Sam, Andrew, James and thread,
Some thoughts about DUV implementation and 0.18um:
1) DUV is not a new technology where chemistries and lasers are
not understood. It has been around for 10+ years and been used
in the "manufacturing" of 0.35um devices for many semiconductor
suppliers. It has now moved into the manufacturing of 0.25um
devices.
2) The issues surrounding DUV are not due to the "rush to market"
mentality, but rather have risen due to the fact that DUV is the
first wavelength of light that is targeted to resolve linewidths
at/and below the wavelength itself. In addition, we are building
lenses with numerical apertures that have never before been
attained. All this to push the "fundamental laws of physics" of
optics, so we don't have to deal with the alternative imaging
technologies (like xray and e-beam).
3) What's the bottom line? The bottom line is that the processes
and equipment required to achieve sub-wavelength resolution are
increasing in difficulty and complexity in order to stay on
the Moore's Law curve for performance improvements of IC's. This
means that the fabrication lines will require increased
discipline to utilize these new technologies to yield functional
ICs. Which means that the "square peg into the round hole" isn't
going to work anymore. The entire workforce/workplace must
improve it's own performance and methodologies to complement
and facilitate the newer processes and equipment.
4) As far as 0.18um goes, you can expect all the leaders to use
248nm (DUV) lithography for manufacturing. It is well accepted
amongst the leaders (IBM, Intel, AMD, Motorola, etc.) that 248nm
with off axis illumination, high numerical aperture lenses,
optical proximity corrected masks, and high gamma resists will
acceptably produce the required lithographic performance for
0.18um ICs. Shrinks of this technology will most likely use
193nm and the alternative technologies (i.e. Xray, SCALPEL, IPL,
EUV, and EBDW) will not be used until the 0.10um node.
Hope this helps some.
TS |
