MEDEA+ Investigates Mask Fabrication for EUVL

Extreme ultraviolet (EUV) lithography is now the most favored technique for production of devices with ultrafine linewidths where 157 nm radiation from fluorine lasers cannot provide the required resolution, even with full use of enhancement techniques.

It seems that EUV lithography will be required before the end of the decade, when circuit features have dimensions of 50 nm or less. It is estimated that EUV is capable of pattern transfer with a resolution down to 15 nm. Ion projection lithography is being developed as an alternative, but it has a much lower level of acceptance in the market. The EUV technique is expected to move from the R&D stage into commercial development.

Under the auspices of the pan-European collaborative venture MEDEA+ (Microelectronics Development for European Applications), researchers are starting to develop new techniques required for EUV lithography. This project, designated T404, is known as EXTUMASK (Extreme UV Lithography Masks).

At EUV photon energies, it is no longer possible to use conventional transmissive optical elements because all materials absorb at such wavelengths. One of the most difficult problems seems to be the development of suitable masks whose performance will be acceptable in an industrial environment.

Conventional optical lithography is performed with quartz masks, but at EUV wavelengths the masks will have to be based on a highly reflecting multi-layer mirror. New maskmaking techniques will be required because the masks must meet the following conditions:

• The thermal expansion of the mask substrate material must be close to zero, and the flatness must be reduced by a factor of 10-100 relative to the mask substrates now in use.
• The change to reflective masks will require an extra defect-free multi-layer mirror for the mask itself. Such mirrors are required for the illumination and demagnifying optics.
• A usable reduction of 100-150 nm is required for mask patterning and characterization purposes.
• The allowable defect size must be <50 nm and maintained without pellicle protection.

To achieve these four objectives, the whole sandwich structure must be closely specified, and appropriate materials and coating techniques developed. In addition, the process and its final results must be monitored with suitable metrology tools.

The EXTUMASK project began in November 2001 and will run until October 2004. The overall goal of EXTUMASK is to prove the manufacturability of EUV masks by providing a solution in each of the problem areas or narrowing the choice of options for future work. It will provide an infrastructure for experimental EUV masks by 2004.

For additional information on materials science, go to www.semiconductor.net/materials.