JPS6055826B2 - photomask pattern - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a photomask for producing fine patterns, and more particularly, to a photomask pattern that can reliably form fine photomask patterns with high precision.

At present, fine patterns for semiconductor L2 and magnetic bubble elements are produced by photolithography using a photomask.

In recent years, the size of fine patterns in semiconductor LSIs and magnetic bubble devices has become smaller and smaller as devices become more densely packed. In the current fine patterns, the gap between the patterns has been reduced to about 1 μm, which is said to be the limit of photolithography technology. In order to reliably produce such a fine pattern on a device over a wide area with high precision, the fine pattern on the photomask used for this needs to be formed with high precision over a wide area.

However, it is difficult to reliably produce such fine patterns, which are close to the limits of photolithography technology, on a photomask over a wide area with high density and precision, and it has been difficult to obtain satisfactory photomasks with good reproducibility. . Specifically, this is a problem in that the size of the gap between fine patterns varies greatly depending on the location, and in some cases, the patterns short-circuit (overlap) at the gap portion. SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a photomask which eliminates the above-mentioned problems with photomasks and which can reliably form fine patterns with high density over a wide area with high accuracy.

To make a photomask, first, a device called a pattern generator is used to create an original J version of the pattern that is twice the actual size.

This original plate is called a reticle. Next, using a photorepeater device called a step repeater, the pattern on the reticle is exposed at a 1:1 reduction to create an actual size pattern on the dry plate. By repeating this operation using a step repeater, a photomask in which a plurality of full-sized patterns are arranged vertically and horizontally is produced. In order to accurately create high-density micropatterns over a wide area in photomasks manufactured using these procedures, it has been considered that it is essential that the pattern on the original reticle be uniform and accurate. The results became clear.

It has been found that if the accuracy of the reticle pattern is poor, it is difficult to obtain a satisfactory photomask with good reproducibility, and conversely, if the accuracy of the reticle pattern is good, a satisfactory photomask can be easily produced from it. The results of the study regarding the relationship between the reticle pattern accuracy and the photomask pattern accuracy will be explained in more detail.

Now, assume that the design value of the gap between the reticle patterns, which is one scale of the actual size, is 10 μm and the accuracy is ±1 μWL (±10%). In this case, the actual gap between the reticle patterns is 9 μm in some places and 1 μm in other places.
It becomes 1 μm. When a photomask is made from such a reticle pattern, assuming that a photomask pattern that is completely faithful to the reticle pattern can be created, the gap between the photomask patterns will be 1.0 pTrL, ±0.1 μ.
m (±10%), so there is no problem. However, in reality, a ~1 μm gap is close to the limit of photolithography technology, so it is impossible to obtain a photomask pattern that is completely faithful to the reticle pattern, and only the following photomask pattern can be obtained. That is, if you try to make a 9 μm gap part on the reticle pattern by 0.9 μml on the photomask pattern, the 11 μm gap part on the reticle pattern will not be 1.1 μm on the photomask pattern, but will be ~1.5 μm. It has spread greatly. Mau. The photomask pattern is 1.0 pT
There is no rL±10% gap. 1.0μm×50%,
- Only 10% can be obtained.

Conversely, if you try to create a part with a gap of 11 μm on the reticle pattern to 1.1 μm on the photomask pattern, the part with a gap of 9 μm on the reticle pattern will not be 0.9 μm on the photomask pattern, but will be shorted by 0 μm. Put it away. 1.0 as a photomask pattern
Only μm±10% and −100% can be obtained. In this way, even if the accuracy on the reticle pattern is ±10%, the accuracy on the photomask pattern is ±10%.
%, but is expanded to ±10% or more. This expansion rapidly increases as the pattern gap approaches the limit of photolithography technology, and as the accuracy on the reticle pattern deteriorates. Contrary to the above, the gap between reticle patterns is 10μ
If the accuracy is as high as m±0.1μ7n, (±1%), a fully satisfactory photomask pattern of 1μm±1 to 2% can be easily obtained.

In magnetic bubble elements, a bubble transfer path 10 consisting of an asymmetrical apron 1 as illustrated in FIG. 1 is currently often used as a bubble transfer path.

FIG. 2 shows an example of the result of actually producing a reticle pattern (actual size 10f8 reticle pattern) of such a magnetic bubble transfer path 10. The figure shows the results of manufacturing a magnetic bubble transfer path in which the repetition period λ of the asymmetrical asymmetrical aprons 1 is 80 μm, and the gap g between the asymmetrical asymmetrical aprons 1 is 10 PTr1. As shown in Figure A, the finished reticle pattern has a bulge at the shoulder part of the asymmetrical apron 1, and the gap g1 at that part is 9 to 8 μm.
It became. FIG. 3 shows an example of the result of actually creating a full-sized photomask pattern using such a reticle pattern.

The gap g″ between the asymmetrical Siprons 1 is 1.0PTrL
When fabricating the reticle pattern as shown in A'' in the same figure, the shoulder portion of the asymmetric asymmetrical apron 1 was short-circuited, and a satisfactory photomask pattern could not be obtained.The cause of this short-circuit was that the reticle pattern A'' This is due to swelling.

The reason why this bulge A occurs will be described below. Creating a reticle pattern using a pattern generator
It is done as follows.

FIG. 4 is a diagram for explaining this, and illustrates the case of the asymmetrical apron 1. As illustrated in the figure, the asymmetrical apron 1 is decomposed into rectangles. In the case of the figure, it is decomposed into four rectangles numbered 1 to 4. A reticle pattern is created by sequentially exposing these four rectangles using a pattern generator. In this case, as shown in FIG. 5, the overlapping portion S of the separated rectangles will be exposed twice.

The exposed portions of the photoresist on the reticle dry plate are exposed to light, but the boundary portions of the photoresist are also exposed as light penetrates into the photoresist. At the boundary of the double exposed portion S, the amount of light penetrating into the photoresist is doubled.

As a result, the finished pattern has bulges A and B, as illustrated in FIG. The reason why the finished reticle pattern causes bulge A is as follows.
As explained above, this was due to double exposure.

Therefore, the basic idea of the present invention is to eliminate the portion A that swells toward the gap by devising the pattern shape of the portion that is exposed multiple times.

Hereinafter, the photomask pattern according to the present invention will be explained using examples.

FIG. 6 is a diagram showing an example of a pattern according to the present invention, and shows an example of a pattern of an asymmetrical apron.

After rectangular decomposition, the part S'' adjacent to the gap part among the parts to be double exposed is a from the boundary 1 of the gap part.
It is designed to go inside only. According to experiments, it is sufficient to select the value of a to be 0.1 to 0.2 μm in terms of actual pattern size. By doing this, even if the reticle pattern bulges in the double-exposed area, the bulge will not go into the gap, and g1 described in FIG. 2 will match g. Furthermore, if the a is large, such as 0.3 μm, the shape of the pattern cannot be secured, and conversely, if it is too small, there is no effect of preventing the occurrence of A. FIG. 7 is a diagram showing another embodiment of the pattern according to the present invention, and shows an example of a pattern of an asymmetrical apron.

The feature of the embodiment shown in FIG. 6 is that in the pattern shown in FIG.
It's in the middle of the day. This makes it possible to reduce the amount by which the reticle pattern swells in the double-exposed area. As explained above, according to the present invention, the problem of the gap g of the reticle pattern being smaller than the design value in the double-exposed area can be solved, and the photomask consisting of a high-density fine pattern over a wide area can be reliably formed with high precision. It can be made.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an example of a photomask pattern, FIG. 2 is a diagram illustrating problems with a conventional reticle pattern, and FIG. 3 is a diagram illustrating problems when a photomask is manufactured from the reticle pattern shown in FIG. 2. 4 and 5 are diagrams for explaining the causes of the above-mentioned problems, and FIGS. 6 and 7 are diagrams showing examples of photomask patterns according to the present invention.

Claims (1)

[Scope of Claims] 1. A reticle pattern is produced by a pattern generator by dividing the pattern into rectangles, and in a photomask pattern produced from this reticle pattern by a step repeater, in a portion S that is double exposed by the pattern generator, A photomask pattern characterized in that a portion S' adjacent to a gap between patterns is located inside a boundary l of the gap. 2. The photomask pattern according to claim 1, wherein the dimension inside the boundary l is 0.1 to 0.2 μm.