When Will MEMS Switches be Ready for Commercial Products

Abstract

of applications due to their broadband operation (i.e. DC-40GHz). The RF MEMS switch, however, has not yet made it to “prime time” due to its perceived limitations in both performance (i.e. stable electrical contact resistance) and reliability (i.e. lifetime). Let me explain the perceived limitations and how I see a way forward to commercialization. First, the original specifications for RF MEMS switches were < 1dB insertion loss and >1×1012 switching cycles. I believe these specifications were originally based on using RF MEMS switches to replace transistor switches onboard satellites. Unfortunately, this original specification has become the “gold standard” for all applications. Next, the majority of RF MEMS switch researchers have primarily focused on either searching for the “holy grail” electrical contact material or on designing innovative, high contact force, mechanical switch designs (i.e. actuators). Both of these approaches are attempts to “engineer away” poor performance and reliability with either a special contact material or a better actuator. Unfortunately, most of these efforts have done little to actually improve micro-switch performance and/or reliability. As of today, I am aware of only one viable company that is pursuing the manufacture of microswitches and several companies that have gone out of business.3 There are, however, a few companies that have in-house research projects to develop micro-switches for their own products.3 I believe the way forward to commercialization MEMS switches is to first study the fundamentals of micro-contact physics and thin film metallurgy in a meaningful way. What does this mean? First, as a community MEMS switch researchers need to focus on the root causes of device failure not just avoiding failures. For example, when studying the electrical contact it should be decoupled from the mechanical switch design so that the micro-contact physics can be studied directly. The results of these fundamental studies could then be used to design micro-contacts suitable for micro-switches.