[extropy-chat] Current State of Photolithography

[Dan Clemmensen]

I decided to check this out to make sure that my recollections of my 
casual reading were approximately correct. Here is the best reference I 
found:
   http://www.icknowledge.com/misc_technology/ Immersion%20Lithography.pdf

Some minor corrections on my prior posts:
    Current technology is 193nm, as I said. This is generated by an ArF 
excimer laser, which I did not say.
The industry is struggling to move to 157nm. I incorrectly stated that 
it was 153nm. 157nm is generated by an Florine (F-F) excimer laser. I 
incorrectly said it was an ArF laser. The  Lens material for 157nm is 
calcium diflouride, not CaF as I said. The major stumbling block is 
manufacturing these lenses, as I said.

The paper gives some nice histories of the progressions in wavelengths 
and feature sizes.

The best thing in the paper is a formula:
      Width=KxL/NA.

Feature width decreases with Lambda (i.e., wavelength).
Feature width decreases as Numerical Aperture increases
Feature width decreases with the fudge factor K.

As discussed previously, pushing Lambda below 193nm looks really hard.

As a practical matter, NA cannot be pushed higher than .93 in air. The 
paper is mostly about "immersion Lithography," meaning that you fill the 
space between the lens and the substrate with water. With water, NA can 
be pushed to 1.47

The paper's description of K is hilarious. K is the factor that accounts 
for all the tricks other than NA that engineers and physicists have 
dreamed up to drive the width below the wavelength, including 
phase-shift masks. The paper asserts that the industry consensus is that 
K cannot be driven below .25.

With all of the above, the paper concludes that fairly straightforward 
extension of existing technology will get us to W=35nm, which is where 
the industry wants to be in 2008 or so.