Photocopying – Projection printing and copying cameras – Methods
Reexamination Certificate
2001-03-23
2004-12-21
Nguyen, Henry Hung (Department: 2851)
Photocopying
Projection printing and copying cameras
Methods
C355S053000
Reexamination Certificate
active
06833908
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the generation and display of images or the recording of patterns for reconstruction using an image transducer, and in particular relates to a computer architecture for and method of generating and displaying images or recording patterns that exhibit line edge placement accuracy to a fraction of the pixel size of the image transducer.
BACKGROUND OF THE INVENTION
A system for a maskless lithography system employing a flashing (pulsed) radiation source, an array of micro-mirror chips each having a programmable two-dimensional array of adjustable micro-mirrors, and a continuously scanning stage for supporting a wafer to be exposed is described in U.S. Pat. No. 5,691,541 (“the '541 patent”), which patent is incorporated herein by reference. An example of a digital micro-mirror device (DMD) suitable for such a lithography system is described in U.S. Pat. No. 4,566,935, and in an article by Mignardi, entitled “Digital micro-mirror Array for Projection TV,” Solid State Technology, July 1994, pg. 63.
In the system of the '541 patent, each DMD in the array is programmed so that a binary (black or white) array pattern is formed, with each micro-mirror in each DMD constituting a pixel. The pattern is briefly illuminated by a flashing radiation source. Radiation reflected from the DMD array is projected through a projection optical system to a substrate located on the scanning stage. The flashing radiation source produces a very short pulse with a relatively long time between pulses. For example, for an excimer laser light source, the temporal pulse length of the flash may be less than 30 ns and the repetition rate may be up to 5000 Hz. Each flash results in a partial exposure of the pattern onto the substrate, which is coated with a photosensitive material (e.g., photoresist). The motion of the substrate during the duration of a flash is negligible. In the relatively long period between radiation pulses the pattern on the mirror array is advanced (“stepped”) by a small increment corresponding to the movement of the substrate on the scanning stage, so that the pattern formed on the DMD array and the pattern formed on the substrate remain aligned. Hence, the designation “step-and-flash.” Complete exposure of the DMD array pattern onto the substrate typically requires a significant number of flashes.
The '541 patent also discussed how a pattern written on a relatively fine grid could be accurately reproduced by a mirror array corresponding to a coarse grid, provided that the pattern can be modified slightly between successive partial exposures.
The basic scheme of the '541 patent has a number of advantages. Since the pattern moves across the micro-mirror array in many steps, an isolated defective micro-mirror (pixel) in the array has a minimal effect on the imaged pattern. This is because the contribution from the defective micro-mirror is outvoted by the remaining micro-mirrors used to represent this pattern element in the other exposures. For example, if 31 exposures are required for a pattern element to advance across the width of the micro-mirror array, then the diffraction-limited pattern element would be correctly represented by 30 exposures and would be distorted slightly by the defective micro-mirror in only one exposure. Furthermore, if a minimum feature in the pattern is represented by a 2 by 2 or 3 by 3 array of micro-mirrors, then the effect of an isolated defective mirror is reduced further by a factor of 4 or 9.
Another advantage of using multiple exposures of a micro-mirror array is that it affords the possibility of shifting the position of an edge between exposures. For example, with a single exposure, an edge position has to correspond to an edge boundary in the micro-mirror array. If these boundaries correspond to 50 nm increments on the substrate, then it is possible to write 100 nm and 150 nm lines but not 125 nm lines. Current mask pattern writing systems approximate a pattern laid out on a fine grid with a pattern on a coarse grid by choosing the edge on the coarse grid that is closest to the edge on the fine grid. This “edge popping” inevitably results in some distortion of the desired pattern. With multiple exposures it is possible to dither the edge position from one exposure to the next and more closely approximate the desired pattern.
By using 2 exposures and by shifting the position of an edge between exposures, it is possible to create a 125 nm line-width. Similarly, with 31 exposures it is possible to divide the mirror spacing by 31 or to achieve 1.6 nm line-width resolution capability and 0.8 nm line edge position resolution.
Generally, a very fine grid is used to represent the patterns required to manufacture an integrated circuit. For example, a 5 nm grid might be used to represent a circuit having a 100 nm minimum feature size. This allows incremental changes in feature size of only 5 nm. It is impractical to build a micro-mirror array corresponding directly to such a fine grid.
Most lithography projection optical systems are designed to be diffraction limited. When used with binary masks, such lens systems are able to resolve minimum features corresponding to about double the diffraction-limited size or those corresponding to a spatial frequency of NA/&lgr;, where NA is the numerical aperture of the projection optical system, and &lgr; is the wavelength of exposure radiation used. If a minimum feature width (typically, 0.5 &lgr;/NA) is formed using 2 or more mirrors in the micro-mirror array, then each of the mirrors is of a size (0.25 &lgr;/NA), which is equal to or smaller than the diffraction limit.
An image of an edge in a diffraction-limited imaging system is not infinitely sharp but rather has a gradual transition from full exposure to no exposure over a distance of about &lgr;/NA. This is larger than the width of an individual micro-mirror sized at or below the diffraction limit of 0.25 &lgr;/NA. Because of this effect, and because of the threshold nature of most photoresists used in lithography, multiple partial exposures can be superimposed with the pattern edge defined on the mirror array varying by as much as a mirror width with only a small but manageable effect on the edge slope of the resultant exposure. This shifts the resultant edge position on the substrate by an amount that is well approximated by the position of the edge averaged over the number of exposures. Thus, with a total of 31 partial exposures for example, a pattern edge can be positioned 7/31 of the distance between adjacent micro-mirror edges A and B by performing 7 exposures with the pattern image edge located at mirror edge A and with 24 exposures with the pattern image edge located at mirror edge B.
Despite the above-identified advantages of the system of the '541 patent, there is a significant drawback to the multiple-exposure approach. Namely, use of multiple exposures greatly increases the volume of data needed to program the DMD array to represent the pattern. For example, 31 exposures increases the amount of data needed to program the array by 31 times. Even with a single exposure, the data needed to represent a 22 mm by 22 mm pattern using a 50 nm-scale grid (i.e., 50 nm micro-mirror size) is an enormous 1.9×10
11
bits.
Thus, for the multiple-exposure approach to be applicable to image formation and printing applications involving large amounts of data, systems for and methods of storing transferring and displaying the data are needed.
SUMMARY OF THE INVENTION
The present invention relates to the generation and display of images or the recording of patterns compatible with using an image transducer. In particular the invention relates to a computer architecture for, and method of, generating and displaying images or recording patterns that exhibit line edge placement resolution to a fraction of the pixel size of the image transducer. The present invention is particularly applicable to the creation of high-resolution images and the patterning of substrates using maskless, step-and-
Jones Allston L.
Nguyen Henry Hung
Ultratech, Inc.
LandOfFree
Computer architecture for and method of high-resolution... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Computer architecture for and method of high-resolution..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer architecture for and method of high-resolution... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3275646