Resolution and total blur: Correlation and focus dependencies in e-beam lithography

Abstract

The resolution limit in e-beam lithography is dependent on the resist process, the proximity effect, and the e-beam writer itself. However, it is difficult to distinguish these contributions just by comparing critical dimensions (CDs) of isolated and dense lines. Since the calculation of the parameter α of the point-spread function is usually based on the CD measurement of resist structures, α should include all mentioned contributions. Therefore, the total blur αtotal can be defined as the square root of the squared sums of αprocess, αforwardscattering, and αbeamblur. With keeping αprocess constant, the authors compared Gaussian beam (GB) and variable shaped beam (VSB) writers via the isofocal-dose method and proved that the isofocal dose is independent of the beam profile. For further comparison, they enhanced the method to the linewidth roughness and the total blur. For the latter, a special frequency modulated pattern was applied, which enables the determination of αtotal in dependency on the focus. The analysis of the total blur in best focus and the resolution of dense lines, exposed in a negative chemically amplified resist on GB and VSB writers, indicates their proportionality, which is supported by the same analysis with a nonchemically amplified resist. The resolution of the VSB writer cannot be directly compared to the GB writers’ resolution since the beam blur of the former changes with the shot size.