- Email: [email protected]
Liquid-state motion sensing
Sensors and Actuators B 154 (2011) 33–40
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Liquid-state motion sensing夽 Hansong Zeng, Yi Zhao ∗ Laboratory for Biomedical Microsystems, Department of Biomedical Engineering, The Ohio State University, 294 Bevis Hall, 1080 Carmack Road, Columbus, OH 43210, USA
a r t i c l e
i n f o
Article history: Available online 6 December 2009 Keywords: Motion detection Droplet dynamics Frequency response Superhydrophobicity
a b s t r a c t This paper demonstrates a liquid droplet-based motion sensing system which has the advantages of simple fabrication, low power consumption and digital signal processing. The sensor consists of a dielectric substrate patterned with an array of microelectrodes, and a saline droplet as the proof mass. Once an external linear acceleration is applied, the inertial force moves the droplet on the micropatterned substrate. The acceleration is determined from the movement profile detected by the microelectrodes. In order to enhance the threshold and the sensitivity of motion sensing, two surface treatment approaches are utilized to create superhydrophobic surfaces. The result shows that the minimal sliding angle that can move a 20 l droplet on the superhydrophobic surface is lower than 1◦ , corresponding to a threshold of lower than 0.017 g. A lumped-parameter model is developed to estimate the dynamic behavior of the proposed system. The result shows that the frequency response of the droplet-based sensor is more significant at low frequencies than at high frequencies, which is distinct from solid-state accelerometers. Measurement under a constant acceleration shows that the predicted value derived from the measurement has a good match with the actual applied acceleration, validating the proposed system as a viable alternative for motion sensing. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Detection of motion parameters, such as linear acceleration and angular change, is critical for a dynamic system. For example, motion detection is a key element of industrial applications including inertial navigation, passive automotive safety systems and simulation of space microgravity [2,3]. Increasingly smaller motion sensors have generated a great deal of interest among the booming gaming industry, where sensors serve as the interface between the gaming systems and the player. There long has been a practice of determining motion parameters using a solid-state MEMS sensor (i.e., accelerometer or gyroscope) that has high precision, low cost and small size [4]. In such a sensor, the motion change induces change of mechanical strain, displacement, or resonant frequency shift of the solid-state proof mass. A piezoresistive, piezoelectric or capacitive sensing component converts such change into electrical signals that can be recognized and processed [5,6]. Although effective, solid-state motion sensors are limited to a certain extent by complex fabrication and packaging processes. In these sensors, free-standing microstructures usually serve as the
夽 This paper was presented at the 15th International Conference on Solid-State Sensors and Actuators and Microsystems held in Denver, CO, June 21–25, 2009, and is an expansion of the abstract as printed in the Technical Digest of this meeting [1]. ∗ Corresponding author. Tel.: +1 614 247 7424; fax: +1 614 292 7301. E-mail address: [email protected] (Y. Zhao). 0925-4005/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2009.11.069
proof mass, and are fabricated by surface micromachining (patterning and releasing). Since the sensing performance is highly sensitive to fabrication imperfection [7], the tools for fabricating, characterizing and calibrating these free-standing structures must have high precision. Moreover, extraordinary attention is paid in the packaging process, not only to reduce damping effects but also to protect the mechanically fragile, free-standing structures. Meanwhile, in sensors using piezo-materials, the sensing performance is largely determined by the homogeneity, sensitivity and linearity of the piezo-sensing materials, which all vary with deposition conditions [8]. In addition, the sensitivity and the measurement range of a solid-state motion sensor often are intrinsically coupled, which makes it difficult to develop a sensor with a very high-level of sensitivity and a large measurement range, simultaneously [9,10]. In this work, we demonstrate a new concept of motion sensing that utilizes a liquid droplet as the inertial proof mass. Externally applied acceleration induces an inertial force on the droplet and moves it over the substrate. The droplet movement is detected by the time sequence of the electrical impedance measurement using microelectrodes patterned on the substrate. The externally applied acceleration is derived from droplet motion characteristics based on a droplet dynamics model. Unlike solid-state counterparts, a liquid-state motion sensing system requires minimal configuration and fabrication complexity. Meanwhile, the acquisition of digital signals solely depends on the electrical impedance between the measuring electrodes, which are not affected by the magnitude of external acceleration. Therefore, the sensitivity of the measurement is decoupled from the measurement range. Digital signal
34
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
Fig. 2. Illustration of the droplet dynamics during motion. The driving force of droplet motion is induced by the external applied acceleration (a with the red arrow) and is proportional to the mass of the liquid droplet. The motion of the droplet is governed by the driving force, the capillary force which is induced by dynamic contact angle hysteresis between the advancing contact angle ( A ) and receding contact angle ( R ), the air damping and the contact-line friction. The relative motion between the droplet and the substrate occurs once the capillary force is overcomed. For analysis simplicity, it is assumed that the damping/friction force is linearly related to the traveling velocity of the droplet. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Fig. 1. The concept of liquid droplet-based motion sensing. (a) The work is inspired by observing a drop of dew moving on the hydrophobic surface of a lotus leaf. (b) Schematic configuration of a liquid-state motion sensor.
processing also makes the sensor less vulnerable to external electric/magnetic disturbance. This is especially important for slow motion detection, e.g. body movement. 2. Design and analysis 2.1. Concept of liquid-based motion sensing The concept of liquid-based motion sensing is inspired by a natural phenomenon shown in Fig. 1(a), where a drop of dew on the hydrophobic surface of a lotus leaf is susceptible to a tiny external perturbation [11]. The relative movement of the droplet with respect to the leaf can be used as a measure of the external perturbation. Based on this observation, an engineered liquid droplet-based motion sensing system is designed, which consists of a hydrophobic substrate patterned with an array of microelectrodes (Fig. 1(b)). Like the dew rolling over the lotus leaf, the motion of the ionic droplet depends on the externally applied acceleration and surface hydrophobicity of the substrate. As it moves over the surface, the conductive ionic droplet changes the electrical connection states of the microelectrodes, thus allowing the motion characteristics of the droplet to be determined. In order to demonstrate the concept of liquid-state motion sensors, a simplified one-dimension prototype with one-time sensing capability is developed in this work. An array of microelectrodes is patterned on a planar dielectric substrate to detect the motion profile of the liquid droplet. Dual-axis and continuous sensing can be implemented by patterning an array of bi-layered microelectrodes on a curved substrate. 2.2. Lumped model for droplet dynamic study A lumped-parameter model is developed to quantitatively understand the droplet dynamics that govern the sensing perfor-
mance (Fig. 2). As seen from a coordinate system attached to the substrate, the driving force is in the opposite direction of the externally applied acceleration and has a magnitude of m × a, where m is the mass of the droplet and a is the acceleration to be determined [12]. In order to move the droplet, the driving force must overcome the capillary force caused by the dynamic contact angle hysteresis that resists the droplet motion [13]. Once the droplet moves on the surface, it is subjected to a damping/friction force which includes air damping as well as contact-line friction [14]. To simplify the analysis without significantly affecting the accuracy, it is assumed that the damping/friction force linearly relates to the droplet velocity. The governing equation of the droplet can be expressed as: m
dx d2 x − (m × a − Fthreshold ) = 0 +b dt dt 2
(1)
where x is the relative displacement, b is the damping/friction coefficient and Fthreshold is the capillary force. For a 20 l ionic droplet (0.9% saline) used in this work, m = 20 mg. The threshold (Fthreshold ) in Eq. (1) is an important parameter that determines the minimal acceleration that can be detected by the sensing system. In order to obtain a small dynamic contact angle hysteresis, a low surface wettability is required. This is usually performed by changing the surface topography, surface chemistry or both [15–17]. In this work, surface hydrophobicity is regulated using two independent approaches. Fig. 3(a) shows the first approach in which a layer of silica nanoparticles (average diameter 7–10 nm) is spin-coated on the surface (at 1000 rpm for 10 s), followed by immersion into an aqueous HDFT solution (3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-1decanethiol, Sigma–Aldrich® , CA, USA). The silica nanoparticles create nano-roughened surface topography, while the HDFT solution assembles a molecular layer of fluorocompound on the surface to further increase surface hydrophobicity [18]. After the treatment, the sliding angle of a 20 l (about 20 mg in weight) saline droplet is measured around 4◦ , corresponding to a threshold of 0.07 g (g being the gravity). Fig. 3(b) shows the second approach for regulating the surface hydrophobicity, in which the surface is treated with atmospheric plasma of CF4 –H2 –He. Both fluorocompound deposition and surface roughening are involved in the process [19]. Using this approach, the sliding angle of a 20 l saline droplet is below 1◦ , corresponding to a threshold lower than 0.017 g. Following the threshold determination, the damping/friction coefficient is estimated by optical motion analysis. Fig. 4(a) illustrates the experimental setup for analyzing the droplet motion characteristics. A precise tilting stage is used to apply a constant acceleration to the saline droplet. The motion of the droplet is recorded by a high-speed digital camera at a frame rate of 40 frames
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
35
Fig. 3. Two approaches are used to modulate surface hydrophobicity. (a) In the first approach, silica nanoparticles are coated on the substrate, which is followed by immersing the surface in HDFT solution. The surface is first roughened by silica nanoparticles, and subsequently decorated by a self-assembled molecular layer of fluorocompound which reduces the surface energy; (b) in the second approach, the surface is treated by CF4 –H2 –He plasma for both surface roughening and fluorocompound deposition; (c) static contact angle measurement of the surfaces treated by the two approaches. The samples treated by the first approach have a static saline contact angle of 153◦ , a dynamic contact angle hysteresis of 6◦ , and a sliding angle of 4◦ . The samples treated by the second approach have a static saline contact angle of 162◦ , a dynamic contact angle hysteresis of 3◦ , and a sliding angle below 1◦ . All the measurements are averaged from 3 samples.
per second (fps). Fig. 4(b) is a series of snapshots showing the droplet movement under a constant acceleration of 0.573 g, corresponding to a tilting angle of 35◦ . The position of the droplet at specific time points is determined from each snapshot. The motion profile of the droplet (displacement–time curve) during a certain period of time can be plotted (Fig. 5), where typical motion parameters including displacement, velocity, and acceleration can be derived. It is clearly seen that in the initial phase of the movement, the traveling velocity of the liquid droplet increases. As a result, the damping/friction force increases. Once the damping/friction force equals the driving force, the traveling velocity reaches a critical value and the droplet starts to move at a constant velocity. This is indicated by a linear displacement–time curve. When the balance is reached, the damping/friction coefficient can be determined based on Eq. (1). For a 20 l saline droplet rolling over a substrate treated by the second surface treatment approach (CF4 –H2 –He plasma), the damping/friction coefficient is determined as 6.73 N (m/s)−1 . Frequency response of droplet-based motion sensing is also analyzed using a lumped-parameter model. Fig. 6 compares the equivalent circuits of the droplet-based motion sensing system and a typical solid-state accelerometer [20]. The threshold is represented by a power source whose polarization is opposite to the source denoting the driving force. Since there is no restoring force in a liquid-state sensing system with a planar surface, the capacitor is absent. As a result, the liquid-state system does not have a resonant frequency. The displacement is infinite under a constant acceleration and reduces as the frequency increases. The frequency response is obtained numerically using Matlab Simulink software. The dimension-less values are plotted against the frequency (Fig. 6(c) and (d)).
Fig. 4. Optical determination of the droplet motion profile. (a) The experiment setup is composed of a tilting stage that generates a constant external acceleration, and a digital camera that records the motion of the droplet; (b) the time sequence images show the motion of the droplet during 1.5 s.
Fig. 5. The damping/friction coefficient is determined from the motion profile of the droplet. In the initial phase, the velocity of the droplet increases. The damping/friction force increases accordingly. Once the driving force and the damping/friction force become the same, the droplet moves at a constant velocity. The damping/friction coefficient can thus be determined.
3. Experiment and results 3.1. Fabrication An array of Au/Ti microelectrodes is patterned on a planar glass substrate using standard photolithography and lift-off processes
36
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
Fig. 8. Electrical impedance measurement from a pair of adjacent microelectrodes while the inter-electrode space is filled with a saline droplet or with air. At a stimulating frequency of 100 Hz, the ratio of the electrical impedance between the microelectrode pairs with air-filled inter-electrode space and with liquid-filled inter-electrode space are about 1000. Such a difference is sufficient for digital detection the presence of a saline droplet. Fig. 6. Comparison of the frequency responses of a liquid-state sensor and of a solidstate accelerometer. The equivalent circuits of the two systems are illustrated in (a) and (b). (c) Since there is no capacitor, the droplet-based system has high response in the low frequency regime. (d) In a solid-state system, high response is around the resonant frequency.
(Fig. 7(a)). In this work, the width of the electrodes is 200 m, and the separation of neighboring electrodes is 1 mm. After patterning, the entire surface of the device is treated with CF4 -H2 -He plasma as described in Section 2. A prototype is shown in Fig. 7(b). 3.2. Impedance analysis Electrical impedance between the adjacent electrodes is measured using an electrical impedance analyzer (Solartron® 1260A). A stimulating AC current of 300 A is applied between the adjacent electrodes without damaging the surface layer. Electrical impedance is obtained as a function of the stimulating frequency (Fig. 8). It is seen that the presence of the saline droplet can be easily identified from the magnitude of the electrical impedance. For example, the impedance magnitude of an air-filled electrode pair at 100 Hz is about 1000 times greater than that of a saline dropletfilled electrode pair. Such a large difference is sufficient for digital detection. Fig. 9(a) shows a representative measurement while a
Fig. 9. Static measurement of the droplet position. (a) The magnitude of the electrical impedance of each adjacent pair of microelectrodes is recorded, which indicates that a liquid droplet is present between the second and third electrodes and (b) the measurement is converted to digital values by applying an impedance threshold (1 × 105 ).
Fig. 7. Prototype fabrication. (a) Surface micromachining is used to create the micropatterned substrate. The superhydrophobic surface is obtained by CF4 –H2 –He plasma treatment. (b) shows a prototype with a saline droplet on top of the substrate. In a typical prototype for the measurement, each microelectrode is 200 m in width. The separation between adjacent microelectrodes is 1 mm.
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
37
Fig. 12. The time sequences of the electrical voltages in all the channels are recorded during the droplet motion. Fig. 10. Experiment setup for acceleration measurement. A constant acceleration is generated by titling the substrate at a certain angle with respect to the horizontal plane.
3.3. Acceleration measurement
Fig. 12 shows the analog signals obtained by the DAQ system at the tilting angle of 35◦ , which corresponds to a constant acceleration of 5.62 m/s2 (0.57 g) along the moving direction of the droplet. The figure shows the time sequence of voltage variations in all five adjacent electrode pairs. In order to demonstrate digital data processing, the analog voltages are converted to digital signals (Fig. 13). In this measurement, a voltage higher than 0.3 mV is considered to be high-level, indicating the presence of a saline droplet. Each rising edge of the voltage curve indicates the moment when the advancing edge of the saline droplet contacts the electrodes. The falling edge of the voltage curve indicates the moment when the receding edge of the saline droplet leaves the electrodes. The position of the droplet as a function of time is therefore obtained. The first order derivative (dx/dt) and the second order derivative (d2 x/dt2 ) of the droplet motion profile correspond to the velocity and acceleration, respectively. By inserting the droplet velocity and droplet acceleration into Eq. (1), the externally applied acceleration that induces relative droplet motion is determined. In order to validate the lumped model, the acceleration measurements are performed at four different externally applied accelerations, namely 5.62, 6.23, 6.92 and 7.51 m/s2 , which correspond to the tilting angles of 35◦ , 40◦ , 45◦ and 50◦ , respectively. The comparison between the values predicted by the lumpedparameter model and the external acceleration actually applied (Fig. 14) shows that the liquid-state sensor can detect linear accelerations with high precision, as evidenced by a high coefficient of determination of 0.982.
In order to validate the utility of the liquid-state system for motion detection, the system is operated under constant accelerations generated by an experimental setup shown in Fig. 10. The setup consists of a tilting stage, which generates a constant external acceleration to the liquid droplet by tilting the planar substrate at a certain angle with respect to the horizontal plane. During the experiments, the electrical signals acquired from the prototypes are recorded through a multi-channel data acquisition (DAQ) interface with the circuit schematized in Fig. 11. An AC power supply (3 V, 100 Hz) is used to provide the electrical stimuli. A reference resistor (1 k) is connected to each adjacent pair of microelectrodes to form a testing branch. The saline droplet behaves as an electrical switch, which either connects or disconnects the adjacent electrodes while it rolls over the planar substrate. Consequently, the voltage drop on the reference resistor in the corresponding branch changes from zero to a finite value. The signals from all of the resistors are simultaneously and continuously recorded. The droplet motion profile is thus obtained.
Fig. 13. Digital values are obtained by imposing a voltage threshold to the analog signals (0.3 mV). The raising edge of the digital signal indicates the moment that the droplet advancing edge starts to contact with the corresponding electrode. The falling edge indicates the moment that the receding edge of the droplet detaches the corresponding electrode.
Fig. 11. Illustration of the electrical connections for acquiring the droplet motion profile. The ionic liquid droplet behaves as an electrical switch connecting or disconnecting the circuit branch formed by adjacent electrodes and the reference resistor. A multi-channel DAQ system simultaneously records the voltage signals from all the channels. The signals are used to indicate the droplet position.
20 l droplet sits on top of two electrodes (#2 and #3). The electrical impedance between the saline-filled electrodes is orders lower than the air-filled ones. The position of the saline droplet is determined by scanning the electrical impedances of all of the electrode pairs (Fig. 9(b)).
38
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
motion sensor with a stimulating current of 300 A at 100 Hz and an electrical resistor of 1 k, the true power consumption is estimated as about 90 W according to Joule’s law. This value is much smaller than the typical power consumption of a commercial solid-state accelerometer (e.g. Analog Devices one-axis accelerometer, ADXL 105, power consumption 13 mW, operation voltage 6 V [21]). Moreover, the exclusion of free-standing structures lowers fabrication complexity and cost, which is expected to yield a lower per-unit cost. Furthermore, two-dimensional acceleration detection can be easily implemented using a liquid-state sensor by patterning bi-layered microelectrodes with minimal additional cost and fabrication complexity. Motion detection along one axis does not interfere with the other, which is vastly different from a two-dimensional solid-state sensor where cross-axis sensitivity is a major design concern [22]. 4.2. The effect of droplet size
Fig. 14. The comparison between the linear accelerations derived from the lumped model and the actual values of the given externally applied acceleration yields a high coefficient of determination (R2 = 0.982).
4. Discussion 4.1. Comparison with conventional solid-state accelerometers In order to have a comprehensive understanding of the characteristics of liquid-state motion sensing, a comparison of conventional solid-state accelerometers and the liquid-state motion sensing system is given in Table 1. One can see that although both of the systems detect linear accelerations based on inertial effects, they have different configurations due to different proof mass materials. In a solid-state accelerometer, the inertial force induces the displacement (or change of strain, or resonant frequency shift) of a free-standing solid proof mass. The measurement is in the form of analog signal. In a liquid-state sensor, the inertial force induces the relative motion of the liquid proof mass (ionic droplet) with respect to the substrate. The externally applied acceleration is determined from the droplet motion profile. Since the measurement is converted into the time domain, digital output is implemented. Because the sensitivity depends on the surface hydrophobicity, the sampling rate and the electrode configuration, the sensitivity and the measurement range are decoupled. Therefore, the beauty of digital measurement is that it enables the development of an inertial sensor with a wide measurement range and a high sensitivity. In addition, one can reduce power consumption by reducing the magnitude of external stimuli without sacrificing the sensing performance. For example, in a liquid-state
In the liquid-state motion sensor, the size of the liquid droplet is a critical parameter in determining sensing performance. One can see from Eq. (1) that the inertial force that drives the droplet motion increases with the droplet size. It is easier for a larger droplet to overcome the surface-related capillary force and move. Accordingly, a smaller threshold can be obtained by increasing the droplet size. On the other hand, according to the scaling law, the inertial body force becomes more predominant than the surface tension as the droplet size increases. Once the droplet size exceeds a critical value, the surface tension that tends to hold the droplet intact becomes smaller than the inertial body force that tends to pull the droplet apart (e.g. the inertial force due to vibration, gravity or mechanical shock). As a result, a large droplet may break apart more easily and ruin the measurement. Therefore, a critical droplet size must be determined not only to attain a satisfactory threshold, but also to avoid droplet splitting. The critical splitting size has been determined from the equilibrium between the surface tension and the driving body force [23]. In this work, the droplet size is experimentally determined as 20 l. With the help of appropriate surface treatment, the measurement threshold is determined as lower than 0.017 g (as elaborated in Section 2.2). For droplets of 20 l or smaller, no splitting is observed during the entire measurement under the stimulating voltage of 3 V at 100 Hz. 4.3. Effect of electrowetting and dielectrophoresis (DEP) The electrical stimulating signals used for detecting droplet motion may induce electrowetting force and DEP force. Both may affect droplet motion characteristics and sensing performance, and therefore need to be investigated. According to a previous study [24], the minimal electrowetting voltage that can be used to actuate
Table 1 Comparison of solid-state accelerometers and the liquid-state motion sensing system.
Proof mass Fabrication and configuration Measurement threshold Cross-axis sensitivity Frequency response Power consumption Sensitivity Measurement range Output a b
Solid-state accelerometer
Liquid-state motion sensing system
Solid A free-standing structure is needed Susceptible to fabrication imperfection [7] Depends on sensitivity, homogeneity, and linearity of the sensing material [8] A finite value, depending on the configuration of the free-standing structure [22] High response around the resonant frequency 13mWa Coupled due to the sensing materials [9,10]
Ionic liquid droplet No free-standing structures Less susceptible to fabrication imperfection Depends on surface hydrophobicity Zero Response decreases with frequency 90 Wb Not coupled.
Analog signal
Digital signal
Analog Devices® , ADXL 105, operation voltage 6 V. True power of the prototype in this work estimated by Joule’s law.
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
a 0.8 l droplet is about 30 V. There is no substantial electrowetting force applied on the droplet if the voltage is lower than the threshold. In our experiment, the electrical stimulating voltage (3 V) is much smaller than the reported threshold. Meanwhile, the volume of the droplet in this work is 25 times larger than the reported value. Therefore, we assume the influence of electrowetting can be negligible. DEP force for droplet manipulation is largely dependent on the frequency of the electrical stimulating signal [25–28]. Previous studies show that the typical operation frequency for droplet manipulation is larger than 1000 Hz [26]. Due to the relatively low frequency used in this work, the DEP force is assumed to be negligible. The above assumptions is validated by experiments that no change of contact angle or droplet motion are observed while applying electrical stimulating signals (3 V, 100 Hz) to a 20 l droplet. 4.4. Droplet evaporation, packaging issue, and continuous measurement It should be noted that in this work, motion detection is performed while placing the liquid droplet in an open environment. Since the motion detection is a one-time measurement (with no restoring force) under a constant external acceleration, and the one-time measurement lasts only for a very short period (i.e. on the order of tens of seconds), the droplet evaporation does not significantly affect the measurement. Nonetheless, droplet evaporation must be considered in long-term operations, where it could change the droplet size and consequently affect measurement threshold and sensitivity. Impermeable materials, such as polyurethane, glass, silicon, and so forth, will be explored to develop the hermetic package to minimize the droplet evaporation. In addition to carefully controlling the vapor pressure inside the chamber to minimize the droplet evaporation without increasing the damping effect, continuous measurement also can be implemented by patterning the array of microelectrodes on a curved surface, where the gravity component along the substrate surface serves as a restoring force. This work currently is underway. 5. Conclusion With the advantages of simple fabrication, low power consumption and digital signal processing, a fully implementable liquid-state motion sensing system has been designed and tested. We demonstrate that linear acceleration is determined from the relative motion of an ionic liquid droplet with respect to a solid substrate, which results in a small measurement threshold lower than 0.017 g after the surface is treated superhydrophobic through CF4 –H2 –He plasma. Through frequency analysis, it is revealed that the liquid-state sensor has a more sensitive response in the low frequency regime, which is distinct from the relatively high resonant frequencies of conventional solid-state accelerometers. Furthermore, digital output in the time domain decouples the sensitivity and measurement range, which addresses one major concern in the design of conventional solid-state accelerometers. In this work, the sensing capability of the liquid-state sensing system is demonstrated by measuring externally applied accelerations ranging from 5.62 to 7.51 m/s2 , which result in a good match between predicted and actual applied acceleration. A comparison of the characteristics of the liquid-state sensor illuminates the key parameters for determining sensing performance, including droplet size, droplet evaporation, and packaging considerations. This work is expected to provide a promising starting point for developing liquid-state inertial sensors complementary to their solid-state counterparts.
39
Acknowledgements The authors acknowledge the Center for Emergent Materials at the Ohio State University, an NSF MRSEC (Award Number DMR-0820414), for providing partial funding for this research. The authors thank Dr. Ronald Xiaorong Xu, his laboratory members, and Mrs. Melanie Senitko at the Ohio State University for their generous support. References [1] H. Zeng, Y. Zhao, Design and implementation of liquid droplet based motion sensing, in: Proceedings of the 15th International Conference on Solid-State Sensors and Actuators and Microsystems, Denver, CO, 2009, pp. 680–683. [2] N. Barbour, G. Schmidt, Inertial sensor technology trends, IEEE Sensor Journal 1 (2001) 332–339. [3] N. Yazdi, F. Ayazi, K. Najafi, Micromachined inertial sensors, Proceedings of the IEEE 86 (1998) 1640–1659. [4] H. Shusen, L. Xinxin, S. Zhaohui, W. Yuelin, Y. Heng, C. Lufeng, J. Jiwei, A high-performance micromachined piezoresistive accelerometer with axially stressed tiny beams, Journal of Micromechanics and Microengineering (2005) 993. [5] B.E. Boser, R.T. Howe, Surface micromachined accelerometers, IEEE Journal of Solid-State Circuits 31 (1996) 366–375. [6] D.W. Burns, R.D. Horning, W.R. Herb, J.D. Zook, H. Guckel, Sealed-cavity resonant microbeam accelerometer, Sensors and Actuators A: Physical 53 (1996) 249–255. [7] E.J. Eklund, A.M. Shkel, Single-mask fabrication of high-G piezoresistive accelerometers with extended temperature range, Journal of Micromechanics and Microengineering 17 (2007) 730–736. [8] N.K. Zayer, R. Greef, K. Rogers, A.J.C. Grellier, C.N. Pannell, In situ monitoring of sputtered zinc oxide films for piezoelectric transducers, Thin Solid Films 352 (1999) 179–184. [9] K.H.L. Chau, S.R. Lewis, Y. Zhao, S.F. Bart, R.G. Marcheselli, R.T. Howe, An integrated force-balanced capacitive accelerometer for low-g applications, Sensors and Actuators A: Physical 54 (1996) 472–476. [10] B.R. Davies, C.C. Barron, S. Montague, J.H. Smith, J.R. Murray, T.R. Christenson, V.I. Bateman, High-G MEMS integrated accelerometer, in: Smart Structures and Materials 1997: Smart Electronics and MEMS, San Diego, CA, USA, 1997, pp. 52–62. [11] T. Sun, L. Feng, X. Gao, L. Jiang, Bioinspired surfaces with special wettability, Accounts of Chemical Research 38 (2005) 644–652. [12] H. Ren, R.B. Fair, M.G. Pollack, E.J. Shaughnessy, Dynamics of electro-wetting droplet transport, Sensors and Actuators B: Chemical 87 (2002) 201–206. [13] O.N. Tretinnikov, Y. Ikada, Dynamic wetting and contact angle hysteresis of polymer surfaces studied with the modified Wilhelmy balance method, Langmuir 10 (2002) 1606–1614. [14] T.N. Krupenkin, J.A. Taylor, T.M. Schneider, S. Yang, From rolling ball to complete wetting: the dynamic tuning of liquids on nanostructured surfaces, Langmuir 20 (2004) 3824–3827. [15] H. Nakae, R. Inui, Y. Hirata, H. Saito, Effects of surface roughness on wettability, Acta Materialia 46 (1998) 2313–2318. [16] M. Mantel, J.P. Wightman, Influence of the surface chemistry on the wettability of stainless steel, Surface and Interface Analysis 21 (1994) 595–605. [17] H.M. Shang, Y. Wang, S.J. Limmer, T.P. Chou, K. Takahashi, G.Z. Cao, Optically transparent superhydrophobic silica-based films, Thin Solid Films 472 (2005) 37–43. [18] I.A. Larmour, S.E.J. Bell, G. Saunders, Remarkably simple fabrication of superhydrophobic surfaces using electroless galvanic deposition, Angewandte Chemie 119 (2007) 1740–1742. [19] S.H. Kim, J.-H. Kim, B.-K. Kang, H.S. Uhm, Superhydrophobic CFx coating via inline atmospheric RF plasma of He–CF4 –H2 , Langmuir 21 (2005) 12213–12217. [20] R.P. van Kampen, R.F. Wolffenbuttel, Modeling the mechanical behavior of bulkmicromachined silicon accelerometers, Sensors and Actuators A: Physical 64 (1998) 137–150. [21] H. Weinberg, Minimizing Power Consumption of iMEMS Accelerometers, Application note, Analog Devices, Inc., 2002. [22] H.L. Wong, M.D. Trifunac, Effects of cross-axis sensitivity and misalignment on the response of mechanical-optical accelerographs, Bulletin of the Seismological Society of America 67 (June) (1977) 929–956. [23] F. Bierbrauer, T.N. Phillips, The numerical prediction of droplet deformation and break-up using the Godunov marker-particle projection scheme, International Journal for Numerical Methods in Fluids 56 (2008) 1155–1160. [24] R. Bavière, J. Boutet, Y. Fouillet, Dynamics of droplet transport induced by electrowetting actuation, Microfluidics and Nanofluidics 4 (2008) 287–294. [25] A. Castellanos, A. Ramos, A. Gonzalez, N.G. Green, H. Morgan, Electrohydrodynamics and dielectrophoresis in microsystems: scaling laws, Journal of Physics D-Applied Physics 36 (2003) 2584–2597. [26] P.R. Gascoyne, J.V. Vykoukal, J.A. Schwartz, T.J. Anderson, D.M. Vykoukal, K.W. Current, C. McConaghy, F.F. Becker, C. Andrews, Dielectrophoresis-based programmable fluidic processors, Lab Chip 4 (Aug 2004) 299–309. [27] R. Ahmed, T.B. Jones, Optimized liquid DEP droplet dispensing, Journal of Micromechanics and Microengineering (2007) 1052.
40
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
[28] K.L. Wang, T.B. Jones, A. Raisanen, Dynamic control of DEP actuation and droplet dispensing, Journal of Micromechanics and Microengineering 17 (2007) 76–80.
Biographies Hansong Zeng received his B.S. and M.S. from Xi’an Jiaotong University. He is currently a Ph.D. candidate in Department of Biomedical Engineering at the Ohio State University. His research focuses on addressing microfluid and nanofluid dynamics involved in miniaturized systems for biosensing and actuation.
Yi Zhao is currently an assistant professor at Department of Biomedical Engineering at the Ohio State University and leads the Laboratory for Biomedical Microsystems. His research interests include development of miniaturized systems for point-of-care diagnostics and treatment, and lab-on-chip systems for exploring the interaction between living organisms and engineering environment. He is also interested in investigating critical mechanical, material and innovative fabrication issues in micro/nanosystems. He has published 50+ papers in journal and peer-reviewed conference proceedings. His research is currently funded by National Institute of Health, NSF MRSEC Center for Emergent Materials, and a number of internal research units.
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Liquid-state motion sensing夽 Hansong Zeng, Yi Zhao ∗ Laboratory for Biomedical Microsystems, Department of Biomedical Engineering, The Ohio State University, 294 Bevis Hall, 1080 Carmack Road, Columbus, OH 43210, USA
a r t i c l e
i n f o
Article history: Available online 6 December 2009 Keywords: Motion detection Droplet dynamics Frequency response Superhydrophobicity
a b s t r a c t This paper demonstrates a liquid droplet-based motion sensing system which has the advantages of simple fabrication, low power consumption and digital signal processing. The sensor consists of a dielectric substrate patterned with an array of microelectrodes, and a saline droplet as the proof mass. Once an external linear acceleration is applied, the inertial force moves the droplet on the micropatterned substrate. The acceleration is determined from the movement profile detected by the microelectrodes. In order to enhance the threshold and the sensitivity of motion sensing, two surface treatment approaches are utilized to create superhydrophobic surfaces. The result shows that the minimal sliding angle that can move a 20 l droplet on the superhydrophobic surface is lower than 1◦ , corresponding to a threshold of lower than 0.017 g. A lumped-parameter model is developed to estimate the dynamic behavior of the proposed system. The result shows that the frequency response of the droplet-based sensor is more significant at low frequencies than at high frequencies, which is distinct from solid-state accelerometers. Measurement under a constant acceleration shows that the predicted value derived from the measurement has a good match with the actual applied acceleration, validating the proposed system as a viable alternative for motion sensing. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Detection of motion parameters, such as linear acceleration and angular change, is critical for a dynamic system. For example, motion detection is a key element of industrial applications including inertial navigation, passive automotive safety systems and simulation of space microgravity [2,3]. Increasingly smaller motion sensors have generated a great deal of interest among the booming gaming industry, where sensors serve as the interface between the gaming systems and the player. There long has been a practice of determining motion parameters using a solid-state MEMS sensor (i.e., accelerometer or gyroscope) that has high precision, low cost and small size [4]. In such a sensor, the motion change induces change of mechanical strain, displacement, or resonant frequency shift of the solid-state proof mass. A piezoresistive, piezoelectric or capacitive sensing component converts such change into electrical signals that can be recognized and processed [5,6]. Although effective, solid-state motion sensors are limited to a certain extent by complex fabrication and packaging processes. In these sensors, free-standing microstructures usually serve as the
夽 This paper was presented at the 15th International Conference on Solid-State Sensors and Actuators and Microsystems held in Denver, CO, June 21–25, 2009, and is an expansion of the abstract as printed in the Technical Digest of this meeting [1]. ∗ Corresponding author. Tel.: +1 614 247 7424; fax: +1 614 292 7301. E-mail address: [email protected] (Y. Zhao). 0925-4005/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2009.11.069
proof mass, and are fabricated by surface micromachining (patterning and releasing). Since the sensing performance is highly sensitive to fabrication imperfection [7], the tools for fabricating, characterizing and calibrating these free-standing structures must have high precision. Moreover, extraordinary attention is paid in the packaging process, not only to reduce damping effects but also to protect the mechanically fragile, free-standing structures. Meanwhile, in sensors using piezo-materials, the sensing performance is largely determined by the homogeneity, sensitivity and linearity of the piezo-sensing materials, which all vary with deposition conditions [8]. In addition, the sensitivity and the measurement range of a solid-state motion sensor often are intrinsically coupled, which makes it difficult to develop a sensor with a very high-level of sensitivity and a large measurement range, simultaneously [9,10]. In this work, we demonstrate a new concept of motion sensing that utilizes a liquid droplet as the inertial proof mass. Externally applied acceleration induces an inertial force on the droplet and moves it over the substrate. The droplet movement is detected by the time sequence of the electrical impedance measurement using microelectrodes patterned on the substrate. The externally applied acceleration is derived from droplet motion characteristics based on a droplet dynamics model. Unlike solid-state counterparts, a liquid-state motion sensing system requires minimal configuration and fabrication complexity. Meanwhile, the acquisition of digital signals solely depends on the electrical impedance between the measuring electrodes, which are not affected by the magnitude of external acceleration. Therefore, the sensitivity of the measurement is decoupled from the measurement range. Digital signal
34
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
Fig. 2. Illustration of the droplet dynamics during motion. The driving force of droplet motion is induced by the external applied acceleration (a with the red arrow) and is proportional to the mass of the liquid droplet. The motion of the droplet is governed by the driving force, the capillary force which is induced by dynamic contact angle hysteresis between the advancing contact angle ( A ) and receding contact angle ( R ), the air damping and the contact-line friction. The relative motion between the droplet and the substrate occurs once the capillary force is overcomed. For analysis simplicity, it is assumed that the damping/friction force is linearly related to the traveling velocity of the droplet. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Fig. 1. The concept of liquid droplet-based motion sensing. (a) The work is inspired by observing a drop of dew moving on the hydrophobic surface of a lotus leaf. (b) Schematic configuration of a liquid-state motion sensor.
processing also makes the sensor less vulnerable to external electric/magnetic disturbance. This is especially important for slow motion detection, e.g. body movement. 2. Design and analysis 2.1. Concept of liquid-based motion sensing The concept of liquid-based motion sensing is inspired by a natural phenomenon shown in Fig. 1(a), where a drop of dew on the hydrophobic surface of a lotus leaf is susceptible to a tiny external perturbation [11]. The relative movement of the droplet with respect to the leaf can be used as a measure of the external perturbation. Based on this observation, an engineered liquid droplet-based motion sensing system is designed, which consists of a hydrophobic substrate patterned with an array of microelectrodes (Fig. 1(b)). Like the dew rolling over the lotus leaf, the motion of the ionic droplet depends on the externally applied acceleration and surface hydrophobicity of the substrate. As it moves over the surface, the conductive ionic droplet changes the electrical connection states of the microelectrodes, thus allowing the motion characteristics of the droplet to be determined. In order to demonstrate the concept of liquid-state motion sensors, a simplified one-dimension prototype with one-time sensing capability is developed in this work. An array of microelectrodes is patterned on a planar dielectric substrate to detect the motion profile of the liquid droplet. Dual-axis and continuous sensing can be implemented by patterning an array of bi-layered microelectrodes on a curved substrate. 2.2. Lumped model for droplet dynamic study A lumped-parameter model is developed to quantitatively understand the droplet dynamics that govern the sensing perfor-
mance (Fig. 2). As seen from a coordinate system attached to the substrate, the driving force is in the opposite direction of the externally applied acceleration and has a magnitude of m × a, where m is the mass of the droplet and a is the acceleration to be determined [12]. In order to move the droplet, the driving force must overcome the capillary force caused by the dynamic contact angle hysteresis that resists the droplet motion [13]. Once the droplet moves on the surface, it is subjected to a damping/friction force which includes air damping as well as contact-line friction [14]. To simplify the analysis without significantly affecting the accuracy, it is assumed that the damping/friction force linearly relates to the droplet velocity. The governing equation of the droplet can be expressed as: m
dx d2 x − (m × a − Fthreshold ) = 0 +b dt dt 2
(1)
where x is the relative displacement, b is the damping/friction coefficient and Fthreshold is the capillary force. For a 20 l ionic droplet (0.9% saline) used in this work, m = 20 mg. The threshold (Fthreshold ) in Eq. (1) is an important parameter that determines the minimal acceleration that can be detected by the sensing system. In order to obtain a small dynamic contact angle hysteresis, a low surface wettability is required. This is usually performed by changing the surface topography, surface chemistry or both [15–17]. In this work, surface hydrophobicity is regulated using two independent approaches. Fig. 3(a) shows the first approach in which a layer of silica nanoparticles (average diameter 7–10 nm) is spin-coated on the surface (at 1000 rpm for 10 s), followed by immersion into an aqueous HDFT solution (3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-1decanethiol, Sigma–Aldrich® , CA, USA). The silica nanoparticles create nano-roughened surface topography, while the HDFT solution assembles a molecular layer of fluorocompound on the surface to further increase surface hydrophobicity [18]. After the treatment, the sliding angle of a 20 l (about 20 mg in weight) saline droplet is measured around 4◦ , corresponding to a threshold of 0.07 g (g being the gravity). Fig. 3(b) shows the second approach for regulating the surface hydrophobicity, in which the surface is treated with atmospheric plasma of CF4 –H2 –He. Both fluorocompound deposition and surface roughening are involved in the process [19]. Using this approach, the sliding angle of a 20 l saline droplet is below 1◦ , corresponding to a threshold lower than 0.017 g. Following the threshold determination, the damping/friction coefficient is estimated by optical motion analysis. Fig. 4(a) illustrates the experimental setup for analyzing the droplet motion characteristics. A precise tilting stage is used to apply a constant acceleration to the saline droplet. The motion of the droplet is recorded by a high-speed digital camera at a frame rate of 40 frames
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
35
Fig. 3. Two approaches are used to modulate surface hydrophobicity. (a) In the first approach, silica nanoparticles are coated on the substrate, which is followed by immersing the surface in HDFT solution. The surface is first roughened by silica nanoparticles, and subsequently decorated by a self-assembled molecular layer of fluorocompound which reduces the surface energy; (b) in the second approach, the surface is treated by CF4 –H2 –He plasma for both surface roughening and fluorocompound deposition; (c) static contact angle measurement of the surfaces treated by the two approaches. The samples treated by the first approach have a static saline contact angle of 153◦ , a dynamic contact angle hysteresis of 6◦ , and a sliding angle of 4◦ . The samples treated by the second approach have a static saline contact angle of 162◦ , a dynamic contact angle hysteresis of 3◦ , and a sliding angle below 1◦ . All the measurements are averaged from 3 samples.
per second (fps). Fig. 4(b) is a series of snapshots showing the droplet movement under a constant acceleration of 0.573 g, corresponding to a tilting angle of 35◦ . The position of the droplet at specific time points is determined from each snapshot. The motion profile of the droplet (displacement–time curve) during a certain period of time can be plotted (Fig. 5), where typical motion parameters including displacement, velocity, and acceleration can be derived. It is clearly seen that in the initial phase of the movement, the traveling velocity of the liquid droplet increases. As a result, the damping/friction force increases. Once the damping/friction force equals the driving force, the traveling velocity reaches a critical value and the droplet starts to move at a constant velocity. This is indicated by a linear displacement–time curve. When the balance is reached, the damping/friction coefficient can be determined based on Eq. (1). For a 20 l saline droplet rolling over a substrate treated by the second surface treatment approach (CF4 –H2 –He plasma), the damping/friction coefficient is determined as 6.73 N (m/s)−1 . Frequency response of droplet-based motion sensing is also analyzed using a lumped-parameter model. Fig. 6 compares the equivalent circuits of the droplet-based motion sensing system and a typical solid-state accelerometer [20]. The threshold is represented by a power source whose polarization is opposite to the source denoting the driving force. Since there is no restoring force in a liquid-state sensing system with a planar surface, the capacitor is absent. As a result, the liquid-state system does not have a resonant frequency. The displacement is infinite under a constant acceleration and reduces as the frequency increases. The frequency response is obtained numerically using Matlab Simulink software. The dimension-less values are plotted against the frequency (Fig. 6(c) and (d)).
Fig. 4. Optical determination of the droplet motion profile. (a) The experiment setup is composed of a tilting stage that generates a constant external acceleration, and a digital camera that records the motion of the droplet; (b) the time sequence images show the motion of the droplet during 1.5 s.
Fig. 5. The damping/friction coefficient is determined from the motion profile of the droplet. In the initial phase, the velocity of the droplet increases. The damping/friction force increases accordingly. Once the driving force and the damping/friction force become the same, the droplet moves at a constant velocity. The damping/friction coefficient can thus be determined.
3. Experiment and results 3.1. Fabrication An array of Au/Ti microelectrodes is patterned on a planar glass substrate using standard photolithography and lift-off processes
36
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
Fig. 8. Electrical impedance measurement from a pair of adjacent microelectrodes while the inter-electrode space is filled with a saline droplet or with air. At a stimulating frequency of 100 Hz, the ratio of the electrical impedance between the microelectrode pairs with air-filled inter-electrode space and with liquid-filled inter-electrode space are about 1000. Such a difference is sufficient for digital detection the presence of a saline droplet. Fig. 6. Comparison of the frequency responses of a liquid-state sensor and of a solidstate accelerometer. The equivalent circuits of the two systems are illustrated in (a) and (b). (c) Since there is no capacitor, the droplet-based system has high response in the low frequency regime. (d) In a solid-state system, high response is around the resonant frequency.
(Fig. 7(a)). In this work, the width of the electrodes is 200 m, and the separation of neighboring electrodes is 1 mm. After patterning, the entire surface of the device is treated with CF4 -H2 -He plasma as described in Section 2. A prototype is shown in Fig. 7(b). 3.2. Impedance analysis Electrical impedance between the adjacent electrodes is measured using an electrical impedance analyzer (Solartron® 1260A). A stimulating AC current of 300 A is applied between the adjacent electrodes without damaging the surface layer. Electrical impedance is obtained as a function of the stimulating frequency (Fig. 8). It is seen that the presence of the saline droplet can be easily identified from the magnitude of the electrical impedance. For example, the impedance magnitude of an air-filled electrode pair at 100 Hz is about 1000 times greater than that of a saline dropletfilled electrode pair. Such a large difference is sufficient for digital detection. Fig. 9(a) shows a representative measurement while a
Fig. 9. Static measurement of the droplet position. (a) The magnitude of the electrical impedance of each adjacent pair of microelectrodes is recorded, which indicates that a liquid droplet is present between the second and third electrodes and (b) the measurement is converted to digital values by applying an impedance threshold (1 × 105 ).
Fig. 7. Prototype fabrication. (a) Surface micromachining is used to create the micropatterned substrate. The superhydrophobic surface is obtained by CF4 –H2 –He plasma treatment. (b) shows a prototype with a saline droplet on top of the substrate. In a typical prototype for the measurement, each microelectrode is 200 m in width. The separation between adjacent microelectrodes is 1 mm.
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
37
Fig. 12. The time sequences of the electrical voltages in all the channels are recorded during the droplet motion. Fig. 10. Experiment setup for acceleration measurement. A constant acceleration is generated by titling the substrate at a certain angle with respect to the horizontal plane.
3.3. Acceleration measurement
Fig. 12 shows the analog signals obtained by the DAQ system at the tilting angle of 35◦ , which corresponds to a constant acceleration of 5.62 m/s2 (0.57 g) along the moving direction of the droplet. The figure shows the time sequence of voltage variations in all five adjacent electrode pairs. In order to demonstrate digital data processing, the analog voltages are converted to digital signals (Fig. 13). In this measurement, a voltage higher than 0.3 mV is considered to be high-level, indicating the presence of a saline droplet. Each rising edge of the voltage curve indicates the moment when the advancing edge of the saline droplet contacts the electrodes. The falling edge of the voltage curve indicates the moment when the receding edge of the saline droplet leaves the electrodes. The position of the droplet as a function of time is therefore obtained. The first order derivative (dx/dt) and the second order derivative (d2 x/dt2 ) of the droplet motion profile correspond to the velocity and acceleration, respectively. By inserting the droplet velocity and droplet acceleration into Eq. (1), the externally applied acceleration that induces relative droplet motion is determined. In order to validate the lumped model, the acceleration measurements are performed at four different externally applied accelerations, namely 5.62, 6.23, 6.92 and 7.51 m/s2 , which correspond to the tilting angles of 35◦ , 40◦ , 45◦ and 50◦ , respectively. The comparison between the values predicted by the lumpedparameter model and the external acceleration actually applied (Fig. 14) shows that the liquid-state sensor can detect linear accelerations with high precision, as evidenced by a high coefficient of determination of 0.982.
In order to validate the utility of the liquid-state system for motion detection, the system is operated under constant accelerations generated by an experimental setup shown in Fig. 10. The setup consists of a tilting stage, which generates a constant external acceleration to the liquid droplet by tilting the planar substrate at a certain angle with respect to the horizontal plane. During the experiments, the electrical signals acquired from the prototypes are recorded through a multi-channel data acquisition (DAQ) interface with the circuit schematized in Fig. 11. An AC power supply (3 V, 100 Hz) is used to provide the electrical stimuli. A reference resistor (1 k) is connected to each adjacent pair of microelectrodes to form a testing branch. The saline droplet behaves as an electrical switch, which either connects or disconnects the adjacent electrodes while it rolls over the planar substrate. Consequently, the voltage drop on the reference resistor in the corresponding branch changes from zero to a finite value. The signals from all of the resistors are simultaneously and continuously recorded. The droplet motion profile is thus obtained.
Fig. 13. Digital values are obtained by imposing a voltage threshold to the analog signals (0.3 mV). The raising edge of the digital signal indicates the moment that the droplet advancing edge starts to contact with the corresponding electrode. The falling edge indicates the moment that the receding edge of the droplet detaches the corresponding electrode.
Fig. 11. Illustration of the electrical connections for acquiring the droplet motion profile. The ionic liquid droplet behaves as an electrical switch connecting or disconnecting the circuit branch formed by adjacent electrodes and the reference resistor. A multi-channel DAQ system simultaneously records the voltage signals from all the channels. The signals are used to indicate the droplet position.
20 l droplet sits on top of two electrodes (#2 and #3). The electrical impedance between the saline-filled electrodes is orders lower than the air-filled ones. The position of the saline droplet is determined by scanning the electrical impedances of all of the electrode pairs (Fig. 9(b)).
38
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
motion sensor with a stimulating current of 300 A at 100 Hz and an electrical resistor of 1 k, the true power consumption is estimated as about 90 W according to Joule’s law. This value is much smaller than the typical power consumption of a commercial solid-state accelerometer (e.g. Analog Devices one-axis accelerometer, ADXL 105, power consumption 13 mW, operation voltage 6 V [21]). Moreover, the exclusion of free-standing structures lowers fabrication complexity and cost, which is expected to yield a lower per-unit cost. Furthermore, two-dimensional acceleration detection can be easily implemented using a liquid-state sensor by patterning bi-layered microelectrodes with minimal additional cost and fabrication complexity. Motion detection along one axis does not interfere with the other, which is vastly different from a two-dimensional solid-state sensor where cross-axis sensitivity is a major design concern [22]. 4.2. The effect of droplet size
Fig. 14. The comparison between the linear accelerations derived from the lumped model and the actual values of the given externally applied acceleration yields a high coefficient of determination (R2 = 0.982).
4. Discussion 4.1. Comparison with conventional solid-state accelerometers In order to have a comprehensive understanding of the characteristics of liquid-state motion sensing, a comparison of conventional solid-state accelerometers and the liquid-state motion sensing system is given in Table 1. One can see that although both of the systems detect linear accelerations based on inertial effects, they have different configurations due to different proof mass materials. In a solid-state accelerometer, the inertial force induces the displacement (or change of strain, or resonant frequency shift) of a free-standing solid proof mass. The measurement is in the form of analog signal. In a liquid-state sensor, the inertial force induces the relative motion of the liquid proof mass (ionic droplet) with respect to the substrate. The externally applied acceleration is determined from the droplet motion profile. Since the measurement is converted into the time domain, digital output is implemented. Because the sensitivity depends on the surface hydrophobicity, the sampling rate and the electrode configuration, the sensitivity and the measurement range are decoupled. Therefore, the beauty of digital measurement is that it enables the development of an inertial sensor with a wide measurement range and a high sensitivity. In addition, one can reduce power consumption by reducing the magnitude of external stimuli without sacrificing the sensing performance. For example, in a liquid-state
In the liquid-state motion sensor, the size of the liquid droplet is a critical parameter in determining sensing performance. One can see from Eq. (1) that the inertial force that drives the droplet motion increases with the droplet size. It is easier for a larger droplet to overcome the surface-related capillary force and move. Accordingly, a smaller threshold can be obtained by increasing the droplet size. On the other hand, according to the scaling law, the inertial body force becomes more predominant than the surface tension as the droplet size increases. Once the droplet size exceeds a critical value, the surface tension that tends to hold the droplet intact becomes smaller than the inertial body force that tends to pull the droplet apart (e.g. the inertial force due to vibration, gravity or mechanical shock). As a result, a large droplet may break apart more easily and ruin the measurement. Therefore, a critical droplet size must be determined not only to attain a satisfactory threshold, but also to avoid droplet splitting. The critical splitting size has been determined from the equilibrium between the surface tension and the driving body force [23]. In this work, the droplet size is experimentally determined as 20 l. With the help of appropriate surface treatment, the measurement threshold is determined as lower than 0.017 g (as elaborated in Section 2.2). For droplets of 20 l or smaller, no splitting is observed during the entire measurement under the stimulating voltage of 3 V at 100 Hz. 4.3. Effect of electrowetting and dielectrophoresis (DEP) The electrical stimulating signals used for detecting droplet motion may induce electrowetting force and DEP force. Both may affect droplet motion characteristics and sensing performance, and therefore need to be investigated. According to a previous study [24], the minimal electrowetting voltage that can be used to actuate
Table 1 Comparison of solid-state accelerometers and the liquid-state motion sensing system.
Proof mass Fabrication and configuration Measurement threshold Cross-axis sensitivity Frequency response Power consumption Sensitivity Measurement range Output a b
Solid-state accelerometer
Liquid-state motion sensing system
Solid A free-standing structure is needed Susceptible to fabrication imperfection [7] Depends on sensitivity, homogeneity, and linearity of the sensing material [8] A finite value, depending on the configuration of the free-standing structure [22] High response around the resonant frequency 13mWa Coupled due to the sensing materials [9,10]
Ionic liquid droplet No free-standing structures Less susceptible to fabrication imperfection Depends on surface hydrophobicity Zero Response decreases with frequency 90 Wb Not coupled.
Analog signal
Digital signal
Analog Devices® , ADXL 105, operation voltage 6 V. True power of the prototype in this work estimated by Joule’s law.
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
a 0.8 l droplet is about 30 V. There is no substantial electrowetting force applied on the droplet if the voltage is lower than the threshold. In our experiment, the electrical stimulating voltage (3 V) is much smaller than the reported threshold. Meanwhile, the volume of the droplet in this work is 25 times larger than the reported value. Therefore, we assume the influence of electrowetting can be negligible. DEP force for droplet manipulation is largely dependent on the frequency of the electrical stimulating signal [25–28]. Previous studies show that the typical operation frequency for droplet manipulation is larger than 1000 Hz [26]. Due to the relatively low frequency used in this work, the DEP force is assumed to be negligible. The above assumptions is validated by experiments that no change of contact angle or droplet motion are observed while applying electrical stimulating signals (3 V, 100 Hz) to a 20 l droplet. 4.4. Droplet evaporation, packaging issue, and continuous measurement It should be noted that in this work, motion detection is performed while placing the liquid droplet in an open environment. Since the motion detection is a one-time measurement (with no restoring force) under a constant external acceleration, and the one-time measurement lasts only for a very short period (i.e. on the order of tens of seconds), the droplet evaporation does not significantly affect the measurement. Nonetheless, droplet evaporation must be considered in long-term operations, where it could change the droplet size and consequently affect measurement threshold and sensitivity. Impermeable materials, such as polyurethane, glass, silicon, and so forth, will be explored to develop the hermetic package to minimize the droplet evaporation. In addition to carefully controlling the vapor pressure inside the chamber to minimize the droplet evaporation without increasing the damping effect, continuous measurement also can be implemented by patterning the array of microelectrodes on a curved surface, where the gravity component along the substrate surface serves as a restoring force. This work currently is underway. 5. Conclusion With the advantages of simple fabrication, low power consumption and digital signal processing, a fully implementable liquid-state motion sensing system has been designed and tested. We demonstrate that linear acceleration is determined from the relative motion of an ionic liquid droplet with respect to a solid substrate, which results in a small measurement threshold lower than 0.017 g after the surface is treated superhydrophobic through CF4 –H2 –He plasma. Through frequency analysis, it is revealed that the liquid-state sensor has a more sensitive response in the low frequency regime, which is distinct from the relatively high resonant frequencies of conventional solid-state accelerometers. Furthermore, digital output in the time domain decouples the sensitivity and measurement range, which addresses one major concern in the design of conventional solid-state accelerometers. In this work, the sensing capability of the liquid-state sensing system is demonstrated by measuring externally applied accelerations ranging from 5.62 to 7.51 m/s2 , which result in a good match between predicted and actual applied acceleration. A comparison of the characteristics of the liquid-state sensor illuminates the key parameters for determining sensing performance, including droplet size, droplet evaporation, and packaging considerations. This work is expected to provide a promising starting point for developing liquid-state inertial sensors complementary to their solid-state counterparts.
39
Acknowledgements The authors acknowledge the Center for Emergent Materials at the Ohio State University, an NSF MRSEC (Award Number DMR-0820414), for providing partial funding for this research. The authors thank Dr. Ronald Xiaorong Xu, his laboratory members, and Mrs. Melanie Senitko at the Ohio State University for their generous support. References [1] H. Zeng, Y. Zhao, Design and implementation of liquid droplet based motion sensing, in: Proceedings of the 15th International Conference on Solid-State Sensors and Actuators and Microsystems, Denver, CO, 2009, pp. 680–683. [2] N. Barbour, G. Schmidt, Inertial sensor technology trends, IEEE Sensor Journal 1 (2001) 332–339. [3] N. Yazdi, F. Ayazi, K. Najafi, Micromachined inertial sensors, Proceedings of the IEEE 86 (1998) 1640–1659. [4] H. Shusen, L. Xinxin, S. Zhaohui, W. Yuelin, Y. Heng, C. Lufeng, J. Jiwei, A high-performance micromachined piezoresistive accelerometer with axially stressed tiny beams, Journal of Micromechanics and Microengineering (2005) 993. [5] B.E. Boser, R.T. Howe, Surface micromachined accelerometers, IEEE Journal of Solid-State Circuits 31 (1996) 366–375. [6] D.W. Burns, R.D. Horning, W.R. Herb, J.D. Zook, H. Guckel, Sealed-cavity resonant microbeam accelerometer, Sensors and Actuators A: Physical 53 (1996) 249–255. [7] E.J. Eklund, A.M. Shkel, Single-mask fabrication of high-G piezoresistive accelerometers with extended temperature range, Journal of Micromechanics and Microengineering 17 (2007) 730–736. [8] N.K. Zayer, R. Greef, K. Rogers, A.J.C. Grellier, C.N. Pannell, In situ monitoring of sputtered zinc oxide films for piezoelectric transducers, Thin Solid Films 352 (1999) 179–184. [9] K.H.L. Chau, S.R. Lewis, Y. Zhao, S.F. Bart, R.G. Marcheselli, R.T. Howe, An integrated force-balanced capacitive accelerometer for low-g applications, Sensors and Actuators A: Physical 54 (1996) 472–476. [10] B.R. Davies, C.C. Barron, S. Montague, J.H. Smith, J.R. Murray, T.R. Christenson, V.I. Bateman, High-G MEMS integrated accelerometer, in: Smart Structures and Materials 1997: Smart Electronics and MEMS, San Diego, CA, USA, 1997, pp. 52–62. [11] T. Sun, L. Feng, X. Gao, L. Jiang, Bioinspired surfaces with special wettability, Accounts of Chemical Research 38 (2005) 644–652. [12] H. Ren, R.B. Fair, M.G. Pollack, E.J. Shaughnessy, Dynamics of electro-wetting droplet transport, Sensors and Actuators B: Chemical 87 (2002) 201–206. [13] O.N. Tretinnikov, Y. Ikada, Dynamic wetting and contact angle hysteresis of polymer surfaces studied with the modified Wilhelmy balance method, Langmuir 10 (2002) 1606–1614. [14] T.N. Krupenkin, J.A. Taylor, T.M. Schneider, S. Yang, From rolling ball to complete wetting: the dynamic tuning of liquids on nanostructured surfaces, Langmuir 20 (2004) 3824–3827. [15] H. Nakae, R. Inui, Y. Hirata, H. Saito, Effects of surface roughness on wettability, Acta Materialia 46 (1998) 2313–2318. [16] M. Mantel, J.P. Wightman, Influence of the surface chemistry on the wettability of stainless steel, Surface and Interface Analysis 21 (1994) 595–605. [17] H.M. Shang, Y. Wang, S.J. Limmer, T.P. Chou, K. Takahashi, G.Z. Cao, Optically transparent superhydrophobic silica-based films, Thin Solid Films 472 (2005) 37–43. [18] I.A. Larmour, S.E.J. Bell, G. Saunders, Remarkably simple fabrication of superhydrophobic surfaces using electroless galvanic deposition, Angewandte Chemie 119 (2007) 1740–1742. [19] S.H. Kim, J.-H. Kim, B.-K. Kang, H.S. Uhm, Superhydrophobic CFx coating via inline atmospheric RF plasma of He–CF4 –H2 , Langmuir 21 (2005) 12213–12217. [20] R.P. van Kampen, R.F. Wolffenbuttel, Modeling the mechanical behavior of bulkmicromachined silicon accelerometers, Sensors and Actuators A: Physical 64 (1998) 137–150. [21] H. Weinberg, Minimizing Power Consumption of iMEMS Accelerometers, Application note, Analog Devices, Inc., 2002. [22] H.L. Wong, M.D. Trifunac, Effects of cross-axis sensitivity and misalignment on the response of mechanical-optical accelerographs, Bulletin of the Seismological Society of America 67 (June) (1977) 929–956. [23] F. Bierbrauer, T.N. Phillips, The numerical prediction of droplet deformation and break-up using the Godunov marker-particle projection scheme, International Journal for Numerical Methods in Fluids 56 (2008) 1155–1160. [24] R. Bavière, J. Boutet, Y. Fouillet, Dynamics of droplet transport induced by electrowetting actuation, Microfluidics and Nanofluidics 4 (2008) 287–294. [25] A. Castellanos, A. Ramos, A. Gonzalez, N.G. Green, H. Morgan, Electrohydrodynamics and dielectrophoresis in microsystems: scaling laws, Journal of Physics D-Applied Physics 36 (2003) 2584–2597. [26] P.R. Gascoyne, J.V. Vykoukal, J.A. Schwartz, T.J. Anderson, D.M. Vykoukal, K.W. Current, C. McConaghy, F.F. Becker, C. Andrews, Dielectrophoresis-based programmable fluidic processors, Lab Chip 4 (Aug 2004) 299–309. [27] R. Ahmed, T.B. Jones, Optimized liquid DEP droplet dispensing, Journal of Micromechanics and Microengineering (2007) 1052.
40
H. Zeng, Y. Zhao / Sensors and Actuators B 154 (2011) 33–40
[28] K.L. Wang, T.B. Jones, A. Raisanen, Dynamic control of DEP actuation and droplet dispensing, Journal of Micromechanics and Microengineering 17 (2007) 76–80.
Biographies Hansong Zeng received his B.S. and M.S. from Xi’an Jiaotong University. He is currently a Ph.D. candidate in Department of Biomedical Engineering at the Ohio State University. His research focuses on addressing microfluid and nanofluid dynamics involved in miniaturized systems for biosensing and actuation.
Yi Zhao is currently an assistant professor at Department of Biomedical Engineering at the Ohio State University and leads the Laboratory for Biomedical Microsystems. His research interests include development of miniaturized systems for point-of-care diagnostics and treatment, and lab-on-chip systems for exploring the interaction between living organisms and engineering environment. He is also interested in investigating critical mechanical, material and innovative fabrication issues in micro/nanosystems. He has published 50+ papers in journal and peer-reviewed conference proceedings. His research is currently funded by National Institute of Health, NSF MRSEC Center for Emergent Materials, and a number of internal research units.