• Tidak ada hasil yang ditemukan

Local electrical conduction in

N/A
N/A
Protected

Academic year: 2023

Membagikan "Local electrical conduction in"

Copied!
7
0
0

Teks penuh

(1)

Nanotechnology

PAPER

Local electrical conduction in polycrystalline La- doped BiFeO 3 thin films

To cite this article: Ming-Xiu Zhou et al 2013 Nanotechnology 24 225702

View the article online for updates and enhancements.

Related content

Photovoltaic effect in transition metal modified polycrystalline BiFeO3 thin films Venkata Sreenivas Puli, Dhiren Kumar Pradhan, Rajesh Kumar Katiyar et al.

-

Variation of electric properties across the grain boundaries in BiFeO3 film Bojan Stojadinovi, Borislav Vasi, Dimitrije Stepanenko et al.

-

Switchable Ferroelectric Diode Effect and Piezoelectric Properties of Bi0.9La0.1FeO3 Ceramics

Gong Yu-Fei, Wu Ping, Liu Wei-Fang et al.

-

Recent citations

High-Performance Photovoltaic Readable Ferroelectric Nonvolatile Memory Based on La Doped BiFeO3 Films

Dong Li et al -

Unipolar resistive switching in porous perovskite-like Bi3.15Nd0.85Ti3O12 thin films

H. J. Song et al -

Local polarization switching in Ba–Ni co- doped BiFeO3 thin films with low rhombohedral-symmetry distortion G. Rojas-George et al

-

This content was downloaded from IP address 218.94.142.95 on 15/12/2019 at 07:28

(2)

Nanotechnology24(2013) 225702 (6pp) doi:10.1088/0957-4484/24/22/225702

Local electrical conduction in

polycrystalline La-doped BiFeO 3 thin films

Ming-Xiu Zhou, Bo Chen, Hai-Bin Sun, Jian-Guo Wan, Zi-Wei Li, Jun-Ming Liu, Feng-Qi Song and Guang-Hou Wang

National Laboratory of Solid State Microstructures, and Department of Physics, Nanjing University, Nanjing 210093, People’s Republic of China

E-mail:wanjg@nju.edu.cn

Received 31 January 2013, in final form 12 April 2013 Published 1 May 2013

Online atstacks.iop.org/Nano/24/225702 Abstract

Local electrical conduction behaviors of polycrystalline La-doped BiFeO3thin films have been investigated by combining conductive atomic force microscopy and piezoelectric force microscopy. Nanoscale current measurements were performed as a function of bias voltage for different crystal grains. Completely distinct conducting processes and resistive switching effects were observed in the grain boundary and grain interior. We have revealed that local electric conduction in a grain is dominated by both the grain boundary and ferroelectric domain, and is closely related to the applied electric field and the as-grown state of the grain.

At lower voltages the electrical conduction is dominated by the grain boundary and is associated with the redistribution of oxygen vacancies in the grain boundary under external electric fields. At higher voltages both the grain boundary and ferroelectric domain are responsible for the electrical conduction of grains, and the electrical conduction gradually extends from the grain boundary into the grain interior due to the extension of the ferroelectric domain towards the grain interior. We have also demonstrated that the conduction dominated by the grain boundary exhibits a much small switching voltage, while the conduction of the ferroelectric domain causes a much high switching voltage in the grain interior.

(Some figures may appear in colour only in the online journal)

1. Introduction

BiFeO3 (BFO) is a kind of lead-free multiferroic material, which has large ferroelectric polarization and weak ferro- magnetism at room temperature, promising to be a candidate for magneto-electric energy conversion devices [1]. Recently, remarkable resistive switching (RS) and switchable ferroelec- tric diode effects have also been observed in monodomain BFO crystals and some BFO-based systems (e.g. Ca-doped BiFeO3 films [2], BiFeO3/Nb–SrTiO3 heterojunctions [3], etc), making BFO even more attractive for applications in some other technologically demanding fields such as non-volatile resistive random access memory (RRAM) [2–4].

Pure BFO is well known to be a perovskite structure;

nevertheless, it is not facile to obtain pure BFO phase in

an actual preparation process due to the easy formation of impurity phases such as Bi2O3 and Bi2Fe4O9 which seriously influence the electric properties of BFO. In order to overcome this obstacle, many efforts have been made to prepare the BFO-based compounds by doping ions into BFO, e.g. the fabrication of A-site-ion-doped BFO compounds such as Bi1−xLaxFeO3 and Bi1−xNdxFeO3, and B-site-ion-doped BFO compounds such as BiFe1−xMnxO3 and BiFe1−xTixO3 [5–8]. The electrical conduction behavior of the BFO-based compound is of great importance for its RS effect since it is directly related to the power consumption of devices and especially affects the duration of charge storage in the application of RRAM devices [3, 9]. In practical applications, polycrystalline BFO-based films are usually used due to the facility of preparation,

(3)

Nanotechnology24(2013) 225702 M-X Zhouet al

nevertheless, their electrical conduction behaviors are quite complicated. The presence of grain boundaries makes their electrical conduction behaviors different from those in single-crystal films. The ferroelectric polarization and defects such as oxygen vacancies also play important roles in the electrical conduction behaviors of polycrystalline BFO-based films [10]. Although a large number of investigations have demonstrated that the doping of ions into BFO is effective for adjusting the electric conduction behavior of the polycrystalline BFO-based films, most of them mainly focus on the macroscopic scale so far. To perform the measurements of electrical conduction, the films are usually covered by a metal electrode with fairly large area. So, it is impossible for such a configuration to spatially resolve the local conduction behaviors at the nanoscale.

Local observations at the nanoscale are invaluable for understanding the electric conduction behaviors of polycrystalline BFO-based films, and are also significant for developing high-performance integrated electronic devices based on BFO. Scanning probe microscopy (SPM) with an electric conducting tip (i.e. conductive atomic force mi- croscopy) has been frequently used in such investigations [2, 11–14]. In recent years, local conduction behaviors in some BFO systems have been studied by this technique [9,15–17];

however, most of them are involved in single-crystal films. Up to now, no clear consensus has emerged on the microscopic origin of conduction in polycrystalline BFO-based films.

Herein we report on a local probe into electrical conduction of polycrystalline La-doped BFO films. The 10% La-doped BiFeO3 (BLFO) thin films are chosen for investigations based on the following considerations. Firstly, the substitution of Bi3+ at the A-site by La3+ is more favorable for stabilizing the perovskite phase because La3+

and Bi3+ have almost the same radius, i.e. 1.032 A˚ and 1.030 ˚A, respectively. Previous investigations have demonstrated that the doping of La at the Bi site in BFO can effectively avoid the formation of impurity phases, and also greatly improve the crystallization and stabilize the crystal structure, consequently leading to the improvement of the electric properties of BFO [18]. Secondly, with a 10%

doping concentration of La, BLFO can still maintain the rhombohedral perovskite structure withR3csymmetry similar to pure BFO while exhibiting better ferroelectric properties and a smaller leakage current [19]. By combining conductive atomic force microscopy (c-AFM) and piezoelectric force microscopy (PFM) to observe local electrical conduction characteristics and the extension of domain polarization at the nanoscale for the BLFO samples, we have revealed that the electrical conduction behaviors of the crystal grains are dominated by both the grain boundary and ferroelectric domain, closely related to the applied electric field and the as-grown state of the grains. Our results point to a feasible avenue for modulating local electrical conduction behaviors of polycrystalline BFO-based thin films, and are likely to be universal for the other ferroelectric and multiferroic materials.

Figure 1. (a) XRD patterns of the 10% La-doped BFO film.

(b) Schematic of the test diagram. (c) Topography of a

3µm×3µm area. (d) c-AFM measurement at a bias voltage of

−6 V. (e) Average current as a function of bias voltage in the whole measured 3µm×3µm area. (f) Current profiles extracted from (c) along the green solid line at various bias voltages.

2. Experimental details

La-doped (10%) BiFeO3 (BLFO) thin films (185 nm thick) were prepared by a sol–gel process on Pt(111)/Ti/SiO2/ Si(100) wafers. The details of preparation can be found elsewhere [20]. Structural characterizations of the films were carried out by x-ray diffraction (XRD) on a D/MAX- RA diffractometer using Cu Kα radiation. The surface morphology of the films was measured using atomic force microscopy (AFM) (NT-MDT Inc.), and local electrical conductivity measurements were performed using a c-AFM equipped with a CrPt-coated tip. Figure 1(b) shows the schematic setup employed for the c-AFM process. The film surface was scanned in contact mode and the scan frequency was 0.5 Hz. The tip which acted as a mobile top electrode was grounded, and various negative bias voltages were applied to the bottom electrode. PFM measurement was also carried out to observe the change of the ferroelectric domains in the film under an external electric field. All the measurements were performed at room temperature and under ambient conditions.

3. Results and discussion

Figure 1(a) presents the XRD patterns of the BLFO film. Except for the (012), (110) and (024) peaks of the rhombohedral perovskite structure of the BFO phase, no

2

(4)

et al

Figure 2. (a) Topography of the test area. (b)–(f) c-AFM images measured at bias voltages from−4 to−6 V with a step of 0.5 V.

(g) Current profiles extracted from (b) to (f) along the yellow line with an arrow in (f). (h)I–Vhysteresis at dot A and B marked in (f).

other additional or intermediate phase peaks are observed.

Moreover, the XRD patterns suggest that the film has a preferential (012) crystallographic orientation. The local electrical conductivity measurements were conducted on an area of 3µm×3 µm. Figure 1(c) presents a typical AFM topographic image of the film. It is seen that the film exhibits hill–valley topography with average grain size of∼100 nm.

We then applied a bias voltage gradually increasing from 0 to−10 V to observe the change in the local c-AFM current map. The measured c-AFM current map series revealed that the c-AFM contrast was bias dependent. Figure1(d) shows a typical local c-AFM current map at a bias of −6 V, in which different conducting paths emerge. Figure 1(e) plots the average current in the whole measured area as a function of bias. It is seen that at lower voltages (below−4 V) the current is small, indicating that few conducting paths form.

The electric conduction quickly rises as the bias is beyond

−4 V and almost reaches saturation near−8 V. To clearly elucidate the local conduction process of the film, the current versus position profiles taken from the c-AFM current maps at various biases along a green solid line (shown in figure1(d)) were further depicted, as shown in figure1(f). It is clear that the local conduction changes with the changes of both position and bias. As the bias changes, the currents in some areas change sharply, while they vary gently in the other areas.

Moreover, the current peaks prefer to concentrate around the grain boundaries. In addition, some grains exhibit large conductivity while the others do not. All of these indicate that the local conduction behavior of BLFO film is strongly related to its polycrystalline microstructures, e.g. surface morphology, grain boundary, defects, and so on [12].

From figure 1(d), we notice that plenty of conduction channels are visible between grain clusters, indicating that the grain boundaries might play an important role in the local conduction process of BLFO film. To understand in depth the microscopic conduction process associated with the grain boundaries, we carried out c-AFM measurements in a typical area only containing a few grain clusters (shown in figure2(a)) with a voltage stepwise increase of 0.5 V from−4 to−6 V. A remarkable bias dependence of local conduction in the grain boundaries was observed, as shown in figures2(b)–(f). With increasing bias, the current contrast in the grain boundary gradually becomes intense. We then chose a smaller area only covering one grain (blue dashed ellipse in figure 2(a)) and constructed conduction line profiles scanned along a yellow solid line which was exactly across the grain. Figure2(g) plots a series of line profiles at various bias voltages. Two sharp current peaks are observed in the grain boundaries, while a small current platform appears in the grain interior. As the bias increases, currents in both grain boundaries and the grain interior increase, and their differences become more evident at high bias. It is worth noticing that the peak widths of currents in the grain boundaries hardly change with the change of bias.

For a deeper understanding of the microscopic electrical conduction process in the grain boundaries, we further performed current (I) versus voltage (V) measurements exactly in the grain boundary (dot A shown in figure 2(f)).

For comparison, similar measurement was also conducted in the grain interior close to the grain boundary (dot B shown in figure2(f)). As plotted in figure2(h), both grain boundary and grain interior show Schottky-likeI–Vbehaviors, nevertheless, the grain boundary exhibits lower switching voltage than the grain interior, indicative of a lower Schottky barrier in the grain boundary [13, 21]. In general, there are a large number of oxygen vacancies in a polycrystalline film, which are induced during the preparation process [10]. Upon the application of external electric fields, the oxygen vacancies will give rise to redistribution, causing the oxygen vacancy concentration in the grain boundary to be much higher than that in the grain interior [10]. As a result, the conduction channels appear first in the grain boundaries, with a lower switching voltage due to a small Schottky barrier dominated by the accumulating oxygen vacancies in the grain boundary.

As the bias increases, the oxygen vacancies will migrate through the film to find a new thermodynamic equilibrium, accompanied by continuous variation of oxygen vacancy distribution [4]. The currents in the grain boundaries increase more rapidly than in the grain interior because the rate of oxygen vacancy diffusion along the grain boundaries is about five to six orders of magnitude larger than that through the crystal-cell interior [22]. In addition, clustering of oxygen vacancies may lower the energy levels of oxygen vacancies by about a factor of two, which has been confirmed in some other material systems such as SrTiO3 [23]. Thereby it is also possible that clustering of oxygen vacancies partially contributes to the increased currents in both the grain boundary and the interior of the film.

From the current map shown in the blue square of figure1(d), we can observe another different conduction mode

(5)

Nanotechnology24(2013) 225702 M-X Zhouet al

Figure 3. (a) Topography and the corresponding c-AFM images at bias voltages of−4,−6 and−8 V. (b) Current profiles extracted from (a) along the green dashed line at various bias voltages.

(c)I–Vhysteresis at dots A, B and C marked in (a).

in the grain boundary, i.e. remarkable widening of conduction channel in the grain boundary. Figures 3(a) and (b) further present the evolution of the width of a conduction channel in the grain boundary with the change of bias. It is seen that as the bias increases to a critical value of about−6 V, the current in the grain boundary quickly rises; meanwhile, the current peak remarkably broadens, which becomes more evident as the bias is beyond−6 V. Moreover, as plotted in figure3(c), when the measuring tip is at the grain boundary (dot A) the current quickly increases even at very low bias, exhibiting Schottky-like characteristics dominated by the oxygen vacancy migration. However, once the tip moves away from the grain boundary and slightly enters into the grain interior (e.g. dot B or dot C), an evident and large threshold voltage appears, implying that another conduction process occurs. This is very different from the case shown in figure2.

Considering that the present BLFO film is a ferroelectric material, we suggest that the electrical conduction process in the above grain boundary might be accompanied by the extension of the ferroelectric domain under the bias, which could greatly affect the conduction behavior. To demonstrate this, we subsequently chose a typical region (in the green dotted circle shown in figure 4(a)) and carried out simultaneous measurements of c-AFM (figure4(b)) and PFM (figure 4(c)). A bias of −6.5 V was applied which was beyond the ferroelectric switching field so that the ferroelectric domains can be reversed. From figure4(c), we can observe a bright ferroelectric domain area, which is similar to the electric conduction area (bright area shown in figure4(b)). For better comparison, we plotted the line profiles of both PFM image and c-AFM current map, as shown in figure 4(d), and also plotted the height profile of the AFM

Figure 4. (a) Topography of the test area. (b) c-AFM image scanned at−6 V and (c) subsequent PFM measurement. (d) Line profiles of c-AFM current and PFM amplitude and

(e) corresponding topography profile as indicated by the blue line in green circle in (a)–(c).

image along the same line, as shown in figure 4(e). It is clear that the peaks of c-AFM conduction and PFM amplitude almost superpose each other, both being exactly located at the boundary between two grains.

Such dependence of electric conduction in grain boundaries on the ferroelectric domain extension at higher voltages is further analyzed as follows. It is known that the formation and extension of ferroelectric domains in a polycrystalline ferroelectric material are usually determined by some factors such as grain growth direction, self-consistent electrostatic interaction, ferroelectric polarization, applied electric field, and so on [24]. Upon the application of external electric fields beyond the coercive field, the growth of different oriented domains occurs, preferably at the surface of existing domain walls or grain boundaries [25–28]. As the voltage increases, the ferroelectric domains gradually grow up, accompanied by the domain wall extension towards the grain interior [26]. The localized conductivity thus appears in the domain wall as a response of the whole polarization structure to the electric field. The superposition of conduction of both ferroelectric domain wall and grain boundary consequently results in an increase in total conduction;

meanwhile, the merging of domain walls in the grain boundary causes the conduction channel in the original grain boundary to become broader. However, on the other hand, the existence of defects such as oxygen vacancies in some grain boundaries might pin the domain and prevent the switching of polarization and domain extension [27]. Therefore, for these grains it is difficult to involve the domain extension, so the electric conduction in them is still dominated by the grain boundaries, which is corresponding to the case shown in

4

(6)

et al

figure2. In addition, it should be pointed out that the existence of impurities may also influence the electrical conduction of the samples. In general, the formation of second phases or parasitic phases such as Bi2O3 and Fe oxides in BFO would cause the appearance of additional defects [6]. These defects can bring about the charge imbalance in BFO and act as conduction channels, leading to an increased leakage current. Nevertheless, according to the XRD results (shown in figure1(a)), the number of impurities in the present BLFO films is greatly reduced due to the doping of La. Therefore, the influence of trace amounts of impurities on the electric properties of the samples is actually very limited in this work.

We now turn to study the local resistive switching effect in the BLFO film, which should be closely associated with the electrical conduction behaviors of grain boundaries and ferroelectric domains. As seen in figure 3(c), the I–V characteristics of both the grain boundary and ferroelectric domain are strongly nonlinear and asymmetric. At positive bias, we do not observe current until much higher voltages, while we observe asymmetric hysteresis loops at negative bias, which should be caused by the different effective Schottky barrier heights of the nonidentical configurations between the tip and bottom electrodes [22]. Meanwhile, we find that the I–V characteristics of grain boundaries are significantly different from those of the ferroelectric domains.

The grain boundary (curve A in figure 3(c)) exhibits a very slim I–V hysteresis loop with a much lower switching voltage of only about−2 V, whereas theI–V characteristics dominated by the ferroelectric domain (curve C in figure3(c)) show strong hysteresis with a much higher switching voltage of about−6 V.

We further carried out local I–V measurements in two different types of grains in order to explore in depth the resistive switching characteristics of the grains in the BLFO film, as shown in figure5(a). For the grain marked by the solid circle (shown in the inset of figure5(a)), it exhibits different I–Vcharacteristics in different regions. The switching voltage continuously varies with the change of measuring position in the grain. Figure5(b) further plots the switching voltage as a function of measuring position (along a line from dot M to N) in this grain. The middle of line MN is assigned as coordinate zero. It is seen that when the tip sweeps from the boundary to the center of the grain, the switching voltage increases, from the smallest value of 3.5 V in the grain boundary, until it reaches the maximum value of 9.7 V at the grain center. Such a gradual increase of switching voltage from the grain boundary to the center reflects the extension of the domain wall towards the grain interior with increasing bias. Furthermore, we characterized the local normalized conductivity for this grain along the line MN at various bias voltages, as shown in figure5(c). It is seen that at lower voltages (below−5 V) the whole grain is only locally conductive, dominated by the grain boundary. At higher voltages (beyond−6 V), however, the conductive region in the grain becomes large, determined by both the grain boundary and the ferroelectric domain. The conduction in the grain boundary quickly reaches saturation, even at a lower voltage of about−6 V, while the conduction in the grain interior gradually increases and extends towards

Figure 5. (a)I–Vhysteresis measured at different positions in one grain which is marked by a solid ellipse in the inset of (a).

(b) Variation of switching voltage of conduction with measurement position in the grain. (c) Normalized conductivity of the grain along line MN at various bias voltages.

the center, which is dominated by the domain wall extension.

At −10 V, the whole grain is highly conductive due to the complete reversal of the domain.

The low switching voltage in the grain boundary is actually determined by the accumulating state of defects such as oxygen vacancies. Under external electric fields, the oxygen vacancies align in an orderly manner as vacancy clusters, accumulating in the grain boundaries and forming one-dimension-like chains or branches [21]. These vacancy clusters could act as preferential conduction channels, thereby causing the grain boundaries to have good electric conductivity even at low voltages due to their good mobility under electric field. The case of I–V switching behaviors dominated by the ferroelectric domains is complicated, and depends on the alignment of polarization vectors associated with the specific orientation of the grains [26]. In general, the grain orientations in the film are initially determined at the stage of nucleation and growth, which greatly influences the configuration of initial domains. Upon the application of external electric fields, the domains start to grow from the existing domain walls or grain boundaries, gradually extending over multiple grains [26–28]. In this process, the conduction occurs at the domain wall [29–31]. The domain wall angle, i.e. the relative angle between the orientations of polarization axes of the adjacent domains, plays a key role in such a conduction process [11,27]. The angular-dependent domain conductance is determined to a great extent by the

(7)

Nanotechnology24(2013) 225702 M-X Zhouet al

as-grown state of the grains [32]. For those grains which can be easily combined with the adjacent grains and reversed together under an external electric field to form a polarized domain, localized conductivity can easily occur in them as a response of the whole polarization structure to the electric field. Maximum conduction of a domain can be expected when the domain wall is oriented parallel to the direction of the electric field. Nevertheless, for those grains which are pinned, it is difficult for them to participate in the domain extension, so they might be almost non-conductive.

To demonstrate it, we also measured the conduction behavior for another grain marked by a dashed circle (shown in the inset of figure 5(a)). We observe almost no current until a high negative voltage of−8 V, implying that the grain does not incur any domain movement due to the pinning effect probably as a result of some factors such as grain growth orientation, restriction from the substrate, defects, and so on. Nevertheless, we found that when the applied voltage reached−10 V the grain became conductive since the voltage was high enough to involve it in reversal of the ferroelectric domain.

4. Conclusion

In conclusion, local electrical conduction behavior and the resistive switching effect of polycrystalline La-doped BFO films have been investigated by combining conductive atomic force microscopy and piezoelectric force microscopy.

At lower voltages the electric conduction behavior is dominated by the grain boundary, which is associated with the redistribution of oxygen vacancies under external electric fields. At higher voltages, both the grain boundary and ferroelectric domain are responsible for the electric conduction of the grains. The current–voltage characteristics dominated by the electric conduction of the grain boundary exhibit a much small switching voltage, while the conduction of the ferroelectric domain causes a much high switching voltage in the grain interior. We have also demonstrated that the electrical conduction and resistive switching behaviors of the grains are related to their as-grown state. From the point of view of applications, our results point out a feasible avenue for modulating local electrical conduction behaviors of polycrystalline BFO-based films, which are also applicable for the other ferroelectric and multiferroic materials. By carefully tailoring the microstructures such as grain growth direction and grain size, one can easily regulate the number of grain boundaries and the ferroelectric domain extension under an electric field, consequently tuning the local conduction behavior as well as the resistive switching effect of the whole film.

Acknowledgments

This work was supported by the National Key Projects for Basic Research of China (grant nos 2010CB923401, 2009CB623303), the National Natural Science Foundation of China (grant nos 11134005, 50972055) and the PAPD project.

References

[1] Garcia V, Fusil S, Bouzehouane K, Enouz-Vedrenne S, Mathur N D, Barthelemy A and Bibes M 2009Nature 46081

[2] Yang C H, Seidel J, Kim S, Rossen P, Yu P, Gajek M, Chu Y H, Martin L W, Holcomb M and He Q 2009Nature Mater.8485

[3] Qu T, Zhao Y, Xie D, Shi J, Chen Q and Ren T 2011Appl.

Phys. Lett.98173507

[4] Choi T, Horibe Y, Yi H, Choi Y J, Wu W and Cheong S W 2010Nature Mater.9253

[5] Singh S K and Ishiwara H 2006Japan. J. Appl. Phys.453194 [6] Kawae T, Terauchi Y, Tsuda H, Kumeda M and Morimoto A

2009Appl. Phys. Lett.94112904

[7] Kim J K, Kim S S, Kim W J, Bhalla A S and Guo R 2006 Appl. Phys. Lett.88132901

[8] Wang Y and Nan C W 2006Appl. Phys. Lett.89052903 [9] Shen W, Bell A, Karimi S and Reaney I M 2012Appl. Phys.

Lett.100133505

[10] Lazenka V, Ravinski A, Makoed I, Vanacken J, Zhang G and Moshchalkov V 2012J. Appl. Phys.111123916

[11] Lang K, Hite D, Simmonds R, McDermott R, Pappas D and Martinis J M 2004Rev. Sci. Instrum.752726

[12] Kremmer S, Wurmbauer H, Teichert C, Tallarida G, Spiga S, Wiemer C and Fanciulli M 2005J. Appl. Phys.97074315 [13] Seidel J, Martin L W, He Q, Zhan Q, Chu Y H, Rother A,

Hawkridge M, Maksymovych P, Yu P and Gajek M 2009 Nature Mater.8229

[14] Seidel J, Maksymovych P, Batra Y, Katan A, Yang S Y, He Q, Baddorf A P, Kalinin S V, Yang C H and Yang J C 2010 Phys. Rev. Lett.105197603

[15] Guyonnet J, Gaponenko I, Gariglio S and Paruch P 2011Adv.

Mater.235377

[16] Farokhipoor S and Noheda B 2011Phys. Rev. Lett.

107127601

[17] He Q, Yeh C H, Yang J C, Singh-Bhalla G, Liang C W, Chiu P W, Catalan G, Martin L, Chu Y H and Scott J 2012 Phys. Rev. Lett.10867203

[18] Cheng Z, Li A, Wang X, Dou S X, Ozawa K, Kimura H, Zhang S and Shrout T R 2008J. Appl. Phys.10307E507 [19] Karimi S, Reaney I, Han Y, Pokorny J and Sterianou I 2009

J. Mater. Sci.445102

[20] Pintilie L, Dragoi C, Radu R, Costinoaia A, Stancu V and Pintilie I 2010Appl. Phys. Lett.96012903

[21] Guo X and Waser R 2006Prog. Mater. Sci.51151 [22] Szot K, Speier W, Bihlmayer G and Waser R 2006Nature

Mater.5312

[23] Marshall M S J, Becerra-Toledo A E, Marks L D and Castell M R 2011Phys. Rev. Lett.10786102

[24] Shur V Y, Rumyantsev E, Nikolaeva E and Shishkin E 2000 Appl. Phys. Lett.773636

[25] Duiker H and Beale P D 1990Phys. Rev.B41490 [26] Pramanick A, Prewitt A D, Forrester J S and Jones J L 2012

Crit. Rev. Solid State37243

[27] Wang Y and Nan C W 2008J. Appl. Phys.103114104 [28] Kuroda A, Kurimura S and Uesu Y 1996Appl. Phys. Lett.

691565

[29] P´erez R, Garc´ıa J E, Albareda A and Ochoa D A 2007J. Appl.

Phys.102044117

[30] Shu W, Wang J and Zhang T Y 2012J. Appl. Phys.

112064108

[31] Zhang W and Bhattacharya K 2005Acta Mater.53199 [32] Meier D, Seidel J, Cano A, Delaney K, Kumagai Y,

Mostovoy M, Spaldin N A, Ramesh R and Fiebig M 2012 Nature Mater.11284

6

Referensi

Dokumen terkait

From result of linearity test by using SPSS 16 obtained degree of significance test of linearity of organizational culture variable with commitment to

The reliability test gives information that all instruments were used to measure the seven traits of critical thinking are reliable. Table 2, which is the reliability