- Email: [email protected]
Strain tuning effects in perovskites
Strain tuning effects in perovskites
2
Zhenxiang Cheng*,¶, Fang Hong*, Tingting Jia†, Hongyang Zhao‡, Zengmei Wang§, Yuanxu Wang¶, Kiyoshi Ozawak, Hideo Kimurak *Institute for Superconducting and Electronics Materials, University of Wollongong, Innovation Campus, North Wollongong, NSW, Australia, †Research Laboratory for Nano-Biomechanics, Institutes of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, CAS, Shenzhen, China, ‡Hubei Key Laboratory of Plasma Chemistry and Advanced Materials, Wuhan Institute of Technology, Wuhan, People’s Republic of China, §School of Materials Science and Engineering, Key Lab. of Construction Materials, Southeast University, Nanjing, People’s Republic of China, ¶Institute for Computational Materials Science, School of Physics and Electronics, Henan University, Kaifeng, People’s Republic of China, kNational Institute for Materials Science, Tsukuba, Japan
2.1
Introduction
It has been recognized that transition metal oxides exhibiting a wide range of phenomena, such as magnetism, superconductivity, ionic conduction and ferroelectricity, host versatile forms of order in their charge, spin, and orbit sectors. These electronic orders meet at the interface and are strongly affected by electronic processes such as charge transfer, hybridization, and exchange interactions, as well as by the built-in scalar or vector potential, to produce emergent properties and functions. Therefore oxide heterostructures present a tremendous opportunity for realizing new phases through controlled synthesis routes [1]. An important approach in designing the heterostructure interface with targeted properties is to tailor the amount of strain at the interface to tune the balance among the competing interactions and structure the degrees of freedom available to oxides. Using epitaxy and the misfit strain imposed by an underlying substrate, it is possible to strain dielectric thin films to levels far beyond the normal crack failure in the bulk form of the active material. Such strains are used to enhance mobility in transistors and increase the superconducting, ferromagnetic (FM), and ferroelectric transition temperatures of many materials [2]. Multiferroic material in which magnetic ordering and ferroelectric ordering can coexist, shows great potential for the next generation of data storage applications and other kinds of spintronic devices because spin freedom can be operated beyond the charge freedom. Therefore exploring new functional materials with multiferroic properties has become a highly topical research field in recent years [3–7]. BiFeO3 has demonstrated its importance due to its room temperature multiferroic property [8]. Bi3+ is not environmentally friendly and is a risk to human health, however, and obtaining single phase on a large scale is still a problem, which could increase the cost of production. Furthermore, the weak magnetization at room temperature, Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications. https://doi.org/10.1016/B978-0-12-814499-2.00002-5
© 2019 Elsevier Inc. All rights reserved.
24
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
weak magnetoelectric coupling, and large electrical coercive field strongly limits its practical application. Rare earth manganites, such as DyMnO3 [9–11] and TbMnO3 [10] (space group: Pnma), are of great interest, as electric polarization is induced by spiral magnetic ordering because of the simultaneously broken spatial inversion and time-reversal symmetries [12]. In addition, the ferroelectric property can also be found in antiferromagnetic (AFM) hexagonal HoMnO3 and YMnO3 (space group: P63cm) [13, 14]. The ferroelectric property of these two materials originates from the buckling of layered MnO5 polyhedral units, driven by the size effect and electrostatic interaction as well. In these manganites, the spiral magnetic ordering temperatures are very low and far from room temperature, which restricts their application in the common ambient environment. Most traditional ferroelectric materials, such as lead zirconate titanate (PZT) and BaTiO3 (BTO), are not magnetic. Therefore to find new multiferroic materials with strong ferroelectricity and a high magnetic transition temperature seems to be an unavoidable goal to achieve real applications. Up to now, there have been few reports on ferroelectric studies of the orthoferrites, except for DyFeO3 and SmFeO3 [15–17]. From the point of view of symmetry, it is nearly impossible to achieve a ferroelectric phase in orthorhombic ferrite. In the case of DyFeO3, the ferroelectric transition takes place below the magnetic transition temperature of the Dy3+ lattice around 3.5 K, which can be explained by the exchange-striction interaction. As for SmFeO3, a small polarization is supposed to occur due to the symmetric exchange striction interaction among the canted AFM spins [18, 19], so it shares similar ferroelectric origins to DyFeO3 [15] and CaMn7O12 [20–22]. Strain engineering has been employed to produce low-dimensional materials, such as functional thin films, and this method has introduced many unique properties that are not possessed by their corresponding bulk forms. The colossal magnetoresistance (CMR) manganite perovskites have attracted considerable attention due to their potential applications and fascinating physical properties, such as phase separation, charge ordering, and the CMR effect. A strong interplay among the degrees of freedom, involving charge, lattice, spin, and orbital ordering, has been found in CMR materials [23–26]. Therefore for a film-on-substrate structure, the substrate-induced lattice strain can modify the physical properties of the CMR manganite through modification of the MndOdMn bond lengths and angles, and thus alter the strength of the double exchange (DE) and Jahn-Teller (JT) latticeelectron coupling [27–34]. In this work we will show that in epitaxial pseudocubic SmFeO3 on (001) SrTiO3 (STO):Nb substrate, prepared by pulsed laser deposition (PLD), the mismatch between the two sets of lattices could introduce static strain and lead to local structural distortion at the film-substrate interface, and that the structure and the physical properties of the thin film are consequently modified. We will also show that in the epitaxial grown La2/3Ca1/3MnO3 (LCMO) CMR thin film on (001) cut BTO substrate, the strain state of LCMO film is subject to large changes at phase transition temperatures of BTO, especially at the rhombedral-orthorombic (R-O) transition of BTO (190 K). Such a change in the strain state will introduce a strong change in the lattice distortion in LCMO film and thus both the transport and magnetic property was tuned by dynamic strain.
Strain tuning effects in perovskites
2.2
25
Experimental
SmFeO3 thin films were deposited at 830°C on (001) Nb-STO in a dynamic flowing oxygen atmosphere of 200 mTorr, using a PLD system. The SmFeO3 target was prepared using the solid-state reaction method, and X-ray diffraction (XRD, model: GBC MMA, Cu Kα radiation) shows that it is single phase with an orthorhombic structure. A Nd:YAG laser source was used with 355 nm wavelength. During the deposition process, the laser was stabilized at 6–7 J/cm2, and pulses were repeated at 10 cycles per second. The crystal structures of the films were examined by XRD at room temperature. The thickness of the films was determined by scanning electron microscope (SEM, model: JEOL JSM-6460A) and was around 450 nm. A high-resolution transmission electron microscope (HRTEM, model: JEOL JEM-4000EX) clarified the epitaxial growth of the thin film at the interface and the existence of the boundaries that separate the majority [010]o and the minority [101]o lattices. The magnetic properties were measured by a superconducting quantum interference device (SQUID) magnetometer. Pt electrodes were deposited on the film for ferroelectric measurements by magnetron sputtering with a standard shadow mask. A ferroelectric analyser (TF2000, aixACCT) was employed to characterize the ferroelectric properties (ferroelectric hysteresis loops and fatigue). Meanwhile, positive-up negative-down (PUND) mode measurements were performed to reveal the real switching process. Nanoscale ferroelectric measurements of as-grown SmFeO3 film were carried out via piezoresponse force microscopy (PFM, model: MFP-3D Asylum Research). The typical scan size and rate were 500 nm and 0.1 Hz, respectively. Positive and negative dc bias of 5 V was applied to the tip to produce the switched domain structure, and subsequent scanning was done at 370 kHz in dual AC resonance tracking (DART) PFM mode. The LCMO thin film samples in this work were deposited at 600°C in a reduced dynamic flowing oxygen atmosphere of 200 mTorr, using a PLD system. The thickness of the obtained thin film was around 100 nm, which is within the strain relaxation thickness of epitaxially grown film. The structure of the film was determined by using Cu Kα radiation on a JEOL 3500 XRD machine. Raman scattering was carried out to examine the structural change in the LCMO film at the BTO substrate phase transition temperature, using a micro-Raman system with a 514.5 nm laser, as Raman spectra have ultra-high spatial resolution, which means that a shift of 1 cm1 in frequency quantitatively measures a change of 0.0003 nm in bond length.19 The magnetic and transport properties of the sample were examined on a Quantum Design 14 Tesla physical properties measurement system (PPMS). For the first principles calculations, an initial distorted structure was used. The redefined crystal axes of the SmFeO3 film were set as: a//[101]o, b//[010]o, and c//[101]o of the orthorhombic structure of bulk SmFeO3. The lattice parameters ˚ ¼ 2c for the STO, while the out-of-plane a and b of the film were set as a ¼ b ¼ 7.81 A lattice parameter c and angle β were relaxed. During this process, other parameters, a, b, and the angles α and γ were fixed. The relaxed and distorted SmFeO3 crystal ˚ , c ¼ 7.44 A ˚ , α ¼ 90°, β ¼ 92.09°, and γ ¼ 90°. The calculastructure is a ¼ b ¼ 7.81 A tions were performed using the project-augmented wave (PAW) method implemented in the Vienna ab initio simulation package (VASP).
26
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
2.3
Results and discussion
2.3.1 Structures The SmFeO3 (SFO) thin film was prepared by PLD on (001) STO:Nb single crystal substrate. The basic structure was examined by XRD, as shown in the right figure of Fig. 2.1A. The XRD pattern reveals that the SmFeO3 is epitaxially grown, and three satellite peaks are found near the (001), (002), and (003) STO:Nb peaks. This is similar to the case of LuFeO3 and BiFeO3 films on (001) STO substrate, which are both “cube-on-cube,” grown epitaxially with pseudocubic structure. According to these three diffraction peaks, we can estimate the out-of-plane pseudocubic lattice param˚ . Considering the similarity between epitaxial BiFeO3 [35, 36] and eter as a ¼ 3.72 A SmFeO3, it is reasonable to tentatively assign the SmFeO3 film to the rhombohedral or hexagonal structure. A structural study based on diffraction peak calculations shows that the three peaks of the pseudocubic structure can successfully be assigned to the ˚, (012), (024), and (217) peaks of the rhombohedral structure (R3c, ar ¼ 5.638 A αr ¼ 52.64°), or to the (102), (204), and (217) peaks of the hexagonal structure ˚ , ch ¼ 14.53 A ˚ , α ¼ β ¼ 90°, γ ¼ 120°). (P63cm, ah ¼ bh ¼ 5.00 A Compared with the sharp diffraction peaks from the substrate, the three peaks from the thin film are relatively broad, which may originate from two sets of reflections from close sets of lattice parameters. To find out how the thin film really grows, HRTEM was employed to determine the interface structure, as shown in Fig. 2.2. The cross-section of the thin film was imaged in bright-field mode, in which a clear
2q (degree)
(A)
(B)
Fig. 2.1 X-ray diffraction pattern of SmFeO3 film deposited on (001) STO:Nb single crystal substrate; lattice parameter: STO cc ¼ 3.905 and SFO cc ¼ 3.790 (A). La2/3Ca1/3MnO3 (LCMO)/BaTiO3 (BTO) sample measured at room temperature (B). Inset of the (B) figure is the crystal structure of the LCMO film on BTO substrate: top view of one LCMO cell above a 2 2 BTO supercell. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society; Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
Strain tuning effects in perovskites
27
interface between the thin film and the substrate can be seen. In Fig. 2.2, a specifically selected film/substrate interface highlighted in the dotted rectangular frame is magnified in the upper left corner. Considering the lattice mismatch between the orthorhombic SmFeO3 and the cubic STO, stress-induced distortion should exist at the film/substrate interface. The TEM image (Fig. 2.2) of the film/substrate interface clearly shows the atomic rearrangement in an exactly cubic structure that is gradually relaxed to orthorombic. In addition, two types of growth along different crystallographic directions of orthorhombic SmFeO3 were found. A structural transition boundary is clearly shown in the TEM image. According to the TEM image and the selected area electron diffraction pattern, the growth direction of the SmFeO3 film in relation to the substrate orientation can be determined, which is shown in Fig. 2.2B and C. The out-of-plane directions of the SmFeO3 film on (100) STO substrate are <101 >o and < 010 >o, respectively. Therefore the three satellite diffraction peaks in the XRD patterns from low to high angles can be assigned to 1/2(101)o, (101)o + (010)o, and 3/2(101)o. This also explains why the diffraction peak of the film at around 2θ ¼ 48° is broad, because it contains two sets of diffraction peaks in very close proximity. The estimated lattice spacing ˚ , which is significantly reduced in comparison with the bulk value of d101 is 3.79 A ˚ 3.89 A. Such a reduction is caused by the elongation of the b axis due to the fact that ˚ , which is less than 2a ¼ 7.806 A ˚ of STO. This estimation is based on bSFO ¼ 7.706 A the part of the film with <101 >o growth, although the part of the film with h010io growth may not exhibit a significant change in the lattice parameter because the in-plane d101 lattice spacing matches well with d100 of the STO substrate. Therefore the so-called epitaxial growth actually includes two sets of growth directions. The part of the film with < 101 >o growth shows significant distortion with reduced a and c, and elongated b. Such a structural distortion will cause orbital and spin reconstruction and result in novel properties. The existence of the two sets of growth will introduce grain boundaries with disorder or rearrangement of the atoms, which can be clearly observed in the TEM image in Fig. 2.2. The boundary between the [010] and [101] lattices should also introduce extra stress and local structural distortion in this system. The epitaxial growth of the film will gradually relax and therefore the structural distortion will be released when the film grows thicker. The boundaries between lattices with two different orientations always exist, however, which is beyond the scale of the stress on the interface. Fig. 2.1B shows the XRD pattern of an LCMO thin film on BTO substrate fabricated by the PLD method. Only the (001) and (002) peaks of LCMO were observed near the BTO (100)/(001) and (200)/(002) diffraction peaks, indicating that LCMO film shows a strong epitaxial growth habit on BTO substrate. The expected 90° domain structure in the BTO substrate is also demonstrated by the presence of (002) and (200) diffraction peaks in the XRD pattern. The lattice parameters of the ˚ and c ¼ 4.001 A ˚ , based on the above BTO substrate are calculated as a ¼ 3.955 A ˚. XRD pattern, while the pseudocubic lattice parameter of LCMO film is c ¼ 3.83 A ˚ In comparison to the standard value, c ¼ 3.86 A, of bulk LCMO, this result shows that the LCMO film is compressed in the out-of-plane direction with large compressive strain.
28 Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
Fig. 2.2 (A) Interface between the SmFeO3 thin film and the SrTiO3 substrate, on which the grain boundary of [010]o and [101]o grains is also shown. The inset in the upper left corner is the high-resolution transmission electron microscope image of the film/substrate interface, magnified from the selected [010] growth area shown in the dotted rectangular frame. The inset in the lower left corner is the selected area electron diffraction pattern, where the high-level diffraction spot shows a splitting, indicating the different structures existing in the film. (B) and (C) contains the schematic structures of the two different growth orientations of the film, h101io and h010io, respectively. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
Strain tuning effects in perovskites
29
2.3.2 Multiferroic property of SmFeO3 Electric hysteresis loops of SmFeO3 films on STO:Nb were collected at room temperature. When the applied voltage exceeds 1.2 V (electric field 22 kV/cm), a quasi-rectangular loop can be observed, as shown in Fig. 2.3A, indicating a strong ferroelectric property. A higher applied voltage will increase the leakage current, and a relatively round ferroelectric loop will be present (not shown). The time dependence of the applied voltage and consequent current are given in Fig. 2.3B, in which an obvious switching current can be seen, which is quite different from typical resistor or capacitor behavior. The small difference in current behavior between opposite voltages is probably due to the electrode effect (Au and Nb-STO). To check the fatigue properties, polarization-electrical (P-E) loops were measured 106 times, and the remanent polarizations are presented in Fig. 2.3C. The remanent polarization is about 0.08 μC/cm2. There was no significant change in the remanent polarization after the fatigue test, indicating the good stability of this ferroelectric material. Considering the leakage effect, which can also contribute to P-E loops, we used the PUND method to confirm the real polarization. The PUND method uses an up-updown-down wave to exclude the leakage effect. The standard PUND wave form is
(A)
(B)
(C) Fig. 2.3 (A) Electrical hysteresis loop of SmFeO3 film on SrTiO3:Nb at room temperature, (B) time dependence of applied voltage and corresponding current, (C) fatigue test results after 106 cycles (10 data points every decade). Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
30
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
Fig. 2.4 Time dependence (left) of applied voltage (ranging from 2 to 2 V) and consequent current (right) measured in the positive-up negative-down mode. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
presented in Fig. 2.4. At the beginning, the system uses a negative writing pulse to polarize the film. After that, a positive pulse V1 is applied to the film, and a current will be produced due to leakage and domain flipping. Then another positive pulse V2 is applied and a current will be produced, but only due to leakage. Therefore the integrity of the current difference produced between the up-up and down-down processes is in accordance with two times the remanent polarization, 2Pr. This also applies to the down-down process. A V3 pulse will switch the ferroelectric domains to the opposite direction, and V4 just measures the leakage current contribution. In Fig. 2.3, a clear difference in the current between I1 and I2, or I3 and I4 can be observed (since the pure leakage current peak is lower and narrower than when it is combined with the switching current), indicating that polarization in the P-E loop is real. The real current contribution from ferroelectric switching is reduced to half, however, meaning that the remanent polarization is about 0.04 μC/cm2. This value is very close to that found in orthorhombic HoMnO3 [37], in which Mn3+ spins are in E-type AFM ordering. To some extent, considering the different magnetic states involved in the orthorhombic HoMnO3 and the orthorhombic SmFeO3, this similarity suggests that the mechanism in the orthorhombic bulk SmFeO3 may not apply to the SmFeO3 thin film. The PFM image presented in Fig. 2.5 shows the presence of FE domains, which are mainly dominated by two types with opposite vectors. This means that the domain walls in this film are mostly 180° domain walls. Meanwhile, information on the surface morphology was also obtained, and the results show an average roughness of about 1.8 nm, indicative of good-quality film. Comparing the morphology with the phase pattern, it is not difficult to see that there is more than one domain in some single grains. The average domain size is around 0.1 μm, much smaller than the domain size in BiFeO3 film. One important characteristic of ferroelectricity is that the ferroelectric domains can be controlled and reversed by an external electric field, the so-called “write.” To do this we carried out PFM lithography on the SmFeO3 film. A square-in-square pattern is obtained, as shown in Fig. 2.5C, in which the inset is the pre-set pattern. The amplitude distribution along the red line across the pattern
Strain tuning effects in perovskites
31 1.0
1.0
80
0.8
10
0.8
60
Deg µm
µm
20 0 0.4
5
0.6
0 0.4
–20 –5
–40 0.2
–60
0.2 –10
–80 0.0
0.0
(A)
nm
40 0.6
0.0
0.2
0.4
0.6
0.8
0.0
1.0
0.2
0.4
0.6
0.8
1.0
µm
(B)
µm
10
8
80
80
40
6
0 4
pm
µm
20
–20 –40 2
Amplitude (pm)
60 40
0
–40
–60 –80 –80 2.0
0
(C)
4.0
6.0
8.0
Position (µm) 0
2
4
6 µm
8
10
(D)
Fig. 2.5 Nanoscale ferroelectric domain structure characterized by piezoelectric force microscope (PFM): (A) amplitude image and (B) phase image; (C) PFM lithography pattern, inset: the pre-set pattern, and (D) amplitude distribution along the red line across the lithography pattern in (C). Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
is presented in Fig. 2.5D. It is clear that these two cubic boxes show opposite amplitude values, suggesting that they are truly two types of domains with opposite polarization directions. To investigate the magnetic properties of this “pseudocubic” SmFeO3 film, we carried out measurements using a magnetic properties measurement system (MPMS). The field-cooling temperature dependence of the magnetic moment was measured from 10 to 300 K, as shown in Fig. 2.6A. A sharp ferromagnetic (FM)-like transition can be found around 185 K. The FM behavior can be confirmed by the magnetic hysteresis loop at 100 K. Curie-Weiss law fitting gives a positive
32
(A)
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
(B)
Fig. 2.6 (A) Temperature dependence of magnetic moment from 10 to 300 K, and magnetic hysteresis loop measured at 100 K (inset); (B) antiferromagnetic interaction of Fe3+ sublattice confirmed by Curie-Weiss law fitting with temperature range from 100 to 300 K. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
intercept, which corresponds to a negative Curie-Weiss temperature, indicating an AFM interaction starting below 240 K, as can been seen in Fig. 2.6B. Hence, the FM-like behavior should originate from a canted AFM ordering. This is quite different from what happens in bulk SmFeO3, in which a spin reorientation takes place around 433 K from a canted AFM weak ferromagnetic ordering to simple AFM ordering. Such a big change may be attributed to Fe3+ spin frustration, which weakens the interaction between Fe3+ ions compared with that in the orthorhombic lattice. The stress-induced distortion on the interface and boundaries should be responsible for the significant change in the magnetic properties. This result indicates that the ferroelectricity in the SmFeO3 at room temperature may not arise from the magnetic exchange striction interaction, but from the structural distortion.
2.3.3 The first principle calculation Bulk SmFeO3 has been reported to show spin-driven ferroelectric polarization at room temperature based on experimental measurements; however, the theoretical calculations deny such a possibility. Therefore the multiferroic property in SmFeO3 is still controversial. Our SmFeO3 thin film, however, shows strong structural distortion at both the film-substrate interface and the grain boundaries. Such structural distortion will cause the reconstruction of the crystal lattice and rearrangement of spin, and thus electrical polarization, driven by both displacement of ions and spin. Adopting a distorted structure of SmFeO3 at the film-substrate interface, the different magnetic states were calculated based on first principles calculations in the framework of density functional theory (DFT) and the result is presented in Table 2.1. The energies of those magnetic states were compared with G-type AFM ordering of Fe for both Hubbard U ¼ 0 and 2, which shows that G-AFM is the most stable state in comparison to A-type AFM, C-type AFM, FM, and paramagnetic (PM) states [38]. The polarization
Strain tuning effects in perovskites
33
Table 2.1 Calculated total energies of the distorted SmFeO3 system with different magnetic structures
Energya (U ¼ 0) Energya (U ¼ 2)
G-AFM
A-AFM
C-AFM
FM
PM
ΔP [100]
0
1.09
0.739
3.599
5.045
15.32 μC/cm2
0
1.66
0.723
5.372
12.274
15.31 μC/cm2
The Hubbard-U values of 0 and 2 were considered in the calculation and compared. The ferroelectric polarization was calculated for G-type AFM. a Total energy (relative to the G-AFM state in eV/f.u.) Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
is calculated as 15.3 μC/cm2 for both U values. The value of U does not show any significant effect on the magnetic state and polarization. Although the calculated polarization value is much larger than the measured value, this result indicates that the observed polarization might be caused by the structural distortion at the filmsubstrate interface. The discrepancy might be caused by the relaxation of the distorted SmFeO3 film structure to a bulk one in a relatively thick film, which has been confirmed in HRTEM images. In addition, G-AFM easily forms a canted magnetic structure [38, 39] and causes the appearance of net moment and spin-driven polarization based on an inverse Dyaloshinski-Moriya (DM) interaction. Although we cannot distinguish the two contributions to the polarization, that is, ionic-displacement-driven polarization and spin-driven polarization, studying the possibility for the existence of large magnetoelectric coupling will be the aim of the next step in our research program. Finally, we compare the calculated electron density isosurface with the value of 0.45 for undistorted and distorted SmFeO3 systems obtained from DFT calculation, which is shown in Fig. 2.7. Electron density contour is known to be an informative tool to distinguish different bonding interaction in solids. It is obvious that, in the distorted system, the electron density between oxygen and iron is strong enough to form a three-dimensional network, indicating a strong asymmetric covalent bonding interaction. While in the undistorted system, the electron density between oxygen and iron is weak. The strong asymmetric covalent bonding is responsible for the observed polarization in the distorted SmFeO3 thin film.
2.3.4 Dynamic strain tuning of the transport and magnetic property of La2/3Ca1/3MnO3 film Besides the SmFeO3, a stress tuning effect is also observed in the CMR manganite perovskites, such as LCMO film. In Fig. 2.8A, the Raman spectra of the film are dominated by the strong Raman scattering peaks from the substrate, and only a few weak peaks from LCMO can be observed due to its pseudocubic symmetry. Three peaks are located at around
34
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
(A)
(B)
Fig. 2.7 The calculated electron density isosurface with the value of 0.45 for undistorted (A) and distorted (B) SmFeO3 systems. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
(A) 153 K 233 K
I
Log intensity (a.u.)
236.4 cm–1 173 K 163 K
(B) 153 K
III II 235.7 cm–1
233 K 200
300
400
(C) 193 K
500
600
466.4 cm–1
700 220
230
240
250
(D) T < 193 K 645.3 cm–1
183 K
T > 193 K
465.2 cm–1 450
460
470
480
560
600
640
680
720
Raman shift (cm–1)
Fig. 2.8 (A) Raman spectra of the La2/3Ca1/3MnO3/BaTiO3 film from 153 to 233 K. (B) Magnified view of peak I. (C) Magnified view of peak II. (D) Magnified view of peak III. Reproduced from Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
Strain tuning effects in perovskites
35
235, 465, and 645 cm1, which are ascribed to the y-rotation, bending, and stretching modes of MndO octahedra, respectively [40]. The two high-frequency phonons (bending and stretching modes) are associated with modifications of the MndO bond, while the low frequency one is assigned to rigid rotations of the octahedra. Fig. 2.8B shows that there is a peak shift of 0.8 cm1 to higher frequency of the y-rotation mode between 163 and 173 K. This shift takes place at a temperature lower than the temperature where the substrate phase transition occurs, and thus, it is difficult to directly attribute this shift to the substrate-phase transition. In Fig. 2.8C, the bending mode shows a shift of 1.2 cm1 to the lower frequency side at temperatures lower than 193 K, with a simultaneous increase in peak intensity. In Fig. 2.8D, the Raman scattering peak from the stretching mode disappears at temperatures above 193 K. The temperatures for the shifting of the bending mode and the vanishing of the stretching mode coincide with the temperature for the substrate structural phase transition, which indicates that the structure of the LCMO film is modulated by the BTO R-O phase transition. Fig. 2.9A shows that a PM to FM transition starts at 244 K for the sample in the 100 Oe DC field, while the transition temperature shifts to 277 K and 296 K when higher fields of 2000 Oe and 1 T are applied, respectively (M-T curve). Distinct anomalies with concomitant thermal hysteresis are observed on both the field cooled and zero-field cooled (FC-ZFC) magnetization curves between 188 and 200 K, measured at 100 Oe, where the BTO R-O phase transition and LCMO Raman scattering abnormalities occur. Similar magnetization anomalies were also observed for the FC magnetization curve at an applied magnetic field of 2000 Oe, but with a much reduced magnitude. However, such anomalies were not clearly observed for the magnetization curve measured under the highest magnetic field of 1 T, which means that a strong external magnetic field can suppress the jump in magnetization. This result
180
2000 Oe 100 Oe
200 T (K)
6.0 ´ 10–4
100
220
60
30
FC 0.0
(A)
90 100 Oe
ZFC 150
200 T (K)
250
2.5
0T
2.0 R (W)
1T
2000 Oe
M (10–6 emu) (out-of-plane)
1.8 ´ 10–3 1.2 ´ 10–3
120
1T M (in-plane)
M (emu) (in-plane)
2.4 ´ 10–3
1.5 1.0 0.5 0.0
0 300
13T 0
(B)
80
160 T (K)
240
Fig. 2.9 (A) Field cooled and zero-field cooled in-plane and out-of-plane magnetization curves of the La2/3Ca1/3MnO3(LCMO)/BaTiO3(BTO) film measured under different magnetic fields; inset is an enlargement of the indicated temperature range. (B) Temperature dependence of resistance of the LCMO/BTO film measured under 0 and 13 T. Reproduced from Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
36
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
demonstrates that we are able to modulate the magnetic response of LCMO via strain, but this can only happen at low magnetic fields. The out-of-plane magnetization FC and ZFC curves of the same sample at a magnetic field of 100 Oe show a magnetic moment two orders of magnitude smaller than for the in-plane case, indicating strong magnetic anisotropy. Similar to the M-T curves, thermal hysteresis is also observed in the resistivity versus temperature measurements (R-T curves), as evidenced by the resistance peak positions at 208 and 193 K for heating and cooling measurements, respectively (shown in Fig. 2.9B). In addition, an abrupt decrease in the resistance jump was observed for both the cooling and the heating processes in R-T curves. The decrease in the magnitude of the jump in the resistance between 0 and 13 T is (2.36–0.5)/2.36 ¼ 78.8%. The temperatures for the occurrence of the abrupt decrease in resistance are the same as for the occurrence of magnetization anomalies, the BTO R-O phase transition and the abnormal LCMO Raman scattering peak. This abrupt change indicates that modulation of the strain in LCMO film by substrate R-O structural transformation modifies the transport behavior as well as the magnetization. In addition, both the magnitude and sign of the resistivity changes remain constant through many thermal cycles, indicating that the structural effects are elastic and reversible. Since cracks can be easily formed in BTO, one cannot completely eliminate the possibility of substrate-induced cracking in the film. Cracking will cause an irreversible increase in the resistivity. However, the observed decrease in the resistivity in this study should not have any relationship to cracks. In CMR materials, the semiconductor-metallic state transition takes place simultaneously with the PM-FM ordering. The relationship between conductivity and magnetic ordering at and below the transition temperature is explained by a double exchange mechanism (DE) and Jahn-Teller electron-lattice coupling, for which the MndOdMn bond length and angle are the key factors [41–43]. Thermal cycling of the BTO crystal during the temperature-dependent measurements leads to the formation of different types of crystal domains. These BTO domains give rise to two different strains in a magnetic film and two different modulations of strain at the BTO ˚ (a and b axes) phase transition. The lattice parameters of BTO change from 4.013 A ˚ ˚ and 3.99 A (c axis) to 4.001 A, with a deviation of 90° to 89.85° for the ab angle at the O-R phase transition. A change in the in-plane lattice parameters will change the MndOdMn bond length and bond angle, which is evidenced by Raman peak shifting, similar to the effect caused by Jahn-Teller distortion, and will eventually change the resistance of the film, according to the DE mechanism. However, a high magnetic field of 13 T was found to eliminate the observed sharp resistance jump in R-T measurements, and the resistance in high magnetic field below the R-O phase transition temperature is much smaller than the value measured without magnetic field. This means that the magnetic field can produce similar structural distortion effects to those that are produced in LCMO film by strain from substrate phase transitions. Such an effect produced by a magnetic field, however, is much stronger than the effect caused by a substrate phase transition. It has been recognized that CMR materials experience a significant magnetostriction effect, with the magnitude of volume change up to 0.1% at the Curie temperature, TC, due to strong electron-lattice
Strain tuning effects in perovskites
37
interaction or Jahn-Teller distortion when the 3d eg electrons start to delocalize at temperatures approaching TC from the high temperature side [44–46]. Such a delocalization of 3d electrons favors DE, and an insulator to metal-like transition takes place. The changing lattice distortion in LCMO film at the BTO phase transition temperature will affect the status of the eg electrons due to strong electron-phonon interaction and accelerates the process of delocalization of eg electrons, eventually leading to a sharp decrease in resistance. Therefore the observed anomaly in R-T measurements at the R-O transition indicates the existence of a strong electron-lattice interaction in CMR materials and strong effects of the interaction on the physical properties. However, the magnitude of the change in the lattice distortion due to the substrate is less than that caused by the magnetostriction effect in LCMO film when a strong magnetic field is applied. Therefore a high magnetic field can easily eliminate the observed resistance jumping in a zero magnetic field. Similarly, the large structural distortion in a high magnetic field caused by a strong electron-lattice interaction can mask the external disturbance, and thus the magnetic moment anomaly is destroyed by a high magnetic field. It has proposed that the changing of the magnetic anisotropy and magnetic domain is the cause for the modification of the magnetization at the BTO substrate structure–phase transition in other kinds of magnetic film (non-CMR film) [47, 48]. However, for those films and such a cause a magnetic field can kill the modification of magnetization but not the resistivity, which is different from what has been observed in this report. Therefore at least the modification of the magnetic anisotropy and the magnetic domain of CMR film at the BTO-phase transition is not the only reason for the observation in this report.
2.4
Summary
In summary, the structure, and the ferroelectric and magnetic properties were studied in a SmFeO3 film on STO:Nb substrate with strong static strain. Clear structural distortion can be identified on the [010]o grown film-substrate interface, as well as boundaries of [010]o and a small fraction of [101]o grains. A strong ferromagnetic-like transition is observed around 185 K, which is likely to be due to a canted AFM ordering. The PFM images clarify the existence of ferroelectricity and the switchable domain structure. There is also a saturating hysteresis loop in the ferroelectric measurements, and the loops are very stable, even after 106 cycles. The PUND method was used to determine the intrinsic ferroelectric polarization, and the remanent polarization is about 0.04 μC/cm2, which is comparable to that in orthorhombic HoMnO3. The structural distortion in this system is probably responsible for the significant change in the magnetic properties and the small polarization as well. The theoretical calculations of the highly distorted SmFeO3 structure have theoretically confirmed the existence of the ferroelectric polarization and net magnetic moment in the thin film form of SmFeO3. In an LCMO film grown on a BTO substrate, anomalies were observed on both M-T and R-T curves at the BTO R-O-phase transition, demonstrating a strong dynamic strain tuning effect. Furthermore, both the anomalies in the magnetization and the resistance drop can be eliminated by a strong magnetic field. All these
38
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
phenomena indicate that colossal resistance change in CMR material at temperatures around the PM-FM transition can be produced solely by strain as well as by magnetic field.
Acknowledgments We thanks the National Natural Science Foundation of China (51571083, 51702351) and the Shenzhen Science and Technology Innovation Committee (JCYJ20170307165829951, JCYJ20170413152832151) for support.
References [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]
A.S. Disa, et al., Phys. Rev. Lett. 114 (2015). 026801. H.Y. Hwang, et al., Nat. Mater. 11 (2012) 103. G. Catalan, J.F. Scott, Adv. Mater. 21 (2009) 2463. C.A.F. Vaz, J. Hoffman, C.H. Anh, R. Ramesh, Adv. Mater. 22 (2010) 2900. P. Jain, V. Ramachandran, R.J. Clark, H.D. Zhou, B.H. Toby, N.S. Dalal, H.W. Kroto, A. K. Cheetham, J. Am. Chem. Soc. 131 (2009) 13625. H. Cui, Z. Wang, K. Takahashi, Y. Okano, H. Kobayashi, A. Kobayashi, J. Am. Chem. Soc. 128 (2006) 15074. S.W. Kim, H.Y. Chang, P.S. Halasyamani, J. Am. Chem. Soc. 132 (2010) 17684. G.A. Smolenskiı˘, I.E. Chupis, Sov. Phys. Uspekhi 25 (1982) 475. N. Zhang, K.F. Wang, S.J. Luo, T. Wei, X.W. Dong, S.Z. Li, J.G. Wan, J.M. Liu, Appl. Phys. Lett. 96 (2010) 252902. T. Kimura, G. Lawes, T. Goto, Y. Tokura, A.P. Ramirez, Phys. Rev. B 71 (2005) 224425. T. Kimura, S. Ishihara, H. Shintani, T. Arima, K.T. Takahashi, K. Ishizaka, Y. Tokura, Phys. Rev. B 68 (2003) 060403. M. Mostovoy, Phys. Rev. Lett. 96 (2006) 067601. H. Lueken, Angew. Chem. Int. Ed. 47 (2008) 8562. T. Matsumoto, R. Ishikawa, T. Tohei, H. Kimura, Q. Yao, H. Zhao, X. Wang, D. Chen, Z. Cheng, N. Shibata, Y. Ikuhara, Nano Lett. 13 (2013) 4594. Y. Tokunaga, S. Iguchi, T. Arima, Y. Tokura, Phys. Rev. Lett. 101 (2008) 097205. Y. Du, Z.X. Cheng, X.L. Wang, S.X. Dou, J. Appl. Phys. 107 (2010) 09D908. J.-H. Lee, Y.K. Jeong, J.H. Park, M.-A. Oak, H.M. Jang, J.Y. Son, J.F. Scott, Phys. Rev. Lett. 107 (2011) 117201. R.D. Johnson, N. Terada, P.G. Radaelli, Phys. Rev. Lett. 108 (2012) 219701. J.-H. Lee, Y.K. Jeong, J.H. Park, M.-A. Oak, H.M. Jang, J.Y. Son, J.F. Scott, Phys. Rev. Lett. 108 (2012) 219702. G. Zhang, S. Dong, Z. Yan, Y. Guo, Q. Zhang, S. Yunoki, E. Dagotto, J.M. Liu, Phys. Rev. B 84 (2011) 174413. R.D. Johnson, L.C. Chapon, D.D. Khalyavin, P. Manuel, P.G. Radaelli, C. Martin, Phys. Rev. Lett. 108 (2012) 067201. X.Z. Lu, M.H. Whangbo, S. Dong, X.G. Gong, H.J. Xiang, Phys. Rev. Lett. 108 (2012) 187204. H.Y. Hwang, T.T.M. Palstra, S.-W. Cheong, B. Batlogg, Phys. Rev. B 52 (1995) 15046. J.M. Coey, M. Viret, S. von Molnar, Adv. Phys. 48 (1999) 167. M.R. Ibarra, P.A. Algarabel, C. Marquina, J. Basco, J. Garcia, Phys. Rev. Lett. 75 (1995) 3541.
Strain tuning effects in perovskites
39
[26] S. Jin, T.H. Tiefel, M. McCormack, H.M. O’Bryan, R. Ramesh, D. Schurig, Appl. Phys. Lett. 67 (1995) 557. [27] C. Zener, Phys. Rev. 82 (1951) 403. [28] A.J. Millis, P.B. Littlewood, B.I. Shraiman, Phys. Rev. Lett. 74 (1995) 514. [29] H.Y. Hwang, S.-W. Cheong, P.G. Radaelli, M. Marezio, B. Batlogg, Phys. Rev. Lett. 75 (1995) 914. [30] L. Nagaev, Phys. Stat. Sol. B 9 (1994) 186. [31] J.M. Teresa, M.R. Ibarra, P.A. Algarabel, et al., Nature 386 (1997) 256. [32] J.B. Goodenough, J.S. Zhou, Nature 386 (1997) 229. [33] H. Aliaga, D. Magnoux, A. Moreo, D. Poilblanc, S. Yunoki, E. Dagotto, Phys. Rev. B 68 (2003) 104405. [34] K. Lai, M. Nakamura, W. Kundhikanjana, M. Kawasaki, Y. Tokura, M.A. Kelly, Z. X. Shen, Science 329 (2010) 190. [35] M.K. Singh, H.M. Jang, S. Ryu, M.-H. Jo, Appl. Phys. Lett. 88 (2006) 042907. [36] J.F. Ihlefeld, A. Kumar, V. Gopalan, D.G. Schlom, Y.B. Chen, X.Q. Pan, T. Heeg, J. Schubert, X. Ke, P. Schiffer, J. Orenstein, L.W. Martin, Y.H. Chu, R. Ramesh, Appl. Phys. Lett. 91 (2007) 071922. [37] Y.S. Chai, Y.S. Oh, L.J. Wang, N. Manivannan, S.M. Feng, Y.S. Yang, L.Q. Yan, C.Q. Jin, K.H. Kim, Phys. Rev. B 85 (2012) 184406. [38] W.H. Meiklejohn, C.P. Bean, Phys. Rev. 102 (1956) 1413. [39] W.H. Meiklejohn, C.P. Bean, Phys. Rev. 105 (1957) 904. [40] M.N. Iliev, M.V. Abrashev, H.G. Lee, V.N. Popov, Y.Y. Sun, C. Thomsen, R.L. Meng, C. W. Chu, Phys. Rev. B 57 (1998) 2872. [41] P.W. Anderson, H. Hasegawa, Phys. Rev. 100 (1955) 675–681. [42] P.-G. de Gennes, Phys. Rev. 118 (1960) 141–154. [43] Y. Tokura, N. Nagaosa, Science 21 (2000) 462. [44] J.M. De Teresa, M.R. Ibarra, J. Blasco, J. Garcı´a, C. Marquina, P.A. Algarabel, Phys. Rev. B 54 (1996) 1187. [45] A.I. Abramovich, L.I. Koroleva, A.V. Michurin, O.Y. Gorbenko, A.R. Kaul, Phys. Solid State 42 (2000) 1494. [46] M.R. Ibarra, P.A. Algarabel, C. Marquina, J. Blasco, J. Garcı´a, Phys. Rev. Lett. 75 (1995) 3541. [47] R.V. Chopdekar, Y. Suzuki, Appl. Phys. Lett. 89 (2006) 182506. [48] H.F. Tian, T.L. Qu, L.B. Luo, J.J. Yang, S.M. Guo, H.Y. Zhang, Y.G. Zhao, J.Q. Li, Appl. Phys. Lett. 92 (2008) 063507.
Further Reading [49] Z.X. Cheng, F. Hong, Y.X. Wang, K. Ozawa, H. Fujii, H. Kimura, Y. Du, X.L. Wang, S. X. Dou, ACS Appl. Mater. Interface 6 (2014) 7356. [50] Z.X. Cheng, X.L. Wang, S.X. Dou, M. Osada, H. Kimura, Appl. Phys. Lett. 99 (2011) 092103.
2
Zhenxiang Cheng*,¶, Fang Hong*, Tingting Jia†, Hongyang Zhao‡, Zengmei Wang§, Yuanxu Wang¶, Kiyoshi Ozawak, Hideo Kimurak *Institute for Superconducting and Electronics Materials, University of Wollongong, Innovation Campus, North Wollongong, NSW, Australia, †Research Laboratory for Nano-Biomechanics, Institutes of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, CAS, Shenzhen, China, ‡Hubei Key Laboratory of Plasma Chemistry and Advanced Materials, Wuhan Institute of Technology, Wuhan, People’s Republic of China, §School of Materials Science and Engineering, Key Lab. of Construction Materials, Southeast University, Nanjing, People’s Republic of China, ¶Institute for Computational Materials Science, School of Physics and Electronics, Henan University, Kaifeng, People’s Republic of China, kNational Institute for Materials Science, Tsukuba, Japan
2.1
Introduction
It has been recognized that transition metal oxides exhibiting a wide range of phenomena, such as magnetism, superconductivity, ionic conduction and ferroelectricity, host versatile forms of order in their charge, spin, and orbit sectors. These electronic orders meet at the interface and are strongly affected by electronic processes such as charge transfer, hybridization, and exchange interactions, as well as by the built-in scalar or vector potential, to produce emergent properties and functions. Therefore oxide heterostructures present a tremendous opportunity for realizing new phases through controlled synthesis routes [1]. An important approach in designing the heterostructure interface with targeted properties is to tailor the amount of strain at the interface to tune the balance among the competing interactions and structure the degrees of freedom available to oxides. Using epitaxy and the misfit strain imposed by an underlying substrate, it is possible to strain dielectric thin films to levels far beyond the normal crack failure in the bulk form of the active material. Such strains are used to enhance mobility in transistors and increase the superconducting, ferromagnetic (FM), and ferroelectric transition temperatures of many materials [2]. Multiferroic material in which magnetic ordering and ferroelectric ordering can coexist, shows great potential for the next generation of data storage applications and other kinds of spintronic devices because spin freedom can be operated beyond the charge freedom. Therefore exploring new functional materials with multiferroic properties has become a highly topical research field in recent years [3–7]. BiFeO3 has demonstrated its importance due to its room temperature multiferroic property [8]. Bi3+ is not environmentally friendly and is a risk to human health, however, and obtaining single phase on a large scale is still a problem, which could increase the cost of production. Furthermore, the weak magnetization at room temperature, Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications. https://doi.org/10.1016/B978-0-12-814499-2.00002-5
© 2019 Elsevier Inc. All rights reserved.
24
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
weak magnetoelectric coupling, and large electrical coercive field strongly limits its practical application. Rare earth manganites, such as DyMnO3 [9–11] and TbMnO3 [10] (space group: Pnma), are of great interest, as electric polarization is induced by spiral magnetic ordering because of the simultaneously broken spatial inversion and time-reversal symmetries [12]. In addition, the ferroelectric property can also be found in antiferromagnetic (AFM) hexagonal HoMnO3 and YMnO3 (space group: P63cm) [13, 14]. The ferroelectric property of these two materials originates from the buckling of layered MnO5 polyhedral units, driven by the size effect and electrostatic interaction as well. In these manganites, the spiral magnetic ordering temperatures are very low and far from room temperature, which restricts their application in the common ambient environment. Most traditional ferroelectric materials, such as lead zirconate titanate (PZT) and BaTiO3 (BTO), are not magnetic. Therefore to find new multiferroic materials with strong ferroelectricity and a high magnetic transition temperature seems to be an unavoidable goal to achieve real applications. Up to now, there have been few reports on ferroelectric studies of the orthoferrites, except for DyFeO3 and SmFeO3 [15–17]. From the point of view of symmetry, it is nearly impossible to achieve a ferroelectric phase in orthorhombic ferrite. In the case of DyFeO3, the ferroelectric transition takes place below the magnetic transition temperature of the Dy3+ lattice around 3.5 K, which can be explained by the exchange-striction interaction. As for SmFeO3, a small polarization is supposed to occur due to the symmetric exchange striction interaction among the canted AFM spins [18, 19], so it shares similar ferroelectric origins to DyFeO3 [15] and CaMn7O12 [20–22]. Strain engineering has been employed to produce low-dimensional materials, such as functional thin films, and this method has introduced many unique properties that are not possessed by their corresponding bulk forms. The colossal magnetoresistance (CMR) manganite perovskites have attracted considerable attention due to their potential applications and fascinating physical properties, such as phase separation, charge ordering, and the CMR effect. A strong interplay among the degrees of freedom, involving charge, lattice, spin, and orbital ordering, has been found in CMR materials [23–26]. Therefore for a film-on-substrate structure, the substrate-induced lattice strain can modify the physical properties of the CMR manganite through modification of the MndOdMn bond lengths and angles, and thus alter the strength of the double exchange (DE) and Jahn-Teller (JT) latticeelectron coupling [27–34]. In this work we will show that in epitaxial pseudocubic SmFeO3 on (001) SrTiO3 (STO):Nb substrate, prepared by pulsed laser deposition (PLD), the mismatch between the two sets of lattices could introduce static strain and lead to local structural distortion at the film-substrate interface, and that the structure and the physical properties of the thin film are consequently modified. We will also show that in the epitaxial grown La2/3Ca1/3MnO3 (LCMO) CMR thin film on (001) cut BTO substrate, the strain state of LCMO film is subject to large changes at phase transition temperatures of BTO, especially at the rhombedral-orthorombic (R-O) transition of BTO (190 K). Such a change in the strain state will introduce a strong change in the lattice distortion in LCMO film and thus both the transport and magnetic property was tuned by dynamic strain.
Strain tuning effects in perovskites
2.2
25
Experimental
SmFeO3 thin films were deposited at 830°C on (001) Nb-STO in a dynamic flowing oxygen atmosphere of 200 mTorr, using a PLD system. The SmFeO3 target was prepared using the solid-state reaction method, and X-ray diffraction (XRD, model: GBC MMA, Cu Kα radiation) shows that it is single phase with an orthorhombic structure. A Nd:YAG laser source was used with 355 nm wavelength. During the deposition process, the laser was stabilized at 6–7 J/cm2, and pulses were repeated at 10 cycles per second. The crystal structures of the films were examined by XRD at room temperature. The thickness of the films was determined by scanning electron microscope (SEM, model: JEOL JSM-6460A) and was around 450 nm. A high-resolution transmission electron microscope (HRTEM, model: JEOL JEM-4000EX) clarified the epitaxial growth of the thin film at the interface and the existence of the boundaries that separate the majority [010]o and the minority [101]o lattices. The magnetic properties were measured by a superconducting quantum interference device (SQUID) magnetometer. Pt electrodes were deposited on the film for ferroelectric measurements by magnetron sputtering with a standard shadow mask. A ferroelectric analyser (TF2000, aixACCT) was employed to characterize the ferroelectric properties (ferroelectric hysteresis loops and fatigue). Meanwhile, positive-up negative-down (PUND) mode measurements were performed to reveal the real switching process. Nanoscale ferroelectric measurements of as-grown SmFeO3 film were carried out via piezoresponse force microscopy (PFM, model: MFP-3D Asylum Research). The typical scan size and rate were 500 nm and 0.1 Hz, respectively. Positive and negative dc bias of 5 V was applied to the tip to produce the switched domain structure, and subsequent scanning was done at 370 kHz in dual AC resonance tracking (DART) PFM mode. The LCMO thin film samples in this work were deposited at 600°C in a reduced dynamic flowing oxygen atmosphere of 200 mTorr, using a PLD system. The thickness of the obtained thin film was around 100 nm, which is within the strain relaxation thickness of epitaxially grown film. The structure of the film was determined by using Cu Kα radiation on a JEOL 3500 XRD machine. Raman scattering was carried out to examine the structural change in the LCMO film at the BTO substrate phase transition temperature, using a micro-Raman system with a 514.5 nm laser, as Raman spectra have ultra-high spatial resolution, which means that a shift of 1 cm1 in frequency quantitatively measures a change of 0.0003 nm in bond length.19 The magnetic and transport properties of the sample were examined on a Quantum Design 14 Tesla physical properties measurement system (PPMS). For the first principles calculations, an initial distorted structure was used. The redefined crystal axes of the SmFeO3 film were set as: a//[101]o, b//[010]o, and c//[101]o of the orthorhombic structure of bulk SmFeO3. The lattice parameters ˚ ¼ 2c for the STO, while the out-of-plane a and b of the film were set as a ¼ b ¼ 7.81 A lattice parameter c and angle β were relaxed. During this process, other parameters, a, b, and the angles α and γ were fixed. The relaxed and distorted SmFeO3 crystal ˚ , c ¼ 7.44 A ˚ , α ¼ 90°, β ¼ 92.09°, and γ ¼ 90°. The calculastructure is a ¼ b ¼ 7.81 A tions were performed using the project-augmented wave (PAW) method implemented in the Vienna ab initio simulation package (VASP).
26
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
2.3
Results and discussion
2.3.1 Structures The SmFeO3 (SFO) thin film was prepared by PLD on (001) STO:Nb single crystal substrate. The basic structure was examined by XRD, as shown in the right figure of Fig. 2.1A. The XRD pattern reveals that the SmFeO3 is epitaxially grown, and three satellite peaks are found near the (001), (002), and (003) STO:Nb peaks. This is similar to the case of LuFeO3 and BiFeO3 films on (001) STO substrate, which are both “cube-on-cube,” grown epitaxially with pseudocubic structure. According to these three diffraction peaks, we can estimate the out-of-plane pseudocubic lattice param˚ . Considering the similarity between epitaxial BiFeO3 [35, 36] and eter as a ¼ 3.72 A SmFeO3, it is reasonable to tentatively assign the SmFeO3 film to the rhombohedral or hexagonal structure. A structural study based on diffraction peak calculations shows that the three peaks of the pseudocubic structure can successfully be assigned to the ˚, (012), (024), and (217) peaks of the rhombohedral structure (R3c, ar ¼ 5.638 A αr ¼ 52.64°), or to the (102), (204), and (217) peaks of the hexagonal structure ˚ , ch ¼ 14.53 A ˚ , α ¼ β ¼ 90°, γ ¼ 120°). (P63cm, ah ¼ bh ¼ 5.00 A Compared with the sharp diffraction peaks from the substrate, the three peaks from the thin film are relatively broad, which may originate from two sets of reflections from close sets of lattice parameters. To find out how the thin film really grows, HRTEM was employed to determine the interface structure, as shown in Fig. 2.2. The cross-section of the thin film was imaged in bright-field mode, in which a clear
2q (degree)
(A)
(B)
Fig. 2.1 X-ray diffraction pattern of SmFeO3 film deposited on (001) STO:Nb single crystal substrate; lattice parameter: STO cc ¼ 3.905 and SFO cc ¼ 3.790 (A). La2/3Ca1/3MnO3 (LCMO)/BaTiO3 (BTO) sample measured at room temperature (B). Inset of the (B) figure is the crystal structure of the LCMO film on BTO substrate: top view of one LCMO cell above a 2 2 BTO supercell. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society; Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
Strain tuning effects in perovskites
27
interface between the thin film and the substrate can be seen. In Fig. 2.2, a specifically selected film/substrate interface highlighted in the dotted rectangular frame is magnified in the upper left corner. Considering the lattice mismatch between the orthorhombic SmFeO3 and the cubic STO, stress-induced distortion should exist at the film/substrate interface. The TEM image (Fig. 2.2) of the film/substrate interface clearly shows the atomic rearrangement in an exactly cubic structure that is gradually relaxed to orthorombic. In addition, two types of growth along different crystallographic directions of orthorhombic SmFeO3 were found. A structural transition boundary is clearly shown in the TEM image. According to the TEM image and the selected area electron diffraction pattern, the growth direction of the SmFeO3 film in relation to the substrate orientation can be determined, which is shown in Fig. 2.2B and C. The out-of-plane directions of the SmFeO3 film on (100) STO substrate are <101 >o and < 010 >o, respectively. Therefore the three satellite diffraction peaks in the XRD patterns from low to high angles can be assigned to 1/2(101)o, (101)o + (010)o, and 3/2(101)o. This also explains why the diffraction peak of the film at around 2θ ¼ 48° is broad, because it contains two sets of diffraction peaks in very close proximity. The estimated lattice spacing ˚ , which is significantly reduced in comparison with the bulk value of d101 is 3.79 A ˚ 3.89 A. Such a reduction is caused by the elongation of the b axis due to the fact that ˚ , which is less than 2a ¼ 7.806 A ˚ of STO. This estimation is based on bSFO ¼ 7.706 A the part of the film with <101 >o growth, although the part of the film with h010io growth may not exhibit a significant change in the lattice parameter because the in-plane d101 lattice spacing matches well with d100 of the STO substrate. Therefore the so-called epitaxial growth actually includes two sets of growth directions. The part of the film with < 101 >o growth shows significant distortion with reduced a and c, and elongated b. Such a structural distortion will cause orbital and spin reconstruction and result in novel properties. The existence of the two sets of growth will introduce grain boundaries with disorder or rearrangement of the atoms, which can be clearly observed in the TEM image in Fig. 2.2. The boundary between the [010] and [101] lattices should also introduce extra stress and local structural distortion in this system. The epitaxial growth of the film will gradually relax and therefore the structural distortion will be released when the film grows thicker. The boundaries between lattices with two different orientations always exist, however, which is beyond the scale of the stress on the interface. Fig. 2.1B shows the XRD pattern of an LCMO thin film on BTO substrate fabricated by the PLD method. Only the (001) and (002) peaks of LCMO were observed near the BTO (100)/(001) and (200)/(002) diffraction peaks, indicating that LCMO film shows a strong epitaxial growth habit on BTO substrate. The expected 90° domain structure in the BTO substrate is also demonstrated by the presence of (002) and (200) diffraction peaks in the XRD pattern. The lattice parameters of the ˚ and c ¼ 4.001 A ˚ , based on the above BTO substrate are calculated as a ¼ 3.955 A ˚. XRD pattern, while the pseudocubic lattice parameter of LCMO film is c ¼ 3.83 A ˚ In comparison to the standard value, c ¼ 3.86 A, of bulk LCMO, this result shows that the LCMO film is compressed in the out-of-plane direction with large compressive strain.
28 Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
Fig. 2.2 (A) Interface between the SmFeO3 thin film and the SrTiO3 substrate, on which the grain boundary of [010]o and [101]o grains is also shown. The inset in the upper left corner is the high-resolution transmission electron microscope image of the film/substrate interface, magnified from the selected [010] growth area shown in the dotted rectangular frame. The inset in the lower left corner is the selected area electron diffraction pattern, where the high-level diffraction spot shows a splitting, indicating the different structures existing in the film. (B) and (C) contains the schematic structures of the two different growth orientations of the film, h101io and h010io, respectively. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
Strain tuning effects in perovskites
29
2.3.2 Multiferroic property of SmFeO3 Electric hysteresis loops of SmFeO3 films on STO:Nb were collected at room temperature. When the applied voltage exceeds 1.2 V (electric field 22 kV/cm), a quasi-rectangular loop can be observed, as shown in Fig. 2.3A, indicating a strong ferroelectric property. A higher applied voltage will increase the leakage current, and a relatively round ferroelectric loop will be present (not shown). The time dependence of the applied voltage and consequent current are given in Fig. 2.3B, in which an obvious switching current can be seen, which is quite different from typical resistor or capacitor behavior. The small difference in current behavior between opposite voltages is probably due to the electrode effect (Au and Nb-STO). To check the fatigue properties, polarization-electrical (P-E) loops were measured 106 times, and the remanent polarizations are presented in Fig. 2.3C. The remanent polarization is about 0.08 μC/cm2. There was no significant change in the remanent polarization after the fatigue test, indicating the good stability of this ferroelectric material. Considering the leakage effect, which can also contribute to P-E loops, we used the PUND method to confirm the real polarization. The PUND method uses an up-updown-down wave to exclude the leakage effect. The standard PUND wave form is
(A)
(B)
(C) Fig. 2.3 (A) Electrical hysteresis loop of SmFeO3 film on SrTiO3:Nb at room temperature, (B) time dependence of applied voltage and corresponding current, (C) fatigue test results after 106 cycles (10 data points every decade). Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
30
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
Fig. 2.4 Time dependence (left) of applied voltage (ranging from 2 to 2 V) and consequent current (right) measured in the positive-up negative-down mode. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
presented in Fig. 2.4. At the beginning, the system uses a negative writing pulse to polarize the film. After that, a positive pulse V1 is applied to the film, and a current will be produced due to leakage and domain flipping. Then another positive pulse V2 is applied and a current will be produced, but only due to leakage. Therefore the integrity of the current difference produced between the up-up and down-down processes is in accordance with two times the remanent polarization, 2Pr. This also applies to the down-down process. A V3 pulse will switch the ferroelectric domains to the opposite direction, and V4 just measures the leakage current contribution. In Fig. 2.3, a clear difference in the current between I1 and I2, or I3 and I4 can be observed (since the pure leakage current peak is lower and narrower than when it is combined with the switching current), indicating that polarization in the P-E loop is real. The real current contribution from ferroelectric switching is reduced to half, however, meaning that the remanent polarization is about 0.04 μC/cm2. This value is very close to that found in orthorhombic HoMnO3 [37], in which Mn3+ spins are in E-type AFM ordering. To some extent, considering the different magnetic states involved in the orthorhombic HoMnO3 and the orthorhombic SmFeO3, this similarity suggests that the mechanism in the orthorhombic bulk SmFeO3 may not apply to the SmFeO3 thin film. The PFM image presented in Fig. 2.5 shows the presence of FE domains, which are mainly dominated by two types with opposite vectors. This means that the domain walls in this film are mostly 180° domain walls. Meanwhile, information on the surface morphology was also obtained, and the results show an average roughness of about 1.8 nm, indicative of good-quality film. Comparing the morphology with the phase pattern, it is not difficult to see that there is more than one domain in some single grains. The average domain size is around 0.1 μm, much smaller than the domain size in BiFeO3 film. One important characteristic of ferroelectricity is that the ferroelectric domains can be controlled and reversed by an external electric field, the so-called “write.” To do this we carried out PFM lithography on the SmFeO3 film. A square-in-square pattern is obtained, as shown in Fig. 2.5C, in which the inset is the pre-set pattern. The amplitude distribution along the red line across the pattern
Strain tuning effects in perovskites
31 1.0
1.0
80
0.8
10
0.8
60
Deg µm
µm
20 0 0.4
5
0.6
0 0.4
–20 –5
–40 0.2
–60
0.2 –10
–80 0.0
0.0
(A)
nm
40 0.6
0.0
0.2
0.4
0.6
0.8
0.0
1.0
0.2
0.4
0.6
0.8
1.0
µm
(B)
µm
10
8
80
80
40
6
0 4
pm
µm
20
–20 –40 2
Amplitude (pm)
60 40
0
–40
–60 –80 –80 2.0
0
(C)
4.0
6.0
8.0
Position (µm) 0
2
4
6 µm
8
10
(D)
Fig. 2.5 Nanoscale ferroelectric domain structure characterized by piezoelectric force microscope (PFM): (A) amplitude image and (B) phase image; (C) PFM lithography pattern, inset: the pre-set pattern, and (D) amplitude distribution along the red line across the lithography pattern in (C). Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
is presented in Fig. 2.5D. It is clear that these two cubic boxes show opposite amplitude values, suggesting that they are truly two types of domains with opposite polarization directions. To investigate the magnetic properties of this “pseudocubic” SmFeO3 film, we carried out measurements using a magnetic properties measurement system (MPMS). The field-cooling temperature dependence of the magnetic moment was measured from 10 to 300 K, as shown in Fig. 2.6A. A sharp ferromagnetic (FM)-like transition can be found around 185 K. The FM behavior can be confirmed by the magnetic hysteresis loop at 100 K. Curie-Weiss law fitting gives a positive
32
(A)
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
(B)
Fig. 2.6 (A) Temperature dependence of magnetic moment from 10 to 300 K, and magnetic hysteresis loop measured at 100 K (inset); (B) antiferromagnetic interaction of Fe3+ sublattice confirmed by Curie-Weiss law fitting with temperature range from 100 to 300 K. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
intercept, which corresponds to a negative Curie-Weiss temperature, indicating an AFM interaction starting below 240 K, as can been seen in Fig. 2.6B. Hence, the FM-like behavior should originate from a canted AFM ordering. This is quite different from what happens in bulk SmFeO3, in which a spin reorientation takes place around 433 K from a canted AFM weak ferromagnetic ordering to simple AFM ordering. Such a big change may be attributed to Fe3+ spin frustration, which weakens the interaction between Fe3+ ions compared with that in the orthorhombic lattice. The stress-induced distortion on the interface and boundaries should be responsible for the significant change in the magnetic properties. This result indicates that the ferroelectricity in the SmFeO3 at room temperature may not arise from the magnetic exchange striction interaction, but from the structural distortion.
2.3.3 The first principle calculation Bulk SmFeO3 has been reported to show spin-driven ferroelectric polarization at room temperature based on experimental measurements; however, the theoretical calculations deny such a possibility. Therefore the multiferroic property in SmFeO3 is still controversial. Our SmFeO3 thin film, however, shows strong structural distortion at both the film-substrate interface and the grain boundaries. Such structural distortion will cause the reconstruction of the crystal lattice and rearrangement of spin, and thus electrical polarization, driven by both displacement of ions and spin. Adopting a distorted structure of SmFeO3 at the film-substrate interface, the different magnetic states were calculated based on first principles calculations in the framework of density functional theory (DFT) and the result is presented in Table 2.1. The energies of those magnetic states were compared with G-type AFM ordering of Fe for both Hubbard U ¼ 0 and 2, which shows that G-AFM is the most stable state in comparison to A-type AFM, C-type AFM, FM, and paramagnetic (PM) states [38]. The polarization
Strain tuning effects in perovskites
33
Table 2.1 Calculated total energies of the distorted SmFeO3 system with different magnetic structures
Energya (U ¼ 0) Energya (U ¼ 2)
G-AFM
A-AFM
C-AFM
FM
PM
ΔP [100]
0
1.09
0.739
3.599
5.045
15.32 μC/cm2
0
1.66
0.723
5.372
12.274
15.31 μC/cm2
The Hubbard-U values of 0 and 2 were considered in the calculation and compared. The ferroelectric polarization was calculated for G-type AFM. a Total energy (relative to the G-AFM state in eV/f.u.) Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
is calculated as 15.3 μC/cm2 for both U values. The value of U does not show any significant effect on the magnetic state and polarization. Although the calculated polarization value is much larger than the measured value, this result indicates that the observed polarization might be caused by the structural distortion at the filmsubstrate interface. The discrepancy might be caused by the relaxation of the distorted SmFeO3 film structure to a bulk one in a relatively thick film, which has been confirmed in HRTEM images. In addition, G-AFM easily forms a canted magnetic structure [38, 39] and causes the appearance of net moment and spin-driven polarization based on an inverse Dyaloshinski-Moriya (DM) interaction. Although we cannot distinguish the two contributions to the polarization, that is, ionic-displacement-driven polarization and spin-driven polarization, studying the possibility for the existence of large magnetoelectric coupling will be the aim of the next step in our research program. Finally, we compare the calculated electron density isosurface with the value of 0.45 for undistorted and distorted SmFeO3 systems obtained from DFT calculation, which is shown in Fig. 2.7. Electron density contour is known to be an informative tool to distinguish different bonding interaction in solids. It is obvious that, in the distorted system, the electron density between oxygen and iron is strong enough to form a three-dimensional network, indicating a strong asymmetric covalent bonding interaction. While in the undistorted system, the electron density between oxygen and iron is weak. The strong asymmetric covalent bonding is responsible for the observed polarization in the distorted SmFeO3 thin film.
2.3.4 Dynamic strain tuning of the transport and magnetic property of La2/3Ca1/3MnO3 film Besides the SmFeO3, a stress tuning effect is also observed in the CMR manganite perovskites, such as LCMO film. In Fig. 2.8A, the Raman spectra of the film are dominated by the strong Raman scattering peaks from the substrate, and only a few weak peaks from LCMO can be observed due to its pseudocubic symmetry. Three peaks are located at around
34
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
(A)
(B)
Fig. 2.7 The calculated electron density isosurface with the value of 0.45 for undistorted (A) and distorted (B) SmFeO3 systems. Reproduced with permission from ACS Appl. Mater. Interface 6 (2014) 7356. Copyright 2014 American Chemical Society.
(A) 153 K 233 K
I
Log intensity (a.u.)
236.4 cm–1 173 K 163 K
(B) 153 K
III II 235.7 cm–1
233 K 200
300
400
(C) 193 K
500
600
466.4 cm–1
700 220
230
240
250
(D) T < 193 K 645.3 cm–1
183 K
T > 193 K
465.2 cm–1 450
460
470
480
560
600
640
680
720
Raman shift (cm–1)
Fig. 2.8 (A) Raman spectra of the La2/3Ca1/3MnO3/BaTiO3 film from 153 to 233 K. (B) Magnified view of peak I. (C) Magnified view of peak II. (D) Magnified view of peak III. Reproduced from Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
Strain tuning effects in perovskites
35
235, 465, and 645 cm1, which are ascribed to the y-rotation, bending, and stretching modes of MndO octahedra, respectively [40]. The two high-frequency phonons (bending and stretching modes) are associated with modifications of the MndO bond, while the low frequency one is assigned to rigid rotations of the octahedra. Fig. 2.8B shows that there is a peak shift of 0.8 cm1 to higher frequency of the y-rotation mode between 163 and 173 K. This shift takes place at a temperature lower than the temperature where the substrate phase transition occurs, and thus, it is difficult to directly attribute this shift to the substrate-phase transition. In Fig. 2.8C, the bending mode shows a shift of 1.2 cm1 to the lower frequency side at temperatures lower than 193 K, with a simultaneous increase in peak intensity. In Fig. 2.8D, the Raman scattering peak from the stretching mode disappears at temperatures above 193 K. The temperatures for the shifting of the bending mode and the vanishing of the stretching mode coincide with the temperature for the substrate structural phase transition, which indicates that the structure of the LCMO film is modulated by the BTO R-O phase transition. Fig. 2.9A shows that a PM to FM transition starts at 244 K for the sample in the 100 Oe DC field, while the transition temperature shifts to 277 K and 296 K when higher fields of 2000 Oe and 1 T are applied, respectively (M-T curve). Distinct anomalies with concomitant thermal hysteresis are observed on both the field cooled and zero-field cooled (FC-ZFC) magnetization curves between 188 and 200 K, measured at 100 Oe, where the BTO R-O phase transition and LCMO Raman scattering abnormalities occur. Similar magnetization anomalies were also observed for the FC magnetization curve at an applied magnetic field of 2000 Oe, but with a much reduced magnitude. However, such anomalies were not clearly observed for the magnetization curve measured under the highest magnetic field of 1 T, which means that a strong external magnetic field can suppress the jump in magnetization. This result
180
2000 Oe 100 Oe
200 T (K)
6.0 ´ 10–4
100
220
60
30
FC 0.0
(A)
90 100 Oe
ZFC 150
200 T (K)
250
2.5
0T
2.0 R (W)
1T
2000 Oe
M (10–6 emu) (out-of-plane)
1.8 ´ 10–3 1.2 ´ 10–3
120
1T M (in-plane)
M (emu) (in-plane)
2.4 ´ 10–3
1.5 1.0 0.5 0.0
0 300
13T 0
(B)
80
160 T (K)
240
Fig. 2.9 (A) Field cooled and zero-field cooled in-plane and out-of-plane magnetization curves of the La2/3Ca1/3MnO3(LCMO)/BaTiO3(BTO) film measured under different magnetic fields; inset is an enlargement of the indicated temperature range. (B) Temperature dependence of resistance of the LCMO/BTO film measured under 0 and 13 T. Reproduced from Appl. Phys. Lett. 99 (2011) 092103 with the permission of AIP Publishing.
36
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
demonstrates that we are able to modulate the magnetic response of LCMO via strain, but this can only happen at low magnetic fields. The out-of-plane magnetization FC and ZFC curves of the same sample at a magnetic field of 100 Oe show a magnetic moment two orders of magnitude smaller than for the in-plane case, indicating strong magnetic anisotropy. Similar to the M-T curves, thermal hysteresis is also observed in the resistivity versus temperature measurements (R-T curves), as evidenced by the resistance peak positions at 208 and 193 K for heating and cooling measurements, respectively (shown in Fig. 2.9B). In addition, an abrupt decrease in the resistance jump was observed for both the cooling and the heating processes in R-T curves. The decrease in the magnitude of the jump in the resistance between 0 and 13 T is (2.36–0.5)/2.36 ¼ 78.8%. The temperatures for the occurrence of the abrupt decrease in resistance are the same as for the occurrence of magnetization anomalies, the BTO R-O phase transition and the abnormal LCMO Raman scattering peak. This abrupt change indicates that modulation of the strain in LCMO film by substrate R-O structural transformation modifies the transport behavior as well as the magnetization. In addition, both the magnitude and sign of the resistivity changes remain constant through many thermal cycles, indicating that the structural effects are elastic and reversible. Since cracks can be easily formed in BTO, one cannot completely eliminate the possibility of substrate-induced cracking in the film. Cracking will cause an irreversible increase in the resistivity. However, the observed decrease in the resistivity in this study should not have any relationship to cracks. In CMR materials, the semiconductor-metallic state transition takes place simultaneously with the PM-FM ordering. The relationship between conductivity and magnetic ordering at and below the transition temperature is explained by a double exchange mechanism (DE) and Jahn-Teller electron-lattice coupling, for which the MndOdMn bond length and angle are the key factors [41–43]. Thermal cycling of the BTO crystal during the temperature-dependent measurements leads to the formation of different types of crystal domains. These BTO domains give rise to two different strains in a magnetic film and two different modulations of strain at the BTO ˚ (a and b axes) phase transition. The lattice parameters of BTO change from 4.013 A ˚ ˚ and 3.99 A (c axis) to 4.001 A, with a deviation of 90° to 89.85° for the ab angle at the O-R phase transition. A change in the in-plane lattice parameters will change the MndOdMn bond length and bond angle, which is evidenced by Raman peak shifting, similar to the effect caused by Jahn-Teller distortion, and will eventually change the resistance of the film, according to the DE mechanism. However, a high magnetic field of 13 T was found to eliminate the observed sharp resistance jump in R-T measurements, and the resistance in high magnetic field below the R-O phase transition temperature is much smaller than the value measured without magnetic field. This means that the magnetic field can produce similar structural distortion effects to those that are produced in LCMO film by strain from substrate phase transitions. Such an effect produced by a magnetic field, however, is much stronger than the effect caused by a substrate phase transition. It has been recognized that CMR materials experience a significant magnetostriction effect, with the magnitude of volume change up to 0.1% at the Curie temperature, TC, due to strong electron-lattice
Strain tuning effects in perovskites
37
interaction or Jahn-Teller distortion when the 3d eg electrons start to delocalize at temperatures approaching TC from the high temperature side [44–46]. Such a delocalization of 3d electrons favors DE, and an insulator to metal-like transition takes place. The changing lattice distortion in LCMO film at the BTO phase transition temperature will affect the status of the eg electrons due to strong electron-phonon interaction and accelerates the process of delocalization of eg electrons, eventually leading to a sharp decrease in resistance. Therefore the observed anomaly in R-T measurements at the R-O transition indicates the existence of a strong electron-lattice interaction in CMR materials and strong effects of the interaction on the physical properties. However, the magnitude of the change in the lattice distortion due to the substrate is less than that caused by the magnetostriction effect in LCMO film when a strong magnetic field is applied. Therefore a high magnetic field can easily eliminate the observed resistance jumping in a zero magnetic field. Similarly, the large structural distortion in a high magnetic field caused by a strong electron-lattice interaction can mask the external disturbance, and thus the magnetic moment anomaly is destroyed by a high magnetic field. It has proposed that the changing of the magnetic anisotropy and magnetic domain is the cause for the modification of the magnetization at the BTO substrate structure–phase transition in other kinds of magnetic film (non-CMR film) [47, 48]. However, for those films and such a cause a magnetic field can kill the modification of magnetization but not the resistivity, which is different from what has been observed in this report. Therefore at least the modification of the magnetic anisotropy and the magnetic domain of CMR film at the BTO-phase transition is not the only reason for the observation in this report.
2.4
Summary
In summary, the structure, and the ferroelectric and magnetic properties were studied in a SmFeO3 film on STO:Nb substrate with strong static strain. Clear structural distortion can be identified on the [010]o grown film-substrate interface, as well as boundaries of [010]o and a small fraction of [101]o grains. A strong ferromagnetic-like transition is observed around 185 K, which is likely to be due to a canted AFM ordering. The PFM images clarify the existence of ferroelectricity and the switchable domain structure. There is also a saturating hysteresis loop in the ferroelectric measurements, and the loops are very stable, even after 106 cycles. The PUND method was used to determine the intrinsic ferroelectric polarization, and the remanent polarization is about 0.04 μC/cm2, which is comparable to that in orthorhombic HoMnO3. The structural distortion in this system is probably responsible for the significant change in the magnetic properties and the small polarization as well. The theoretical calculations of the highly distorted SmFeO3 structure have theoretically confirmed the existence of the ferroelectric polarization and net magnetic moment in the thin film form of SmFeO3. In an LCMO film grown on a BTO substrate, anomalies were observed on both M-T and R-T curves at the BTO R-O-phase transition, demonstrating a strong dynamic strain tuning effect. Furthermore, both the anomalies in the magnetization and the resistance drop can be eliminated by a strong magnetic field. All these
38
Nanoscale Ferroelectric-Multiferroic Materials for Energy Harvesting Applications
phenomena indicate that colossal resistance change in CMR material at temperatures around the PM-FM transition can be produced solely by strain as well as by magnetic field.
Acknowledgments We thanks the National Natural Science Foundation of China (51571083, 51702351) and the Shenzhen Science and Technology Innovation Committee (JCYJ20170307165829951, JCYJ20170413152832151) for support.
References [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]
A.S. Disa, et al., Phys. Rev. Lett. 114 (2015). 026801. H.Y. Hwang, et al., Nat. Mater. 11 (2012) 103. G. Catalan, J.F. Scott, Adv. Mater. 21 (2009) 2463. C.A.F. Vaz, J. Hoffman, C.H. Anh, R. Ramesh, Adv. Mater. 22 (2010) 2900. P. Jain, V. Ramachandran, R.J. Clark, H.D. Zhou, B.H. Toby, N.S. Dalal, H.W. Kroto, A. K. Cheetham, J. Am. Chem. Soc. 131 (2009) 13625. H. Cui, Z. Wang, K. Takahashi, Y. Okano, H. Kobayashi, A. Kobayashi, J. Am. Chem. Soc. 128 (2006) 15074. S.W. Kim, H.Y. Chang, P.S. Halasyamani, J. Am. Chem. Soc. 132 (2010) 17684. G.A. Smolenskiı˘, I.E. Chupis, Sov. Phys. Uspekhi 25 (1982) 475. N. Zhang, K.F. Wang, S.J. Luo, T. Wei, X.W. Dong, S.Z. Li, J.G. Wan, J.M. Liu, Appl. Phys. Lett. 96 (2010) 252902. T. Kimura, G. Lawes, T. Goto, Y. Tokura, A.P. Ramirez, Phys. Rev. B 71 (2005) 224425. T. Kimura, S. Ishihara, H. Shintani, T. Arima, K.T. Takahashi, K. Ishizaka, Y. Tokura, Phys. Rev. B 68 (2003) 060403. M. Mostovoy, Phys. Rev. Lett. 96 (2006) 067601. H. Lueken, Angew. Chem. Int. Ed. 47 (2008) 8562. T. Matsumoto, R. Ishikawa, T. Tohei, H. Kimura, Q. Yao, H. Zhao, X. Wang, D. Chen, Z. Cheng, N. Shibata, Y. Ikuhara, Nano Lett. 13 (2013) 4594. Y. Tokunaga, S. Iguchi, T. Arima, Y. Tokura, Phys. Rev. Lett. 101 (2008) 097205. Y. Du, Z.X. Cheng, X.L. Wang, S.X. Dou, J. Appl. Phys. 107 (2010) 09D908. J.-H. Lee, Y.K. Jeong, J.H. Park, M.-A. Oak, H.M. Jang, J.Y. Son, J.F. Scott, Phys. Rev. Lett. 107 (2011) 117201. R.D. Johnson, N. Terada, P.G. Radaelli, Phys. Rev. Lett. 108 (2012) 219701. J.-H. Lee, Y.K. Jeong, J.H. Park, M.-A. Oak, H.M. Jang, J.Y. Son, J.F. Scott, Phys. Rev. Lett. 108 (2012) 219702. G. Zhang, S. Dong, Z. Yan, Y. Guo, Q. Zhang, S. Yunoki, E. Dagotto, J.M. Liu, Phys. Rev. B 84 (2011) 174413. R.D. Johnson, L.C. Chapon, D.D. Khalyavin, P. Manuel, P.G. Radaelli, C. Martin, Phys. Rev. Lett. 108 (2012) 067201. X.Z. Lu, M.H. Whangbo, S. Dong, X.G. Gong, H.J. Xiang, Phys. Rev. Lett. 108 (2012) 187204. H.Y. Hwang, T.T.M. Palstra, S.-W. Cheong, B. Batlogg, Phys. Rev. B 52 (1995) 15046. J.M. Coey, M. Viret, S. von Molnar, Adv. Phys. 48 (1999) 167. M.R. Ibarra, P.A. Algarabel, C. Marquina, J. Basco, J. Garcia, Phys. Rev. Lett. 75 (1995) 3541.
Strain tuning effects in perovskites
39
[26] S. Jin, T.H. Tiefel, M. McCormack, H.M. O’Bryan, R. Ramesh, D. Schurig, Appl. Phys. Lett. 67 (1995) 557. [27] C. Zener, Phys. Rev. 82 (1951) 403. [28] A.J. Millis, P.B. Littlewood, B.I. Shraiman, Phys. Rev. Lett. 74 (1995) 514. [29] H.Y. Hwang, S.-W. Cheong, P.G. Radaelli, M. Marezio, B. Batlogg, Phys. Rev. Lett. 75 (1995) 914. [30] L. Nagaev, Phys. Stat. Sol. B 9 (1994) 186. [31] J.M. Teresa, M.R. Ibarra, P.A. Algarabel, et al., Nature 386 (1997) 256. [32] J.B. Goodenough, J.S. Zhou, Nature 386 (1997) 229. [33] H. Aliaga, D. Magnoux, A. Moreo, D. Poilblanc, S. Yunoki, E. Dagotto, Phys. Rev. B 68 (2003) 104405. [34] K. Lai, M. Nakamura, W. Kundhikanjana, M. Kawasaki, Y. Tokura, M.A. Kelly, Z. X. Shen, Science 329 (2010) 190. [35] M.K. Singh, H.M. Jang, S. Ryu, M.-H. Jo, Appl. Phys. Lett. 88 (2006) 042907. [36] J.F. Ihlefeld, A. Kumar, V. Gopalan, D.G. Schlom, Y.B. Chen, X.Q. Pan, T. Heeg, J. Schubert, X. Ke, P. Schiffer, J. Orenstein, L.W. Martin, Y.H. Chu, R. Ramesh, Appl. Phys. Lett. 91 (2007) 071922. [37] Y.S. Chai, Y.S. Oh, L.J. Wang, N. Manivannan, S.M. Feng, Y.S. Yang, L.Q. Yan, C.Q. Jin, K.H. Kim, Phys. Rev. B 85 (2012) 184406. [38] W.H. Meiklejohn, C.P. Bean, Phys. Rev. 102 (1956) 1413. [39] W.H. Meiklejohn, C.P. Bean, Phys. Rev. 105 (1957) 904. [40] M.N. Iliev, M.V. Abrashev, H.G. Lee, V.N. Popov, Y.Y. Sun, C. Thomsen, R.L. Meng, C. W. Chu, Phys. Rev. B 57 (1998) 2872. [41] P.W. Anderson, H. Hasegawa, Phys. Rev. 100 (1955) 675–681. [42] P.-G. de Gennes, Phys. Rev. 118 (1960) 141–154. [43] Y. Tokura, N. Nagaosa, Science 21 (2000) 462. [44] J.M. De Teresa, M.R. Ibarra, J. Blasco, J. Garcı´a, C. Marquina, P.A. Algarabel, Phys. Rev. B 54 (1996) 1187. [45] A.I. Abramovich, L.I. Koroleva, A.V. Michurin, O.Y. Gorbenko, A.R. Kaul, Phys. Solid State 42 (2000) 1494. [46] M.R. Ibarra, P.A. Algarabel, C. Marquina, J. Blasco, J. Garcı´a, Phys. Rev. Lett. 75 (1995) 3541. [47] R.V. Chopdekar, Y. Suzuki, Appl. Phys. Lett. 89 (2006) 182506. [48] H.F. Tian, T.L. Qu, L.B. Luo, J.J. Yang, S.M. Guo, H.Y. Zhang, Y.G. Zhao, J.Q. Li, Appl. Phys. Lett. 92 (2008) 063507.
Further Reading [49] Z.X. Cheng, F. Hong, Y.X. Wang, K. Ozawa, H. Fujii, H. Kimura, Y. Du, X.L. Wang, S. X. Dou, ACS Appl. Mater. Interface 6 (2014) 7356. [50] Z.X. Cheng, X.L. Wang, S.X. Dou, M. Osada, H. Kimura, Appl. Phys. Lett. 99 (2011) 092103.