Skip to Main Content
Your organization might have access to this article on the publisher's site. To check, click on this link:http://dx.doi.org/+10.1116/1.583311
One of the difficulties found in working with the positive e‐beam resists in the problem of controlling the linewidth of critical geometries using a fixed time development process. Using the method of interrupted development, where development continues until a particular pattern ‘‘looks right,’’ only somewhat improves linewidth control. Furthermore, the improvement in the sub‐half‐micron domain remains questionable. Another approach, given here, correlates the disappearance of small islands of unexposed resist (in a ‘‘sea’’ of exposed resist) to the critical dimension (CD) of geometries in the sub‐half‐micron region. This approach, referred to as SCALES, is a technique that can use an optical microscope for linewidth determination. This, in turn, aids in linewidth control. The existence of these small islands and their subsequent disappearance in the development process can be seen quite easily under the optical microscope. Correlation studies were done comparing the sequential disappearance of the small islands with the CD’s of isolated single‐pass lines exposed in polymethylmethacrylate (PMMA). Both random and systematic errors and their sources are discussed and experimental results given. Results based on one carefully controlled in‐house e‐beam machine yielded a standard deviation (sigma) of less than 0.03 μm. On another machine, the sigma was found to be 0.04 μm. Here, dosage was varied from 40 to 100 μC/cm2, two types of resist were used, and the spot size varied from 0.24 to 0.27 μm during the study. SAMPLE simulations were used to help illuminate certain trends in CD variations due to systematic variations in resist thickness, dosage, and spot size. For comparison, simulations were done for a fixed time development. For the three cases observed, the interrupted development procedure u- - sing SCALES showed a superior ability to track CD changes due to process variations over the fixed time development process.
Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures (Volume:4 , Issue: 1 )
Date of Publication: Jan 1986