High Aspect Ratio Photolithography for MEMS Application

Review There are several methods that have been used in photolithography and fabrication of molds for use in metal microstructures. LIGA is one of the most common methods that have been previously used for this purpose. this approach has some advantages in that it can be used with tall microstructures that ranged from 100µm to 1mm. The effectiveness of this method is that it has no effect on the lateral dimensional accuracy of the long microstructures making it a versatile method, and this explains why it is popular among other photolithographic methods. The limitation of this method is that fabrication of LIGA involves a considerably high cost, and lack of an x-ray synchronized source that would give hard x-rays of high intensity and with low divergent properties. Reactive Ion Etching (RIE) s another method that involves polyamide, which is sued to fabricate low cost high aspect ratio structures as compared to LIGA above. With this process, a thickness of 100µm is achieved with the aspect ratio being 10. The limitation of this process is the tediousness of modifying the RIE machine. In addition to these approaches, there are several other methods that have been used for this purpose and include the high aspect commercial photolithography with photoresists and a UV light source, and the use of a photosensitive polyamide with a UV exposure in fabricating metal molds. The research by Miyajima Mehregany (1995) involved integrating the above methods in previous studies, and solving the limitations of the studies to achieve low cost high-aspect-ration structures in photolithography. To achieve this, commercially available positive photoresists and UV exposure increasing molds were used together with electroless nickel plating. In the research, a 2-3µm line width was maintained and increasing the photoresists thickness, that the above LIGA approach in photolithography lacked. The research used electrostatic actuator micro machined at 1.5µ active gaps with a polysilicon of up to 5µm. The difference in this approach was that instead of increasing the gap height, the gap side was increased. This resulted to reduced use of high aspect ratio structures due to reduced force/torque. This is the main difference comparing this approach and the previous approaches in photolithography. Coating speed was very important in the cause of this research. A lower speed of 1000 rpm portrayed rough surfaces in the photoresists and mask-wafer spacing was small to result to a high resolution pattern. The optimum speed for high resolution was set at 2000rpm and a uniform coating, thickness and a smooth final surface was achieved after the final layer. The prebake temperatures were essential in that an extended prebake time resulted to high-resolution with diluted developer and standard exposure. The prebake temperature ranged between 95 and105 oC, with the optimum temperature being set at 100oC. After each layer, prebake was done at the same temperature to enhance drying. This was because, prebake after applying the second layer was observed to result to cracking mainly due to uneven heating in the two layers that resulted to stresses, and eventually cracking. The optimum prebake temperature of 100oC at 30 min was selected as the optimum prebake conditions after each layer. This time was important in that a short prebake time could have resulted to defects caused by the remaining solvent in the photoresis