MEMS sensors have come a long way over the past few decades. The late 1990s brought us the mass production of both MEMS accelerometers for automotive airbag crash sensors and MEMS gyros for rollover detection and anti-locking braking systems (ABS). In the early 2000s, MEMS sensors made the jump from automotive to mobile and consumer electronics, first with a MEMS microphone in the wildly successful Motorola RAZR phone and then with a MEMS accelerometer in the first Nintendo Wii remote.
Following this initial period of MEMS’ commercialization, the timetable for the mass proliferation of both MEMS and non-MEMS sensors accelerated dramatically. Just take Apple iPhone. Released in 2007, the first iPhone had one MEMS accelerometer and one proximity sensor. Released 10 years later, iPhone X included four MEMS microphones, a barometer, three-axis gyro, MEMS accelerometer, proximity sensor, ambient light sensor, and infrared (IR) sensor, magnetometer, and multiple image sensors. For perspective’s sake, well over two billion iPhones have been sold since 2007, making iPhone a major growth-driver in MEMS. According to Yole Développement[i] (Yole), MEMS will generate $10.9 billion in 2020 alone (revenue from non-MEMS sensors will be even higher), spanning automotive, consumer and mobile, Internet of Things (IoT), medical and healthcare, aerospace, industrial and other markets.
With so much growth behind us, what’s ahead? As executive vice president, Automotive Electronics at Robert Bosch GmbH, Jens Fabrowsky shared his insights on the future of MEMS during his MSEC 2020 keynote, The Next 10 Years of MEMS: An Outlook on Opportunities and Challenges. SEMI recently spoke with Fabrowsky about his October 15 presentation at SEMI’s first virtual MEMS & Sensors Executive Congress, October 6-8 and 13-15, 2020.
SEMI: What are some of the primary challenges facing the MEMS industry?
Fabrowsky: Development costs for new generations of MEMS sensors are increasing, leading to several major shifts. To compensate for rising development costs and reduce risk, MEMS sensors suppliers are pursuing wider, diverse markets instead of just targeting high-volume applications. At the same time, end-device manufacturers are demanding greater product differentiation, but they don’t want to pay a premium for it or wait for new hardware iterations. To stay competitive, sensor suppliers are providing software solutions that support new features and functionality. That approach is more cost-effective and speeds design-to-production cycles.
SEMI: What factors are increasing development costs for new MEMS sensors, and what can companies do to mitigate their R&D risk?
Fabrowsky: As with most electronic components, MEMS’ costs are driven by development and capital expenditures. The increasing complexity of the content, especially in interface ASICs and software, makes MEMS development a multidisciplinary feat, requiring several competencies across multiple design centers to meet ever-demanding timelines.
Manufacturing also plays a role. We often see dedicated manufacturing lines built for new MEMS products, which stresses both investments and capacity planning. Working together as an industry, we can reduce risk and costs by applying the same manufacturing process to more than one generation of products, which will speed time to market, increase volumes and improve ROI.
SEMI: To what degree will the COVID-19 pandemic continue to affect sensors suppliers?
Fabrowsky: MEMS manufacturing flows have been affected by disruptions in the supply chain. While the benefits of multiple sourcing and more direct ownership of the flow itself (on-shoring, vertical integration) have helped us, no one in the industry can claim they are out of danger, especially if a new wave of contagion occurs. Our industry relies heavily on just-in-time manufacturing and logistics, and we are all watching for influences that could alter flow. The pandemic has reminded us all that an important competitive advantage is a predictable, secure supply — which also comes at a cost that the end customer must value.
SEMI: Why and how are traditional hardware companies like Robert Bosch differentiating their platforms for end-device manufacturers?
Fabrowsky: On-shoring was already a trend before the pandemic. We’ve always believed in and are still investing in our own manufacturing facilities. That includes the 12-inch ASIC fab in Dresden, Germany, where we expect to manufacture future generations of power and control electronics to satisfy the growing appetite for silicon that vehicle electrification demands.
We think that one of our biggest differentiators is that our portfolio includes more than just components: Close collaboration with our internal partner divisions gives us comprehensive system know-how across the automotive supply chain. On the consumer-electronics side, we have extensive partnerships with makers of application processors, wireless systems, and sensor processing software. With this expertise behind us, we can provide flexible system-integration options to our end customers — who also benefit from a mature supply chain that supports high volumes and field-tested quality.
SEMI: What does customer demand for software solutions mean for sensor suppliers and how will suppliers evolve to meet this need?
Fabrowsky: In some silicon product business units, the R&D effort to develop software is higher than the effort to design the hardware! Software is not only what’s needed on the application layer. It also runs the interface to the processors – the drivers. In addition, increasingly complex testing software ensures high yield and minimizes defects.
On the application layer, we are increasingly using and promoting open-source platforms to encourage better collaboration throughout the ecosystem. In contrast, companies that charge fees to access their own proprietary software environments are missing the opportunity to remain competitive in the long run.
SEMI: Why are end-device manufacturers looking for plug-and-play solutions instead of standalone devices?
Fabrowsky: Consumers of electronic devices always want products with more features and lower prices. Their requirements produce a trickle-down effect that reaches all the way to component suppliers such as ourselves. This requires us to manage a healthy innovation pipeline and to choose products and technologies that promise growth and high volumes. This isn’t always simple, however, and many times the component itself is not enough.
Think of our Light Drive projector for Bosch Smartglasses. The only way we can hope to win designs in this market is by realizing a fully integrated module, with our own scanning mirrors and driver chips, as well as our integration of laser modules and the display system. This lets us offer an individually tested and calibrated end product ready for assembly.
SEMI: What would you like MSEC 2020 attendees to take away from your presentation?
Fabrowsky: We’ll be looking at what’s driving the next decade of MEMS applications. For example, the embedded computing inside the sensors, together with enhancements in integration, materials, and packaging, will increase the pervasiveness of MEMS sensors and actuators as touchpoints between electronics and the physical world. This will create a new form of intimacy between us and the machines, which we call Artificial Empathy.
Robert Bosch GmbH is a longtime member of MEMS & Sensors Industry Group® (MSIG), a SEMI technology community that connects the MEMS and sensors supply network in established and emerging markets, enabling members to grow and prosper. Visit us today.
Nishita Rao is product marketing manager at SEMI.
[i] Status of the MEMS Industry report, Yole Développement, 2020.