The invention of the scanning tunnel microscope (STM) opened the door to nanoworld, enabling humankind to actually touch and feel the individual atoms at the surface of a material. It has inspired a series of inventions such as the atomic force microscope (AFM), the lateral force microscope (LFM), the magnetic force microscope (MFM), etc. Recently, the scanning probe microscope (SPM), which encompasses all these inventions, has evolved into the millipede technology which has a potential to make competing data storage technologies obsolete. A common denominator of these exciting developments is a microscopic structure fabricated by MEMS techniques: the microcantilever.

Microcantilevers are not confined to imaging microscopy. They can also be found in the nanopatterning efforts using near-field scanning optical microscopy (NSOM), scanning electrochemical microscopy (SECM), dip-pen nanolithography (DPN), as well as in microsensors. Microcalorimeters offering ultra-high sensitivity have been demonstrated, and ultra-low (ppb) concentrations of toxic gases in air detected. Moreover, surface stresses were measured for the self-assembly of alkanethiols on gold, DNA hybridization or receptor-ligand binding. By using a nanoscale resonator, a mass sensitivity of a few femtograms was reported.

Microcantilevers certainly have fundamental advantages as microsensors. Thousands of microcantilevers can be prepared in a single wafer, so they are well-suited for miniaturization. The detection mechanisms are very simple – just measuring the bending deformation or the shift of the resonance frequency of a beam. Chemical or biological stimuli are directly converted to mechanical responses, resulting in a high degree of selectivity and efficiency.

We have devised novel microcantilevers with a radically different material and structure, as well as fabrication method. Figures 1(a) through (c) show SEM photographs of the proposed microcantilever array at different magnifications. The overall shape of the microcantilever array can be seen in Figure 1(a). All cantilevers have the same thickness (2μm) and length (50μm), but different widths ranging from 10μm to 50 μm. Figure 1(b) is the magnified image of the edge of one microcantilever in Figure 1(a). Straight nanochannels are clearly visible at this magnification. Further magnification in Figure 1(c) unveils the hexagonal arrangement of nanochannels on the surface of the microcantilever. This photograph was taken from the bottom surface to prove the presence of nanochannels all the way through the thickness of the microcantilever. These nanochannels with extremely high aspect ratio are impossible to obtain by conventional lithographic techniques.

The centerpiece of this novel development is the wellknown nanomaterial, anodic aluminum oxide (AAO). It typically has parallel channels of tens of nanometers in diameter up to hundreds of microns in length. It has been commonly used as molds for nanomaterials such as nanotubes and nanowires, taking advantage of these high-aspect-ratio nanochannels. Moreover, replicas of these nanochannels can also be applied in many important fields such as nanoimprint lithography, superhydrophobic surface, and photonic crystals. Recently, AAO technology has experienced a major technological leap by adopting photolithography. This combination will result in various MEMS/NEMS structures with nanochannels in the years to come.

To our knowledge, microcantilevers have been made exclusively of isotropic materials until now. Since the operation of a microcantilever relies on the bending of the beam, Young’s modulus is an important material property. A microcantilever made of an isotropic material has a fixed Young’s modulus, so cantilever design is rather limited. In contrast, our microcantilevers possess tunable nanochannels. Since the nanochannels are arranged in one direction, the AAO is not isotropic, and the Young’s modulus is by no means a unique material property. In fact, our microcantilever has multiple Young’s moduli depending upon the direction, and, best of all, the Young’s moduli can be controlled by varying the dimensions of the nanochannels. This is an extremely important advantage in the design of microcantilevers, which has the potential to greatly expand their applicability.

The mere presence of nanochannels is beneficial to some applications. Microcantilevers are frequently operated in a vibration mode which has a high sensitivity. The resistance from the surrounding fluid consumes much energy, and eventually causes damping of the vibration of a microcantilever. It becomes a serious problem when microcantilevers are used as remote sensors with microbattery power sources. With the microcantilevers developed here, gas molecules freely move through the nanochannels, so that the resistance is reduced greatly.

The deflection of a cantilever stems from the change in surface energy. Therefore, a large surface area is generally advantageous in most sensor applications. Cantilevers based on AAO can provide surface areas several orders of magnitude larger than conventional silicon cantilevers with flat surfaces.

The fabrication method for the microcantilevers proposed here is quite different from that of conventional silicon counterparts. The proposed method consists of (1) AAO fabrication by anodization, (2) patterning on the AAO layer by photolithography, (3) fabrication of the beam structures by anisotropic etching of the AAO layer, and (4) removal of the substrate below the AAO layer by electrochemical etching.

The fabrication process of AAO microcantilevers is outlined with the corresponding SEM images in Figure 2. The first step is to form an AAO layer on top of an aluminum layer. The particular sample shown in Figure 2(a) has hexagonally ordered nanochannels 30nm in diameter. The channel diameter, the channel-tochannel distance, and the thickness of the AAO layer can be controlled by changing the anodization conditions such as reaction time, applied voltage, temperature, etc.

Patterns of microcantilevers are made in the second step by conventional photolithography. A thin layer of aluminum was deposited on the surface of the AAO as a transfer layer, followed by the spin coating of a photoresist layer. The transfer layer is necessary to provide a smooth surface, which generates sharper boundaries. Patterns for the microcantilevers were formed on the photoresist layer using a photomask, and transferred to the transfer layer by removing the exposed area using aluminum etchant. Since the aluminum etchant also dissolves alumina, the channels in the underlying AAO layer were slightly widened during this step. The SEM image in Figure 2(b) shows the microcantilever patterns formed on the AAO. The brighter area is the area with the exposed AAO, though the channels are not visible at this magnification. Note that the microcantilevers were intentionally fabricated with different widths for later measurements.

After the pattern transfer, the sample was immersed in a phosphoric acid solution to remove the AAO in the exposed area. The acid solution penetrated into the nanochannels of the exposed area, while the area blocked by the photoresist was protected from the acid solution. As a result, the AAO in the exposed area was selectively etched away. The SEM image in Figure 2(c) shows the structures after AAO etching. Although the AAO layer is rather thick, 10μm, the microcantilever patterns of the AAO were successfully formed with vertical sidewalls which are hard to obtain with the wet etching of silicon. The presence of nanochannels in the AAO made the anisotropic etching of the complex patterns possible. Note that the photoresist layer and the transfer layer still exist on top of the AAO structures.

Suspended cantilevers are fabricated by removing the aluminum under the patterns without damaging the AAO. Electrochemical etching was chosen for this. Suspended microcantilvers are clearly observed in the SEM image of Figure 2(d). The conventional fabrication method for suspended microcantilevers requires a much more complex procedure including deposition, patterning, and etching. Low pressure chemical vapor deposition (LPCVD), plasma enhanced CVD (PECVD) as well as reactive ion etching (RIE) are frequently used in the fabrication of silicon microcantilevers. Our method greatly simplifies the fabrication process through electrochemical etching.

The huge surface area of an AAO microcantilever offers an enormous number of binding sites. The adsorption of chemical or biological molecules on the binding sites of a cantilever shifts the resonance frequency. The relation between the mass of adsorbed materials and the shift of resonance frequency is given as , where f1 and f0 are the resonance frequencies after and before mass loading. The more molecules adsorbed on the surface of a cantilever, the larger the shift of the resonance frequency, as obvious from this equation.

Figure 3 shows the shifts of the observed resonance frequencies after injecting saturated dodecanthiol on two different microcantilevers. Both cantilevers have the same thickness(2μm), length(250μm) and width(35μm), but the AAO microcantilever has vertically parallel pores with 50nm diameter while the silicon cantilever has a flat surface. The red line in this figure is the observed frequency shift of the AAO microcantilever as a function of time, and the black one is those of the Si cantilever. The shift of resonance frequency of the AAO microcantilever (68Hz) is about 6 times as large as that of the Si cantilever (11Hz) at the saturated condition. This large frequency shift is partly due to the large surface area of the AAO cantilever, and partly due to the different physical properties such as smaller Young’s modulus. Therefore, the AAO microcantilever is much more sensitive than the Si cantilever of identical overall dimensions.

Moreover, the presence of nanochannels provides not only a large surface area, but also selectivity among different analytes based on a sieving mechanism. Thus, the AAO microcantilever seems to be a promising candidate for biosensors.

The fabrication method proposed here offers more freedom than conventional methods. For example, AAO microcantilevers can be fabricated with thicknesses up to hundreds of micrometers. Figure 4(a) shows ultra-thick AAO microcantilevers of 100μm in thickness. The spring constant of a microcantilever increases in proportion to the cube of the thickness of the beam. Therefore, such thick microcantilevers are favorable for applications which require very stiff beams, i.e., they can operate at much higher frequencies. Note that these structures have vertical sidewalls which are hard to obtain with conventional microcantilever fabrication methods.

The fabrication process of AAO microcantilevers also guarantees structural versatility at a larger scale. Such versatility is demonstrated in Figures 4(b) and 4(c) with rectangular and triangular loop microcantilevers. The former is routinely adopted in the microcalorimetric detection, while the latter is found in most AFMs.

The nanochannels of an AAO cantilever can be made with only one end open. A sensing material with a special affinity to a specific analyte can be deposited inside the bottom of the nanochannels. A strand of DNA may attach to the sensing material without the interference of other DNAs, which is hard to achieve with microcantilevers with flat surfaces. The enormous difference in the surface area between the upper and the lower surfaces may create other interesting applications.

In summary, we have reported microcantilevers which have a radically different material and structure, as well as fabrication method. They are made of alumina, and have hexagonally-arranged parallel channels tens of nanometers in diameter. They can provide a surface area several orders of magnitude larger than those of conventional microcantilevers. Precise control of the dimensions of the nanochannels is made possible through the versatile fabrication method, resulting in fine tuning of the physical properties. This fact gives us an additional handle for the design of microsensors. These revolutionary microcantilevers are made possible by combining photolithography with nanotemplates. This technique should greatly expand our ability to fabricate other 2-D structures with nanochannels, and has great potential to be the general fabrication method of various NEMS devices in the years to come.

Professor Kun-Hong Lee
Department of Chemical Engineering
Tel: +82-54-279-2271
Fax: +82-54-279-8298
E-mail: ce20047@postech.ac.kr

Professor Sangmin Jeon
Department of Chemical Engineering
Tel: +82-54-279-2329
Fax: +82-54-279-5528
E-mail:jeons@postech.ac.kr

Professor Hyun Chul Park
Department of Mechanical Engineering
Tel: +82-54-279-2167
Fax: +82-54-279-5899
E-mail: hcpark@postech.ac.kr

Professor Woonbong Hwang
Department of Mechanical Engineering
Tel: +82-54-279-2174
Fax: +82-54-279-5899
E-mail: whwang@postech.ac.kr