Scientists from Dow Corning Electronic and IBM have created a new flexible polymer material which can be used to make optical waveguides on printed circuit boards that will enable supercomputers to perform faster while using energy more efficiently.
The material’s chief advantage is its high reliability. Waveguides made with the material can deliver sustained good performance past 2,000 hours under high humidity and temperature over 500 thermal cycles between -40 degrees centigrade and 120 degrees centigrade, according to Dr. Bert Jan Offrein, manager of the Photonics Research Group at IBM Research.
“The robustness and flexibility of the new material make it and ideal replacement for traditional copper waveguides,” said Simon Jones, business builder for Dow Corning. “It has the potential to drastically reduce the cost of circuit board production and at the same time enable supercomputers to perform at faster data rates while consuming less energy.”
RELATED CONTENT
A non-super chip for a supercomputer
Big Data pushing faster, better decision
With exabytes of structured and unstructured data growing annually at a rate of 60 per cent, he said, there a big push from Big Data users to cut down the energy required to move data from the processor to the printed board within a computer. Dow projects that machines using the new polymer will likely roll out by 2016.
Conventional copper waveguides using electrical signals are prone to producing “crosstalk” or signal interferences that radiates from one copper link to the one adjacent to it. This slows down performance and eats up energy. This, Jones said, does not occur with silicone-based waveguides using light signals.
Another advantage of the material is its manufacturability. The silicone can survive the manufacture process better than most materials and its flexibility allows it to be easily shaped and made to fit various patterns.
These properties also made the polymer ideal for 3D assembly.
“This means manufacturers can stack more boards on top of each other, essentially allowing them to pack more data density in a smaller space,” explained Jones.