Surface Acoustic Wave Atomizer with Pumping Effect

Transcription

1 Surface Acoustic Wave Atomizer with Pumping Effect Minoru KUROSAWA, Takayuki WATANABE and Toshiro HIGUCHI Dept. of Precision Machinery Engineering, Faculty of Engineering, University of Tokyo Hongo, Bunkyo-ku, Tokyo 113, JAPAN Abstract We propose a surface acoustic wave device for miniaturization of ultrasonic atomizers and for functional system construction. Actually, ultrasonic atomizers are rather smaller than other method, so that nebulizers which are handy to carry type use ultrasonic vibration. But they are to large for a pocket or an endoscope. We have obtained a fine mist with a first trial of surface acoustic wave atomizing device whose dimensions are pocketable size. This functional device is able to reduce the dimensions to mm-order as a system. The atomizer consists of a vibrator, a cover and a tube. The vibrator and cover material is 127.8" Y-cut LiNbO, to generate Rayleigh wave, a kind of surface acoustic wave. The Rayleigh wave is the driving force of pumping effect and atomizing effect. The pumping effect is caused by the wave motion of the elastic material and viscosity of the fluid like an ultrasonic motor. The atomizing rate varied according to the driving voltage. The maximum rate was about 0.1 ml per minute. The mean diameter was 19 ym. 1. Introduction Atomization of liquid causes increase of a surface area. For applications such as vaporing, drying, burning, chemical reactions and others, reaction speed is accelerated with this effect. Although, there are several ways for atomization, ultrasonic atomization technique is investigated [1][2] and widely used [3] especial for rather compact equipments. A merit of ultrasonic atomizers is based on the high energy conversion density of the piezoelectric materials. A lot of ultrasonic atomizers have been proposed and developed for some applications [4]-[6]. Ultrasonic atomizers which use capillary wave (surface acoustic wave of fluid) are well-known. With this method, the particle size depends on the driving frequency of ultrasonic vibration and the distribution of the particle size is in rather narrow range. To generate fine fog whose diameter is around 10 pm, the driving frequency become high about 1 MHz. In that case, a fountain type atomizer is applied. However, this method requires some depth of fluid to concentrate the acoustic power. So it is difficult to miniaturize and increase the frequency. We propose a surface acoustic wave atomizer for miniaturization of ultrasonic atomizers and functional system construction. Actually, ultrasonic atomizers are rather smaller than other method, so that nebulizers which are handy to carry type utilize ultrasonic vibration. But they are to large for a pocket or an endoscope. We have obtained a fine mist with a first trial of surface acoustic wave atomizing device whose dimensions are pocketable size. This functional device is able to reduce the dimensions to mm-order as a system. 2. Surface acoustic wave device for atomizer A fabricated surface acoustic wave (SAW) atomizer by way of experiment is shown in Fig. 1. The atomizer consists of a vibrator, a cover and a tube. The atomizing unit size is 52 mm x 22 mm x 2 mm as shown in Fig. 2. The vibrator and cover material is degree Y-cut X-propagation LiNbO, to generate the Rayleigh wave, a kind of surface acoustic wave. In the case of the surface k 52 Fig.1 Schematic shape of the surface acoustic wave atomizer. tor IEEE

2 acoustic wave, the propagating energy of the elastic wave is concentrated in the vicinity of the surface of the elastic medium. As shown in Fig.3, the vibration amplitude decreases in the depth direction. Usually, therefore, substrate thickness is the several times of the wave length so that the elastic wave will not come to the back surface. The vibration of the back surface and the side are negligible. The motion of the surface elastic media is indicated in the Fig.3. When the wave propagates to the left, a point of the surface turns counterclockwise as indicated. The elastic wave was generated by 20 pairs of inter digital transducer (IDT) which was supplied RF power. The dimensions of the IDT were 200 Fm electrode pitch, 100 Fm electrode strip width and 10 mm aperture. The Rayleigh wave length was about 400 pm so that the driving frcquency was about 9.5 MHz. The Rayleigh wave propagated from the IDT to the side of watcr in a gap between the vibrator and thc cover around 10 pi. This narrow gap prevented thc water coming out excessively. Through the water, the wave transmit to thc cover surface so that the Rayleigh wave propagated on the surface of the cover also. The Rayleigh wave is the driving force of pumping effect and atomizing effect. The pumping effcct is caused by the wave motion of the elastic material and viscosity of the fluid like an ultrasonic motor as shown in Fig. 4. The fluid is pulled out from the narrow gap to the side of the IDT, then spread out in thin layer onto the vibrator. On the way to the IDT, the fluid is atomized from the crest of the capillary wave as shown in Fig. 5. For atomization by ultrasonic vibration, rather high t=lm I particle mntion COVER C? I VIBRATOR Rayleigh Wave Fig. 2 The surface acoustic wave device for the atomizer: the substrate is degree Y-rotated LiNb03; the IDT has 10 mm aperture and 20 pairs of strip to generate 10 MHz Rayleigh wave. Fig. 4 Principle of the pumping effect through the morion of the Rayleigh wave. A + propagatlon '. '.. * * -. :. -* *... * '. '. * -. *. - MIST,.. a a. : m I * *. - 0 *. ' *. Fig. 3 Wave motion of the Rayleigh Fig. 5 Principle of the atomization with the capillary wave. 26

3 ~ intensity ultrasonic is required. In low frequency range ultrasonics below 1 MHz, therefore, some vibration systems to emphasize the vibration amplitude, such as solid mechanical horns, are necessary. Because the maximum vibration velocity of piezoelectric ceramics is about 30 to 50 cm/sec at most. In the case of above 1 MHz, power concentration mechanism in liquid is used. Hence, the depth of the liquid should be controlled so that the liquid surface come to the focal point to make atomize efficiently as illustrated in Fig.6. However, since the lithium niobate is superior material than ceramics such as PZT, the maximum vibration velocity is much larger than that of PZT as indicated in Fig.7. In the case of the vibrator as shown in Fig.2, the maximum vibration velocity was beyond 2 &sec. It seems that the vibration velocity is enough large to make fine droplets directly from the vibrating surface. The vibration amplitude of normal direction of the substrate was measured with a Laser Doppler vibrometer which is able to measure up to 20 MHz. 3. Characteristics of the atomizer Fluid: : PZT In experiments, tap water was used. To make the water spread on the vibrator in a thin layer, a few soap was added in the water. It was found that if the surface condition of the vibrator was appropriate, the soap was Fig. 6 A fountain type ultrasonic atomizer; frequency range from 1 to 5 MHz. Fig. 8 Picture of the surface acoustic wave atomizer Voltage (Vo-p) 1 1 cycle Fig. 7 Normal direction vibration velocity of the SAW device. Fig. 9 Driving wave form of the RF generator. 27

4 ... not required. In these experiments, however, a few soap was used. The atomizer worked as shown in Fig. 8. To produce some mists, large vibration velocity amplitude was required such as 1.2 &sec or more. If we drove the SAW element in continuously, the mean electric input power resulted in a huge amount of 25 W or more. Such a huge power causes a damage to the element. Therefore, intermittent drive of 1 m second interval (frequency of 1 khz) and about 10 % driving period was taken to operate at high vibration amplitude with low power input. Actual driving signal form is shown in Fig.9. In a result, the mean driving power was about 3 W, and then a half of this power was wasted at the opposite side of the IDT without being used for atomization. The actual power used for operation of the " Voltage (Vo-p) Fig. 10 Atomizing rate vs. driving voltage. atomizer was about 1.5 W. The atomizing rate varied accordingly to the driving voltage as shown in Fig. 10. At the low vibration amplitude up to 0.9 &sec, the mist was not generated. Although the surface of the water vibrated and the capillary wave was observed. To obtain the mist at stable condition, the water layer should be maintained to be thin. If the water layer become thick such as 1 mm or more, the mist was not generated any more. The maximum atomizing rate was about 0.1 ml per minute. In this measurement, the number of wave train was IOOO. The particle size distribution was measured with a Laser Doppler particle analyzer at several mm above the atomizing surface. The measurement was carried out at rather thin mist condition so that the particle analyzer was not suitable to measure the thick mist. Hence, the driving voltage of the IDT was maintained to be 50 Vo-p in the following experiments. Through the influence of the intermittent drive, a distribution of the particle diameters had two peaks as shown in Fig. 11. The first peak was about 10 pm which was considered to be caused by the capillary wave length determined by the driving frequency. The second peak was about 40 km. This peak was duc to the intermittent burst drive. It has been reported that this particle size was varied by the intermittent frequency. By the measurement shown in Fig. 11, the linear mean diameter was 19.2 pn and the sauter mean diameter was 34.3 pn. From the measurement of the flight velocity with the particle analyzer, the velocity of the small particles around 10 pm was about 2 m/s and the large ones was below 1 m/s * 100 d U Diameter (pm) Fig. 11 Distribution of the particle diameters. 5 IO- +Unear Mean - s- Sauter Mean 40 I - ~ j ;... <... >... i < >...,..._. n I I I I I I " Wave number Fig. 12 The mean particle size depence on the driving duration. 28

5 ~ -~ To examine the particle size dependence on the driving condition, the number of the burst wave trains was changed. The wave number was varied from 500 to The intermittent frequency of 1 khz was held. When we drove the atomizer at a short number of wave trains, the mean diameters became large as shown in Fig. 12. The mean diameters both of linear and sauter decreased about 30 % with increase of the burst wave number in the measured range. This is because the beginning of the each burst drive, large droplets are generated. On the contrary, the small particle size peak was not changed so much by the driving condition of the burst duration. To measure only small particles, the sampling point was fixed at about 3 cm high above the vibrator surface. The measured result is shown in Fig. 13. The mean diameters both of linear and sauter decreased about only 10 % in the measured range of the wave number from 400 to It is clear that to generate small particles, the wave number should be large. 4. Discussion Ultrasonically driven pumps which utilize a radiation pressure at finite amplitude have been proposed. Shiokawa proposed usage of a surface acoustic wave device for a pump whose operation frequency was 50 MHz[7], and recently, Bradley used a flexural plate wave@]. However, pumping mechanism that we propose is entirely different from these devices. This is a novel application of ultrasonic vibration to control fluid flow Linear Mean -e- Sauter Mean ~ -... ~~ ~.i... : I I I I I I I Wave number 1 We found that at rather low vibration level, fluid was driven by friction when thin liquid layer. As illustrated in Fig.14, the driven direction of the fluid was changed dependent on the thickness of the fluid layer and the intensity of the vibration. When the fluid was thin layer and the vibration level was low, the water moved opposite direction of the wave propagation. In the case of the radiation pressure drive, the water is drive to the propagating direction of the elastic wave. The phenomenon of the friction drive was observed at an open condition as shown in Fig. 14. It would be possible to apply this phenomenon to the closed condition as shown in Fig. 4. In this case, the Rayleigh wave transmits to the cover which is the same material as the vibrator through the water as illustrated in the figure. Usually, if a dimension is small, the viscosity of the liquid disturbs the flow due to the resistance at the boundary of a channel wall. This mechanism, however, uses the motion of the channel wall. Therefore, this mechanism is expected to be suitable for micro devices. Ordinary fountain type atomizers in MHz range require some depth of the liquid to concentrate the acoustical power as shown in Fig.6. In such a case, it will be difficult to reduce the size of the system. The SAW atomizing element dose not use focusing like fountain type ultrasonic atomizers, so that the depth of the fluid is very small and not critical. Hence, the element requires only electrodes and low in profile. With this method, the atomizer will be miniaturized by using higher driving frequency. If we can apply 100 MHz for the driving frequency, the dimensions will become one tenth. Miniaturization will enlarge application fields of the ultrasonic atomization. Due to the intermittent driving method to save the consumed power in the system, the generated droplet radiation direction SAW propagation Fluid I capillary wave cosity ve >>> I IDT Vibrator Fig. 13 The mean particle size of the smaller particles with the change of driving duration. Fig. 14 Driving force on the fluid at the condition of the liquid. 29

6 size distribution had two peaks. The larger peak around 40 pn is far kom the particle size expected by a capillary wave theory. These larger droplets would not be suitable for applications. To prevent the generation of these large particles, the driving wave form should be improved. About the smaller droplets, the particle size was larger than expected value of 3 pm. The reason of the difference should be examined in theoretically. The investigation of the efficiency of the atomizer from the input electrical power to the output result is interesting subject. The design of the IDT and other part should be improved for practical use. 5. Conclusion A new ultrasonic atomizer using surface acoustic wave at operation frequency of 10 MHz is described. The operation frequency is higher than the current fountain type ultrasonic atomizers. The increase of the operation frequency has been enabled with introduction of superior material and novel way of atomization. This construction of the atomizer has advantage of miniaturization. The atomizer is expected to have a function of a pump by using the same wave guide. That is also a merit to realize a functional fluid operation device in a small dimensions. The particle size of the fog seems to be rather larger than that of expected by the capillary wave theory. The linear mean diameter was 19 pm and the seater diameter was 34 pm The atomized quantity was 0.1 ml per minute. These characteristics will be improved. With this method, we will be able to make a miniature atomizer including a pump to feed a fluid within several mm. Since the pumping effect utilizes the viscosity of fluid, reduction of the dimensions dose not lower the pumping performance like other micro pump. For miniaturization, the driving frequency should be increased. If we apply 100 MHz for the driving frequency, the dimension of the atomizer become onetenth. Such a small atomizer will be useful for medical application together with an endoscope. References [l] R. J. Rang, "Ultrasonic atomization of liquid," J. Acoust. SOC. Am., VO1.34, pp.6-8 (1962). [2] L. D. Rozenberg, "Physical principles of ultrasonic technology vol. 2," Plenum Press, pp (19973). [3] C. Chiba, "Tyouonnpa funnmu (Ultrasonic spray)," Tokyo, Sunnkaidou, (in Japanese). [4] S. Ueha, N. Maehara and E. Mori, "Mechanism of ultrasonic atomization using a multi-pinhole plate," J. Acoust. Soc. Jpn. (E), v01.6, no.1, pp (1985). [5] B. Hadimioglu, et. al., "Acoustic ink printing," in Roc. of 1992 Ultrasonics Symposium, pp , [6] H. Iida, M. Kurosawa and S. Ueha, "An ultrasonic atomizer using squeeze film," in Roc. of Ultrasonics International 93, pp (1993). [7] S. Shiokawa, Y. Matsui and T. Ueda, "Study on SAW streaming and its application to fluid devices," Jpn. J. Appl. Phys., vo1.29, Suppl.20-1, pp (1990). [8] C. E. Bradley, R. M. White and R. M. Moroney, "Acoustically driven flow in flexural plate wave devices: theory and experiment," Presented at 1994 IEEE Ultrasonics symposium, Nov. 1-4, Cannes, France. Acknowledgments The authors would like to thank the engineers of Nippondenso Co., Ltd. in making the SAW devices. The help of Japan Laser Corp. in measuring particles is gratefully acknowledged. 30