Article Index

. What is MEMS? This abbreviation stands for Micro-Electro-Mechanical Systems: "Micro" means small dimensions of elements (about 0.5-1 micron, but not smaller — this is one of the reasons to design such devices exactly in Fab 8); "Electro-Mechanical" means the importance of both electric and mechanical properties; "Systems" implies MEMS integration with other chips. MEMS is sometimes called MST — Micro Systems Technology.




One of the methods to manufacture MEMS is surface micromachining. This method consists in consecutively applying films (layers) on the silicon wafer and subsequent processing. There are two layer types: (1) the structural layer has the desired electrical, mechanical, and thermal properties; (2) the sacrificial layer. This layer is the first to be deposited. It supports the structural layer until it's etched and then it's removed from a wafer.




The key application of MEMS is wireless devices. In the wireless system block diagram above, MEMS can act as a pre RF unit, which contains band pass filters, amplifiers, tunable elements, etc. MEMS advantages lie in using silicon to make all components of the system, that is the best integration, which leads to smaller form factors, better performance, and low power consumption.


The second line of research in the R&D department of Fab 8 is semiconductors or silicon photonics. Photonics is a technology of emitting, transferring, controlling, and detecting photons using principles similar to those in optical fibre and optoelectronic devices. Advantages of using optical technologies for data transfer lie in miniature dimensions and simplicity of such devices (using a few signal lines), absolute insensibility to electromagnetic interference, low losses at long-range transmissions even of high theoretical throughput (up to 100 TB/s). Nevertheless, they are not without drawbacks, high costs in the first place due to exotic and expensive materials (gallium arsenide, indium phosphide, etc), expensive manufacturing and packaging technologies. In this connection, Intel sees its semiconductor photonics task in developing CMOS-compatible photonics — photonic devices based on usual silicon and standard streamlined technologies for semiconductor production in large volumes in the existing plants. Even though Si integration into photonics offers some problems, as silicon does not have some necessary properties like electro-optic effect and effective photon emission, Intel's research and development in this area has made great progress since the first Raman-effect continuous silicon laser designed in 2004.


 

Please publish modules in offcanvas position.