«UNIVERSITY OF CALIFORNIA Santa Barbara A Micro/Nano-Fabricated Gecko-Inspired Reversible Adhesive A Dissertation submitted in partial satisfaction of ...»
The adhesion mechanics of the fine hair adhesive system relies on a large number of weak, short-range interactions to create a large amount of adhesion. The fundamental interaction seems to be a van der Waals force, while it may still be possible that water plays a role through other effects. Using JKR contact mechanics, it is possible to demonstrate the improved adhesion due to contact splitting. Just as important as the contact mechanics at the nanoscale is the compliance of the surface at the micro and mesoscales.
B. MEMS Fabrication Techniques The crux of the doctoral work presented here is the implementation and innovation of MEMS (microelectromechanical systems) processing techniques.
However this thesis is aimed at a broader audience as much of the work relates to mimicking a biological system. It is therefore important to establish a baseline of processing knowledge so as to not inhibit the flow of this document with cumbersome descriptions of standard fabrication techniques. This section will give a brief and to the point fundamental overview of the core MEMS processing techniques.
Arguably the most important technological innovation of this age, photolithography, has brought about the computer revolution, and subsequently the internet. Simply put, photolithography is the transfer of a pattern into a photoactive polymer. This photoactive polymer then acts as a protective shield for the underlying layer, allowing for selective degradation, or coating, of the material below. Very large scale integration (VLSI) technology uses photolithography to densely pattern wafers with transistors and interconnects. In fact, the density of this packing is what dictates the speed and efficiency of computer chips. Refined photolithography has allowed for smaller minimum feature sizes (MFS) and correspondingly smaller, faster and cheaper computer chips.
The two basic forms of photolithography are contact and projection lithography.
Both techniques are used in this work. Contact lithography uses a chrome/quartz mask defined with the desired pattern, outsourced for about $1,000. A wafer spin coated with a photoactive is brought into contact with this mask and irradiated with UV light. The UV light either serves to degrade the photopolymer (positive resist) or to cross link it (negative resist). The wafer can then be washed in a developer to remove unwanted resist, leaving exposed vias in the polymer coating.
As the name implies, projection lithography does not bring the wafer and mask into contact, but instead projects the pattern on the substrate. This is typically done in a stepper tool which projects the pattern with UV light, then moves to a new location projecting again. In this manner the pattern is stepped across a wafer.
Since the pattern is being projected there is a refined control of the focal plane of the image. Additionally steppers usually operate at a 5X reduction in size. This means that if a feature size is 5 μm at the mask level it will be reduced to 1 μm in the resist.
As a result of these factors, projection lithography offers enhanced feature resolution over contact lithography.
Chemically etching materials in a reactive liquid medium has been in use for centuries, and in recent decades has become quite prevalent in semiconductor and MEMS processing. Samples are simply placed in a solution containing a reactive species with the material that is to be removed. Wet etching is generally isotropic as the solution conformably contacts the entire surface. There are a few exceptions to this where chemicals will selectively etch along certain crystallographic planes. In this work, wet etching will be used to remove thin metal layers difficult to remove by other means.
Reactive ion etching (RIE), also commonly referred to as dry etching, is the hallmark of MEMS processing. As the name implies RIE uses reactive ions (and free radicals) generated in a plasma to chemically and mechanically remove material from a surface. The chemical removal of material is performed by appropriately selecting species that are chemically reactive with the substrate. Mechanical etching is accomplished by applying a bias between the plasma and the substrate. Ions are then accelerated in this electric field gradient toward the substrate, removing material. This electric field gradient is of particular importance in this work and relates directly to one of the major innovations produced in this body of work.
Because of the directed nature of the ions bombarding the surface, an anisotropic etch pattern is produced. This allows for the fabrication of high aspect ratio structures with near vertical etch profiles.
Electron beam evaporation, or e-beam deposition, is a method for depositing material evenly across a surface with nanometer precision. Samples to be coated are placed in high vacuum facing a crucible containing the material to be deposited. The material, generally a metal (nickel, titanium, gold in this work), in the crucible is then ablated using an electron beam. The material deposits evenly (due to the vacuum and a long mean free path) on the sample surface. The thickness is monitored by measuring the resonance change of a quartz crystal positioned next to the sample. When the desired thickness is reached the electron beam is turned off.
Another deposition technique, sputtering offers wider materials selection, better step coverage and improved adhesion over e-beam lithography. Instead of evaporating material, as in e-beam deposition, the source material is bombarded with high energy ions effectively knocking off material which then travels to the sample surface and adheres. The positive ions, generally argon, are created in plasma and directed toward a ring of source material using a negative bias.
Indentation testing has long been one of the most important materials characterization techniques. Standard indentation involves indenting a sample with a known force, P, using a tip of particular geometry. After the indentation, the area of the residual indent, A, is measured, and the hardness of the material can be calculated using the relationship H = P/A.
Nanoindentation, as the name implies, uses a much smaller and sharper tip typically in the sub-micrometer regime. While this allows the tip to probe a much smaller volume of material, assessing the size of the indent becomes unreasonable.
To circum-indent this problem instrumentation has been developed to record displacement along with load. Once a characteristic penetration versus load function has been established for a particular tip, load versus displacement data can be analyzed to determine the hardness and modulus of a test surface.
Decreasing the size of the tip also means decreasing the penetration depth and load. To detect these small forces and displacements a high resolution transducer must be used. MTS and Hysitron both make commercially available nanoindenters, with the key difference between them the transduction method. MTS uses an inductor to apply a force and measure the displacement of the tip. Hysitron, the system here at UCSB, utilizes a movable parallel plate capacitor strategy. The Hysitron Triboindenter was utilized for all the nanoindentation described in this document.
The fundamental workings of the parallel plate transducer in the Hysitron Triboindenter is a movable plate mounted between two fixed plates, Fig. II-6. To detect movement, alternating currents 180 degrees out of phase are applied between the fixed plates and the center plate. Any small increment in movement is then detected by measuring the change in voltage between the respective plates. To move the tip downwards, applying a load, a DC bias is applied between the lower plate and the movable plate. This bias draws the plate down and by knowing the applied voltage and the spring constant of the movable plate, the force can be determined.
Figure II-6 Diagram of the parallel plate sensing mechanism in the Hysitron™ nanoindenter system.
A reasonable amount of development has been done on using nanoindentation to measure material properties. However, the use of a nanoindenter to probe the mechanical properties of micromechanical devices has been virtually unexplored.
Some of the first work done and still unpublished is presented here in the appendix.
In this work a variety of three dimensional structures has been fabricated. These structures predominantly consist of 10-200 micrometer platforms supported by a single pillar. While the in-plane shape of the platforms is well defined by the lithography and easily seen through optical microscopy, the out-of-plane shape of the platforms is more illusive. Given the flexibility of the platforms, contact surface profilometry is not possible. Scanning electron microscopy (SEM) provides an excellent means of visualizing the structures, Fig. II-2, but is a time consuming process and offers little quantitative information. To quickly and quantitatively measure the curvature of the platforms, as well as deposition height and etch depths, an optical profiling methodology was employed.
The Wyko NT110 optical profiling system combines optical microscopy with interferometry to produce accurate vertical measurements of surfaces with micrometer lateral resolution. The system consists of an optical microscope, a vertical optical scanning system, a Mirau interferometer and CCD detector array, Fig. II-7.
Figure II-7 Interference microscope design for the Wyko NT1100 optical profiler.
The basic theory of operation is that a white light beam is focused onto a surface.
Prior to reaching the surface, half of the beam is split and focused on an interferometer. The other half of the light reflects off of the surface and onto the interferometer. When the sample is in focus, the path length between the two beams is zero and the intensity is at a maximum. When the sample is moved away from the optics, a series of dark and light interference fringes are passed while the intensity decreases. Another way to conceptualize the fringe effect is to consider a flat sample tilted relative to the optical projection, Fig. II-8. The tilted surface will create an alternating dark and light fringing pattern, with the maximum intensity located in the middle.
Figure II-8 Fringing pattern see on a tilted surface (top). Illustration of the intensity versus height for interfering light on a tilted surface.
So when the intensity is a maximum, the microscope is focused on the surface.
By then focusing the interference pattern on a CCD it is possible to collect the intensity of each individual pixel. The optical system is then scanned vertically relative to the sample. Through calibration of the motor it is then possible to determine the height of each maxima of each pixel creating a map of the surface.
Thus the resolution is limited by the optical resolution and the CCD pixel size.
III. Detailed Micro/Nanofabrication As the main thrust of the work presented here was to use micro/nano-fabrication techniques to mimic a biological system, and although already published, a more thorough treatise of the fabrication processes used is given here. In addition, commentary on the various techniques is given, including motivation, limitations and future directions for development. The chapter is broken up into three sections, representing the three major advancements in fabrication development. Although the final process flow seems logical and straightforward, this work has been a path of discovery, pioneering new areas of fabrication techniques.
A. Silicon Dioxide Platforms Supported by Single Crystal Silicon
The initial concept driving the development of these structures was to create a micro-mirror. While conventional micromirrors utilize four springs, in two different planes, to allow for three dimensional mirror rotation, these structures were designed to have a single spring in the middle of the mirror. By placing the spring in the middle of the mirror, and making it flexible enough, three dimensions of rotation would be achieved through a greatly simplified fabrication scheme. Additionally, the location of the pillar directly below the mirror would allow for greater mirror density, desirable in many applications – e.g. projectors. Unfortunately, or fortunately, about this time, the micro-mirror problem was “solved” and commercial products became available, reducing the driving force for fundamental research.
Nonetheless these structures would find a purpose integrated into a hierarchical adhesive system.
The fabrication of centrally supported released platforms, also called single high aspect ratio pillar support structures (SHARPS), was accomplished using a single photolithographic step, a single oxidation, and two deep reactive ion etches (DRIE). The structures, Figure III-1, were fabricated with silicon dioxide platforms with thicknesses varying from 150 nm to 2 μm (see chapter IV, also the paper titled Single High Aspect Ratio Pillar Support Structures).
Figure III-1 Electron micrographs of SHARPS structures showing the silicon dioxide platform supported by a single crystal silicon pillar. Right image offers a magnified view of the pillar. Scale bars are 20 μm and 10 μm left and right respectively.
Fabrication begins with oxidation of a silicon wafer, Fig. III-2. Generally (100) n-doped 100 mm wafers were used. In one case a heavily p-doped (111) was used, in a failed attempt at a fringing field actuation mechanism. Using (111) wafers is discouraged due to the cleaving direction, not orthogonal, making die separation challenging.
Figure III-2 SHARPS fabrication process flow schematic.
Wafers were oxidized in the Tystar™ furnace in the MEMS cleanroom on the second floor of Engineering II. Oxidation times were calculated using the Stanford website: http://www.lelandstanfordjunior.com/grovedeal.html. Using this calculator based on the Deal-Grove Oxidation Model, an oxide thickness of 2 μm takes 15 hours 15 minutes and 30 seconds using the wetox2 recipe in tube 2 of the furnace.
This thickness was then verified using the Filmetrics white light reflection dielectric characterization tool in the nanofabrication facility.