«UNIVERSITY OF CALIFORNIA Santa Barbara A Micro/Nano-Fabricated Gecko-Inspired Reversible Adhesive A Dissertation submitted in partial satisfaction of ...»
This paper describes the realization and characterization of nanofabricated organic looking polymer nanorods, “organorods,” for use in a biomimetic adhesion system. The adhesion system is inspired by the fine hair adhesive motif found in nature and best exemplified by the gecko. The meso- to nanostructure of the gecko’s foot is designed to maximize inelastic surface contact to enhance van der Waals interactions. In this work fabrication and characterization of nanostructures has been performed for inclusion in a multiscale system mimicking the natural adhesive using cleanroom based processing techniques. The system consists of flexible silicon dioxide platforms, supported by a single silicon pillar, coated with ~200 nm in diameter and ~4 μm tall polymeric organorods. The organorod surface is altered between hydrophilic and highly-hydrophobic. The adhesive properties between the artificial surface and a 3.175 mm aluminum sphere are measured in a modified nanoindenter.
Initial results indicate improved adhesion with the hydrophobic surface over the hydrophilic, further corroborating van der Waals interactions to be the operative force of adhesion and suggesting a reduced cut-off distance in the van der Waals theory.
The emergence of microsensors has come about due to increased microfabrication abilities and gain in understanding of microscale mechanics.
Sensors for chemical, mass, pressure, and inertial sensing are among the most common thus far (42, 47, 51, 52). Significant developments in microscale sensors have enjoyed 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.
Found in nature, and best demonstrated in the pad of the gecko’s foot, the fine hair adhesive system is an excellent example of convergent evolution in biology(1-4, 6-8, 54). Researchers have puzzled over the phenomenon back all the way till the days of Aristotle. Recently much work has been done to better understand the mechanism behind the adhesion (6-8, 15, 56, 67), and convincing data has shown that the adhesion is primarily due to short-range weak van der Waals interactions between the fine hairs on the adhering surface and the external surface(7). In order for these weak interaction forces to become significant a large amount of inelastic surface contact must be made between the surfaces. If the adhesive were to elastically conform to the surface then there would be a repulsive force present from the strain in the material. To avoid this the gecko has a multilevel conformation structure which allows for a large amount of surface contact with out creating a repulsive force. The multi-level structure consists of toes containing blood sinuses supporting rows of imbricated lamellae of densely packed keratinous setae, approximately 100 µm in length and 20 μm in diameter, which then split into finer 200 nm in diameter bristles (2).
Previous work in this area has focused on creating the final terminal bristles(7, 11, 12, 58). While individual nanorods demonstrated expected amounts of adhesion, larger arrays failed to produce larger amounts of adhesion – unless removed from the substrate and placed on a compliant backing(11). This illuminates the need for a multi-level conformation scheme. Additionally, due to processing constraints, prior work has not fully duplicated the nano-structure, done so by any mass production means, or emulated the super-hydrophobic nature of the gecko pad (until now only thought to aid in self-cleaning of the surface). In this paper a new fabrication technique to produce, in a massively parallel fashion, submicrometer polymeric organorods will be discussed. To test the relative functionality of the nanostructures an adhesion test technique utilizing nanoindenter instrumentation will be described and results compared with a simple analytical adhesion model.
For incorporation into the multi-scale system a compatible nanorod fabrication technique was developed to create 50-150 nm diameter and ~4 μm tall organorods (Fig. 1). The process is compatible with standard microelectromechanical fabrication utilizing a modified reactive ion etch (RIE) process. To create the organic looking nanorods a 4 inch silicon wafer is first soaked with HMDS for 30 seconds and spun ‘dry’ at 3500 rpm. The wafer is then coated with photoresist (Shipley™ SPR 220-7, a diazoquinone ester and a phenolic novolak resin resist) spinning at 3500 rpm and baking at 95˚C for 90 seconds. The wafer is then placed into an inductively coupled plasma (ICP) reactive ion etcher using an oxygen plasma with an applied bias between the sample and the plasma (Fig. 2). For the nanorods tested in this paper a RF power of 300 W, an oxygen flow rate of 40 sccm and a 10 minute growth time is used to coat a 4 inch wafer. Further transformation of the hydrophilic organorods to a highly-hydrophobic surface is achieved by placing the sample in a fluorine plasma. The fluorine deposits a conformal coating on the organorods creating a Teflon® like surface.
A Hysitron Triboindenter™ is used to measure adhesion between the test surfaces and a 3.175 mm diameter aluminum sphere (see schematic Fig. 2). In this experiment the aluminum sphere is glued to the tip of a nanoindenter probe. The device is placed in the indenter and the indenter tip pressed against the test surfaces.
The nanoindenter is operated in displacement control and records the load vs.
displacement. To determine the adhesion the loading and unloading curve is analyzed extracting the minimum normal load, taken to be the maximum adhesion.
In the case of no adhesion this value is zero or a positive number. In the case of an adhesion event the number would be negative, representing a pulling force on the retracting tip.
Silicon wafers were coated with three different surfaces and tested with the nanoindenter. The first surface was a simple photoresist surface with no other modifications, the next a hydrophilic organorod surface, and the third a hydrophobic organorod surface.
The contact angle of the hydrophilic and hydrophobic surfaces were measured by placing a drop of water on the test surfaces and taking a picture through a magnifying glass using a Canon A95 digital camera. The high resolution images were then magnified and printed, where lines could be drawn to measure the angle with a protractor. The organorod surface was found to initially have a hydrophilic surface with a contact angle of 42.5° ± 2. After the fluorocarbon treatment the surface became highly-hydrophobic with a contact angle of 145° ± 2 (compared to
160.9° ± 1.4 for the gecko (7)),
The organorod formation may best be explained by the interaction of the dielectric polymer and the electric field gradient. Schäffer et al. [ref. 20] observed a similar phenomenom when they placed a polymer in an electric field created by two parallel plates. They postulated that the dielectric polymer experienced a force created by the electric field gradient. This force would then be accentuated by an instability wavelength ultimately determining the size of the pillars(60). While qualitatively this explanation fits, the pillars formed using the plasma growth method produced much smaller dimensions. This suggests that the plasma growth method may have other benefits in addition to the ease of alignment and growth. Two factors may be contributing to the reduced geometry of ~50-150 nm the organorods (Fig. 1).
One a bilayer of photoresist and hexamethyldisilazane (HMDS) may reduce the interfacial energy, shrinking the instability wavelength(61), and decreasing the final pillar size. Secondly, the oxygen ions may reduce the surface energy of the polymer, decreasing the surface force opposing the electric field force, and further reducing the size.
Adhesion values were obtained by observing loading and unloading curves in the indenter testing described previously and visually represented in figure 3.
Initially a plot shows a no load versus displacement until the sphere comes into contact with the surface. There is then a positive increase in the load with further displacement. Upon retraction the load curve falls below zero to a critical point, where it suddenly jumps back up to zero load. 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 in the case of the organorod surfaces, but not for the smooth photoresist surface, Figure 4. This increase in adhesion with load is due to a greater number of organorods coming into contact with the indentation tip. The flexible organorods offer little restoring force and the net adhesion is much higher. Whereas the solid photoresist surfaces produces an elastic restoring force, repulsing the indenter tip and reducing the adhesion.
To better understand and verify the adhesion data taken in the nanoindenter a basic analytical treatise was given to the system considering van der Waals (vdw) interactions, where the force between a sphere and a flat surface is taken to be FvdW = HRo/6do2, where H is the Hamaker constant (estimated to be H = 2.1x10-21 W =
2.8x10-19 J), Ro is the diameter of the organorods, do is a characteristic cut-off distance, and W is the work of adhesion. The Hamaker constant is estimated to be H = 2.1x10-21 W [ref. 22]. The cut-off distance was initially estimated to be 0.165 nm(64), but was then reduced as will be discussed. The work of adhesion, W, was taken to be √γAl γpolymer, where γAl = 500 mJ/m2 and γpolymer = 35 mJ/m2 and 19 mJ/m2, for the hydrophilic and hydrophobic respectively. In this system the radius of the terminus of the organorods is taken to be much smaller than that off the aluminum sphere. In this manner the tips of the organorods are taken to be many small spheres contacting a flat surface, the aluminum indenter tip. The radius of the organorods was taken to be 200 nm for the hydrophilic and 350 nm for the hydrophobic. The number of organorods in contact was estimated using JKR contact theory(68). The contact theory was used to estimate the radius of contact between the spherical aluminum tip and the organorod surface. This of course requires the elastic modulus of the organorod surface to be known. By indenting and measuring the contact radius at different indent loads it was possible to adjust the equation appropriately. The radius indent and density of organorods was measured in a scanning electron microscope, which also allowed for verification of the number of organorods that contacted the sphere. Using the density, and contact radius as a function of load, the number of organorods in contact at any given load could be determined. For example at a contact load of 10 mN approximately 1200 organorods were estimated to be in contact. Then knowing the number of organorods in contact it is possible to determine the total adhesion based on the van der Waals expression above.
Plotting the predicted adhesion force against applied load there is reasonable agreement between the organorod surface and the van der Waals prediction of above (Fig. 4 curve a), note that this test was meant to simulate a “real life” adhesion situation by using a rough indenter tip and scattered data is to be expected. When the organorods are given a hydrophobic coating there is an unpredicted increase in adhesion. The vdw force equation is highly dependent on the characteristic cut-off distance, which is indicative of the interaction distance between the two surfaces. It may be possible that the hydrophobic coating brings the two surfaces into more intimate contact, reducing the cut-off distance. To test this, the model was modified with different cut-off distances, do. A do equal to 0.09 nm, instead of 0.165 nm, yielded a much better prediction of the empirical data (Fig. 4 curve b).
Unfortunately there is a great deal of scatter due to the testing procedure.
Considering the potential scientific merits of this find a more ideal test will be performed to test influence of the hydrophobic coating on adhesion.
A massively parallel fabrication technique for producing ~200 nm diameter, ~4 um tall polymeric pillars has been developed. The technique is compatible with “dry” cleanroom processing and allows for inclusion into a multi-scale system mimicking natures fine hair adhesive(69). The nanostructures have been tested using a novel nanoindenter adhesion test to confirming their enhancement of adhesion over a smooth surface. Adhesion was further increased by the addition of a hydrophobic coating. This increase in adhesion is attributed to more intimate surface interaction due to a reduced van der Waals interaction distance.
E. Acknowledgments and Correspondence We would like to thank the UCSB nanofabrication staff for all their support and insight. Funding for this work was provided by the AFOSR contract number FA9550-05-1-0045 and the University of California, Santa Barbara.
Correspondence should be addressed to MTN at firstname.lastname@example.org
Figure VI-1 Scanning electron micrographs of the organorod morphology, scale bars 5 μm and 1 μm, left and right respectively.
Figure VI-2 Schematic of the organorod growth mechanism. A bias is applied between oxygen plasma and the substrate, creating an electric field gradient, which acting on the dielectric polymer draws it in the direction of the gradient.
Figure VI-3 Schematic of the nanoindenter test setup. The sphere represents the 3.175 mm aluminum sphere which is fixed to a shank fitted into the nanoindenter transducer. When the sphere is pressed into the platform structures a loading and unloading curve is produced. At the bottom of this curve there is a sudden jump from a negative force value to zero. This jump is indicative of pull-off associated with adhesion, and correspondingly taken to be the adhesive force.