A flat type touch probe sensor using PZT thin film vibrator

Ž. Sensors and Actuators 83 000 67 75 www.elsevier.nlrlocatersna A flat type touch probe sensor using PZT thin film vibrator T. Kanda ), T. Morita, M.K. Kurosawa, T. Higuchi Department of Precision Machinery Engineering, Graduate School of Engineering, UniÕersity of Tokyo, 7-3- Hongo, Bunkyo-ku, Tokyo 3-8656, Japan Received 9 June 999; received in revised form 7 December 999; accepted 9 December 999 Abstract We fabricated a touch probe sensor, having a flat configuration and using a PZT thin film vibrator. This sensor is intended for use in high-precision surface shape measuring tools at low contact force, for example, scanning probe microscopes Ž SPMs.. The sensor consists of an oscillator vibrating in the longitudinal direction. According to our calculations higher sensitivity can be obtained by using longitudinal vibration than by using bending vibration. The vibration was excited and detected by a hydrothermally deposited PZT thin film device. The length of the vibrator was 9.8 mm, and its resonance frequency was 304.35 khz. When the driving voltage was 3 Vp p at the resonance frequency, the vibration amplitude at the tip of the sensor was 6 nm o p. We used the flat configuration and miniaturized the oscillator to increase the quantity ratio of piezo film in comparison to the base material, thus improving the sensitivity. The sensitivity and the resolution were evaluated experimentally, with the vertical resolution estimated to be.4 nm. This sensor device will be effective for use in high-speed and high-resolution surface shape measuring tools without damage to nano scale construction. q 000 Elsevier Science S.A. All rights reserved. Keywords: Touch probe sensor; PZT thin film; Hydrothermal method. Introduction Touch probe sensors have been used as probes in surface shape measuring tools. For example, vibrating cantilevers are employed as touch probes in scanning probe microscopes Ž SPM. as well as in atomic force microscopes Ž AFM. wx. Most of these microscopes use bending vibrating cantilevers. However, a longitudinally vibrating probe has also been proposed for use in such measuring tools w,3 x. Since a longitudinal vibrator has a much higher Q value, and is not as influenced by the damping of air viscosity compared to a bending vibrator. When measurements are taken in fluid, longitudinal vibrators in particular will have much higher sensitivity than bending vibrators. In addition, a longitudinal vibrator has a much higher resonance frequency than cantilevers, thus enabling quick scanning of the probe. However, a much more important matter is that the longitudinal vibrator has much higher sensitivity than bending vibrators. This sensitivity can be defined as the transfer ratio from the vibration amplitude to the pickup voltage. We fabricated a touch probe sensor using a longitudinal vibrator. The vibration was generated by hydrothermally deposited PZT thin film wx 4. This previous sensor consisted of a rod vibrating in the axial direction. Its length was 7.8 mm and its resonance frequency was 6 khz. Its vertical resolution was higher than 67 nm wx 3. However, the pickup voltage contained the same frequency signal that was considered to be caused by induced interference from the driving current to the pickup wire and electrode. Thus the reduction of this interference signal will improve the sensitivity. In this paper, in order to obtain much higher sensitivity, we examine a much smaller and flat-type sensor. ) orresponding author. Tel.: q8-3-584-6466; fax: q8-3-5800-6968; http:rrwww.intellect.pe.u-tokyo.ac.jp. E-mail address: kanda@intellect.pe.u-tokyo.ac.jp Ž T. Kanda.. Present address: The Institute of Physical and hemical Research Ž RIKEN., Hirosawa -, Wako-shi, Saitama 35-098, Japan. Present address: Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, 459 Nagatsuta-cho, Midoriku, Yokohama 6-850, Japan.. Principle and structure.. Principle The schematic view of the touch probe sensor in the process of measurement is shown in Fig.. The sensor 094-447r00r$ - see front matter q 000 Elsevier Science S.A. All rights reserved. Ž. PII: S094-447 00 0098-3

68 ( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 Fig.. Schematic view of surface shape measurement by a touch probe sensor. consists of a half wavelength, longitudinally vibrating oscillator. An exponential horn enlarges the amplitude of vibration. When the tip end of the horn touches the surface of the workpiece, the resonance frequency of the vibrator shifts. By continuously detecting the tip s contact to the workpiece surface, the surface profile of the workpiece can be obtained. If we can use tapping mode, namely, a cyclic contact mode, we can minimize the damage to the surface. By using the tapping mode, the contact force will be smaller than existing tools for measuring surface roughness... Structure The structure of the touch probe sensor is shown in Fig.. The vibrator length is 9.8 mm. The step-up ratio of the exponential horn is 3.. The base material of the vibrator is titanium Ž Ti. substrate. The thickness of this substrate was 00 mm. On the surface of the Ti substrate, we deposited PZT thin film using the hydrothermal method w5,6 x. The thickness of the PZT film was about 0 mm on each side. The electrodes deposited on the surface of the PZT film were used to drive and pickup. Those electrodes were made of gold, and were deposited by vapor deposition. In order to reduce the induced interference, we used a differential amplifier. This is because the interference s value would be mainly due to the leakage between the power supply and the pickup wires. In addition, the differential amplifier circuit decreased the influence of the high impedance of PZT film. As shown in Fig. 3, the pickup voltage was defined as the differential potential between the voltage of the pickup electrode and that of the reference electrode. The differential amplifier circuit gain was set to 0. Ž. Ž. Fig.. Structure of the touch probe sensor: a photo, b schematic view. To compare the sensitivity of the sensor consisting of the longitudinal vibrator and the bending vibrator, we estimated the sensitivity of a bending vibrator sensor as shown in Fig. 4. Electrodes are positioned on each side, one is for driving, and the other is for pickup. The length, width, and thickness of the Ti substrate and that of PZT film are l, b, t, and t, respectively. There are PZT thin film layers on both surfaces of the Ti substrate..3. omparison of the sensitiõity Longitudinal vibrating sensors can effectively realize much higher sensitivities and much higher scanning speeds than can bending vibrating sensors. Fig. 3. Touch probe sensor with the deferential amplifier circuit.

( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 69 Fig. 4. Vibrator for evaluating sensitivity. To evaluate the pickup voltage of the sensor, an equivalent circuit of the vibrator is effective. Fig. 5 shows the equivalent circuit of the oscillator in Fig. 4 wx 4. Lump element circuit components of L, r, R, A and A are equivalent mass, equivalent elasticity modules, equivalent viscosity coefficient, and force factors in the driving piezo element and that in the pickup piezo element. d and d indicate the capacitance in the driving element and that in the pickup element due to the PZT film s ferroelectricity. By using these parameters, the pickup voltage can be described as wx 4 A A < V < s < V <. out ( ž / d d d A y v L q q v R in Ž. In Eq. Ž., v is the angular frequency. Vin is the driving voltage, and Vout is the pickup voltage. When the tip of the sensor is brought close to the workpiece, the pickup voltage of the sensor changes as shown in Fig. 6. In the region of the cyclic contact, we can measure the derivation of the gap between the sensor tip and the workpiece. In this region, in simplified model, the vibration amplitude is limited by the gap. Hence the vibration amplitude decreases linearly when the sensor is brought close to the workpiece. The declination of the vibration amplitude corresponds to the change of the pickup voltage. Sensitivity can be defined as the tangent of the slope in Fig. 6. Namely, sensitivity P is defined by Fig. 5. Equivalent circuit of the vibrator shown in Fig. 4. Fig. 6. Relationship between dd and dv. dvrdd as shown in Fig. 6. From the appendix, sensitivity P can be described as ( X AA AA y 4 ( Aq vr R d d X ž Ay vr Ldq / qvr R d P s. A A y v r R X ( ž X y vr L/ q vr R Ž. In Eq. Ž., vr is the angular resonance frequency at the freely vibrating condition. R X and X mean, R X srqr Ž 3. X s q. Ž 4. R and r indicate the equivalent elasticity modules and the equivalent viscosity coefficient at the surface of the workpiece. We calculated the sensitivity of each vibrator at the ratio between l, b, and t as 40:0:. The thickness of the PZT film, t, was constant at 0 mm. When the PZT thin film was deposited using the hydrothermal method, the thickness of the film was constant at about 0 mm. Fig. 7 shows the results of the calculation. The sensitivity of the bending vibrator does not change against the change of length l. However, the sensitivity of the longitudinal vibrator increases as the sensor miniaturized. This is because the strain becomes larger when the vibrator is miniaturized with the same vibration displacement is unchanged. This result means the miniaturization of the vibrator can effectively to obtain higher sensitivity, especially in the longitudinal vibrator. From the comparison of the sensitivity of both vibrators, the sensitivity of the longitudinal vibrator is higher than that of the bending vibrator in an order of 0 4.

70 ( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 Fig. 7. Relationship between the width of the vibrator and its sensitivity using the calculation. The calculation did not include the contribution of air viscosity and other factors. Air viscosity has a larger influence on the bending vibrator than on the longitudinal vibrator. Then, if we take into account the air damping effect, the difference will be much larger. The factors between the tip of both vibrators and the sample surface equally effect the sensitivity of each vibrator sensor. In this case, the sensitivity of the longitudinal vibrator is much higher than that of the bending vibrator. However, the longitudinal vibrator has much higher rigidity than the bending vibrator. Hence the contact force between the tip of the vibrator and the workpiece will be much larger in the longitudinal vibrator. However, the contact force also depends on the vibration amplitude. Therefore, by using the smaller vibration amplitude, the vibrator can operate using a smaller contact force. The longitudinal vibrator has smaller vibration amplitude than a bending vibrator when the driving or pickup voltage is the same value. Then as the step-up ratio of the horn becomes larger, a smaller contact force can be obtained. In addition to the comparison of the vibrator s sensitivity, the resonance frequencies can also be compared. When the length l s 4 mm, the resonance frequency of the longitudinal vibrator and the bending vibrator were 90 khz and 4. khz, respectively. The scanning frequency cannot exceed the resonance frequency. Thus, from the comparison of the resonance frequency of both vibrators, the longitudinal vibrating sensor has the possibility of realizing high-speed scanning. 3. Experiments 3.. Admittance of the Õibrator We measured the admittance change of the vibrator under a non-contact condition using an impedance analyzer. The results are shown in Fig. 8. The deviation of the admittance curve was detected around 304.5 khz. This shift would be caused by the resonance of the vibrator. The resonance frequency was 0 khz lower than the result of the analysis by the finite element method, which was 35 khz. Fig. 8. Admittance of the vibrator. The measurement showed that the admittance increased as the frequency increased. The phase of the admittance also changed. This means that the value tan d Ždielectric loss tangent. increased as the driving frequency increased, namely, the loss in the vibrator increased. In such a vibrator, the loss has to be considered in the equivalent circuit. The equivalent circuit in Fig. 5 must be converted to the circuit in Fig. 9. In this circuit, the resistance, R d and R d, describes the dielectric loss in the piezoelectric material. 3.. Vibration amplitude and pick-up To evaluate the relationship between the vibration amplitude and the pickup voltage, the velocity of the vibrator tip and the output voltage of the amplifier were measured as the function of driving frequency. A laser Doppler vibrometer and vector voltage meter were used for the measurement. The driving voltage was maintained at a constant value of 3 V. The results are shown in Fig. 0. p p The plots of the pickup voltage indicate the measurement value results divided by amplifier gain. The vibration amplitude and the pick-up voltage peaked at the resonance frequency of 304.35 khz. At that frequency, the vibration amplitude and pickup voltage were 6 nm o p and 3.36 mv. The mechanical Q value was 705. As can be seen rms from the plots of the pickup voltage, the induced interference voltage was small enough. Hence, the resonance frequency can be detected from the peak of the pick-up voltage. 3.3. EÕaluation by equiõalent circuit From the measurement of the vibration amplitude and mechanical Q value, we can obtain the value of each Fig. 9. Equivalent circuit of the touch probe sensor.

( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 7 Fig. 0. Relationship between the amplitude of the vibration and the pick-up voltage of the sensor by measurement: the vibration amplitude was measured by laser Doppler vibrometer. The pick-up voltage was measured using the vector voltage meter. parameter in the equivalent circuit. Parameters are shown in Table. Resistance Rd and Rd were calculated from the relationship between the admittance and the frequency shown in Fig. 8. By using this equivalent circuit in the way shown in Fig. 9, the pick-up voltage can be estimated. Fig. shows the calculated results of the pickup voltage. At the resonance frequency of 304.35 khz, the calculated pickup value was.4 mv rms. This value is smaller than the result of the measurement, although the frequency of the resonance is the same as the measurement. The source of error will be the exponential horn. The equivalent mass, L, is especially influenced by the step-up ratio of the exponential horn. The step-up ratio used in the calculation is that derived from the width of both ends of the horn. However, the ratio was actually diminished by the supporting part. With this, there was a restriction in the machining process to make the form of the element, especially in regard to the accuracy of the wired discharge cutting. So the horn did not consist of the exponential curve but of straight lines. Therefore the step-up ratio was different from that of the design. If the step-up ratio in the calculation were larger than the actual ratio, the equivalent mass and force factors would be smaller than actual values. Then the pickup voltage would then be smaller than the measurement result. As shown in Eq. Ž., especially under the condition that the term including the viscoelasticity of the sensor is large against the term including the force factor, the sensitivity of this sensor depends on the force factor. From the force factor in the equivalent circuit, the piezoelectric constant of the PZT film, e 3, can be esti- Fig.. Results regarding the relationship between the driving frequency and the pick-up voltage arrived at using the equivalent circuit. mated. When the shape of the oscillator is as shown in Fig. 4, the force factor can be described by using e w0 x; Asbe. Ž 5. 3 From Eq. Ž. 5 and the step-up ratio of the exponential horn, the piezoelectric constant e3 of the PZT film was y0.3 rm. The constant of bulk PZT is y3. rm Žcalcu- wx. 7. The pick-up voltage lated from the constants on Ref. as well as sensitivity depends on the force factor. The force factor is proportional to e, as shown in Eq. Ž. 3 5. So it is important to improve the piezoelectric constant in order to obtain high sensitivity. These results indicate that the sensitivity will be made much higher by improving e 3 of the PZT thin film. 3.4. Vibration in the bending direction This sensor has a flat configuration, and the electrode for driving is on one side. It is likely then that the thin structure causes bending vibration. If the bending amplitude is large, the resolution of the sensor is diminished not only laterally but also vertically. To evaluate the vibration amplitude in the bending direction, we measured longitudinal and bending vibration amplitude using laser Doppler vibrometers. Fig. shows the relationship between the amplitude in the longitudinal direction and that in the bending direction. 3 Table Parameters of the equivalent circuit based on measurement L Equivalent mass wkgx.0=0 R Equivalent viscoelastic loss wnsrmx 5.4=0 Equivalent compliance wmrnx.4=0 A Force factor wnrvx 8.7=0 y6 y3 y7 y5 Fig.. Relationship between longitudinal vibration amplitude and bending vibration amplitude.

7 ( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 At the longitudinal resonance frequency, 307.6 khz, the bending vibration velocity was 7 mmrs, which was measured with the laser Doppler vibrometer ŽPolytec OFV-300, OFV-50.. The bending vibration amplitude was only 36 pm although the longitudinal amplitude was 38 nm. The bending direction amplitude was only 0.03% of the longitudinal vibration. The bending amplitude was thus small enough and it was not so serious for the resolution of the sensor. 3.5. SensitiÕity and resolution To evaluate the sensitivity and the resolution of this sensor, we measured the relationship between pickup voltage and tip workpiece distance. Fig. 3 shows the experimental set up. The workpiece is an aluminum plate. The plate was driven by a layered type piezo element. The sensor output signal was amplified by the differential amplifier. The output voltage from the amplifier was measured using a vector voltage meter Ž HP8508A.. The results are shown in Fig. 4. The pickup voltage was measured when the workpiece was brought close to the vibrator tip. The plots of the pickup voltage are not divided by the differential amplifier gain. The plots dewx 8. From the left side of the graph, scribe the force curve each area corresponds to Ž A. the freely vibration, Ž B. the tapping vibration, and Ž. the contact regions. The variation of the pickup voltage is similar to that seen in Fig. 6. In the area Ž B. of Fig. 6, the change is linear. However, it is non-linear in Fig. 4. This will be because the vibrator body in the experiment has elasticity although it was defined as a rigid body in the theoretical model. From the change of the pickup voltage in region Ž B., the sensitivity of this sensor is obtained. From this sensitivity the vertical resolution of this sensor can be estimated. The sensitivity can be defined as the tangent of the tapping Ž. mode in region B of Fig. 4. The slope will change for Fig. 3. Experimental setup for estimation of the relationship between the tip sample distance and the pickup voltage. Fig. 4. Relationship between the pick-up voltage of the sensor and the displacement of the workpiece by the measurement: each area Ž A., Ž B., and Ž. corresponds to the free vibration, the tapping vibration, and the contact regions. the workpiece, which in this estimation was aluminum. From the slope of the curve in Fig. 4, the sensitivity was y.0=0 mvrnm. From the calculation by Eq. Ž., the sensitivity was estimated to be 7.3=0 y mvrnm. This difference would be mainly mechanical loss such as the loss by supporting. The resolution of the sensor corresponds to the minimum detectable voltage divided by the sensitivity. The minimum detectable voltage is defined as the equivalent background noise level at the input terminal of the amplifier. This means the signal-to-noise ratio is 0 db when the minimum detectable voltage equals the equivalent background noise level. The equivalent background noise included the sensor noise, the amplifier noise, and others. The resolution can be defined as the minimum detectable scale corresponding to the minimum detectable voltage. When V min, Õ m, G, P, and Pr indicate the minimum detectable voltage, the noise level measured by the vector voltage meter in Fig. 3, the gain of the differential amplifier, the sensitivity of the sensor, and the resolution. The resolution can be estimated by the following equations, n m Vmin s Ž 6. G V min Pr s. Ž 7. P From the measurement of the noise level 0.4 mvrms at the output of the differential amplifier measured using the vector voltage meter, and the sensitivity as mentioned above, the resolution was.4 nm. This value was smaller than that of our previous probe sensor wx 4, thus improving the resolution. In this experiment, the operational amplifier devices Ž LF356. in differential amplifier were not the low-noisetype, and the bandwidth Ž 40 khz. was wide. However, the noise level was measured using a vector voltage meter that detected only the referred frequency. Hence the wide bandwidth is not significant. That noise would be caused by the thermal noise of the PZT film, and by the mechanical vibration of the oscillator. The noise caused by the

( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 73 capacitance of the film would be especially serious because the impedance of the PZT film of the oscillator is very high. 4. Discussion Based on the measurements, the piezoelectric constant e3 of the hydrothermally deposited PZT film was only 4.3% of the bulk PZT. On the other hand, from the measurement, the piezoelectric constant d3 was y34. rn wx 9. This value corresponds to 37% of the bulk PZT material d, y93.6 rn wx 3 7. Yet the piezoelectric constant e3 is an important factor in regard to sensitivity. The piezoelectric constant e3 is derived from the product of d constants and modulus of elasticity. Then, by comparing the measured constants d3 and e 3, it can be said that the modulus of elasticity of the PZT film is too small. The reason of the low elasticity of the PZT film seems to be the material s structure. Fig. 5 shows the surface of the PZT thin film of the touch probe sensor. As this photograph shows, there is a lot of void space between PZT crystals. It can be considered that the density of the film, including void space, will be small, and the binding force between crystals will be weaker. These matters correspond to the fact that the modulus of elasticity is also small. To increase the modulus of elasticity, the amount of void must be reduced. If the crystal can be grown much larger, the amount of void will be reduced, and the binding force can be improved. The modulus of elasticity will then be much larger. The piezoelectric constant, e 3, has an important relationship to the sensitivity and the resolution of the sensor. From the measurements, the pickup voltage and the amplitude under the damped condition are much smaller than Fig. 5. Photo of the surface of the PZT thin film deposited using the hydrothermal method. those in the freely vibrating condition are. Hence, Eq. Ž. can be transferred as A A 4 (Aqvr R d Ps. Ž 8 A. v r R Then, under the condition that A equals A, the sensitiv- ity can be described as A Ps ). Ž 9. A q d ž v R / r d From the results of the measurement shown in Table, the relationship A ž v R / r d y6-0 is obtained. Hence the sensitivity can be said to be in proportion to the force factor. From Eq. Ž. 5, the sensitivity can be said to be in proportion to the piezoelectric constant e 3, and the resolution is in inverse proportion to sensitivity. So, the piezoelectric constant e3 of the PZT film should be improved. If the constant, e 3, of the PZT film is about 3 times as large as the present film constant, which is the same as that of the bulk PZT material, the resolution will be up to 0. nm. 5. onclusion We estimated and compared the sensitivity of the longitudinal and the bending vibrating sensor. The results demonstrated that the longitudinal vibrating sensor has much higher sensitivity than the bending vibrating sensor as the vibrator was miniaturized. As well, the resonance frequency of the longitudinal sensor is higher than that of the bending sensor. Hence, the longitudinal vibrating sensor device will be effective for use in high-speed and high-resolution surface shape measuring tools. We fabricated a touch probe sensor using hydrothermal deposited PZT thin film, and evaluated its sensitivity and resolution. The resonance frequency and the tip amplitude at that frequency were 304.35 khz and 6 nm o p, respec- tively. The sensitivity of the sensor was.0=0 y mvrnm. At this sensitivity, the vertical resolution of this sensor was estimated to be.4 nm. We could achieve higher sensitivity and resolution than those described in our previous paper wx 4 by miniaturizing the vibrator and using a flat configuration. In order to increase the sensitivity and the resolution of this touch probe sensor, the force factor must be increase.

74 ( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 Measurements indicated that the piezoelectric constant e 3 was only 4.3% of the bulk PZT material, while d3 was 37%. These results show that the piezoelectric constant, especially constant e 3, must be improved. The discussion indicates as well that the modulus of elasticity has to be increased. By improving the piezoelectric constant of the PZT film, we will obtain higher sensitivity and higher resolution. Appendix A Sensitivity is described as Vout Ž R. yvoutž F. Ps. Ž A.. už R. yuž F. The parameter ur Ž. indicates the vibration amplitude at the tip and V Ž R. out means the pickup voltage when the oscillator vibrates freely at its resonance frequency. The amplitude ur Ž. equals the distance between the surface of the workpiece and the tip of the sensor in the start point of the cyclic contact mode. The parameters uf Ž. and V Ž F. out are the vibration amplitude and the pickup voltage under the continuous contact condition. The continuous contact condition indicates the condition in which the tip maintains constant contact with the workpiece. This condition also indicates the condition at the end of the cyclic contact mode. In this contact condition, the factors such as the equivalent elasticity modules and the equivalent viscosity coefficient at the surface of the workpiece influence the pickup voltage and the vibration amplitude. In Eq. Ž., the relationship between the driving voltage and the pickup voltage is described. Especially in the resonance frequency, the output voltage can be described as A A < V < s < V < out in. Ž A.. 4 (v R qa r d In Eq. Ž A.., vr is the angular resonance frequency. From the electro-acoustic transfer equation, the vibration amplitude at the tip of the oscillator is described as wx AV in us. Ž A.3. vz In Eq. Ž A.3., u and Z signify the amplitude and the mechanical impedance, respectively. When the oscillator vibrates freely, the mechanical impedance equals R. When the oscillator is in the contact condition, the mechanical impedance is described as ž / ZsRqj v Ly. Ž A.4. v The vibration amplitude can then be described as A < už R. < s < V < in Ž A.5. v R < už F. < s A < V < in. Ž A.6. yv L qv R (ž / Then, when the tip of the sensor is brought close to the workpiece with the frequency maintained at the resonance frequency of the freely vibration, from Eqs. Ž., Ž A.., Ž A.., Ž A.5. and Ž A.6., sensitivity P can be described as ( X AA AA y 4 ( Aq vr R d d X ž Ay vr Ldq / qvr R d P s. A A y v r R X ( ž X y vr L/ q vr R Ž. X X In Eq. A.7, R and indicates, Ž A.7. R X srqr Ž A.8. X s q. Ž A.9. R and r signify the equivalent elasticity modules and the equivalent viscosity coefficient at the surface of the workpiece, respectively. When the contact takes place, the values of R and r change from R sr s0 to not equal 0, and the condition shifts from the resonance. Ž. So sensitivity P can be described by Eq.. References wx G. Binnig,.F. Quate, h. Gerber, Atomic force microscope, Phys. Rev. Lett. 56 Ž 986. 930 933. wx S.M. Harb, M. Vidic, Resonator-based touch-sensitive probe, Sens. Actuators A 50 Ž 995. 3 9. wx 3 M. Nishimura, K. Hidaka, M. Teraguti, Dependence on directions of the sensitivity of the vibrational touch sensor, Proc. Ann. Spring Meeting JSPE, 994 Ž 994. 795 796 Ž in Japanese.. wx 4 T. Kanda, T. Morita, K. Kurosawa, T. Higuchi, A rod-shaped vibro touch sensor using PZT thin film, IEEE Trans. UFF 46 Ž.Ž 4 999. 875 88. wx 5 K. Shimomura, T. Tsurumi, Y. Ohba, M. Daimon, Preparation of lead zirconate titanate thin film by hydrothermal method, Jpn. J. Appl. Phys. 30 Ž 9B. Ž 99. 74 77. wx 6 T. Morita, T. Kanda, M. Kurosawa, T. Higuchi, Single process to deposit lead zirconate titanate Ž PZT. thin film by a hydrothermal method, Jpn. J. Appl. Phys. 36 Ž 5B. Ž 997. 998 999. wx 7 B. Jaffe, W.R. ook, H. Jaffe, Piezoelectric eramics, Academic Press, London, 97. wx 8. Lee, T. Itoh, T. Suga, Self-excited piezoelectric PZT microcan-

( ) T. Kanda et al.rsensors and Actuators 83 000 67 75 75 tilevers for dynamic SFM-with inherent sensing and actuating capabilities, Sens. Actuators A 7 Ž 999. 79 88. wx 9 T. Morita, M.K. Kurosawa, T. Higuchi, A cylindrical micro ultrasonic motor using PZT thin film deposited by single process hydrothermal method Ž.4 mm, L s0 mm stator transducer., IEEE Trans. UFF 45 Ž.Ž 5 998. 78 87. w0x T. Ikeda, Fundamentals of Piezoelectricity, Oxford University Press, Oxford, 990, p. 86. wx T. Ikeda, Fundamentals of Piezoelectricity, Oxford University Press, Oxford, 990, p. 98. Biographies Takefumi Kanda was born in Fukuoka, Japan, on June 8, 97. He received the B. Eng. and the M. Eng. in precision machinery engineering from The University of Tokyo, Japan in 997 and 999. He is currently a doctoral course student of the Graduate School of Engineering. His research interests are micro sensor and PZT thin film. He is a member of the Japan Society for Precision Engineering. Takeshi Morita was born in Tokyo, Japan, on August 4, 970. He received the B. Eng., the M. Eng. and Dr. Eng. in precision machinery engineering from The University of Tokyo, Japan in 994, 996 and 999. He is currently a special postdoctoral researcher in The Institute of Physical and hemical Research Ž RIKEN.. His research interests are micro ultrasonic motor and PZT thin film. He is a member of the Institute of Electrical Engineers of Japan, the eramic Society of Japan, and the Japan Society for Precision Engineering. Ž. Minoru Kuribayashi Kurosawa formerly Kuribayashi was born in Nagano, Japan, on April 4, 959. He received the B. Eng. in electrical and electronic engineering, and the M. Eng. and Dr. Eng. from Tokyo Institute of Technology, Tokyo, in 98, 984, and 990, respectively. He was a Research Associate at the Precision and Intelligence Laboratory, Tokyo Institute of Technology, Yokohama, Japan, since 984, and an Associate Professor at the Graduate School of Engineering, The University of Tokyo, Tokyo, Japan, since 99. Since 999, he has been an Associate Professor at the Interdisciplinary Graduate School of Science and Engineering, The Tokyo Institute of Technology, Yokohama, Japan. His current research interests include ultrasonic motor, micro actuator, PZT thin film, SAW sensor and actuator, and single bit digital signal processing and its application to control systems. Dr. Kurosawa is a member of the Institute of Electronics Information and ommunication Engineers, the Acoustical Society of Japan, IEEE, the Institute of Electrical Engineers of Japan and the Japan Society for Precision Engineering. Toshiro Higuchi was born in Ehime, Japan on February 6, 950. He received the B. Eng., MS, and PhD in precision engineering from The University of Tokyo, Japan, in 97, 974, and 977, respectively. He was a Lecturer at the Institute of Industrial Science, The University of Tokyo in 977 and an Associate Professor in the same institute from 978. Since 99, he has been a Professor at the Graduate School of Engineering, The University of Tokyo. His present interests include mechatronics, magnetic bearing, stepping motors, electrostatic actuator, robotics and manufacturing. Dr. Higuchi is a member of the Japan Society for Precision Engineering, the Japan Society of Mechanical Engineers, and the Society of Instrument and ontrol Engineers.