Appl Phys A (2012) 108:337–342DOI 10.1007/s00339-012-7011-6

Enhanced memory window of Au/BaTiO3/SrTiO3/Si(001) MFISstructure with high c-axis orientation for non-volatile memoryapplications

Gasidit Panomsuwan · Osamu Takai · Nagahiro Saito

Received: 15 February 2012 / Accepted: 28 May 2012 / Published online: 12 June 2012© Springer-Verlag 2012

Abstract Metal–ferroelectric–insulator–semiconductor(MFIS) structures with BaTiO3 (BTO) as a ferroelectric filmand SrTiO3 (STO) as an insulating buffer layer were fabri-cated on p-type Si(001) substrates using an ion beam sputterdeposition technique. The effect of out-of-plane orientationon the electrical properties of the MFIS structures, includingleakage current density and memory window behavior, werestudied using the growth of the BTO ferroelectric film on Sisubstrate buffered by highly c-axis-oriented and random-oriented STO buffer layers. The experimental results showthat the out-of-plane orientations of the BTO films were al-most identical to those of the STO buffer layers. The MFISstructure with a high c-axis orientation exhibited a max-imum clockwise capacitance-voltage memory window of1.17 V with a low leakage current of 1.05 × 10−7 A/cm2 atan applied voltage of 4 V, which is a significant improve-ment compared to the MFIS structure with a random ori-entation. The difference in the electrical properties of theMFIS structures with both types of orientation is discussedin detail. The results obtained from this study indicate that

G. Panomsuwan (�) · O. Takai · N. SaitoDepartment of Materials, Physics, and Energy Engineering,Graduate School of Engineering, Nagoya University, Furo-cho,Chikusa-ku, Nagoya, Japan, 464-8603e-mail: gasidit@eco-t.esi.nagoya-u.ac.jpFax: +8152-789-3260

N. SaitoEcoTopia Science Institute, Nagoya University, Furo-cho,Chikusa-ku, Nagoya, Japan, 464-8603e-mail: hiro@eco-t.esi.nagoya-u.ac.jp

N. SaitoGreen Mobility Collaborative Research Center,Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Japan,464-8603

the Au/BTO/STO/p-Si MFIS structure with high c-axis ori-entation has good potential for use in non-volatile memoryapplications.

1 Introduction

Recently, ferroelectric random access memory (FRAM) hasbeen received much attention due to its considerable ad-vantages such as non-destructive readout, high-speed oper-ation, and high packing density [1–3]. However, the metal–ferroelectric–semiconductor (MFS) structure has some se-rious problems because of the formation of a SiO2 amor-phous layer and inter-diffusion at the interface between theferroelectric layer and the Si substrate, leading to deteriora-tion of the device performance. To solve these problems, themetal–ferroelectric–insulator–semiconductor (MFIS) struc-ture has been evolved. The insertion of an insulating bufferlayer plays an important role in preventing reaction andinter-diffusion between the ferroelectric film and the Si sub-strate as well as improving the device performance due tothe suppression of charge injection from the Si substrateto the ferroelectric film. Up to the present time, severalinsulating materials such as CeO2, HfO2, Al2O3, ZrO2,Y2O3, and SrTiO3 (STO), have been extensively studiedas possible insulating layers for the MFIS structures [4–9].Among these materials, STO perovskite oxide is one of thebest candidate materials for the insulating buffer layer be-cause of its excellent barrier property, good thermal stabil-ity, low leakage current, and high dielectric constant. Onthe other hand, a wide class of ferroelectric materials suchas SrBi2Ta2O9, YMnO3, BaTiO3 (BTO), PbZrTiO3, andBi2VO5.5 has been studied concurrently with the insulatingbuffer layers [10–14]. When their properties are compared,the BTO perovskite oxide has many advantages over other

338 G. Panomsuwan et al.

ferroelectric materials, including low crystallization temper-ature, high fatigue endurance, good compatibility with inte-grated circuits, and low environmental load [15]. Therefore,the STO and BTO were selected for use as the insulatingbuffer layer and the ferroelectric layer in the current studyof the MFIS structure.

To obtain wider memory window width, the high voltageapplied to the ferroelectric layer should be maximized to in-crease the polarization effect, while the low voltage appliedto the unwanted SiO2 layer should be minimized to suppresscharge injection. Therefore, a low dielectric constant ferro-electric film and a high dielectric constant insulating bufferlayer are essential for further development of MFIS struc-tures [16–18]. This has led us to pay particular attentionto optimizing dielectric constants for both the ferroelectricBTO layer and the insulating STO buffer layer. Crystal ori-entation has been found to play a key role in the alterationof the material dielectric constants, as reported in many pre-vious studies [19–25]. It has been shown that larger dielec-tric constants can be obtained for the highly c-axis-orientedSTO film compared to that for random-oriented STO film[25, 26]. In case of the ferroelectric BTO, the dielectric con-stant of a ferroelectric BTO single crystal is 4100 alongthe a-axis (non-polar) and 160 along the tetragonal c-axis(polar) at room temperature [27]. Spontaneous polarizationswhich lie in an out-of-plane direction (c-axis) are essen-tial for the memory applications. Therefore, a highly c-axisoriented BTO ferroelectric layer with a low dielectric con-stant and a STO buffer layer with a large dielectric constantwould be expected to enhance out-of-plane polarization anddecrease the voltage applied to the SiO2 amorphous layer,leading to an improvement in memory window width.

In this work, we fabricated the BTO ferroelectric filmson Si(001) substrate using a highly c-axis-oriented and arandom-oriented STO layer as an insulating buffer layer.The structural properties, surface morphology, and electri-cal properties of the two types of out-of-plane orientationwere investigated and discussed in detail.

2 Experimental

Both the highly c-axis-oriented and the random-orientedSTO buffer layers (20 nm) were grown on p-type Si (001)substrates using the electron cyclotron ion beam sputter de-position (ECR-IBSD) technique at substrate temperatures of700 ◦C and 800 ◦C, respectively. The experimental detailsfor the preparation of the STO buffer layers can be found inour previous work [26]. Subsequently, 120 nm-thick BTOferroelectric films were grown on both the highly c-axis-oriented and the random-oriented STO buffered Si(001) sub-strate at 700 ◦C. The sputtering target was a stoichiometric

BTO ceramic pellet (Furuuchi Chemical Corporation, pu-rity 99.9 %). The chamber was initially evacuated to a back-ground pressure of 2 × 10−6 Torr. The microwave outputpower applied to the ion source was 180 W at a frequencyof 2.45 GHz and the ions were then accelerated to the targetwith a high voltage of 1.8 kV. The BTO films were grownunder argon and oxygen pressures of 3 × 10−4 Torr and1.5 × 10−4 Torr, respectively. After growth, the films werefurther annealed at 700 ◦C for 30 min and freely cooled toroom temperature under high oxygen pressure of 750 mTorrin order to re-oxidize the grown films.

The structural properties of the films were examinedusing an X-ray diffractrometer (Rigaku SmartLab) op-erating at 9 kW (45 mA, 200 V) with CuKα radiation(λ = 0.15418 nm). The out-of-plane orientation was in-vestigated by high-resolution X-ray diffraction (HRXRD)technique (2θ -ω and ω scans), whereas the in-plane ori-entation was examined by in-plane pole figure measure-ment. The surface morphology of the BTO films was ob-served using an atomic force microscope (AFM Seiko In-strument Inc., SPA-300HV+SPI-3800N) in contact mode.For fabrication of MFIS structure, Au top electrode area2.5 × 10−3 cm2 was deposited using direct current (dc)sputtering through a metallic mask at room temperaturewithout thermal treatment. The leakage current density–voltage (J -V ) and capacitance–voltage (C-V ) character-istics were measured using a source-measure instrument(Keithley 2636) coupled with a LCR meter (HIOKI 3532-50) at room temperature.

3 Results and discussion

Figures 1(a) and 1(b) show the HRXRD 2θ -ω scans ofthe STO buffer layers grown on the Si(001) substrate at700 ◦C and 800 ◦C, respectively. The HRXRD reflection

Fig. 1 HRXRD 2θ-ω scans of the STO buffer layers grown at:(a) 700 ◦C and (b) 800 ◦C

Enhanced memory window of Au/BaTiO3/SrTiO3/Si(001) MFIS structure with high c-axis orientation 339

Fig. 2 HRXRD 2θ -ω scans of the BTO films grown on: (a) the highlyc-axis-oriented STO buffer layer and (b) the random-oriented STObuffer layer

peaks showed only 00l reflections corresponding to the STOphase and the Si substrate. The full width at half maximum(FWHM) for the rocking curve (ω scan) on the 002STO re-flection of the STO buffer layer grown at 700 ◦C is approxi-mately 2◦, indicating a high degree of c-axis orientation. Onthe other hand, the presence of 011STO and 002STO reflec-tions with low peak intensities was observed for the STObuffer layer grown at 800 ◦C, which indicates random ori-entation. The HRXRD 2θ -ω scans of the BTO films grownon the highly c-axis-oriented and the random-oriented STObuffer layers are shown in Figs. 2(a) and 2(b), respectively.It can be clearly seen that the orientations of the BTO filmsare almost identical to those of the STO buffer layers. Thissuggests that the orientation of the buffer layers has a stronginfluence on the orientation of the BTO films. The BTOfilm grown on the highly c-axis-oriented STO buffer layeris preferentially oriented in the [001] direction and onlya c-axis 002BTO reflection is observed at 2θ = 44.9◦. Incontrast, the BTO film grown on the random-oriented STObuffer layer exhibits random orientation with a mixture ofa-axis and c-axis-oriented grains, as indicated by the split-ting of 200BTO (2θ = 45.3◦) and 002BTO (2θ = 44.4◦) re-flections. The out-of-plane lattice constant determined fromthe 002BTO reflection of the highly c-axis-oriented BTOfilm is 0.4024 nm, which is closer to the c-axis lattice con-stant (c = 0.4038 nm) than to the a-axis lattice constant(a = 0.3992 nm) of a bulk tetragonal BTO single crystal.The FWHM for the rocking curve measured on the 002BTO

reflection of the c-axis-oriented BTO film is approximately2.1◦, which corresponds well to the value measured on the002STO reflection of the STO buffer layer. The HRXRD2θ -ω scans only provide the data on an orientation perpen-dicular to the substrate surface. It is also necessary to investi-gate in-plane orientation of the films using in-plane pole fig-ure measurement. The in-plane pole figure was determined

Fig. 3 In-plane pole figures measured on: (a) {002}BTO and(b) {011}BTO

for the {011}BTO and {002}BTO planes of the highly c-axis-oriented BTO film, as depicted in Figs. 3(a) and 3(b), re-spectively. The central maximum of the {002}BTO pole fig-ure and the concentric ring at α = 45◦ for the {011}STO polefigure are apparent, which correspond to the diffraction from(001)-oriented grains parallel to the substrate. However, in-tense spots in the pole figure measured on both planes arenot visible, except for the intense spots from the Si substratein the {002}BTO pole figure. These patterns are evidence thatthe films grew uni-axially along the c-axis perpendicular tothe substrate surface and are randomly oriented in the in-plane direction. On the other hand, there are no defined pat-terns observed for the {011}BTO and {002}BTO pole figuresof the random-oriented BTO film (not shown here), thus in-dicating that it is a random polycrystalline film.

Figures 4(a) and 4(b) depict the surface morphology ofthe highly c-axis-oriented and the random-oriented BTOfilms, respectively. Both BTO films are well crystallizedand dense. The highly c-axis-oriented film is composed ofsmall rounded grains (∼80 nm) and flat plate-shaped grains(∼500 nm). The flat plate regions correspond to the (001)-oriented grains, which is in good agreement with previousreports [25, 26, 28]. This is attributed to the inhomogene-ity of grain growth nucleation driven by the minimization ofsurface and grain boundary energies [29, 30]. In contrast,the randomly oriented BTO film has only small roundedgrains (∼100 nm) without any flat plate-shaped grains. Av-erage root-mean-square (RMS) roughness of the highly c-axis-oriented and random-oriented BTO films deduced from

340 G. Panomsuwan et al.

Fig. 4 Topography AFM images (2 × 2 µm2 scan area) of: (a) thehighly c-axis-oriented BTO film and (b) the random-oriented BTO film

Fig. 5 J -V characteristics of the Au/STO/p-Si MIS structures and theAu/BTO/STO/p-Si MFIS structures

five scan areas (2 × 2 µm2) were found to be 1.8 ± 0.1 nmand 2.5 ± 0.1 nm, respectively.

To confirm the insulating properties of the STO bufferlayers, the J -V characteristic was measured for the Au/STO/p-Si metal-insulator-semiconductor (MIS) structures atroom temperature, as shown in Fig. 5. Their curves exhibitnon-linear resistance with rectifying behavior that graduallyincreases with increasing applied voltage. The leakage cur-rent densities of the highly c-axis-oriented and the random-oriented STO buffer layers are about 3.2 × 10−5 A/cm2 and4.5 × 10−4 A/cm2 at 5 V, respectively. This demonstratesthat the STO buffer layer with a high c-axis orientation ismore effective in preventing leakage and charge injection.The Au/BTO/STO/p-Si MFIS structures were also fabri-cated to study their electrical properties, including leakagecurrent and memory window behavior. The leakage current

Fig. 6 High-frequency C-V characteristics of the Au/BTO/STO/p-SiMFIS structures with high c-axis orientation (a) and random orienta-tion (b) at different applied voltages

density of the MFIS structure with high c-axis orientationwas found to be less than 10−7 A/cm2 when the applied volt-ages was lower than 5 V, which is smaller in magnitude thanthat of the MFIS structure with random orientation, as canbe seen in Fig. 5. The leakage current behavior can be de-scribed in terms of grain boundaries according to the reportof Guillain et al. [31]. They found that the leakage current al-most vanished in the grain regions and was high at the grainboundaries. Thus, a large number of grain boundaries in therandom-oriented BTO film lead to higher current flow alonggrain boundary paths.

The C-V characteristics were measured by varying ap-plied voltage from ±2 V to ±6 V with a sweeping rateof 0.1 V/s at a high frequency of 1 MHz. Figures 6(a)and 6(b) demonstrate the C-V characteristic curves for theMFIS structures with the high c-axis orientation and ran-dom orientation at different applied voltages (±2, ±4, and±6 V), respectively. All C-V curves clearly show a sym-metrical hysteresis loop in a clockwise direction due to fer-roelectric polarization effect and consist of three main re-gions, including accumulation, depletion, and inversion. Thememory window width is defined by the difference in flat-band voltages between the forward (Vfbf ) and the reverse

Enhanced memory window of Au/BaTiO3/SrTiO3/Si(001) MFIS structure with high c-axis orientation 341

Fig. 7 The memory window width of the Au/BTO/STO/p-Si MFISstructures with high c-axis orientation (a) and random orientation(b) as a function of the applied voltage

C-V curves (Vfbr). To investigate the effect of film orienta-tion, the memory window width of the MFIS structures withthe high c-axis orientation and the random orientation as afunction of applied voltage were determined and shown inFig. 7. The memory window width is strongly dependent onthe magnitude of the applied voltage and initially increaseswith an increase in the applied voltage due to polarizationenhancement. However, it starts to saturate and drop at anapplied voltage higher than 4 V for both the MFIS struc-tures with a high c-axis orientation and a random orienta-tion. The maximum values of the memory window widthare about 1.17 V and 0.76 V for the MFIS structures witha high c-axis orientation and a random orientation, respec-tively. This suggests that the BTO film which crystallizedalong the c-axis has better ferroelectric switching propertiesbecause of single out-of-plane polarization along the c-axis,whereas the random-oriented BTO film has polarization inmany directions, resulting in the suppression of net polariza-tion along the c-axis. A decrease of memory window widthat higher applied voltage occurs because the polarization ef-fect is dominated by other effects, including trapped chargeand charge injection. Under high applied negative voltage,holes in the Si substrate start to be injected and are trappedin the oxide and the ferroelectric layer. This causes a re-verse polarity in the field effect and subsequently reducesthe charge contributed by the ferroelectric polarization of theBTO films. When the applied voltage is increased to greaterthan ±4V , a decrease in the Vfbf (shift to left) and an in-crease of Vfbr (shift to right) can be observed. This eventu-ally results in a decrease in the memory window width. It istherefore suggested that large ferroelectric polarization andsmall charge injection are required for a small reduction inthe memory window width, which can be obtained by min-imizing the voltage applied to the SiO2 layer and maximiz-ing the voltage applied to the ferroelectric layer. The voltage

applied to each layer of the MFIS structures will now be dis-cussed in more detail.

The capacitance in the accumulation region of the C-Vcurves consists of the series capacitance of the BTO(∼120 nm), STO (∼20 nm), and SiO2 (∼2 nm) capacitors,according to the ratio of the dielectric constant and the thick-ness of each layer. Therefore, the dielectric constant of theBTO films can be determined using the following equation:

εf = df

(ε0AC

− di

εi− dno

εno), (1)

where ε, A, and d are the dielectric constant, the areaof the Au top electrode (2.5 × 10−3 cm2), and the layerthickness, respectively. ε0 is the permittivity of free space(8.85 × 10−12 F/m). The subscripts of f , i, and no referto the ferroelectric, insulator and native oxide, respectively.The dielectric constants of the highly c-axis-oriented andrandom-oriented STO films are 102 and 67, respectively.The dielectric constant of SiO2 is assumed to be about 3.9[32]. Therefore, the dielectric constants of the highly c-axis-oriented and the random-oriented BTO films can be deter-mined to be about 93 and 201, respectively. By knowing thedielectric constant and the thickness of each layer, the effec-tive voltage applied to each layer of the MFIS structures canbe extracted using the following equations:

Vf =(

εiεnodf

εf εnodi + εiεnodf + εf εidno

)Vg, (2)

Vi =(

εf εnodi

εf εnodi + εiεnodf + εf εidno

)Vg, (3)

Vno =(

εf εidno

εf εnodi + εiεnodf + εf εidno

)Vg, (4)

where V is the effective applied voltage to a layer and Vg

is the gate voltage applied to the MFIS structure. The effec-tive voltages applied to the highly c-axis-oriented and therandom-oriented BTO films are 0.59Vg and 0.33Vg , respec-tively. As a result, the polarization in the c-axis-orientedBTO film can be saturated more easily at a lower appliedvoltage because of the low dielectric constant and the higherapplied voltage. Moreover, the effective voltage applied tothe SiO2 layer was found to be about 0.28Vg and 0.43Vg

for the MFIS structures with a high c-axis orientation anda random orientation, respectively. Smaller voltages appliedto the SiO2 layer and higher voltages applied to the BTOlayer result in the suppression of charge injection and a smallreduction in the memory window width for the MFIS struc-tures with a high c-axis orientation. This suggests that anenhancement of memory window width can be achieved byusing a highly c-axis-oriented BTO ferroelectric layer (lowdielectric constant) and a highly c-axis-oriented STO bufferlayer (high dielectric constant), which is in good agreement

342 G. Panomsuwan et al.

with our prediction. These obtained results are clear evi-dence that the crystal orientation has a strong influence onthe electrical properties and the memory window behaviorof the MFIS structures.

4 Conclusions

We fabricated ferroelectric BTO thin films with a highly c-axis-oriented and a random-oriented STO buffer layer usingthe ECR-IBSD technique on p-Si(001). The orientations ofBTO films correspond to those of the STO buffer layers. TheAFM results reveal a smaller surface roughness and flattersurface for the highly c-axis-oriented BTO film. The leak-age current density of the MFIS structure with high c-axisorientation is 1.05 × 10−7 A/cm2 at a positive applied volt-age of 4 V, which is about one magnitude lower than that ofthe MFIS structure with random orientation. The maximumvalues of memory window widths are 1.17 V and 0.76 V forMFIS structures with a high c-axis orientation and a randomorientation, respectively. The electrical properties and mem-ory window width of the MFIS structure with a high c-axisorientation can be significantly by insertion of a STO bufferlayer with a high dielectric constant and high polarization ofthe BTO along the c-axis with a low dielectric constant. Thisleads to the enhancement of polarization and the suppres-sion of charge injection from the Si substrate. The resultsobtained from this study suggest that the Au/BTO/STO/p-SiMFIS structure with a high c-axis orientation has potentialfor non-volatile memory applications as well as for use indeveloping ferroelectric-based devices under the role of ori-entation.

Acknowledgements This work was supported by the Micro-NanoGlobal COE (Center of Excellence), Nagoya University. The authorswould like to thank Prof. Yoshimura at Graduate School of Engineer-ing, Toyota Technological Institute (TTI) for providing the equipmentfor electrical measurements.

References

1. S.Y. Wu, IEEE Trans. Electron Devices 21, 499 (1974)2. J.F. Scott, C.A. Paz de Araujo, Science 246, 1400 (1989)3. J.P. Han, T.P. Ma, Appl. Phys. Lett. 72, 1185 (1998)

4. Y.T. Kim, D.S. Shin, Appl. Phys. Lett. 71, 3507 (1997)5. A. Roy, A. Dhar, D. Bhattacharya, S.K. Ray, J. Phys. D, Appl.

Phys. 41, 095408 (2008)6. S.Y. Chen, C.L. Sun, S.B. Chen, A. Chin, Appl. Phys. Lett. 80,

3168 (2002)7. H.Y. Chen, J.M. Wu, H.E. Huang, H.Y. Bor, Appl. Phys. Lett. 90,

112907 (2007)8. T. Yoshimura, N. Fujimura, D. Ito, T. Ito, J. Appl. Phys. 87, 3444

(2000)9. X. Lu, H. Ishiwara, X. Gu, D. Lubyshev, J. Fastenau, R. Pelzel,

J. Appl. Phys. 105, 024111 (2009)10. M.H. Tang, Z.H. Sun, Y.C. Zhou, Y. Sugiyama, H. Isihwara, Appl.

Phys. Lett. 94, 212907 (2009)11. H.N. Lee, Y.T. Kim, Y.K. Park, Appl. Phys. Lett. 74(25), 3887

(1999)12. G. Niu, S. Yin, G. Saint-Girons, B. Gautier, P. Lecoeur, V. Pillard,

G. Hollinger, B. Vilquin, Microelectron. Eng. 88, 1232 (2011)13. J. Yu, Z.J. Hong, W. Zhou, G. Cao, J. Xie, X. Li, S. Li, Z. Li, Appl.

Phys. Lett. 70, 490 (1997)14. N.K. Kumari, J. Parui, K.B.R. Varma, S.B. Krupanidhi, Solid State

Commun. 137, 566 (2006)15. K.H. Chen, Y.C. Chen, Z.S. Chen, C.Y. Yang, T.C. Chang, Appl.

Phys. A 89, 533 (2007)16. E. Tokumitsu, R. Nakamura, H. Ishiwara, IEEE Electron Device

Lett. 18, 160 (1997)17. H.N. Lee, Y.T. Kim, S.H. Choh, Appl. Phys. Lett. 76, 1066 (2000)18. T. Yoshimura, N. Fujimura, D. Ito, T. Ito, Appl. Phys. Lett. 87,

3444 (2000)19. K.H. Yoon, B.D. Lee, J.H. Park, Appl. Phys. Lett. 79, 1018 (2001)20. C.L. Li, Z.H. Chen, Y.L. Zhou, D.F. Cui, J. Phys. Condens. Matter

13, 5261 (2001)21. S. Schmidt, J. Lu, S.P. Keane, L.D. Bregante, D.O. Klenov,

S. Stemmer, J. Am. Ceram. Soc. 88, 789 (2005)22. I.D. Kim, Y. Avrahami, H.L. Tuller, Y.B. Park, M.J. Dicken, H.A.

Atwater, Appl. Phys. Lett. 86, 192907 (2005)23. H.S. Kim, T.S. Hyun, H.G. Kim, I.D. Kim, T.S. Yun, J.C. Lee,

Appl. Phys. Lett. 89, 052902 (2006)24. J. Hiltunen, D. Seneviratne, H.L. Tuller, J. Lappalainen, V. Lantto,

J. Electroceram. 22, 395 (2009)25. X. Wang, X. Lu, H. Bo, Y. Liu, Y. Shen, X. Wu, W. Cai, Y. Kan,

C. Zhang, F. Huang, J. Zhu, Solid State Commun. 150, 1637(2010)

26. G. Panomsuwan, S.P. Cho, N. Saito, O. Takai, Cryst. Res. Technol.47, 187 (2012)

27. W.J. Mertz, Phys. Rev. 76, 1221 (1949)28. C.J. Lu, Y. Qaio, Y.J. Qi, X.Q. Chen, Appl. Phys. Lett. 87, 222901

(2005)29. J. Yan, G.D. Hu, X.M. Chen, W.B. Wu, C.H. Yang, J. Appl. Phys.

104, 076103 (2008)30. C.V. Thompson, Annu. Rev. Mater. Sci. 30, 159 (2000)31. J. Guillan, G. Tartavel, E. Defa, L. Ulmer, L. Galera, B. André,

F. Baume, Integr. Ferroelectr. 67, 93 (2004)32. D.S. Shin, H.N. Lee, Y.T. Kim, I.H. Choi, B.H. Kim, Jpn. J. Appl.

Phys., Part 1 37, 4373 (1998)