The marketplace for MEMS foundry services is getting increasingly more competitive and globalized. Recently, large companies such as TSMC and GLOBALFOUNDRIES have been more aggressive in the MEMS space and, perhaps, are even subsidizing some projects to get more experience with the relevant process technologies. At the same time, intellectual property (IP) protection is becoming even more important due to the commoditization of some MEMS devices, especially in the consumer electronics segment. We recently spoke with Edvard Kälvesten, CEO of Silex Microsystems, to discuss his vision for the company, some of the recent changes that were made, his main concerns at this time, and the upcoming milestones for the company in the next few years.
MEMS Journal: What's your long term vision for Silex? How do you see the company in five years?
Edvard Kälvesten: We believe the MEMS market is now moving into the second wave of adoption and we see a larger crowd of engineers in various industries applying the potential of this rather new engineering discipline to create innovative solutions in the micro scale technology space that were not possible before.
Silex’ truly pure play MEMS foundry offer attracts the innovation leaders looking to create novel and ground breaking micro scale solutions. We have the most solid track record of protecting our customers IP and technology while achieving the greatest results in quickly taking customers’ unique product designs to volume manufacturing.
This track record, combined with our process technology modules in the TSV interconnect and wafer level packaging space, puts us in a good shape to serve this next wave of MEMS foundry demand. We built our 200mm fab in 2009 and I am hoping that five years from now we will bring up our third manufacturing line.
MEMS Journal: Silex' former CEO and VP of marketing recently left the company. Why were these moves made?
EK: With our headquarters based in Stockholm, Sweden, our former CEO decided after his tenure of 3.5 years with Silex that monthly intercontinental travel away from his family and home in California was taking too much of a toll on his personal life. With this change at the end of 2013, it was a natural move to also consolidate the marketing function back to the headquarters.
MEMS Journal: What are your plans for 2014? What are the major milestones on the agenda?
EK: We are continuing our dedication to a strict pure play MEMS foundry offer; we do not develop any application specific product or design IP. That said, Silex’ dedication to the true pure play model and protection of customer IP does not imply staying shy of advanced technology development. But our technology nodes must at all times stay clear of competing in the actual application space of the foundry customers.
Examples of such generic process technology nodes are technologies pertaining to the packaging and interconnect space such as TSVs (Sil-Via® and Met-Via®), interposers, IPDs, wafer level packaging, and wafer level integration of MEMS and CMOS. Other areas where Silex is offering leading production solutions to its foundry customers are in the processing capability of novel materials such as AlN, PZT, thick polymers, magnetic materials and a wide range of alloys and noble materials.
The key milestone for Silex in 2014 is to successfully launch a through glass via technology (TGV). Silex assumed a leading position in the TSV space when we launched our first TSV technology (Sil-Via®) back in 2005. It has since become the widest adopted TSV technology in the MEMS space. As a foundry service provider to the world’s leading innovators, it is important to always stay at the forefront of technology innovation and be in position to offer leading edge process technologies that help our customers to create the most competitive designs and product solutions.
MEMS Journal: What are the most interesting trends that you are seeing with 3D integration?
EK: We see significant interest for low entry barrier solutions to 2.5D system integration. Many chip companies cannot allocate million dollar budgets to design custom ICs, so for these companies an affordable 2.5D interposer solution will allow them to select and combine standard technologies with the best possible value using chips in different nodes for IO, logic and memory.
MEMS Journal: Are you seeing demand for your TSV process technology for non-MEMS applications? If so, what types of applications?
EK: Yes, the 2.5D interposer platform we are developing this year is not really a MEMS application even though it rests on our ability to process and handle mechanically thin and structured wafers. The technology builds on our thick wafer metal via technology and we are working with an IC foundry partner to combine our robust metal via technology with a 4-layer copper damascene process. With this approach to form a 2.5D interposer, we avoid the need for complex temp-bonding, de-bonding and also the need for an intermediate BGA laminate. At the top layer of the interposer, we enable micro bumps with 2µm line/space and at the bottom of the interposer our silicon substrate has a PCB pitch of down to 400µm. Most of the high speed signals will run in the 4-layer copper damascene layers on the top side.
MEMS Journal: Aside from introducing the through glass via (TGV) technology, what are some of the improvements for your existing TSV processes that you will be introducing this year?
EK: A key requirement to enter the 2.5D interposer market is to offer the copper damascene process required for the signal routing. Our current TSV technology is good enough, so the focus in this year will be on the integration of copper processing on top of our metal via wafers.
MEMS Journal: What's the main threat to Silex' current business?
EK: One of our main challenges is a perception issue within the investor community where, quite often, we hear from our customers that their Board would expect them to work, for costs reasons, with the larger Asian IC foundries. Some of our customers will need to argue with their Boards that labor cost is typically only between 8-10 % of the manufacturing cost. On the other hand, we offer a world-leading MEMS team that can build wafers at high yields on a dedicated 200mm wafer MEMS volume production line. Therefore, we believe that we would often be a more suitable partner than a converted CMOS fab in Asia.
It is hard to write a “blue book” on making MEMS, so one of the most important qualities of a MEMS foundry services provider stems from their ability to attract and retain top expertise all the way from engineering to technicians in the fab. Silex has been active in the past 13 years, and even we still have some ways to improve in terms of refining and defining engagement models, and expectations for working with a MEMS foundry.
The global marketplace for MEMS foundry services is quite mixed today, where one has everything from product companies offering spare capacity as MEMS foundry services, to mixed model foundries offering turn-key services including product design and product platforms, all the way to Silex’ pure play model with a strict focus to fully protect customer IP. I believe that, with the next wave of MEMS adoption, the engagement principles and preferences will be better understood in the industry as a whole.
MEMS Journal: Large semiconductor foundries such as TSMC have been much more active in the MEMS technology sector recently. This has resulted in their taking away business from mid-size MEMS foundries. Do you believe that companies such as TSMC will be a threat to take the most lucrative and profitable MEMS projects from the smaller foundry players?
EK: I would say that Silex’ true pure play model plays a key part in the overall picture. Large semiconductor foundries such as TSMC are, by definition, not pure play MEMS foundries; they are giant IC foundries and, to our knowledge, act more like IDMs in the MEMS space.
We may have reasons to believe that some large semiconductor foundries are temporarily subsidizing their MEMS foundry efforts to learn MEMS, causing a negative impact on some of the smaller foundries. We do, however, see evidence that prior MEMS engagements with large semiconductor foundries has generated the emergence of corresponding “product platforms”, at least in the microphone and inertial sensor space. We are convinced that a typical Silex customer will try to avoid these types of situations.
True innovation leaders need to make sure they do not enable their competitors too easily in their market and Silex’ true pure play foundry model grants its customers a strong competitive advantage, compared to the approach of the larger semiconductor foundries in the MEMS space. Silex’ mission is to be the pure play and IP-safe foundry service provider for innovation leaders in the MEMS space, just like we believe TSMC once was to the IC industry.
MEMS Journal: What are some of the key industry initiatives or groups that Silex is a part of?
EK: Silex has been an active member of the MEMS Industry Group for many years. We are also quite active in forums pertaining to our focus technologies in the through wafer interconnect and wafer level packaging space.
This year we will launch our low resistivity silicon based 2.5D interposer solution combined with a fine pitch copper damascene process, so we expect to significantly pick up our activity in the 3D integration forums. Additionally, this year we are also launching our first through glass via (TGV) technology in response to the need of our most demanding RF customers. We expect to become much more engaged in the RF and microwave technology community going forward.
MEMS Journal: Which industry events did you personally attend in the past year? Which one was your favorite and why?
EK: I have not personally been a very frequent visitor of the industry events, as we normally cover those with our sales and marketing teams. In March of this year, however, I gave a presentation at the Global Semiconductor Forum in Singapore where I was given the opportunity to speak about the latest advances with our foundry service and process technology offerings.
This article is a part of MEMS Journal's ongoing market research project in the area of MEMS fabrication and manufacturing technologies. If you would like to receive our comprehensive market research report on this topic, please contact Dr. Mike Pinelis at firstname.lastname@example.org for more information about rates and report contents.
Copyright 2014 MEMS Journal, Inc.