Extreme ultraviolet lithography (EUVL) is a leading next generation lithography technology. Significant progress has been made in developing mask fabrication processes for EUVL. The mask blank for EUVL consists of a low thermal expansion material substrate having a square photomask form factor that is coated with Mo/Si multilayers. SEMI standards are being developed for mask substrates and mounting. Several commercial suppliers are developing polishing processes for LTEM substrates, and they are progressing toward meeting the requirements for flatness, surface roughness, and defects defined in the a draft SEMI standard. One of the challenges in implementing EUVL is to economically fabricate multilayer-coated mask blanks with no printable defects. Significant progress has been made in developing mask blank multilayer coating processes with low added defect density. Besides lowering the added defect density, methods to reduce defect printability, such as defect compensation and buffer layer smoothing, are being developed. Experiments indicate that Mo/Si multilayers that are deposited with ion beam deposition tend to smooth substrate defects, and buffer layer films are being designed to enhance this effect. Targets for buffer layer smoothing are being defined using defect printability simulations. A method for using an electron beam to repair substrate defects after multilayer coating is also being investigated. The mask patterning process for EUVL is nearly the same as that for conventional binary optical lithography masks. EUVL mask patterning efforts are focused on developing the EUV-specific aspects of the patterning process. Eight absorbers have been evaluated against the requirements for EUVL masks, and two absorbers appear most promising. Conventional membrane pellicles are not practical for EUVL, so thermophoretic protection is being developed. Experiments have indicated that thermophoretic protection is effective for >125 nm particles down to at least 50 mTorr pressure. A removable pellicle will be used to protect the mask from defects at all times except during wafer exposure.

No abstract

Improvements in mask blank yield by the reduction in defects dramatically reduces the reticle cost of any lithography. Extreme ultraviolet lithography (EUVL) masks typically consist of a substrate coated with reflective multilayer (ML) materials (e.g., Mo and Si), followed by a sacrificial or “repair” layer (e.g., SiO2) and top absorber layer material (e.g., Cr or TaSiNx). Defects of the following two types exist: substrate and absorber defects. Substrate defects are generally below the absorber and buffer layers, i.e., at the substrate–ML interface or within the ML stack, whereas the absorber defects are located at the absorber layer. In addition, the printability of substrate-type defects may be reduced by coating the defects with a planarizing or smoothing layer. In this work, we discuss simulation and experimental results that compare the printability of programmed defect reticles with and without smoothing layers covering the programmed defect. We propose several criteria for smoothing layer specifications that are necessary to mitigate a size range of defects for an EUV imaging system.

Electron projection lithography (EPL) is one possible successor to conventional optical lithography. One type of EPL mask, a SCALPEL mask, consists of an array of rectangular membranes on a 200 mm silicon support wafer. An image of a die is formed by scanning and stitching the patterns resident on the membrane array. Key areas of concern for controlling the mask critical dimension (CD) include temperature uniformity during the resist post exposure bake (PEB) process, heating issues during resist exposure, fogging effects caused by electron scattering from the mask chuck and variations resulting from the pattern transfer of the mask scattering layer. A finite element model was used to evaluate heating issues during the mask writing step and PEB process. Masks were then written to verify the models. A Monte Carlo model was used to evaluate CD variations caused by electrons scattering from the chuck during the mask writing process.