News & Analysis
Lithography experts remain skeptical about EUV
R Colin Johnson
12/17/2012 12:01 AM EST
Long immersion history
Immersion lithography was so quickly adopted because it was based on the proven principles of immersion microscopy that stretched back to the 1600s when the English natural philosopher Robert Hooke predicted it. In the 1800s it was demonstrated by Giovanni Battista Amici an Italian astronomer and microscopist. And in the 1900s it was perfected to a science for microscopy.
The principle is that light bends at the interface to a liquid medium--the imitable bend stick in a glass of water--thus magnifying a microscope's image as it comes up through an immersed lens. Likewise, when light passes down through a lithographic reduction lens immersed in a liquid, it shrinks an image by its index-of-refraction. Today we know that combining immersion lithography with multiple-patterning--splitting up a mask into parts that can be exposed in separate steps--the resolution of standard 193-nanometer lithography can be extended to 32-nanometer. And with more sophisticated multiple patterning, and higher index fluids, 193-nanomoter lithography could go all the way down.
"Intel has already achieved the 32-nanometer node with triple-patterning, and many engineers are talking about making it all the way to 14-nanometer with multiple patterning," said Blatchford. "There are also some other tricks you can play with pitch doubling, making it feasible to achieve 10-nanometer with immersion lithography."
EUV is still being developed, and many semiconductor houses have expressed a willingness to move to EUV when it becomes available, but others are now predicting that immersion lithography, multi-patterning and high-index fluids, will allow semiconductors to make it all the way to the end of the International Technology Roadmap for Semiconductors at eight-nanometer using 193-nanometer lithography.
"Its hard to say whether some revolutionary new architecture will evolve that enables scaling of silicon beyond the end of the current roadmap," said Blatchford. "But Intel has publicly said that even if EUV never works, it will be able to make immersion lithography work all the way to the end of the current roadmap."
Nikon, Canon and ASML have all been hard at work trying to make EUV work for almost a decade, since it uses light wavelengths as short as 10-nanometer and thus theoretically could enable sub-five nanometer features only a few molecules wide. However, by then carbon-based electronics may start us down a new roadmap altogether--one that discards lithography's subtractive masks, in favor of additive self-assembly that begins with individual atoms from the bottom-up.
Immersion lithography was so quickly adopted because it was based on the proven principles of immersion microscopy that stretched back to the 1600s when the English natural philosopher Robert Hooke predicted it. In the 1800s it was demonstrated by Giovanni Battista Amici an Italian astronomer and microscopist. And in the 1900s it was perfected to a science for microscopy.
The principle is that light bends at the interface to a liquid medium--the imitable bend stick in a glass of water--thus magnifying a microscope's image as it comes up through an immersed lens. Likewise, when light passes down through a lithographic reduction lens immersed in a liquid, it shrinks an image by its index-of-refraction. Today we know that combining immersion lithography with multiple-patterning--splitting up a mask into parts that can be exposed in separate steps--the resolution of standard 193-nanometer lithography can be extended to 32-nanometer. And with more sophisticated multiple patterning, and higher index fluids, 193-nanomoter lithography could go all the way down.
"Intel has already achieved the 32-nanometer node with triple-patterning, and many engineers are talking about making it all the way to 14-nanometer with multiple patterning," said Blatchford. "There are also some other tricks you can play with pitch doubling, making it feasible to achieve 10-nanometer with immersion lithography."
EUV is still being developed, and many semiconductor houses have expressed a willingness to move to EUV when it becomes available, but others are now predicting that immersion lithography, multi-patterning and high-index fluids, will allow semiconductors to make it all the way to the end of the International Technology Roadmap for Semiconductors at eight-nanometer using 193-nanometer lithography.
"Its hard to say whether some revolutionary new architecture will evolve that enables scaling of silicon beyond the end of the current roadmap," said Blatchford. "But Intel has publicly said that even if EUV never works, it will be able to make immersion lithography work all the way to the end of the current roadmap."
Nikon, Canon and ASML have all been hard at work trying to make EUV work for almost a decade, since it uses light wavelengths as short as 10-nanometer and thus theoretically could enable sub-five nanometer features only a few molecules wide. However, by then carbon-based electronics may start us down a new roadmap altogether--one that discards lithography's subtractive masks, in favor of additive self-assembly that begins with individual atoms from the bottom-up.
|
|
|
|
|
Navigate to related information