Processing and properties of PSZT 95/5 ceramics with varying Ti and Nb substitution

Abstract

Niobium doped lead‐tin‐zirconate‐titanate ceramics near the PZT 95/5 orthorhombic AFE – rhombohedral FE morphotropic phase boundary Pb_1‐0.5y(Zr_0.865‐xTi_xSn_0.135)_1‐yNb_yO₃ were prepared according to a 2²⁺¹ factorial design with x = 0.05, 0.07 and y = 0.0155, 0.0195. The ceramics were prepared by a traditional solid‐state synthesis route and sintered to near full density at 1250°C for 6 h. All compositions were ∼98% dense with no detectable secondary phases by X‐ray diffraction. The ceramics exhibited equiaxed grains with intergranular porosity, and grain size was ∼5 µm, decreasing with niobium substitution. Compositions exhibited remnant polarization values of ∼32 µC/cm2, increasing with Ti substitution. Depolarization by the hydrostatic pressure induced FE‐AFE phase transition was drastically affected by variation of the Ti and Nb substitution, increasing at a rate of 113 MPa /1% Ti and 21 MPa/1% Nb. Total depolarization output was insensitive to the change in Ti and Nb substitution, ∼32.8 µC/cm2 for the PSZT ceramics. The R3c‐R3m and R3m‐Pm3m phase transition temperatures on heating ranged from 90 to 105°C and 183 to 191°C, respectively. Ti substitution stabilized the R3c and R3m phases to higher temperatures, while Nb substitution stabilized the Pm3m phase to lower temperatures. Thermal hysteresis of the phase transitions was also observed in the ceramics, with transition temperature on cooling being as much as 10°C lower.

INTRODUCTION

The perovskites lead zirconate (PbZrO₃) and lead titanate (PbTiO₃) form a continuous series of solid solutions over the entire composition range.¹ The lead-zirconate-titanate (PZT) system contains a ferroelectric to anti-ferroelectric phase boundary on the lead zirconate rich side of the phase diagram, near the Zr/Ti ratio of 95/5.^1–3 PZT-based ferroelectrics near the 95/5 morphotropic phase boundary (PZT 95/5) have found use in explosively actuated stand-by power supplies.^4,5 These devices can supply voltage or high current to devices through utilization of the pressure induced phase transformation from the FE-rhombohedral phase to the AFE-orthorhombic phase.^1,6

Modification of PZT 95/5 with niobium (Nb) has been shown to stabilize the FE Phase, promote densification, enhance breakdown strength, and reduce dielectric loss.^3,7–11 Additionally, Nb⁵⁺ acts as a donor dopant, reducing the coercive field (a “soft” PZT), improving breakdown strength, and improving aging characteristics.³ Modification by tin (Sn) has been shown to stabilize the AFE phase, enhance densification, and reduce the strain associated with the R3c-R3m and R3m-Pm3m phase transitions.^12–16 PZT 95/5 is typically doped with Nb or co-doping the of Sn4+ and Nb5+ to tailor the electrical properties and pressure-induced FE-AFE phase transition response in these materials.^4,6,13,14,17 Berlincourt and Krueger proposed a pseudo-binary phase diagram for the Pb_0.995[(Zr_0.85Sn_0.15)_1-xTi_x]_0.99Nb_0.01O₃ system, showing a near vertical morphotropic phase boundary for a Ti concentration of ∼4.5% (Figure 1).¹³ Limited studies on these systems have been reported since Berlincourt and Krueger, thus the full effects of chemical substitution in these systems is relatively unknown. Studies in the Nb-modified PZT 95/5 system have reported on the effects of Nb substitution¹⁸ and Ti substitution⁸; however, examples for PSZT 95/5 were not forthcoming, but comparable trends to these studies are expected.

FIGURE 1. Pb(Zr,Sn)O3-PbTiO3 pseudo binary phase diagram in the Pb0.995[(Zr0.85Sn0.15)1-xTix]0.99Nb0.01O3 system13

This manuscript explores the effect of varying the Ti and Nb concentration on the processing, thermal, electrical, and mechano-electrical properties of a niobium doped lead-tin-zirconate-titanate (PSZT)-based ceramic.

EXPERIMENTAL PROCEDURE Processing

This study used commercially available powders (Table 1). PSZT ceramics were prepared with a nominal composition of Pb_1.00025-0.5y(Zr_0.865-xTi_xSn_0.135)_1-yNb_yAl_0.0005O₃, with a with a 2²⁺¹ factorial design, where x is between 0.05 and 0.07, and y is between 0.0155 and 0.0195 (Table 2). The zirconium amount was compensated for the Hf impurity, 1.68 wt.%, such that the sum of actual zirconium and hafnium content matched the stoichiometric zirconium content. All compositions were prepared at a 1 kg nominal batch size. Aluminum nitrate was dissolved in 2-propanol, the batched powders (ZrO₂, PbO, SnO₂, etc.) added, then mixed by ball milling in 1 L baffled HDPE jars, in 2-propanol, using 5 mm diameter YSZ grinding media, at ∼90 RPM for 4 h. The slurry was then pan dried and sieved to −20 mesh. The powder was then placed in covered high-purity alumina crucibles (AD998, CoorsTek, Golden, CO) and calcined at 800°C for 4 h in convective air (Thermolyne F48028-80, Thermo Fisher Scientific, Waltham, MA). The calcined powder was ground using a YSZ auto mortar and pestle (RM 200, Retsch), followed by ball milling as previously described. After milling for 4 h, the slurry was pan dried, then sieved to −20 mesh. The powder was then calcined in covered high-purity alumina crucibles at 800°C for 4 h in convective air. The calcined powder was ground using a YSZ auto mortar and pestle, followed by ball milling in 2-propanol. After 4 h, a binder solution was added and mixed by ball milling for an additional 30 min. The binder was 1 wt.% PEG300, 2.5 wt.% PMMA (Dispex CX 4204, BASF Corp. Florham Park, NJ), 0.5 wt.% PEG6000, and 0.5 wt.% oleic acid. In this binder system, PMMA and PEG6000 acted as the binder, PEG300 as a plasticizer, and oleic acid as a lubricant. The PSZT-organics slurry was dried by automated rotary evaporation, (R-215, Buchi, Flawil, Switzerland) using the 2-propanol setting, at 60 rpm. The dried powder was granulated by lightly grinding with a high-purity alumina mortar and pestle and sieving to −40 mesh. The granulated powder was pressed into approx. 25.4 mm × 25.4 mm × 12.7 mm rectangular prisms in polished steel dies at 28 MPa. The green preforms were then vacuum bagged and cold isostatically pressed (CP92466, Avure Technologies, Västerås, Sweden) to 210 MPa for 2 min. The green compacts were then heat treated under flowing air to thermally decompose the binder and presinter the compacts (bisque). The bisque parts were then sintered in seasoned high-purity alumina crucibles in a sealed triple crucible configuration (Figure 2). PZT powder with the same composition as the parts was used as the PbO vapor source during sintering, and PZT spacers with the same composition as the parts were used to separate the billets from the crucible. The parts were sintered in a SiC element bottom loading furnace (DT-25-SBL-AE-1088, Deltech Furnaces, Golden, CO) in stagnant air by heating at 1°C/min to 1250°C, isothermally dwelling for 6 h, cooling at 0.5°C/min to 250°C, isothermally dwelling for 2 h, finally cooling to room temperature at 0.2°C/min.

TABLE 1 Summary of raw materials used. Particle size is the D50 reported from the supplier CoA

Material	Vendor	Grade	Purity	Particle size (µm)
PbO	Hammond	100Y	99.95%	< 6
ZrO2	Tosoh	TZ-0	99.97% excl. Hf 1.68 wt.% Hf	0.027
SnO2	Materion	T-1137	99.9%	0.23
TiO2	Materion	T-1156	99.9%	0.52
Nb2O5	H.C. Starck	Ceramics Grade	99.9%	0.6
Al(NO3)3·9H2O	Sigma-Aldrich	ACS	≥ 98%	–
PEG300	BASF	Kollisolv 300	–	–
PEG6000	Alfa-Aesar	A17541	–	–
PMMA	BASF	Dispex CX 4204	–	–
Oleic Acid	Sigma-Aldrich	Technical	> 95%	–

TABLE 2 Summary of density and open porosity for the sintered PSZT ceramics

ID	Ti(x)	Nb(y)	Sintered bulk density (g/cm3)	Relative density (%)	N
00	0.06	0.0175	7.988 ± 0.001	98.10	3
++	0.07	0.0195	7.999 ± 0.008	98.30	3
−+	0.05	0.0195	7.990 ± 0.012	98.05	3
−−	0.05	0.0155	7.980 ± 0.006	98.08	3
+−	0.07	0.0155	7.984 ± 0.004	98.01	3

FIGURE 2. Schematic of the sealed triple crucible configuration used to sinter the PSZT ceramics

Characterization

Bulk density of the sintered parts was measured in accordance with ASTM C830, using water as the immersion media and saturation time under vacuum reduced to 5 min. Samples were prepared for X-ray diffraction (XRD) by machining into 10 mm × 10 mm × 1 mm coupons. The coupons were annealed at 400°C for 1 h in air, with a heating and cooling rate of 1°C/min. XRD (D2 Phaser, Bruker, Madison, WI) was performed using Cu-Kα radiation, 10–80° 2Θ, 0.02° step, 16.2 s/step. Structural refinement was conducted with TOPAS v5. Relative density is reported as the measured bulk density divided by the calculated crystallographic density.

Specimens were prepared for microstructural analysis by cross sectioning the billets parallel to the pressing direction, followed by polishing to 1 µm using progressively finer diamond abrasives, and finished by vibratory polishing using colloidal silica. The polished cross-sections were etched in 36% HCl for ∼2 min at room temperature to reveal grain boundaries. Imaging was performed using light optical microscopy (LOM; MA100N, Nikon Metrology Inc, Brighton, MI), under Köhler illumination with polarized white light. The images were segmented using digital image analysis software (MIPAR v3.2.4, MIPAR Software, Worthington, OH), and the grain size, equivalent area diameter (EAD), was measured.

Electrical test specimens of 7 mm × 7 mm × 3 mm were machined from the sintered billets. Gold electrodes were sputter-coated (Q150R Plus, Quorum Technologies, Lewes, UK) on the 7 mm × 7 mm faces and the edges chamfered. Specimens were poled at room temperature by applying an electric field of ± 45 kV/cm at 0.04 Hz in Flourinert FC-33 at room temperature. After poling, the specimens were immediately looped again to generate the P-E hysteresis loop. From the P-E curve, the positive bias coercive field (E_C), positive bias saturated polarization (P_S), and initial remnant polarization (P_R) were calculated. The specimens were then depoled under hydrostatic loading conditions. Hydrostatic pressure was applied to the specimens using a custom cold isostatic press with Isopar H as the dielectric fluid. Charge was measured using a digital multimeter with a 4 µF integrating capacitor. Total depolarization (Q_H) was calculated from the cumulative depolarization versus hydrostatic pressure (Q-P) curve, and the hydrostatic depoling pressure (P_H, pressure at which FE to AFE transformation is induced) was calculated as the P-axis position of the peak maximum in the 1st derivative of the Q-P curve (dQ/dP versus P).

Dielectric spectroscopy (HP4294A, Agilent Technologies, Santa Clara, CA) was performed as a function of temperature (EC1A, Sun Electronic Systems, Inc., Titusville, FL) on the unpoled specimens. Specimens were heated at a rate of 2°C per min to 250°C then cooled to room temperature at 2°C/min. The R3c ↔ R3m and R3m ↔ Pm3m phase transformation temperatures were calculated from the peak in the Tan δ versus T curve and the κ versus T curve, at a frequency of 1 kHz, respectively. Data analysis was performed using DIAdem 2019 (National Instruments, Austin, TX), JMP 14.2 (SAS Institute Inc, Cary, NC), and OriginPro 2020b (Origin Labs, Northampton, MA).

RESULTS AND DISCUSSION

All results presented were analyzed by fitting the full factorial model to each parameter separately. The fit parameter terms, p-values, and full model fitting descriptors can be found online in the supplemental materials (Table S1). A value of α = 0.05 is used to determine statistical significance, while a coefficient of determination value of >0.9 is used to determine a good fit with the independent variable.

The sintered bulk densities are summarized in Table 2. After sintering, all parts exhibited thermal stress cracks extending approx. 1–2 mm into the billets (not shown). Density increased slightly with increasing Ti and Nb concentration (p = 0.02, R² = 1), from 7.98 to 8.00 g/cm3 (Figure 3). All compositions were identified as having the R3c structure, with no additional phases (i.e., ZrO₂, PbO, the Pbam structure, etc.) detected by XRD (Figure 4). Rietveld refinement was used to calculate the lattice parameters (Table 3). The a- and c-axis parameters both decreased with increasing Ti and Nb concentration (Figure 5). The a-axis decreased from 5.837 to 5.834 Å (p = 0.35, R² = 0.92), and the c-axis decreased from 14.378 to 14.368 Å (p = 0.26, R² = 0.96) from the – to the ++ treatments, respectively. These minute changes in lattice parameters are consistent with the differences in ionic radii for Zr⁴⁺, Sn⁴⁺, Ti⁴⁺, and Nb⁵⁺ with 6-fold coordination of 0.72, 0.69, 0.605, and 0.64 Å, respectively. Crystallographic density was calculated using the nominal stoichiometry for each composition, shown in Table 2, and the measured lattice parameters, shown in Table 3. From the calculated crystallographic densities, all compositions were approximately 98% dense.

FIGURE 3. Sintered bulk density as function of Ti and Nb concentration in the PSZT ceramic

FIGURE 4. X-ray diffraction (XRD) patterns for the PSZT ceramics.

TABLE 3 Summary of calculated R3c lattice parameters and crystallographic density from structure refinement of the X-ray diffraction (XRD) patterns for the PSZT ceramics

ID	a-axis (Å)	c-axis (Å)	Crystallographic density (g/cm3)
00	5.8366 ± 0.0002	14.3733 ± 0.0006	8.143
++	5.8343 ± 0.0002	14.3681 ± 0.0005	8.137
−+	5.8367 ± 0.0001	14.3714 ± 0.0005	8.149
−−	5.8374 ± 0.0001	14.3778 ± 0.0005	8.136
+−	5.8344 ± 0.0001	14.3687 ± 0.0005	8.146

FIGURE 5. Calculated lattice parameters as a function of Ti and Nb concentration for the PSZT ceramics

Representative polished and chemically etched microstructures of the sintered PSZT ceramics are shown in Figure 6. Table 4 summarizes the grain size, EAD and the DeBrouckere mean diameter ([D4,3]), and distribution statistics. A finer grain size is exhibited near the surface of the billets, approximately 1–3 mm thick, and a larger grain size in the core of the billet. Reported grain sizes were taken from the core region of the sintered billets. Grain size (Figure 7) was approximately 5.5 ± 2.3 µm, unchanging with Ti content (p = 0.89), decreasing to approximately 4.0 ± 1.5 µm as Nb increased from 0.05 to 0.07 (p = 0.33). A similar trend was observed for [D4,3], with size decreasing from approximately 8 to 6 µm as Nb concentration was increased (p = 0.28, R² = 0.95). The observed decrease in grain size with Nb addition is consistent with previous reported observation in the non-Sn modified PZT 95/5 reported by Yang et al.⁸ Conversion to the volume weighted distribution shows similar trends for the grain sizes for the compositions. The Dv10, Dv50, and Dv90 all decrease with addition of Nb. However, it is interesting that the span of the distribution ((Dv90-Dv10)/Dv10) also increased with Nb addition, increasing from ∼0.8 to 1.0. This demonstrates that reducing Nb content results in a larger but more narrow distribution. Porosity was predominantly intergranular and uniformly distributed. Changes in Ti and Nb content did not appear to significantly change the porosity. Thus, grain size of the PSZT ceramics appears to be relatively unchanging with changes in Ti content and decreases with increasing Nb Content.

FIGURE 6. Light optical micrographs of the polished and chemically etched cross-sections of the PSZT ceramic

TABLE 4 Summary of the equivalent area diameter for the PSZT ceramics sintered at 1250°C for 6 h

	Grain size (µm)
ID	EAD	Dv10	Dv50	Dv90	[D4,3]	Pore size (µm)	N
00	3.8 ± 1.9	3.7	6.1	9.9	6.5	1.6 ± 1.3	1343/384
++	3.8 ± 1.5	3.4	5.4	9.1	5.9	2.6 ± 1.0	2019/143
−+	4.2 ± 1.7	3.8	6.0	8.6	6.1	3.0 ± 1.5	1641/169
−−	5.4 ± 2.2	4.9	7.8	11.5	8.0	3.2 ± 1.6	1011/106
+−	5.5 ± 2.3	5.1	7.9	11.2	8.1	3.2 ± 1.7	1896/226

FIGURE 7. Grain size descriptors as function of Ti and Nb concentration in the PSZT ceramic

P-E hysteresis looping was used to measure polarizability and coercive field (Figure 8, 9). Saturated polarization, remnant polarization, and coercive field values are summarized in Table 5. Saturated polarization (Ps+) increased slightly with Ti addition, 36.0 to 36.5 µC/cm², while increasing Nb decreased Ps+, from 36.6 to 36.0 µC/cm2 but was not statistically significant (R² = 0.97, p = 0.213). Remnant polarization (Pr-) increased with Ti content, from 33.0 to 33.2 µC/cm2, and decreased with Nb addition, from 33.2 to 32.9 µC/cm², and is statistically significant (R² = 1.00, p = 0.03). The initial remnant polarization (Pr_i; taken as the starting polarization after ∼5 min between tests) is ∼0.4 µC/cm² below the remnant polarization and follows the same trends but was not statistically significant (p = 0.482). Thus, charge storage decreases with increasing Nb substitution but increases with Ti substitution since polarization also decreases with Nb substitution and increases with Ti substitution, but density remains relatively unchanged. The coercive field was ∼8.5 kV/cm for all compositions, unchanging with Ti and Nb substitution (p = 0.955). Thus, substitution of Ti from 0.05 to 0.07 and Nb from 0.0155 and 0.0195 (stoichiometric values) resulted in an increase and decrease in polarization values, respectively, of the samples and no change in coercive field value.

FIGURE 8. P-E hysteresis loops for the PSZT ceramics

FIGURE 9. Saturated polarization, remnant polarization, and coercive field as function of Ti and Nb concentration in the PSZT ceramic

TABLE 5 Summary of saturated polarization, initial remnant polarization, and coercive field for the PSZT ceramics sintered at 1250°C for 6 h

ID	Pr- (µC/cm2)	Pri (µC/cm2)	Ps+ (µC/cm2)	Ec+ (kV/cm)	N
00	33.08 ± 0.40	32.80 ± 0.21	36.53 ± 0.34	8.62 ± 0.01	6
++	33.25 ± 0.44	32.74 ± 0.27	36.52 ± 0.48	8.54 ± 0.02	5
−+	32.70 ± 0.17	32.19 ± 0.17	36.04 ± 0.15	8.54 ± 0.01	5
−−	33.29 ± 0.32	32.73 ± 0.30	36.68 ± 0.37	8.49 ± 0.02	6
+−	33.12 ± 0.34	32.67 ± 0.32	36.52 ± 0.36	8.50 ± 0.03	6

Room temperature pressure induced depolarization via the FE to AFE phase transition was measured for the PSZT ceramics (Figure 10, 11). As pressure is increased, a small amount of charge is released, due to the direct piezoelectric effect, until reaching the transformation pressure, at which the R3c phase converts to the Pbam phase, releasing all stored space charge. Further increasing pressure does not result in the release of any additional charge. The samples tested all exhibited a sharp transition, with relatively little curvature before and after the transition, indicating chemical homogeneity within the samples. Table 6 summarizes the calculated depolarization pressure and maximum total depolarization for the PSZT ceramics. The ++ composition did not depole during testing, indicating its pressure induced transformation pressure is above 480 MPa. In this case, the interaction term was excluded when fitting the statistical model. Depolarization occurred at 235 MPa for the “–” treatment, increasing to 278 MPa for “-+”, further increasing to 460 MPa for the “+-” treatment. As expected, the composition has a strong effect on the pressure needed to drive the FE-AFE phase transformation (R² = 1.00, p = 0.005). Within the range of stoichiometric substitution investigated, Ti increases the FE-AFE transformation pressure at a rate of 113 MPa / %, while Nb increases the transformation pressure at a rate of 21.5 MPa / %. Addition of Ti and Nb to PZT 95/5 are known to stabilize the FE phase further away from the FE-AFE transition.^3,8,13,18 The measured rate of change of 113 MPa / % Ti in the present PSZT 95/5 system is identical to that previously reported for PZT 95/5 by Starcher.¹⁸ The rate of change for the Nb substitution is drastically different from that previously reported by Yang et al. for the PZT 95/5 family of 180 MPa / %.⁸ Thus, substitution of Ti into the PSZT 95/5 ceramic more strongly stabilizes the FE phase compared to the Nb substitution. This difference in substitution response between Ti and Nb could allow for more fine tuning of the pressure needed to induce FE-AFE transformation compared to the PZT 95/5 family without Sn modification.

FIGURE 10. Representative waveforms of total depolarization as a function of increasing hydrostatic pressure for the PSZT ceramics

FIGURE 11. Hydrostatic depoling pressure and total depolarization as function of Ti and Nb concentration in the PSZT ceramic

TABLE 6 Summary of depoling pressure (at 15 µC/cm2 depolarization) and maximum depolarization for the PSZT ceramics sintered at 1250°C for 6 h

ID	Ph (MPa)	Qh (µC/cm2)	N
00	370.3 ± 4.0	33.11 ± 0.19	5
++	>483	ND	5
−+	277.7 ± 3.2	32.46 ± 0.42	3
−−	234.9 ± 5.5	32.72 ± 0.19	3
+−	460.9 ± 3.2	32.82 ± 0.37	5

Phase transition temperatures were also measured for the Hf-doped PSZT ceramics (Figure 12, 13). Table 7 summarizes the calculated phase transition temperatures on heating and cooling. On heating, the R3c to R3m phase transition temperature increased from 90 to 105°C (p = 0.001) and the R3m to Pm3m transition temperature slightly increased from 184 to 190°C (p = 0.130) as Ti substitution increased from 0.05 to 0.07. Similar on heating, Nb substitution between 0.0155 and 0.0195 had no effect on the R3c-R3m transition on heating (p = 0.263) and the R3m to Pm3m transition temperature slightly decreased from 186 to 183°C (p = 0.273). Upon cooling, the Pm3m to R3m transition temperature slightly increased from 181 to 186°C (p = 0.001) and the R3m to R3c transition temperature increased from 79 to 101°C (p = 0.003) as Ti substitution increased from 0.05 to 0.07. Also, upon cooling, Nb substitution between 0.0155 and 0.0195 decreased the Pm3m to R3m transition temperature slightly decreased from 188 to 185°C (p = 0.001) and had negligible impact on the R3m to R3c transition (p = 0.272). The material exhibits thermal hysteresis, with the R3c to R3m transition occurring as much as 10°C higher on heating than on cooling, and the R3m to Pm3m phase transition occurring only about 3°C higher on heating than cooling. In the present PSZT 95/5 system, increasing Ti substitution is seen to stabilize the FE phases, as expected from the phase diagram, while the increasing Nb addition is seen to stabilize the paraelectric phase.

FIGURE 12. Permittivity and loss tangent as a function of temperature, upon cooling, for the PSZT ceramics, measured at 1000 Hz

FIGURE 13. Phase transition temperatures as a function of Ti and Nb concentration in the PSZT ceramics

TABLE 7 Summary of R3c-R3m-Pm3m phase transition temperature hysteresis for the PSZT ceramics sintered at 1250°C for 6 h

	Transition temperature (°C)
ID	R3c to R3m	R3m to R3c	R3m to Pm3m	Pm3m to R3m
00	98	91	186	184
++	105	102	189	185
−+	90	80	183	180
−−	90	78	186	183
+−	106	100	191	188

Note: Measurement error is ± 1.2°C for all measurements.

These results demonstrate that adjusting the titanium and niobium substitution in niobium doped lead-tin-zirconate titanate has negligible effects on processing but noticeable effects on the microstructure and properties of the ceramics. While bulk density was relatively unchanged with the Nb and Ti concentrations, small but significant changes in the grain size, polarizations, coercive field, and R3c-R3m and R3m-Pm3m phase transition temperatures are observed. Depolarization pressure was drastically affected by adjustments to the Ti and Nb concentrations, while depolarization output remained relatively unchanged. Utilization of design of experiments and a statistical approach to the results analysis allows for the creation of a statistical model to explore the composition range. This allows for optimization of the chemistry within the experimental boundaries to obtain the desired properties.

SUMMARY

Niobium doped lead-tin-zirconate-titanate ceramics near the PZT 95/5 morphotropic phase boundary were prepared with varying stoichiometric substitution amounts of titanium (0.05–0.07) and niobium (0.0155–0.0195), utilizing a 2²⁺¹ factorial design. The ceramics were prepared by a traditional solid-state synthesis route and sintered to near full density at 1250°C for 6 h in sealed alumina crucibles with self-same material as the lead vapor source. The sintered ceramics exhibited the following properties:

• Sintered density of 7.99 g/cm³ for all compositions, with ∼98% relative density and no detectable secondary phases.

• Grain size was between 4 and 6 µm, decreasing with Nb addition, and consisting of equiaxed grains with intergranular porosity.

• Saturated polarization values of ∼36 µC/cm², remnant polarization values of ∼32 µC/cm², slightly decreasing with Nb addition. Coercive field values of ∼8.5 kV/cm, insensitive to the Ti and Nb concentration.

• Depolarization by the pressure induced FE-AFE phase transition occurred at 370 MPa for the center point composition. The depolarization pressure changed at a rate of 113 MPa / 1% Ti and 22 MPa / 1% Nb.

• Evaluation of the R3c-R3m and R3m-Pm3m phase transition temperatures showed temperatures on heating ranging from 90 to 105°C and 183 to 191°C, respectively. The R3c-R3m and R3m-Pm3m transition temperature increased with Ti addition, and the R3m-Pm3m transition temperature decreased with Nb addition. Thermal hysteresis was also observed, with transition temperature on cooling being ∼10 and ∼3°C lower for the R3c-R3m and R3m-Pm3m, respectively.

The study demonstrates that the PSZT 95/5 ceramics are sensitive to variations in the titanium and niobium contents over small percentage regions. This outcome is consistent with expectations from the limited studies reported in the literature. This demonstrates that manufacturing processes for PSZT for ferroelectric generators need to exercise precise control over their chemistry to obtain the desired properties.

ACKNOWLEDGMENTS

The authors would like to thank the Active Ceramics Value Stream organization, Analytical Technologies Value Stream organization, The Ferroelectric Neutron Generator, Neutron Tube, and Active Ceramics Lifecycle Engineering organization, and The Material Characterization and Performance department of Sandia National Laboratories for their assistance with sample preparation and characterization. The authors also thank Dr. Jonathon Bock for assistance with technical review of the manuscript. This article has been authored by an employee of National Technology & Engineering Solutions of Sandia, LLC under Contract No. DE-NA0003525 with the U.S. Department of Energy (DOE). The employee owns all right, title and interest in and to the article and is solely responsible for its contents. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this article or allow others to do so, for United States Government purposes. The DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan https://www.energy.gov/downloads/doe-public-access-plan.