Computer graphics processing and selective visual display system – Computer graphics processing – Three-dimension
Reexamination Certificate
1997-05-15
2001-08-07
Vo, Cliff N. (Department: 2671)
Computer graphics processing and selective visual display system
Computer graphics processing
Three-dimension
C345S440000, C250S492220
Reexamination Certificate
active
06271852
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a electron beam writing data creating device which can remarkably reduce generation of small graphic data causing deterioration of dimensional accuracy.
2. Description of the Background Art
Due to high fine workability and high controllability, electron beam writing is widely employed for preparing masks which are employed for fabricating semiconductor devices, particularly large-scale integrated circuits (LSIs). Electron beam writing systems can be classified into raster scan and vector scan systems. The raster scan system, which is widely employed for preparing masks in general due to simple device structures and easiness in creation of writing data, has such a disadvantage that the writing speed significantly depends on the minimum grid size (address unit size) for specifying the sizes of the written graphics and the writing positions. In practice, it may be impossible to write graphics in preparation of masks for a 64-Mbit dynamic random access memory (DRAM) or the like requiring small address units by the raster scan system, due to excessively long writing times.
A variable shaping system, which is one of vector scan systems, is recently watched with interest. The variable shaping writing system is adapted to shape an electron beam in correspondence to the sizes of written graphics for irradiating only necessary areas with the electron beam. Thus, the writing speed is increased, and the address unit size can be reduced. Thus, it is considered that the variable shaping system forms the mainstream of electron beam direct writing employed for preparing masks for devices following the 64-M DRAM and development of devices following the 1-G DRAM.
While the variable shaping writing system has the aforementioned characteristics, the processing for creating writing data in this system is disadvantageously complicated and requires a long processing time as compared with the raster scan system. The procedure of creating writing data employed for the variable shaping writing system and the writing procedure are now described.
FIG. 20
illustrates a part of design pattern data
20
for an LSI, which is generally expressed as a polygon (enumeration of vertex coordinates). In order to write the pattern data
20
by a variable shaping electron beam writing system, it is necessary to split the same into basic graphics such as rectangles, squares, triangles and/or trapezoids (hereinafter simply referred to as “trapezoids”).
FIGS. 21 and 22
illustrate the pattern data
20
split into trapezoids by horizontally and vertically inserting split lines from vertices of the pattern data
20
respectively.
First, the procedure of actually writing a graphic with the pattern data
20
shown in
FIG. 21
is described. When the pattern data of an area
20
a shown in
FIG. 21
is inputted in an electron beam writing system, an electron beam shaping deflector forms an electron beam having a width Ws
1
and a height Hs
1
.
Then, an electron beam irradiation position specifying deflector moves the beam to an irradiation position (x
1
, y
1
), for irradiating a resist film which is applied onto a sample such as a mask substrate or a silicon wafer with the electron beam by a time corresponding to an exposure necessary for photosensitizing the resist film. Then, the pattern data corresponding to an area
20
b
in
FIG. 21
is inputted in the electron beam writing system, which in turn executes electron beam writing system in a similar manner to the above. This processing is successively repeated as to all patterns of the LSI, thereby writing all patterns of the LSI.
Problems in creation of writing data through the aforementioned variable shaping writing system are now described.
FIG. 23
shows a resist pattern
21
which is formed on a mask by converting the pattern data
20
shown in
FIG. 20
to writing data, writing the pattern on the mask with the writing data, and developing a resist film. Note the pattern dimensional accuracy in a portion of this resist pattern
21
having a width Wm, and consider the factor deteriorating the accuracy. No factor related to pattern formation process conditions such as resist development are taken into consideration here.
When the pattern data
20
shown in
FIG. 20
is split along the horizontal split line as shown in
FIG. 21
, the dimension of the width Wm is influenced only by shaping accuracy of the electron beam which is shaped in response to the graphic of the area
20
a
shown in
FIG. 21
, i.e., the dimensional accuracy of the electron beam in the portion corresponding to the width Ws
1
.
When the pattern data
20
is split along the vertical split line as shown in
FIG. 22
, however, the dimension of the width Wm is influenced by irradiation positional accuracy of electron beams corresponding to areas
20
c
and
20
d
shown in FIG.
22
and the shaping accuracy. Thus, two factors additionally influence the resist pattern dimensional accuracy, as compared with the case of splitting the pattern data
20
along the horizontal split line shown in FIG.
21
.
FIGS. 24 and 25
show results obtained by forming a plurality of resist patterns with writing data prepared by different split methods shown in
FIGS. 21 and 22
and measuring the dimensions thereof respectively.
FIG. 24
shows the result in case of employing the writing data shown in
FIG. 21
for forming the resist pattern of the portion having the width Wm by one-shot electron beam irradiation.
FIG. 25
shows the result in case of employing the writing data shown in
FIG. 22
for forming the resist pattern of the portion having the width Wm by two-shot electron beam irradiation in different positions.
As shown in
FIGS. 24 and 25
, the dimensional dispersion is about ±0.025 &mgr;m in case of forming the resist pattern by one-shot electron beam irradiation, while that in case of forming the pattern by two-shot electron beam irradiation is about ±0.075 &mgr;m. Namely, it is understood that the dispersion of the pattern dimensions, i.e., the dimensional accuracy is deteriorated by the two-shot electron beam irradiation. This is one of the problems in the variable shaping writing data creation.
When a resist pattern is formed by irradiation with an electron beam of two or more shots on different positions and an edge of the resist pattern is formed with a small-sized shot among the plurality of shots, the dimensional accuracy of the formed resist pattern is further deteriorated. This problem is described with reference to
FIGS. 26A
,
26
B,
27
A and
27
B.
A graph shown in
FIG. 26B
illustrates the intensity distribution of an electron beam which is shaped as shown in FIG.
26
A. The intensity distribution of this electron beam is not completely rectangular, but spread on edge portions. The slopes (hereinafter referred to as “beam sharpness”) of the intensity distribution at the edges vary with the size of the shaped beam. In general, Coulomb repulsion in the electron beam is increased as the size of the shaped beam is increased, leading to reduction of the beam sharpness and dulling of the intensity distribution on the edges.
On the other hand,
FIG. 27B
illustrates intensity distribution of an electron beam of two shots shown in
FIG. 27A
for forming a pattern of the same size as that shown in
FIG. 26A
with a small-sized left shot. As shown in
FIGS. 27A and 27B
, the intensity distribution on the left edge is different from that shown in
FIG. 26B
, leading to difference between the dimensions of the formed resist patterns.
FIG. 28
shows a result obtained by forming a plurality of resist patterns with two-shot electron beam irradiation including a small graphic as shown in
FIGS. 27A and 27B
and measuring resist dimensions. As compared with the result (
FIG. 24
) in case of writing the graphic with one-shot beam irradiation, dimensional dispersion is increased while difference results in the average value. This difference of the average value results from the electron beam irradiation of the smal
Kamiyama Kinya
Moriizumi Koichi
Taoka Hironobu
McDermott & Will & Emery
Mitsubishi Denki & Kabushiki Kaisha
Vo Cliff N.
LandOfFree
boundary processing of oblique overlapping graphics to... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with boundary processing of oblique overlapping graphics to..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and boundary processing of oblique overlapping graphics to... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2505011