Navigation bar: Home > Miscellaneous Technology > Lithography Options 32nm and 22nm
Lithography Options for 32nm and 22nm
The following discussion is based on more in depth coverage in the forthcoming 2008 IC Technology report.
At the 90nm process node the 193nm exposure wavelength entered use. Based on the argon fluoride excimer laser the 193nm exposure wavelength required the development of a whole new family of photoresists, a huge industry wide undertaking. Once 193nm photoresists and exposure systems ramped up into production the industry had a strong incentive to continue to utilize the same technology for as many process generations as possible, for example the prior generation 248nm exposure wavelength was used for the 350nm, 250nm, 180nm and 130nm process nodes.
So called “dry”193nm exposure where air is used between the exposure system lens and the wafer has a theoretical resolution limit of approximately 52nm and with process improvements 193nm dry exposure was relatively easy to extend down to the 65nm process node.
At the 45nm process node a new approach was once again required. Most companies have turned to immersion or “wet” lithography where water is used in place of air between the exposure system lens and the wafer. The practical numerical aperture limit of an air based system is approximately 0.93 whereas a water based system can reach approximately 1.3. The net result is that a 193nm exposure system using water between the final lens and the wafer can achieve a resolution of approximately 40nm, good enough for the 45nm process node.
At the 32nm process node immersion lithography using water will not meet the required minimum resolution and there are basically three options:
The first option is to use existing immersion lithography technology and do double patterning. This technology exists today and is currently in production at two companies. The downside is cost; we estimate that immersion double exposure is 3 times the cost of a single immersion exposure.
The second option is higher index fluids in place of water. So called generation 2 fluids can reach a resolution of approximately 33nm and generation 3 fluids can reach approximately 31nm resolution. Unfortunately generation 2 and generation 3 fluids with the required supporting materials such as high index lens and photoresists are both unlikely to be ready in time for 32nm to ramp up in 2009.
The third option is Extreme Ultraviolet (EUV) that uses 13nm exposing wavelengths. EUV needs a new photoresist family and a lot of laser development is still required. So once again as with high index exposure EUV is unlikely to be ready for 2009. The bottom line is that at 32nm we expect double exposure with water based exposure systems to be the standard for critical layers.
At the 22nm process node there are only two main options although one option has two variants.
The first option is double patterning. There have been some claims that double patterning with water based exposure systems can meet the 22nm process node resolution requirements. Alternately double patterning with high index fluids could be used. High index fluids have a fairly good chance of being ready in the 2011-2012 time frame when the 22nm process node is expected to ramp up.
The second option is EUV. EUV would likely be less expensive than double patterning even with the expected high cost of EUV systems and EUV has a fairly good chance of being ready in the 2011-2012 time frame.
In conclusion we believe that double patterning will be used with water based exposure systems for the 32nm process node simply because no alternatives will be ready. For the 22nm process node EUV is the preferred solution with double patterning using water or high index fluids as the fall back. 32nm and 22nm process node lithography are discussed in much more detail in the forthcoming 2008 IC Technology Report.
Related Products
Copyright © 2000 - 2008 IC Knowledge - all rights reserved.
|
|||