Nanotechnology Now

Our NanoNews Digest Sponsors
Heifer International



Home > Nanotechnology Columns > Martini Tech Inc. > The long-term implications of the Canon-Molecular Imprints deal

Martini Tech Inc
Martini Tech

Abstract:
The deal, announced about two months ago, has the potentiality of re-shaping the whole nanoimprint industry by moving its focus from niche applications to IC-patterning

April 20th, 2014

The long-term implications of the Canon-Molecular Imprints deal

About two months ago, important news shook the small but lively Japanese nanoimprint community: Canon has decided to acquire, making it a wholly-owned subsidiary, Texas-based Molecular Imprints, a strong player in the nanotechnology industry and one of the main makers of nanoimprint devices such as the Imprio 450 and other models.

Molecular Imprints` devices are aimed at the IC patterning market and not just at the relatively smaller applications market to which nanoimprint is usually confined: patterning of bio culture substrates, thin film applications for the solar industry, anti-reflection films for smartphone and LED TV screens, patterning of surfaces for microfluidics among others.
While each one of the markets listed above has the potential of explosive growth in the medium-long term future, at the moment none of them is worth more than a few percentage points, at best, of the IC patterning market.
The mainstream technology behind IC patterning is still optical stepper lithography and the situation is not likely to change in the near term future.

However, optical lithography has its limitations, the main challenge to its 40-year dominance not coming only from technological and engineering issues, but mostly from economical ones.
While from a strictly technological point of view it may still be possible for the major players in the chip industry (Intel, GF, TSMC, Nvidia among others) to go ahead with optical steppers and reach the 5nm node using multi-patterning and immersion, the cost increases associated with each die shrink are becoming staggeringly high.
A top-of-the-notch stepper in the early 90s could have been bought for a few millions of dollars, now the price has increased to some tens of millions for the top machines
The next big thing of the optical stepper industry, that is EUV technology steppers, has been in the roadmap of ASML since 1995~1996 and was assumed to become a viable technology for the 65nm logic manufacturing node.

EUV stepper development has been far more challenging than previously thought and ASML still has to deliver high volume manufacturing solutions using EUV, the main stumbling block being the lack of a viable high-power source
Even if HVM EUV steppers may one day see the light, early estimates would put their price in the range of 100~150 million dollars, more or less the price of a medium-sized jet.
It is no wonder, therefore, that the industry is looking for cheaper alternatives.

Optical lithography still has enormous advantages if compared with other alternatives: non-contact between the mask and the substrate ensures that masks can be used for a long time before a replacement is needed, the pattern on the masks can be created as needed by the circuit designers (even if with multiple patterning some limitations may apply) and optical stepper lithography is a proven, reliable technology that has been used by the industry for decades.

Multi-beam technology, using over 10,000 E-beams simultaneously to pattern one single substrate, was heralded as the next big thing in the lithography industry as it was hoped it may have been able to overcome the main issue with traditional E-Beam lithography: the very long time needed to pattern a large area.
Companies like Mapper in collaboration with TSMC worked hard on the development of this solution but so far multi-beam technology has failed to live up the promises.

Direct self-assembly, another potential candidate for pattern fabrication at nanometer level, uses block copolymers to create patterns that self-assembly directly on the substrate surface. This technology is considered by many as very promising, but until now it is just little bit more than a laboratory experiment and cannot be employed in full production as structures still show an excessive variation in size and shape if compared with traditional optical lithography created patterns.

Other candidates for replacement of stepper lithography are there: one of the most promising is nanoimprint.
Nanoimprint is a relatively old technology as it was first demonstrated as viable by Prof. Chou back in 1996, but so far has only been employed in patterning of substrates for bio, anti-reflection films and other applications that do not require complex pattern shapes but only periodical patterns with lines-and-spaces and pillar/hole arrays.
Nanoimprint can be seen as the egg of Columbus of nanopatterning: instead of relying on complex technology such as steppers and E-beam, an imprint machine works very similarly to a press, using a mold as master for imprinting a pattern on a substrate covered with resist.
The cost of an imprint machine can be in the order of tens of thousands of dollars for a lab or small-scale manufacturing machine up to hundreds of thousands or a million dollar for a production scale machine, a small fraction of the price required for a stepper or an E-beam machine.

Molecular Imprints has been a company on the forefront of commercialization of nanoimprint-based solutions for IC manufacturing, but so far their solutions have yet to become a viable alternative HVM IC manufacturing market.
The main stumbling blocks for IC patterning using nanoimprint technology are: the occurrence of defects on the mask that inevitably replicates them on each substrate and the lack of alignment precision between the mold and the substrate needed to pattern multi-layered structures.
Therefore, applications for nanoimprint have been limited to markets where no non-periodical structure patterning is needed and where one-layered patterning is sufficient.
But the big market where everyone is aiming for is, of course, IC patterning and this is where much of the R&D effort goes.

While logic patterning with nanoimprint may still be years away, simple patterning of NAND structures may be feasible in the near future, and the purchase of Molecular Imprints by Canon is a step in this direction
Patterning of NAND structures may still require multi-layered structures, but the alignment precision needed is considerably lower than logic.
Moreover, NAND requirements for defectivity are more relaxed than for logic due to the inherent redundancy of the design, therefore, NAND manufacturing is the natural first step for nanoimprint in the IC manufacturing market and, if successful, it may open a whole new range of opportunities for the whole sector.

The technology development will also involve Toshiba and will be focused on 15nm patterning using Canon devices at Toshiba` s facilities in Mie prefecture.
The purchase of Molecular Imprints and the collaboration with Toshiba for developing 15nm NAND memory via nanoimprint manufacturing makes definite sense for Canon as the Japanese company is more or less cut off from EUV stepper development, a technology monopolized by ASML, and therefore has its only hopes in betting on a dark-horse technology such as nanoimprint to remain a big player in the nanopatterning business.
Whether the Japanese company will be able to deliver on promises remains to be seen.



Bookmark:
Delicious Digg Newsvine Google Yahoo Reddit Magnoliacom Furl Facebook

NanoNews-Digest
The latest news from around the world, FREE




  Premium Products
NanoNews-Custom
Only the news you want to read!
 Learn More
NanoStrategies
Full-service, expert consulting
 Learn More











ASP
Nanotechnology Now Featured Books




NNN

The Hunger Project