Domain Evolution and Piezoelectric Response ... - ACS Publications

Subscriber access provided by University of South Dakota

Article

Domain evolution and piezoelectric response across thermotropic phase boundary in (K,Na)NbO-based epitaxial thin films 3

Jin Luo, Wei Sun, Zhen Zhou, Yu Bai, Zhan Jie Wang, Guo Tian, Deyang Chen, xingsen gao, Fangyuan Zhu, and Jing-Feng Li ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b02263 • Publication Date (Web): 03 Apr 2017 Downloaded from http://pubs.acs.org on April 5, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Domain evolution and piezoelectric response across thermotropic phase boundary in (K,Na)NbO3-based epitaxial thin films Jin Luo,† Wei Sun,† Zhen Zhou,† Yu Bai, ‡ Zhan Jie Wang,‡ Guo Tian,§ Deyang Chen,§ Xingsen Gao,§ Fangyuan Zhu,⊥ and Jing-Feng Li*, † †

State Key Laboratory of New Ceramics and Fine Processing, School of Materials Science and

Engineering, Tsinghua University, Beijing 100084, China ‡

Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese

Academy of Sciences, Shenyang 110016, China §

Institute for Advanced Materials and Guangdong Provincial Key Laboratory of Quantum

Engineering and Quantum Materials, South China Normal University, Guangzhou 510006, China ⊥Shanghai

Institute of Applied Physics, Chinese Academy of Sciences, No.239 Zhangheng

Road, Pu-dong District, Shanghai 201204, China

KEYWORDS: lead-free, piezoelectric, domain evolution, phase transition, thermotropic phase boundary

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

ABSTRACT

Recent research progress in (K,Na)NbO3 (KNN) based lead-free piezoelectric ceramics has attracted increasing attention for their applications to microsystems or microelectromechanical systems (MEMS) in the form of thin films. This work demonstrates that high-quality KNN-based epitaxial films can be synthesized by a conventional sol-gel method, whose phase structure and domain characteristics have been investigated with emphasis on the temperature effect. A monoclinic MC structure is observed at room temperature in KNN-based epitaxial films, which is close to but different from the orthorhombic phase in bulk counterparts. Piezoresponse force microscopy (PFM) at elevated temperatures reveals continuous changes of ferroelectric domains in KNN films during heating and cooling cycles between room temperature and 190 ℃. A distinct change in domain morphology is observed upon heating to 110 ℃, accompanied with a clear variation of dielectric permittivity suggesting a thermotropic phase transition, which is revealed to belong to a MC - MA phase transition on the basis of structural and PFM analysis on local ferroelectric and piezoelectric behaviors. Enhanced piezoelectric response at the thermotropic phase boundary is observed, which is attributed to active domains and/or nanodomains formed across the boundary. Domain engineering by utilizing the phase transition should be important and effective for property enhancement not only in KNN-based films but also for its textured ceramics. INTRODUCTION High-performance piezoelectric materials, which allow effective conversion of mechanical strain and electrical signal, play a pivotal role in medical imaging, telecommunication and ultrasonic devices.1-2 Excellent piezoelectric properties are usually observed in the Pb(Zr,Ti)O3


2

Page 3 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


system due to the effect of morphotropic phase boundary (MPB), where two competing ferroelectric phases exist. Characteristics of the MPB have been mimicked in lead-free counterparts to achieve competitive piezoelectric properties by chemical modification or epitaxial strain.3-8 Among them, (K,Na)NbO3 (KNN)-based ceramics are particularly a promising candidate as lead-free piezoelectric materials, which attract extensive attention since Saito et al. reported that highly textured KNN-based ceramics exhibited high d33 value over 400 pC/N in 2004.3 KNN-based ceramics with high piezoelectric properties exhibit the thermotropic phase boundary (TPB) at room temperature,9-11 which is thermally induced (‘thermo-’) phase transition instead of compositional (‘morpho-’) phase boundaries.2 For example, a large d33 up to 490 pC/N was achieved in KNN-based ceramics by obtaining a rhombohedral - tetragonal phase boundary at room temperature through shifting both rhombohedral - orthorhombic and orthorhombic tetragonal phase transition temperature.10,

12

Recently, a giant d33 as high as 570 pC/N was

reported in the KNN-based ceramic by nanodomain engineering at the rhombohedral - tetragonal phase boundary, demonstrating promising prospective for the future replacement of lead-free based counterparts.13 Understanding the nature of the TPB in KNN-based materials is a very difficult yet interesting work. Most literatures are focused on KNN-based bulk materials. However, KNNbased films have received increasing attentions as one of promising alternatives in various applications including sensors, actuators, energy harvesting systems and microelectromechanical systems (MEMS) because of the high piezoelectric properties in their bulk counterparts.14-19 Previously, KNN-based films with good ferroelectric and piezoelectric properties on Pt/Ti/SiO2/Si wafers or single crystal SrTiO3 (STO) substrates have been synthesized by sol-gel method and pulsed laser deposition method.14-15, 20 However, epitaxial thin films grown on single


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

crystalline substrates are technically important to exhibit high performance.14, 21 Compared with bulk ceramics and polycrystalline thin films, where the presence of the grain boundary and random orientation of the grains makes it complicated for the study on the microscopic physical properties, the epitaxial thin films have a similar structure to the single crystals with a preferred orientation, which can be regarded as ideal specimens for studying the mechanism of high piezoelectricity at the TPB. The domain structure as an important physical parameter in piezoelectric materials plays an essential role in piezoelectric properties. Additionally, domain wall movement and the change of domain structure at TPB are correlated with the coexistence of the two phases, which are responsible for enhanced piezoelectric properties due to substantially increased ferroelectric variants. While extensive researches have investigated the effects of chemical modifications on piezoelectric properties, more detailed investigation has been devoted to the domain evolution of KNN across the TPB, despite the importance for the fundamental physics of piezoelectric materials and the origins of high piezoelectricity. In this work, the phase structure, domain evolution and piezoelectric property, especially across the TPB, were investigated in Li-doped (K,Na)NbO3/STO (001) thin films synthesized by the sol-gel method. The KNN-based thin films demonstrated high-quality epitaxial characteristics and exhibited monoclinic MC structure instead of an orthorhombic structure due to the elastic strain imposed by the substrate at room temperature. The thermally induced domain evolution investigated using piezoresponse force microscopy (PFM) in KNN-based epitaxial thin films suggested a kind of monoclinic MC - MA phase transition. The active domains and nanodomains across the thermotropic phase transition are closely associated with the enhanced piezoelectric response.


4

Page 5 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


EXPERIMENTAL SECTION Lead-free (Na0.47K0.47Li0.06)NbO3 epitaxial thin films were prepared by a sol-gel method having the advantages of low cost and high chemical homogeneity. Raw materials of sodium ethoxide (NaOC2H5), potassium ethoxide (KOC2H5), lithium acetylacetonate (LiO2C5H7), 2methoxyethanol solvent, acetylacetone as chelating agent and acetic acid were mixed in the stoichiometric ratio and refluxed at 106 ℃ for 1 h. Niobium pentaethoxide (NbO5C10H25) was added to the resultant solution and refluxed at 106 ℃ for another one hour. 10 mol% Na and K excess were introduced to the precursor solution to compensate the loss caused by the volatilization of the elements22. The concentration of the precursor of KNN solution is 0.4 M. After being aged for one day, the precursor solution was spin-coated on (001) orientated Nb:STO single crystal substrates at a speed of 4000 rpm for 30 s. Each layer of the film was dried on a hotplate at 150 ℃ for 30 s and pyrolyzed at 450 ℃ for 3 mins. The deposition-dryingpyrolysis cycle was repeated five times and the film was annealed at 800 ℃ for 3 mins. X-ray diffraction (XRD, D/max-RB, Rigaku; Tokyo, Japan) using Cu Kα radiation was carried out to analyze the crystallographic structure and orientation of the films. Reciprocal space mapping (RSM) was performed to examine the epitaxial growth and the crystal structure of the thin film using beam line 14B1 (λ = 1.2378 Å) at Shanghai Synchrotron Radiation Facility and by X-ray diffraction (X′Pert PRO, PANalytical, Holland). The cross section and the crystal structure of the epitaxial thin film were analyzed using the transmission electron microscope (TEM, F20, Tecnai, Hillsboro, Oregon). The surface morphology and the domain structure were investigated using a piezoresponse force microscope (PFM, MFP-3D, Asylum Research, USA) with a Pt and Ir-coated cantilever (NanoWorld EFM, force constant 2.8 N/m, resonance


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

frequency 75 KHz). The piezoelectric response was analyzed using switching spectrometry PFM (SSPFM). Platinum electrodes with diameter 0.4 mm were sputtered for the dielectric measurement. The dielectric properties as a function of temperature were measured from room temperature to 145 ℃ using a semiconductor device analyzer (B1500A, Agilent, USA) connected with a probe system (Summit 11000, Cascade Microtech, USA). RESULTS AND DISCUSSION The XRD pattern of Li-doped KNN thin films on the Nb:STO (001) substrate is shown in Figure 1 (a). Only diffraction peaks of (001), (002) and (003) crystal planes were observed, which indicate the possibility of epitaxial relationship between the film and the substrate. The diffraction angles of the Li-doped KNN film were lower than the corresponding angles of the STO substrate, which reveal a larger lattice parameter of the film than that of the substrate along the [001] direction. The full widths at half-maximum (FWHMs) of the (001) and (002) peaks in the rocking curves of the thin films were estimated to be 0.17° and 0.40° respectively, and these small FWHMs demonstrated good crystallization of Li-doped KNN films along the out-of-plane orientation. From the (001), (002) and (003) peaks of Li-doped KNN films, an averaged c-lattice parameter of 4.012 Å is estimated, which is close to the lattice parameter in KNN-based bulk ceramics (4.00 Å).23 Phi scan of (110) peaks of the thin film was performed to analyze the in-plane orientation as shown in Figure 1 (b). The (110) peaks of the KNN-based thin film demonstrated 4-fold symmetry with an equal 90° interval, which confirms the same preferential in-plane orientation of the KNN-based thin film as the substrate and the epitaxial growth of the thin film. The


6

Page 7 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


FWHMs of the (110) peak in the rocking curve was 0.42°, which also indicates the good crystallinity of the film along the in-plane orientation.

Figure 1. (a) The XRD pattern of Li-doped KNN epitaxial thin films on the Nb:STO (001) substrate; (b) Phi scan of (110) peaks of Li-doped KNN epitaxial thin films; (c) and (d) are the RSMs of 103pc and 013pc peaks for the Li-doped KNN thin film respectively. In order to further analyze the structure of the thin films, reciprocal space mappings were also carried out. Figure 1 (c) and (d) show the X-ray RSMs of 103pc and 013pc peaks for the Lidoped KNN thin film. Figure S1 shows the synchrotron RSMs of 103pc, 113pc and 002pc family peaks for the Li-doped KNN thin film. RSMs of 002pc, 103pc and 113pc family peaks for the


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

Li-doped KNN thin film all showed single broadened spots, with characteristic diffraction patterns of epitaxial thin films. RSMs of the KNN-based thin films shifted to the lower reciprocal space coordinates, indicating compressive stress between the film and the substrate. The lattice parameters of Li-doped KNN thin films, a-KNN, b-KNN and c-KNN, were estimated to be 3.954 Å, 3.949 Å and 4.018 Å respectively, which are close to but different from that of the orthorhombic phase in KNN-based bulk ceramics (a=c=4.00 Å, b=3.94 Å).23 Since a=c for the perovskite type unit cell of KNN-based bulk ceramics, the unit cell has an orthorhombic symmetry.24 However, these three lattice parameters were no longer equal in KNN-based epitaxial thin films, which suggests epitaxial KNN-based thin films are not orthorhombic phase. Since the lattice mismatch along horizontal a-axis and b-axis in the (001) plane were 1.24% and 1.11% respectively in KNN-based epitaxial thin films, the elastic strain between the thin film and the substrate could lead to a different phase. In order to further analyze the fundamental information on crystal structure of the thin film, the transmission electron microscopy observation was conducted at room temperature as shown in Figure 2. Figures 2 (a) – (c) show both high resolution and low resolution TEM of the interface between the thin film and the substrate. The low resolution TEM image in Figure 2 (a) demonstrates flat and sharp interface between the thin film and the substrate, and the thickness of the thin film is estimated to be 200 nm. The high resolution TEM images of the interface clearly demonstrate epitaxial growth of the thin film with rigid lattice matching and the same orientation as the substrate. The insects of Figures 2 (b) and (c) show the fast Fourier transformation for the thin film and the substrate along different zone axes. The film exhibited the same electron pattern as the substrate along the (100) zone axis. However, supperlattice spots were observed in the [101] direction along the zone axis (010) in the thin film as marked by red cycles as shown in the


8

Page 9 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


insect of Figure 2 (c), which indicates that the crystal structure of the thin film is monoclinic MC phase with the pseudocubic cell shearing in the (010) or (100) plane. From the fast Fourier transformation along the (100) and (010) zone axes, the lattice parameters of the monoclinic pseudocubic unite cell can be estimated: a-KNN=3.957 Å; b-KNN=3.951 Å; c-KNN=4.001 Å, which is consistent with that determined by RSM results. The big difference of monoclinic phase in the epitaxial KNN-based thin film is that a-KNN is no longer equal to c-KNN compared with the orthorhombic phase,23 which could arise from the elastic strain imposed by the substrate.

Figure 2. (a) The low resolution TEM cross section image of the thin film, showing the interface between the thin film and the substrate. The scale bar is 100 nm. (c) and (d) the high resolution TEM cross section images of the thin film on the STO substrate taken along the 100 and 010 zone axes respectively, and the insect images show the fast Fourier transformation corresponding to the thin film and the substrate. The scale bar is 5 nm; In addition to crystal structure, the piezoresponse of epitaxial thin films is closely related with the domain structure25, which can be revealed by PFM. Both the surface topography and domain structure of the Li-doped epitaxial thin film were obtained using PFM technique as shown in Figure S2. The root-mean-square roughness was estimated to be 1.9 nm over a square


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

area of 2×2 µm2, indicating a high quality of the thin film synthesized by the sol-gel method. Domain structures were revealed in the amplitude and phase images along both in-plane (IP) and out-of-plane (OP) directions as shown in Figure S2 (b) - (e). Both OP and IP amplitude images exhibited obvious contrast and the differences of the phase angles in both IP and OP phase images were close to 180°, being indicative of the existence of different polarization directions. Since the internal bias field or aligned defects may cause no piezoresponse or small piezoresponse along the upward or downward direction in OP PFM mapping obtained by a small tip bias,26 the domains with a relatively small amplitude in the OP amplitude image had a different phase as the domains with a high amplitude as shown in Figures S2 (d) and (e). Only the IP PFM amplitude and phase were measured in the heating and cooling cycle from 26 ℃ to 190 ℃; OP PFM amplitude and phase measurement were not conducted because it is too sensitive to the surface topography especially at high temperatures. Figure 3 shows the IP PFM amplitude evolution with temperature, and corresponding temperature dependence of IP PFM phase is presented in Figure 4. No obvious changes were observed in the IP PFM amplitude and phase images from 26 ℃ to 90OC. Parallel stripe domains with domain wall intersecting along the [110] direction in the (100) plane were observed at 26 ℃ and 90 ℃. When the temperature reached 110 ℃, parallel stripe domain walls, used to be parallel at T < 110 ℃, became curved and some became crossed at a point as marked by red arrows in Figure 3 (c) and Figure 4 (c), indicating the domain wall movement and the change of domain structure at 110 ℃. Dramatic changes occurred in domain structure including both amplitude and phase at 130 ℃ and 150 ℃. The parallel stripe domains completely disappeared at 130 ℃, and another change was that the area with yellow color in the phase mapping shrank


10

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


obviously. Additionally, a different domain wall intersecting with the (100) plane along the [100] direction showed up in both amplitude and phase images at 150 ℃ as shown in Figure 3 and Figure 4, implying emergence of a new phase. The domain walls along the [100] direction still existed at temperature up to 190 ℃ and remained at 150 ℃ when cooled from 190 ℃ as shown in Figure 3 (h) and Figure 4 (h). Meanwhile, the small domains with size less than 100 nm in width coalesced into larger domains at 170 ℃ as marked by red squares in Figure 4, resulting in greatly decreased nanodomains.

Figure 3. (a) – (l) The IP amplitude of Li-doped KNN epitaxial thin films during heating and cooling cycles from 26 ℃ to 190 ℃. Scan size = 1 µm.


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

Figure 4. (a) - (l) The IP phase of Li-doped KNN epitaxial thin films during heating and cooling cycles from 26 ℃ to 190 ℃. Scan size = 1 µm. Dramatic changes occurred when the thin film was cooled from 130 ℃ to 110℃, indicating active domain wall movement at 130 ℃. Moreover, the domain wall intersecting the (100) plane along the [110] direction reappeared and the area with yellow color increased substantially at 110 ℃. The domain structures gradually changed from 110 ℃ to 26 ℃ during the cooling and the domain structure at room temperature was similar to the original one before the heating and cooling cycle.


12

Page 13 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. (a) The ratios of the two different phases in the IP phase mappings of Li-doped KNN epitaxial thin films at various temperatures; (b) Temperature dependence of relative dielectric constant of Li-doped KNN epitaxial thin films; (c) X-ray diffraction map / (ω - θ , 2θ ) of (002) peaks for Li-doped KNN epitaxial thin films at 150 ℃. The ratios of the yellow colored area over that with purple color as a function of temperature are shown in Figure 5 (a). The ratio increased abruptly after 110 ℃, which confirms that 110 ℃ is the turning point for the phase transition. And then the ratio became stabilized at 150 OC, which implies that the domain structure at 130 ℃ was the most active. Figure 5 (b) shows the temperature dependence of the dielectric constant at 1 kHz from 26 ℃ to 140 ℃. Only one phase transition was observed at 108 ℃, which is consistent with the phase transition temperature determined by the domain evolution. The results also confirmed that the change of the domain structure was indeed caused by the phase transition. The phase transition temperature of the lithium doped KNN-based epitaxial thin film was 110 ℃, while the phase transition temperature of the pure phase KNN-based ceramic was about 200 ℃,27-28 which is because that lithium doping in KNN-based films is an effective way to lower the phase transition temperature29. However, the phase transition temperature of 6% Li-doped KNN-based thin films was higher than that of Li-doped bulk ceramics,30 which could be due to the evaporation of lithium during


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 29

the annealing or owing to the clamping effect induced by the substrate since the ferroelectric transition temperature could be altered by the elastic strain.7, 31 The changes in domain structure with temperature were related with the change of spontaneous polarizations due to the phase transition. In the KNN pure phase bulk ceramic, the orthorhombic - tetragonal phase transition occurs with spontaneous polarizations rotated from the directions in the orthorhombic phase to the directions in the tetragonal phase. However, domain walls along the [100] direction in the (001) plane with domains on both sides of those domain walls having180° phase difference were observed at T > 130 ℃, which indicates that the new phase could be a rhombohedral, MA or MB but tetragonal phase. For the tetragonal phase of (001) oriented epitaxial thin films, it could only have lateral or vertical polarization component when PFM tips scan along the [100] or [010] direction, and it could not have domain walls in the (100) plane, thus the new phase at high temperatures is not tetragonal. Since the rhombohedral phase is a low temperature phase, the new phase should not be rhombohedral either. Figure 5 (c) shows the X-ray diffraction map of (002) peaks for Li-doped KNN epitaxial thin films above the phase transition temperature at 150 ℃, and c-KNN=4.031 Å can be estimated, which is similar to that at room temperature (c-KNN=4.018 Å). And there was no new diffraction spot shown up between that of the KNN film and the substrate at ω=0 in Xray diffraction map, which also confirms the new phase at high temperature is not rhombohedral. Thus the new phase could be MA or MB phase. If c>a, it is MA. If c a-KNN=3.954 Å at room temperature and c-KNN=4.031 Å at 150 ℃, it would result in a high stain if lattice a-KNN became larger then c-KNN at 150 ℃. Thus, the new phase could only be MA.


14

Page 15 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The emergence of monoclinic MC - MA phase transition could be ascribed to the elastic strain imposed by the substrate. The well-defined elastic strain in epitaxial thin films could induce fundamental changes in not only the crystal structure but also the ferroelectric properties in lead-free ferroelectrics. For example, strain-free rhombohedral phase of BiFeO3 (BFO) thin films changed into monoclinic MA phase by compressive strain or monoclinic MB structure by tensile strain, and MA - MC phase transition occurs at large enough compressive strain, which was reported to have a higher ferroelectric polarization than any other perovskites.32-33 The thermally induced phase transition from MC to MA instead of rhombohedral to tetragonal phase transition was also observed in highly-strained BFO thin films.34-35 One fascinating phenomena of those monoclinic phases is that they have been reported as bridging structures with strongly enhanced functional properties within the MPB in lead-based ferroelectrics, and considered as the origin for enhanced piezoelectricity due to the symmetry-allowed polarization rotation in those monoclinic phases.1,

36

Previously, the emergence of monoclinic MA phase in

K0.75Na0.25NbO3 thin films was observed by the incorporation of a highly anisotropic strain.37 However, the monoclinic MC phase and MC - MA phase transition were discovered for the first time in lead-free KNN-based epitaxial thin films. To understand how the monoclinic MC - MA phase transition effects the piezoelectric properties, the SSPFM was applied to analyze the piezoelectricity. The local piezoelectric response and the ferroelectric switching behavior were recorded at more than six positions from 26 ℃ to 170 ℃. In order to measure the amplitude and phase loops versus tip bias, a series of DC voltage with a maximum value of 20 V and 3 V AC voltage were applied simultaneously to the tip. Figure 6 shows the typical amplitude and phase loops at various temperatures. The amplitude loops of KNN-based epitaxial thin films exhibited typical butterfly shapes from 26 ℃


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 29

to 170 ℃, and all domains demonstrated 180° phase switching under an external tip bias as shown in phase loops, indicating typical ferroelectric-like piezoelectric response. In addition, all amplitude loops and phase hysteresis loops of the KNN-based thin film were reversible from 26 ℃ to 170 ℃. Additionally, the butterfly loop was intensively asymmetric at 170 ℃, with the amplitude at maximum positive tip bias much smaller than that at maximum negative tip bias.

Figure 6. Typical butterfly loops and phase hysteresis loops of Li-doped KNN epitaxial thin films measured at various temperatures: 26 ℃, 90 ℃, 110 ℃, 130 ℃, 150 ℃, 170 ℃. To better understand the temperature dependence of the local piezoresponse, local VPFM hysteresis loops are shown in Figure 7 (a) and various relevant parameters as a function of temperature are presented in Figures 7 (c) - (f). The local maximum piezoresponse Dmax was


16

Page 17 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


defined as the averaged absolute value of forward saturation response R and reverse saturation response R : Dmax=(R -R )/2 Remnant response Drem and coercive field VC were derived as the same as that of Dmax, and the work of switching A was the area of the hysteresis loop as shown in Figure 7 (b). The Dmax, with a week peak at 130 ℃, retained high values from 26 ℃ to 170 ℃. At T > 110 ℃, the Drem and AS increased abruptly and reached the highest value at 130 ℃ and 150 ℃ respectively and decreased monotonically after the maximum value. Generally, an increase in work of switching indicated the enhanced pinning. Different from hysteresis in macroscopic samples, the local PFM hysteresis loop was measured in a fine point (tip apex radius at about 10 nm), and the PFM hysteresis mainly occurred due to the nucleation of a single domain not the nucleation growth or the interaction of the multiple different domains.38 Thus, not only the domain wall energy and spontaneous polarization but also the nucleation of a new domain determined the PFM hysteresis. However, only the nucleated new domains were related with the pinning strength. Therefore the change in different nucleated new domains across the TPB could be attributed to the change in working of switching, remnant response and maximum switchable piezoresponse.39 As discussed above, the domain wall movement was the most active at 130 ℃, which is correlated with the increase in Drem and the slight peak in Dmax. In addition, the coexistence of the low symmetry MA and MC phases, leading to the change in domain wall energy and spontaneous polarization, could also be one of the factors that lead to the enhanced piezoelectric response at T > 110 ℃.


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 29

Figure 7. (a) Local VPFM hysteresis loops at different temperatures; (b) Definition of various piezoelectric parameters in VPFM hysteresis loop: forward coercive voltages V+ and reverse coercive voltage V-, forward saturation response R and reverse saturation response R , forward remnant response R and reverse remnant response R ; Various piezoelectric parameters as function of temperature: (c) maximum piezoresponse D , (d) remnant response D , (e) work of switching A and (f) coercive field V . The change in spontaneous polarization and corresponding change in domain structures also contributed to the switching kinetics. The coercive field increased drastically at T > 110 ℃, which concurs with the phase transition. Since the thermal expansion coefficient of KNN film (8×10-6 K-1)40 is close to that of STO substrate (11×10-6 K-1),41 a thermal expansion induced mismatch, ~ 0.03%, is continuous and considerably small in the temperature range from 26 ℃ to 190℃ , which could not induce the abrupt change of the domain structure and piezoelectricity with temperature. The increase in coercive field indicated a hardening effect, which could be


18

Page 19 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


related with the change of elastic strain between the film and the substrate. Both the polarization change and the lattice parameter difference across TPB could result in a change of clamping effects between the STO substrate and the KNN-based epitaxial film. Thus, enhanced coercive field at temperatures higher than 110 ℃ could be ascribed to a stronger substrate clamping effect originating from the phase transition.42-44 Imprint bias Im is defined as the averaged value of forward coercive voltages V+ and reverse coercive voltage V-. At T < 170 ℃, imprint bias had a negative value as shown in Figure S3. However, imprint bias changed into a positive value at 170 ℃, which indicates the existence of a negative internal bias field. When a negative tip bias is applied, the negative polarization is enhanced leading to an increased piezoelectric amplitude. While a positive tip bias is applied, the positive polarization is smaller than that at the negative tip bias due to the negative internal bias field. Therefore, the existence of internal bias field contributed to the intensively asymmetric butterfly loop and the suppressed piezoelectric response at 170 ℃, as shown in Figure 6. This internal bias field was usually generated by the aligned or accumulated defects.39 Although a positive internal bias field existed at T < 170 ℃ as shown in Figure S3, the butterfly loops were not highly asymmetric, which could be due to the existence of nanodomains. Nanodomains

Domain Evolution and Piezoelectric Response ... - ACS Publications

Recommend Documents