JP2001174974A - Method for correcting optical proximity effect and light intensity simulation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an OPC method which can suppress an increase of the time required for OPC by implementing the OPC to all mask patterns within a chip. SOLUTION: The simulation base OPC is implemented by changing light intensity simulation models according to the shapes of the mask patterns and the positional relations of the mask patterns with each other and changing the number of repetition of the light intensity simulations.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical proximity correction (OPC) method and a light intensity simulation method, and more particularly to the manufacture of semiconductor integrated circuit devices, and more particularly, to the use of optical proximity correction in lithographic images. And an improvement in the accuracy of a reactive ion etching (RIE) image.

[0002]

2. Description of the Related Art A semiconductor device is manufactured by accurately replicating a mask pattern generated by a CAD for designing a semiconductor mask layout on the surface of an element substrate. In order to improve the performance of advanced semiconductor integrated circuits, it is essential to reduce the size of mask patterns. However, it is difficult to form a mask pattern having a minute size on an element substrate due to a deformation of the mask pattern from an image and a change in element dimensions in a series of optical lithography steps and an RIE step.

In recent years, as one of the methods for reducing the deformation of the mask pattern from the image and the fluctuation of the element size in the series of optical lithography steps and the RIE step, a series of optical lithography steps and RI
In order to correct the deformation of the mask pattern from the image and the fluctuation of the element dimensions in the E step, the shape of the mask pattern is selectively biased in advance to shift the edge of the mask pattern, that is, to deform the shape of the mask pattern.
PC is used.

In the OPC, two methods are generally used depending on the method of determining a correction bias amount. One is a rule-based OPC that determines the amount of bias for each edge of the mask pattern from the attributes of the mask pattern such as the size and shape of the mask pattern, the proximity to the adjacent mask pattern, and the like. This is a simulation-based OPC in which a light intensity simulation is performed to extract a difference between the shape of the mask pattern after deformation or dimensional change and the shape of the original mask pattern, and the bias amount of each side of the mask pattern is determined from the result.

In general, in rule-based OPC, a bias amount is determined from the size and shape of a mask pattern and a positional relationship with a peripheral mask pattern, and a bias is applied to each edge of the mask pattern. As described above, the light intensity simulation is not performed for each mask pattern. Therefore, the calculation processing time of OPC is faster than that of simulation-based OPC.
Also, since each edge of the mask pattern is not subdivided, O
It is possible to reduce an increase in the amount of data when recording the shape of the mask pattern as compared with the mask pattern before the PC processing.

However, in order to reduce the OPC calculation processing time and the data amount, it is necessary to reduce the subdivision of each edge of the mask pattern. Therefore, it is necessary to use the mask pattern after the OPC processing. However, there is a case where the pattern accuracy in forming the semiconductor element cannot be ensured. In addition, highly accurate OPC for diversifying mask patterns
It is necessary to define in detail the size and shape of the mask pattern for determining the bias amount, the positional relationship with the peripheral mask pattern, and the like, so that a rule-based OPC using a complicated calculation algorithm is required. Become.

In general, a simulation-based OP
In C, as shown in FIG. 7, each edge of the mask pattern is subdivided to perform pattern edge division (pattern edge division step 101), light intensity simulation is performed (simulation execution step 105), and a bias amount is calculated based on the result. Is determined, and a corrected OPC pattern is generated (Step 106) and output (Step 107), so that the shape of the mask pattern can be reproduced on the photomask with high accuracy.

In addition, since it is not necessary to finely define the size and shape of the mask pattern and the positional relationship with the peripheral mask pattern in order to determine the bias amount as in the rule-based OPC, it depends on the shape of the mask pattern. A high-precision OPC can be realized without the need.

However, in an actual LSI circuit, mask pattern data is very complicated and enormous.
Usually, it is composed of hundreds of thousands to millions of figures. In order to optimize the fine processing accuracy for such a pattern having a huge amount of data, it is practically impossible to perform OPC by performing light intensity simulation on the entire mask pattern in terms of time and cost. Was.

Conventionally, there is a method disclosed in Japanese Patent Application Laid-Open No. H11-184064, for example, in order to overcome the above problems. In this method, as shown in FIG. 7, each edge of a mask pattern is subdivided to perform pattern edge division (pattern edge division step 101), light intensity simulation is performed (simulation execution step 105), and specific pattern data is extracted. A light intensity simulation is performed on the specific pattern data, and a light intensity simulation is performed to determine a pattern intensity distribution when exposure is performed using the specific pattern data. This pattern data is corrected so that the difference between them becomes equal to or smaller than the allowable value.

[0011]

However, in such a conventional technique, simulation-based OPC is performed only on a mask pattern having a specific shape, and OPC is not performed on a mask pattern other than a specific shape. .
However, as the miniaturization of semiconductor elements progresses, OP
The shape of the mask pattern that needs to perform C increases,
Since the ratio of the mask pattern for performing the OPC to all the mask patterns in the chip increases, it takes an enormous amount of time to perform the OPC.

The present invention has been made in view of the above-mentioned circumstances, and the present invention has been made in consideration of the fact that an OP
An object of the present invention is to provide an OPC method capable of performing C and suppressing an increase in the time required for OPC.

[0013]

According to a first aspect of the present invention, there is provided a mask pattern edge dividing step of dividing each edge included in each mask pattern of a semiconductor integrated circuit according to a desired rule. A mask pattern edge classification step of classifying each edge divided in the mask pattern edge division step, and a simulation model for selecting a simulation model to be used when performing light intensity simulation for each edge classified in the mask pattern edge classification step A first simulation execution step of performing a light intensity simulation on the edge classified by the mask pattern edge classification step using the simulation model selected in the simulation model selection step; A first OPC pattern for generating an OPC pattern for a mask pattern by comparing the positions of edges before and after the simulation in the first simulation execution step and moving the edge before the simulation based on the difference; And a generating step.

According to the first configuration of the present invention, edges are classified and classified based on characteristics unique to each mask pattern, such as the shape and arrangement position of edges included in the mask pattern and the positional relationship with other mask patterns. A simulation model to be used for light intensity simulation is determined for each edge, so a high-precision simulation model is used for edges that require high-precision OPC, and a low-precision simulation model is used for edges that do not require very high OPC accuracy. , A light intensity simulation is performed at high speed, whereby high-precision and high-speed OPC can be realized for all mask patterns included in the chip.

According to a second aspect of the present invention, in the optical proximity correction method according to the first aspect, the mask pattern edge classifying step includes the step of determining an edge length, an edge arrangement position in the mask pattern, and other parameters. This is a classification step of performing classification based on at least one of the farmland overlapping with the mask pattern data.

According to this configuration, patterns are likely to change, and patterns are classified into patterns that need to increase simulation accuracy and patterns that do not need to be improved, and the simulation is executed according to the classification result. A highly reliable simulation can be performed.

According to a third aspect of the present invention, in the optical proximity effect correction method according to the first or second aspect, the mask pattern edge classification step further includes the step of further selecting the mask pattern edge according to the accuracy required on the actual device. The method includes a step of performing weighting and classification.

For example, since the gate length has a great influence on the response speed, high-precision control is required, whereas the element isolation region width does not necessarily require high-precision control. According to such a configuration, the pattern accuracy has a great effect on the device accuracy. Therefore, the pattern is classified into a pattern that needs to increase the simulation accuracy and a pattern that does not need to be improved, and the simulation is executed according to the classification result. Therefore, a highly reliable simulation can be performed efficiently.

According to a fourth aspect of the present invention, claims 1 to 3 are provided.
In the optical proximity correction method according to any one of
The simulation model selection step includes, for each edge classified in the mask pattern edge classification step, a repetition number determination step of determining the number of repetitions of light intensity simulation, the simulation execution step
The light intensity simulation is repeatedly performed on the edges classified in the mask pattern edge classification step for the number of times determined in the repetition number determination step, and the OPC pattern generation step is performed before and after the simulation is performed in the simulation execution step. And generating the OPC pattern for the mask pattern by moving the edge before performing the simulation based on the difference by the number of times determined in the repetition number determination step. It is characterized by the following.

According to this configuration, edges are classified based on characteristics unique to each mask pattern, such as the shape and arrangement position of the edges included in the mask pattern and the positional relationship with other mask patterns, and the light intensity is classified for each classified edge. Since the number of times of repeating the simulation and the generation of the OPC pattern is determined, the light intensity simulation and the generation of the OPC pattern are repeatedly performed many times for an edge that requires high-precision OPC, and the edge that does not require much OPC accuracy is executed. , The light intensity simulation and the generation of the OPC pattern are performed with a small number of repetitions.
C can be performed, and OPC can be performed with high accuracy and high speed.

According to a fifth aspect of the present invention, claims 1 to 4 are provided.
In the optical proximity correction method according to any one of
The mask pattern edge dividing step includes a step of dividing each edge included in each mask pattern based on a distance from a vertex of the mask pattern.

Since the optical proximity effect varies depending on the distance from the vertex, it is possible to perform OPC with higher precision by performing correction according to the distance from the vertex.

In the sixth aspect of the present invention, claims 1 to 5 are provided.
In the optical proximity correction method according to any one of
The mask pattern edge dividing step is a step of dividing an edge included in each mask pattern when the distance between the edge and the opposing edge is within a predetermined distance or less. It is characterized by including.

Since the optical proximity effect differs depending on the distance to the opposing edge, according to this configuration, it is possible to perform OPC with higher precision by performing correction in accordance with this distance.

According to a seventh aspect of the present invention, claims 1 to 6 are provided.
In the optical proximity correction method according to any one of
The mask pattern edge dividing step is characterized in that, for each edge included in each mask pattern, if the length of the edge is equal to or longer than a certain length, the edge is split to be equal to or shorter than this length.

According to such a configuration, since the optical proximity effect varies depending on the length of the edge, it is possible to perform OPC with higher accuracy by performing correction according to this length.

According to an eighth aspect of the present invention, claims 1 to 7 are provided.
In the optical proximity correction method according to any one of
The mask pattern edge dividing step is a case where the distance between each edge included in each mask pattern and the opposing edge is equal to or less than a predetermined distance, and the length of the opposing edge is the length. It is characterized in that the edge is divided in the case of more than that.

According to this configuration, since the optical proximity effect varies depending on the length of the edge, a more accurate OPC can be performed by performing correction in accordance with the distance from the opposing edge.
Can be performed.

According to a ninth aspect of the present invention, claims 1 to 8 are provided.
In the optical proximity correction method according to any one of
In the mask pattern edge dividing step, for a shape in which 90-degree vertices sandwich an edge having a certain length or less among the edges included in each mask pattern, a certain length or less is used for a 90-degree vertex The edge which is opposite to the edge is divided by a certain length.

According to this configuration, since the optical proximity effect of the shorter edge and the opposing edge differs depending on how short the edge is, correction is performed in accordance with the length of the opposing edge and its distance.

According to a tenth aspect of the present invention, in the optical proximity correction method according to any one of the first to ninth aspects, the mask pattern edge dividing step is performed for each edge included in each mask pattern. And a step of dividing an edge at an intersection with an edge included in another mask pattern data.

Since the base step differs depending on the presence / absence of an intersection with another mask pattern data, particularly an edge included in a lower-layer mask pattern, a higher correction can be made by taking into account the presence / absence of this intersection. OPC with high accuracy can be performed.

In the eleventh light intensity simulation method according to the present invention, each edge included in each mask pattern of the semiconductor integrated circuit is divided into a mask pattern edge dividing step and each edge divided in the mask pattern edge classifying step is divided. A mask pattern edge classification step of classifying, a simulation model selection step of selecting a simulation model to be used when performing light intensity simulation for each edge classified by the mask pattern edge classification step, and a classification by the mask pattern edge classification step. The method is characterized by comprising a simulation execution step of performing a light intensity simulation using the simulation model selected in the simulation model selection step for the edge thus set.

According to the eleventh configuration of the present invention, edges are classified based on characteristics unique to each mask pattern, such as the shape and arrangement position of edges included in the mask pattern and the positional relationship with other mask patterns, and the classified edges are classified. A simulation model to be used for the light intensity simulation is determined every time. Therefore, high accuracy is required for edges that require high precision light intensity simulation, and high accuracy is required for edges that do not require high precision of light intensity simulation. The light intensity simulation is performed at a high speed using a low simulation model, so that the light intensity simulation can be performed with high accuracy and at a high speed for all the mask patterns included in the chip.

According to a twelfth aspect of the present invention, in the light intensity simulation method according to the eleventh aspect, the mask pattern edge classification step includes the step of determining an edge length, an edge arrangement position in a mask pattern, and other mask pattern data. It is a classification process that classifies based on the overlapping situation of

According to a thirteenth aspect of the present invention, claim 11 or claim 1 is provided.
2. In the light intensity simulation method according to 2, the mask pattern edge classification step further performs weighting according to the accuracy required of the pattern on an actual device,
The method includes a step of classifying.

In the fourteenth aspect of the present invention, claims 11 to 13
In the light intensity simulation method according to any one of the above, the simulation model selection step includes, for each edge classified in the mask pattern edge classification step, a repetition number determination step of determining the number of repetitions of the light intensity simulation, The simulation performing step is a step of repeatedly performing the light intensity simulation on the edges classified in the mask pattern edge classification step for the number of times determined in the repetition number determining step,
The OPC pattern generation step is a step of generating an OPC pattern for a mask pattern by comparing the positions of the edges before and after performing the simulation in the simulation execution step and moving the edges before performing the simulation based on the difference. Is performed for the number of times determined in the repetition number determination step.

According to the fourteenth aspect of the present invention, edges are classified based on characteristics unique to each mask pattern, such as shapes and arrangement positions of edges included in the mask pattern, and positional relationships with other mask patterns, and for each classified edge. Since the number of times the light intensity simulation is repeated is determined, the light intensity simulation is repeated many times for edges that require high-precision simulation results. Light intensity simulation can be performed with high accuracy and high speed for all mask patterns included in the chip by reducing the number of repetitions of intensity simulation.

According to a fifteenth aspect of the present invention, claims 11 to 14 are provided.
In the light intensity simulation method according to any one of the above, the mask pattern edge dividing step includes a step of dividing each edge included in each mask pattern based on a first distance from a vertex of the mask pattern. It is characterized by the following.

In the sixteenth aspect of the present invention, claims 11 to 15
In the light intensity simulation method according to any one of the above, the mask pattern edge dividing step may be configured such that, when a distance between each edge included in each mask pattern and an opposite edge is equal to or less than a second distance, The step of dividing an edge within a distance of.

In the seventeenth aspect of the present invention, claims 11 to 16
In the light intensity simulation method according to any one of the above, in the mask pattern edge dividing step, for each edge included in each mask pattern, if the edge length is equal to or greater than a certain length, the mask pattern edge division step It is characterized in that edges are divided.

In the eighteenth aspect of the present invention, claims 11 to 17 are provided.
In the light intensity simulation method according to any one of the above, the mask pattern edge dividing step is a case where a distance between each edge included in each mask pattern and an opposite edge is equal to or less than a predetermined distance, and When the length of the edge to be processed is equal to or longer than a predetermined length, the edge is divided.

In the nineteenth aspect of the present invention, claims 11 to 18 are provided.
In the light intensity simulation method according to any one of the above, the mask pattern edge dividing step may include a step of dividing the edge included in each mask pattern into a shape having 90 degrees of vertices sandwiching an edge having a predetermined length or less. An edge paired with an edge having a length equal to or less than a certain length with respect to the vertex of the degree is divided into edges having a certain length.

In the twentieth aspect of the present invention, claims 11 to 19 are provided.
In the light intensity simulation method according to any one of the above, the mask pattern edge dividing step may include dividing each edge included in each mask pattern at an intersection with an edge included in another mask pattern data. It is characterized by.

[0045]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a flowchart showing an optical proximity effect correction method according to a first embodiment of the present invention. 2 to 4 are explanatory diagrams of the optical proximity correction method according to the first embodiment of the present invention.

In the first embodiment of the present invention, as shown in FIG. 1, a mask pattern edge dividing step 1 for dividing each edge included in each mask pattern of a semiconductor integrated circuit, Are classified by the mask pattern edge classifying step 2 and the mask pattern edge classifying step, which classify each edge divided by the length of the edge, the arrangement position of the edge in the mask pattern, and the state of overlap with other mask pattern data. A simulation model selection step 3 for selecting a simulation model to be used when performing a light intensity simulation for each edge; and a simulation model selected in the simulation model selection step for the edges classified in the mask pattern edge classification step. A first simulation execution step 5 for performing a light intensity simulation using a light source and a position of an edge before and after the simulation is performed in the first simulation execution step are compared with each other. And a first OPC pattern generation step 7 for generating an OPC pattern for the mask pattern by moving.

As shown in FIG. 1, in the pattern edge division step 1, the shape of each mask pattern, the positional relationship with the adjacent mask pattern, and the positional relationship with other mask pattern data for the input mask pattern. The edge included in each mask pattern is divided according to.

More specifically, among the edges included in the mask pattern, a 90 ° vertex sandwiches an edge of a first length or less from an edge of a first length or less. The edge paired with the mask pattern is divided by the second length from the vertex of the mask pattern, or the edge is divided by the third length from the vertex of the mask pattern, or the distance from the opposing edge is the fourth length. In the following cases, the edge is divided inside and outside the fourth length, or when the edge length is equal to or longer than the fifth length, the edge is divided into the fifth length or less, or the opposite edge When the distance from the edge is less than or equal to the sixth length and the length of the opposing edge is greater than or equal to the seventh length, the edge is divided or the intersection with the edge included in other mask pattern data Split the edge with.

Here, the first to sixth predetermined values are values determined from factors such as lithography and RIE processes in semiconductor manufacturing. Resist type, thickness, ground conditions,
In addition to the exposure condition and the etching condition in RIE, and further considering what kind of film pattern is to be formed, a set condition having several values is determined in advance. Edge division is performed based on the edge division. For example, a required accuracy in device operation is different between a simple wiring pattern and a gate electrode pattern. As described above, it is necessary to perform classification according to a request on an actual device.

In the pattern edge classification step 2, edges are classified according to the conditions when the edges output in the pattern edge division step 1 are divided in the pattern edge division step 1.

In the simulation model selection step 3, a simulation model to be used for the light intensity simulation is determined for each edge classified in the pattern edge classification step 2 according to the accuracy of the light intensity simulation required for each edge. Since a high-precision simulation model is composed of multidimensional functions, the simulation requires a long time and the simulation speed is low. On the other hand, the simulation model with low accuracy uses a simplified function, so that the simulation speed is high and the simulation is completed in a short time.

Furthermore, in the simulation model selection step 3, specifically, among the edges classified in the pattern edge classification step 2, for example, the edges near the vertices of the mask pattern are difficult to completely correct by OPC. In addition, a simulation model that has low accuracy but can operate at high speed is used because even if it is completely corrected, the characteristics on the actual device are not greatly affected. Also, CM
In the case of a mask pattern of a polysilicon mask which forms a transistor on an actual device in an OS semiconductor, an overlap portion with a mask pattern indicating an active region forms a transistor gate, so that gate length and gate width greatly affect transistor characteristics. Therefore, it is necessary to perform high-precision OPC. In this case, an edge included in the polysilicon mask is divided at an intersection with an edge included in the mask pattern indicating the active region. Use

In the repetition number determination step 4, the number of repetitions of the simulation execution step and the OPC pattern generation step in the steps after the repetition number determination step 4 is determined for each edge classified in the pattern edge classification step 2. In the simulation-based OPC, generally, after performing a light intensity simulation, a mask pattern input to the pattern edge division step 1 is compared with a pattern obtained as a result of the light intensity simulation, and an OPC pattern is generated from the difference. By repeating the light intensity simulation and the generation of the OPC pattern, the difference between the mask pattern before the light intensity simulation is performed and the pattern obtained as a result of the light intensity simulation is reduced, and highly accurate OPC is realized. The smaller the number of repetitions of the light intensity simulation and the generation of the OPC pattern, the lower the accuracy of the OPC.

In step 4 for determining the number of repetitions, specifically, when the intervals between edges of all edges included in the mask pattern among the edges classified in pattern edge classification step 2 are smaller than or equal to the second distance, and When each edge is included in the same mask pattern, that is, in the case of an edge constituting a narrow line width in the wiring mask pattern, unless the OPC is performed with high accuracy, a disconnection occurs on the actual device, which causes adverse effects in the actual device. The number of repetitions of the light intensity simulation and the OPC pattern generation is increased. Further, when the distance between the edges of all the edges included in the mask pattern is equal to or greater than the second distance, and when each edge is included in the same mask pattern, that is, when the edge forming the thick line width in the wiring mask pattern is used. In this case, even if the mask pattern is deteriorated in the actual device because the accuracy of the OPC is low, the ratio of the deterioration to the line width is lower than that in the case of the thin line width and does not have a large adverse effect on the actual device. To reduce the number of repetitions.

In the simulation execution step 5, the simulation model selected in the simulation model selection step 3 is used for each edge classified in the pattern edge classification step 2, and the light intensity is determined by the number of repetitions determined in the repetition number determination step 4. Perform a simulation.

In the OPC pattern generation step 6, the mask pattern input in the pattern edge division step 1 is compared with the pattern obtained by performing the light intensity simulation in the simulation execution step 5, and the difference is used as the OPC pattern. The mask pattern generated and input to the simulation
A PC pattern is added.

Finally, in the OPC pattern output step 7, the OPC pattern output in the OPC pattern generation step 6 is output.

FIG. 2 is a diagram showing an example of the mask pattern 20 of the semiconductor integrated circuit. Reference numerals 21 to 23 denote mask patterns for performing OPC. These mask patterns 2
Reference numerals 1 to 23 denote mask patterns of polysilicon masks constituting transistors on actual devices in a CMOS semiconductor. Reference numeral 24 denotes a mask pattern of mask data different from the mask patterns 21 to 23,
3 shows a mask pattern indicating an active region constituting a transistor on an actual device in a semiconductor.

FIG. 3 is a diagram showing the pattern edge dividing step 1, wherein P1 to P20 indicate the points of division when the mask pattern 21 is divided in the pattern edge dividing step 1, and E1 to 20 are the divided points. Indicates the edge that was set.
P30 to P49 show division points when the mask pattern 22 is divided in the pattern edge division step 1.
E30 to 51 indicate the divided edges. P60-65
Is the pattern edge division step 1 of the mask pattern 23.
Indicates the points of division at the time of division at E60 to E66.
Indicates a divided edge.

FIG. 4 is a diagram showing the O in the first embodiment of the present invention.
It is a figure which shows the example of the mask pattern 40 output after PC implementation, and 41-43 has shown the mask pattern after performing OPC.

Here, as shown in Table 1 below, classification 1-
The edges grouped into 10 are classified according to the simulation accuracy and the number of repetitions of the simulation model to be used, and a simulation according to the necessity of each group is performed.

[0063]

[Table 1]

Table 1 shows edges E1 to E20, E30 to 51, and E60 to 66 classified in pattern edge classification step 2.
3 shows a list of the models selected in the simulation model selection step 3 and the number of repetitions determined in the repetition number determination step 3.

Next, an optical proximity effect correction method according to the first embodiment will be described. First, the semiconductor integrated circuit 20 is input to the pattern edge division step 1. The pattern edge division step 1 first includes the mask pattern 2
The edges 0 to 23 are divided by the following (conditions 1 to 6).

(Condition 1) A 90-degree vertex sandwiches an edge having a predetermined length equal to or less than the first length.
Determining division points P3, P20, P61, and P64 for dividing an edge paired with an edge having a length equal to or less than the first length from the vertex of the mask pattern by a length of a second predetermined value from the vertex of the degree; Edges E1, E2, E20, E62, E63, E
64 is output. Here, the edges E1 and E63 are the first
, And edges E2, E20, E6
2, E64 is the second length.

(Condition 2) Third from the top of the mask pattern
Points P5, P8, P9, P that divide an edge by the length
11, P12, P14, P15, P18, P32, P3
4, P37, P39, P42, P44, P45, P47
Are determined, and edges E5, E6, E7, E9, E10, E
12, E13, E15, E16, E17, E33, E3
4, E38, E39, E44, E45, E47, E48
Is output. Here, edges E5, E7, E9, E10,
E12, E13, E15, E17, E33, E34, E
38, E39, E44, E45, E47 and E48 are the third length.

(Condition 3) The distance from the opposing edge is the fourth
When the length is equal to or less than the length, the dividing points P35 and P36 for dividing the edge inside and outside the fourth length are determined, and the edge E36 is output. Here, between the dividing points P35 and P63, P36 and P
The length between 62 is equal to or less than the fourth length.

(Condition 4) When the length of the edge is equal to or longer than the fifth length, the dividing point P41 for dividing the edge to the length equal to or shorter than the fifth length is determined, and the edges E42 and E43 are output. Here, the total length of the edges E42 and E43 is equal to or greater than the fifth length, and each of the edges E42 and E43 is equal to or less than the fifth length.

(Condition 5) The distance from the opposing edge is the sixth
And the division point P40 for dividing the edge when the length of the facing edge is equal to or greater than the seventh length.
And outputs an edge E41. Here, branch point P4
The length between 0 and P62 is equal to or less than the sixth length, and the length of the edge E41 is equal to or less than the seventh length.

(Condition 6) A division point P3 for dividing an edge at an intersection with an edge included in other mask pattern data,
P4, P19, P20, P30, P31, P48, P4
9, P60, P61, P64, and P65 are determined, and edges E3, E14, E31, E50, E61, and E65 are output.

Next, in the pattern edge classification step 2, the edges divided in the pattern edge division step 1 are classified in the following (classification 1 to 7).

(Category 1) Pattern edge division step 1
Among the edges divided under (condition 1), edges E1, E2, and E20 having no other opposing edge in a range equal to or less than the fourth length shown in (condition 3).

(Category 2) Pattern edge division step 1
Among the edges divided under (condition 1), edges E62, E63, and E64 having other opposing edges in a range equal to or less than the fourth length shown in (condition 3).

(Category 3) Pattern edge division step 1
Edges E5, E6, E of the edges divided in (condition 2)
7, E9, E10, E12, E13, E15, E16,
E17, E33, E34, E38, E39, E44, E
45, E47, E48.

(Category 4) Pattern edge division step 1
The edge E36 divided in (condition 3).

(Category 5) Pattern edge division step 1
Edges E42 and E43 divided by (condition 4).

(Category 6) Pattern edge division step 1
The edge E41 divided in (condition 5).

(Category 7) Pattern edge division step 1
Edges E3, E19, E31,
E50, E61, E65.

(Category 8) The distance from the opposite edge in the same mask pattern is not less than the seventh length, and the length of the edge is not less than the eighth length, that is, not less than the eighth length. Edges E35, E37, E46 forming a linear mask pattern having a length equal to or greater than the seventh length.

(Category 9) The distance from the opposing edge in the same mask pattern is greater than or equal to the seventh length, and the length of the edge is less than or equal to the eighth length, that is, the seventh rectangular shape. Edges E8, E11, E14 forming a mask pattern having a size equal to or greater than the length.

(Category 10) Edges E4, E18, E30, E32, E40, E that do not fall under (Categories 1 to 9)
49, E51. Here, the seventh and eighth values are values determined from factors such as lithography and RIE processes in semiconductor manufacturing.

Next, a simulation model selection step 3 selects a simulation model according to the accuracy of the light intensity simulation required for each classification for the edges (classifications 1 to 10) classified in the pattern edge classification step 2. .

The edge of (Category 1) is an edge located at the end of the line shape, and the deformation from the image of the mask pattern in the optical lithography process and the RIE process at this portion is the end of the line (hereinafter referred to as the end). Line end), so that on the actual device, for example, the line end of the polysilicon mask pattern recedes, and the line end of the polysilicon mask pattern enters the mask pattern of the active region, and the transistor gate is formed on the actual device. Since there is a possibility that a problem of being unable to be formed may occur, high-precision OPC is required.

The edge of (Category 2) constitutes the line end of the mask pattern in the same manner as the edge of (Category 1).
Edge exists. This portion has a smaller pattern variation margin allowed in semiconductor device manufacturing than the (Category 1) portion.
Unless OPC is performed with higher precision than PC, the line end of the mask pattern 23 and the mask pattern 22 may be short-circuited on the actual device. Therefore, a simulation model with relatively high simulation accuracy is used for the edge of (class 1), and a simulation model with even higher accuracy is used for the edge of (class 2).

Similarly, the OPC required for each part of the mask pattern on the actual device for (Categories 3 and 4)
Simulation model is selected according to the accuracy of.

Next, a simulation model selecting step 3 selects a simulation model according to the accuracy of the light intensity simulation required for each classification for the edges (classifications 1 to 10) classified in the pattern edge classification step 2. .

The edge of (Category 1) is the edge located at the end of the line shape, and the deformation of the mask pattern from the image in the optical lithography process and the RIE process at this portion is the end of the line (hereinafter referred to as the line end). Line end), the line end of the polysilicon mask pattern recedes on the actual device, and the line end of the polysilicon mask pattern enters the mask pattern of the active region, so that a transistor gate can be formed on the actual device. OPC with high accuracy is required because of the problem of disappearance.

The edge of (Category 2) constitutes the line end portion of the mask pattern in the same manner as the edge of (Category 1).
Edge exists. Therefore, unless OPC is performed with higher accuracy than the OPC performed for the edge of (class 1), the line end of the mask pattern 23 and the mask pattern 22 may be short-circuited on the actual device.

Therefore, a simulation model with relatively high simulation accuracy is used for the edge of (class 1), and a simulation model with even higher accuracy is used for the edge of (class 2).

Similarly, OPs required for each part of the mask pattern on the actual device for (Categories 3 to 10)
A simulation model is selected according to the accuracy of C.

Next, in the number-of-repetitions determination step 4, classification is performed in the pattern edge classification step 2 (classifications 1 to 10).
The number of repetitions of the simulation execution step 5 and the OPC pattern generation step 6 is determined according to the OPC accuracy required for each classification for the edge of.

The edge of (Category 7) has a line width of a line shape (hereinafter referred to as CD: Critical Dimension).
n), the deformation from the image of the mask pattern in the optical lithography process and the RIE process in this region appears as a CD deformation,
For this reason, on the actual device, the gate length of the transistor gate is different from that at the time of gate design, and there is a problem in that the transistor characteristics are changed.
The edge of (Category 10) forms a CD similarly to the edge of (Category 7), but this edge does not constitute a transistor gate and serves as a wiring pattern. There is no problem with low accuracy OPC.
Therefore, since the edge of (Category 7) needs high-precision OPC, it is necessary to increase the number of repetitions of the simulation execution step 5 and the OPC pattern generation step 6, and the edge of (Category 10) has a low-precision OPC. Therefore, the number of repetitions of the simulation execution step 5 and the OPC pattern generation step 6 may be small.

Similarly, for (classifications 1 to 6, 8, and 9), simulation execution step 5 and OPC pattern generation step 6 are performed according to the OPC accuracy required for each part of the mask pattern on the actual device. Determine the number of repetitions.

Table 1 shows the accuracy of the simulation model used for the edges of (classification 1 to 10) and the number of repetitions of the simulation execution step 5 and the OPC pattern generation step 6.

Next, in simulation execution step 5,
Light intensity simulation is performed using the simulation model selected in the simulation model selection step 3 for each edge classified in the pattern edge classification step.

Next, in the OPC pattern generation step 6,
The mask pattern input to the pattern edge division step 1 is compared with the pattern obtained by performing the light intensity simulation in the simulation execution step 5, and the difference is generated as an OPC pattern. To the OPC pattern.

For the edges classified in the pattern edge classification step 2, the simulation execution step 5 is performed by the number of times determined in the repetition number determination step 4;
The OPC pattern generation step 6 is repeatedly performed, and finally, the mask pattern after the OPC is performed is output from the OPC pattern output step 7.

As described above, according to the first embodiment of the present invention, the edge shape and the arrangement position of the edge included in the mask pattern, the positional relationship with other mask patterns, etc. And a simulation model to be used for the light intensity simulation is determined for each of the classified edges. Since the high-precision simulation model is composed of multidimensional functions, the simulation speed is low. On the other hand, a simulation model with low accuracy simplifies the function, so that the simulation speed is high. Therefore, the light intensity simulation is performed at a high speed by using a simulation model with low accuracy for an edge requiring high accuracy OPC, and at a low accuracy for an edge requiring low OPC accuracy. Thus, high-precision and high-speed OPC can be performed on all mask patterns included in the chip.

Edges are classified based on characteristics unique to each mask pattern, such as the shape and arrangement position of edges included in the mask pattern and the positional relationship with other mask patterns, and light intensity simulation and OPC pattern generation are performed for each classified edge. Is determined the number of times to repeat
Light intensity simulation and OPC pattern generation are performed repeatedly for edges that require PC, and light intensity simulation and OPC pattern generation are repeated for edges that do not require very high OPC accuracy. This makes it possible to perform OPC with high accuracy and high speed on all mask patterns included in the chip.

(Embodiment 2) Next, a light intensity simulation method using the method of the present invention will be described. FIG. 5 is a flowchart showing a light intensity simulation method according to the second embodiment of the present invention. In this method, FIGS.
In the optical proximity correction method according to the first embodiment of the present invention, the simulation is executed up to the simulation execution step 5, and the obtained simulation result is output in the simulation pattern output step 50, whereby the light intensity simulation is executed. Is done.

Next, a light intensity simulation method according to the second embodiment of the present invention will be described with reference to FIG. Steps 1 to 5 in FIG. 5 are the same as steps 1 to 5 shown in FIG. 1 in the first embodiment of the present invention. Reference numeral 50 denotes a file indicating the result of the simulation performed in the simulation execution step 5 or output as image data.

The dividing and classifying method is described in FIG. 2 and FIG.
Since the third embodiment is the same as the first embodiment shown in FIG.

FIG. 6 shows image data 60 showing an example of a light intensity simulation result output after the light intensity simulation according to the second embodiment of the present invention. Reference numerals 61 to 63 denote mask patterns 21 to 63 in FIG. 23 shows an output pattern after a light intensity simulation has been performed on 23.

Next, the light intensity simulation operation of the second embodiment will be described. The operation is started by first inputting the semiconductor integrated circuit 20 to the pattern edge division step 1, but as apparent from the comparison with FIG.
The steps up to the simulation execution step 5 are executed in the same manner as in the first embodiment. Then, as in the first embodiment, in the simulation execution step 5, a light intensity simulation is performed using the simulation model selected in the simulation model selection step 3 for each edge classified in the pattern edge classification step.

At this time, the simulation execution step 5 is performed on the edges classified in the pattern edge classification step 2 by the number determined in the repetition number determination step 4.
Are repeated, and finally, in the simulation result output step 50, the mask pattern after the OPC is performed is output.

As described above, according to the second embodiment of the present invention, the edge shape included in the mask pattern, the arrangement position thereof, the positional relationship with other mask patterns, etc., are used to determine the edge characteristics of each mask pattern. And a simulation model to be used for the light intensity simulation is determined for each of the classified edges.

Here, a high-accuracy simulation model is composed of multidimensional functions, so that the simulation speed is slow, and a low-accuracy simulation model has a simplified function, so that the simulation speed is high. Accordingly, the light intensity simulation is performed at high speed by using a simulation model with high accuracy for edges that require high-precision light intensity simulation, and at a low accuracy for edges that do not require much accuracy of light intensity simulation. . As a result, the light intensity simulation can be performed with high accuracy and at high speed for all the mask patterns included in the chip.

Edges are classified based on characteristics unique to each mask pattern, such as the shape and arrangement position of the edges included in the mask pattern and the positional relationship with other mask patterns, and the number of times light intensity simulation is repeated for each classified edge Is determined, the light intensity simulation is repeated more times for edges where high-precision light intensity simulation is required, and the light intensity simulation is performed for edges where less accurate light intensity simulation is required. Adjustments such as performing a small number of simulation iterations are possible. As a result, a high-precision and high-speed light intensity simulation can be executed for all the mask patterns included in the chip.

[0110]

As described above, according to the present invention, each edge included in the mask pattern is determined by determining the length of the edge, the arrangement position of the edge in the mask pattern, the state of overlap with other mask pattern data, and the like. The simulation model used for the light intensity simulation is changed according to the accuracy required when performing the OPC or the light intensity simulation from the features, and the number of repetitions of the light intensity simulation is changed. A simulation-based OPC can be performed on all mask patterns, and an optical proximity effect correction method that suppresses an increase in the time required for OPC can be provided.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating an optical proximity effect correction method according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram of an optical proximity effect correction method according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram of an optical proximity effect correction method according to the first embodiment of the present invention.

FIG. 4 is an explanatory diagram of an optical proximity effect correction method according to the first embodiment of the present invention.

FIG. 5 is a flowchart illustrating a light intensity simulation method according to a second embodiment of the present invention.

FIG. 6 is an image data diagram showing a light intensity simulation result according to the second embodiment of the present invention.

FIG. 7 is an explanatory diagram of a conventional optical proximity correction method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Pattern edge division step 2 ... Pattern edge classification step 3 ... Simulation model selection step 4 ... Repetition number determination step 5 ... Simulation execution step 6 ... OPC pattern output step 20 ... Mask pattern example of semiconductor integrated circuit 21-23 ... Mask pattern 24: mask pattern of mask data different from mask patterns 21 to 23 P1 to P20: division points of mask pattern 21 P30 to P49: division points of mask pattern 22 P60 to P65: division points of mask pattern 23 E1 to E20: mask Edges of pattern 21 E30 to E51 E60 to E66 of mask pattern 21 Edge 40 of mask pattern 21 OPC output patterns 41 to 43 according to the first embodiment of the present invention 41 to mask patterns 21 to 23 Output pattern after OPC is performed according to the first embodiment of the present invention 50: Simulation pattern output step 60: Light intensity simulation output pattern according to the second embodiment of the present invention 61 to 26: The present invention for mask patterns 21 to 23 Light intensity simulation output pattern according to the second embodiment of the present invention

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Seiji Mukai 1006 Kazuma Kadoma, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. F-term (reference) 2H095 BB01 5B046 AA08 BA05 DA02 FA07 GA06 JA04

Claims (20)

[Claims] A mask pattern edge dividing step of dividing an edge included in each mask pattern of the semiconductor integrated circuit; a mask pattern edge classifying step of classifying each edge divided in the mask pattern edge dividing step; A simulation model selecting step of selecting a simulation model to be used when performing light intensity simulation for each edge classified by the mask pattern edge classifying step; and for the edge classified by the mask pattern edge classifying step, the simulation model selecting step. A simulation execution step of performing light intensity simulation using the simulation model selected by the simulation, and comparing the positions of edges before and after the simulation is performed in the simulation execution step, and performing simulation based on the difference. Optical proximity correction with respect to the mask pattern by moving the front edges of performing ® emission (OP
C) an OPC pattern generation step of generating a pattern. 2. The mask pattern edge classification step,
2. The optical system according to claim 1, wherein the classification is performed based on at least one of an edge length, an edge arrangement position in the mask pattern, and an overlapping state with other mask pattern data. 3. Proximity effect correction method. 3. The mask pattern edge classification step,
3. The optical proximity correction method according to claim 1, further comprising the step of weighting and classifying the patterns according to the accuracy required on an actual device. 4. The simulation model selecting step includes a repetition number determining step of determining a repetition number of a light intensity simulation for each edge classified in the mask pattern edge classification step. The light intensity simulation is repeated for the edges classified in the pattern edge classification step for the number of times determined in the repetition number determination step. The OPC pattern generation step includes an edge before and after performing the simulation in the simulation execution step. Generating the OPC pattern for the mask pattern by moving the edge before performing the simulation based on the difference, and performing the number of times determined in the repetition number determining step. Special Optical proximity correction method according to any one of claims 1 to 3,. 5. The mask pattern edge dividing step,
5. The optical proximity correction method according to claim 1, further comprising a step of dividing each edge included in each mask pattern based on a distance from a vertex of the mask pattern. 6. The mask pattern edge dividing step,
Claims: A method comprising: for each edge included in each mask pattern, a step of dividing an edge within a range of the distance when a distance between the edge and a facing edge is equal to or less than a predetermined distance. Item 6. The optical proximity effect correction method according to any one of Items 1 to 5. 7. The mask pattern edge dividing step includes:
For each edge included in each mask pattern, if the length of the edge is equal to or greater than a predetermined length, the length of each divided edge is set to be equal to or less than this length. 7. The optical proximity correction method according to claim 1, further comprising a dividing step. 8. The mask pattern edge dividing step includes:
For each edge included in each mask pattern, the distance from the opposing edge is equal to or less than a predetermined distance, and the length of the opposing edge is equal to or more than a predetermined length. 8. The optical proximity correction method according to claim 1, further comprising a step of dividing an edge in the case. 9. The mask pattern edge dividing step includes:
Among the edges included in each mask pattern, a shape in which a 90-degree vertex sandwiches an edge having a predetermined length or less is opposed to an edge having a predetermined length or less with respect to the 90-degree vertex. 2. The method according to claim 1, further comprising the step of dividing the edge into predetermined lengths.
9. The optical proximity effect correction method according to any one of claims 1 to 8. 10. The mask pattern edge dividing step includes a step of dividing an edge included in each mask pattern at an intersection with an edge included in another mask pattern data. Item 10. The optical proximity correction method according to any one of Items 1 to 9. 11. A mask pattern edge dividing step of dividing each edge included in each mask pattern of the semiconductor integrated circuit, and a mask pattern edge classifying step of classifying each edge divided in the mask pattern edge classifying step. A simulation model selecting step of selecting a simulation model to be used when performing light intensity simulation for each edge classified by the mask pattern edge classifying step; and a simulation model selecting step for the edges classified by the mask pattern edge classifying step. A light intensity simulation method comprising a simulation execution step of performing a light intensity simulation using the simulation model selected in (1). 12. The mask pattern edge classification step is a classification step of performing classification based on at least one of an edge length, an edge arrangement position in a mask pattern, and an overlapping state with other mask pattern data. 12. The light intensity simulation method according to claim 11, wherein: 13. The light according to claim 11, wherein the mask pattern edge classifying step further includes a step of weighting and classifying the pattern according to accuracy required on a real device of the pattern. Strength simulation method. 14. The simulation model selection step includes a repetition number determination step of determining a repetition number of light intensity simulation for each edge classified in the mask pattern edge classification step. The light intensity simulation is repeated for the edges classified in the pattern edge classification step for the number of times determined in the repetition number determination step. The OPC pattern generation step includes an edge before and after performing the simulation in the simulation execution step. Generating the OPC pattern for the mask pattern by moving the edge before performing the simulation based on the difference, and performing the number of times determined in the repetition number determining step. Light intensity simulation method according to any one of claims 1 to 3, symptoms. 15. The mask pattern edge dividing step includes a step of dividing each edge included in each mask pattern based on a distance from a vertex of the mask pattern. The light intensity simulation method according to any one of the above. 16. The mask pattern edge dividing step includes, when a distance between each edge included in each mask pattern and an opposite edge is equal to or less than a predetermined distance, is within the range of the distance. 16. The light intensity simulation method according to claim 11, further comprising a step of dividing an edge. 17. The method according to claim 17, wherein in the mask pattern edge dividing step, when an edge length of each edge included in each mask pattern is equal to or longer than a predetermined length, the length of each divided edge is determined. 17. The light intensity simulation method according to claim 11, further comprising a step of dividing an edge so that the length is equal to or less than the length. 18. The method according to claim 18, wherein the mask pattern edge dividing step is performed when a distance between each edge included in each mask pattern and the opposing edge is equal to or less than a predetermined distance. 18. The light intensity simulation method according to claim 11, further comprising a step of dividing an edge when the length is equal to or longer than a predetermined length. 19. The method according to claim 19, wherein, in the mask pattern edge dividing step, a 90-degree vertex is sandwiched between a 90-degree vertex and an edge having a predetermined length or less among the edges included in each mask pattern. 19. The light intensity simulation method according to claim 11, further comprising a step of dividing an edge facing the edge having a length equal to or less than the predetermined length into a predetermined length. 20. The mask pattern edge dividing step includes a step of dividing an edge included in each mask pattern at an intersection with an edge included in other mask pattern data. Item 20. The light intensity simulation method according to any one of Items 11 to 19.