A new concept for proximity effect correction (the representative figure method) is proposed in order to greatly reduce the correction time. The method assigns only one representative rectangle within a small area whose size is sufficiently smaller than the backscattering range of electrons. Proximity effect correction is carried out by using representative figures instead of the original patterns. This method eliminates the pattern density dependence of the correction time because the pattern density of the representative figures does not exceed the number of small areas. The relative error caused by the representative figure method is at most ±1% for a Ghost-scheme exposure and ±3-4% for the dose correction method.

We construct a kink solution on a non-BPS D-brane using Berkovits' formulation of superstring field theory in the level truncation scheme. The tension of the kink reproduces 95% of the expected BPS D-brane tension. We also find a lump-like solution which is interpreted as a kink-antikink pair, and investigate some of its properties. These results may be considered as successful tests of Berkovits' superstring field theory combined with the modified level truncation scheme.

No abstract

No abstract

The generation of a triangular and rectangular shaped beam is very useful in increasing throughput for writing ULSI patterns which often include many oblique lines. To make use of these shaped beams in ULSI pattern formation, a new rectangular and triangular shaped beam calibration method has been developed on the EB exposure system EX-7. The shaped beam calibration method is established by analyzing the beam current of shaped beams and the backscattered electron signal from a fine gold particle on the target. Resultant accuracies were 0.013 µm for a beam size of 1.6 µm and 0.025 %micro;m for the relative beam position on the target. Using this method, 0.2 µm ULSI patterns including oblique lines have been accurately formed.