Image-based autonomous micromanipulation system for arrangement of spheres in a scanning electron microscope

Abstract

The micromanipulation technique in a scanning electron microscope (SEM) has been attracting interest as a technique to produce microstructures such as three-dimensional photonic crystals or advanced high-density electronic circuits. However, it is difficult to fabricate a large-scale structure or to conduct a systematic experiment using numbers of structures, as long as we rely on manually operated micromanipulation. In this study, we constructed an automatic system which arranges 10-μm-sized microspheres into a given two-dimensional pattern in a SEM. The spheres are picked up by touching with the center of the planar tip of a probe (needle), and placed on the substrate by moving the contact point to the edge of the tip and inclining the probe. The positions of the probe and the spheres are visually recognized from the SEM image from above and the optical microscope image from the side. The generalized Hough transform, which can robustly detect arbitrary shape from the edge fragments, is employed for the image recognition. Contact force information obtained by a force sensor with a resolution of 14 μN is also utilized for the control. Completely automatic rearrangement of randomly sprinkled metal spheres with a diameter of 30 μm into arbitrary patterns was successfully demonstrated. Autonomous micromanipulation technique under the observation of a SEM would contribute not merely to laboratories but also to the opto-electronics industry.