MIT (the Massachusetts Institute of Technology) has publicly announced some of the results obtained working on DSA (direct self assembly) patterning technology in an article published on Nature magazine.
DSA is one of the promising technologies that may enable the electronics and chip industry to keep pace with the requirements dictated by Moore` s Law.
Current lithography technologies like immersion and multi-patterning lithography, EUV among others, are becoming increasing expensive and the industry is exploring alternative solutions.
Electron beam, nanoimprint are among the alternatives, but electron beam technology is simply too slow to pattern large areas with sub-20nm features. Nanoimprint has several issues with alignment of planes for patterns on different layers.
Among the candidates we have direct self-assembly, a technology by which a block copolymer self-assemblies itself in order to produce small patterns in the order of few tens of nanometers. The technology, until recently considered very far from practical applications in volume production for IC manufacturing, is however heavily studied and worked upon. Direct self-assembly is now on the ITRS roadmap as next-generation lithography candidate
While DSA has already been proved on 300mm-wafers, the patterns produced so far had not enough resolution and very limited control on geometry. One of the main issues with direct self-assembly is that pattern size is not strictly determined and instead tends to variate between a lower value and a higher value.
Differently from traditional lithography, where the pattern is carefully produced on a mask and then copied in a step-and-repeat fashion on several wafers and in an exact manner, direct self-assembly relies on a statistical approach to patterns. Whether this new way to produce patterns can be applied in an industrial scale for IC manufacturing remains to be seen.
MIT has developed a simple but effective method to create a pre-defined pattern over a large area using direct self-assembly with the help of electron-beam lithography to work as a trigger for the self-assembly process.
“We think that our current work will enable Moore’s Law to continue to be able to scale up transistors density and to help shrink feature size. Other patterning methods and technologies are available but the issue of the investments involved may prove prohibitive. Our research represents a possible future solution for this problem.” said Caroline Ross, MIT professor of Materials Science and Engineering.
Bob Havemann, Director of Nanomanufacturing Sciences at Semiconductor Research Corporation, a consortium of companies funding university research in energy-related areas such as photovoltaics, power electronics, energy storage and smart grid is also bullish on direct self-assembly as a potential enabler of Moore` s Law in the medium-long term. “The request for more computer power due to increasing bandwidth and more power-crunching applications will simply keep increase in the next years. While traditional lithography has so far been able to cope with such increase in demand, the cost of keeping this growth going has also increased exponentially and this trend will not be sustainable in the long term. Therefore, the industry has to look for alternatives. I believe the work done by MIT represents a step in this direction.”