Photocopying – Projection printing and copying cameras – Illumination systems or details
Reexamination Certificate
2002-03-11
2004-10-26
Mathews, Alan (Department: 2851)
Photocopying
Projection printing and copying cameras
Illumination systems or details
C355S053000, C355S070000, C438S487000
Reexamination Certificate
active
06809801
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention generally relates to liquid crystal display (LCD) and integrated circuit (IC) fabrication and, more particularly, to a silicon film and fabrication process to laser irradiate silicon film in making polycrystalline silicon thin film transistors (TFTs) for Active Matrix (AM) LCDs using a 1:1 laser projection mechanism.
2. Description of the Related Art
Lateral crystallization by excimer-laser anneal (LC-ELA) is a desirable method for forming high quality polycrystalline silicon films having large and uniform grains. Further, this process permits precise control of the grain boundary locations.
FIG. 1
illustrates steps in an LC-ELA annealing process (prior art). As seen in Step 1, initially amorphous silicon film is irradiated by a laser beam that is shaped by an appropriate mask to an array of narrow “beamlets”. The shape of the beamlets can vary. In Steps 1-4, each beamlet is shaped as a straight slit of narrow width, approximately 3-5 microns (&mgr;m). This slit is represented in the figure as the two heavy lines. The width of the slit is the distance between these two lines. This width can vary, but ultimately it is dependent upon the attainable lateral growth length (LGL), which is defined as the distance crystals can grow laterally (inwardly) from the edges of the irradiated area. Typically, the beamlet width is designed to be slightly less than twice the corresponding LGL.
The sequence of steps in
FIG. 1
illustrates the growth of long polysilicon grains by LC-ELA process. A step-and-repeat approach is used. The laser beamlet width (indicated by the 2 parallel, heavy black lines) irradiates the film and, then steps a distance (d), smaller than half of the lateral growth length (L), i.e. d<L/2. Using this step-and-repeat process, it is possible to continually grow crystal grains from the point of the initial irradiation, to the point where the irradiation steps cease. L is dependent upon a combination of film thickness and substrate temperature. For 50 nanometer-thick films at room temperature, L is approximately 1.2 microns (&mgr;m). Due to this slow beamlet advancement, at each step, grains are allowed to grow laterally from the crystal seeds of the polycrystalline silicon (poly-Si) material formed in the previous step.
FIG. 2
illustrates a plan view of a conventional mask (prior art). The initially amorphous silicon film is irradiated by a very narrow laser beamlet, with typical width of a few microns (i.e. 3-5 &mgr;m). Such small beamlets are formed by passing the original laser beam through such a mask, that has open spaces, and projecting the formed beamlets on the surface of the annealed Si-film.
FIG. 3
is partial cross-sectional view of
FIG. 1
illustrating the surface topography of laser-irradiated domains prior art). After the completion of the lateral growth, the two crystal fronts meet at the center of the domain where they form a “boundary” between the two crystal regions developing from each opposing edge of the domain. As a result of the grain boundary formation, a “ridge” develops at the surface of the film at the boundary, corresponding to the planned congruence of the two crystal fronts. Since the substrate steps under the beam a distance of d, where d is less than L/2, the ridge is irradiated is a subsequent shot. This ridge remelts and locally planarizes. However, as part of the same process, another ridge is formed at a new location. Therefore, the ridge location will “march” across the substrate in response to the scans under the beam.
If the angle of rotation between the lattice mismatch on the two sides of the boundary is less than approximately 15 degrees, the boundary is considered to be a low-angle boundary. An angle of rotation between 15 and 90 degrees is considered to be a high-angle boundary. Electron mobility between high-angle boundaries is impaired, while mobility between low-angle boundaries is usually insignificant. The step-and-repeat annealing typically promotes low-angle boundaries. However, the film regions corresponding to the mask edges, not being subject to the step-and-repeat process, are likely to form high-angle boundaries.
When high-angle boundaries are formed, the TFT channels need to be arranged to avoid these regions. That is, the TFTs need to be formed in the planar regions between neighboring ridges to avoid performance deterioration. Even more undesirable is the formation of neighboring TFTs with different performance parameters, resulting from the random formation of TFT channels with ridges adjacent TFT channels without ridges. Hence, some sort of alignment is necessary between the crystallized domains and the position of the TFT channels within these domains. This alignment process introduces additional processing steps, hence, increasing the cost of the process. It would be desirable to eliminate these additional processing steps so that TFT channels can be placed on the processed (laterally crystallized) film without the requirement of calculating ridge alignments.
Typically the beam, after being shaped by a mask such as the one shown in
FIG. 2
, is projected on the surface of the sample after being “demagnified” by several times. As a result of this demagnification, the beam area shrinks considerably by comparison to the beam area available at the mask plane. One unfortunate result of a demagnified beamlet is an increase in the number of high-angle boundaries. Since the laser energy remains the same, while the beam area shrinks, the laser energy density on the surface of the substrate is much higher than on the mask. The typical demagnification factor is ×3-×5, which means that the beam area on the substrate is 9 to 25 times smaller than on the mask. Consequently, the laser energy density on the mask is 9 to 25 times lower than on the substrate. This allows very low energy density levels on the mask for a given requirement of energy density on the substrate.
Generally, for low cost masks, it is important to avoid exposing the mask at high laser energy density levels. The demagnification of ×3-×5 associated with the conventional use of masks minimizes the exposure of masks to high energy densities. For example, if a 1:1 projection were used with the masks, instead of 3:1 to 5:1, the amount of laser energy density impinging on the mask would necessitate frequent mask changes to maintain dimensional stability and to reduce optical artifacts. Even with frequent mask, it is doubtful that such a scheme would be practical in a mass production application. However, a 1:1 projection concept would have certain advantages. Larger areas of film can be exposed to each laser shot, so throughput would be increased. Likewise, control over the location of crystallized areas would be enhanced. Further, simple schemes would be available to reduce material directionality issues, as electrical performance is not isotropic, but depends upon the relative position of the device channel with respect to the lateral growth direction. Hence, it would be desirable to use a 1:1 projection laser irradiation, if a means could be found of reducing the possibility of mask damage.
It would be advantageous to develop a 1:1 laser projection system that did not increase the likelihood of damaging masks. To that end, it would be advantageous if a 1:1 projection system could be used with a lower power laser energy density.
It would be advantageous if additional energy could be directed to the substrate in a controlled manner to decrease the energy that must be directed to any particular area of the substrate through the mask.
It would be advantageous to increase the “footprint” of the laser beamlet on the substrate to increase process times and to reduce to formation of high-angle boundaries.
SUMMARY OF THE INVENTION
The present invention is a system and method that uses a lamp, in additional to the mask-projected laser beam, to anneal a semiconductor substrate. The use of the lamp as an auxiliary heat source permits a 1:1 projection lens to be used to d
Hartzell John W.
Voutsas Apostolos
Curtin Joseph P.
Mathews Alan
Rabdau Matthew D.
Ripma David C.
Sharp Laboratories of America Inc.
LandOfFree
1:1 projection system and method for laser irradiating... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with 1:1 projection system and method for laser irradiating..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and 1:1 projection system and method for laser irradiating... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3291887