Driving Characteristics of a Surface Acoustic Wave Motor using a Flat-Plane Slider

Advanced Robotics 24 (2010) 1407 1421 brill.nl/ar Full paper Driving Characteristics of a Surface Acoustic Wave Motor using a Flat-Plane Slider Koki Sakano, Minoru Kuribayashi Kurosawa and Takashi Shigematsu Interdisciplinary Graduate School of Engineering, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Japan Received 11 September 2009; accepted 11 November 2009 Abstract We discuss the use of a flat-plane slider for a surface acoustic wave (SAW) motor. For stable driving, the SAW motor requires a silicon slider with many microprojections on its surface. Previously, sliders with no projections on the flat-plane surface could not provide stable drive because of contact electrification occurring due to the low electrical conductivity of the lithium niobate stator. In this paper, we present a flatplane slider that can provide stable drive by using for the stator lithium niobate with conductivity improved by chemical reduction. The flat-plane slider can drive under a higher preload than a slider with projections because the contact area of the flat-plane slider is larger than that of the slider with projections. Thus, the flat-plane slider is expected to obtain a higher output force than the slider with projections. We report herein the results of measurements of the flat-plane slider driving characteristics such as the no-load speed and output force. The SAW motor with our proposed flat-plane slider achieves a high output force of 18 N and a no-load speed of 0.9 m/s. Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2010 Keywords Linear motor, ultrasonic motor, surface acoustic wave, friction drive, microactuator 1. Introduction A surface acoustic wave (SAW) motor is a traveling wave-type ultrasonic linear motor that uses Rayleigh waves, which is a type of SAW, for friction drive [1 9]. As the driving frequency of the SAW motor, approximately 10 MHz, is much higher than other ultrasonic motors, the Rayleigh wave amplitude attains the nanometer level. Therefore, the SAW motor achieves high speed, high output force and high resonance specificity. In previous publications, 10 N output force, 1.5 m/s no-load speed and 0.5 nm stepping motions were reported [10, 11]. * To whom correspondence should be addressed. E-mail: sakano.k.aa@m.titech.ac.jp Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2010 DOI:10.1163/016918610X505549

1408 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 To date, the SAW motor, and the slider in particular, has been developed to realize stable drive and higher output force. In the initial stage of research into SAW motors, three ruby balls were used as a slider to obtain high contact pressure at a contact area between the slider and the stator [1]. Subsequently, multiple steel balls were used to obtain higher output force [2]. Furthermore, using a silicon micromachining process, silicon sliders with many minute projections were incorporated into SAW motors [3]. In a previous study, the projection parameters, such as the projection diameter or the distance between projections, were examined in the expectation of achieving higher speed or higher output force [6]. It was found that smaller projections and higher projection density lead to better driving characteristics. When we push the projection density to the limit, we ultimately arrive at the flat-plane geometry. We, therefore, expect the flat-plane slider to improve the driving characteristics of SAW motors. In addition, the flat-plane slider also has the advantage that silicon micromachining processes are not required. However, slider projections achieve high contact pressure for the friction drive and good contact between the stator and the slider. With the flat-plane slider, it is difficult to achieve stable drive. In this paper, we report the use of lithium niobate as stator whose conductivity is improved by chemical reduction, thus improving the driving stability. We perform driving experiments using the flat-plane slider and we discuss its driving characteristics. 2. Principle The SAW motor consists of a thin-plate transducer and a thin-friction material, which are the stator and the slider, respectively, as illustrated in Fig. 1. The stator, which is made of lithium niobate, has interdigital transducers (IDTs) at both ends to excite traveling waves in both directions. When RF electrical power is input into the IDT, a propagating Rayleigh wave is generated on the stator surface by the piezoelectric effect. The propagating Rayleigh wave drives the slider in one direction via the frictional force [5]. The opposite linear motion of the slider is achieved by changing the active IDT. Details of the friction drive are shown in Fig. 2. To clearly communicate the actuator principle, the waveform in Fig. 2 is enlarged in the lengthwise direction. For Figure 1. Schematic view of the SAW motor.

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1409 Figure 2. Schematic view of frictional force transmission at the contact area. a 9.61 MHz driving frequency, the actual wavelength and the vibration displacement are about 400 µm and 20 nm, respectively, suggesting that the stator surface is almost flat. As shown in Fig. 2, the particles on the stator surface move in elliptical paths. When the slider (with preload) is placed onto the stator, it is driven in the direction opposite to the direction of propagation of the traveling wave by the frictional force generated at the wave slider boundary. Thus, the maximum speed and output force for the SAW motor is determined by the vibration velocity and the friction force at the wave crest. The slider is made of silicon. A large number of microprojections are fabricated on the contact surface using the dry-etching process, as shown in Fig. 3. These projections create high contact pressure for the friction drive, so that a good contact between the stator and the slider may be established. 3. Experimental Setup The stator was a 1-mm thick, 60 15 mm 2 rectangular plate of chemically reduced, 128 y-rotated x-propagation lithium niobate (see Fig. 4). IDTs were fabricated at both ends of the stator using vacuum vapor deposition, and chromium and aluminum were the electrode materials. The IDT consisted of 20 strip-electrode pairs, and was 400 µm in pitch with a 100 µm electrode strip width, a 100 µm electrode strip space and a 9 mm aperture. The dimensions of the electrode imposed a 9.61 MHz resonance frequency on the IDT. The flat-plane slider was a square silicon substrate with dimensions of 5 5 0.5 mm 3. In this paper, we examine both the flat-plane slider and the silicon slider with projections, shown in Fig. 3, to compare the two. The slider with projections used for the driving experiments had the same dimensions as those of the flat-plane slider. The projections were fabricated in a 4 4mm 2 central square region of the surface using a dry-etching process. The cylindrical projections were 10 µm in diameter and 2 µm high, and were arranged at a 20 µm pitch.

1410 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 Figure 3. Photograph of a silicon slider with projections fabricated on the stator surface using a dry-etching process. Figure 4. SAW motor stator fabricated from chemically reduced lithium niobate. A photograph of an experimental SAW motor is shown in Fig. 5. The motor consists of a fixed part and a movable part, and the fixed part consists of the stator and a linear guide rail. The stator is fixed in a steel jig and the linear guide rail is fixed to a steel block. The movable part consists of the slider, a slider block and a linear guide block. The total mass of the movable parts was 2.0 g. To obtain good contact between the stator and the slider, we used the slider block shown in Fig. 6, which was glued with epoxy resin onto the slider. The contact area between the slider and the slider block was 4.6 4.6 mm 2, whereas the slider is 5 5mm 2, thus preventing stress concentration at the edge of the flat-plane slider. The upper side of the slider block was shaped into a hemisphere, as shown in Fig. 6. The hemispherical part of the slider block mates with a washer, which is glued with

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1411 Figure 5. Photograph of an experimental SAW motor with a flat-plane slider. Figure 6. Photograph of the slider block glued onto the slider. epoxy resin onto the linear guide block. The contact between the hemispherical part of the slider block and the washer allows parallel contact between the slider surface and the stator surface. Figure 7 shows a photograph of the experimental setup. The preload of the slider was provided by a coiled spring that was connected to a micrometer. Therefore, the slider preload could be easily adjusted using the micrometer. In this experiment, we adjusted the preload from 10 to 120 N.

1412 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 Figure 7. Photograph of the experimental setup. 4. Stator for Flat-Plane Slider 4.1. Chemically Reduced Lithium Niobate To date, no reports exist of stable driving by a flat-plane slider. By analyzing the friction drive mechanism [12 16], it was found that the root cause of unstable driving is electrical charge that accumulates during actuation at the contact area between the slider and the stator. In the present study, to improve the driving stability we use for the stator lithium niobate, whose conductivity is improved by chemical reduction. Typical properties of chemically reduced lithium niobate and unreduced lithium niobate are given in Table 1 [17]. The electrical conductivity of chemically reduced lithium niobate is 4.17 10 11 1 cm 1, which is approximately 10 4 times higher than that of normal lithium niobate. Other properties of chemically reduced lithium niobate and unreduced lithium niobate, such as the SAW velocity and the coupling coefficient, are the same; therefore, a wave generated on the chemically

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1413 Table 1. Typical properties of chemically reduced and unreduced lithium niobate Unreduced Reduced Conductivity ( 1 cm 1 ) 2.63 10 15 4.17 10 11 SAW velocity (m/s) 3980 3980 Coupling coefficient k 2 5.5 5.5 reduced lithium niobate stator will have the same properties as a wave generated on the unreduced lithium niobate stator. Thus, by using chemically reduced lithium niobate for the stator, the driving instability of the flat-plane slider is improved. 4.2. Particle Vibration Velocity The performance of the SAW motor, and in particular the no-load speed, depends on the particle movement at the stator surface. We measured the particle vibration velocity for the chemically reduced lithium niobate stator. The horizontal vibration velocity v H is given by: v H = 2πfA H, (1) where f and A H denote the driving frequency and the horizontal vibration amplitude. The horizontal vibration amplitude of the particle at the stator surface is then decided by the vertical vibration amplitude from the particle trajectories. Thus, the horizontal vibration velocity can be estimated from the vertical vibration amplitude. Using a laser Doppler vibrometer, we measured the vertical vibration amplitude of the Rayleigh wave as a function of the driving voltage at the center of the stator, as illustrated in Fig. 8, and the results are shown in Fig. 9. The vibration amplitude is 25 nm at a driving voltage of 125 V peak and the horizontal vibration velocity is estimated to be approximately 1.3 m/s at a driving voltage of 125 V peak. 5. Flat-Plane Slider 5.1. Transient Responses To examine the SAW motor speed response using the flat-plane slider, we measured the transient response of the motor using a laser Doppler vibrometer. The transient responses of the flat-plane slider are shown in Fig. 10. The preload was 20 N and the driving voltage was varied from 35 to 125 V peak. The speed depends on the driving voltage because the horizontal vibration velocity at the stator surface is proportional to the driving voltage, as shown in Fig. 9. The maximum speed of the flat-plane slider is 0.81 m/s at 125 V peak driving voltage. The minimum driving voltage required for actuation is 35 V peak with a 20 N preload. For a comparison, we also measured the transient response of the slider with projections (explained in the experimental section), and show the results in

1414 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 Figure 8. Method to measure vibration amplitude on the stator surface. Figure 9. Vertical vibration amplitude and horizontal vibration velocity of the chemically reduced lithium niobate stator. Fig. 11. The maximum speed of the slider with projections is 0.95 m/s and the minimum driving voltage is 30 V peak. The difference in speed between the flatplane slider and the slider with projections is less than 20%. From Figs 10 and 11, we find that the flat-plane slider can attain a speed response similar to that of the slider with projections. 5.2. Driving Characteristics as a Function of Driving Voltage From the transient responses of the motor as shown in Figs 10 and 11, we examine the mechanical outputs such as the no-load speed and the output force. The no-

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1415 Figure 10. Transient response of the flat-plane slider with a 20 N preload. Figure 11. Transient response of the slider with projections with a 20 N preload. load speed can be estimated from the saturated speed of the transient response. The output force can be calculated from the rise in acceleration of the transient response and the mass of the movable parts (2.0 g). The no-load speeds of the flat-plane slider and the slider with projections are shown in Fig. 12 for a preload of 20 N. The driving voltage was varied from 30 to 125 V peak. To clarify the reproducibility of the driving characteristics, three measurements were made for each driving condition at each driving voltage. Figure 12 shows that the no-load speed of the flat-plane slider and the slider with projections increases as the driving voltage increases. Although the no-load speeds of the slider

1416 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 Figure 12. No-load speed characteristics of the flat-plane slider and slider with projections with a 20 Npreload. Figure 13. Output force characteristics of the flat-plane slider and slider with projections with a 20 N preload. with projections are higher than those of the flat-plane slider for each driving voltage, there are no significant differences between the flat-plane slider and the slider with projections. The maximum difference in no-load speed between the flat-plane slider and the slider with projections is less than 30% at 110 V peak driving voltage. The driving characteristics vary little between the results of the slider with projections and the flat-plane slider. Thus, we conclude that the flat-plane slider can provide stable drive using a chemically reduced lithium niobate stator. The output forces of the flat-plane slider and the slider with projections are shown in Fig. 13. Overall, the output forces of the flat-plane slider and the slider with projections are almost the same. The characteristics of the output force, however, differ slightly between the flat-plane slider and the slider with projections. The out-

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1417 put force of the slider with projections increases as the driving voltage increases from 35 to 125 V peak, whereas the output force of the flat-plane slider peaks at a driving voltage of 95 V peak. The output force generated from the friction force has a maximum value that is the product of the preload and the frictional coefficient between the stator and the slider. With a preload of 20 N and a frictional coefficient less than 0.2, the output force is less than 4 N. Thus, we conclude that the flat-plane slider delivers sufficient output force at a driving voltage of 95 V peak, whereas the slider with projections delivers sufficient output force at a driving voltage of 125 V peak. 5.3. Driving Characteristics as a Function of Preload To obtain higher output force, it is necessary to increase the preload. We therefore performed experiments with varying preloads to achieve a higher output force for the SAW motor. The preload was varied from 10 to 120 N and the driving voltage was set at 125 V peak. Two measurements under the same driving conditions were made with each preload. The no-load speeds of the flat-plane slider and the slider with projections are shown in Fig. 14, and decrease with increasing preload. Although for less than 70 N preload, the no-load speeds of the flat-plane slider are lower than those of the slider with projections, the flat-plane slider is faster than the slider with projections for preloads over 70 N. The maximum no-load speed of the flat-plane slider is 0.90 m/s with a 10 N preload, which is approximately 69% of the vibration velocity of 1.3 m/s, whereas the maximum no-load speed of the slider with projections is 0.93 m/s with a 10 N preload, which is approximately 72% of the vibration velocity of 1.3 m/s. We, thus, find that the no-load speeds of the flat-plane slider and the slider with projections are very similar. The output forces of the flat plane and the slider with projections are shown in Fig. 15. The output forces were calculated from the same transient responses used in the no-load speed experiment. For a preload below 60 N, the output forces of the flat-plane slider and the slider with projections are almost the same, and increase with increasing preload. Above a 60 N preload, the output forces of the flat-plane slider are higher than those of the slider with projections. The maximum output force of the flat-plane slider is 18 N with a 120 N preload, which is over twice that of the slider with projections. The reason the output force differs is related to the stiffness at the slider contact surface. Numerical simulations for sliders with projections [14, 18] indicate that a stiff surface is suitable for high-output-force operation. The stiffness of the slider surface is proportional to the number of projections, and small-diameter projections are stiffer than large-diameter projections. It was also demonstrated experimentally that smaller and denser projections lead to better driving properties [6]. If we extrapolate the smaller projection diameter and the higher projection density, we ultimately arrive at the flat-plane geometry.

1418 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 Figure 14. No-load speed characteristics of the flat-plane slider and slider with projections with a driving voltage of 125 V peak and a horizontal vibration velocity of 1.3 m/s. Figure 15. Output force characteristics of the flat-plane slider and slider with projections for a driving voltage of 125 V peak. Figure 16 illustrates the differences between the flat-plane slider and the slider with projections. When the preload becomes high, the surface of the slider with projections reaches (or is close to) the bottom of the wave where the particles move in the direction opposite to the slider motion. However, the surface of the flat-plane slider, which is stiffer than that of the slider with projections, only makes contact near the top of the wave under the same preload. Thus, under high preload, the flat-plane slider can obtain a higher output force than the slider with projections.

K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1419 Figure 16. Contact area between the stator and slider surface under high preload. 6. Conclusions To obtain stable drive, the SAW motor required projections on the slider surface; the flat-plane slider, which has no projections on its surface, could not provide stable drive because of contact electrification. To achieve stable drive by the flatplane slider, we used lithium niobate with its conductivity improved by chemical reduction for the stator. This paper reports the experiments involving driving the flat-plane slider. The flat-plane slider can provide stable drive by using chemically reduced lithium niobate for the stator. The flat-plane slider delivers 18 N output force at a driving voltage of 125 V peak with a 120 N preload, which is about twice that of the slider with projections. The no-load speed of the flat-plane slider is 0.9 m/s at a driving voltage of 125 V peak, which is about 69% of the vibration velocity of 1.3 m/s. Thus, in conclusion, the flat-plane slider not only simplifies the slider manufacturing process but also offers a higher output force to the SAW motor. Acknowledgements This work was supported by a grant-in-aid for the Development of Innovative Technology of the Ministry of Education, Culture, Sports, Science and Technology of Japan.

1420 K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 References 1. M. Kurosawa, M. Takahashi and T. Higuchi, Ultrasonic linear motor using surface acoustic waves, IEEE Trans. Ultrason. Ferroelect. Freq. Control 43, 901 906 (1996). 2. M. K. Kurosawa, M. Chiba and T. Higuchi, Evaluation of a surface acoustic wave motor with a multi-contact-point slider, Smart Master. Struct. 7, 305 311 (1998). 3. N. Osakabe, M. Kurosawa, T. Higuchi and O. Shinoura, Surface acoustic wave linear motor using silicon slider, in: Proc. IEEE Workshop Micro Electro Mechanical Systems, Heidelberg, pp. 390 395 (1998). 4. M. Kurosawa, M. Takahashi and T. Higuchi, Elastic contact conditions to optimize friction drive of surface acoustic wave motor, IEEE Trans. Ultrason. Ferroelect. Freq. Control 45, 1229 1237 (1998). 5. K. Asai, M. K. Kurosawa and T. Higuchi, Evaluation of the driving performance of a surface acoustic wave linear motor, in: Proc. IEEE Ultrasonics Symp., San Juan, pp. 675 679 (2000). 6. M. K. Kurosawa, H. Itoh, K. Asai, M. Takasaki and T. Higuchi, Optimization of slider contact face geometry for surface acoustic wave motor, in: Proc. IEEE Int. Micro Electro Mechanical Systems, Interlaken, pp. 252 255 (2001). 7. P. Helin, V. Sadaune, C. Druon and J. B. Tritsch, Linear ultrasonic motors using surface acoustic waves mechanical model for energy transfer, in: Proc. Transducers 97, Chicago, IL, pp. 1047 1050 (1997). 8. M. Vermeulen, F. Peeters, H. Soemers, P. Feenstra and P. Breedveld, Development of a surface acoustic wave planar motor under closed loop control, in: Proc. EUSPEN, Eindhoven, pp. 107 110 (2002). 9. P. J. Feenstra and P. C. Breedveld, Analysis of surface acoustic wave motor, in: Proc. IEEE Ultrasonics Symp., Honolulu, HI, pp. 1133 1136 (2003). 10. Y. Nakamura, M. K. Kurosawa and T. Shigematsu, Effects of ceramic thin film coating on friction surfaces for surface acoustic wave linear motor, in: Proc. IEEE Ultrasonics Symp., Honolulu, HI, pp. 1766 1769 (2003). 11. T. Shigematsu, M. K. Kurosawa and K. Asai, Sub-nanometer stepping drive of surface acoustic wave motor, in: Proc. IEEE Conf. on Nanotechnology, San Francisco, CA, pp. 299 302 (2003). 12. T. Shigematsu and M. K. Kurosawa, Friction drive of an SAW motor part I: measurements, IEEE Trans. Ultrason. Ferroelect. Freq. Control 55, 2005 2015 (2008). 13. T. Shigematsu and M. K. Kurosawa, Friction drive of an SAW motor part II: analyses, IEEE Trans. Ultrason. Ferroelect. Freq. Control 55, 2016 2024 (2008). 14. T. Shigematsu and M. K. Kurosawa, Friction drive of an SAW motor part III: modeling, IEEE Trans. Ultrason. Ferroelect. Freq. Control 55, 2266 2276 (2008). 15. T. Shigematsu and M. K. Kurosawa, Friction drive of an SAW motor part IV: physics of contact, IEEE Trans. Ultrason. Ferroelect. Freq. Control 55, 2277 2287 (2008). 16. T. Shigematsu and M. K. Kurosawa, Friction drive of an SAW motor part V: design criteria, IEEE Trans. Ultrason. Ferroelect. Freq. Control 55, 2288 2297 (2008). 17. Yamaju Ceramics, Specification Sheet: Black-LT Black-LN, Yamaju Ceramics, Aichi (2010). Available online at http://www.yamajuceramics.co.jp/english/product/blt.pdf 18. M. K. Kurosawa and T. Shigematsu, Friction drive simulation of surface acoustic wave motor characteristics based on contact mechanics, Jpn. J. Appl. Phys. 47, 4287 4291 (2008).

About the Authors K. Sakano et al. / Advanced Robotics 24 (2010) 1407 1421 1421 Koki Sakano received the BE degree in Electrical and Electronic Engineering from the Tokyo Institute of Technology, Tokyo, Japan, in 2008. He is currently a Master s course student of the Interdisciplinary Graduate School of Engineering, Tokyo Institute of Technology. He is a Student Member of the Acoustical Society of Japan. Minoru Kuribayashi Kurosawa received the BE degree in Electrical and Electronic Engineering, and the ME and DE degrees from the Tokyo Institute of Technology, Tokyo, in 1982, 1984 and 1990, respectively. Beginning in 1984, he was a Research Associate of the Precision and Intelligence Laboratory, Tokyo Institute of Technology, Yokohama, Japan. From 1992, he was an Associate Professor at the Department of Precision Machinery Engineering, Graduate School of Engineering, University of Tokyo, Tokyo, Japan. From 1999, he has been an Associate Professor at the Department of Advanced Applied Electronics, Interdisciplinary Graduate School of Engineering, Tokyo Institute of Technology, Yokohama, Japan. Since 2003, he belongs to the Department of Information and Processing in the same Institute. His current research interests include ultrasonic motors, micro actuators, PZT thin films, SAW actuators, and single-bit digital signal processing and its application to control systems. He is a Member of the Institute of Electronics Information and Communication Engineers, Acoustical Society of Japan, IEEE, Institute of Electrical Engineers of Japan, and Japan Society for Precision Engineering. Takashi Shigematsu received the BE, ME and DE degrees in Electrical and Electronic Engineering from the Tokyo Institute of Technology, Tokyo, Japan, in 2001, 2003 and 2007, respectively. He was a Visiting Postdoctoral Research Fellow at Heinz Nixdorf Institute, University of Paderborn, Paderborn, Germany, from April 2007 to March 2008, and then he was a Postdoctoral Research Fellow at RIKEN, Saitama, Japan. He joined Bosch Corporation, Saitama, Japan, in 2008. He is a Member of the IEEE.