Retaining field-dependent properties of NBT-6BT ceramics sintered under reducing atmospheres via defect-level engineering

Abstract

The development of lead-free multilayer ceramic actuators (MLCAs) compatible with cost-effective Cu base-metal electrodes (BMEs) remains challenging due to the degradation of piezoceramic properties under the low oxygen partial pressure (P_O2), required for co-sintering. In particular, Na_0.5Bi_0.5TiO₃-based systems suffer from severe conductivity and loss of functional response under reducing conditions. In this work, donor-doped Na_0.5Bi_0.5TiO₃–6BaTiO₃ (NBT–6BT) ceramics are designed to retain their electromechanical performance under such atmospheres. Nb and Ta doping effectively suppress defect-related conductivity, stabilize the defect chemistry, and preserve the ferroelectric response. As a result, Nb- and Ta-modified NBT–6BT exhibits a large strain of up to 0.36% even after sintering in reducing conditions, whereas undoped NBT–6BT fails completely under identical processing. These results highlight an effective strategy for defect engineering in NBT-based ceramics and demonstrate a viable pathway toward high-performance, Cu-compatible, lead-free MLCAs.

---

1. Introduction

In recent years, considerable effort has been devoted to advancing high-power electronic systems. In particular, piezoceramic materials have become indispensable in applications that demand precise positioning and motion control. Some of the many applications of piezoceramics include microscopes,¹ cameras,² fuel injection systems,³ and ultrasound transducers,⁴ where volumetric efficiency, high-temperature, and high-power durability are of great importance.^1–3 Among the potential lead-free materials for such applications, NBT stands out for its higher Curie temperature (∼325 °C) and moderate piezoelectric constant (73 pC N⁻¹).^4,5 However, in its pristine form, it possesses a higher tan δ, higher leakage current, and therefore higher conductivity and cannot be directly used for such applications. The higher conductivity has been attributed to oxygen vacancy formation, which may result from nonstoichiometric conditions at the A site during processing and sintering.^6–10 However, solid solutions of NBT have shown potential for applications requiring high strain. For instance, in earlier studies, Takenaka et al. reported that introducing BT to NBT can promote relaxor behavior in NBT due to the mixed occupancy on the A-site (Na⁺, Bi³⁺, and Ba²⁺). They also reported that a 6% to 7% addition of BT to NBT can generate a morphotropic phase boundary (MPB), thereby enhancing the d₃₃ from 73 pC N⁻¹ to 125 pC N⁻¹.¹¹ Later studies showed even higher d₃₃ values (180 pC N⁻¹) can be obtained for NBT-6BT.¹² Motivated by the exceptional properties of NBT-based solid solutions, researchers and industry professionals have devoted their efforts to developing MLCAs with PMEs.^13–18 Achieving the desired characteristics and performance of NBT-based MLCAs requires precise control over several key parameters, including defect chemistry, particle size, tape casting formulation, electrode paste rheology, and the processes of debinding and sintering.¹⁹ As noted above, the majority of reported studies have focused on MLCAs employing PMEs. To the best of our knowledge, only a single study has demonstrated the co-sintering of a 0.88Bi_0.5Na_0.5TiO₃–0.08Bi_0.5K_0.5TiO₃–0.04BaTiO₃ ceramic pellet sandwiched between two Cu electrodes.²⁰ In contrast, a study by Chen et al. investigated the electromechanical properties of NBT–7%BaTiO₃ solid solutions sintered in a reducing N₂ atmosphere; however, the functional properties were found to be severely degraded under these conditions.²¹ This further highlights the pronounced sensitivity of NBT to defect formation during sintering under low oxygen partial pressures. In our previous study, we demonstrated that sintering NBT at oxygen partial pressures below 10⁻⁹ bar results in phase instability, increased electrical conductivity, and ultimately renders the material unsuitable for piezoelectric applications. In contrast, to prevent oxidation of copper electrodes during co-firing, the oxygen partial pressure must be maintained below ∼10⁻⁷ bar. This mismatch indicates that the stability window of undoped NBT is insufficient for reliable MLCA fabrication and NBT has to show stability at even lower partial pressures.²² Generally, NBT-based materials exhibit unusual conduction behavior when the oxygen vacancy concentration is increased.^23,24 Even for small variations, the concentration of the material can transition from low-conducting electronic transport to high oxygen conductivity, changing the total conductivity by up to 5 orders of magnitude. This has been found to correlate with the binding energy of acceptor-oxygen vacancy defect complexes.²⁵ This also affects the room temperature ferroelectric behavior. High leakage and fast electro-degradation can be observed in these cases.^26,27 Even in regularly sintered NBT-materials, which lost bismuth oxide during the sintering process, resulting in higher oxygen vacancy concentration, this behavior could be observed. However, the incorporation of donor dopants such as Nb and Ta effectively mitigated this issue, preserving the phase stability and functional properties of NBT. Remarkably, the doped NBT ceramics maintained their performance even when sintered at oxygen partial pressures as low as 10⁻¹² bar.²²

Whether the same stabilization strategy can be applied to the NBT–6BT system remains unclear. To elucidate this, it is essential to understand the mechanisms by which donor dopants such as Nb and Ta influence the structural and electrical stability of NBT–6BT during sintering under low oxygen partial pressure. Moreover, maintaining an appropriate oxygen partial pressure in the furnace atmosphere is critical to prevent the reduction of volatile or reducible species such as Bi₂O₃.

Therefore, this study aims to investigate whether pristine (undoped) NBT–BT ceramics retain their electrical properties after sintering at low oxygen partial pressure, compared with those sintered in air. Should these materials fail to maintain their functional stability, their defect chemistry will be tailored by introducing suitable dopants to enhance phase and electrical stability under reducing conditions. In-depth structural studies will be performed to identify defect species and their impact on the functional properties of NBT–6BT.

2. Experimental procedures

2.1. Synthesis

To synthesize 0.94[Na_0.5Bi_0.5Nb_xTi_1−xO₃]–0.06[BaNb_yTi_1−yO₃] compositions (0.005 ≤ x + y ≤ 0.02), a conventional solid-state reaction route was employed. Prior to weighing, the starting materials – Na₂CO₃ (99.50%), Bi₂O₃ (99.975%), BaCO₃ (99.8%), Nb₂O₅ (99.90%), and Ta₂O₅ (99.85%) – were pre-dried at 300 °C to remove moisture, whereas TiO₂ (99.60%) was heat-treated at 800 °C due to removal of residual organics or adsorbed species that cannot be eliminated at lower temperatures and stabilization of the powder by reducing surface hydroxyl groups.²⁸ The purity of all precursors was accounted for in batch calculations to ensure stoichiometric accuracy. The weighed powders (all supplied by Alfa Aesar GmbH, Karlsruhe, Germany) were mixed in ethanol and ball-milled in a nylon container using yttria-stabilized zirconia media in a planetary mill (Pulverisette 5, Fritsch, Germany) at 250 rpm for 24 h. The resulting slurry was dried overnight at 100 °C, and the dried cake was gently crushed, placed in an alumina crucible, and calcined at 900 °C for 3 h with a heating rate of 5 °C min⁻¹. The calcined powders were then manually ground and subjected to a second ball-milling cycle under identical conditions for 24 h to ensure homogeneity and fine particle size. The final powder was pressed into pellets of 10 mm diameter using a uniaxial press at 64 MPa for approximately 2 min. To further enhance the green density, the pellets were subsequently cold isostatically pressed (KIP 100E, Weber, Germany) at 380 MPa for 1.5 min. For sintering, the green pellets were embedded in sacrificial (master) powder to minimize Bi volatility and placed in alumina crucibles. Sintering was conducted at 1150 °C for 2 hours under two conditions: (i) ambient air and (ii) a low oxygen partial pressure environment (10⁻¹² to 10⁻⁸ bar).

2.2. Sintering atmosphere

A tube furnace (HTM Reetz GmbH, Germany) equipped with a turbo pump with a rotation speed of up to 90 000 rpm (D-35614 ASSLAR series, Pfeiffer Vacuum GmbH, Germany), oxygen sensor (Oxygen Analyzer SGM5, ZIROX GmbH, Germany), and two digital mass flow meters (red-y compact series, Vögtlin Instruments GmbH, Germany) with a maximum flow rate of 270 mln min⁻¹ was employed for this purpose. To control the desired partial pressure, a mixture of N₂ and a forming gas of 0.1 vol% H₂-99.9 vol% N₂ (supplied by Air Liquide) was used. To establish the precise partial pressure of P_O2 = 10⁻⁷ to P_O2 = 10⁻⁸ bar, 5 mln min⁻¹ of 0.1 vol% H₂-99.9 vol% N₂, 1 mln min⁻¹ wet N₂, and 270 mln min⁻¹ dry N₂ were mixed. To supply wet N₂, a double-gas bubbler filled with deionized water was used. To prevent any potential delamination of the stacks, the heat treatment commenced up to 600 °C at a heating rate of 0.5 °C min⁻¹, followed by a heating rate of 2.5 °C min⁻¹ up to 1020 °C. A dwell time of 1 hour was maintained at the sintering temperature, after which the furnace naturally cooled to RT.

2.3. Characterization

The relative density of the sintered samples was evaluated using the Archimedes method. Phase identification and crystallinity were analyzed through X-ray diffraction (XRD) with Cu-Kα radiation (λ = 1.5406 Å), conducted on a Bruker AXS D8 Advance diffractometer (Karlsruhe, Germany). For detailed microstructural and compositional analysis, a field-emission scanning electron microscope (FE-SEM, JEOL JSM-7600F, Japan) equipped with an energy-dispersive X-ray spectroscopy (EDS) system (X-Max80, Oxford Instruments, UK) was utilized. To prepare the samples for imaging, sequential grinding was carried out using SiC papers with mesh sizes of 800, 1200, 2500, and 4000, followed by polishing with diamond pastes of 15 µm, 6 µm, 3 µm, and 0.25 µm. Finally, thermal etching was performed at 1050 °C to reveal the grain structure. The grain size measurements were performed using ImageJ software (National Institutes of Health, USA). For transmission electron microscopy (TEM) and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), specimens were mechanically thinned to approximately 30 µm, followed by final thinning with an Ar⁺ ion beam system (GATAN 695, GATAN, Pleasanton, USA). Samples were mounted on copper rings with a 1.5 mm inner diameter prior to ion milling. Bright-field TEM and HAADF-STEM imaging, along with EDS elemental mapping, were performed using a TALOS F200X microscope (FEI, now Thermo Fisher Scientific) at 200 kV, equipped with a Super-X EDS detector system.

Impedance spectroscopy measurements were conducted using a Novocontrol Alpha-A High-Performance Frequency Analyzer over a frequency range of 0.1 Hz to 1 MHz, with an AC signal amplitude of 0.1 V. Temperature-dependent measurements were performed from 125 °C to 600 °C, and the data were analyzed using RelaxIS software (rhd instruments, Germany). Poling of ceramic pellets was executed in a silicone oil bath at 25 °C under a DC electric field of up to 5 kV mm⁻¹ for 20 minutes. To examine the field-dependent polarizability of the samples, a triangular electric field was applied using a modified Sawyer–Tower circuit combined with an optical displacement sensor (D63, Philtec Inc., USA), reaching fields of up to 6 kV at 1 Hz. The d₃₃ was measured 24 hours post-poling using a Berlincourt-type d₃₃-meter (PiezoMeter System PM300, Piezotest Pte Ltd., Singapore) operating at 110 Hz. The effective piezoelectric coefficient was further derived from the unipolar strain–electric field (S–E) loops.

3. Results and discussion

The XRD patterns of undoped NBT-6BT and NBT-6BT doped with 0%, 0.5%, 1.0%, and 2.0% mol Nb and Ta, sintered at 1150 °C in both P_O2 = air and P_O2 = 10⁻¹² bar is illustrated in Fig. S1(a–d), respectively. All the examined samples exhibit a pseudo-cubic structure, in line with previous studies on NBT materials.^5,29 A detailed examination of the patterns did not reveal any detectable secondary phases within the XRD device's sensitivity limits. Additionally, peak splitting at around 40.0° and 46.5° is obvious for all the patterns, indicative of the presence of both rhombohedral and tetragonal structure at the MPB (shown in Fig. 1).^30,31 However, increasing the Nb and Ta content reduces the peak shoulder at rhombohedral and tetragonal regions, indicating an increase in symmetry and the transition of the material from a complex, non-ergodic ferroelectric/relaxor state to a more structurally homogeneous, ergodic relaxor (or pseudocubic) state.^32–34

Fig. 1 XRD patterns in the 38° to 50° angle range for NBT-6BT samples doped with 0%, 0.5%, 1.0%, and 2.0% mol Nb and Ta, sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

Furthermore, substitution of Nb⁵⁺ for Ti⁴⁺ slightly shifts all diffraction peaks toward lower 2θ values (i.e., to the left). This shift occurs because the incorporation of larger Nb⁵⁺ ions (r = 0.640 Å)³⁵ compared to Ti⁴⁺ (r = 0.605 Å)³⁵ at octahedral coordination expands the lattice, thereby increasing the interplanar spacing d. Since the X-ray wavelength (λ) remains constant, Bragg's law requires a smaller diffraction angle (θ) to satisfy the equation nλ = 2d sin θ. Consequently, the 2θ values decrease, resulting in a leftward shift of the peaks. This indicates successful incorporation of the dopants. Additionally, the sintering atmosphere influences the 2θ positions (see Fig. S2a and b), causing a slight shift toward lower angles that indicates minor lattice expansion. This behavior is likely associated with an increased concentration of oxygen vacancies under low oxygen partial pressure, as well as local lattice distortions induced by defect dipoles. Nevertheless, the magnitude of the shift is minimal, suggesting that no phase transition occurs and the perovskite structure remains stable.

Fig. 2 shows the SEM images of undoped NBT-6BT and NBT-6BT doped with 0.5, 1.0, and 2.0 mol% Nb and Ta, sintered at 1150 °C under both ambient air and a reducing atmosphere (P_O2 = 10⁻¹² bar). The air-sintered NBT-6BT exhibits an average grain size of 1.37 µm (see Fig. S3). In contrast, sintering undoped NBT-6BT under reducing conditions results in a grain size of 8.67 µm (see Fig. S3) – over six times larger than that of the air-sintered sample. Some studies claim that this pronounced grain growth is likely due to the generation of a higher concentration of oxygen vacancies in the reducing atmosphere, which enhances grain boundary mobility.^36–38 Our previous TEM analysis for NBT revealed that the grain boundaries are devoid of Bi ions under a reducing atmosphere.²² Typically, the segregation of Bi ions at the grain boundary would induce a solute drag effect, resulting in a finer grain structure; however, the lack of such refinement here is consistent with the absence of Bi.

Fig. 2 SEM images of 0%, 0.5%, 1.0%, and 2.0% mol Nb and Ta doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

Incorporating a small amount of Nb or Ta (0.5 mol%) markedly refines the microstructure, yielding submicron grains (∼0.75 µm), i.e., more than eleven times smaller than the undoped sample sintered at P_O2 = 10⁻¹² bar. Similar trends were observed at higher dopant concentrations, where sintering under reducing conditions did not promote abnormal grain growth. However, in our previous work, we found that sintering Nb/Ta-doped NBT under a reducing atmosphere still led to grain coarsening. To reconcile this apparent discrepancy, a closer examination of the grain boundary chemistry in NBT-6BT is necessary. Fig. 3 presents the STEM images of 2.0% Nb-doped NBT-6BT sintered in P_O2 = air. Careful analysis of the images reveals that, similar to NBT from our previous study,²² the grain boundaries in 2.0% Nb-doped NBT-6BT are enriched with Bi, but deficient in Na, Ti, and O. No clear, distinguishable differences in Ba and Nb concentrations were observed between the bulk and the grain boundaries. The charge compensation mechanism appears consistent with NBT, where Nb is compensated by the formation of Bi vacancies and Bi segregation to the grain boundaries (eqn (1)). Such a segregation leads to finer grains due to the solute drag effect.

Conversely, a closer examination of the grain boundaries in the 2.0 mol% Nb-doped NBT-6BT sample sintered under P_O2 = 10⁻¹² bar (shown in Fig. S4) reveals enrichment of at least Ti, and possibly Ba, at the grain boundaries. In contrast, the boundaries appear to be free from Bi, Na, and O. Under these conditions, the charge compensation mechanism becomes more straightforward, primarily involving the formation of cation vacancies on Ti⁴⁺ site, as illustrated by eqn (2).The segregation of Ti from the grain interior toward the grain boundary could hinder grain boundary mobility, thereby resulting in a finer microstructure.

Fig. 3 STEM mapping of Na, Bi, Ba, Ti, O, and Nb elements of 2.0 mol % Nb-NBT-6BT, which was sintered at P_O2 = air.

To further analyze the impact of various dopants on the temperature-dependent conductivity of NBT-6BT, Arrhenius-type plots of the bulk conductivity were generated as a function of the reciprocal of temperature and are illustrated in Fig. 4(a–e). The Arrhenius plots for undoped NBT-6BT, analyzed under both P_O2 = air and P_O2 = 10⁻¹² bar, show a dramatic increase in conductivity – by approximately three orders of magnitude – when sintered in a lower oxygen environment (Fig. 4a).

Fig. 4 The Arrhenius plots of 0%, 0.5%, 1.0%, and 2.0% mol Nb and Ta doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

This pronounced increase in conductivity is accompanied by a notable change in the Nyquist response: while the air-sintered NBT-6BT exhibits a single semicircle, samples sintered at P_O2 = 10⁻¹² bar display two distinct semicircles (see Fig. S5). According to the brick-layer model, the high-frequency semicircle corresponds to the bulk response, whereas the low-frequency one is attributed to grain boundary conductivity.³⁹ It should be noted that all impedance measurements were carried out in air. Prolonged or repeated measurements at elevated temperatures could enable oxygen from the atmosphere to diffuse into the samples, thereby partially reoxidizing them and reducing the concentration of oxygen vacancies created during reducing sintering. In this study, however, only a single measurement cycle was performed, and no significant decrease in conductivity or deviation from typical NBT-related behavior was observed. This result aligns with previous observations for NBT-6BT containing higher levels of acceptor dopants such as Mg²⁺, where the formation of oxygen vacancies enhances ionic conductivity⁴⁰ by multiple orders of magnitude in a non-linear fashion, as well as with our earlier findings on undoped NBT sintered under reducing conditions,²² which illustrated partial phase instability of NBT in these conditions. It can be rationalized that a similar mechanism is at play during the sintering of NBT-6BT in a low-oxygen partial-pressure environment. The reduced-oxygen atmosphere during sintering likely promotes the formation of a higher concentration of oxygen vacancies, thereby increasing ionic conductivity. This mechanism explains the elevated ionic conductivity observed when NBT-6BT is sintered in P_O2 = 10⁻¹² bar. The activation energies, derived from the Arrhenius plots, are summarized in Table 1. The oxygen-conducting sample, sintered at P_O2 = 10⁻¹² bar, does not exhibit a constant activation energy. Two regions can be separated, in which the high temperature region has an activation energy of 0.44, and in the low temperature region, it is 0.77 eV. This has been rationalized with the help of DFT calculations and attributed to different activation energies for oxygen transport in rhombohedral and tetragonal structures.^24,26 Increasing the Nb content to 1.0 mol% in air-sintered samples results in a decrease in conductivity compared to the 0.5 mol% Nb-doped NBT-6BT (Fig. 4a). However, increasing the Nb content to 2.0 mol% does not significantly affect conductivity beyond the level observed at 1.0 mol%. Notably, for NBT-6BT doped with 0.5, 1.0, and 2.0 mol% Nb and sintered in P_O2 = 10⁻¹² bar, the conductivity shows only a slight increase compared to its air-sintered (still more than 3 orders of magnitude smaller than undoped NBT-6BT sintered under reducing atmosphere) counterparts (Fig. 4b). A similar trend is observed in NBT-6BT doped with Ta. For 0.5, 1.0, and 2.0 mol% Ta-doped NBT-6BT, the conductivity is only slightly affected during sintering in P_O2 = 10⁻¹² bar, as illustrated in Fig. 4c and d. However, Ta doping exhibits an inconsistent trend in activation energy with increasing doping (Table 1). This reveals further complex behavior. Even though the drastic change in conductivity with increasing oxygen vacancy concentration can already be rationalized mechanistically, the conductivity mechanism of both pure NBT and donor-doped NBT remains only partially elucidated. It has recently been shown that, rather than the previously assumed intrinsic electronic conductivity, polarons are responsible for the transport.⁴¹ The energy band gap for charge transport does not coincide with the optical band gap. The observed changes in activation energy, especially for the Ta-doped material, can thus be attributed to variations in polaronic behavior. However, the origin and nature of the mid-gap states could, nevertheless, not be revealed so far. It remains a relevant topic of ongoing research. Interestingly, there are differences when comparing NBT and NBT-BT. Pure NBT is much more sensitive to increases in oxygen vacancy concentration, whereas NBT-BT shows greater variation in activation energy for the low-conductivity samples.²⁶

Table 1 Activation energies of 0%, 0.5%, 1.0%, and 2.0% mol Nb- and Ta-doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar

Sample	Sintering PO2	Activation energy (eV)	Sample	Sintering PO2 (bar)	Activation energy (eV)
NBT-6BT	Air	0.90	NBT-6BT	10^−12	0.44–0.77
0.5% Nb-NBT-6BT	Air	1.29	0.5% Nb-NBT-6BT	10^−12	1.43
1.0% Nb-NBT-6BT	Air	1.55	1.0% Nb-NBT-6BT	10^−12	1.30
2.0% Nb-NBT-6BT	Air	1.59	2.0% Nb-NBT-6BT	10^−12	1.44
0.5% Ta-NBT-6BT	Air	1.29	0.5% Ta-NBT-6BT	10^−12	1.32
1.0% Ta-NBT-6BT	Air	1.54	1.0% Ta-NBT-6BT	10^−12	1.32
2.0% Ta-NBT-6BT	Air	1.59	2.0% Ta-NBT-6BT	10^−12	1.36

---

Fig. 5(a–d) and Fig. S6(a–d) display the polarization vs. electric field (P–E) loops and strain vs. electric field (S–E) plots of undoped NBT-6BT, alongside 0.5 mol%, 1.0 mol%, and 2.0 mol% Nb- and Ta-doped NBT-6BT, sintered in both P_O2 = air and P_O2 = 10⁻¹² bar, respectively. The undoped, air-sintered NBT-6BT, shown in Fig. 5a, exhibits a typical P–E loop characteristic of NBT-6BT, consistent with observations reported in other studies.^42,43

Fig. 5 Field-dependent polarization and strain behavior of 0%, 0.5%, 1.0%, and 2.0% mol Nb-doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

Conversely, sintering undoped NBT-6BT in P_O2 = 10⁻¹² bar significantly increases the leakage current, resulting in egg-shaped P–E loops (Fig. 5c), higher remnant polarization (P_r), and reduced maximum strain (S_max). Additionally, the material experiences a breakdown at 5 kV cm⁻¹ due to electrodegradation by oxygen vacancy migration, leading to a short circuit of the sample.²⁷ This behavior aligns with the findings of Chen et al., who reported a similar suppression of piezoelectric properties in NBT-7% BT when sintered in a N₂ atmosphere.²¹ In another study, Bi_0.47Na_0.47Ba_0.06Cu_0.015Ti_0.985O₃ was investigated under O₂, air, and N₂ atmospheres, revealing a pronounced degradation of electrical properties when sintered in an N₂ atmosphere. Specifically, d₃₃ decreased from 227 pC N⁻¹ in O₂ to 155 pC N⁻¹ in N₂.⁴⁴ Investigating the P–E loops of air-sintered Nb-doped samples (shown in Fig. 5a) demonstrates that increasing the Nb concentration to 1.0 mol% results in slightly slimmer P–E loops, indicative of softening behavior and excellent stability. Further increasing Nb to 2.0 mol% significantly alters the P–E loop to a pinched shape, where polarization decreases slightly but the coercive field decreases substantially. A similar behavior is observed with Ta-doped NBT-6BT (see Fig. S6a). This phenomenon is likely due to the collapse of the ferroelectric long-range order, induced in the non-ergodic state, transitioning into the ergodic state. This transition is typically observed in undoped NBT-6BT near or above the depolarization temperature (T_d).^11,45 Additionally. it indicates the field-induced ferroelectric phase becomes increasingly difficult to activate due to the higher entropy associated with these systems. This behavior closely resembles that of an antiferroelectric material. Furthermore, our results on Nb-doped system demonstrate that Nb significantly contributes to the suppression of the ferroelectric phase. In essence, Nb-doped NBT-BT exhibits characteristics similar to those of NBT-ST compositions, further reinforcing the inhibition of formation of ferroelectric phase.⁴⁶ Additionally, the S–E response of air-sintered Nb-doped samples is depicted in Fig. 5b. The plots indicate that adding Nb up to 1.0 mol% slightly shifts the S–E curves toward positive strain. However, increasing the Nb concentration to 2.0 mol% markedly alters the S–E behavior from a butterfly shape to a sprout shape with a total S_max of 0.36%. This change is likely due to a reversible electric field-induced phase transition, a phenomenon commonly observed in ergodic relaxors. In this state, the ferroelectric long-range order is established by a sufficiently high external electric field, resulting in characteristic sprout-shaped S–E plots.⁴⁷ Almost similar properties were attained for Ta-doped samples shown in Fig. S6b.

When undoped NBT-6BT is sintered at P_O2 = 10⁻¹² bar, S_max exhibits a pronounced deterioration (Fig. 5d). However, the introduction of a small amount of Nb (0.5 mol%) substantially restores the strain to values comparable to those of samples sintered in air, and a further increase to 1.0 mol% Nb leads to an additional enhancement. For 2 mol% Nb, both polarization and strain responses remain comparable to the air-sintered state.

It is important to note that these changes are not related to increased electrical conductivity. On the contrary, donor doping with Nb (and similarly Ta) is known to reduce conductivity compared to undoped NBT.^22,48 The observed changes in both P–E and S–E responses are instead intrinsic to the material's electromechanical behavior. Specifically, the S–E loops evolve from the typical butterfly shape toward a sprout-like shape, while the P–E loops become pinched. Too high doping concentrations disturb the long-range rhombohedral ferroelectric state and promote an ergodic relaxor state with polar nanoregions.⁴⁹ The strain behavior of Ta-doped samples (up to 1.0 mol%) under both air and P_O2 = 10⁻¹² bar conditions follows a similar trend to Nb-doped compositions, as shown in Fig. S6b and d. However, at 2.0 mol% Ta, a slight reduction in the S–E response is observed under reducing conditions. These results further emphasize the critical role of defect chemistry in tailoring the functional properties of NBT-6BT, particularly under low oxygen partial pressure.

The results of d₃₃ measurements for Nb- and Ta-doped NBT-6BT in both P_O2 = air and P_O2 = 10⁻¹² bar is presented in Fig. 6a and b. The undoped, air-sintered NBT-6BT exhibited a d₃₃ value of approximately 150 pC N⁻¹. However, as previously discussed, poling undoped NBT-6BT sintered in P_O2 = 10⁻¹² bar resulted in a breakdown due to higher leakage currents. Therefore, no value could be assigned for the respective sample.

Fig. 6 The d₃₃ results of 0%, 0.5%, 1.0%, and 2.0% mol Nb- and Ta-doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

In Nb- and Ta-doped air-sintered samples, increasing the donor dopant concentration up to 1.0 mol% raised the d₃₃ to 180 pC N⁻¹. Table S1 summarizes a comparison of the d₃₃ values of pristine and doped NBT-6BT ceramics reported in the literature with the results obtained in the present work.^44,50–58 This shows that the d₃₃ values obtained in this work are generally at the higher end of reported results. However, all presented values fall more or less in a quite narrow region. Given that multiple examples of doping are illustrated, it can thus be highlighted again that significant hardening or softening of ferroelectric properties by doping is not possible in NBT-based material. The difference between lead-based ceramics and NBT material is far too different. However, when the dopant concentration was increased to 2.0 mol%, the d₃₃ value became negligible. This reduction is attributed to the collapse of the ferroelectric long-range order and the transition from a non-ergodic state to an ergodic state, similar to relaxor ferroelectrics, where no significant d₃₃ can be recorded. The samples sintered in P_O2 = 10⁻¹² bar showed similar d₃₃ values, with only a slight reduction compared to their air-sintered counterparts.

Additionally, Fig. 7a and b illustrate the values for both Nb- and Ta-doped NBT-6BT in P_O2 = air and P_O2 = 10⁻¹² bar. The undoped, air-sintered sample exhibited a value of approximately 258 pm V⁻¹, while the sample sintered in P_O2 = 10⁻¹² bar experienced breakdown and no value could be recorded. The introduction of Nb and Ta to air-sintered samples resulted in a continuous increase in values, reaching 630 pm V⁻¹ for 2 mol % Nb-doped and 606 pm V⁻¹ for 2 mol% Ta-doped NBT-6BT, respectively. The samples sintered in P_O2 = 10⁻¹² bar showed similar values, with only a slight reduction compared to their air-sintered counterparts. However, it should be noted that these values are derived from S–E measurements and are based on a limited number of data points. Therefore, while the results indicate a clear improvement with donor doping, the evolution of with composition should be interpreted as a general trend rather than a strictly quantitative progression, particularly for the 2 mol% compositions.

Fig. 7 The results of 0%, 0.5%, 1.0%, and 2.0% mol Nb- and Ta-doped NBT-6BT sintered in P_O2 = air and P_O2 = 10⁻¹² bar.

4. Conclusion

This study presented findings on the behavior of Nb- and Ta-doped NBT-6BT, highlighting the effects of sintering atmospheres on structure, microstructure, and piezoelectric properties. STEM images revealed distinct segregation patterns: in Nb-doped NBT-6BT sintered in air, Bi³⁺ segregates to the grain boundaries, while in a reducing atmosphere P_O2 = 10⁻¹² bar, Ti⁴⁺ segregates. In terms of only Bi, this behavior mirrors that of NBT, where Bi³⁺ is prevalent at grain boundaries in air but deficient under reducing conditions. Both sintering atmospheres showed that introducing 2.0 mol% Nb⁵⁺ and Ta⁵⁺ donor dopants lead to the collapse of ferroelectric long-range order, transitioning the material from a non-ergodic to an ergodic phase. This phase transition was corroborated by pinched-shaped P–E loops (AFE-like behaviour) and sprout-shaped S–E plots. The addition of 2.0 mol% Nb or Ta, irrespective of the sintering atmosphere, rendered NBT-6BT nonresponsive to direct d₃₃ measurements, indicating a significant alteration in its piezoelectric properties. However, these doped samples exhibited a high field-dependent response, achieving values of 630 pm V⁻¹ for 2 mol% Nb-NBT-6BT and 606 pm V⁻¹ for 2 mol% Ta-NBT-6BT, respectively. These findings confirm that Nb and Ta can be effectively utilized as donor dopants in NBT-6BT, significantly impacting its structural and piezoelectric properties depending on the sintering atmosphere. Concerning a possible use in MLCA with Cu-inner electrodes, a dopant concentration of about 1 mol% of either Ta or Nb would be optimal. In this case, the ferroelectric properties of NBT-6BT are retained even when the ceramic is sintered in heavily reducing atmosphere.

Conflicts of interest

Hamed Salimkhani and Till Frömling have filed a patent application under number DE 10 2025 121 761.0.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d6ma00198j.

Acknowledgements

Hamed Salimkhani and Till Frömling would like to thank the German Ministry of Education and Research (BMBF) for funding the Young Investigator Group HTL-NBT within the program “NanoMatFutur” [Grant No. 03XP0146]. The authors are deeply grateful to the Nichtmetallisch-Anorganische Werkstoffe (NAW) group led by Prof. Jürgen Rödel for providing us with the opportunity to conduct experiments in their laboratories. We sincerely appreciate their support and generosity. The authors also acknowledge the support of the Alexander von Humboldt Foundation through the Georg Forster Research Fellowship, which funded Ahmad Sayyadi-Shahraki's research stay at TU Darmstadt, Germany. Additionally, the authors appreciate Phadcalc (https://www.phadcalc.com) for their TEM service.

References

1. K. Uchino and S. Takahashi, Multilayer ceramic actuators, Curr. Opin. Solid State Mater. Sci., 1996, 1(5), 698–705.
2. S. Takahashi, Multilayer piezoelectric ceramic actuators and their applications, Jpn. J. Appl. Phys., 1985, 24(S2), 41.
3. X. Gong and Z. Suo, Reliability of ceramic multilayer actuators: a nonlinear finite element simulation, J. Mech. Phys. Solids, 1996, 44(5), 751–769.
4. D. Damjanovic, et al., What can be expected from lead-free piezoelectric materials? Functional, Mater. Lett., 2010, 3(01), 5–13.
5. K. Reichmann, A. Feteira and M. Li, Bismuth sodium titanate based materials for piezoelectric actuators, Materials, 2015, 8(12), 8467–8495.
6. F. Yang, et al., Defect chemistry and electrical properties of sodium bismuth titanate perovskite, J. Mater. Chem. A, 2018, 6(13), 5243–5254.
7. Y. Sung, et al., Effects of Bi nonstoichiometry in (Bi_0.5+xNa)TiO₃ ceramics, Appl. Phys. Lett., 2011, 98(1), 012902.
8. J. Carter, et al., Structure and ferroelectricity of nonstoichiometric (Na_0.5Bi_0.5)TiO₃, Appl. Phys. Lett., 2014, 104(11), 112904.
9. I.-T. Seo, S. Steiner and T. Frömling, The effect of A site non-stoichiometry on 0.94(Na_yBi_x)TiO_3-0.06BaTiO₃, J. Eur. Ceram. Soc., 2017, 37(4), 1429–1436.
10. S. A. Khan, et al., Achieving high electromechanical response in lead-free BNT-BT ceramics through synergistic A/B-site doping, Appl. Phys. Lett., 2024, 124(25), 252904.
11. T. Takenaka, K.-i Maruyama and K. S. K. Sakata, (Bi_1/2Na_1/2)TiO₃-BaTiO₃ system for lead-free piezoelectric ceramics, Jpn. J. Appl. Phys., 1991, 30(9S), 2236.
12. Q. Xu, et al., Synthesis and piezoelectric and ferroelectric properties of (Na_0.5Bi_0.5)_1−xBa_xTiO₃ ceramics, Mater. Chem. Phys., 2005, 90(1), 111–115.
13. M. Gehringer, et al., Prototyping Na_0.5Bi_0.5TiO₃-based multilayer ceramic capacitors for high-temperature and power electronics, J. Eur. Ceram. Soc., 2023, 43(14), 6122–6129.
14. P. Ren, et al., Super-stable permittivity and low dielectric loss of (1 − x)Na_0.5Bi_0.5+yTiO_3−xNaTaO₃ ceramics within an ultra-wide temperature range, J. Materiomics, 2023, 9(3), 625–634.
15. P. Fan, et al., Progress and perspective of high strain NBT-based lead-free piezoceramics and multilayer actuators, J. Materiomics, 2021, 7(3), 508–544.
16. K. Liu, et al., Design and development of outstanding strain properties in NBT-based lead-free piezoelectric multilayer actuators by grain-orientation engineering, Acta Mater., 2023, 246, 118696.
17. P. Fan, et al., Low-temperature sintered (Na_1/2Bi_1/2)TiO₃-based incipient piezoceramics for co-fired multilayer actuator application, J. Materiomics, 2019, 5(3), 480–488.
18. H. Zhang, et al., (Na_1/2Bi_1/2)TiO₃-based lead-free co-fired multilayer actuators with large strain and high fatigue resistance, J. Am. Ceram. Soc., 2019, 102(10), 6147–6155.
19. C. Randall, et al., High strain piezoelectric multilayer actuators—a material science and engineering challenge, J. Electroceram., 2005, 14, 177–191.
20. G. Yesner and A. Safari, Development of a lead-free copper co-fired BNT-based piezoceramic sintered at low temperature, J. Am. Ceram. Soc., 2018, 101(12), 5315–5322.
21. C.-S. Chen, et al., The effects of sintering atmosphere on microstructures and electrical properties of lead-free (Bi_0.5Na_0.5)TiO₃-based ceramics, Ceram. Int., 2014, 40(7), 9591–9598.
22. H. Salimkhani, et al., Overcoming the Sensitivity of Sodium Bismuth Titanate Towards Sintering in a Reducing Atmosphere by Defect Chemistry Engineering. Available at SSRN 5654872.
23. M. Li, et al., Dramatic influence of A-site nonstoichiometry on the electrical conductivity and conduction mechanisms in the perovskite oxide Na_0.5Bi_0.5TiO₃, Chem. Mater., 2015, 27(2), 629–634.
24. L. Koch, et al., Ionic conductivity of acceptor doped sodium bismuth titanate: influence of dopants, phase transitions and defect associates, J. Mater. Chem. C, 2017, 5(35), 8958–8965.
25. L. Koch, et al., Revealing the impact of acceptor dopant type on the electrical conductivity of sodium bismuth titanate, Acta Mater., 2022, 229, 117808.
26. S. Steiner, et al., The effect of Fe-acceptor doping on the electrical properties of Na_1/2Bi_1/2TiO₃ and 0.94 (Na_1/2Bi_1/2)TiO_3–0.06BaTiO₃, J. Am. Ceram. Soc., 2019, 102(9), 5295–5304.
27. P. Ren, et al., High field electroformation of sodium bismuth titanate and its solid solutions with barium titanate, J. Mater. Chem. C, 2021, 9(9), 3334–3342.
28. R. Mueller, et al., OH surface density of SiO₂ and TiO₂ by thermogravimetric analysis, Langmuir, 2003, 19(1), 160–165.
29. E. Aksel, et al., Phase transition sequence in sodium bismuth titanate observed using high-resolution X-ray diffraction, Appl. Phys. Lett., 2011, 99(22), 222901.
30. M. Chen, et al., Structure and electrical properties of (Na_0.5Bi_0.5)_1−xBa_xTiO₃ piezoelectric ceramics, J. Eur. Ceram. Soc., 2008, 28(4), 843–849.
31. B.-J. Chu, et al., Electrical properties of Na_1/2Bi_1/2TiO₃–BaTiO₃ ceramics, J. Eur. Ceram. Soc., 2002, 22(13), 2115–2121.
32. W. Jo, et al., On the phase identity and its thermal evolution of lead free (Bi_1/2Na_1/2)TiO3–6 mol% BaTiO₃, J. Appl. Phys., 2011, 110(7), 074106.
33. A. Glazounov, A. Tagantsev and A. Bell, Evidence for domain-type dynamics in the ergodic phase of the PbMg_1/3Nb_2/3O₃ relaxor ferroelectric, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 53(17), 11281.
34. H. S. Han, et al., Effect of Nb Doping on the Dielectric and Strain Properties of Lead–free 0.94 (Bi_1/2Na_1/2)TiO_3–0.06BaTiO₃ Ceramics, J. Korean Ceram. Soc., 2016, 53(2), 145–149.
35. R. D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Found. Crystallogr., 1976, 32(5), 751–767.
36. R. Zuo, et al., Influence of A-site nonstoichiometry on sintering, microstructure and electrical properties of (Bi_0.5Na_0.5)TiO₃ ceramics, Mater. Chem. Phys., 2008, 110(2–3), 311–315.
37. I.-T. Seo, S. Steiner and T. Frömling, The effect of A site non-stoichiometry on 0.94 (Na_yBi_x)TiO_3-0.06BaTiO₃, J. Eur. Ceram. Soc., 2017, 37(4), 1429–1436.
38. Y. Sung, et al., Effects of Bi nonstoichiometry in (Bi_0.5+xNa)TiO₃ ceramics, Appl. Phys. Lett., 2011, 98(1), 012902.
39. J. T. Irvine, D. C. Sinclair and A. R. West, Electroceramics: characterization by impedance spectroscopy, Adv. Mater., 1990, 2(3), 132–138.
40. S.-M. An and S.-J. L. Kang, Boundary structural transition and grain growth behavior in BaTiO₃ with Nd₂O₃ doping and oxygen partial pressure change, Acta Mater., 2011, 59(5), 1964–1973.
41. P. Hu, et al., How semiconducting are ferroelectrics: the fundamental, optical and transport gaps of Na_0.5Bi_0.5TiO₃–BaTiO₃ and NaNbO₃, Rep. Prog. Phys., 2026, 89(2), 028004.
42. T. Karthik and S. Asthana, Polarization extension mechanism revealed through dynamic ferroelectric hysteresis and electric field driven structural distortions in lead free Na_0.5Bi_0.5TiO₃ ceramic, J. Phys. D: Appl. Phys., 2017, 50(38), 385601.
43. W. Cao, et al., Defect dipole induced large recoverable strain and high energy-storage density in lead-free Na_0.5Bi_0.5TiO₃-based systems, Appl. Phys. Lett., 2016, 108(20), 202902.
44. J. HU, et al., Microstructure and Electrical Properties of Bi_0.47Na_0.47Ba_0.06Cu_0.015Ti_0.985O₃ Ceramics Sintered in Different Atmospheres, J. Ceram., 2023, 44(6), 1128–1138.
45. C. Groh, W. Jo and J. Rödel, Frequency and temperature dependence of actuating performance of Bi_1/2Na_1/2TiO₃-BaTiO₃ based relaxor/ferroelectric composites, J. Appl. Phys., 2014, 115(23), 234107.
46. S. Steiner, et al., Influence of oxygen vacancies on core-shell formation in solid solutions of (Na,Bi)TiO₃ and SrTiO₃, J. Am. Ceram. Soc., 2021, 104(9), 4341–4350.
47. W. Jo, et al., Giant electric-field-induced strains in lead-free ceramics for actuator applications-status and perspective, J. Electroceram., 2012, 29, 71–93.
48. L. Li, et al., Controlling mixed conductivity in Na_1/2Bi_1/2TiO₃ using A-site non-stoichiometry and Nb-donor doping, J. Mater. Chem. C, 2016, 4(24), 5779–5786.
49. J. Hao, et al., Structure evolution and electrostrictive properties in (Bi_0.5Na_0.5)_0.94Ba_0.06TiO₃–M₂O₅ (M = Nb, Ta, Sb) lead-free piezoceramics, J. Eur. Ceram. Soc., 2016, 36(16), 4003–4014.
50. W. Zhao, et al., Fabrication of Na_0.5Bi_0.5TiO₃–BaTiO₃-textured ceramics templated by plate-like Na_0.5Bi_0.5TiO₃ particles, J. Am. Ceram. Soc., 2009, 92(7), 1607–1609.
51. G. A. Tina and R. Ranjan, Simultaneous enhancement of d₃₃ and depolarization temperature of the morphotropic phase boundary composition of the Pb-free piezoceramic Na_1/2Bi_1/2TiO₃–BaTiO₃, J. Eur. Ceram. Soc., 2025, 45(7), 117271.
52. X. Cheng, L. K. Venkataraman and Y. Li, Simultaneous enhancement of piezoelectric constant and thermal stability in lead-free Fe-doped 0.94 (Na_1/2Bi_1/2)TiO_3-0.06BaTiO₃ ceramics, J. Alloys Compd., 2022, 891, 161880.
53. D. K. Khatua, et al., Enhanced thermal stability of dielectric, energy storage, and discharge efficiency in a structurally frustrated piezoelectric system: Erbium modified Na_0.5Bi_0.5TiO₃-BaTiO₃, J. Appl. Phys., 2018, 124(10), 104101.
54. L. KV, et al., Hardening of electromechanical properties in piezoceramics using a composite approach, Appl. Phys. Lett., 2017, 111(2), 022905.
55. L. Cangini, et al., Effect of thermal depolarization on the poling-induced domain texture and piezoelectric properties in Mg-doped NBT-6BT, J. Am. Ceram. Soc., 2023, 106(11), 6879–6890.
56. O. Turki, et al., Lanthanides effects on the ferroelectric and energy-storage properties of (Na_0.5Bi_0.5)_0.94Ba_0.06TiO₃ ceramic: Comparative approach, Solid State Sci., 2021, 114, 106571.
57. J. Gu, et al., Effect of excess Bi₂O₃ on the microstructure and electrical properties of NBT-6BT lead free piezoceramics, Mater. Today Commun., 2025, 114055.
58. W. Zhao, et al., Influence of different dopants on the piezoelectric properties of the Na_1/2Bi_1/2TiO₃-BaTiO₃-lead-free ceramics, Key Eng. Mater., 2007, 336, 105–108.

---