12 June 2012
Will MEMS devices ever be the stars of the show?
Micro electromechanical systems, or MEMS, have been around for years. But despite the fact that the technology now plays a central role in a range of applications, it isn't what you might term a 'star'; in the Oscars, it might be a candidate for best supporting actor.
Karen Lightman, managing director of the MEMS Industry Group, said: "It's wrong to say MEMS hasn't broken through because it's an enabling technology. It might not be the 'bright shiny thing', but MEMS technology is saving lives around the world, playing a central role in air bags, tyre pressure monitoring and blood pressure monitoring. People don't realise where it's being used."
If one application has pushed MEMS towards star status, it's probably the iPhone. But the technology has also revolutionised movie projection through TI's Digital Light Processor. "And it enabled ink jet printing," Lightman continued.
Currently, the MEMS market is worth around $9billion a year, but is forecast to grow to $13bn by 2015. But it's a market currently dominated by TI, Hewlett-Packard, Robert Bosch and STMicroelectronics. Beyond that, it's a classic 'long tail', with some 300 companies developing products. Alissa Fitzgerald, founder of MEMS engineering services company AM Fitzgerald and Associates, said: "The top 30 companies generate 80% of the revenue. Beyond that, a range of companies are developing products for niche applications, including pressure sensors and microfluidics."
Lightman believes innovation in the MEMS market relates to the way the technology is applied, rather than development of the technology itself. "What MEMS is doing," Lightman believes, "is allowing designers to realise functionality which they have always wanted in their products. It's bringing a lot of things together, but there are challenges in terms of packaging and testing."
As with other technologies, the main market for MEMS is consumer electronics products. "Consumer electronics will provide the main growth for the near term," Lightman believes. One of the potential growth areas is health care. "Technologies which help people take care of themselves as they grow older – being able to monitor their health, for example – will be revolutionary and MEMS will help to do that." She also believes the broader medical market will see bigger growth in the long term. "And there's promise in areas such as robotics."
The fact that Intel is now involved with MEMS gives some illustration of the technology's potential. Steve Whalley is Intel's director of sensors. "We're interested in MEMS because we build full function reference designs; doing the 'heavy lifting' for our OEM customers. There's a wealth of MEMS devices available and we have to learn about them and see if we can add value beyond what's already there." Intel has been investigating MEMS devices since 2007, said Whalley. "It's mainly been from the conceptual point of view," he explained.
Fitzgerald believes the fact that MEMS is an enabling technology makes it harder to appreciate. "What makes MEMS difficult to characterise – and difficult to explain – is the fact that it isn't a device technology; it's a 'catch all' for a broad range of devices."
She believes that it's possible to aggregate around 20 different sensor technologies under the broad MEMS heading. "And the technology is applicable across a wide range of markets."
She's keen to point out that MEMS is anything but a new technology. "It's been around for many years; more than 20years in the automotive industry. Even though a top end car today may feature 40 MEMS devices, it's still an 'under the hood' technology. It's a component technology and will always be part of a larger product."
Whalley gave an indication of MEMS' potential in the consumer market. "With the advent of the Windows 8 operating system, computers are set to bring touch capability. That will be a new area for the computing market and we're showing OEMs how to put MEMS sensors into ultrabooks. It's a pioneering time because the industry hasn't done anything like this before. But we have to do it cost effectively."
He said any technology had to pass three basic tests before its use could be considered. "What's its form factor? How much does it cost and what is the power consumption? If we get satisfactory answers to those questions, we can start thinking about what we can do with the technology.
"But it's not just about the technology; we also have to explore areas such as sensor placement because it doesn't always make sense for them all to be in the same place."
MEMS has the benefit of not requiring bleeding edge manufacturing technology. "Moore's Law doesn't apply to MEMS," Fitzgerald asserted. "You can't shrink them because their functionality changes."
Peter Himes, vice president, marketing and strategic alliances, with MEMS foundry Silex Microsystems explained some of the process work underway. "MEMS has always been about process innovation to create new structures. A lot of what we do is to develop new capabilities which help our customers to make the best products.
"Our areas of research focus mainly on materials," he continued. "We're looking at how we can use piezoelectric materials for actuators."
While MEMS has been a standalone technology since its inception, work is underway to discover ways of integrating it into mainstream cmos manufacture. "The big foundries are showing more interest and are developing cmos MEMS platforms," Himes observed. "They are looking to take a cmos wafer and add modules on top, using the metallisation layers as a mask element.
"While the market is evolving, we believe a basic MEMS process will be needed for better performance. However, there will probably be some kind of hybrid."
But Whalley said this wasn't an area of interest for Intel. "They are two different technologies and incompatible. We have no plans to try to integrate them."
So innovation is focused on packaging. "Packaging has always been an area of innovation," Himes claimed. "At the moment, there's growing use of through silicon vias (tsv) to reduce package and footprint sizes and to enable device stacking."
Silex has been using interposers since 2006. "The interposer handles the tsvs and has bumps, so it can be mounted directly on a pcb," Himes added.
Like many other technologies, MEMS is being pushed by the mobile phone sector. "MEMS devices are smaller, have lower power consumption and can do things older technologies can't," Fitzgerald noted. "You can't put a mechanical component into a mobile phone." Himes agreed: "Much of the development is being driven by the mobile world."
But now MEMS devices are consumer products, Fitzgerald sees their effect being noticed by the user. "The average engineer now has a mobile phone with a motion sensor. They know there's 'something in there' and, hopefully, will think about how they can use the technology in robotics and machines. There's more awareness of MEMS now and it'll get parts in more movies," she concluded.