«UNIVERSITY OF CALIFORNIA Santa Barbara A Micro/Nano-Fabricated Gecko-Inspired Reversible Adhesive A Dissertation submitted in partial satisfaction of ...»
For the thinner, and more uniform evaporated nickel samples, a deep etch and release could be performed to create the nickel platforms. However, upon organorod growth, the force required to grow organorods would also act on the flexible platforms permanently deforming them into an undesirable state. Unfortunately, thinner platforms also mean more compliant platforms. The more compliant the structures the more conformation they are capable of, subsequently enhancing adhesion. From an optimization standpoint, this presented a dead end. Fortunately, however, there was a simple solution. It was found that organorod growth could be performed prior to the final releasing of the structures. This innovation proved to be essential to this work and will allow for further optimization and development.
Figure III-6 Electron micrographs of released organorod coated nickel paddle structures. Opposing paddles coated with organorods scale bar 20 μm (top). Magnified view with dimensions of organorods, scale bar 2 μm (bottomleft). Array of paddles, scale bar 100 μm (bottom-right).
To create thin flexible platforms, the platform pattern was deep etched into the silicon and a partial undercut was performed. This partial undercut etched away the bulk of undesired silicon, but left a support for the thin platforms. Organorod growth was then performed as above. The samples were then transferred back to the PlasmaTherm system to complete the undercut etch and release the flexible platforms. While in the system, the fluorocarbon coating could also be applied.
IV. Single High Aspect Pillar Support Structures
This paper describes the realization of a new three dimensional dry processing technique used to create structures supported by a single high aspect ratio support. This process allows for the creation of submicron single crystal silicon features, with tight geometric control, attached to much larger (10-200μm) silicon dioxide platforms. These single high aspect-ratio pillar support structures, or SHARPS, lend themselves to testing of submicron features and open new avenues of device design. One application for SHARPS as a passive adhesive is also investigated. The unique geometry created by SHARPS structures allows the structures to conform to non-planar
For several decades researchers have been extending IC fabrication techniques into the rapidly growing field of micro electro mechanical systems (MEMS).
Through a variety of processing techniques, the MEMS community has been able to carve microdevices out of bulk materials or build them up layer by layer. Each innovation in fabrication is added to the design toolbox furthering the capabilities of the MEMS community as a whole. Also helping further device innovation is an increased understanding of microscale phenomenon, specifically dynamics and micromechanics. Micromechanical testing is not new to the MEMS community (30-38)), however, little work has been done to test the mechanics of refined threedimensional geometries at the micron size. Here the ability to create refined highaspect-ratio geometries at the micron and submicron size is reported. The fabrication techniques described add yet another tool to the MEMS fabrication toolbox and allow for new opportunities in testing.
The emergence of microsensors has come about due to increased microfabrication abilities and gain in understanding of microscale mechanics.
Sensors for location determination, chemical sensors, mass sensors, pressure sensors, inertial sensors, are among the most common thus far (39-52)). Significant developments in microscale sensors have enjoyed significant interest because they possess important properties, including high sensitivity and the ability to be fabricated in large arrays of thousands to millions. However, there still exists a question in the deployment of these arrays. Adhesion of sensor arrays is one area necessary for many applications that require the deployment of microdevices. In addition chip-scale recognition and adhesion systems will be of great interest for use in fabrication, self-assembling microdevices and immerging miniature aerospace applications. The idea of micromechanical Velcro was laid out over a decade ago by Han et al. (53)). In the work the idea of recognition and mechanical interlocking on two separate chips was introduced. In this paper, the initial steps to create a more universal chip-scale microadhesive are discussed.
A number of different single high aspect ratio support (SHARPS) structures have been produced, Figure 1. They all follow the same fundamental process flow, shown in Figure 2. Fabrication begins by growing 2 μm of oxide on a 4 inch silicon wafer in the (100) orientation. The top SHARPS platform is then defined in the photoresist using an I-Line stepper. This pattern is then transferred into the oxide through ICP reactive ion etching using CHF3 chemistry. The remaining photoresist is stripped from the wafer and the oxide is then used as a mask in etching the silicon below.
This is accomplished by placing the wafer in a PlasmaTherm ICP etcher and performing a modified Bosch etch process, in which the plasma gas is cycled between a reactive etching chemistry, SF6, and a polymer producing species, C4F8, producing high aspect vertical trenches. The duration of each etch or passivation step can be modified to control the vertical profile of each trench, the extreme of this being a sustained SF6 where the silicon is nearly isotropically etched. By controlling these parameters it is possible to create silicon dioxide platforms supported in a single point by a silicon pillar, Figure 1.
Single crystal silicon wafers 100 mm in diameter in the (100) orientation were used for all processing. Silicon dioxide was grown to a thickness of 2 μm using the wet oxidation process at 1050 C in a Tystar Tytan Furnace (Tystar Corporation, Torrance, California). Standard stepper lithography was carried out using a GCA Mann 5x I-line stepper (Hampton, New Hampshire). The photoresist pattern was transferred into the oxide using a Panasonic E640-ICP dry etching system using CHF3 chemistry (Panasonic Factory Solutions, Osaka, Japan). The deep etching and release was carried out using a modified Bosch Process in a PlasmaTherm 770 ICP reactive ion etcher (Plasmatherm, North St. Petersburg, Florida). In the case where titanium was deposited, DC reactive sputtering was used using an Endeavor 3000 cluster sputter tool (Sputter Films, Santa Barbara, CA). Titanium was transformed into titania through ICP etching of the samples in Cl2/Ar.
Scanning electron micrographs were taken in a S2400 Hitachi scanning electron microscope (Hitachi Instruments Inc., San Jose, California) or on a FEI XL40 Serion FEG Digital Scanning Microscope (FEI, Hillsboro, Oregon). Surface profilometry was performed with Wyko NT100 Optical Profiler (Veeco, Woodbury, New York).
Nanoindentation and adhesion experiments were performed in a Hysitron Triboindenter (Hysitron Inc., Minneapolis, Minnesota).
As seen in Figure 1, a variety of different SHARPS structures have been fabricated. The structures have top platforms varying in size from 10 to 200 μm and pillar diameters from 0.5-10 μm, and heights from 5 to 40 μm. It should be noted that the extents of dimensions are not limiting, with the exception of the pillar diameter. This dimension is not immediately defined by the oxide mask. However,
extending the notion of pattern transfer from the planar oxide into the 3dimensionally etched silicon a variety of pillar geometries have been produced:
round, triangular, square, hexagonal and octagonal pillars, Figure 3. In addition to altering the radial geometry of the pillar the shape along the long axis has been controlled, though to a lesser degree, by creating a bend along the long axis of the pillar, Figure 4.
The planar shape of the top platform is easily controlled. This shape is directly controlled by the resist pattern and gives the designer great flexibility. A variety of different shapes including platforms, slotted platforms, serpentines of different varieties, branched fingers and radial meanders have been produced, see Figure 1.
As can be seen in Figure 1 the platforms stay remarkably planar over large distances, ~50 nm height change over 40 μm. The planarity of the platform was disrupted by DC reactive ion sputtering of titanium on the top surface. This served to curl the platform upwards due to a stress gradient. With increased thickness of titanium there was an increase in the radius of curvature of the platforms, shown in Figure 5.
Characterization of the structures created has primarily been carried out using scanning electron microscopy. The electron micrographs give us detailed information on the structures created, including shape and dimensions. This information is invaluable when trying to correlate data from other tests, specifically nanoindentation testing. Preliminary micromechanical testing has been performed using a nanoindenter to apply a force at a specific location on the top platform. By monitoring the force versus displacement the bending mechanics predicted from simple bending beam theory correlated with indentation distances far from the center, Equation 1 and Figure 6.
Where b is the thickness of the top surface, w is the total width of support arms, L is the distance from the center to the point load, a is the pillar size and h is the height of the pillar. Moving closer to the center of the structure the theoretical and experimental stiffnesses deviated. This is quite easily explained by the significance of other bending modes, specifically lateral displacement of the platform and off axis bending, as the tip moves closer to center. Off axis bending may be a result of the diamond shape of the pillar, bending about the diagonal as opposed to the orthogonal. Since this is the direction of maximum stiffness the structure may concede to the less stiff off axis modes. This is illustrated in Figure 6, where the top plot shows the diamond shape and the bottom plot shows the square shape, the latter more accurately capturing the bending mechanics closer to the center.
Qualitatively these structures have been found to be quite robust. Although more compelling in video form, a few snapshots of pushing on one of the structures in a probe station are shown, Figure 7. In the figure the structures bend a large amount before ultimately failing, depicted in Figure 7. Also, while pushing on the platform fingers large amounts of elastic deformation are observed with the fingers deforming in every direction and then snapping back into place.
Characterization of the surface topology of the supported platforms was carried out using a Wyko Profiler. Using the light interferometer system quantitative surface plots of the silicon dioxide platforms with and without titanium sputtered on the surface were obtained, Figure 5. In looking at the electron micrographs there is a dramatic change in curvature of the top platform with the addition of the sputtered titanium. Whereas the initial platform was flat an addition of approximately a half micron of titanium causes the 80 micrometer structure to rise 7 μm on either side.
With the addition of approximately 5 μm the edges rise up 23.5 μm. In the latter structure it appears as though the structure has come under some self-limiting condition where the titanium on fingers are interacting with each other. Looking to alternative structures deflections could be further increased by choosing different platform geometries, Figure 8. Here two structures with a radial meander platform geometry are shown, wherein four radial spokes serpentine outward from the center.
The use of the nanoindenter to measure adhesion between SHARPS structures and a 1.5 mm diameter polyamide sphere is reported. In this experiment a polyamide sphere was glued to the tip of a nanoindenter probe. The device was placed in the indenter and the indenter tip pressed against SHARPS devices and the solid substrate. Initial efforts were made comparing a radial meander SHARPS structure array and the substrate, both covered with 5 μm of titanium under a thin layer of titania. The nanoindenter was operated in displacement control and recorded the load vs. displacement. Initially a plot would show a no load versus displacement until the sphere came into contact with the surface. There would then be a positive increase in the load with further displacement. Upon retraction the load curve would fall below zero to a critical point when it would then return to the origin. The minimum load value is taken to be the maximum adhesive force. The maximum adhesion was found to be highly dependent upon the maximum positive load value, Figure 9. In the case of the SHARPS structures, the indenter tip would have to travel much further to achieve significant positive normal forces. This ultimately limited the maximum force applied to the SHARPS structures, as the instrument has a 5 micrometer maximum travel limitation. However looking at the trend in figure 10 it appears that there is a significant increase in adhesion with the structures as opposed to the solid substrate. This increase in adhesion can be attributed to an increase in contact area. Any adhesive system is going to have an overall increase in adhesion when there is more contact area. In the case of these structures they are more capable of conforming to the spherical morphology than a solid substrate.
Future work on this project will entail continued development of novel processing to fabricate new SHARPS structures. Helping guide the development of new structures will be new and refined characterization techniques. Initial work in indentation and adhesion testing has been discussed. Further of development of these techniques to make robust test methodologies is needed. These techniques may include continued development of nanoindentation test techniques and integrated manipulation and SEM systems. With improved fabrication and testing techniques, the integration of the SHARPS structures into a working device will be explored. Of particular interest is the use of the SHARPS structure as a chip-scale microadhesive. To this end, investigations of integration and reliability will be explored.