10-14 MEMS: Applications 50 µs 55 µs 60 µs 65 µs 70 µs 75 µs 80 µs 85 µs Satellite droplet Break-off of droplet 10V (50-85 µs) 100 µm FIGURE 10.17 Droplet ejection sequence of HP 51626A printhead, courtesy Tseng (1998c). Satellite droplets Main droplets FIGURE 10.18 A printed vertical line smeared by satellite droplets, courtesy Tseng (1998c). One way to eliminate puddle formation is to coat the chamber’s outer surface with a nonwetting material.(The inner surface of the chamber needs to remain hydrophilic for liquid refill.) However, even with this coating, there is still no guarantee that puddle will not form. More research is underway to fully understand the mechanism in the puddle formation process. © 2006 by Taylor & Francis Group, LLC Microdroplet Generators 10-15 100 µm 20 µs 30 µs 40 µs 50 µs 60 µs 80 µs 100 µs 120 µs No satellite droplet FIGURE 10.19 Droplet ejection without satellite droplets, courtesy Tseng (1998a). Nozzle Liquid puddle Droplet FIGURE 10.20 Liquid puddle formation outside the micronozzles, courtesy Tseng (1998d). © 2006 by Taylor & Francis Group, LLC 10.3.5 Material Issues Material issues including stress, erosion, durability, and compatibility are very complex problems in the design of microdroplet generators. In the area of processing, material compatibility, stress, and durability problems are commonly discussed. Material compatibility issues result from the processing temperature, processing environment (oxidation, reactive gas, etc.), etching method used, and adhesion ability; stress issues usually concern the processing temperature as well as the doping condition; material durability issues are due either to the material’s intrin- sicproperties or the mechanical forces induced during the process (i.e., fluid flow force, surface tension force, vacuum forces, or handling force). To eliminate the material issues, much care needs to be taken in designing the process flow, such as compensating for material stress during or after the fabrication process; performing high-temperature processes before introducing the low temperature material; finishing aggres- sive wet etching before metal film deposition; or using low temperature bonding material and processes to protect IC and microdevices. In selecting chamber materials, metals such as nickel and stainless steel have been widely employed for different microdroplet generators due to their ease of fabrication (such as by wet etching or electroplating), high mechanical strength, resistance to erosion by certain bases or acids, imperviousness to solvents, and high durability in cycling operations, etc. However, owing to cost and fabrication-precision considerations, the chamber materials of many commercial ink-jets have been changed to polymers, such as polyimide or PMMA in recent years. As a result, chamber formation by bonding a polymer thin plate on the actuator substrate and nozzle-array formation by laser drilling on those polymer thin plates are nowverycommon in the ink-jet industry. Nevertheless, the bonding and laser drilling processes may encounter issues such as bonding nonuniformity, time consuming serial laser drilling processes, and alignment limitations. Therefore, various polymer-MEMS processes for fabricating multiple embedded chambers integrated with nozzles have been developed; these not only simplify the fabrication process but also improve the accuracy of align- ment and nozzle fabrication on the chamber structures. Employing double exposures (one full power, and the other partial power) on a single SU-8 resist layer with antireflection coating in the resist–substrate inter- face, Chuang, et al. (2003) demonstrated that embedded microchannel structures can be fabricated easily, as shown in Figure 10.21.Thethickness of chamber roof can be controlled from 14 to 60 µm in a 2 µm resolution. Durability, stress, and erosion issues are the major problems concerning operation. Due to the cycling nature of the droplet generation process, the materials chosen for actuation face challenges not only from stress but also from fatigue. The HP Corporation reported that possible reasons for failure of the heater pas- sivation material are cavitation and thermal stress [Bhaskar et al., 1985]. Silicon, low-stress silicon nitride, sil- icon carbide, silicon dioxide, and some metals, are usually used to overcome these problems. In addition to selecting proper materials, reducing sharp corners in the design is an important key to eliminating stress concentration points and thus preventing material from cracking. The working fluid’s erosion of structural materials is another serious issue. Lee et al. (1999) reported the erosion of the spacer material in a com- mercial ink-jet head while using diesel fuel as working fluid. In contrast, materials including silicon and silicon nitride used by Tseng et al. (1998c, 1998d) and Lee et al. (1999) in the microinjector are free of this problem and can also be used with a wide variety of fluids including solvents and chemicals. Selecting materials wisely, ordering them correctly in the process, and properly designing the materials in the microstructures are the three primary measures for reducing material issues. 10.4 Fabrication of Micro-Droplet Generators The structures in microdroplet generators commonly include a manifold for storing liquid, microchannels for transporting liquid, microchambers for holding liquid,nozzles for defining droplet size and direction, and actuation mechanisms for generating droplets. Occasionally, droplet generators may not have nozzles but gen- erate droplets locally by energy focusing means, such as acoustic wave droplet generators [Zhu et al., 1996]. Before micromachining processes became widely used, most processes for fabricating microdroplet generators 10-16 MEMS: Applications © 2006 by Taylor & Francis Group, LLC followed the same general method: nozzle plates, fluid handling plates, and actuation plates are manufactured separately and then integrated into a single device. However, as the nozzle resolution becomes finer, bond- ing processes pose severe alignment, yield, and material problems as well as IC compatibility issues. On the other hand, the interconnection lines may not have enough space to fan out from each chamber when nozzle resolution is higher than 600 dpi. As a result, monolithic methods for fabricating high resolution IC integrated droplet generators have become very important. The following sections introduce examples of different fabrication techniques. 10.4.1 Multiple Pieces Figure 10.22 schematically shows the traditional method of fabricating microdroplet generators by bonding separately fabricated pieces [Tseng, 1998d]. In this process, actuation plates are fabricated separately from the nozzle plates. In the thermal bubble jet, heaters are usually sputtered or evaporated and then patterned with an IC circuit on the bottom plate; piezo, thermal buckling, electrostatic, and inertial actuators consist of more complex structures, such as piezo disks, thin plate structures, or cantilever beams. Nozzles are fabri- cated by electroforming [Ta et al., 1988], molding, or laser drilling [Keefe et al., 1997]. These separately processed pieces are assembled either by using intermediate layers of polymer spacer material [Siewell et al., 1985; Askeland et al., 1988; Hirata et al., 1996; Keefe et al., 1997] or directly adhering several pieces through anodic bonding [Kamisuki et al., 1998, 2000], fusion bonding [Gruhler et al., 1999], eutectic bond- ing, or low temperature chemical bonding. However, most of the bonding methods are chip-level rather than wafer-level processes and face challenges of alignment, bonding quality, and material–process compati- bility. As the nozzle resolution becomes higher than 600dpi, alignment accuracy approaching 4 µm (10% of the nozzle pitch) becomes difficult to attain. Higher alignment accuracy significantly increases the fabrication cost, especially for the chip-level process. Bonding quality is another important issue affecting the fabrication yield of large array and high-resolution devices. Additionally, the bonding materials (mostly polymers) Microdroplet Generators 10-17 Anti-reflection coating SU-8 Anti-reflection coating Substrate Standard UV exposure Partial UV exposure (a) (b) (c) (d) Stacked channel Channel release (e) (f) 46.5 µµm 23 µµm (g) FIGURE 10.21 (a)–(f) Fabrication process of polymer multiple embedded microchambers and (g) fabricated chambers, courtesy Chuang et al. (2003). © 2006 by Taylor & Francis Group, LLC chosen must be suitable for the application environments and working fluids. Finally, bonding processes involving heat, pressure, high voltage, or chemical situations restrict IC integration with the droplet gen- erators, and the IC integration is essential for large-array and high resolution applications. 10.4.2 Monolithic Fabrication To address the problems inherent in using multiple pieces, monolithic processes utilizing micromachin- ing technology have been widely employed since the early 1990s. Two primary monolithic methods have been introduced: one combines bulk and surface micromachining and the other uses bulk microma- chining and the deep UV lithography associated with electroforming (or UV lithography only). For example, Tseng et al. (1998d) combined surface and bulk micromachining to fabricate a micro- droplet generator array with potential nozzle resolution up to 1200 dpi (printing resolution can be 2400 dpi or higher). This design used double bulk micromachining processes to fabricate the fluid handling system, including the manifold, microchannels, and microchambers. Surface micromachining, on the other hand, was used for fabricating heaters, interconnection lines, and nozzles. The whole process was finished on (100) crystal orientation silicon wafers. Figures 10.23 and 10.24 show the three-dimensional structure of the microinjectors and the monolithic fabrication process respectively. The ejection of 0.9 pl droplets has also been demonstrated by Tseng et al. (2001b) using the high-resolution microinjectors. The structural materials used in the microinjector are silicon, silicon nitride, and silicon oxide, which are durable in high temperature and suitable for various liquids (even some harsh chemicals). Using this device, ICs can be easily integrated on the same silicon substrate. 10-18 MEMS: Applications a1: Metal mold fabrication a2: Sacrificial film pattern a3: Nozzle plate plating a4: Nozzle plate de-mold Nozzle plate and actuator plate bonding b1: Substrate passivation b2: Actuator fabrication b3: Spacer pattev FIGURE 10.22 Conventional fabrication process flow of microdroplet generators. © 2006 by Taylor & Francis Group, LLC The second primary method can be found in Lee’s (1999) work. This process used multiexposure and sig- nal development (MESD) lithography to define microchannel and microchamber structures (photoresist as sacrificial layer) and constructed the physical structures with electroformed metal. The manifold was manufactured from the wafer’s backside by electrochemical methods [Lee et al., 1995]. This device also demonstrated a capacity for very high-resolution arrays and compatibility with the IC process. Another method, using photoresist as sacrificial layer and polyimide as structure layer, was introduced by Chen et al. (1998) for high resolution and IC compatible applications. 10.5 Characterization of Droplet Generation Droplet trajectory, volume, ejection direction, and ejection sequence/velocity are four important quanti- tative measures of the ejection quality of microdroplet generators. The following sections briefly introduce the basic methods for testing droplet generation. 10.5.1 Droplet Trajectory Droplet trajectory can be visualized by directing a flashing light on the ejection stream, as shown in Figure 10.25 [Tseng et al., 1998a]. The white dots in Figure 10.26 show the visualized droplet stream. The visualized droplet trajectory follows an exponential curve that is very different from the parabolic curve expected for normal sized objects with a similar initial horizontal velocity. Tseng et al. (1998a) also esti- mated droplet trajectory by solving a set of ordinary differential equations from the balance of horizon- tal and vertical forces on a single droplet flying through air. From this analysis, the vertical position Y and horizontal position X of the droplet can be expressed by the following equations: Y ϭ U v ∞ ΄ t Ϫ ᎏ U g v ∞ ᎏ 1 Ϫ e ᎏ Ϫ6 m πµ r 0 ᎏ t ΅ (10.1) X ϭ ᎏ U 6 π H µ 0 m r 0 ᎏ 1 Ϫ e ᎏ Ϫ6 m πµ r 0 ᎏ t (10.2) where g is the acceleration due to gravity, t is the time, m is the mass and r 0 is the radius of the droplet, µ is the viscosity of air,U v ∞ ϭ ᎏ 6 m πµ g r 0 ᎏ is the droplet terminal velocity,and U H0 is the initial horizontal veloc- ity. The trajectory determined by the experiment is drawn in Figure 10.26 and fits the visualized tra- jectory well except at the end, suggesting the interaction among droplets. From this simple analysis, the maximum flying distance of a droplet with a known diameter can be estimated as: X max ϭ ϭ (U H0 r 2 0 ), (10.3) 2 ρ liquid ᎏ 9 µ air U H0 m ᎏ 6 πµ r 0 Microdroplet Generators 10-19 Liquid entrance Manifold Narrow heater Wide heater Nozzle Chamber Common line Electrode Liquid FIGURE 10.23 Schematic three-dimensional structure view of microinjectors, after Tseng (1998c). © 2006 by Taylor & Francis Group, LLC when t ϳ ∞. Here the maximum distance is proportional to the droplet velocity and droplet radius to the second power. For different droplet sizes with the same initial velocity, the maximum flying distance of smaller droplets decreases very fast. To obtain 1 mm flying distance, the droplet with 10 m/s initial velocity needs a minimum radius of 2.7 µm. From the above estimation, droplet size should be maintained above a certain value to ensure enough flying distance for printing. Printing with very fine droplets (diameter smaller than a couple of micrometers) requires either increasing the droplets’ initial velocity or printing in a special vacuum environment to overcome air drag. 10-20 MEMS: Applications (d) ( e ) Nozzle formation & pad open KOH etch manifold, PSG remove (b) Heater & interconnection formation KOH etch enlarge chamber depth (c) PSG & low-stress-nitride deposition (a) Si Top view Cross section A-A′ A A′ FIGURE 10.24 Fabrication process flow of monolithic microinjectors, after Tseng (1998c). © 2006 by Taylor & Francis Group, LLC Microdroplet Generators 10-21 CCD camera Flashlight Microinjector Droplet trajectory FIGURE 10.25 Experimental setup for droplet stream visualization. U H 0 0 1 2 3 4 5 cm Satellite droplets Main droplets Calculated curve Calculated curve 0 2 4 5 6 8 cm 731 FIGURE 10.26 Visualized droplet stream and estimated trajectory of a microdroplet generator, courtesy Tseng (1998a). 10.5.2 Ejection Direction Droplet direction can be determined by the visualized trajectory. Many parameters, including nozzle shape, roughness, aspect ratio, and wetting property as well as actuation direction and chamber design,affect droplet direction. In general, symmetric structure design and accurate alignment can help control droplet direction. 10.5.3 Ejection Sequence/Velocity and Droplet Volume To characterize the detailed droplet ejection sequence, a visualization system [P H. Chen et al., 1997b, Tseng et al., 1998c] as shown in Figure 10.27 has been widely used. In this system, an LED was placed under the © 2006 by Taylor & Francis Group, LLC droplet generator to back-illuminate the droplet stream. Two signals synchronized with adjustable time delay were sent to a microinjector and an LED respectively. Droplets were ejected continuously from a droplet generator, and the droplet images were frozen by the LED’s flashing light at specified time delays, as shown in Figure 10.17.Droplet volume can be determined from the images by assuming the droplet is axi-symmetric, or from weighing certain numbers of droplets. Droplet velocity can be estimated by meas- uring the difference in flying distance of the droplet fronts in two succeeding images. 10.5.4 Flow Field Visualization Flow field visualization is one of the most direct and effective tools to better understand flow properties such as cross talk, actuation sequence, liquid refill, and droplet formation inside microdroplet generators. Flow visualization in small scale presents some difficulties that do not occur in large scale, such as lim- ited viewing angles, impossible to generate light sheet, reflection from the particles trapped on the wall, short response time, and small spatial scale. Meinhart et al. (2000) adopted a micrometer resolution par- ticle image velocimetry system to measure instantaneous velocity fields in an electrostatically actuated ink-jet head. The system introduced 700 nm diameter fluorescent particles for flow tracing; the spatial as well as temporal resolutions of the image velocimetry are 5–10 µm and 2–5 µs respectively. The four pri- mary phases of ink-jet operation — infusion, inversion, ejection, and relaxation — were clearly captured and quantitatively analyzed. 10.6 Applications More than a hundred applications for microdroplet generators have been explored. This section sum- marizes some of them. 10-22 MEMS: Applications CCD Camera VCR Microscope LED Microinjector Signal A Signal B Synchronizer FIGURE 10.27 Experimental setup for droplet ejection sequence visualization. © 2006 by Taylor & Francis Group, LLC 10.6.1 Ink-jet Printing Ink-jet printing, which involves arranging small droplets on a printing medium to form texts, figures, or images, is the most well-known microdroplet application. The smaller and cleaner the droplets are, the sharper the printing is. However, smaller droplets cover a smaller printing area and thus require more printing time. Therefore in printing, high-speed microdroplet generation with stable and clean micro sized droplets is desired for fast and high quality printing. The printing media can be paper, textile, skin, cans or other surfaces that can adsorb or absorb printing solutions. Ink-jet printing generated revenues of more than $10 billion worldwide in 2000 and will keep growing in the future. 10.6.2 Biomedical and Chemical Sample Handling The application of microdroplet generators in biomedical sample handling is an emerging field that has drawn much attention in the past few years. Many research efforts have focused on droplet volume control, droplet size miniaturization, compatibility issues, the variety of samples, and high-throughput parallel methods. Luginbuhl et al. (1999), Miliotis et al. (2000), and Wang et al. (1999) developed piezo- and pneumatic- type droplet injectors respectively for mass spectrometry. Figure 10.28 schematically shows the design of the injectors,which generate submicron to micron sized bioreagent droplets for sample separation and analy- sis in a mass spectrometer, as shown in Figure 10.29. Luginbuhl et al. (1999) employed silicon bulk micro- machining to fabricate silicon nozzle plates and Pyrex glass actuation plates, while Wang et al. (1999) employed a combination of surface and bulk micromachining to fabricate the droplet generator. These injectors are part of the lab-on-the-chip system for incorporating microchips with macroinstruments. Microdroplet generators were also used by Koide et al. (2000), Nilsson et al. (2000), Goldmann et al. (2000), and Szita et al. (2000) for the accurate dispensing of biological solutions. Piezo- and thermal-type injectors were used in those investigations for protein, peptide, enzyme, or DNA dispensing. With an oper- ation principle similar to ink-jet printing, the devices provided for precisely dispensing and depositing a Microdroplet Generators 10-23 Pressure chamber Nozzle Liquid path PZT disc transducer Silicon Glass FIGURE 10.28 The injector design for mass spectrometry, after Luginbuhl (1999). MS inlet Droplet generators Droplets FIGURE 10.29 Operation principle of mass spectrometry using microdroplet generators. © 2006 by Taylor & Francis Group, LLC [...]... Honolulu, pp 1 32 37 , February 22 26 Chen, P.-H., Chen, W.-C., Ding, P.-P., and Chang, S.-H (1998) “Droplet Formation of a Thermal Sideshooter Inkjet Printhead,” Int J Heat Fluid Flow, 19, pp 3 82 90 Chen, P.-H., Peng, H.-Y., Liu, H.-Y., Chang, S.-L., Wu, T.-I., and Cheng C.-H (1999) “Pressure Response and Droplet Ejection of a Piezoelectric Inkjet Printhead,” Int J Mec Sci., 41, pp 23 5–48 Chuang, Y.-J., Tseng,... Techniques,” J MEMS, 8, pp 601–04 Lee, J.-D., Yoon, J.-B., Kim, J.-K., Chung, H.-J., Lee, C.-S., Lee, H.-D., Lee, H.-J., Kim, C.-K., and Han, C.-H (1999) “A Thermal Inkjet Printhead with a Monolithically Fabricated Nozzle Plate and Self-Aligned Ink Feed Hole,” J MEMS, 8, pp 22 9 36 Lee, Y.-K., Yi, U., Tseng, F.-G., Kim, C.J., and Ho, C.-M (1999) “Diesel Fuel Injection by a Thermal Microinjector,” Proc MEMS, ... Dev., 21 , pp 10 20 Darling, R.H., Lee, C.-H., and Kuhn, L (1984) “ Multiple-Nozzle Ink Jet Printing Experiment,” IBM J Res Dev., 28 , pp 30 0–6 Fromm, J.E (1984) “Numerical Calculation of the Fluid Dynamics of Drop-On Demand Jets,” IBM J Res Dev., 28 , pp 32 2 33 Fuller, S., and Jacobson, J (20 00) “Ink Jet Fabricated Nanoparticle MEMS, ” Proc IEEE MEMS Conference, pp 138 –41, January 23 27 , Kyoto, Japan © 20 06... Turbulent-incompressible Rev Ͼ 23 00, Mv Ͻ 0 .2 2( rh,v)q (fv Rev) ϭ 0. 038 ᎏ Avµvλ 3/ 4 (11 .20 ) C ϭ 1.00 Turbulent-compressible Rev Ͼ 23 00, Mav Ͼ 0 .2 2( rh,v)q (fv Rev) ϭ 0. 038 ᎏ Avµvλ ΄ 3/ 4 ΅ γv Ϫ 1 2 C ϭ 1 ϩ ᎏ Mav 2 Ϫ1 /2 (11 .21 ) The solution procedure is to first assume laminar, incompressible flow and then to compute the Reynolds and Mach numbers Once these values have been found, the initial assumptions... “Spherical Particle Sorting by Using Droplet Deflection Technology,” J Electrostat., 35 , pp 3 12 Ashley, C.T., Edds, K.E., and Elbert, D.L (1977) “Development and Characterization of Ink for an Electrostatic Ink Jet Printer,” IBM J Res Dev., 21 , pp 69–74 Askeland, R.-A., Childers W.-D., and Sperry, W.-R (1988) The Second-Generation Thermal InkJet Structure,” Hewlett-Packard J., August, pp 28 31 Bharathan,... F.E (1984) The Application of Drop-On Demand Ink Jet Technology to Color Printing,” IBM J Res Dev., 28 , pp 30 7– 13 Lee, H.C (1977) “Boundary Layer Around a Liquid Jet,” IBM J Res Dev., 21 , pp 48–51 Lee, J.-D., Lee H.-D., Lee H.-J., Yoon, J.-B., Han, K.-H., Kim, J.-K., Kim, C.-K., and Han, C.-H (1995) “A Monolithic Thermal Inkjet Printhead Utilizing Electrochemical Etching and Two-Step Electroplating... pipe results in the vaporization of a portion of the working fluid The vapor then flows through the central portion of the channel cross-section The return flow of the liquid formed in the condenser is accomplished by utilizing the capillary action at the narrow corner regions of the passage Thus, in the micro heat pipe, wicking is provided by the corners of the passage, thus avoiding the need for a... Bar-Cohen (19 83) have demonstrated that the following combinations of these conditions can be used with reasonable accuracy © 20 06 by Taylor & Francis Group, LLC Micro Heat Pipes and Micro Heat Spreaders 1 1-7 Laminar-incompressible Rev Ͻ 23 00, Mav Ͻ 0 .2 (fv Rev) ϭ 16 (11.18) C ϭ 1.00 Laminar-compressible Rev Ͻ 23 00, Mav Ͼ 0 .2 (fv Rev) ϭ 16 ΄ (11.19) ΅ γv Ϫ 1 C ϭ 1 ϩ ᎏ Ma2 v 2 Ϫ1 /2 Turbulent-incompressible... ceramic particles can be plasma sprayed for surface coating, as introduced by Blazdell and Juroda [Blazdell et al., 20 00] © 20 06 by Taylor & Francis Group, LLC 1 0 -2 8 MEMS: Applications The operation principle is shown in Figure 10 .34 A continuous jet printer was used for droplet formation from ceramic solution The ceramic stream was delivered into the hottest part of the plasma jet and then sprayed onto the. .. many of the parameters is quite different Perhaps the most significant difference was the relative sensitivity of the micro heat pipes to the amount of working fluid present These early steady-state models later led to the development of both transient numerical models and 3- D numerical models of the velocity, temperature, and pressure distribution within individual micro heat pipes [Peterson, 19 92, 1994; . analyzed. 10.6 Applications More than a hundred applications for microdroplet generators have been explored. This section sum- marizes some of them. 1 0 -2 2 MEMS: Applications CCD Camera VCR Microscope LED Microinjector Signal. “Numerical Calculation of the Fluid Dynamics of Drop-On Demand Jets,” IBM J. Res. Dev., 28 , pp. 32 2 33 . Fuller, S., and Jacobson, J. (20 00) “Ink Jet Fabricated Nanoparticle MEMS, ” Proc. IEEE MEMS. ceramic solution. The ceramic stream was delivered into the hottest part of the plasma jet and then sprayed onto the working piece. Splats from the plasma spray are claimed to be similar in morphology