MEMS devices are being developed with high mechanical, thermal and electrical power densities. These "Power MEMS" are being considered for portable electrical power applications as alternatives to battery technology. The mechanical power densities that can be achieved are directly proportional to the strength of the materials that are used as the prime movers. Furthermore, for a given power density, the total power available scales with the volume of material under stress. These system requirements drive towards the creation of high strength materials and structures with scales which push the upper limits of what can be realized by traditional microfabrication technologies. Deposition processes are attractive means of creating such structures, however there are significant mechanics issues associated with their use. In particular the control of residual stress is key. The presence of residual stress affects the ability to control down stream processes, such as planarization and wafer bonding and in extreme cases can lead to fracture of the deposited material.
Experimental results are presented from two material development efforts, for silicon carbide and silicon dioxide. In both cases chemical vapor deposition processes were used to create layers in the range 10-100 µm thick on silicon wafers. Wafer curvature measurements at varying temperatures were used to quantify the thermal and "intrinsic" components of the residual stress. Microscopy and elemental analysis were used to identify the factors affecting the residual stress levels and this information was utilized to achieve process control. For the case of silicon oxide, a systematic study of the thickness dependence of film cracking was also carried out. The implications of these material and process development activities on the MIT Microengine will be discussed.