Double exposure for the contact layer of the 65-nm node

Abstract

The critical dimension (CD) of contact holes for the 65-nm application specific integrated circuit (ASIC) is 100 nm according to the 2002 update of the International Technology Roadmap for Semiconductors. The common through-pitch depth of focus (DOF) of such contact holes is very small using the current ArF exposure tool. High-numerical-aperture (NA) ArF exposure tools are not expected to improve the common DOF that scales by the square of the numerical half aperture. High-transmission attenuated phase-shifting masks increase the DOF of isolated contact holes. Off-axis illumination such as annular or quadrupole illumination improves the DOF of dense contact holes. Nonetheless, both the isolated and the dense contact holes need to be printed within spec on logic circuit. To delineate 100-nm contact holes at several different pitches, we proposed the pack-and-unpack (PAU) process which employs double exposures. First, dummy holes are added to the surroundings of isolated contact holes facilitating the patterning of the resultant dense pattern with a resolution enhancement technique that favors dense contact holes. For example, dense holes are packed to 180-nm pitch and imaged with high-NA lens setting and quadrupole illumination. Then, the second image is used to open the desired holes or block the dummy contact holes. The purpose of this study was to develop new methods and new materials for the patterning of the second image. Three approaches were investigated. The first approach was forming an isolation layer to protect the first image; second, applying UV curing to harden the first image; third, using alcohol-based resists to pattern the second image. Among those three approaches of printing the second image, using resist in alcohols is the most convenient method. Even though the CD control of the second image is not so critical, resolution and process window of resists may need further improvement for 45-nm node and below. Using the second approach allows conventional ArF resists, which does not raise as many concerns as the alcohol-based resists. With the first approach, a lot more work is needed to prevent intermixing and reactions between the isolation layer and the resist for the second image. The results of this work point to the directions for material developments of the PAU process. Both the alcohol-based resists and UV curing are good approaches for PAU. Further characterizations such as DOF, exposure latitude (EL), and mask error factor (MEF) on them will be carried out in the near future.