“City of Things” to be venue of upcoming IMEC Tech Forum
That changing emphasis will be in evidence at the IMEC Technology Forum, which this year has different location, Antwerp, and takes place May 16 and 17 at the Elisabeth Center with an expected audience of 2,000 people.
IMEC’s remit is work on micro and nanoelectronic technologies that are somewhere between three to ten years out in terms of full commercialization. We asked Van den Hove what was leaving, and what was entering, IMEC’s target zone. For example, is extreme ultraviolet (EUV) lithography so close to market as to be of reduced interest to IMEC?
“We think EUV will enter manufacturing in the next two years but that does not mean that everything has been solved. The pellicle issue has not been completely solved; we need higher energy and we think the next stage is higher NA (numerical aperture). There are a lot of questions in EUV and it remains an important program for IMEC. We need to research the future of EUV and how to extend it to 3nm and beyond.”
What is happening on the lithography front is a winding down of research into 193nm wavelength optical lithography. IMEC will continue with multi-patterning research but it will gradually shift the emphasis to EUV, said Van den Hove.
Similarly, he said, in transistor structures the move is on from FinFETs to nanowires. “Our effort goes to nanowires, both horizontal and vertical which we think will come in after the 10nm node.”
Van den Hove’s discussion of the 10nm, 5nm and 3nm nodes and beyond – even though node definitions have become notoriously loose and arbitrary – would seem to suggest that lateral scaling in IC manufacture is alive and well. This contrasts with the views of some who have indicated that Moore’s Law is coming to an end and migrating to future nodes, while being technically possible, may become economically undesirable.
Next: Moore’s law not ending
Van den Hove said: “I wouldn’t say Moore’s Law is coming to an end. True geometry scaling gets harder and there may be a slowing down in moving from one node to the next but other technologies will compensate such as 3D stacking. In memory 3D-NAND flash is already here. But we are not saying scaling is going to stop.”
It is clear that the silicon semiconductor industry has seen an explosion of complexity in terms of methods of scaling, integration and materials. That produces an even broader front that IMEC and other research institutes are needed to investigate. Silicon itself is becoming the carrier medium with different materials being inserted into the transistor channel to provide optimized switching or memory performance.
“Clearly silicon-germanium and germanium is a given, but we are looking at III-V materials in vertical nanowires, magnetic materials in the memory space and possibly spin-torque transfer gates for logic,” said Van den Hove. There are some more advanced molecular and atomic possibilities that could enable extreme density. “We are starting to look at organic memories such as DNA but it’s a long way out,” he added.
Other things that have moved into IMEC’s area of interest include neuromorphic and quantum computing and van den Hove said that silicon will be the platform on which these concepts will be introduced to the market, but again with specialized architectures and materials. “For example, we see that spin torque transfer MRAM and resistive RAM and their underlying materials systems may be important for neuromorphic computing.”
Van den Hove added that machine learning and neuromorphic computing would have a strong presence at the upcoming IMEC Technology Forum in Antwerp.
Next: More design
It can be argued that with the slowing down of the pace and increased cost, of node-to-node transitions, industry will strive to gain more value from R&D on design. Van den Hove agreed and said: “We have kept a strong design competence at IMEC and that has been demonstrated in wireless programs, in RF design. We are increasing the design activity with regard to new computing concepts but now we have to think more at the systems level.”
“It is not so much about design rule optimization but how can we design systems in a more effective way. That is true for both high-end computer systems and for lean IoT devices.”
And IMEC has also been increasing its fields of endeavour with work and collaborations into sensors, life sciences, medical electronics and sustainable electronics. “The opportunities are endless so we need to specialize. We should pursue those applications where we can leverage our high-end technology expertise in photonics, in microfluidics and so on.
IMEC’s expanding horizon was one of the reasons it chose to merge with iMinds, a Belgian digital research and incubation center, in 2016. This move brought together the technology expertise of 2,500 IMEC researchers worldwide and the digital and software competencies of some 1,000 iMinds researchers drawn from nearly 50 countries. The iMinds network includes the iStart entrepreneurship program supporting start-up businesses. Creating spin-off businesses that, in turn, create high-value, wealth-generating employment is part of the remit of IMEC which receives financial support from the Flemish government.
“We merged with iMinds as they specialized in IoT networking and have groups working on artificial intelligence, security, privacy and encryption,” said Van den Hove. These are all vital ingredients to the smart cities, smart homes, smart vehicles and smart health that people expect technology to provide in the future, said Van den Hove.
Next: Flanders, Europe, the world
The iMinds group also had locations in Ghent and Antwerp and an even more distributed research model which brought in skills from many locations. One of the collaborative projects undertaken by the newly enlarged IMEC is the “City of Things” smart city project with the city of Antwerp and the Flanders government. Between 2017 and 2019, the Antwerp city authorities intend to invest €650,000 in IoT sensor networks to improve mobility, security, sustainability and digital interaction with citizens.
This all provides good reasons to stage this year’s IMEC Technology Forum in that city.
Europe has acquired a reputation in electronics over the last few decades. On the one hand Europe punches above its weight in applied research and attracts global players to collaborate, largely due to such institutes as IMEC, Leti in Grenoble and the Fraunhofer Institutes across Germany. On the other hand, Europe fails to invest sufficiently heavily or effectively in commercialization – both within established companies and at startups – to take the benefit of this excellent research. Ultimately this has allowed companies based in other continents to capture the opportunities to generate wealth and profits, particularly in the area of semiconductors.
Does Van den Hove see it that way?
“We consider ourselves a global citizen but we want to have an impact in the region, in Europe and the world – in that order. There are lots of opportunities in Europe; in health, energy, in IoT, agriculture. There are lots of SMEs [small- and medium-sized enterprises] and lots of opportunities for new initiatives not dominated by big companies. IMEC should be a platform to allow low-barrier access to new technologies. In Europe let’s focus where we can be excellent.”
With excellent research institutes within Europe and events of global significance such as the IMEC Technology Forum happening here, there is an opportunity to make the best of this situation. In any case, it is expected that many executives and researchers from both inside and outside Europe will be making their ways to Antwerp in May.
imec and iMinds merge to focus on the digital economy
IMEC leverages Si platform toward quantum computing
IMEC reports on 8nm MRAM device
Imec stacks Gate-all-Around Si nanowires vertically in CMOS MOSFETs