Silicon crystal is too expensive, the future MEMS materials are paper and plastic

In the view of the International Semiconductor Industry Association (SEMI), microelectromechanical systems (MEMS) technology has grown the fastest in the semiconductor field in recent years, so how to accurately predict the future of MEMS? After learning about the history of MEMS components and reviewing 500 of the most innovative MEMS papers, Alissa Fitzgerald, founder of MEMS design and development company AM Fitzgerald and Associates LLC, at this year's MEMS & Sensors Executive Congress ) Sharing optimistic views and predictions about the future development of MEMS.

According to Fitzgerald, "The next billion-dollar product is hidden in the university's research literature." The 2017 academic paper reveals passive and near-zero sensors, as well as paper and plastic based products. The program replaces the latest advances in expensive silicon-based solutions as consumer products and special products for single use.

AM Fitzgerald is committed to the future of MEMS, and they are committed to applying innovative academic and entrepreneurial ideas to small MEMS fabs, benefiting from the use of Soitec's commercial silicon and silicon-on-insulator (SOI). Wafers are the same as Rogue Valley Microdevices (RVM).

In his speech, Fitzgerald talked about the history of MEMS technology, dating back to the development of acid-etched three-dimensional (3D) force sensors in the 1980s, which led Kurt Petersen to invent a pressure sensor based on bulk silicon micromachining technology. The pressure sensor eventually realized the inkjet nozzle and prompted the emergence of digital light processing (DLP) MEMS, and soon the first manufacturer used the accelerometer from ADI to trigger the airbag, which is more than the traditional in-line ball mechanical network. Technology is faster.

“Since then, Bosch's deep reactive ion etching (DRI) process has opened up a whole new era, enabling the world's first MEMS gyroscope, film bulk acoustic resonator (FBAR), and MEMS piezoelectric. The widespread use of aluminum nitride (AlN) films has also spawned the various MEMS components we have today."

Another important invention, Fitzgerald said, is "precisely aligned eutectic bonding, enabling InvenSense to bond its own ASIC wafers to MEMS chips for automatic sealing, eliminating the need for additional capping steps."

According to Fitzgerald, in the early days, major companies such as ADI and Bosch met more than 50% of the market demand, and the remaining 400 small companies divided the remaining market. But with the popularity of smartphones, the huge consumer market has made these 400 small companies the main force in the market.

So where do all these consumer market ideas come from? Fitzgerald believes that to a large extent traceable to academia, they "cultivate creativity in university laboratories" as a solution to the problem. Organizations such as AM Fitzgerald put the ideas of scholars into design and develop them into products that are suitable for sale, providing momentum for today's global mega dollar consumer market.

Looking to the future, and then working hard to find out the technology that is being developed in university laboratories. Fitzgerald said in his speech, "After reviewing the top 500 papers in 2017, we conducted a commercial feasibility screening. It is expected that some technologies will change the global rules of the game."

Future MEMS - paper or plastic?

According to Fitzgerald, the first technology to rewrite the rules of the game will come from new uses for FBAR and surface acoustic wave (SAW) sensors.

Currently, FBAR and SAW technologies are mainly used for radio frequency (RF) filters. Fitzgerald said: "Based on literature data, they can also be used to produce passive sensors that do not require batteries; such battery-free sensors can still wake up the processor when certain parameters are reached." In addition, the sensor provides height. Accurate extreme temperature detection also works at pressure limits and can even detect specific gases.

"These passive sensors are ideal for harsh environments where you can't or can't replace batteries," she said. "And they also have high performance that provides zero standby power."

After further research on the 2017 MEMS literature, she also discovered near-zero power components, sometimes referred to as "event-driven" sensors. They are similar to passive components but use very small μA currents and consume less than 1pW in standby mode. When they sense that a particular event occurs, they wake up and trigger the application processor.

Figure 1: AM Fitzgerald and other MEMS chip designers can use 8, 46, 4, 3 or even 2 吋 wafers based on pure silicon or SOI (right to left)

For example, Fitzgerald said: "Northeastern University has proven that near-zero-power infrared (IR) sensors can achieve wavelength-sensitive functions and can wake up processors in Internet of Things (IoT) devices or security monitors. Even when applied to large arrays, they can still use small energy harvesting technology as a backup power source."

Many of today's new MEMS components use piezoelectric materials, not only for energy harvesting, but also for wide-range microspeakers, magnetometers, and even transformers that do not require authorization for high efficiency but are expensive. DRI process.

Fitzgerald said: "For low-cost devices and the Internet of Things, the consumer market is mature because it can be used once for mass production."

At the same time, MEMS researchers are working to explore ways to replace expensive silicon crystals. According to Fitzgerald, in 2004, 90% of MEMS components worldwide were fabricated using bulk silicon or silicon substrates; however, half of the next-generation components described in the literature are plastic or even paper substrates.

"Paper-based technology is increasingly replacing multi-billion dollar expensive silicon fabs, especially for disposable applications that use only once, and usually only require sensors that cost less than a cent." Or the components of the paper substrate are not as fast or precise as silicon-based components, but their performance is sufficient for consumer products that are used for short periods of time or frequently, as well as disposable disposable applications.

For example, paper sensors can be used to detect specific types of bacteria. These components can reduce the need for various antibiotics, especially since many antibiotics may contribute to the evolution of super bacteria. Similarly, paper-based food packaging can be embedded in paper-based components to inform consumers whether the food has actually deteriorated to replace the less accurate “expiration date” stamp.

Figure 2 The final MEMS chip produced by RVM (20 samples here) is ready to be supplied to customers.

Fitzgerald said: "After 2020, people will see a series of new sensors driven by piezoelectric events; and by 2030, we will see the growth of paper and plastic sensors."

She said that the CMOS+ sensor design with built-in readings still requires silicon. However, "as the research on silicon technology slows down and favors cheaper paper components, the risk of stagnation in silicon technology is there."