THE DESIGN AND FABRICATION OF CAPILLARY FORCE MICROACTUATORS FOR DEFORMABLE MIRRORS Alexander Russomanno University of Virginia Advisor: Carl Knospe Adaptive optics (AO) is a revolutionary technology that enhances the quality of astronomical images by correcting aberrations in them caused by atmospheric turbulence and distortions in the primary mirrors of large telescopes. The technology relies on manipulating arrays of microactuator devices mounted behind a deformable secondary mirror to correct errors in wave-fronts of incoming light. Current microactuators used in deformable mirrors lack sufficient force and stroke needed for higher resolution space imaging. This paper focuses on the fabrication of a novel type of microactuator called a capillary force actuator that is theoretically capable of producing 10 to 100 times more force than current actuators of similar size. Prototype devices currently being fabricated at the University of Virginia are too large for adaptive optics applications due to the subtractive fabrication process used. This paper discusses the design and implementation of a new fabrication process that uses nickel electroplating to manufacture capillary force actuators. The unique fabrication techniques developed allow the actuators to be scaled down, while maintaining necessary force and stroke needed for adaptive optics. Adaptive Optics Adaptive optics (AO) technology is the most revolutionary technological breakthrough in astronomy since Galileo pointed his telescope skyward to explore the heavens 400 years ago. 1 Astronomers are now using ground-based telescopes to capture images comparable to those collected by space telescopes by employing adaptive optics to compensate for the distortions in light wavefronts caused by the Earth s turbulent atmosphere. 2 While ground-based telescopes are limited to capturing infrared and lower wavelengths of light, with adaptive optics technology, their capacity to be built larger and more robust allows them to obtain three to four times better spatial resolution than space telescopes. 3 Additionally, future large space telescopes, implementing segmented mirrors, will benefit significantly from AO technology; the large wave-front errors introduced by their primary mirrors can be corrected using a secondary mirror equipped with adaptive optics. 4 With the support of AO technology, the capabilities of ground-based and space telescopes can improve significantly and together, their complementary advantages will contribute greatly to the study of the Universe, and particularly extrasolar planets. 5 Figure 1: Diagram of an AO system 3 Russomanno 1
Top Wafer Contact Side Flexures Liquid Bridge Non- contact Side Bottom Wafer Figure 2: Schematic of a capillary force actuator (not to scale). Top and bottom wafers are fixed and the square suspended platen moves. Deformable mirrors (DM) are an essential component of adaptive optics systems. AO technology uses a closed-loop control system to sense and correct light wave-front errors by manipulating a deformable secondary mirror in real-time see Figure 1 for details. 6 The deformation is achieved by an array of actuators mounted behind the mirror s surface. The performance of the DM is dependent on the spatial frequency, magnitude and dynamic response of its displacement. These attributes are directly related to the stroke, spacing and dynamics of the array of actuators behind the DM. In the past, DMs relied on manually assembled electric or magnetic actuators, which had large mass, high power consumption and limited stroke. 7 Recently, however, the use of Micro Electro Mechanical Systems (MEMS) technology has sparked a revolution in deformable mirror design and fabrication. 8 MEMS technology allows for actuation with much lower cost, power and size. 9 Up until this point, research into actuation using MEMS devices has focused on parallel plate electrostatic actuators and piezoelectric thin film materials, which both have certain limitations for applications in astronomical adaptive optics. 10 A 2008 NSF Roadmap for the Development of United States Astronomical Adaptive Optics cited the need for high stroke deformable mirrors with 10 μm stroke and actuator spacing less than 500 μm to enable correction at high spatial frequencies over narrower fields of view. 11 However, electrostatic actuators require high operating voltages; they are able to produce 3 μm of stroke, but at voltages greater than 200 V. 12 In addition, piezoelectric actuators are achieving strokes as high as 4.5 μm at 100 V, but only at low spatial frequencies (2.5 mm spacing). 10 Due to their limitations, it would be very difficult to upgrade electrostatic or piezoelectric actuators to achieve the stroke and spatial frequency desired for improving astronomical adaptive optics. Capillary Force Actuation Capillary force actuation is a novel MEMS technology developed at the University of Virginia by Prof. Carl Knospe that has distinct advantages over other conventional actuators in deformable mirror applications. 13 As illustrated in Figure 2, the capillary force actuators (CFA) consist of a liquid bridge connecting two surfaces, both of which consist of an electrode covered with a very thin dielectric layer. When voltage is applied across the electrodes, the contact angle of the liquid bridge changes, a phenomenon already recognized and referred to as Russomanno 2
Figure 3: Current microfabrication process for CFA platen. electrowetting on dielectric (EWOD). 14 The change in contact angle results in a decrease in capillary pressure in the liquid and creates an attractive force normal to the two surfaces. 15 When compared to similarly sized conventional MEMS actuators, CFAs are capable of producing significantly greater forces (10 to 100 times) when voltage levels are within the range commonly employed in integrated circuits. 13 Further, CFAs can achieve greater strokes at voltages one tenth to one half that of other actuators in MEMS. 14 Lastly, the out-of-plane forces achieved by CFAs are normal to the device plane, which is ideal for AO technology. 14 Current Fabrication Process Capillary force actuators are being fabricated in the UVAs Microfabrication Laboratory by Huihui Wang (Ph.D. canidate). The current process involves a flip-chip assembly. Figure 2 shows a model of the assembly (moving surface). The square suspended platen, part of the top wafer in the model, is 2 mm across and supported by eight 200 μm long flexures, which are only 3 μm thick. The platen is microfabricated from a silicon-on-insulator (SOI) wafer using an eleven-step process, shown in Figure 3. The process relies on the embedded SiO 2 as a stop layer to carry out front and backside etches. The contact sides (see Figure 2) of the top and bottom wafers are coated with a 30 nm dielectric layer of SiO 2 and a 22 nm hydrophobic fluoropolymer topcoat (Cytop). When the top and bottom wafers are assembled together, one end of the capillary bridge contacts the center of the square platen and the other end contacts the bottom wafer. The liquid wets a circle with a diameter of about 1 mm on both contact surfaces. Voltage is then applied across the two wafers that make up the assembly. A stroke of about 30 μm is achievable with this design. The design of CFAs is well understood at this point, however, for applications in adaptive optics, there are several problems that need to be solved. The main problem, which encompasses many of the others, is that the current size of the CFA is too large. The design shown in Figure 2 is at least four times larger than that required for applications in adaptive optics. Further, the current flexure design is not compact and will not be sufficient if the size of the actuator is to be scaled down. Lastly, the flexures in the current design are anchored to the surrounding wafer, which will not be accessible in constructing device arrays. Future Fabrication Plan The next step in facing these technical hurdles is to develop a new fabrication process. Currently, the fabrication of the CFA requires an isotropic etch through the noncontact side of the top wafer. The purpose of the etch is to leave a 100 μm thick mass attached to the suspended platen to ensure that it maintains its stiffness during electrowetting. Without the mass, the platen could warp or break under the pressure developed in the process. However, the need for a back side etch limits the size and closeness of the features that can be attained. Using this method, it would be impossible to scale the CFA for adaptive optics because closely Russomanno 3
Figure 4: Example of SU-8 Photoresist. 10 μm features, 50 μm thick. packed features would be lost in the isotropic etch process. One promising direction is to move toward a more additive fabrication process by electroplating nickel into a photoresist mold. SU-8 photoresist, an epoxy-based ultra thick negative photoresist manufactured by MicroChem, is known for having excellent imaging characteristics, producing high aspect ratio structures, and its high optical transmission makes it ideally suited for imaging near vertical sidewalls in very thick films see Figure 4 for an example of the unique characteristics of SU-8. 16 These attributes are perfect for creating the structures necessary for CFA scaling, but SU-8 photoresist is also known to be difficult to remove and is said to be best suited for permanent application. 16 However, a technique is being used at the University of Virginia that uses OmniCoat and NANO Remover PG, an NMP based solvent stripper manufactured by MicroChem, that allows for the removal of SU-8 in a lift off procedure. 16 A thin coat of OmniCoat applied before spinning SU-8 allows for the liftoff of the photoresist using MicroChem s NANO Remover PG. This makes it possible to use SU-8 as a mold for electroplating high aspect ratio nickel structures. Being able to electroplate into a mold will allow the Figure 5: Fabrication plan using SU-8 photoresist as a mold for electroplating. fabrication of smaller and more compact CFAs suitable for adaptive optics. The end goal is to create the entire CFA structure using an additive nickel electroplating process. Using an additive process has several distinct advantages over the current SOI back side and front side etching technique. First, the subtractive fabrication process currently being employed requires considerable manipulation and handling of the wafer, resulting in a fabrication yield rate below 50%. This yield rate will drop even lower when the actuators are scaled down to AO size due to greater fragility. Introduction of an additive process will reduce the amount of handling necessary and thereby increase the fabrication yield rate. The SU-8 photoresist mold will also allow for more complex folded flexures necessary for scaling. The plated nickel will also conveniently serve as the electrical connection needed for the electrode on the platen, removing the need for depositing aluminum for the electrode. Intermediate Fabrication Plan In order to test the proposed electroplating fabrication methods, a new intermediate fabrication process was designed, shown in Figure 5, which focuses on building Russomanno 4
CFAs using nickel electroplating on the same scale as the original CFA design. The contact side of the top wafer is build up from a silicon substrate by depositing a seed layer and electroplating the platen structure using an SU-8 mold. The process also uses a new method of creating the dielectric layer. Aluminum is evaporated onto the surface and fully anodized to produce a thin layer of Al 2 O 3. Testing has yielded successful changes in contact angles on the anodized surface. The non-contact side of the top wafer is still fabricated using a backside XeF 2 isotropic etch. Further Research The point of the intermediate fabrication process is to determine the properties of electroplated nickel and see how it works in CFAs, specifically as a material for EWOD. Fabrication is currently underway, but there are still a few technical hurdles to overcome before the devices are ready for testing. The next step is to scale down the CFA by building up the entire structure using nickel electroplating. There are some important factors that need to be considered in the ultimate array design required for AO: (1) the flexures need to be anchored to something other than the surrounding wafer for an array design; (2) the current design for the flexures is not compact and therefore not suitable for small-scale actuation; (3) the volume needed for liquid deposition is significantly smaller than what the micropipette currently used is capable of. We will approach these problems once we have fabricated some of the larger scale nickel-based devices. Conclusion The development of microactuator arrays with high stroke (>10 μm) and high spatial frequency (<500 μm spacing) at voltage levels commonly employed in integrated circuits would contribute significantly to adaptive optics technology. This paper discussed the research of a novel type of microactuator called a capillary force actuator that is theoretically able to attain higher stroke and force than microactuators normally used in adaptive optics. CFAs would contribute significantly to improving adaptive optics technology. Therefore, they are an important technology for next generation ground-based and space telescopes. Improving these will allow scientists and astronomers to attain higher resolution images of outer space and help them in the search and study of exosolar planets. Acknowledgements First, I would like to thank the Virginia Space Grant Consortium for giving me the opportunity to do this research. I would also like to thank Professor Carl Knospe and Professor Michael Reed for letting me work with them on this extremely interesting project. I would also like to thank Huihui Wang for teaching me all about fabrication and capillary force actuators. Lastly, I would like to thank the University of Virginia for the opportunity to do research as an undergraduate. This project has inspired me to continue research in pursuit of my doctorate in mechanical engineering. References [1] R. W. Duffner, The adaptive optics revolution: A history, University of New Mexico Press, 2009. [2] M.C. Roggemann, B.M. Welsh, and R.Q. Fugate, Improving the Resolution of Ground Based Telescopes, Review of Modern Physics, Vol. 69, No. 2, April 1997. [3] C. Max, Introduction to Adaptive Optics and its History, American Russomanno 5
Astronomical Society 197 th Meeting, January 2001. [4] E. Yang and K. Shcheglov, A Piezoelectric Uniform Deformable Mirror Concept for Wafer Transfer for Ultra Large Space Telescopes, SPIE International Symposium on Astronomical Telescopes and Instrumentation, Waikoloa, Hawaii, August 22-28, 2002E. [5] B. MacIntosh, J. Graham, B. Oppenheimer, L. Poyneer, A. Sivaramakrishnan, and J. Veran, MEMS-Based Extreme Adaptive Optics for Planet Detection, Proceedings SPIE, Vol. 6113, 2006. [6] J. Gibson, C. Chang and B. Ellerbroek, Adaptive Optics: Wave-Front Correction by use of Adaptive Optics Filtering and Control, Applied Optics, Vol. 39, No. 16, 2000. [7] D. Gavel, New Adaptive-Optics Technology for Ground-based Astronomical Telescopes, SPIE Newsroom, February 10, 2011, http://spie.org/x44333.xml. [8] S. Cornelissen, P. Bierden, T. Bifano, and J. Stewart, MEMS Deformable Mirrors for Adaptive Optics in Astronomical Imaging, The Advanced Maui Optical and Space Surveillance Technologies Conference, Wailea, Maui, Hawaii, September 10-14, 2006. [9] Yang, K. Shcheglov, and S. Troiler- McKinstry, Concept, Modeling, and Fabrication Techniques for Large- Stroke Piezoelectric Uniform Deformable Mirrors, SPIE Photonics West, Micromachining and Microfabrication Conference, San Jose, CA, January 2003. [10] X. Xu, B. Li, Y. Feng, and J. Chu, Design, Fabrication and Characterization of a Bulk-PZTactuated MEMS Deformable Mirror, Journal of Micromechanics and Microengineering, Vol. 17, No. 12, December 2007. [11] A Roadmap for the Development of Astronomical Adaptive Optics, National Optical Astronomy Observatory, http://www.noao.edu/, April 2008. [12] T. Bifano, S. Cornelissen, and P. Bierden, MEMS Deformable Mirrors in Astronomical Adaptive Optics, 1 st AOELT Conference, EDP Sciences, 2010. [13] C. Knospe, Method and System for Capillary Force Actuators, US Provisional Patent June 2007. [14] S.A. Nezamoddini, Capillary Force Actuators, PhD. Dissertation, Mechanical and Aerospace Engineering, University of Virginia, 2008. [15] C. Knospe and S.A. Nezamoddini, Capillary Force Actuation, Journal of Micro - Nano Mechatronics, DOI 10.1007/s12213-009-0023-4, January 2010. [16] MicroChem, SU-8 2000: Permanent Epoxy Negative Photoresist, http://www.microchem.com/pdf/su- 82000DataSheet2000_5thru2015Ver4. pdf. Russomanno 6