Identification of Phase Boundaries and Electrical Properties in Ternary

Jul 12, 2016 - Department of Materials Science, Sichuan University, Chengdu, 610064, ... piezoelectric activity endures a leapfrog development from âˆ...
1 downloads 0 Views 6MB Size
Research Article www.acsami.org

Identification of Phase Boundaries and Electrical Properties in Ternary Potassium-Sodium Niobate-Based Ceramics Xiang Lv, Jiagang Wu,* Shuang Yang, Dingquan Xiao, and Jianguo Zhu Department of Materials Science, Sichuan University, Chengdu, 610064, P.R. China ABSTRACT: A large piezoelectric constant (d33) of ∼480 pC/N was attained in new ternary (1−x−y)K0.5Na0.5Nb0.96Sb0.04O3-xBaSnO3-yBi0.5Na0.5ZrO3 ceramics by forming rhombohedral-orthorhombic-tetragonal (R-O-T) phase boundary using the variations of x and y, and such a phase boundary was successfully confirmed by the convergent beam electron diffraction (CBED) patterns. For (1−x)K0.5Na0.5Nb0.96Sb0.04O3-xBaSnO3, the orthorhombic (O) phase is well-maintained for 0 ≤ x ≤ 0.015, and both the R and T phases can be introduced to (0.99−y)K0.5Na0.5Nb0.96Sb0.04O3-0.01BaSnO3-yBi0.5Na0.5ZrO3 with y = 0.025−0.04 by simultaneously tailoring their compositions (x and y); then, R-O-T multiphases can be well-established. The CBED patterns strongly support the existence of R-O-T multiphases in the ceramics with y = 0.035. When the phase transitions endure from O to R-O-T, their piezoelectric activity endures a leapfrog development from ∼165 to ∼480 pC/N. In the region of the R-O-T phase boundary, a large d33 of ∼480 pC/N was attained in the ceramics with x = 0.01 and y = 0.035. In addition, the ceramics with x = 0.01 and y = 0.04 possess a high strain of ∼0.274% due to the multiphases coexistence. According to the variations of dielectric and ferroelectric properties, the enhancement in εr and Pr plays a part in the improved d33 except for the RO-T phase boundary. We believe that the (K, Na)NbO3 ternary systems can be used to promote piezoelectric activity by forming new phase boundaries. KEYWORDS: KNN, ternary system, phase boundary, piezoelectricity, CBED

1. INTRODUCTION Recently, we systematically reviewed the development of phase boundaries in (K, Na)NbO3 (KNN)-based materials1,2 and found that new phase boundaries can effectively promote their piezoelectric properties. According to our previous reports, the piezoelectric constant (d33) of KNN-based ceramics is dependent not only on phase boundary types but also on the corresponding compositions.1−15 For example, the same orthorhombic-tetragonal (O-T) phase boundary is shown in the (K 0 . 4 5 Na 0 . 5 5 ) 0 . 9 8 Li 0 . 0 2 (Nb 0 . 7 7 Ta 0 . 1 8 Sb 0 . 0 5 )O 3 and Li0.02(Na0.55K0.45)0.98(Nb0.77Ta0.18Sb0.05) O3 ceramics, but there is a dramatic difference in their d33 values.1,3 In addition, the rhombohedral-tetragonal (R-T) phase boundary can be used to further promote the piezoelectric properties of KNN-based ceramics, but the corresponding compositions will strongly affect their d33 values.1,2 For example, a relatively low d33 of 380−400 pC/N can be observed in KNNS ceramics using the addition of Bi0.5Li0.5ZrO3 or Bi0.5(Na0.7K0.2Li0.1)0.5ZrO3;4,5 however, the ceramics with Bi0.5(Na1−wKw)0.5ZrO3 can attain a higher d33 of ∼490 pC/N.6 As a result, an obvious difference in d33 of KNN-based ceramics can be driven by changing the compositions. Previously, it was reported that the intermediate phase in the R-T phase boundary plays an important role in the origin of piezoelectricity for Pb(Zr,Ti)O 3 (PZT), Pb(Mg1/3Nb2/3)O3-PbTiO3, and (Ba0.7Ca0.3TiO3)(BaZr0.2Ti0.8O3) materials.16−18 It is well-accepted that there © XXXX American Chemical Society

is no group/subgroup relationship between the T and R symmetries and that an intermediate phase is expected to be involved in the T and R phases.16−18 Therefore, it has been questioned whether the different phase compositions of KNNbased ceramics can be induced by changing the doping element types. As a result, it is necessary to further modify the piezoelectric activity of KNN-based ceramics by refining the compositions, and then actual phase compositions should be further identified. Over the past decade, we tried to promote the piezoelectricity of KNN-based ceramics by constructing new phase boundaries using the ternary material systems;1,2,4−8 unfortunately, a limited improvement in d33 (400 pC/N) in the KNN ternary materials by refining their compositions. Previously, BaSnO3 was used to improve the piezoelectric properties of a material.19,20 For example, an ultrahigh d33 of ∼697 pC/N can be realized in (1−x)BaTiO3xBaSnO3 ceramics due to multiphase coexistence,19 and its TR‑O gradually increases as x rises.20 As a result, BaSnO3 may be considered as a rhombohedral additive. In this work, we examine one example of the (1−x−y)K0.5Na0.5Nb0.96Sb0.04O3Received: April 11, 2016 Accepted: July 5, 2016

A

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. XRD patterns of the ceramics as a function of (a−c) x and (d−f) y. scattering geometry of λ = 632 nm. The detailed procedures for the measurement of other electrical properties (e.g., crystal structure, dielectric, ferroelectric, and piezoelectric) can be found in our previous references.4−8

xBaSnO3-yBi0.5Na0.5ZrO3 ternary system to illuminate how composition affects the piezoelectric properties of KNN-based ceramics by designing new phase boundaries. Here, the addition of BaSnO3 can increase TR‑O values of KNN, while its TO‑T and TC remain almost unchanged. As a result, the piezoelectric properties of KNN ceramics cannot be effectively improved by constructing phase boundaries using the addition of only BaSnO3. Facing this serious situation, we doped Bi0.5Na0.5ZrO3 to K0.5Na0.5Nb0.96Sb0.04O3-BaSnO3 to drive different phase boundaries using the ternary material systems (i.e., (1−x−y)K0.5Na0.5Nb0.96Sb0.04O3-xBaSnO3yBi0.5Na0.5ZrO3) where Bi0.5Na0.5ZrO3 (BNZ) can be used to lower TO‑T and increase TR‑O of KNN.1,2,6,21 Importantly, a high d33 of ∼480 pC/N can be attained in such a KNN ternary system by constructing a new R-O-T phase boundary using the refined compositions, where the phase compositions can be identified by the convergent beam electron diffraction (CBED) patterns. We importantly emphasize the physical mechanisms for the composition-driven phase boundary and illuminate the relationships among composition, phase boundary, and piezoelectricity of KNN ternary systems.

3. RESULTS AND DISCUSSION First, the crystal structures of the samples were analyzed by their corresponding X-ray diffraction (XRD) patterns.1−20 Figure 1a shows the XRD patterns of the (1−x)K0.5Na0.5Nb0.96Sb0.04O3-xBaSnO3 ceramics measured at 20 °C and 2θ = 20−70°. One can observe from Figure 1a that all ceramics mainly displayed a perovskite structure with a trace of secondary phases (marked by the red diamonds), and the intensity of the secondary phase slightly increased with increasing x values. Figures 1b and c show the expanded XRD patterns of the ceramics located at 2θ = 21−23° and 44− 47°. One can see that the ceramics with x ≤ 2.0% possess an O phase, which is similar to the previously reported results of a pure KNN ceramic.9,10 However, the phase structure of the ceramics shifts from O to the cubic phase as x increases (x ≥ 4.0%). Usually, the O phase cannot cause a dramatic increase in the piezoelectric properties of KNN,1,9,10,14 making it necessary to construct other phase boundaries by doping different additives. As a result, we doped Bi0.5Na0.5ZrO3 to this material and found that different phase boundaries can be driven by changing its corresponding contents.1,6 Figure 1d shows the XRD patterns of ternary (1−x−y)K0.5Na0.5Nb0.96Sb0.04O3xBaSnO3-yBi0.5Na0.5ZrO3 (x = 0.01) ceramics measured at 20 °C and 2θ = 20−70°. The perovskite structure together with very few traces of secondary phases can be attained by the addition of BNZ, as marked by the red diamond (Figure 1d.1,6 To clearly identify the secondary phases, the XRD data of the ceramics with x = 0.04 and y = 0.04 were retrieved, as shown in Figure 2. It can be seen from Figure 2a that the secondary phases of the ceramics with x = 0.04 mainly belong to SnO2 and

2. EXPERIMENTAL PROCEDURES (1−x−y)K0.5Na0.5Nb0.96Sb0.04O3-xBaSnO3-yBi0.5Na0.5ZrO3 (x = 0− 0.04 and y = 0−0.06) ceramics were prepared by normal sintering, and raw materials included K2CO3 (99%), Na2CO3 (99.8%), Nb2O5 (99.5%), Sb2O3 (99.99%), BaCO3 (99%), SnO2 (99%), Bi2O3 (99.999%), and ZrO2 (99%). The detailed fabrication process can be found in our previous work.4−8 Before transmission electron microscopy (TEM) observation, the samples were prepared by following a conventional method of mechanical thinning, dimpling, and ion milling to electron transparency. TEM measurement was carried out using a JEM-2010 microscope with an acceleration voltage of 200 kV and a CBED measurement with a 5 nm probe size used to determine the crystal symmetry. Raman spectra of the samples were carried out using a Renishaw InVia Raman spectrometer under a back B

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

matrix, which is responsible for the occurrence of secondary phases in the ceramics with x = 0.04. As shown in Figure 2b, the secondary phases of the ceramics with y = 0.04 are due to the formation of K6(Nb10.88O30). Such a small amount of secondary phases were also observed in other work.9 More importantly, we can see from Figures 1e and f that the XRD peak shapes are strongly dependent on the doped BNZ contents. For the ceramics with x ≤ 0.01, the O phase still exists, and the phase structure of the ceramics with x > 0.01 will be explained by considering both their XRD patterns and temperature dependence of dielectric constant (εr−T) curves. It has been confirmed that εr−T can be used to characterize the phase transition temperatures of KNN-based ceramics.1 To clearly show their phase evolutions, we characterized the corresponding εr−T curve of the ceramics with the variations of x and y. Figures 3a−m show the εr−T curves of the ceramics with different x and y values, measured at −150 to 200 °C and 100 kHz. Their TR‑O and TO‑T values can be clearly observed in all ceramics, which are dependent on the doped compositions as well as their contents. Figures 3a−f plot the εr−T curves of the ceramics with different x values (y = 0). It was found that their TR‑O values gradually shift to high temperatures as x increases, while their TO‑T slightly increases. Previously, it was reported that the variations of TR‑O and TO‑T can be caused by doping BaSnO3.23 As a result, R-T or other phase boundaries cannot be driven by introducing only BaSnO3; other additives must be doped if new phase boundaries are wanted. Figures 3h−m plot the εr−T curves of the ceramics with different y values (x = 0.01). We can observe from Figures 3h−m that

Figure 2. Analysis for the secondary phases in the ceramics with (a) x = 4.00% and (b) y = 4.00%.

other counterparts. As the contents of SnO2 exceed the solid solubility, excess SnO2 cannot diffuse into the KNN-based

Figure 3. Temperature dependence (−150 to 200 °C) of the dielectric constant of the ceramics as a function of (a−f) x and (h−m) y. C

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 4. Retived refinement for the XRD pattern of the ceramics with y = 0.035. Inset shows the results of the Retived refinement.

Figure 5. (a−c) CBED patterns of the ceramics with y = 0.035 and [001] and [11̅0] beam incidence. (d) Domain structure of the samples with y = 0.035.

Figure 4. One can see from Figure 4 that the fitting data is consistent with the original. According to the results (see inset of Figure 4), the space groups (SG) obtained in this work are R3m (rhombohedral), Amm2 (orthorhombic), and P4mm (tetragonal), indicating the existence of the R-O-T phase boundary.15 Therefore, the results of the Rietveld refinement are consistent with the ones derived from XRD patterns and εr−T curves. More importantly, detailed cell parameters and phase contents were also obtained (see inset of Figure 4); that is, 33% for the R phase, 28% for the O phase, and 39% for the T phase. As a result, the results of the Rietveld refinement are also consistent with the CBED patterns (Figures 5a−c), as shown in the following discussion.

their TR‑O and TO‑T values gradually approach room temperature by modifying BNZ contents, and finally, the R-T phase boundary may be driven by controlling y values. It was reported that BNZ can simultaneously increase TR‑O and decrease TO‑T of KNN materials,1,2,6,21 and an intermediate phase between T and R was often observed in the morphotropic phase boundary (MPB) regions.16−18 Therefore, the question arises of whether there is an intermediate phase in such a material. In this work, we detected the phase structure in detail using CBED patterns. As we know, the Rietveld refinement method is a good way to obtain detailed information on phase structure.15 In this work, the Rietveld refinement for the ceramics with y = 0.035 was conducted using the Maud software package as shown in D

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 6. Temperature dependence (60−500 °C) of dielectric constant of the ceramics as a function of (a) x and (b) y. Phase diagrams of the ceramics with different (c) x and (d) y values.

Figure 7. (a) Raman scatterings of the ceramics with variations of y. Raman shift and fwhm of (b) ν5 and (c) ν1. (d) polymorphic phase boundary (PPB) of the ceramics with variations of y. Fittings of Raman spectra of the ceramics with (h) y = 0.01, (g) y = 0.02, (f) y = 0.035 and (e) y = 0.05 in the wavenumber range of 500−700 cm−1. (i) Raman shift of Eg (υ2) and A1g (υ1) as a function of y.

the same grains, strongly verifying the existence of R-O-T phase boundaries in the samples with y = 0.035. According to previous literature involving Pb-based and Pbfree materials, the distinct domain structures are always observed in high piezoelectric compositions lying in MPB or polymorphic phase transition (PPT).21,25−30 In the present work, we also measured the domain structure using highresolution TEM. The bright field (BF) TEM image (Figure 5d) shows the microstructure of the samples with y = 0.035. One can see that the submicron domains (one of which is shown by the dashed line in Figure 5d) with widths of around 200−300 nm can be observed in the samples. Moreover, inside the submicron domain, there are many miniaturized domains or nanodomains with domain widths of 10−30 nm, as shown by solid lines in Figure 5d. This phenomenon, which is the miniaturized nanodomain configuration in a domain hierarchy, was also observed in other Pb-based or Pb-free ceramics with high piezoelectric effects.21,25−30 As a result, except for phase boundaries, another contribution coming from ferroelectric

To directly verify the three coexisting ferroelectric phases (i. e., R, O, and T) in the samples, the CBED measurements of the samples with y = 0.035 were carried out as shown in Figures 5a−c. First, the CBED patterns were observed along [010] beam incidence as shown in Figure 5a. One can see from Figure 5a that the CBED pattern shows the mirror symmetry with mirror plane//[001], which is consistent with the 4mm symmetry (tetragonal phase).25−28 Then, the CBED patterns were observed from [110] beam incidence in the same grain as shown in Figures 5b and c. As shown in Figure 5b, there are two mutually perpendicular mirror planes in the CBED pattern, which are respectively parallel to the [001] and [−110] directions, suggesting the existence of mm2 symmetry (orthorhombic phase).25−28 As shown in Figure 5c, only one mirror plane//[1−10] was observed in the CBED pattern which matches the 3m symmetry feature (rhombohedral phase).25−28 Therefore, all three ferroelectric phases involving tetragonal, orthorhombic, and rhombohedral were observed in E

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 8. FE-SEM images of the ceramics with (a) x = 0, (b) x = 0.01, (c) x = 0.02, (d) x = 0.04, (e) y = 0, (f) y = 0.02, (g) y = 0.035, and (h) y = 0.05.

increase in Raman shifts indicates a shortening of the distance between Nb5+ and its coordinated oxygen caused by the incorporation of a small amount of Bi0.5Na0.5ZrO3. On the contrary, a lengthening in distance can be confirmed by a decrease in Raman shifts, and the anomaly of the change of Raman shifts suggests a phase transition happens. It is noted that the ν1 model of the ceramics with y ≥ 0.035 inversely decreases, implying the lattice stretches along the [100] and [011] directions, directly resulting in a shift of phase structure to the R phase. In addition, variations in full width at half maximum (fwhm) reflect the crystallization of the ceramics, which is consistent with the densification observed in scanning electron microscopy (SEM) images. When the changes in Raman shifts with the variations of y in Figure 7d are systematically analyzed, R-O-T can be defined for 0.025 ≤ y ≤ 0.04 because the Raman shifts exhibit a unique change in the composition range. Notably, the Raman shift of ν1 has nearly zero change for 0.025 < x < 0.035, and ν5 increases obviously for x = 0.03, suggesting the lattice mainly collapses along the [100] and [011] direction, while lattice deformation in the (110) plane remains nearly unchanged. The results further demonstrate R-O-T for 0.025 ≤ y ≤ 0.04, which agree with a previous analysis of phase structure. Furthermore, the fittings of Raman spectra of the ceramics with y were conducted using the Lorentzian method (Figures 7e−h). According to the previous reference, the observed two models are Eg (υ2) and A1g (υ1), respectively.31−34 A1g (υ1) first increases and then decreases with an increase in y values (Figure 7i), showing a similar change tendency with ν1, which can further prove the existence of R-O-T for the ceramics with 0.025 ≤ y ≤ 0.04. Subsequently, we investigated the microstructure evolutions of the ceramics with different x and y values by characterizing FE-SEM images. Figures 8a−d show the FE-SEM images of the ceramics with different x values. It was found that a small amount of BaSnO3 can promote the grain growth of KNNS ceramics (Figures 8a and b) and that their grain sizes dramatically drop as BaSnO3 content further inceases. Similar phenomenon can be observed in Sn-doped BaTiO3-based ceramics.19,23,24,35 More interestingly, the ceramics show a typical bimodal grain size distribution after being doped with BaSnO3 (Figure 8b), and small grains fill in the gaps of large ones. Figures 8e−h show the FE-SEM images of the ceramics with different y values. When the ceramics are doped with BNZ, their grain sizes can be effectively promoted, and then the small grains almost disappear (Figures 8e−g). However, a refined grain size can be found in the ceramics with y = 0.05 when the

nanodomain structure also can be responsible for the enhanced d33. Figures 6a and b show the εr−T curves of the ceramics measured at 60−500 °C and 100 kHz. We can see from Figures 6a and b that there are different effects of the additives on the TC values of the ceramics. In addition, some diffused phase transitions can be clearly observed in the ceramics with x ≥ 0.02 and y ≥ 0.05, as shown in Figures 6a and b. It may be thought that doping with Sn4+ or Zr4+ was also responsible for their diffused phase transition.23,24 The addition of BaSnO3 slightly decreases the TC values of the ceramics, which is similar to the previously reported results for BaTiO3 and BaSnO3.19,20 However, their TC values gradually decreased by being doped with BNZ (x ≤ 0.04) and then greatly decreased for x > 0.04.1,2,6 The phase diagrams can be used to clearly address the phase evolution. Figures 6c and d show the phase diagrams of the ceramics with varied x and y values. According to Figure 6c, their TC and TO‑T values are slightly dependent on x content, and their TR‑O increases as x content increases.19,20 Such a phase diagram also shows that the ceramics with x ≤ 0.02 possess an O phase at room temperature and that new phase boundaries cannot be achieved by being doped with only BaSnO3. However, there is a different phenomenon for the ceramics with different y values (Figure 6d). After being doped with BNZ, their TC, TO‑T, and TR‑O values were strongly dependent on the y content; that is, their TC and TO‑T values decrease and TR‑O increases as the y values increase. Finally, the R-O-T phase boundary of the ceramics can be tuned to near room temperature by when the y values are changed. As a result, the phase boundaries can be identified according to our analysis as shown here: (i) O phase for x ≤ 0.02 and cubic phase for x = 0.04, (ii) O phase for 0 ≤ y ≤ 0.01, (iii) O-T phase boundary for 0.01< y < 0.025, (iv) R-O-T phase boundary for 0.025 ≤ y ≤ 0.04, and (v) R phase for y = 0.05− 0.06. It is well-known that Raman spectroscopy can be used to characterize the phase structure of a material because it is very sensitive to the change of combination bonds directly resulting from molecular vibrations.22,31−34 Here, we measured the composition dependence of Raman spectra of the ceramics at room temperature. As shown in Figure 7a, all modes of 1A1g(ν1) + 1Eg(ν2) + 2F1u(ν3,ν4) + F2g(ν5) + F2u(ν6) can be observed in a wide range of 200−900 cm−1, which corresponds to the vibrations of NbO6 octahedra.22,31−34 Besides, ν5 (concerning bending modes) and ν1 (concerning stretching modes) were investigated, as shown in Figures 7b and c. The F

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 9. Element mapping of the ceramics with different (a) x and (b) y values. (c) energy-dispersive X-ray spectroscopy analysis.

Figure 10. Evolution of ferroelectric properties of the ceramics with different (a) x and (b) y values and their corresponding Pr and EC values (c and d) against the compositions, respectively.

0.035 shown in Figure 9. These data confidently confirms the existence of all elements in the two ceramics. Figures 10a and b show the ferroelectric properties (P−E curves) of the ceramics, where all samples were measured at 10 Hz at room temperature. From Figures 10a and b, their ferroelectric properties are strongly dependent on the actual compositions; that is, the ceramics with x > 0.02 or y > 0.04 show very poor ferroelectric behavior together with a slim P−E loop. The ferroelectric properties almost disappear for x = 0.04 (Figure 10a), confirming the existence of a cubic phase. To clearly demonstrate the composition-dependent ferroelectric properties, their Pr and EC values against x and y are shown in Figures 10c and d. Similar Pr values can be observed in the ceramics with x ≤ 0.015 or y ≤ 0.035. However, there is an obvious difference in Pr values, which are dependent on x and y values. One can see that the transition points of the ceramics with different x and y values are 0.015 and 0.04, respectively. In addition, we can also observe from Figures 10c and d that the addition of y (≤ 0.035) cannot degrade the ferroelectric properties of the ceramics. However, their EC values show a decreasing trend almost similar to the increase in x or y values.

BNZ contents further increase (Figure 8h). It is usually accepted that there is a close relationship between microstructure and electrical properties of a material. First, the refined grain sizes can result in broader εr−T curves of the ceramics, and then the piezoelectricity will be degraded. Second, large grain sizes also affect the electrical properties of a material;1 that is, a large grain size helps promote the piezoelectric activity of a material. As a result, a larger grain size in this work also contributes to the improvement of d33; that is, a very poor d33 is associated with a small grain size. In this work, many elements were used to modify the ceramics, and it was necessary to show whether the involved elements were unevenly distributed in the ceramic matrix. Figure 9 shows the element mapping of the ceramics with x = 0.01 and y = 0 as well as x = 0.01 and y = 0.035. From these figures, we note that K, Na, Nb, Sb, O, Ba, and Sn were homogeneously distributed. Even if BNZ is doped, the state of homogeneous distributions still could not be changed; that is, the elements (e.g., K, Na, Nb, Sb, O, Ba, Sn, Bi, and Zr) were well-distributed. We also provide the EDS patterns of the ceramics with x = 0.01 and y = 0 as well as x = 0.01 and y = G

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 11. Dielectric constant and dielectric loss of the ceramics as a function of (a) x and (b) y.

Figure 12. Evolution of piezoelectric properties of the ceramics with different (a) x and (b) y values.

Figure 13. Evolution of εrPr and d33 of the ceramics with different (a) x and (b) y values.

the ceramics. As shown in Figure 12a, d33 gradually increases and then linearly decreases as x increases, reaching a maximum value for x = 0.01. Their kp value shows a similar changed behavior. However, the addition of BNZ strongly affects piezoelectric activity of the ceramics. One can see from Figure 12b that their d33 value linearly increases and then dramatically decreases as y content increases; a peak d33 value was found in the ceramics with y = 0.035. In this work, we confirmed the existence of an intermediate O phase between T and R in the phase boundary region,18,36 and a large d33 value in this work was shown to be a result of the intermediate orthorhombic phase. In the past, we analyzed the underlying physical mechanism for the enhanced piezoelectric properties when the compositions are located at R-T phase boundary.1,2 It is well-accepted that the R-T phase boundary can contribute to the enhancement of d33.1,2,6 Previously, it was confirmed that the involvement of an intermediate phase (i. e., monoclinic or orthorhombic) is responsible for a large d33 value in both leadfree and lead-based materials in the vicinity of the R-T phase boundary.16−18,37 Generally, the intermediate phase could be

Figure 11 shows the composition dependence of dielectric properties of the ceramics measured at 100 kHz at room temperature. Figure 11a gives the dielectric properties vs x values of the ceramics. Their dielectric constant (εr) gradually increases as x continues to increase, while their dielectric loss (tan δ) remains almost unchanged for x ≤ 0.02 and then dramatically increases for x > 0.02. The increase in εr in this work should be due to the BaSnO3 doping.19,23,24 However, there are different trends in the ceramics with different y values. As shown in Figure 11b, εr reaches a peak value for x = 0.035 and decreases when the compositions deviate to 0.035. With the addition of y, their tan δ value almost decreases. Comparing Figure 11a with Figure 11b, we find that their εr increases from 708 to 2803 as y is doped, showing that the addition of BNZ will greatly promote εr of KNNS-BaSnO3 (BS) ceramics. As shown in Figures 10 and 11, their dielectric properties can be modified by doping both x and y, and their ferroelectric properties can be maintained by controlling the x and y values. Subsequently, we discussed the composition-dependent piezoelectric activity of the ceramics. Figures 12a and b show the composition-dependent piezoelectric properties (d33 and kp) of H

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 14. Evolution of strain hysteresis of the ceramics with different (a) x and (b) y values and their corresponding strain values (c and d) against the compositions.

induced by the external effect, such as electric field and composition design.16−18,37 The reasons why the intermediate phase can enhance piezoelectric properties were mainly explained by two opinions. (i) The intermediate phase (i. e., O phase) induced by composition design can contribute to the construction of the R-O-T phase boundary that consisted of RO and O-T phase boundaries, which enhances the piezoelectric properties.18,38 (ii) The intermediate phase (i. e., monoclinic or orthorhombic) induced by the applied electric field could bridge the R and T phases by modifying the polarization rotation path along the R-intermediate phase-T. The intermediate has a low symmetry, which could be attributed to enhanced piezoelectric properties.45−48 In the present work, a seemingly R-T phase boundary was obtained by co-doping the ceramics with BaSnO3 and Bi0.5Na0.5ZrO3, and the CBED patterns (Figure 5) disclosed the existence of the O phase induced by composition design, which demonstrated the R-T phase boundary was constructed by the intermediate O phase as a bridge. As a result, the R-O-T phase boundary may play an important role in the piezoelectric activity of this work. The equation εrPr ∝ d33 is also used to unveil the underlying physical mechanisms for a high piezoelectricity in the KNNbased ceramics.6,10 It can be found that this equation can illuminate the relationships among the piezoelectric, dielectric, and ferroelectric properties of a material. In this work, we gave the curves of εrPr against x and y in the ceramics with different x and y values, as shown in Figures 13a and b. A similar phenomenon can be observed in εrPr or d33 against x and y. As a result, it was thought that the enhancement in εrPr played a part in the improvement of d33. As far as d33 is concerned, the microstructure cannot be neglected. A larger grain can be developed in the ceramics with x = 0.01 or y = 0.035. Previously, it was reported that the ferroelectric domain variations of a material would be limited as their grain sizes decrease,1,39,40 and the coupling effects between grain boundaries and domain walls can also be strengthened due to the decreased grain sizes, thus prohibiting the motions of the domain wall. As a result, a larger grain can rotate the domains of the ceramics as compared with a smaller grain, which benefits the piezoelectric activity.

Figures 14a and b address the composition dependent strain hysteresis in the ceramics, measured at 10 Hz and room temperature. From Figures 14a and b, one can see that the ceramics with x = 0.01 or y = 0.04 possess a higher strain as compared with other compositions. To clearly address the changed trends in strain, we plotted the strain values against the compositions of the ceramics with different x and y, as shown in Figures 14c and d. A peak strain value can be observed in the ceramics with x = 0.01−0.02 or y = 0.04. For BS-doped ceramics, an enhanced strain may be due to the introduction of optimum Ba and Sn elements.19,23,24,35 In the past, the multiphases coexistence can largely contribute to the improvement of strain value in KNN ceramics.1,41,42 In this work, the ceramics with x = 0.04 possess an R-O-T phase boundary, which may result in an enhanced strain value. There is some controversy concerning the effects of poling conditions on piezoelectric properties of KNN-based ceramics,43,44 although a large amount of attention has been focused on these effects. In the past, some authors have investigated the relationships between poling conditions and the O-T phase boundary of KNN-based ceramics.43 Recently, new phase boundary was constructed in KNN-based ceramics;1 however, there are few systematic reports on the influence of poling conditions on the piezoelectric activity of KNN-based ceramics. In this work, we investigated the relationships between poling conditions (temperature and electric field) and d33 in the ceramics with x = 0.01 and y = 0.035. According to our experiments, a lower (0.5 kV/cm) or higher (4.0 kV/ cm) electric field cannot warrant a large d33 value, and a high d33 of ∼480 pC/N can be observed only when the applied electric field is 1.5 kV/cm. In addition, d33 is sensitive to their poling temperatures; that is, d33 dramatically drops from 480 to 230 pC/N when the poling temperatures increase from 20 to 120 °C. In the past, the changes in XRD patterns during the poling process were found in some materials.45−48 In this work, we measured the XRD patterns of the ceramics under different electric fields, as shown in Figure 15. The reorientation of the diffraction peaks is induced along the applied electric field direction. As shown in Figure 15, it is difficult to clearly identify some new phases, but the diffraction peaks positions were I

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

ACS Applied Materials & Interfaces



REFERENCES

(1) Wu, J.; Xiao, D.; Zhu, J. Potassium−Sodium Niobate Lead-Free Piezoelectric Materials: Past, Present, and Future of Phase Boundaries. Chem. Rev. 2015, 115 (7), 2559−2595. (2) Wu, J.; Xiao, D.; Zhu, J. Potassium-Sodium Niobate Lead-Free Piezoelectric Ceramics: Recent Advances and Perspectives. J. Mater. Sci.: Mater. Electron. 2015, 26 (12), 9297−9308. (3) Gao, Y.; Zhang, J.; Qing, Y.; Tan, Y.; Zhang, Z.; Hao, X. Remarkably Strong Piezoelectricity of Lead-Free (K0.45Na0.55)0.98Li0.02(Nb0.77Ta0.18Sb0.05)O3 Ceramic. J. Am. Ceram. Soc. 2011, 94 (9), 2968−2973. (4) Zheng, T.; Wu, J.; Cheng, X.; Wang, X.; Zhang, B.; Xiao, D.; Zhu, J. Wide Phase Boundary Zone, Piezoelectric Properties, and Stability in 0.97(K0.4Na0.6) (Nb1‑xSbx)O3 −0.03 Bi0.5Li0.5ZrO3 Lead-free Ceramics. Dalton. Trans. 2014, 43, 9419−9426. (5) Cheng, X.; Wu, J.; Wang, X.; Zhang, B.; Zhu, J.; Xiao, D.; Wang, X.; Lou, X. Giant d33 in (K,Na) (Nb,Sb)O3-(Bi,Na,K,Li)ZrO3 Based Lead-Free Piezoelectrics with High TC. Appl. Phys. Lett. 2013, 103, 052906. (6) Wang, X.; Wu, J.; Xiao, D.; Zhu, J.; Cheng, X.; Zheng, T.; Zhang, B.; Lou, X.; Wang, X. Giant Piezoelectricity in Potassium-Sodium Niobate Lead-free Ceramics. J. Am. Chem. Soc. 2014, 136 (7), 2905− 2910. (7) Cheng, X.; Wu, J.; Wang, X.; Zhang, B.; Zhu, J.; Xiao, D.; Wang, X.; Lou, X.; Liang, W. Lead-free Piezoelectric Ceramics Based on (0.97-x)K 0.48 Na 0.52 NbO 3 −0.03 Bi 0.5 (Na 0.7 K 0.2 Li 0.1 ) 0.5 ZrO 3 xB0.5Na0.5TiO3 Ternary System. J. Appl. Phys. 2013, 114, 124107. (8) Cheng, X.; Wu, J.; Lou, X.; Wang, X.; Wang, X.; Xiao, D.; Zhu, J. Achieving Both Giant d33 and High TC in Patassium-Sodium Niobate Ternary System. ACS Appl. Mater. Interfaces 2014, 6 (2), 750−756. (9) Guo, Y.; Kakimoto, K.; Ohsato, H. Phase Transitional Behavior and Piezolectric Properties of (Na0.5K0.5)NbO3-LiNbO3 Ceramics. Appl. Phys. Lett. 2004, 85, 4121−4123. (10) Shrout, T. R.; Zhang, S. J. Lead-free Piezoceramics: Alternatives for PZT? J. Electroceram. 2007, 19, 113−126. (11) Hollenstein, E.; Davis, M.; Damjanovic, D.; Setter, N. Piezoelectric Properties of Li-and Ta-modified (K0.5Na0.5)NbO3 Ceramics. Appl. Phys. Lett. 2005, 87, 182905. (12) Lin, D.; Kwok, K. W.; Lam, K. H.; Chan, H. L. Structure and Electrical Properties of K0.5Na0.5NbO3-LiSbO3 Lead-Free Piezoelectric Ceramics. J. Appl. Phys. 2007, 101, 074111. (13) Akdoğan, E. K.; Kerman, K.; Abazari, M.; Safari, A. Origin of High Piezoelectric Activity in Ferroelectric (K0.44Na0.52Li0.04)(Nb0.84Ta0.1Sb0.06)O3 Ceramics. Appl. Phys. Lett. 2008, 92, 112908. (14) Rödel, J.; Jo, W.; Seifert, K.; Anton, E. M.; Granzow, T.; Damjanovic, D. Perspective on the Development of Lead-Free Piezoceramics. J. Am. Ceram. Soc. 2009, 92 (6), 1153−1177. (15) Zuo, R.; Fu, J. Rhombohedral-Tetragonal Phaes Coexistence and Piezoelectric Properties of (Na,K) (Nb,Sb)O3-LiTaO3-BaZrO3 Lead-Free Ceramics. J. Am. Ceram. Soc. 2011, 94 (5), 1467−1470. (16) Noheda, B.; Gonzalo, J. A.; Cross, L. E.; Guo, R.; Park, S. E.; Cox, D. E.; Shirane, G. Tetragonal-to-Monoclinic Phase Transition in a Ferroelectric Perovskite: the Structure of PbZr0.52Ti0.48O3. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61, 8687−8695. (17) Bokov, A. A.; Ye, Z. G. Domain Structure in the Monoclinic Pm Phase of Pb(Mg1/3Nb2/3)O3-PbTiO3 Single Crystals. J. Appl. Phys. 2004, 95, 6347. (18) Keeble, D. S.; Benabdallah, F.; Thomas, P. A.; Maglione, M.; Kreisel, J. Revised Structural Phase Diagram of (Ba0.7Ca0.3TiO3)(BaZr0.2Ti0.8O3). Appl. Phys. Lett. 2013, 102, 092903. (19) Yao, Y. G.; Zhou, C.; Lv, D. C.; Wang, D.; Wu, H. J.; Yang, Y. D.; Ren, X. B. Large Piezoelectricity and Dielectric Permittivity in BaTiO3-xBaSnO3 System: the Role of Phase Coexisting. EPL 2012, 98, 27008. (20) Zhu, L.; Zhang, B.; Zhao, L.; Li, J. High Piezoelectricity of BaTiO3−CaTiO3−BaSnO3 Lead-Free Ceramics. J. Mater. Chem. C 2014, 2 (24), 4764−4771. (21) Qin, Y.; Zhang, J.; Yao, W.; Lu, C.; Zhang, S. Domain Configuration and Thermal Stability of (K0.48Na0.52) (Nb0.96Sb0.04)O3-

Figure 15. XRD patterns of the ceramics poled under different (a) electric fields and (b) temperatures.

shifted when the external electric fields (10−20 kV/cm) were applied. As we know, the reorientation of domains (i.e., 90° domain) plays an important role in the poling process, which can be responsible for the changes in XRD patterns under electric fields.45−48 As a result, it may be thought that the changes in d33 with electric fields are intrinsically ascribed to the reorientation of domain (i. e., 90° domain), and further investigation will be continued.



CONCLUSION In this work, we confirmed the phase compositions (R-O-T) of phase boundaries of the ceramics by CBED patterns and that a large d33 value of 480 pC/N can be attained by controlling x and y values. The phase boundary types can be controlled by modifying x and y values; that is, the ceramics with 0 ≤ x ≤ 0.015 and y = 0 maintain an O phase, and the R-O-T phase boundary is seen in the ones with x = 0.01 and y = 0.025−0.04. More importantly, their piezoelectric activity is strongly dependent on their corresponding compositions, and there is a leap in d33 from ∼165 to ∼480 pC/N as the phase transitions change from O to R-O-T. In the region of the R-O-T phase boundary, a large d33 of ∼480 pC/N was attained in the ceramics with x = 0.01 and y = 0.035. In addition, the ceramics with the R-O-T phase boundary (x = 0.01 and y = 0.04) have a high strain of ∼0.274%. We believe that the KNN ternary system can be used to promote piezoelectric activity by forming the R-O-T phase boundary.



Research Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] or [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors gratefully acknowledge support from the National Science Foundation of China (Grants 51102173, 51272164, 51332003, and 51472169) and the College of Materials Science and Engineering of Sichuan University. The authors thank Xiangjian Wang and Xiaojie Lou for measuring the strain properties of the ceramics and Ms. Hui Wang for characterizing the FE-SEM images of all ceramics. J

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces Bi0.50(Na0.82K0.18)0.50ZrO3 Piezoceramics with High d33 Coefficient. ACS Appl. Mater. Interfaces 2016, 8 (11), 7257−7265. (22) Klein, N.; Hollenstein, E.; Damjanovic, D.; Trodahl, H. J.; Setter, N.; Kuball, M. A. Study of the Phase Diagram of (K,Na,Li)NbO3 Determined by Dielectric and Piezoelectric Measurements, and Raman Spectroscopy. J. Appl. Phys. 2007, 102 (1), 014112. (23) Upadhyay, S. K.; Reddy, V. R.; Bag, P.; Rawat, R.; Gupta, S. M.; Gupta, A. Electro-Caloric Effect in Lead-Free Sn Doped BaTiO3 Ceramics at Room Temperature and Low Applied Fields. Appl. Phys. Lett. 2014, 105, 112907. (24) Shi, T.; Xie, L.; Gu, L.; Zhu, J. Why Sn Doping Significantly Enhances the Dielectric Properties of Ba(Ti1‑xSnx)O3. Sci. Rep. 2015, 5, 8606. (25) Zuo, J. M.; Spence, J. C. H. In Electron Microdiffraction; Springer Science & Business Media: Berlin, 2013. (26) Gao, J. H.; Xue, D. Z.; Wang, Y.; Guo, S. W.; Ren, X. B. Microstructure Basis for Strong Piezoelectricity in Pb-Free Ba(Zr0.2Ti0.8)O3-(Ba0.7Ca0.3)TiO3 Ceramics. Appl. Phys. Lett. 2011, 99, 092901. (27) Gao, J. H.; Zhang, L. X.; Xue, D. Z.; Kimoto, T.; Ren, X. B. Symmetry Determination on Pb-Free Piezoceramic 0.5Ba(Zr0.2Ti0.8)O3-0.5(Ba0.7Ca0.3)TiO3 using Convergent Beam Electron Diffraction Method. J. Appl. Phys. 2014, 115, 054108. (28) Gao, J. H.; Hao, Y. S.; Ren, S.; Kimoto, T.; Ren, X. B. Large Piezoelectricity in Pb-Free 0.96(K0.5Na0.5)0.95Li0.05Nb0.93Sb0.07O30.04BaZrO3 Ceramic: A Perspective from Microstructure. J. Appl. Phys. 2015, 117, 084106. (29) Woodward, D. I.; Knudsen, J.; Reaney, I. M. Review of Crystal and Domain Structures in the PbZrxTi1‑xO3 Solid Solution. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 72, 104110. (30) Wang, H.; Zhu, J.; Lu, N.; Bokov, A. A.; Ye, Z. G.; Zhang, X. W. Hierarchical micro-/nanoscale Domain Structure in MC Phase of (1x)Pb(Mg1/3Nb2/3)O3-xPbTiO3 Single Crystal. Appl. Phys. Lett. 2006, 89, 042908. (31) Rubio-Marcos, F.; Bañares, M. A.; Romero, J. J.; Fernández, J. F. Correlation between the Piezoelectric Properties and the Structure of Lead-Free KNN−Modified Ceramics, Studied by Raman Spectroscopy. J. Raman Spectrosc. 2011, 42 (4), 639−643. (32) Rubio-Marcos, F.; López-Juárez, R.; Rojas-Hernandez, R. E.; Del Campo, A.; Razo-Pérez, N.; Fernandez, J. F. Lead-Free Piezoceramics: Revealing the Role of the Rhombohedral−Tetragonal Phase Coexistence in Enhancement of the Piezoelectric Properties. ACS Appl. Mater. Interfaces 2015, 7 (41), 23080−23088. (33) Ramajo, L.; Castro, M.; del Campo, A.; Fernandez, J.; RubioMarcos, F. Revealing the role of Cationic Displacement in Potassium− Sodium Niobate Lead-Free Piezoceramics by Adding W6+ ions. J. Mater. Chem. C 2015, 3 (16), 4168−4178. (34) Gao, J.; Ren, S.; Zhang, L.; Hao, Y.; Fang, M.; Zhang, M.; Dai, Y.; Hu, X.; Wang, D.; Zhong, L. Phase Transition Sequence in Pb-Free 0.96(K0.5Na0.5)0.95Li0. 05Nb0.93Sb0.07O3−0.04BaZrO3 Ceramic with Large Piezoelectric Response. Appl. Phys. Lett. 2015, 107 (3), 032902. (35) Lin, D.; Kwok, K. W.; Chan, H. L. W. Structure, Dielectric and Piezoelectric Properties of Ba0.90Ca0.10Ti1‑xSnxO3 Lead-Free Ceramics. Ceram. Int. 2014, 40 (5), 6841−6846. (36) Lummen, T. T.; Gu, Y.; Wang, J.; Lei, S.; Xue, F.; Kumar, A.; Barnes, A. T.; Barnes, E.; Denev, S.; Belianinov, A. Thermotropic Phase Boundaries in Classic Ferroelectrics. Nat. Commun. 2014, 5, 3172. (37) Jaffe, B.; Cook, W. R.; Jaffe, H. In Piezoelectric Ceramics; Academic: New York, 1971. (38) Zhang, L.; Zhang, M.; Wang, L.; Zhang, L. X.; Xue, D. Z.; Ren, X. B. Phase Transitions and the Piezoelectricity around Morphotropic Phase Boundary in Ba(Zr0.2Ti0.8)O3-x(Ba0.7Ca0.3)TiO3 Lead-Free Solid Solution. Appl. Phys. Lett. 2014, 105, 162908. (39) Zhao, Z.; Buscaglia, V.; Viviani, M.; Buscaglia, M. T.; Mitoseriu, L.; Testino, A.; Nygren, M.; Johnsson, M.; Nanni, P. Grain-size Effects on the Ferroelectric Behavior of Dense Nanocrystalline BaTiO3 Ceramics. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 70, 024107.

(40) Randall, C. A.; Kim, N.; Kucera, J. P.; Cao, W. W.; Shrout, T. R. Intrinsic and Extrinsic Size Effects in Fine-Grained MorphotropicPhase-Boundary Lead Zirconate Titanate Ceramics. J. Am. Ceram. Soc. 1998, 81 (3), 677−688. (41) Hao, J.; Shen, B.; Zhai, J.; Liu, C.; Li, X.; Gao, X. Switching of Morphotropic Phase Boundary and Large Strain Response in LeadFree Ternary (Bi0.5Na0.5)TiO3−(K0.5Bi0.5)TiO3−(K0.5Na0.5)NbO3 System. J. Appl. Phys. 2013, 113, 114106. (42) Yao, F. Z.; Wang, K.; Jo, W.; Lee, J.; Li, J. Effect of Poling Temperature on Piezoelectricity of CaZrO3-modified (K,Na)NbO3based Lead-Free Ceramics. J. Appl. Phys. 2014, 116, 114102. (43) Yao, F. Z.; Wang, K.; Jo, W.; Lee, J.; Li, J. Effect of Poling Temperature on Piezoelectricity of CaZrO3-modified (K,Na)NbO3based Lead-Free Ceramics. J. Appl. Phys. 2014, 116, 114102. (44) Rubio-Marcos, F.; Romero, J. J.; Ochoa, D. A.; García, J. E.; Perez, R.; Fernández, J. F. Effects of Poling Process on KNN-modified Piezoceramic Properties. J. Am. Ceram. Soc. 2010, 93 (2), 318−321. (45) Ullah, A.; Malik, R. A.; Ullah, A.; Lee, D. S.; Jeong, S. J.; Lee, J. S.; Kim, I. W.; Ahn, C. W. Electric-Field-Induced Phase Transition and Large Strain in Lead-Free Nb-doped BNKT-BST Ceramics. J. Eur. Ceram. Soc. 2014, 34 (1), 29−35. (46) Iamsasri, T.; Tutuncu, G.; Uthaisar, C.; Wongsaenmai, S.; Pojprapai, S.; Jones, J. L. Electric Field-Induced Phase Transitions in Li-Modified Na0.5K0.5NbO3 at the Polymorphic Phase Boundary. J. Appl. Phys. 2015, 117, 024101. (47) Kovacova, V.; Vaxelaire, N.; Le Rhun, G.; Gergaud, P.; SchmitzKempen, T.; Defay, E. Correlation between Electric-Field-Induced Phase Transition and Piezoelectricity in Lead Zirconate Titanate Films. Phys. Rev. B: Condens. Matter Mater. Phys. 2014, 90, 140101. (48) Ge, W.; Li, J.; Viehland, D.; Chang, Y.; Messing, G. L. ElectricField-Dependent Phase Volume Fractions and Enhanced Piezoelectricity nea r the Polymorphic Pha se Boundary of (K0.5Na0. 5)1‑xLixNbO3 Textured Ceramics. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83 (22), 224110.

K

DOI: 10.1021/acsami.6b04288 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX