A Hydrothermally Deposited Epitaxial PbTiO 3 Thin Film on SrRuO 3 Bottom Electrode for the Ferroelectric Ultra-High Density Storage Medium

Transcription

1 Integrated Ferroelectrics, 64: , 2004 Copyright C Taylor & Francis Inc. ISSN: print/ online DOI: / A Hydrothermally Deposited Epitaxial PbTiO 3 Thin Film on SrRuO 3 Bottom Electrode for the Ferroelectric Ultra-High Density Storage Medium TAKESHI MORITA and YASUO CHO Research Institute of Electrical Communication, Tohoku University, Katahira, Aoba-ku, Sendai , Japan (Received April 30, 2004; In final form May 21, 2004) A lead titanate epitaxial thin film is one of the important ferroelectric films and lattice matching with strontium titanate is preferable. However, it was thought to be difficult to obtain high insulation property and very few superior hysteresis curves were measured with lead titanate. The hydrothermal method can avoid a lead vacancy and residual strain at the boundary of the substrate due to low reaction temperature of 150 C. Using strontium ruthenium oxide bottom electrode on strontium titanate, the lead titanate thin film was successfully deposited by the hydrothermal method. The single crystal-like DE hysteresis curve showed a remanent polarization of 96.5 µc/cm 2. The domain direction was controlled by an applied electric field using a metal coated AFM cantilever probe and the domain pattern was observed by scanning nonlinear dielectric microscopy (SNDM). Minimum dot radius of inverted domain was 12 nm corresponding to data storage density of 1 Tbit/inch 2. Keywords: hydrothermal method; ferroelectricity; lead titanate; epitaxial thin film; scanning nonlinear dielectric microscopy I. INTRODUCTION A domain wall thickness of the ferroelectric material is very thin compared to that of the ferromagnetic materials [1], so that a ferroelectric ultra-high density data storage system attracts researcher s attentions [2 5] To write and pick up the data bit a scanning nonlinear dielectric microscopy system is superior to piezoelectric force microscopy due to its superior resolution [2 5]. Up to now, 1.5Tbit/inch 2 data storage density has been demonstrated using congruent lithium tantalate thin films [2]. To obtain lithium tantalate tmorita@ieee.org 247

2 248 T. MORITA AND Y. CHO thin films, mechanical polishing and dry etching were carried out. This preparation scheme was complicated so that it is needed a lot of fabrication costs to obtain a large media with uniform thickness. Therefore a suitable epitaxial thin film is strongly required to realize the ferroelectric ultra-high density data storage system. For data storage medium, the ferroelectric film must not contain any defect such as a-domain, domain boundary and grain boundary even in nano scale. In this study, we focused on a hydrothermally deposited epitaxial lead titanate. Lead titanate is an important perovskite-type ferroelectric material, although high quality crystal was previously thought to be quite difficult to obtain [6 8]. The remanent polarization for lead titanate was predicted to be 81 µc/cm 2 [9] and thus far, very little experimental data has been provided for this fundamental value, with the exception of that reported by Tabata [10]. The hydrothermal method [11 18], in contrast to the other methods, enables the deposition of a perovskite thin film in aqueous solution at very low temperatures, typically C. The high purity film of the film achieved is attributed to the absence of a phase change during deposition at this low temperature of crystallization. In methods using a high temperature process, a lead defect must be taken into consideration, because the vapor pressure of lead oxide is low therefore an excess of lead must compensate. Such lead evaporation does not occur in the hydrothermal method. In addition to the low reaction temperature, the hydrothermal method is versatile in that it is a simple process procedure, polarization is aligned automatically and it achieves a three-dimensional deposition. Barium titanate [11 13], lead zirconate titanate [14 7], and lead titanate [18] were previously reported to be successfully deposited by this method. Morita et al. applied PZT polycrystalline PZT thin films for cylindrical micro ultrasonic motors [19 21]. The substrate was cylindrical shaped titanium and the motivation to use the hydrothermal method was the three-dimensional deposition. Different from the polycrystalline films, the epitaxial thin films can be obtained using a single crystal substrate, such as a SrTiO 3 substrate. However, in previous studies of PZT epitaxial films, only deposition methods and crystal orientations have been investigated; few reports have been made on the ferroelectric properties of deposited films. Morita et al. recently modified Chien s process [12] and measured the DE hysteresis properties of a hydrothermally deposited PZT film using a SrRuO 3 bottom electrode [14] on (100) SrTiO 3 single crystal. It was the first measurement results about the DE hysteresis curve of the hydrothermally deposited epitaxial ferroelectric thin films. From the DE hysteresis measurement of PZT thin film deposited by the hydrothermal method on

3 A HYDROTHERMALLY DEPOSITED EPITAXIAL PbTiO 3 THIN FILM 249 SrRuO 3 /SrTiO 3 substrate, an imprint was too large to reverse and maintain the intended poling direction as shown later. Hence; this PZT thin film was not suitable for a FeRAM application and ultra-high density storage medium to construct a nano dot pattern with a conductive atomic force microscopy (AFM) probe [2 5]. The composition ratio of Zr/Ti in the deposited PZT film was controlled in an effort to achieve the morphotropic phase boundary (MPB). The lattice constant mismatch between the PZT and substrate was relatively large. This mismatch was thought to be responsible for the imprint property of the PZT film. The lattice match with a SrRuO 3 /SrTiO 3 substrate indicated that a lead titanate may be promising compared to the MPB PZT. The bulk a-lattice constant of lead titanate is nm, which is very close to that of SrTiO 3 ( nm). The lattice constant of a thin SrRuO 3 bottom electrode is considered to be almost equal to that of SrTiO 3. Therefore, in this study we assumed lead titanate had little distortion at the boundary surface with substrate. Thus, superior ferroelectric performance is expected. II. DEPOSITION PROCESS To improve the ferroelectric crystal quality, the uniformity and flatness of the substrate is indispensable. In this study, a 0.5 mm thick SrTiO 3 single crystal substrate (100) was treated with buffered hydrofluoric acid (BHF) and annealing was performed at 950 C [22] in order to obtain an atomic step surface. Atomic force microscope observation revealed an atomic step of 0.4 nm at the SrTiO 3 substrate surface. A 50 nm thick epitaxial SrRuO 3 film was RF magnetron sputtered on this substrate as a bottom electrode [14, 23]. Before applying SrRuO 3 /SrTiO 3,weused Nb doped SrTiO 3 substrate to provide conductivity as a bottom electrode for PZT deposition, although measured DE hysteresis curve had been strongly distorted. This phenomenon must be related to the conductive mechanism of Nb doped SrTiO 3, and some researches about them has been already reported [24, 25]. When the metal coated AFM cantilever is used to control the ferroelectric domain, applied electric field is concentrated just below the probe tip, and the incomprehensible conductive mechanism between ferroelectric film and the bottom electrode can be neglected. However, to study the fundamental properties of deposited ferroelectric films, such these strange conductivity with Nb doped SrTiO 3 bottom electrode must be considered, Therefore, the Nb doped SrTiO 3 substrate was rejected and the SrRuO 3 bottom electrode on SrTiO 3 was utilized for lead titanate deposition in our study.

4 250 T. MORITA AND Y. CHO TABLE I Reaction conditions for an epitaxial PTO thin film. Pb(NO 3 ) 3 TiO 2 KOH Temp. Time 2.26 g g 10N 15 ml 150 deg 3 h or 24h The hydrothermal reaction conditions for epitaxial lead titanate thin film were modified from those for PZT deposition, as listed in Table I. An asreceived SrTiO 3 substrate was used as the substrate for the PZT deposition in the previous study [14]. This substrate had no atomic step surface. After adding the reagents (Table I) to the pressure vessel (Parr Model 4744), the substrate was placed in the reaction vessel using a Teflon holder in the same manner as that for PZT deposition. The substrate was maintained horizontally, 8.6 mm above the bottom of the pressure vessel, facing the polished surface bottom. After either 24 hours or 3 hours of deposition, the substrate was removed, cleaned ultrasonically with distilled water, and dried for 1 hour at 200 C under atmospheric conditions. It is important to note that the highest temperature achieved in this study was that used for drying, 200 C, which is far below the Curie temperature of lead titanate (490 C). III. CRYSTAL QUALITY AND FERROELECTRIC PROPERTIES OF PbTiO 3 THIN FILM The deposited lead titanate film had no cracks and appeared smooth under both the optical microscopic and SEM observation. The thickness of the lead titanate film was 100 nm and 50 nm for 24 and 3 hours deposition, respectively. An XRD measurement revealed that the lead titanate film had only (00l) peaks, as shown in Fig. 1, indicating the film was oriented with the c-axis. Figure 1 shows the XRD measurement of the 50 nm film. The thicker film (100 nm, not shown), was also c-axis oriented. The c-axis lattice parameter of deposited film was nm for 24 hours of deposition and nm for 3 hours of deposition. These values had less than 0.2% difference from the single crystal data ( nm) [2]. This agreement indicated that the strain between lead titanate film and substrate was small compared to that of the PZT thin film on SrRuO 3 /SrTiO 3.In addition, scanning nonlinear dielectric microscopy (SNDM) measurement

5 A HYDROTHERMALLY DEPOSITED EPITAXIAL PbTiO 3 THIN FILM 251 FIGURE 1 XRD spectra for PbTiO 3 thin film on a (100) SrRuO 3 /SrTiO 3 substrate. did not show an a-domain film property in the lead titanate film, even at nano-meter scale resolution. The full width at half-maximum value of the (002) rocking curve was 0.85 degrees. A pole figure measurement showed a 45 degree in-plane shift between (111) SrTiO 3 and FIGURE 2 Phi san to confirm in-plane epitaxial deposition.

6 252 T. MORITA AND Y. CHO FIGURE 3 Hysteresis curves for PbTiO 3 thin films deposited on SrRuO 3 /SrTiO 3 (100). Reaction time; 24 hours, thickness; 100 nm. (104) PbTiO 3, which indicates an in-plane epitaxial property as shown in Fig. 2. A platinum top electrode (0.083 mm 2 )was evaporated on the lead titanate by a shadow masked metal process, and the DE hysteresis property was subsequently measured. The thickness of the lead titanate film was 100 nm. As shown in Fig. 3, a single crystal like DE hysteresis curve with wellsaturated polarization was obtained. The remanent polarization value was 96.5 µc/cm 2 and the coercive electrical field was 290 kv/cm. A superior lattice constant matching to the substrate, the low temperature deposition and the direct deposition with ionic reaction were thought to contribute to the high purity and low conductivity of the crystal. The initial imprint direction of the film was from the bottom electrode to the surface, which was in agreement with that observed in SNDM measurements. It is interesting that the imprint electrical field was opposite to that of PZT thin film as shown in Fig. 4 [14]. This is thought to be related to the lattice mismatching to the substrate, because in the PZT film the a-lattice constant was larger than the substrate, whereas that of lead titanate was smaller than the substrate.

7 A HYDROTHERMALLY DEPOSITED EPITAXIAL PbTiO 3 THIN FILM 253 FIGURE 4 Hysteresis curves for PZT thin films deposited on SrRuO 3 /SrTiO 3 (100) [14]. IV. DOMAIN PATTERNING AND NANO DOMAIN PATTERNS Using the epitaxial lead titanate thin film deposited for 3 hours (50 nm thickness), the inverse domain was patterned using a platinum-iridium coated atomic force probe. The curvature radius of the probe was 25 nm and the electrical potential of the probe was grounded. The domain pattern was drawn by applying a plus or minus DC voltage with amplitude of 5 volts to the SrRuO 3 bottom electrode, in three successive scanning processes. The scanned areas were 800 nm, 400 nm and 200 nm square, respectively. After three scanning procedures, the domain pattern was observed by SNDM. As shown in Fig. 5, the expected domain pattern was obtained, however, an area about 50 nm larger than that expected was reversed due to the spread of the electrical field below the scanned probe. As seen in the figure, the absence of an a-domain, a grain boundary, and a domain boundary was confirmed. A ferroelectric signal from SNDM was detected over the entire area and the entire area responded to the DC applied electrical field. This outstanding performance indicated that this lead titanate was the appropriate medium for ultra-high density system. Pulse applied voltage was

8 254 T. MORITA AND Y. CHO FIGURE 5 Domain patterns with DC plus or minus 5 V observed using scanning nonlinear dielectric microscopy (SNDM). subsequently applied to pattern domain dots with various pulse parameters, as shown in Fig. 6. The minimum dot radius of 12 nm was achieved under the trial parameters with 5 V and a 40 µsec pulse. This size corresponds to a data size of 1 Tbit/inch 2, and is a match for the data storage density using a congruent lithium tantalate (CLT) medium [2 5]. The high crystal quality of a CLT single crystal was suitable for an ultra-high density medium, although the process to reduce the thickness to less than 100 nm was complicated compared to the lead titanate film in this study. Furthermore, the large permittivity of lead titanate is an additional advantage from the view of higher resolution of SNDM measurements. However, the poling reversal response of lead titanate to the short pulse was relatively inferior. At a voltage of 5 V,the 40 µsec pulse was the limitation for the reversal of poling. To date, CLT has been confirmed to respond to a 4 nsec pulse [5]. This property determines the writing speed for the data storage system; thus,

9 A HYDROTHERMALLY DEPOSITED EPITAXIAL PbTiO 3 THIN FILM 255 FIGURE 6 Nano dot patterns with various pulse parameters observed using SNDM and cross section of SNDM signal. it is required in order to investigate which material parameter is related to the fast response against the electrical field. V. CONCLUSIONS In conclusion, an epitaxial lead titanate thin film was successfully deposited on a SrRuO 3 /SrTiO 3 substrate. By adopting a SrRuO 3 bottom electrode, the ferroelectric properties of lead titanate film could be estimated. This thin film lacks an a-domain and a grain boundary. A DE hysteresis curve confirmed that a high quality single crystal film was obtained. The remanent polarization was 96.5 µc/cm 2 and coercive electrical field was 290 kv/cm; this data confirms the parameters of the lead titanate single crystal. This suitable process for achieving high quality crystalline materials was attributed to its low reaction temperature. The inversed poling pattern was obtained using

10 256 T. MORITA AND Y. CHO SNDM. From this domain pattern, uniform and no defect ferroelectric performance was confirmed, even at the nanometer level. The obtained nano dot pattern verified the suitability of lead titanate film as an ultra-high density data storage medium. At present, the pulse parameter is being optimized in order to obtain a smaller dot radius. ACKNOWLEDGEMENTS This work was financially supported by Foundation for C&C promotion. REFERENCES [1] F. Jona and G. Shirane, Ferroelectric Crystals, APergamon Press Book (London), Vol. 46, [2] Y. Cho, K. Fujimoto, Y. Hiranaga, Y. Wagatsuma, A. Onoe, K. Terabe, and K. Kitamura, Tbit/inch 2 ferroelectric data storage based on scanning nonlinear dielectric microscopy, Appl. Phys. Lett. 81, (2002). [3] Y. Cho, S. Kazuta, and K. Matsuura Scanning, nonlinear dielectric microscopy with nanometer resolution, Appl. Phys. Lett. 75, (1999). [4] Y. Hiranaga, K. Fujimoto, Y. Wagatsuma, Y. Cho, A. Onoe, K. Terabe, and K. Kitamura, Basic study on high-density ferroelectric data storage using scanning nonlinear dielectric microscopy, Proceedings of Mat. Res. Soc. Symp U5.5, 1 6 (2002). [5] K. Fujimoto and Y. Cho, High-speed switching of nanoscale ferroelectric domains in congruent single-crystal LiTaO 3, Appl. Phys. Lett. 83, (2003). [6] B. Jaffe, W. R. Cook, and H. Jaffe, Piezoelectric Ceramics, Academic Press 115 (1971). [7] F. Jona and G. Shirane, Ferroelectric crystals, A Pergamon Press Book (London) 236 (1962). [8] K. Kakuta, T. Tsurumi, and O. Fukunaga, Dielectric property of flux-grown PbTiO 3 single crystal, Jpn. J. Appl. Phys. 34, (1995) [9] Y. N. Venevtsev, G. S. Zhdanov, S. P. Solov ev, and VV. Ivanova, The internal fields in ferroelectric crystals of PbTiO 3, Krystallografiya 4, (1959). [10] H. Tabata, O. Murata, T. Kawai, S. Kawai, and M. Okuyama, c-axis preferred orientation of laser ablated epitaxial PbTiO 3 films and their electrical properties, Appl. Phys. Lett. 64, (1994). [11] K. Kajiyoshi, N. Ishizawa, and M. Yoshimura, Heteroepitaxial growth of BaTiO 3 thin films on SrTiO 3 substrates under hydrothermal conditions, Jpn. J. Appl. Phys. 30, L (1991). [12] A. T. Chien, X. Xu, J. H. Kim, J. Sachleben, J. S. Speck, and F. F. Lange, Electrical characterization of BaTiO 3 heteroepitaxial thin films by hydrothermal synthesis, J. Mater. Res. 14, (1999). [13] W. Zhu, S. A. Akbar, R. Asiaie, and P. K. Dutta, Synthesis, microstructure and electrical properties of hydrothermally prepared ferroelectric BaTiO 3 thin films, Journal of Electroceramics 2, (1998)

11 A HYDROTHERMALLY DEPOSITED EPITAXIAL PbTiO 3 THIN FILM 257 [14] T. Morita, Y. Wagatsuma, Y. Cho, H. Morioka, H. Funakubo, and N. Setter Appl. Phys. Lett. 84 (2004). [15] A. T. Chien and F. F. Lange, Hydrothermal synthesis of heteroepitaxial Pb(Zr x Ti 1 x )O 3 thin films at C, J. Mater. Res. 12, (1997). [16] K. Shimomura, T. Tsurumi, Y. Ohba, and M. Daimon, Preparation of lead zirconate titanate thin film by hydrothermal method, Jpn. J. Appl. Phys. 30, (1991). [17] T. Morita, T. Kanda, Y. Yamagata, M. K. Kurosawa, and T. Higuchi, Single process to deposit lead zirconate titanate (PZT) thin film by a hydrothermal method, Jpn. J. Appl. Phys. 36, (1997). [18] A. T. Chien, J. Sachleben, J. K. Kim, J. S. Speck, and F. F. Lange, Synthesis and characterization of PbTiO 3 powders and heteroepitaxial thin films by hydrothermal synthesis, J. Mater. Res. 14, (1999). [19] T. Morita, M. K. Kurosawa, and T. Higuchi, An ultrasonic micromotor using a bending cylindrical transducer based on PZT thin film, Sensors and Actuators A50, (1995). [20] T. Morita, M. K. Kurosawa, and T. Higuchi, A cylindrical micro ultrasonic motor using PZT thin film deposited by single process hydrothermal method (φ 2.4 mm, L = 10 mm stator transducer), IEEE Trans. on Ultrasonic, Ferroelectrics and Frequency Control 45, (1998). [21] T. Morita, M. K. Kurosawa, and T. Higuchi, A cylindrical shaped micro ultrasonic motor utilizing PZT thin film (1.4 mm in diameter and 5.0 mm long stator transducer), Sensors and Actuators 83, (2000). [22] G. Koster, B. Kropman, Rijnders GJHM, Blank DHA, and H. Rogalla, Quasi-ideal strontium titanate crystal surfaces through formation of strontium hydroxide, Appl. Phys. Lett. 73, (1998). [23] K. Takahashi, T. Oikawa, K. Saito, S. Kaneko, H. Fujisawa, M. Shimizu, and H. Funakubo, Effect of strain in epitaxially grown SrRuO 3 thin films on crystal structure and electric properties, Jpn. J. Appl. Phys. 41, (2002). [24] Y. Watanabe, Reproducible memory effect in the leakage current of epitaxial ferroelectric/conductive perovskite heterostructures, Appl. Phys. Lett. 66, (1995). [25] Y. Watanabe, Electrical transport through Pb(Zr,Ti)O 3 p-n and p-p heterostructures modulated by bound charges at a ferroelectric surface, Ferroelectric p-n diode. Physical Review B 59, (1999).