MIT researchers developed a new technique that allows chip structures to “self-assemble” at microscopic levels. The technique is said to be more cost-effective than extreme ultraviolet (EUV) lithography, another technique used in chip fabrication that’s expected to be used in 7nm and 5nm chips around 2020 or later.
Difficulties At 7nm And Below
Moore’s Law has slowed significantly over the past few years, and we’re now seeing a process node last 3 years, if not longer. It’s becoming more difficult, as well as more expensive, to make transistors smaller as we approach the size of atoms.
EUV lithography has been researched for a long time because it uses tiny 13.5nm extreme ultraviolet wavelengths that can help create those smaller 7nm or 5nm transistors. However, some chip makers, such as TSMC, have complained that using EUV soon would add too much to a chip’s cost, so it’s not yet too practical. Others, such as Samsung, expect to start using it on their 7nm process by 2020.
Direct Self-Assembly Technique
A team of MIT researchers have developed a new technique called Direct Self-Assembly (DSA). This technique may help chip makers, especially those skeptical about EUV’s practicality in the short-term, to create 7nm or smaller chips.
DSA uses a novel integration of three existing methods. First, a pattern of lines is created on the chip’s surface using well-established lithography techniques, such as using an electron beam to print the chip’s pattern.
The second step involves creating a layer of material, called a block copolymer, that’s made up of a mix of two different polymer materials that naturally separate into alternating layers or other predictable patterns. The block copolymers contain chain-like molecules, where each consists of two polymer materials connecting end-to-end.
Another protective polymer layer is then added on top of the previous layer using chemical vapor deposition (CVD). The top layer is what forces the block copolymers to self-assemble into vertical layers, rather than horizontal ones.
The MIT researchers said the underlying lithographic pattern determines the positioning of these layers, but the natural tendencies of the copolymers is what causes their width to be much smaller than that of the base lines.
The result is four or more lines, each of them a fourth as wide, in place of the original pattern line. The combination of the lithographed layer and the top layer is what controls both the orientation and the alignment of the resulting finer lines, according to the researchers.
The top layer can also be additionally patterned, so the system can be used to build complex patterning, as needed for the interconnections of a microchip.
According to the MIT team, the new process can be implemented using existing machinery with only small changes. This could make the technology more appealing to chip manufacturers, compared to using new lithography techniques to print much finer patterns.
“Being able to create sub-10-nanometer features with polymers is major progress in the area of nanofabrication,” said Joerg Lahann, a professor of chemical engineering at the University of Michigan, who was not involved in this work.
“The quality and robustness of this process will open an entirely new area of applications, from nanopatterning to nanotribology,” he added.
One roadblock that the DSA technique could face is the fact that the lines formed by the chemical reactions may form structures that are too regular. The DSA technique may not work very well for chips with more irregular patterns, such as CPUs, but may work well enough for chips with much more regular patterns, such as memory chips.