MEMS are Micro-Electro-Mechanical Systems and they are an enabling technology for so many of the devices we interact with everyday. While their near ubiquity may come as a surprise to some, their economic impact should not be understated. MEMS and the components or products they empower fuel markets in the tens of billions of dollars per year. Recently, Knowles Corporation announced that they’ve shipped eight billion MEMS microphones globally. Think about that for a minute – That’s more microphones than people living today. And that’s just one MEMS supplier and one of the myriad applications of MEMS…
MEMS represent an interesting combination of materials science, electrical engineering, and chemistry. Almost every major class of advanced materials are used in concert to produce today’s MEMS. Similar to semiconductor device fabrication, silicon remains the workhorse. Beyond the immense photolithography patterning infrastructure for crystalline silicon in place today, silicon has good mechanical properties for many MEMS. It is highly resistant to fatigue. This allows for repeatable and predictable motion at the micro-scale enabling billions of cycles. CMOS-MEMS integration is also possible using silicon-based materials, which can improve MEMS performance and lower both packaging size and cost.
Various fluorine-containing chemicals find use in MEMS fabrication for etching processes. Light sensitive polymers are employed in the photoresists that are used to pattern MEMS substrates. Polymeric structures can also be used directly in MEMS devices. One example is in polymer microfluidic chips. Like polymers, ceramics are used in both MEMS processing and end devices. Silicon dioxide is often a key material in etching processes. Silicon carbide, various nitrides, and piezoelectric ceramics like lead zirconate titanate (PZT) find use due to their electrical and/or mechanical properties.
MEMS production leverages many of the same techniques used in the mature integrated circuit industry which means low production costs per-device and high levels of scalability (See Chris Erickson’s blog on Moore’s Law to learn more). While modern semiconductor device fabrication equipment can have staggeringly high upfront price tags, these costs can be spread over many die per wafer in a batch fabrication process. This often leads to price per device counted in terms of pennies or dimes. As such, MEMS-based solutions have seen commercial success in inertial, chemical, and pressure sensing along with electrical applications (e.g. RF or bulk acoustic wave (BAW) filters) and optical switching. They’re also increasingly being explored and utilized in the medical space for ultrasonic transducers and lab-on-chip microfluidics.
The ability to rapidly ramp production and drive down costs using established IC fabs is a bit of a double edged sword. Profitability can be elusive for MEMS companies, both large and small. Thin margins and continued price erosion are very real concerns for MEMS components that have become commoditized, such as inertial sensors (i.e. accelerometers, magnetometers, and gyroscopes) in smartphones. Compared to many other materials businesses we look at here at Pangaea, the economies of scale effects on profitability are typically not as strong in MEMS. Furthermore, the lion’s share of the profits are often captured by the systems companies rather than the pioneering component suppliers.
So how does a resource-constrained startup successfully compete in this arena? Relatedly, how does a venture capital investor make “smart” investments in companies innovating in the “smart” sensor space? To me, it comes down to product differentiation. Sure, running lean, improving production yields, and continuous technology innovation are important too and should not be overlooked, but if you’re selling a component that doesn’t sufficiently move the performance or new functionality needle, it’s going to be a struggle. It also will be a business that VCs will (rightly) pass on.
Budding MEMS entrepreneurs should ask themselves the simple yet oft-overlooked question: Will customers pay for the innovation I am developing? If the answer is “no”, it may be time to rethink the plan. If it’s “I’m not sure”, then it’s time for some serious value chain and end market research. If it’s a resounding “Yes!”, well, that’s a pretty good place to start.
Next the real work begins. Processes must be designed and optimized, wafers must be processed, dies need to be separated, interconnects must be made, and it all has to be packaged into a component ready for integration into an end product. Positioning the company within the value chain adds more complexity to this equation but it’s a critical variable. Can additional value be captured by the company by integrating more functionalities and/or proprietary software or firmware? If so, is it realistic or economical for a hardware startup to undertake these endeavors in-house?
MEMS development cycles can be daunting and these timelines can be challenging for a returns-focused VC investor. However, a company commercializing a highly differentiated MEMS product in a large and growing market with a strong IP position and scalable process makes for an interesting proposition. We at Pangaea relish the opportunity to help make the next generation of MEMS companies a huge success.