«UNIVERSITY OF CALIFORNIA Santa Barbara A Micro/Nano-Fabricated Gecko-Inspired Reversible Adhesive A Dissertation submitted in partial satisfaction of ...»
Figure VI-4 Nanoindenter test data plotted against analytical van der Waals (vdw) predictions. The circles represent a smooth polymer surface without modification. The squares represent a hydrophilic organorod surface on a solid substrate. The triangles represent a hydrophobic organorod surface on a solid substrate. Curve a represents a simple vdw model based on the number of organorods contacting the test surface, the radius of curvature of the organorod tips, and using a cut-off distance of 0.165 nm. The number of contacts is estimated by knowing the density of organorods and calculating the contact area of the indenter tip using JKR contact mechanics. Curve b is identical to curve a, but with a reduced cut-off distance in the vdw equation of
VII. Meso-Scale Adhesion Testing of Integrated Micro- and Nano-Scale Structures
Many insects and lizards display the amazing ability to climb and stick to just about any surface. Recent research has honed in on these systems to better understand how they work(6, 7, 15, 70), particularly on how fine sub-micron hairs enhance van der Waals, or other short-range interactions. Provided enough intimate surface contact these “weak” forces can add up to produce significant amounts of adhesion. Additionally, the attractive interaction must be much larger than repulsive forces, due to elastic deformation, for the adhesive to be effective. To achieve this, nature has created a hierarchal structure to conform over a range of size scales(2). In previous work, microfabrication techniques were used to create a synthetic dry adhesive modeled after the fine hair adhesive motif found in nature(71). The artificial structure consists of a silicon dioxide platform, covered with organic looking polymeric nanorods (“organorods”), and supported by a single single-crystal silicon pillar. The Multiscale Integrated Compliant Structures (MICS) offer three levels of surface compliance: (1) the organorods on the surface (also necessary for enhancing surface adhesion), (2) the fingers of the platforms, and (3) the flexibility of slendor silicon pillar supporting the platform. In this work large arrays, 1cm x 1cm, have been batch fabricated across an entire 10 cm wafer.
Additionally, to characterize mesoscale adhesion, a nanoindentation adhesion test technique was extended to measure the adhesion between the microfabricated samples and a rough 5 mm diameter aluminum flat punch. Results indicate improved adhesion with the integration of the nano- and micro-structures. The multi-scale structures also demonstrated improved wear characteristics over solid supported organorods.
Keywords: adhesion, nanorod, nanoindenter, 3-D, MEMS, gecko, dry-adhesive
In the emerging field of biomimetics, the ever growing knowledge base of biology is brought together with the rapidly developing ability to measure and manipulate properties at very small length scales. Since the time of Aristotle, scientists have been fascinated by the gecko’s ability to scale virtually any surface, and under completely different environmental conditions(2, 6). In the last hundred years scientists have speculated that the adhesion force in the gecko pad is a result of suction, secretions or capillary forces. Recently Autumn et al. performed a series of experiments giving convincing evidence that van der Waals interactions are the dominant interaction force (6).
An excellent example of convergent evolution, the gecko has developed highly refined 200 nm protrusions to maximize van der Waals interactions (2, 6, 7).
However, for this surface contact to be effective there needs to be a minimum amount of repulsive force from the surface. To achieve this, the fine hair adhesive motif has a multi-scale compliant structure designed to conform to varying levels of surface roughness.
The largest scale of conformation, in the case of the gecko, is the gecko itself with a body and legs able to move in and around tens of centimeter size objects.
Moving down in size scale are the toes with the ability to wrap around curved surfaces. Within these toes there are blood sinuses acting as a hydraulic suspension, deforming with little elastic response to millimeter scale roughness. These sinuses support rows of imbricated lamellae composed of rows of keratinous setae 30-130 μm in length and approximately 20 μm in diameter(2). These slender setae can deform to micrometer scale roughness, and offer the densely packed array of fibers necessary to maintain large amounts of surface contact. The terminus of the setae subdivide into 200 nm diameter spatulae, capable of achieving the last level of surface intimacy necessary for van der Waals forces to become significant.
Prior work has succeeded in replicating the final terminal structure of the spatulae, but centimeter scale testing resulted in negligible adhesion(11). The low values of adhesion were attributed to reduced surface conformation across multiple length scales, and to the inability of the surface interface to absorb energy and arrest interfacial crack growth. Additionally, it was seen that the hydrophilic nanorods would adhere to each other reducing wear characteristics.
The emergence of the field microelectromechanical systems (MEMS) over the last three decades has brought with it a variety of microsensors and transducers(42, 44-48, 50, 52, 72). One of the many challenges still remaining in the microsensors field is the deployment and placement of microdevices. The development of a technique to microfabricate an adhesive, capable of sticking to virtually any surface, could greatly enhance the potential for “fly on the wall” distributed sensing.
In previous work the authors have used microprocessing techniques to create novel micro- and nano-structures mimicking the hierarchal structure of the gecko(71). Here, this fabrication technique has been extended to create larger 1 cm x 1 cm arrays and a new testing methodology has been implemented to measure mesoscale adhesion characteristics of the system. The processing technique is fully compatible with standard microprocessing, requires only a single lithographic step, and uses only dry etch techniques. The structures produced follow a similar motif to the fine hair adhesive in nature by creating multiple levels of compliance. The Multiscale Integrated Compliant Structures (MICS) consist of a single single-crystal silicon pillar supporting a silicon dioxide platform coated by polymeric organorods (Fig. 1). The silicon pillars can have high aspect ratios and diameters as small as 1 μm. The silicon dioxide platforms are 1 μm thick and consist of four radial meandering fingers extending 50 μm from a central square platform. Atop these platforms are arrays of vertically aligned ~250 nm diameter, ~4 μm tall organorods composed of positive photoresist (Fig. 2). Combining these three structures an analogous system to the fine hair adhesive is created mimicking the multiple levels of compliance on a chip. The first level of compliance, like the toe of a gecko, is the small size of the chips that can be produced, allowing the chip itself to fit within centimeter scale roughness. The next level of compliance is the flexible silicon pillars, allowing the entire platforms to rotate, accommodating sub-millimeter scale roughness. The fingers of the oxide platforms allow for conformation to tens to hundreds of micron size features. And finally to maximize surface area contact, and enhance van der Waals interactions, are 250 nanometer diameter organorods.
B. Fabrication The MICS structures were fabricated using a single lithographic step and multiple etch process(59, 71). Single crystal silicon wafers in the (100) orientation were used for all fabrication. Wafers were first coated with 1μm of thermal oxide using a wet oxidation process at 1150°C. Positive photoresist (Shipley™ SPR220-7) was spun on the wafers to a thickness of 7 μm. The top platforms were then patterned in the resist using projection lithography. This pattern was then transferred into the underlying oxide using an inductively coupled plasma (ICP) etch with CHF3 chemistry defining the platforms in the silicon dioxide. Deep reactive ion etching, using the Bosch process, was then used to extend the oxide pattern vertically ~35 μm into the bulk silicon. Subsequent to the extension etch an extended isotropic SF6 etch is run, undercutting the silicon dioxide, and creating oxide platforms supported by a single silicon pillar (Fig. 1).
To create the polymer organorods, the photoresist coated structures were placed in an inductively coupled oxygen plasma. By controlling the oxygen pressure, RF bias, and time the photoresist surface was transformed into first a roughened morphology and then into a coating of organorods(71) (Fig. 2). Different organorod morphologies were observed by adjusting the various growth parameters.
In this work better spacing and isolation of the organorods was achieved by increasing the RF bias and growth time than in previous work. Additionally, uniform coatings of organorods have been created over a complete 100 mm wafer, with and without structures. As can be seen in Figure 1, incomplete coverage was seen on the platform fingers as a result of previous degradation of the photoresist during the silicon dioxide and silicon etch steps. As will be discussed momentarily, this incomplete coating may have a significant impact on the adhesion strength of the system.
Using this fabrication method, 100 mm wafers have been created with 80 separate 10 mm x 10 mm arrays of 2,500 platforms, each containing ~50,000 organorods. The process technique is true batch fabrication allowing for simultaneous creation of 200,000 platforms and ten billion organorods. This massively parallel approach highlights the potential of this fabrication technique for future low cost device implementation.
C. Experimental Previous testing of artificial fine hair dry adhesives has focused on measuring the adhesion of individual nanorods(7, 11, 12). Scaling of individual nanorod adhesion greatly overestimates the adhesion of larger scale (50μm) adhesion testing (11). This suggests the need to develop alternate test techniques for measuring the meso-scale adhesion of these systems. Additionally, a rough surface was desired to test the MICS in a simulated application environment, e.g. an aircraft wing. To accomplish these testing requirements a Hysitron Triboindenter™ was retrofitted with 5 mm aluminum flat punch tip. The tip was fabricated using a stock indenter shaft and a 5 mm aluminum puck (RMS roughness of 2.5 μm). To bring the two components into registration the puck was placed on the sample stage, with a drop of glue on the top surface, and the shaft was fitted into the Triboindenter™ transducer.
By performing a standard load controlled indent, the tip was brought into contact with the puck and allowed to rest there while the glue dried.
Operating the nanoindenter in displacement control, the flat punch tip was brought into contact with the test surfaces and withdrawn at a constant rate recording the normal force and displacement. Upon unloading an adhesion event would cause a negative normal load. The maximum negative force was taken to be a measurement of adhesion. Repeated tests were performed in the same location to assess wear characteristics of the structures.
D. Result and Discussion Previous testing used a spherical tip geometry with a diameter of ~3mm(71).
While the spherical geometry avoids the requirement for near perfect tip sample registration, the spherical tip produced several complications in testing. The spherical geometry, even at scales as large as 3 mm, would only contact a single platform structure, making it difficult to ascertain the conglomerate effects of arrays of structures. Additionally, the tip added strong normal load dependence to adhesion (more indentation force would increase the surface contact area and subsequent adhesion). The incorporation of the 5 mm flat punch into the testing aimed to alleviate these two complications.
Using the aluminum flat punch, adhesion testing was performed on three test surfaces including: photoresist coated silicon wafers, organorod coated wafers, and arrays of MICS on a silicon substrate (Fig. 3). The photoresist surface showed very little adhesion with the average adhesion strength of 0.1 Pa. The organorod surface demonstrated improved adhesion strength of 6.5 Pa. Integrating the pillar supported platforms with the organorods enhanced adhesion by more than a factor of three to
21.8 Pa. No significant dependence was seen on the maximum positive applied normal load.
Upon repeated adhesion tests in the same location, the MICS demonstrated no reduction in adhesion by the fifth iteration. Whereas the adhesion to the organorod surface decreased nearly to zero by the fifth indent (Fig. 4). The diminished adhesion is likely a result of damage to the organorod surface by the indenting puck.
Individual organorods are either pushed down to the surface where they remain, are permanently deformed moving them out of the contact zone, or condense into clumps of organorods. The integration of the flexible platforms may improve wear by flexing before a critical load is placed on the organorods, causing nonrecoverable damage. This behavior is not unlike that of the gecko, where the long shafts of the setae can bend before the nanoscopic spatulae receive an excessive load.
The experimental values of adhesion strength for the MICS of ~20 Pa is much lower than those reported for the gecko, Gekko gecko, of ~100 kPa(6, 73). Although there is much work left to be done to rival nature, there are several significant testing differences that account for the orders of magnitude difference in adhesion strength.