JPH07288224A - Operating method for mask pattern data of integrated circuit - Google Patents

Abstract

PURPOSE:To prevent the change in graphics after operation by performing doughnut spreading for a mask pattern data used in a semiconductor integrated circuit having the hierarchical structure, providing graphics in close proximity to a cell boundary at both master and slave cells, performing geometrical operation, and correcting the values after the processing. CONSTITUTION:At first, a hard zone frame is set for all cells (a). Then, the processing for forming the spreading data such as doughnut spreading is performed for the upper cells for every cell from the bottom to the root cell (b). The results are sequentially recorded (c). Then, the spread data obtained for every cell are read from the bottom cell (cell A) to the root cell (cell B) (d). When the data are the ordinary data, the data are sent for the geometrical operation; and when the data the cut-zone-frame data, the data are sent for correcting operation (e). Then, the geometrical operation is performed for every cell (f). The correcting operation is performed on the obtained geiometrically operated data (g). The obtained processed data are outputted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to processing of mask pattern data used for layout verification and electron beam exposure of integrated circuits such as LSI and VLSI, and more particularly, to processing of mask pattern geometric operations. The present invention relates to arithmetic processing of mask pattern data.

[0002]

2. Description of the Related Art Conventionally, mask pattern data of an integrated circuit is composed of a set of cells each composed of unique pattern data and cell references, and has a hierarchical structure based on cell references. The highest cell in this hierarchy is called a root cell and corresponds to an LSI chip. Then, replacing the cell reference with the pattern data is called expansion, and performing an operation such as enlarging or reducing the pattern expressed by each pattern data (graphic data) is called a geometric operation. This was done by flat processing in which all pattern data were expanded to the root cell without any structure and then operated. However, in recent years, with the high integration (large-scale) of semiconductor integrated circuits, mask pattern data of large-scale integrated circuits used for manufacturing this circuit has come to exceed 1 million in the number of mask patterns. For the mask pattern represented by such pattern data,
In the case of performing a geometric operation, the amount of data to be processed increases in the conventional flat processing, and the processing time becomes long.

In order to deal with this, cell unit processing using the cell structure of the mask pattern data has been performed, but this processing has the following problems. The first problem is the problem of deformation of the figure at the cell boundary of the mask pattern data when enlarging the mask pattern represented by the mask pattern data. 12
(A) shows a state in which the figure has changed compared to the result of the flat processing at the time of enlargement. It can be seen that the figures overlap and are deformed. Even if the overlapping portion of the figures is removed, the shape of the figure is deformed. This is
The pattern 1 figure of (a) is processed in the upper cell in the figure, and
The pattern 2 figure of (a) is individually enlarged in the lower cells of the figure, resulting in pattern 1 being 1A and pattern 2 being 2A. However, since the figure boundary at the cell boundary is not taken into consideration. The figure in which 1A and 2A are combined is deformed, unlike the figure in which patterns 1 and 2 before processing are combined. The second problem is the problem of separation of figures at the cell boundaries of the mask pattern data when the reduction process is performed on the mask pattern represented by the mask pattern data. FIG. 12B shows a state in which graphic separation occurs as compared with the result of the flat processing at the time of reduction. This also occurs mainly because no consideration is given to the graphic contact at the cell boundary. This is because the pattern 3 figure of FIG. 12B is processed in the upper cell of the figure, and the pattern 4 figure of FIG.
The pattern 4 becomes 4A in A, but the figure in which 3A and 4A are combined is deformed unlike the figure in which patterns 3 and 4 before processing are combined.

[0004]

Therefore, when performing geometric calculation on mask pattern data representing a mask pattern of an integrated circuit, the mask pattern data can be processed in a cell unit, and a figure of the obtained mask pattern can be obtained. There has been a demand for eliminating shape deformation. Under the circumstances as described above, the present invention can perform the cell unit processing when performing the geometric operation on the mask pattern data of the integrated circuit, solve the problem occurring at the cell boundary, and obtain the shape of the figure of the obtained mask pattern. It is intended to provide a data processing method capable of eliminating deformation.

[0005]

According to a method of processing mask pattern data of the present invention, a mask pattern for an integrated circuit represented by mask pattern data in which data is formed in a hierarchical structure of a plurality of cells is enlarged. Or a data processing method using a hierarchical structure of the mask pattern data in performing a geometrical operation of reduction, which is (1) represented by another cell frame or upper pattern data for all cells Setting a hard zone frame for each cell, which is a frame that represents an area in an independent cell that does not interfere with the figure, and recording rectangular data that represents the hard zone frame, (2) bottom-up, in cell units, Donut expansion that expands the data of the donut area, which is the area between the zone frame and the original cell frame obtained by reducing the specified width from the hard zone frame to the upper cell Expansion of rectangular data representing a cut zone consisting of a predetermined area into an upper cell to remove the overlap of pattern data with the cell from an upper cell area that cites the cell, and expansion including pattern data unique to the cell The step of creating data, the expansion data created in (3) and (2) are read in the order of coordinates, and it is judged whether the data is normal pattern data or rectangular data representing a cut zone. Passing to the arithmetic processing step, if the rectangular data representing the cut zone is passed to the correction processing step, (4) Geometric arithmetic processing is performed on the normal pattern data to create geometric arithmetic processing data. To the correction processing step,
(5) When there is a subordinate cell in the processing cells, a logical difference operation is performed using the geometrical operation processing data of (4) and the rectangular data representing the cut zone of each cell, and the geometrical operation of (4) above is performed. Logical difference calculation processing data is created by removing the data located in the cut zone area from the calculation processing data. If there is no subordinate cell, the geometric calculation processing data of (4) above is used as the logical difference calculation processing data. When there is a higher-order cell for the logical difference operation processing data, the AND operation is performed using rectangular data that represents the AND zone of the area corresponding to the cut zone of the upper cell of each cell, and the AND operation is performed in addition to the AND-zone area. The logical product arithmetic processing data is created by removing the located data, and when there is no higher-order cell, the logical difference arithmetic processing data is directly used as the logical product arithmetic processing data and is passed to the output processing step. Correction process, (6)
The output step of outputting the logical product operation processing data obtained in (5). The mask pattern data arithmetic processing method according to the present invention, in the above, sets the enlargement width or the reduction width in the geometric calculation of enlargement or reduction to 2
^If the value multiplied by ^{1/2 is used} as the size amount, the area where the donut is expanded is the area between the zone frame and the cell frame that are twice the size reduced by the hard zone frame, and the outermost circumference of the cut zone is set. The cut zone frame to be represented is a frame in which the hard zone frame is shrunk inward by the size amount.

In the above, after the hard zone frame is set, the pattern data of the donut area around the lower cell to be quoted is incorporated into the pattern data of the higher cell to be quoted. The calculation processing is duplicated for the pattern data of the donut area. Here, the processing of the mask pattern data for performing the geometric calculation of the mask pattern is referred to as geometric calculation processing. Therefore, as the donut area, it is necessary to have a predetermined width that can correspond to the enlargement and reduction width for enlarging and reducing, but the smaller the area, the smaller the amount of data to be processed redundantly. When the enlargement width or reduction width at the time of the geometric calculation is set to 2 ^1/2 times, the area where the donut is expanded is reduced by twice the size of the hard zone frame and the original area including the hard zone frame. The area between the cell frame and the cell frame. So in this case,
The donut area is set as shown in FIG. The cut zone is an area that is provided to eliminate pattern data that has been processed redundantly and to eliminate graphic deformation due to overlap. The setting of this area also corresponds to the enlargement and reduction width for enlargement and reduction processing. Although it is sufficient if it is a predetermined area that can be done, in general, the donut area is the area between the zone frame in which the hard zone frame is doubled in size and the original cell frame. The cut zone frame to be represented is a hard zone frame that is shrunk inward by the size amount. However, for the root cell, the cut zone corresponds to both expansion and contraction operations, so the cut zone frame is obtained by expanding the hard zone frame by the size amount to the outside. The AND zone is also provided in order to eliminate pattern data that has been processed redundantly and to eliminate graphic deformation due to overlap, but this area is set to the area corresponding to the cut zone in the cell above each cell. And The geometric calculation processing data, the logical difference calculation processing data, and the logical product calculation processing data refer to the data obtained as a result of the geometric calculation processing, the logical difference calculation processing, and the logical product calculation processing, respectively.

[0007]

The geometric operation method of the integrated circuit mask pattern using the hierarchical structure of the present invention is configured as described above, so that the process for obtaining the mask pattern without figure deformation can be performed in the cell unit process of the mask pattern data. It is possible.
As a result, the number of data to be processed can be significantly reduced as compared with the flat processing, and processing of mask pattern data for a large scale integrated circuit can be performed at a practical level. By setting the hard zone, the interference between cells and the interference between the figures of cells and upper cells are eliminated, and independent processing can be performed in cell units. By expanding the donuts, both the parent and child cells (upper cell and lower cell) have shapes that are close to the cell boundary, and both the parent and child cells process it redundantly. However, it is assumed that the shape contact at the cell boundary can be recognized. However, at this time, the donut area is defined as an area between the hard zone frame and the original cell frame that is twice the size of the hard zone frame, and the cut zone frame that represents the outermost circumference of the cut zone is the hard zone frame. By increasing the size of the frame to the inside, the increase in the amount of data is minimized. By the correction processing, the protruding portions of the upper cell and the lower cell of the geometric calculation result are cut, whereby the generation and separation of abnormal figures can be suppressed.

[0008]

Embodiments of the present invention will be described below with reference to the flow chart of FIG. The processing in the embodiment will be described with reference to a case where a mask pattern reduction operation represented by mask pattern data having a cell hierarchical structure is performed. FIG. 2A shows the entire mask pattern used here. For simplicity, the mask pattern data is 2
It has hierarchical structure data having two cells A and B, where A is an upper cell of B. The cell A has graphic data of a1, a2, a3, and the cell B has graphic data of b1, b2, b3, b4. The cell A and the cell B are graphically displayed in FIGS. 2B and 2C, respectively. In this case, the general view (a) of the mask pattern matches the combined view of (b) and (c).

First, hard zone frames are set for all cells. (A) The hard zone frames of cell A and cell B are HA and HB that do not interfere with each other, as shown in FIG. Here, the interfering state means a state in which cells overlap with each other as shown in FIG.
As shown in (b), the cell and the upper cell figure overlap each other. The hard zone frames of the cell A and the cell B in 3 do not overlap with each other and therefore do not interfere. In such setting of the hard zone frame, first, the minimum rectangle including the pattern data unique to the cell is obtained. 14
As in (a), this minimum rectangle is called a cell boundary or cell frame. Next, with respect to all the cells, the interference between the cell and the cell, and the cell and the upper cell graphic is examined, and the original cell frame is reduced to a position where there is no interference, and a new frame is obtained. Specifically, the maximum interrupt widths in the horizontal and vertical directions of the graphic of the upper cell are respectively obtained, and the original cell frame is reduced accordingly. This new frame is called a hard zone frame and is as shown in FIG. FIG. 4 shows the graphic data b1, b2, b for the cell B.
3 and b4 are graphic displays and a heart zone frame S1 and the like. S2 is a zone frame in which S1 is reduced in size,
The data in the S2 frame is used when setting the cut zone area of the upper cell A and when setting the AND zone area of the own cell B. S3 is a zone frame obtained by reducing S1 by twice the size amount, and the donut area is an area between S1 and S3.

Next, a process of creating expanded data is performed for each cell (b) and sequentially recorded in the recording section.
(C) In cell B, since there is no lower cell, there is no expanded data, and the graphic inside the hard zone frame, that is, the inside of HB is the expanded data. Processing for creating expanded data is performed for cell A. The data in the donut area of the cell B shown in FIG. 4 is expanded in the cell A, which is the upper cell. The shaded portions b1, b2, and b4 in FIG. 5 display this data.
B1, b2, and b4 are selected as the graphic data in the donut area near the cell boundary of the cell B, which is a lower cell of the cell A. The expanded data of the cell A after the donut expansion are a1, a2, a3, b1, b2, b4. This is shown in FIG. Further, the zone frame of the cell B, which is graphically displayed as S2 in FIG. 4, is used as the cut zone frame of the cell A, and its rectangular data is expanded into the cell A and the cut zone expansion data S
Get 2. This is shown in FIG. Therefore, the expanded data after expansion of the donut and cut zone of cell A is a
1, a2, a3, b1, b2, b4 and S2 data.

Next, the following processing is performed for each cell from the bottom (cell B) toward the root cell (cell A). By the above processing, the expanded data obtained is read (d),
If the data is normal data, pass it to the geometric operation processing,
If the data is cut zone frame data, the data is passed to the correction process. (E)

Next, a geometric calculation process is performed for each cell (f), a correction process (g) is performed on the obtained geometric calculation process data, and the obtained process data is output. (H) First, the geometric operation is performed on the pattern data of the cell B.
When the geometric calculation processing data resulting from the geometric calculation is graphically displayed, it becomes like the shaded area in FIG. A correction process is performed on this geometric calculation result. Since there is no lower cell in cell B, logical difference calculation processing with rectangular data representing the quad zone is not performed,
The geometric calculation processing data is directly used as the logical difference calculation data, and the logical product calculation processing (g2) of this and the rectangular data S2 representing the AND zone is performed. This logical product operation processing (g
In 2), the overlap and the overlap with the upper cell are eliminated. FIG. 10 is a diagram for explaining the logical product process on the cell B after the geometric calculation. For the mask pattern data graphically displayed in (a) after the geometric calculation process, the rectangular data S2 representing the AND zone and The result of performing the logical product calculation process of is the figure in the shaded area shown in (b). The obtained AND operation processing data is output. (H)

Next, the cell A after being expanded into a donut
, Which is graphically displayed in FIG. 7 as described above, performs geometric calculation on this cell A. The geometric calculation result of the cell A is the shaded area in FIG. 7. (B4 has disappeared due to the reduction operation.) As can be seen from this display, since the geometric calculation processing data includes data that overlaps with the lower cell B, this is removed by the correction processing. Specifically, redundant data is removed by performing a logical difference calculation process (g1) using the geometric calculation process data and the S2 data of the cut zone frame. FIG. 9 is a diagram for explaining the logical difference calculation process for the cell A after the geometric calculation. However, the cut-zone data S2 and the logical difference for the mask pattern data graphically displayed in (a) after the geometric calculation process. The result of the calculation (g1) is the hatched portion graphically displayed in (b).
Since cell A is the root cell, the AND operation with AND Thorn is not performed. The obtained logical difference calculation processing data is output. (H)

Therefore, the shaded portions in FIGS. 9 and 10 are the results to be obtained, and when these are combined, the shaded portions in FIG. 11 are obtained, and there is no separation or deformation of the figure, and the correct reduction operation has been performed. I understand. It should be noted that, in the present embodiment, it was mainly explained that the figures were not separated and transformed at the time of data processing, but not limited to the cells A and B, when the number of cells is large, each cell is Even when there is a large amount of graphic data, the data processing method of the present invention can be performed in the same manner as this embodiment. Therefore, the data processing method of the present invention can be applied at a practical level by utilizing the cell structure of the mask pattern data even if the mask pattern data becomes large in scale.

Although the reduction operation of the geometric operation has been described above as an example, the same applies to the enlargement operation, and a correct result can be obtained.

[0016]

As described above, the geometric operation processing method of integrated circuit mask pattern data using the hierarchical structure of the present invention performs the mask operation on the mask pattern data in the cell unit processing when performing the geometric operation on the mask pattern data. In addition, it is possible to provide a method that can perform an operation that can be performed without causing the shape deformation of a figure at a cell boundary, and as a result, enable geometrical operation processing of mask pattern data of a large-scale integrated circuit at a practical level. It is said that it is possible to meet the demands for miniaturization and high integration of the photomask in accordance with the high integration and miniaturization of semiconductor elements.

[Brief description of drawings]

FIG. 1 is a flow chart of a method for geometrical operation processing of an integrated circuit mask pattern according to the present invention.

FIG. 2 is a mask pattern diagram used for explaining an embodiment of the present invention.

FIG. 3 is a diagram showing a hard zone of each cell A of the mask pattern data used for explaining the embodiment of the present invention.

FIG. 4 is a diagram for explaining each zone frame in cell B.

FIG. 5 is a graphic representation of cell A after the donut expansion.

FIG. 6 is a diagram in which a cut zone frame developed in cell A is graphically displayed.

FIG. 7 is a diagram for explaining the geometric calculation process of cell A.

FIG. 8 is a diagram for explaining the geometric calculation processing of cell B.

FIG. 9 is a diagram for explaining the logical difference processing of cell A.

FIG. 10 is a diagram for explaining the logical product process of cell B.

FIG. 11 is a diagram after processing of mask pattern data used for explaining the embodiment of the present invention.

FIG. 12 is a diagram for explaining problems in the conventional cell unit processing.

FIG. 13 is a diagram for explaining interference between cells and between a cell and a graphic of an upper cell.

FIG. 14 is a diagram illustrating a relationship between a cell frame and a hard zone.

FIG. 15 is a diagram for explaining a hard zone and a donut area.

[Explanation of symbols]

HA cell A hard zone frame HB cell B hard zone frame a1, a2, a3, a4 cell A pattern data b1, b2, b3, b4 cell B pattern data S1 cell B hard zone frame S2 cell B cell Cut zone frame for A 1, 1A pattern figure 2, 2A pattern figure 3, 3A pattern figure 4, 4A pattern figure

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H01L 21/30 502 P

Claims (2)

[Claims] 1. A mask pattern data for performing a geometric operation for enlarging or reducing an integrated circuit mask pattern represented by mask pattern data in which data is formed in a hierarchical structure of a plurality of cells. A data processing method using a hierarchical structure of (1) for all cells,
Set a hard zone frame for each cell, which is a frame that represents an area within an independent cell that does not interfere with other cell frames or figures represented by higher-level pattern data, and record rectangular data that represents the hard zone frame. Process, (2) bottom-up,
For each cell, perform donut expansion to expand the data of the donut area, which is the area between the zone frame obtained by reducing the specified width from the hard zone frame and the original cell frame, to the upper cell, Rectangle data representing a cut zone consisting of a predetermined area for removing the overlap of pattern data with the cell from the upper cell area to be quoted is expanded to the upper cell, and expanded data including pattern data unique to the cell is created. Process, the expanded data created in (3) and (2),
A process of reading in the order of coordinates, determining whether the data is normal pattern data or rectangular data representing a cut zone, and passing it to the geometric calculation processing step if it is normal pattern data, and passing it to the correction processing step if it is rectangular data representing a cut zone. ,
(4) A geometric calculation processing step of performing geometric calculation processing on the normal pattern data to create geometric calculation processing data, and passing this geometric calculation processing data to a correction processing step, (5)
When there is a subordinate cell in the processing cells, the geometrical operation processing data of (4) above is used to perform a logical difference operation using the geometrical operation processing data of (4) and the rectangular data representing the cut zone of each cell. The logical difference calculation processing data in which the data located in the cut zone area is removed is created, and when there is no subordinate cell, the geometric calculation processing data of (4) is used as the logical difference calculation processing data, and this logical difference calculation processing is performed. If there is a higher-order cell for the processed data, rectangular data that represents the AND zone of the area corresponding to the cut zone of the higher-order cell of each cell is used to perform the logical product operation and the data located outside the AND-zone area. When the logical product arithmetic processing data is removed, and there is no higher-order cell, the logical difference arithmetic processing data is directly used as the logical product arithmetic processing data and is passed to the output processing step. Management step, (6) (5) in the obtained arithmetic management processing method for an integrated circuit mask pattern data and having an output process, which outputs a logical product operation processing data. 2. In claim 1, when the size amount is a value obtained by multiplying the expansion width or reduction width by 2 ^{1/2 in} the geometric calculation of expansion or contraction, the area for donut expansion is set inside the hard zone frame. In the area between the zone frame and the cell frame that are twice as small as the size amount, the cut zone frame that represents the outermost periphery of the cut zone is a frame that is made by shrinking the hard zone frame by the size amount. A method of arithmetic processing of integrated circuit mask pattern data.