Professor, Department of Micro-Nano Systems Engineering, and Director, Center for Creative Engineering, Graduate School of Engineering, Nagoya University
Three-dimensional fine structures created by etching
—Micro/nano processing using crystal anisotropy—
Prof. Sato, who was researching metal forming, started research on micro electro mechanical systems (MEMS) in 1983. The technology was not called MEMS at that time, but the concept was beginning to emerge in the United States. He says, “For a long time, I thought that machines were fabricated based on conventional concepts, although they had progressed, but then, I began to have the feeling that machines could be smaller by one or two orders of magnitude in size with a substantial change in processing. Then, I decided to shift my research to MEMS.”
Prof. Sato, who studied anisotropy in metal polycrystals as the subject of his doctoral thesis, thought of using anisotropic etching of single crystal silicon for fabrication of MEMS devices. He investigated the anisotropic etching of single crystal silicon in potassium hydroxide (KOH) and tetramethylammonium hydroxide (TMAH) aqueous solutions using a hemispherical single crystal specimen with a diameter of 44 mm (Fig. 1). He says, “Crystal planes with arbitrary orientations appear on the surface of a hemispherical single crystal specimen. When the specimen surface is etched, the crystal planes where the etching rate is high are deeply etched, while the planes that are barely etched remain. When the shape of the hemispherical specimen is measured with a three-dimensional measuring instrument before and after etching, an etching rate contour map (Fig. 2) can be obtained. The accumulated experiments enable us to predict how to obtain a desired shape using calculations.” The database, ODETTE, and an etching simulation system, MICROCAD, which were developed through accumulated experiments using different types of etchants, and temperatures and concentrations of the etchants, have been used by Japanese companies.
Prof. Sato says, “The experiments using the hemispherical single crystal silicon specimens dealt with fabrication in the range of millimeters to micrometers. However, the etching rate contour map that is experimentally obtained is quite variable according to changes in etching conditions. We actually found significant differences in the map pattern between KOH and TMAH. In order to find out how the difference occurs, knowledge of nanotechnology is required.” He has experimentally confirmed that the orientations of atomic steps that are easily etched on a crystal plane differ with the types and concentrations of the etchants. He says, “It has been thought that the etching rate can be determined by the energy that is required to take an atom out of a defectless surface. However, taking an atom out causes a defect, from which an atomic step moves laterally along the surface, and the etching process proceeds.”
Prof. Sato introduced a defect, which becomes a nucleus of atomic steps, on a (111) silicon plane and conducted an etching experiment by changing the types and concentrations of the etchants. This defect may be regarded as a hexagonal seed pit on the (111) plane, of which the sidewalls are composed of three equivalent atomic steps each with a single dangling bond that make a triangular shape and another three atomic steps each with two dangling bonds that make a triangular shape rotated by 60 degrees. Since the atomic steps with two dangling bonds are etched faster than the ones with a single dangling bond in a 40% KOH aqueous solution (9.5M), as predicted from conventional theories, the seed pit becomes a triangle comprised of the three equivalent atomic steps each with a single dangling bond in the KOH solution. In contrast, since the atomic steps with a single dangling bond are etched faster in a 25% TMAH aqueous solution (2.7M), the seed pit becomes a triangle comprised of the three equivalent atomic steps each with two dangling bonds in the TMAH solution. In a highly concentrated 60% KOH (16.7M) solution, the steps with a single dangling bond are etched faster and in a low concentrated 10% TMAH (1.1M) solution, the steps with two dangling bonds are etched faster. Therefore, the seed pit becomes a triangle in the low concentrated KOH solution and also becomes a triangle that is rotated by 60 degrees in the highly concentrated TMAH solution (Fig. 4). This experiment was extremely important, because it showed that the etching activity depends on crystal plane orientations and changes largely due to the types and concentrations of the etchants.
Prof. Sato says, “It has been said that only a hydroxyl group in the etchant controls the etching phenomenon. However, I think positive ions in the etchant have something to do with it. It is likely that positive ions stably attach somewhere on a stepped silicon surface, which blocks the movement of atomic steps. I have just started to research the interaction between positive ions and surfaces using first-principle calculations under collaboration with Helsinki University of Technology, Finland.” The critical molar concentration of hydroxyl group, at which the etching activity is reversed, changes with the etchant. “When the volume fraction of the positive ions in the etchant is used instead of the concentration of hydroxyl group, the critical concentration is 10 vol% in both KOH and TMAH solutions. This phenomenon is not understood yet,” says Prof. Sato. Although the diffusion of the ions in the etchant has not been considered as a factor controlling the etching rate, a localized diffusion phenomenon in the etchant around the atomic steps may also be one of the factors. “So, I have been trying to build a mechanism that enables us to understand the phenomenon in the range of the atomic level to the millimeter level together with physicists and chemists,” says Prof. Sato.
