisotropic dielectric permittivity inherent to perovskite structures. For example, in tetragonal FEs, the dielectric constant reaches a maximum along the [001] direction. By preferentially aligning grains along this axis, as in [001]-textured BT-based systems, one can achieve rapid polarization response at relatively low electric fields, thereby realizing high ESP under moderate operating conditions. Alternatively, the fabrication of [111]-textured ceramics can delay polarization saturation, allowing for higher W_rec under elevated electric fields. These examples underscore the transformative role of crystallographic orientation engineering in tailoring polarization dynamics and dielectric behavior of ES ceramics, providing a versatile platform for optimizing performance across diverse application scenarios.

Recent advances have demonstrated that tailoring crystallographic orientation through texturing can serve as a powerful method to modulate electrostrictive effects in RFE ceramics, as shown in Figure 19A, thereby enhancing ES capability and operational reliability of dielectric devices. For instance, Li et al. successfully employed grain orientation engineering to exploit the intrinsic anisotropy of crystal structures in RFE ceramics, as shown in Figure 19B-D. They reported that <111>-oriented, high-quality textured BNT-based MLCCs exhibit significant suppression of electric-field-induced strain^[152], as shown in Figure 19E. This substantial reduction in electrostrain mitigates mechanical failure risks in MLCCs and contributes to a pronounced improvement in BDS, as shown in Figure 19F. The W_rec of such textured NBT-SBT MLCCs reached up to 21.5 J/cm³, surpassing the performance benchmarks of many state-of-the-art dielectric systems [Figure 19G].

Figure 19. Enhancement of BDS and ESP in polycrystalline ceramics through grain orientation control in textured MLCCs. (A) Finite-element simulations of electric field-induced local displacements and elastic energy densities for <100>-, <110>-, and <111>-oriented perovskite layers in a single MLCC ceramic layer; (B) Cross-sectional SEM image of the textured MLCC; (C) XRD patterns of textured ceramics; (D) Grain orientation mapping of <111>-textured NBT-SBT MLCC obtained by SEM-EBSD analysis; (E) E-induced strain, (F) Weibull distribution of BDS, and (G) ESP as a function of electric field for <111>-textured versus nontextured 0.65BNT-0.35Sr_0.7Bi_0.2TiO₃ MLCC. This figure is quoted with permission from Li et al.^[152]. BDS: Dielectric breakdown strength; ESP: energy storage performance; MLCCs: multilayer ceramic capacitors; XRD: X-ray diffraction; SEM: scanning electron microscopy; EBSD: electron backscatter diffraction; BNT-SBT: (Bi_0.5Na_0.5)TiO₃-(Sr_0.7Bi_0.2)TiO₃.

To elucidate the fundamental mechanisms underpinning this behavior, Wang et al. developed an electromechanical breakdown model that incorporates the electrostrictive contributions to breakdown dynamics in textured ceramics, as shown in Figure 20A. Their investigations revealed that BDS is strongly influenced by local electric field and strain energy distributions in the polycrystalline matrix. Through high-throughput computational simulations coupled with machine learning, they established rational guidelines for texture design to alleviate electromechanical breakdown, as shown in Figure 20B^[153]. Subsequently, Li et al. employed the templated grain growth technique to fabricate <111>-oriented MLCCs. Compared to their nontextured counterparts, these ceramics demonstrated a 37% reduction in electrostrain and a 42% enhancement in BDS^[154]. As a result, the MLCCs achieved an ultrahigh W_rec of 15.7 J/cm³ with an outstanding η exceeding 95% under a high field of 850 kV/cm.

Figure 20. Texture engineering modulates electromechanical breakdown in MLCCs. (A) Schematic illustrating electromechanical breakdown mechanisms in dielectric ceramics based on coupled electrical and mechanical responses; (B) Distributions of local electric field and local stress under 20 MV/m in the dielectric layers of <111>-textured and nontextured MLCCs. This figure is quoted with permission from Wang et al.^[153]. MLCCs: multilayer ceramic capacitors.

Taken together, these studies highlight the pivotal role of texturing as an advanced microstructural design strategy. By exploiting the anisotropic nature of FE crystals, texturing not only mitigates detrimental electrostrictive effects but also enables superior ESP in both bulk ceramics and MLCCs. Nevertheless, despite its demonstrated efficacy, the practical implementation of texturing techniques remains relatively complex and costly. Consequently, the pursuit of simpler, more cost-effective compositional modification strategies continues to attract considerable research interest^[71,155,156].

Domain wall pinning

Large polarization hysteresis and energy loss render conventional FEs suboptimal for dielectric ES applications. Consequently, minimizing polarization hysteresis has become a primary objective in microstructural engineering targeting improved ESP. The hysteretic behavior of FEs is intimately linked to the energy barriers that separate different macroscopic domain orientations^[26]. At the microscopic level, point defects can interact strongly with the polar structure, thereby altering the macroscopic dielectric and polarization response. A particularly well-studied mechanism involves the pinning of FE domain walls by defect dipoles. For instance, acceptor-oxygen vacancy defect dipoles can act as localized pinning centers that restrict domain wall mobility and reshape the polarization switching pathway^[26].

This domain wall pinning effect provides a promising approach for suppressing FE hysteresis and enhancing recoverable polarization, thereby contributing to improved W_rec and η. A representative example is provided by Zhang et al., who demonstrated that defect dipole engineering through B-site substitution in BT-based ceramics can significantly influence the polarization behavior. Specifically, the formation of [Li_Ti-V_O]-dipoles^[159], where Li⁺ ions correlate with neighboring oxygen vacancies, effectively modulates domain dynamics. As schematically illustrated in Figure 21, Li-doped BT exhibits a reduced P_r and an enhanced polarization difference (∆P), collectively leading to increased ESP.

Figure 21. Impact of Li doping on domain behavior and polarization in BaTi_1-xLi_xO₃ ceramics. (A and B) Schematic illustrating the doping mechanism: transition between nonpolar/centrosymmetric tetragonal FE state and polar tetragonal FE state, domain switching process, and corresponding P-E loops before and after doping; (C and D) Domain wall pinning by defect dipoles leads to reduced remnant polarization. This figure is quoted with permission from Zhang et al.^[159]. FE: Ferroelectric.

However, under high electric fields, elevated temperatures, or prolonged cycling, the defect dipoles may become depinned, causing a reversion to the classical square-shaped FE P-E loop. Additionally, activation of these defects can generate substantial leakage currents, thereby compromising electrical insulation reliability. Taken together, while domain wall pinning via defect dipole engineering presents a viable pathway for modulating FE response, its applicability in high-field pulsed ES devices remains limited due to stability and leakage constraints.

Inducing PNRs

One of the most widely employed and effective strategies for developing advanced dielectric materials involves disrupting long-range FE order through compositional modification. Introducing nanoscale chemical heterogeneities, including compositional fluctuations and gradients, fragments spontaneous macroscopic ferroelectric domains into PNRs. This transformation typically induces a RFE state near room temperature, characterized by dynamic, short-range polar order rather than static, long-range ferroelectricity.

RFEs have emerged as highly promising candidates for dielectric ES due to their unique combination of structural and functional properties. The nanoscale polar regions retain the ability to support significant dipolar reorientation under an electric field, thereby contributing to a large dielectric permittivity and high induced polarization^[160,161]. Simultaneously, the energy barriers for polarization switching in these systems are substantially lower compared to conventional FEs. As a result, RFEs exhibit exceptionally slim P-E hysteresis loops, indicative of minimal energy dissipation and negligible hysteresis. This behavior closely mirrors that of ideal paraelectric or linear dielectric materials^[162].

The RFE state is commonly achieved through compositional tuning in traditional FE systems, employing techniques such as isovalent or aliovalent substitution and solid-solution engineering. At the macroscopic level, this leads to a thermally activated relaxor or SPE phase near ambient conditions. Microscopically, the hallmark of this transformation is the breakdown of long-range FE domains into a disordered ensemble of PNRs, as shown in Figure 22A. The transition is typically accompanied by a reduction in the Curie temperature, broadening of the dielectric anomaly, and pronounced frequency dispersion in the dielectric response. Collectively, these features indicate a diffuse phase transition, which enhances thermal stability and expands the operational temperature window for dielectric applications.

Figure 22. Evolution from macroscopic domains to PNRs during composition-driven FE to RFE transition, enhancing ESP. (A) Schematic of evolution of dielectric response, P-E hysteresis loop, and polar microstructure during the composition-driven FE to RFE transition; (B) The characterization of the polar microstructure by piezoelectric force microscopy (PFM) for BNT-BT and BNT-BT-BMN. This figure is quoted with permission from Guo et al.^[166]; (C-E) Composition-dependent evolution of P-E loops for (1-x)BNT-xSrTiO₃, (1-x)(K_0.45Na_0.49Li_0.06)NbO₃-xSrTiO₃, and (1-x)AgNbO₃-x(Sr_0.7Bi_0.2)HfO₃ systems, respectively. This figure is quoted with permission from Xu et al.^[169]. PNRs: Polar nanoregions; FE: ferroelectric; ESP: energy storage performance; RFE: relaxor ferroelectric; BNT-BT-BMN: (Bi_0.5Na_0.5)TiO₃-BaTiO₃-Bi(Mg_2/3Nb_1/3)O₃.

Central to the behavior of RFEs is the emergence of random local electric fields. In a long-range ordered FE, internal fields are typically coherently aligned. However, the introduction of chemical disorder transforms these fields, rendering them spatially inhomogeneous and temporally dynamic. This disruption inhibits the development of coherent FE domains and instead promotes the formation of locally correlated, reorientable dipoles, which manifest as PNRs, as shown in Figure 22B. These local fields may arise from static sources, such as quenched compositional disorder, lattice vacancies, and impurities, or from dynamic phenomena, including non-ferroactive ion displacements away from their ideal lattice positions. Such field-induced frustration plays a crucial role in suppressing FE long-range order while enabling high dielectric tunability and low-loss polarization response^[161].

Inducing PNRs has emerged as the most widely adopted and effective microstructural strategy for designing high-performance dielectric ES materials, as shown in Figure 22. Among the various implementation methods, compositional engineering remains the most fundamental and versatile approach to achieve this transformation. In perovskite-structured inorganic dielectrics, a broad array of chemical modification strategies has been successfully applied to disrupt long-range FE order and promote the formation of RFE states under ambient conditions. These approaches include single-ion substitution, incorporation of complex ionic groups, isostructural solid-solution blending, and co-doping with multiple elements, each contributing to the local compositional heterogeneity necessary for RFE behavior. Representative systems such as BT^[163,164], BNT^[165,166], BiFeO₃ (BFO)^[167], KNN^[168], and even AFE^[169] matrices have all been extensively modified using these techniques to enhance their ES capabilities through relaxor behavior. The evolution of P-E loops typically shows a decrease in residual polarization and a reduction in hysteresis, as shown in Figure 22C-E^[169].

The resulting diversity in composition and structure is one of the primary reasons behind the remarkable complexity and variety of current dielectric ES materials. Despite this diversity, the design principle for constructing effective relaxor states remains consistent: the introduction of compositional disorder must avoid incorporating components that promote stronger FE ordering. Furthermore, achieving the desired relaxor characteristics requires the elemental distribution within the lattice to remain disordered at the nanoscale. Long-range chemical ordering, mesoscale segregation, and macroscopic phase coexistence are detrimental to the formation of a genuine relaxor state and typically lead to the reappearance of FE domains, thereby diminishing the slim-loop polarization behavior and associated ES benefits.

Through careful regulation of ionic size mismatch, charge compensation mechanisms, and local lattice strain, the incorporation of disorder can effectively destabilize FE long-range order while maintaining strong local dipolar interactions. This delicate balance enables high polarizability with minimal hysteresis loss, which is essential for achieving large W_rec and high η in capacitive ES applications.

In the design of RFEs for ES applications, a key strategy involves doping with ions that are not ferroelectrically active. Elements such as Bi^3+[170] or Pb²⁺, as well as Ti⁴⁺ or Nb⁵⁺, which typically enhance FE long-range order, are either excluded entirely or introduced only in trace amounts. The rationale behind this approach is to disrupt rather than reinforce the preexisting long-range polar order. This principle has been widely implemented in perovskite systems, where ionic substitution serves as a critical pathway for tuning polar microstructures. A representative example includes the Na_0.7Bi_0.1NbO₃ system, in which compositional modifications effectively suppress FE ordering and induce a relaxor state^[171]. Similar effects have been observed in non-perovskite structures as well^[172-174].

Among the various doping strategies, heterovalent ion substitution at ferroelectrically active lattice sites is one of the most effective methods for breaking long-range FE order. However, this approach comes with a trade-off: while it enhances relaxor behavior, it may simultaneously dilute the spontaneous dipole density, leading to a reduction in the maximum attainable polarization. As a result, simple doping strategies are typically insufficient to induce robust PNR formation on their own and are best employed in combination with other compositional or structural modifications to finely tune both relaxor characteristics and ESP. A further consideration is that heterovalent doping commonly introduces charged point defects such as oxygen or A-site cation vacancies. These defects can significantly compromise electrical insulation and η, especially under conditions of high temperature or strong electric fields^[175]. Therefore, while heterovalent ion doping is an indispensable tool in engineering dielectric properties, its application must be carefully balanced to avoid undesirable leakage conduction and thermal instability in practical device environments.

To avoid the formation of charge-compensating vacancies typically associated with heterovalent ion doping, an increasingly effective strategy involves the use of complex ion groups composed of two heterovalent ions. These co-doped configurations enhance local electric field fluctuations and disrupt long-range FE order more efficiently, thereby promoting the formation of PNRs, not only in perovskite-structured FE matrix^[176-178], but also in the tungsten bronze FE matrix^[179,180].

An even more sophisticated approach involves simultaneous A-site and B-site doping with carefully selected ion pairs. This co-substitution can promote relaxor behavior at relatively low dopant concentrations by inducing strong local field disorder across the crystal lattice^[181-183]. Additional examples include Ce-Mn co-doped Sr_0.4Ba_0.6Nb₂O₆ FEs^[184], as well as systems where Ba²⁺, Sr²⁺, and Sm³⁺ occupy the A-site, accompanied by partial Nb⁵⁺ substitution with Zr⁴⁺ in the B-site^[185]. These multi-ion doping strategies result in a compositionally disordered lattice, closely resembling that of solid solutions, in which randomly distributed cations serve as the origin of strong local electric fields. This disorder is essential in generating PNRs and promoting an RFE state. The resulting microstructure exhibits both high dielectric tunability and enhanced ES characteristics, making it a promising design route for high-performance dielectric materials.

Beyond simple ion substitution, the incorporation of nonpolar, weakly polar, or quantum paraelectric components into FE matrices has proven highly effective in disrupting long-range FE ordering. This disruption facilitates the formation of PNRs, enabling the emergence of ergodic relaxor (ER) states at room temperature, suppressing polarization hysteresis, and enhancing ESP. Quantum paraelectrics have been widely adopted as solid solution components to this end^[186-192]. Similarly, non-ferroelectric perovskite-structured compounds are frequently utilized to induce relaxor behavior by diluting long-range dipole alignment^[193-203].

An even more promising compositional design strategy involves the introduction of complex perovskite components based on heterovalent cation groups. Representative examples include A(B1,B2)O₃ and (A1,A2)BO₃^{[158,167,204]}. These systems incorporate compositional complexity that enhances local electric field randomness and strongly suppresses long-range FE correlations. In particular, multicomponent systems such as (1-x)Bi_0.5(Na_0.9Li_0.1)_0.5TiO₃-xSr(Al_0.5Nb_0.25Ta_0.25)O₃ demonstrate an advanced level of compositional design^[205]. These systems simultaneously contain ferroelectrically active ions to preserve dipole density and support high polarizability, alongside heterovalent ions that generate pronounced local field fluctuations. Such a combination not only enhances relaxor behavior but also effectively reduces hysteresis, leading to improved dielectric ES characteristics.

Microstructural engineering strategies aimed at inducing nanoscale domains or PNRs are equally applicable to AFE materials and can substantially reduce polarization hysteresis while enhancing η^[206,207]. In contrast to FEs, AFEs also exhibit long-range polar ordering, although the configuration differs in its antiparallel dipolar nature^[208]. For example, classical AFEs such as PbZrO₃ (PZ) accommodate antiferrodistortive instabilities, in which Pb²⁺ ions undergo A-O coupling to generate local spontaneous polarization arranged in an antiparallel fashion across neighboring unit cells.

While electric field-induced AFE-to-FE phase transitions contribute to an abrupt polarization increase and ES capability, they also introduce substantial hysteresis losses due to the large-scale structural transformation. This not only reduces η during charge-discharge cycles but also increases the risk of stress concentration and crack propagation, which are detrimental to material reliability. To mitigate these issues, substitution or doping at both A- and B-sites of the perovskite lattice can be employed to tune the relative stability between FE and AFE phases in PZ-based systems. Such modifications effectively reduce the scale of polarization ordering, enabling the formation of partial nanoscale domains or embedded PNRs. The choice of dopants or solid solution components follows similar empirical guidelines as those established for RFEs, focusing on enhancing local electric field fluctuations and disrupting long-range order.

However, it is crucial to note that the functional requirements for inducing the relaxor state in AFEs differ fundamentally from those in FEs. In FE systems, a complete transition to an ER state is typically desired to minimize hysteresis. In contrast, for AFEs, optimal ESP requires the retention of AFE structural features and polarization characteristics, while introducing moderate relaxor behavior to reduce hysteresis without eliminating the field-induced phase transition mechanism. As a result, the design strategies for inducing relaxor features in AFEs must be carefully balanced to preserve the beneficial aspects of the AFE-FE transition while improving cycling efficiency.

The most representative system of RAFEs is the Pb(Zr,Sn,Ti)O₃ family, in which equimolar or non-equimolar substitution of B-site ions such as Sn⁴⁺ and Ti⁴⁺ in the PZ matrix forms a continuous solid solution. This strategy preserves antiparallel spontaneous polarization at the A-site while introducing polar nanodomains or PNRs, thereby maximizing η^[209-211]. The incorporation of isovalent ions primarily serves to disrupt long-range A-O coupling without completely destroying the AFE ordering. Building upon the Pb(Zr,Ti,Sn)O₃ ternary solid solution, further improvements in relaxor characteristics and ESP can be achieved by introducing trace amounts of aliovalent dopants at the A-site. These dopants modulate the relaxor and AFE features in the P-E hysteresis loop. A notable example is (Pb_0.94La_0.04)(Zr_0.49Sn_0.5Ti_0.01)O₃^[212], which exhibits a distinctive double P-E loop with both AFE and relaxor characteristics. Numerous PZ-based RAFE systems reported with superior ESP follow similar compositional strategies^[213-217]. Moreover, due to the structural and polar behavior similarities between PbZrO₃ and PbHfO₃^[218], the microstructural design strategies established for PZ-based RAFEs are also applicable to PbHfO₃-based systems. These approaches include simple ionic doping and the incorporation of secondary components through solid solution to induce polar nanodomains and PNRs, enabling further enhancement of ES capability in PbHfO₃-based relaxor AFEs^{[107,217,219-221]}.

In lead-free AFE systems, similar compositional modification and structural design strategies can be applied to enhance ESP^[222-226]. Beyond inducing relaxor characteristics, several specialized nanoscale microstructural design approaches have also been explored. Notably, inspired by core-shell architectures in glass ceramics, Tang et al. engineered ceramic-like AFE regions within a self-generated glass-ceramics-like structure. Benefiting from simultaneous improvements in P_max and BDS, the (Eu_xAg_1-3x)NbO₃ compounds achieved ultrahigh W_rec and η^[227]. This work provides a compelling paradigm for advancing the ESP of AFE-based ES MLCCs to meet the stringent demands of next-generation applications.

In the composition-driven induction of PNRs and nanoscale domains, a variety of ionic dopants and solid-solution modifications have been employed to construct new RFE systems with enhanced dielectric ESP. This approach has led to a proliferation of reported dielectric materials with increasingly complex yet compositionally similar formulations. Although many of these systems share the common goal of disrupting long-range FE ordering to induce PNRs, there are still distinct compositional design principles that can be followed to achieve targeted performance enhancements.

During this component-driven process, many specialized strategies have been proposed and validated, including polarization mismatch and reconstruction^{[160,163,228,229]}, high-entropy engineering^[230-242], and PNR characteristic regulation^[243-245]. These strategies provide guiding mechanisms from multiple dimensions for inducing PNRs. They not only include enriching PNR types^{[165,243,246-257]}, reducing PNR size and improving dynamics^{[244,246,252-254,258-260]}, but also regulating PNR shape^[261-265], weakening coupling between PNRs^[266-268], and even combining multiple PNR characteristic regulations^[269,270]. In general, component design follows the empirical principles outlined above and remains within the broader category of “component-driven PNR induction”^{[142,181,197,243,248,249,269,271-273]}. Specific cases are therefore not detailed here.

AFE periodic/phase modulation

AFE materials offer a unique strategy for controlling polarization response pathways through the design of ordered AFE periodic structures. Analogous to FE domains, these periodic structures exist at comparable length scales, making periodic modulation a representative approach for domain and microregion manipulation.

The modulation of AFE polarization ordering in these periodic structures can be interpreted as a phase transition, where the microstructural design strategy primarily governs the sequence of field-induced phase transitions and determines the critical electric field required for such transitions^[274,275]. Typically, once the polarization in an AFE material undergoes a sharp increase upon reaching a certain field, further increases in the electric field contribute little to enhancing the polarization response. Consequently, the most effective working electric field for AFE-based dielectric materials is slightly above the phase transition-driving field. This periodic modulation strategy is particularly effective in optimizing ESP under specific electric fields; the key lies in matching the phase transition-driving electric field with the operating field. For example, to achieve high ESP under elevated electric fields, the driving field should be tuned toward higher values^[275,276]. Conversely, for cost-effective ESP under lower or moderate fields, the driving field should be shifted to lower values^{[219,277,278]}. In recent work, Ge et al. demonstrated the effectiveness of this approach by stabilizing the AFE phase through La³⁺ and Cd²⁺ doping, thereby modulating the incommensurate phase of the material^[216].

Furthermore, AFE materials exhibit complex behaviors, including the coexistence and competition of multiple phases, field-induced multistage phase transitions, and a degree of diffuse phase transition^[214,275]. Li et al. have described the field-induced multistage phase transition in Pb-containing AFE materials, driven by competing FE and AFE interactions, as the “electric devil’s staircase.” This unique transition, characterized by numerous degenerate dipole configurations, facilitates superior ESP. For instance, in (Pb_0.97La_0.02)(Zr_0.50Sn_0.50)O₃ AFE ceramics, the modulated structure can be readily adjusted through chemical substitution. Moreover, the temperature-driven electric devil’s staircase behavior enhances W_rec and thermal stability across a broad temperature range, making it highly suitable for high-power AFE capacitor applications^[279]. Similarly, Li et al. proposed a phase modulation strategy to optimize ESP. In Pb-containing AFE ceramics, multistage phase transition behaviors were effectively controlled via chemical modification, enhancing both BDS and the phase-switching field, resulting in excellent ESP for (Pb_0.92Gd_0.02Sr_0.05)(Zr_0.87Sn_0.12Ti_0.01)O₃ ceramics^[280].

Hu et al. have made significant contributions to the understanding of phase transitions and periodic modulation in AFE materials. Using in-situ biasing TEM, they investigated two representative AFE compositions exhibiting multiple and double hysteresis loops. Their findings revealed that as modulation periods increase, AFE materials undergo additional transitions between various AFE phases prior to the final AFE-FE phase transition, as illustrated in Figure 23A. This progression leads to the formation of a favorable multiple-loop configuration, thereby enhancing ESP^[281,282].

Figure 23. Periodic modulation of AFE materials at the domain scale. (A) Schematic of defect-driven phase transitions in the (Pb,La)(Zr,Sn,Ti)O₃ AFE system. Colors indicate modulation periods of 4, 5, and 6 in the AFE phase (blue, yellow, purple) and FE phase (pink). This figure is quoted with permission from Hu et al.^[281]; (B) Model of phase transition engineering with electric-field-induced polarization switching in Pb_0.97La_0.02(Zr_0.35Sn_0.55Ti_0.10)O₃ AFE ceramics, where N denotes the modulation period; (C) Structure-property correlation diagram illustrating the evolution of multilevel structures under varying electric field stimuli. This figure is quoted with permission from Hu et al.^[283]. AFE: Antiferroelectric; FE: ferroelectric.

In 2024, Hu et al. proposed a phase transition engineering strategy by applying pulsed electric stimuli near the critical electric field. This approach resulted in a ~54.3% enhancement and rapid stabilization of capacitance density in Pb_0.97La_0.02(Zr_0.35Sn_0.55Ti_0.10)O₃ AFE ceramics^[283]. The electric stimuli induced irreversible recovery of the modulated structures, driving successive structural evolution, including domain transformation from multidomain to monodomain states and a shift in the modulation period, as illustrated in Figure 23B. As shown in Figure 23C, Hu et al. effectively employed component- and field-induced periodic modulation/phase transition control to enhance ESP in AFE materials^[283]. Collectively, this series of studies serves as a valuable reference for similar research endeavors.

In conclusion, periodic and phase-transition modulation not only allows for precise tuning of polarization response pathways to optimize ESP but also mitigates polarization hysteresis arising from field-induced phase transitions^[284,285], thereby enhancing η^[217]. This microstructural design strategy provides an effective approach for enhancing the ESP of AFE materials at the domain and microregion scales.

Suppression of charged defects and conductivity

An ideal dielectric material should behave as a perfect insulator. However, due to doping defects or element volatilization, point defects-primarily A-site and oxygen vacancies-are inevitably introduced into the crystal lattice^[286]. Additionally, changes in element valence can generate conductive pathways within the lattice. For instance, in BT-based dielectric materials, the reduction of Ti⁴⁺ to Ti³⁺ facilitates electron hopping between Ti⁴⁺ and Ti³⁺ sites, contributing to long-range conductivity and compromising the material’s insulating properties.

Effective strategies for mitigating leakage conduction, reducing polarization hysteresis, and improving stability involve lowering defect concentration, reducing defect mobility, and increasing activation energy barriers^[287-290]. The most common approach is optimization of the sintering process. Techniques such as buried sintering or the addition of volatile elements can alleviate defects caused by element loss^[291]. Advanced sintering methods, including reaction-sintering, have been shown to reduce energy losses relative to conventional solid-state sintering^[292]. Lowering the sintering temperature can further mitigate volatilization, which can be achieved by employing chemical additives, highly reactive raw materials, or high-energy ball milling to reduce particle size^[293]. Hydrothermal synthesis can also produce uniform spherical nanoparticles that reduce defect formation. Additionally, charge compensation by substituting higher-valence ions for lower-valence ions can suppress defects-for example, Ta⁵⁺ replacing Ti⁴⁺ reduces oxygen vacancies in BNT-based RFE ceramics, enhancing dielectric ESP^[294]. Controlled atmospheres during sintering, such as oxidizing environments or post-sintering oxygen annealing, also minimize element reduction and defect formation.

These defect suppression strategies are directly relevant to enhancing dielectric ESP. For example, Lou et al. introduced an excess of K to suppress volatile Schottky defects, preventing conduction losses and enhancing ESP in tetragonal tungsten bronze-structured dielectric bulk ceramics^[286]. Similarly, Che et al. synthesized (1-x)BNT-xAgNb_0.5Ta_0.5O₃ RFE ceramics in a flowing oxygen atmosphere, eliminating defects such as metallic silver and oxygen vacancies, which significantly improved BDS and reduced hysteresis^[295]. Nb doping has also been used to increase resistivity by eliminating hole conduction and promoting microstructural homogeneity^[296].

For industrial MLCC production, a synergistic approach integrating chemical additives, tailored sintering protocols, and atmospheric control is essential to ensure dielectric reliability and performance^{[287,297,298]}. A central strategy involves aliovalent ion doping for charge compensation, mitigating the formation of charge carriers and suppressing electronic conduction^{[290,299,300]}. Stabilizing defect associations has also been shown to inhibit the proliferation of detrimental point defects during high-temperature processing^[288]. Complementary compositional modifications can enhance redox resistance, providing further protection against oxygen vacancy formation under reductive conditions^[301,302].

Beyond chemical methods, sintering techniques play a critical role. Multistage sintering, applying distinct atmospheric conditions across temperature regimes, has proven effective in reducing defect density while preserving phase integrity^[297]. Importantly, defect suppression often functions synergistically with other microstructural design strategies, such as grain size refinement and composition-driven transitions from long-range FE domains to PNRs, thereby further enhancing the dielectric and ESP of the ceramic^[303].

Defect dipoles and defect engineering

Defect engineering has emerged as a powerful lattice-level strategy for tailoring polarization behavior in dielectric materials. In bulk dielectric ceramics, this approach is primarily associated with extrinsic point defects introduced via the incorporation of dopant atoms or ions into the host lattice, a process commonly referred to as doping. It is important to note, however, that from the standpoint of defect chemistry, even isovalent substitutions such as those in Ba(Ti_1-xSn_x)O₃ or (Ba_1-xSr_x)TiO₃ formally generate point defects. Yet, within the field of electronic ceramics, these substitutions are generally excluded from conventional defect engineering frameworks. Instead, the term “defect engineering” is typically reserved for aliovalent doping and the intentional creation of charged vacancies, most notably oxygen or A-site vacancies. Accordingly, this section focuses on the narrower definition of defect engineering widely adopted in dielectric research.

To harness defect engineering effectively while preserving insulation reliability, it is essential to limit the defect concentration and ensure that defect activation occurs well above room temperature. This strategy enables the modulation of polarization dynamics, particularly in RFEs, while suppressing unwanted leakage. For example, Li et al. reported an A-site defect engineering approach that enhanced the dielectric ESP of Ba_0.105Na_0.325Sr_0.245-1.5x□_0.5xBi_0.325+xTiO₃ RFE ceramics (x = 0.06). Here, the controlled introduction of A-site vacancies and cation disorder disrupted A-site ion alignment and stabilized the polarization response under low fields, ultimately yielding a higher W_rec^[304].

The influence of extrinsic point defects on polarization and ES behavior is primarily mediated through their interaction with microscopic polar structures. A comprehensive understanding of these mechanisms is therefore critical for advancing dielectric ceramics. A representative example is found in (Pb,La)(Zr,Sn,Ti)O₃ (PLZST)-based RAFEs^{[214,219,305]}. In this system, partial substitution of Pb²⁺ with La³⁺ at the A-site introduces a charge imbalance, which is compensated by the formation of A-site vacancies. These defect dipoles subtly disturb the long-range polar order of Pb²⁺, while B-site compositional fluctuations generate strong local fields. Together, these effects promote dielectric relaxation and suppress macroscopic ferroelectricity, resulting in a distinctive RAFE state characterized by both double hysteresis loops and RFE properties.

Defect engineering has also been widely employed in RFEs to modulate polarization response and thereby optimize dielectric ESP. A notable case is the BNT-based RFE system, which exhibits field-induced RFE-FE transitions within a critical temperature window and yields constricted P-E loops. In 2018, Li et al. proposed a compositional design involving the incorporation of (Sr_0.7Bi_0.2)TiO₃ (SBT), a perovskite component with intrinsic A-site vacancies, into the BNT matrix. This strategy introduced nanoscale chemical and structural heterogeneity, producing AFE-like double hysteresis loops. As a result, the engineered material achieved a W_rec of 9.5 J/cm³ with an η of 92% under moderate fields^[306]. Subsequent studies further confirmed the beneficial role of A-site vacancies in BNT-based RFEs. For instance, Liu et al. demonstrated that deliberate A-site deficiencies could effectively suppress remnant polarization and enhance η^[307]. The local fields and structural frustration induced by defect engineering facilitated the field-driven assimilation of PNRs with diverse local symmetries, enabling a recoverable polarization response with minimal hysteretic loss^[248].

The modulation of polarization dynamics through defect engineering relies critically on the interactions between charged point defects and dipolar entities within the lattice. In functional electronic materials, such defect-dipole interactions are widely exploited to tune local energy barriers, thereby governing dipole responses to external electric fields^{[189,308-310]}. Numerous studies have employed aliovalent ion substitution to introduce localized defects into dielectric matrices as a route to optimize ESP. In RFEs, particularly those based on BNT, the inherently low energy barriers associated with polarization reversal of PNRs, as well as transitions between PNRs of different symmetries, often lead to rapid polarization saturation and relatively low critical electric fields. These features constrain their effectiveness in applications requiring both high energy density and robust dielectric performance. Defect engineering offers a promising pathway to overcome these limitations by raising the energy barriers between distinct PNR orientations. By increasing the activation field required for polarization switching, AFE-like features in the P-E loops of BNT-based RFEs can be significantly enhanced^[311].

A representative example is provided by Zhang et al., who developed a defect-engineering strategy based on aliovalent substitution with rare-earth La³⁺ ions^[311]. The introduction of 3 at.% La³⁺ into BNT ceramics markedly modified local energy barriers, elevating both the critical and saturation fields during polarization switching [Figure 24A]. These effects are attributed to the reinforcement of energy barriers associated with lattice torsion and dipole alignment [Figure 24B]. As a result, La³⁺-modified BNT ceramics exhibited enhanced AFE-like relaxor behavior, achieving a W_rec of 8.58 J/cm³ and an η of 94.5%. Despite this progress, such defect-mediated approaches primarily deliver quantitative improvements rather than qualitative breakthroughs. The double hysteresis loops observed in La-doped BNT-based RFEs remain distinct from those of true AFEs, both in loop shape and switching dynamics. This distinction underscores the need for innovative defect-dipole design strategies to bridge the performance gap between relaxor systems and ideal AFEs.

Figure 24. Defect engineering-enabled polarization regulation in AFE-like RFEs for enhanced ESP. (A) Schematic illustrating polarization modulation via defect engineering; (B) Effects of La³⁺ doping on the local crystal lattice structure. This figure is quoted with permission from Zhang et al.^[311]. AFE: Antiferroelectric; RFE: relaxor ferroelectric; ESP: energy storage performance.

At the lattice scale, defect structures also play a critical role in advancing the ESP of thin-film capacitors^[312]. One key insight is that the formation of defect complexes can alleviate the intrinsic trade-off between P_max and BDS, a major limitation in high-performance dielectric thin films. A notable example is provided by Luo et al., who used Mn doping to tailor the defect chemistry of Sr_0.7Bi_0.2TiO₃-based RFE thin films. In this system, Mn²⁺ ions formed stable dipoles with oxygen vacancies, thereby mitigating oxygen and titanium vacancy concentrations^[313]. This defect configuration stabilized a slush-like “single-domain” state characterized by fluctuating B-site cation displacements, ultimately leading to a substantial improvement in ESP.

A similar influence of defect engineering has also been demonstrated in bulk dielectric materials^[314], extending beyond RFE-based dielectrics to AFE systems^[315,316]. The central effect is the reduction of hysteresis losses and the simultaneous enhancement of dielectric response. For instance, to address A-site volatility and Ti⁴⁺ reduction in BNT-based RFEs, Li et al. introduced Mn²⁺ dopants, which effectively suppressed Ti⁴⁺ reduction while promoting off-centering displacements of cations^[154]. This strategy not only minimized energy dissipation under high fields but also boosted net polarization, delivering a remarkable enhancement in ESP. Such lattice-level defect engineering has proven highly effective in reducing energy loss and improving η in BNT-SBT-based MLCCs operated under strong electric fields^{[154,306,317]}.

Dipolar glass and quasi-linearization

In certain dielectric systems, when dipoles fail to establish long-range or even mesoscopic correlations and polar order remains confined to the scale of individual unit cells, the material is typically classified as a dipolar glass. Prototypical examples include doped quantum paraelectrics such as (K_1-xLi_x)TaO₃, K(Ta_1-xNb_x)O₃, and Sr(Ti_1-xNb_x)O₃, in which defect dipoles are sparsely embedded within a nonpolar host matrix^[318]. Beyond inorganic paraelectrics, numerous polymer systems also exhibit glassy dipolar states, particularly those where polarization arises from pendant polar side chains grafted onto the polymer backbone^[319].

From a physical standpoint, dipolar glasses can be regarded as an extreme limit of RFEs, in which the dipole concentration is heavily diluted. Consequently, their dielectric permittivity exceeds that of conventional nonpolar dielectrics but remains far below that of classical FEs. A key distinction between dipolar glasses and relaxors lies in the degree of correlation among polar entities. Whereas RFEs may undergo field-induced or spontaneous FE transitions below the freezing temperature, dipolar glasses lack sufficient interaction strength to sustain such transformations. In these materials, the spatial extent of dipolar coupling remains shorter than the average dipole-dipole separation, even at cryogenic temperatures^[318,320].

This absence of collective polarization implies that even under strong external fields, dipolar glasses cannot develop FE long-range order. Their dielectric response therefore closely resembles that of linear dielectrics, characterized by modest permittivity and limited field-induced polarization. Nonetheless, the highly stable and rigid local structure endows dipolar glasses with exceptionally high BDS. These attributes make them attractive candidates for ultrahigh-voltage ES capacitors, where minimal polarization loss and superior dielectric reliability under extreme fields are critical.

A promising design route for next-generation dielectric ES materials is to exploit the intrinsic features of quantum paraelectrics by deliberately introducing randomly distributed, uncorrelated dipoles at the lattice scale. Such dipolar glasses are defined by high η, nearly linear polarization, and excellent thermal and field stability^{[102,321,322]}. This approach circumvents nanoscale polar correlations and instead harnesses lattice-level disorder for functional performance. For example, Zhang et al. reported that co-doping Mg²⁺ and Nb⁵⁺ into Ca_0.5Sr_0.5TiO₃, substituting for Ti⁴⁺ at the B-site, produced a dielectric with nearly ideal linear polarization. The material achieved W_rec = 7.62 J/cm³ and η = 92% at 640 kV/cm, ranking among the most promising bulk LDs to date^[323].

Building on this concept, Fu et al. proposed a dipolar-glass framework based on high-entropy design [Figure 25A]. By substituting multiple heterovalent, ferroelectrically active cations onto equivalent lattice sites in (Bi_1/3Ba_1/3Na_1/3)(Fe_2/9Ti_5/9Nb_2/9)O₃, they generated a dense network of uncorrelated dipoles with dissimilar polarization vectors across neighboring unit cells [Figure 25B]^[324]. This structure permitted highly diffused dipole reorientation under applied fields while suppressing long-range growth of polarization, thus preventing domain formation. The resulting ceramic exhibited outstanding performance, with W_rec = 15.9 J/cm³ and η = 93.3%.

Figure 25. Design strategies for high-performance dielectric capacitors based on highly polarizable concentrated dipole glass (HPCDG) and QLD behavior. (A and B) HPCDG design: (A) Temperature-dependent dielectric response sketches of typical RFEs at various frequencies, with corresponding polar structures and polarization responses for nonergodic relaxor (NR, or FE), ergodic relaxor (ER), superparaelectric ergodic relaxor (SPE ER), canonical dipole glass (DG), paraelectric (PE) states, and HPCDG. The inset shows the schematic design strategy of HPCDG; (B) Composition strategy and structural mechanism of (BiBaNa)(FeTiNb)O HPCDG, illustrated by phase-field-simulated two-dimensional domain structures of BF-BT-2/9BNT-2/9NN. This figure is quoted with permission from Fu et al.^[324]; (C and D) Novel RFEs with enhanced ESP through high-permittivity QLD behavior: (C) Schematic microstructures and unipolar P-E loops (with identical internal electric fields represented by red, green, and blue loops) comparing different dielectrics including FEs with macroscopic domains and QLD materials featuring lattice distortions; the shaded areas represent electrostatic ES capacity; (D) Advanced transmission electron microscopy of QLD-type composition 0.88NaN_b0.9Ta_0.1O₃-0.10ST-0.02La(Mg_1/2Ti_1/2)O₃: [001]-oriented atomic-resolution HAADF-STEM images, polar states shown by individual arrows representing B-site cation displacement vectors in each unit cell, projected polarization angle, and polarization magnitude mappings. This figure is quoted with permission from Wang et al.^[325]. RFE: Relaxor ferroelectric; ESP: energy storage performance; HAADF-STEM: high-angle annular dark field- scanning transmission electron microscopy; ES: energy storage; QLD: quasi-linear dielectrics.

Similarly, Wang et al. developed a quasi-linear dielectrics (QLD) paradigm inspired by dipolar-glass dynamics but realized through compositional and structural modulation. In a ceramic system of 0.88NaN_b0.9Ta_0.1O₃-0.10ST-0.02La(Mg_1/2Ti_1/2)O₃, they achieved an ultrahigh W_rec approaching 43.5 J/cm^3[325]. This exceptional result was enabled by a controlled polar-evolution process: the system transformed from a FE state with long-range order to a RFE state via targeted compositional design, and subsequently into a QLD state through dopant-induced suppression of polar coupling down to only a few unit cells [Figure 25C]. Structural evidence confirmed this progressive transformation [Figure 25D], highlighting its close alignment with the defining features of dipolar glass systems.

It is noteworthy that although dipolar glasses exhibit remarkable ESP and superior stability, their intrinsically low dielectric permittivity and restricted field-induced polarization fundamentally limit their achievable energy density. Overcoming these constraints requires the realization of ultrahigh BDS. Consequently, the development of advanced ceramic processing routes and precise defect management strategies is imperative to fully unlock the potential of dipolar glass dielectrics for practical high-energy-density capacitor applications.

Lattice distortion engineering

An alternative lattice-scale strategy for enhancing the electrical properties of dielectric ceramics involves the intentional induction of lattice distortions through ionic doping. As noted in the context of domain- and nano-region-level design, doping-induced fluctuations in local electric and strain fields can disrupt long-range FE order. This effect is particularly significant in displacive FEs, where polar fluctuations are strongly coupled to structural distortions. Here, we focus on the broader and potentially beneficial impacts of doping-induced lattice distortion on dielectric ESP, extending beyond their role in modulating polar order.

Lattice distortion refers to deviations from the periodic atomic arrangement in a crystal, arising from intrinsic or extrinsic perturbations. These include ionic substitutions that induce lattice expansion or contraction, thermal or mechanical stresses that cause structural deformations, and spontaneous lattice reconstructions associated with FE phase transitions. The influence of such distortions on functional properties is multifaceted. First, lattice distortion enhances phonon scattering and reduces carrier mobility, thereby improving insulation, minimizing leakage losses, and boosting η. In addition, structural distortions can amplify ionic displacement polarization, reshape potential barriers for dipole reorientation, and create spatial inhomogeneities in local electric fields, each contributing complex effects to dielectric response and polarization behavior. Consequently, the overall impact of lattice distortion on FE polarization remains nontrivial, often requiring a delicate balance between competing mechanisms to optimize performance.

Lattice distortion offers an effective route to tune dielectric response and polarization. A striking example is provided by Zhang et al., who imposed substantial epitaxial strain on PbTiO₃ via fabrication of a PbTiO₃/PbO composite. Owing to the large lattice mismatch between the two phases, interfacial strain generated strong lattice distortion in PbTiO₃, yielding an enhanced tetragonality (c/a ≈ 1.238) and an extraordinarily high P_r of ~236.3 μC/cm², underscoring the promise of interphase strain engineering for amplifying polarization strength^[326]. A similar concept has been applied in dielectric ceramics. Li et al. engineered core-shell ST-BT nanoparticles through epitaxial growth, confining lattice distortion within interfacial regions. Compared with conventional Ba_0.6Sr_0.4TiO₃ solid solutions, these core-shell ceramics exhibit a broadened dielectric anomaly, enhanced polarization, and reduced nonlinearity, thereby demonstrating superior ESP^[327].

In high-entropy dielectric systems, pronounced lattice distortion is often intrinsic due to the coexistence of multiple cation species with different ionic radii and valence states. In KNN-based high-entropy RFEs, strong lattice distortion has been proposed to temporarily absorb part of the applied electric energy, resulting in delayed polarization saturation and improved ESP^[328].

Zhu et al. further demonstrated that local polarization configurations can be effectively tailored by controlling local lattice distortion in BNT-based high-entropy RFEs, as shown in Figure 26. Their approach yielded a remarkable ESP, attributed to the combined effects of suppressed domain growth and stabilized polar disorder^[329]. More recently, Xi et al. proposed a high-entropy quasi-linear dielectric system characterized by locally diverse lattice distortions, including the coexistence of multiphase PNRs within a pseudo-cubic matrix and multiple modes of BO₆ octahedral tilting^[330]. These heterogeneously distorted regions reduce polarization anisotropy, suppress hysteresis, delay saturation, and reinforce BDS through a synergy of intrinsic lattice hardening and extrinsic grain refinement.

Figure 26. Engineering local lattice distortion via a high-entropy strategy to enhance ESP. (A) Ionic radii and theoretical polar displacements of various A-site cations; (B) Chemical composition design of high-entropy Bi_0.5K_0.5TiO₃ (BKT)-based RFEs; (C and D) Atomic configurations showing A-site cation distributions and corresponding local lattice distortion fields from density functional theory (DFT) calculations; (E and F) Schematics of energy profiles and evolution of P-E loops illustrating improved ES behavior. This figure is quoted with permission from Zhu et al.^[329]. ESP: Energy storage performance; RFE: relaxor ferroelectric.

Beyond its direct influence on dielectric response and polarization behavior, lattice distortion also plays a critical indirect role in enhancing ESP and device reliability by mitigating electromechanical breakdown. In FE materials, a strong intrinsic coupling exists between polarization and strain under external electric fields, with multiple mechanisms contributing simultaneously to both responses^[26,331]. In practical dielectric devices, particularly MLCCs, such electromechanical coupling is often detrimental to performance^[332,333]. For instance, in the AFE material AgNbO₃ (AN), field-induced AFE-FE phase transitions produce large electrostrain. Zhu et al. proposed a heterovalent doping strategy to generate low-displacement regions within the AFE matrix, which alleviate spontaneous lattice distortion and reduce macroscopic strain during field-driven transitions, thereby enhancing both BDS and ESP^[156].

In RFEs, electrostrain is primarily governed by the electrostrictive effect. Li et al. observed that materials with higher dielectric permittivity and lower strain under electric fields generally exhibit smaller electrostrictive coefficients, indicating that weak polarization-strain coupling is advantageous for ES applications^[331]. Building on this principle, Zhang et al. introduced a “weak polarization-strain coupling” strategy^[155], which suppresses electrostrain in RFE-based dielectrics without compromising polarization. In (Bi_0.5Na_0.5)TiO₃-Pb(Mg_1/3Nb_2/3)O₃ (BNT-PMN) systems, the coexistence of A- and B-site cations with multiple valences and radii introduces both strong local electric fields and non-synergistic lattice distortions. This duality promotes large ionic-displacement polarization while simultaneously suppressing macroscopic lattice stretching. As a result, the material achieves exceptional polarization response and outstanding ESP with minimal strain output^[155]. This work demonstrates that lattice distortion engineering provides a cost-effective and broadly applicable route to suppress electrostrain in MLCCs, thereby improving performance, stability, and operational lifetime.

In summary, lattice distortion in FE ceramics acts as a double-edged regulator of electrical properties. While moderate distortion can optimize polarization dynamics and dielectric behavior, excessive distortion may trigger performance degradation or structural instability. Achieving precise control over lattice distortion remains a grand challenge, requiring advances in in situ characterization and multiscale theoretical modeling.

Thin-film deposition and nanostructuring

Thin-film deposition techniques, including atomic layer deposition and chemical vapor deposition, have enabled angstrom-level control over thickness, stoichiometry, and interfacial quality in dielectric materials. Such precision is essential for optimizing dielectric performance and ensuring seamless device-level integration. These methods also facilitate rational nanostructuring, allowing porosity, grain orientation, and crystallinity to be tailored at the nanoscale to enhance dielectric constants. Furthermore, the incorporation of nano-engineered architectures, including nanowires, nanotubes, and nanoparticle-assembled composites, has shown great potential to enhance both energy density and discharge efficiency in dielectric systems.

Additive manufacturing

Additive manufacturing offers a transformative route for fabricating complex dielectric structures with programmable control over geometry and hierarchical internal architecture. This technique enables the design of capacitors with engineered spatial distributions of dielectric phases, thereby enhancing ESP through improved polarization pathways and breakdown resistance. It also provides a versatile platform for producing multicomponent composites that integrate disparate dielectric chemistries across multiple length scales, creating new opportunities to overcome the long-standing trade-off between high polarization strength and high BDS.

Hybrid processing techniques

Hybrid processing approaches that integrate conventional and advanced synthesis routes are emerging as an effective strategy for tailoring dielectric performance. For instance, combining sol-gel synthesis with rapid sintering techniques such as spark plasma sintering can achieve near-theoretical ceramic densities and microstructural homogeneity, thereby yielding superior dielectric properties. Embedding such hybrid strategies into scalable manufacturing workflows offers a viable pathway to produce high-performance dielectric materials with enhanced precision, reproducibility, and cost efficiency, accelerating their transition toward advanced energy storage applications.

Microstructure-property relationships

Phase-field simulations have proven highly effective in capturing the physics of FE and RFE systems, where domain evolution and polarization dynamics dictate energy storage characteristics. By modeling domain nucleation, wall motion, and switching under applied electric fields, these simulations can predict dielectric permittivity, hysteresis behavior, and the impact of defects and grain boundaries on energy-storage efficiency. When combined with advanced TEM-based characterization and macroscopic electrical testing, such approaches enable a closed feedback loop in which microstructural features are quantitatively correlated with performance, thereby refining simulation accuracy and predictive capability. This integration provides a unified view of how composition, defect chemistry, and microstructure converge to govern dielectric behavior, laying the foundation for rational design of materials with targeted ESP.

Design of multiscale materials

A unique strength of phase-field simulations lies in their ability to bridge length scales, spanning from atomic distortions to device-level behavior. This capability enables the rational design of materials with optimized grain sizes, engineered interfaces, and tunable domain architectures, ensuring both performance and reliability under realistic operating conditions. For instance, in MLCCs, simulations can capture the electromechanical interplay of individual dielectric layers and predict how interfacial strain or charge accumulation affects device breakdown. Similarly, in RFEs, they can guide the tuning of PNR size, density, and spatial distribution to maximize recoverable energy storage. Such multiscale insights not only clarify the origins of observed phenomena but also provide quantitative guidelines for the fabrication of next-generation capacitors with unprecedented energy-storage capabilities.

Future prospects for simulations and characterizations

With rapidly expanding computational power, phase-field models are becoming increasingly predictive and capable of resolving intricate electromechanical interactions across multiple scales. Advanced TEM techniques provide the essential foundation for supplying realistic model parameters, while experimental validation remains indispensable for ensuring reliability. The interdependence between simulations and characterizations is thus intensifying, and the two are poised to evolve into inseparable counterparts, much like complementary perspectives on the same physical reality. Looking ahead, the integration of phase-field simulations with machine learning algorithms and high-throughput computational screening offers a transformative pathway for the rapid identification of dielectric materials with superior ESP. Such integration promises to accelerate materials discovery, optimization, and device-level translation, thereby driving the emergence of next-generation high-performance capacitors.

The need for consistent testing protocols

Variations in testing procedures frequently lead to substantial discrepancies in reported ESP parameters. For instance, dielectric BDS can vary dramatically with changes in voltage ramp rate, sample geometry, or testing temperature. Likewise, measured energy density is strongly influenced by factors such as electrode composition, sample thickness, and the frequency of the applied field. Studies have further demonstrated that geometric scaling of samples and electrodes can markedly affect polarization behavior and attainable ESP^[337]. Establishing standardized protocols with clearly defined conditions and metrics is therefore essential. Such harmonization would enable the development of comprehensive, reliable performance databases, facilitating cross-study comparisons and guiding the rational selection of optimal dielectric materials for energy-storage applications.

Developing comprehensive databases

The creation of comprehensive, publicly accessible databases is central to advancing standardized evaluation of dielectric materials. Such repositories should couple quantitative performance metrics with detailed metadata on fabrication routes, microstructural characteristics, and processing conditions. Integrated datasets of this nature would allow researchers to discern systematic performance trends, identify composition-structure-property correlations, and accelerate knowledge exchange. Beyond cataloguing performance, these databases would serve as critical platforms for machine-learning-driven predictions and data-informed materials discovery. Sustained progress in dielectric ES will therefore depend not only on innovations in fabrication and computational design but also on the widespread adoption of standardized protocols and open-access data infrastructures. Addressing these needs will enable the rapid discovery, optimization, and translation of next-generation dielectric materials capable of meeting the stringent demands of modern ES systems.