JP4830135B2 - How to selectively modify layout patterns - Google Patents

Description

  The present invention relates to a method for correcting a layout pattern, and more particularly to a method for selectively correcting a layout pattern.

  In the semiconductor manufacturing process, basic technologies such as photolithography and etching are often used. Photolithographic techniques typically include the step of transferring a complex integrated circuit pattern onto the surface of a semiconductor wafer for processes such as etching and implantation. This pattern must be very precise to form a fine integrated circuit that matches the pattern from previous and subsequent steps.

  In the photolithography process, when the pattern on the reticle is transferred to the wafer surface, the deviation often affects the performance of the semiconductor device. Such deviations are usually related to the characteristics of the pattern to be transferred, the topology of the wafer, the light source, and various process parameters.

  There are many methods for verifying, correcting, and compensating for deviations due to optical proximity effects, process rules, and lithography rules in order to improve image quality after transfer. Some of them are called optical proximity correction (OPC), process rule check (PRC), and lithography rule check (LRC). Commercially available OPC software typically inspects issues such as layout pattern pitch, bridges, critical dimension uniformity, and the like. Such software uses the theoretical image to correct the standard layout pattern on the reticle and create an accurate exposure image pattern on the wafer. This method not only inspects the problem of the layout pattern, but also corrects the layout pattern on the reticle using a theoretical image. If the corrected image pattern is available, it is output and used to manufacture a reticle to create an accurate image pattern on the wafer.

  In summary, the verification, correction, and compensation methods already have an operation process established as a standard. For example, in a conventional process for verifying a layout pattern on a reticle by OPC, a layout pattern is first input, and then a Boolean pretreatment is performed on the layout pattern to obtain a preliminary layout pattern. Thereafter, OPC is executed to correct any specific pattern, and PRC and LRC are executed separately, followed by error screening and checking. If the acquired pattern is accurate and usable, this is output. Otherwise, the pattern correction is executed again, and when no error is detected, the pattern is output.

  Therefore, in the semiconductor manufacturing process, the process of obtaining a usable layout pattern by correcting the layout pattern with the OPC model and refining the pattern transfer is an important operation process.

  One object of the present invention is to provide a method for selectively modifying a layout pattern.

  In the method according to the present invention, a first layout pattern including at least a first group and a second group is provided. Each of the first group and the second group includes a plurality of members. Next, a simulation process and a correction process are executed for all members of the first group and the second group, respectively, and finally all the members of the first group and the second group are made to achieve the target. Such selective modification of the layout pattern, which executes the simulation process and the modification process for the first group and the second group in a different manner, improves the accuracy and speed of the conventional OPC method, and the available layout patterns. Can be obtained.

  In the method of selectively correcting a layout pattern according to the present invention, first, a first layout pattern including at least a first group and a second group is provided. Each of the first group and the second group includes a plurality of members. Next, a simulation process and a first group correction process are executed for each member of the first group, and the corrected first group is acquired. Further, the simulation process and the second group correction process are executed for the members of the second group, respectively, and the corrected second group is acquired. The modified first group and the modified second group are then verified until both the modified first group and the modified second group achieve the goal. Later, a second layout pattern including a modified first group and a modified second group that have both achieved the goal is output. If necessary, it can be verified in advance whether both the first group and the second group have achieved their goals. Alternatively, the priority order selection process is performed on the first group and the second group, both of which have not achieved the target, and the priority group and the subordinate group are acquired.

  In the method according to the present invention, before performing the actual correction process or priority selection process, the plurality of groups are classified into the first group and the second group, so that all members of the first group and the second group are classified. When executing the simulation process and the correction process, respectively, the independent simulation process and the correction process can be executed separately for all members of the first group, the second group / priority group, and the subordinate group. Then, a more suitable simulation process and correction process can be performed for all members of the first group and the second group having different attributes, so that all the members of the first group and the second group are quickly set to the target. Can be achieved. Such a selective modification of the layout pattern can improve the accuracy and speed of the conventional OPC method and obtain a usable layout pattern.

  The foregoing and other objects of the invention will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment as illustrated in the various drawings.

  The present invention relates to a method for selectively modifying a layout pattern. Prior to performing the actual correction process, the different groups are classified into a first group and a second group, or a priority selection process is performed first if necessary. Later, independent simulation and correction processes are performed for all members of the first group and the second group, respectively. In the method according to the invention, the priority selection process divides the groups into priority groups and subordinate groups, so that all members of the priority and subordinate groups achieve their goals as quickly as possible. Such selective modification of the layout pattern can improve accuracy and speed when acquiring a layout pattern that can be used in the conventional OPC method.

FIG. 1 is a flowchart showing a main flow of a method for selectively correcting a layout pattern according to the present invention. The layout pattern selective correction method 100 according to the present invention includes the following steps.
Step 110: Provide a first layout pattern.
Step 120: Verify whether the first group and the second group have achieved their goals as necessary.
Step 130: A priority selection process is performed on the first group and the second group that have not achieved the target as necessary to obtain the first group and the second group.
Step 140: A simulation process and a first group correction process are executed for all members of the first group, respectively, and a corrected first group is obtained.
Step 150: A simulation process and a second group correction process are executed for all members of the second group, respectively, and a corrected second group is obtained.
Step 160: Validate both the modified first group and the modified second group repeatedly until the goal is achieved.
Step 170: Output a second layout pattern including a modified first group that has achieved the goal and a modified second group that has achieved the goal.

  The first layout pattern in step 110 is a pattern to be transferred, such as a contact hole pattern, a via hole pattern, a doped region, a polysilicon gate, or a SRAM (Static Random Access Memory). Alternatively, the first layout pattern may be a preliminary layout pattern subjected to a conventional OPC method. The first layout pattern includes at least a first group and a second group having different attributes. Each of the first group and the second group includes a plurality of members.

  A preferred embodiment illustrating the relationship between a given layout pattern group and member is described below. Please refer to FIG. FIG. 2 shows a given layout pattern including a plurality of groups and members belonging to them. For example, the layout pattern includes a plurality of templates 210/211. Each template 210, 211 includes a plurality of shapes 220, 221, 222, 223, 224, 225, 226. Each shape 220, 221, 222, 223, 224, 225, 226 includes at least one edge. For example, the shape 221 includes at least representative edges 231, 232, 233, 234. Desirably, the individual edges 231, 232, 233, 234 are either perpendicular to each other or sandwich a 135 degree angle. In the operation of the OPC model, each edge can be divided into a plurality of segments 241, 242, 243, 244 as necessary. Each segment 241, 242, 243, 244 is regarded as a minimum operation unit of the selective OPC model according to the present invention.

  As shown in FIG. 3, the edges 231 and 232 perpendicular to each other can be regarded as a group having different attributes, and the edges 231 and 232 parallel to each other can be regarded as a group having the same attribute. Therefore, the edges 231 and 232 are divided into separate groups, that is, a first group 261 and a second group 262. Therefore, the segments 241 and 243 belonging to the edges 231 and 233 are regarded as members 271 and 273, and the segments 242 and 244 belonging to the edges 232 and 234 are regarded as members 272 and 274.

  In other words, the first layout pattern 200 includes a first group 261 and a second group 262. The first group 261 includes members 271 and 273, and the second group 262 includes members 272 and 274. On the other hand, in the layout pattern 200, the edges parallel to any of the edges 231 and 232 are all divided into the same group.

  In the lithography process in which the pattern on the reticle is transferred to the wafer surface, deviation occurs due to process parameters such as light source or pattern characteristics. For example, when an asymmetric light source such as QUASAR irradiation is used to transfer a standard pattern on a reticle to the wafer surface, edges 232/231 perpendicular to each other in the same shape have different responses to the same QUASAR irradiation. Results in different sensitivities and results. For example, different exposure intensities can occur, i.e. first and second intensities, or different contrasts, i.e. first contrast and second contrast. As shown in FIG. 4, one of them is sensitive to irradiation and the other is insensitive (ie, insensitive to irradiation). Therefore, the simulation process and the correction process can be executed independently for the edges 312/313 having different sensitivity attributes, and the different simulation criteria and correction criteria can be distinguished. For example, a group with good sensitivity can obtain a clear result with a slight modification, and a group with low sensitivity needs a larger correction to obtain a similar result.

  Next, at step 120, the method according to the present invention first verifies if both the first group 261 and the second group 262 of the first layout pattern 200 have achieved the goal, if necessary. “Achieving the goal” may vary depending on the OPC criteria. For example, “achieving the goal” can mean that the group is within the tolerance of critical dimension errors. The concept of critical dimension error is as shown in FIG. FIG. 3 illustrates a circular pattern as a contact hole pattern. The contact hole pattern is desirably 84 nm × 84 nm in size. However, after the OPC simulation is performed, the dimension of the pattern is evaluated as 89 nm × 76 nm. Since the absolute value of the long side | 76−84 | = 8> the absolute value of the short side | 89−84 | = 5, it is determined whether the group is within the tolerance of the critical dimension error based on the large value of 8 nm. to decide.

Other parameters can also be used to verify whether the first group and the second group of the first layout pattern achieve the goal. FIG. 4 shows exposure intensities of a plurality of shapes in the layout pattern. The layout pattern 301 has a plurality of shapes 311, 312, 313, and 314. Contrast intensity = (maximum intensity value−minimum intensity value) / (maximum intensity value + minimum intensity value), the first contrast intensity between shapes 312 and 313 is 0.5476, and the second contrast between shapes 313 and 314 If the contrast intensity is set to 0.3122, it can be used as a criterion for “target achievement” that a parameter (that is, intensity difference) relating to both intensities falls within a specific range (for example, 1 × 10 ⁻⁴ ). In this case, the difference between the first contrast intensity and the second contrast intensity is 0.2354, which is clearly not within the specific range.

  If both the first group and the second group of the first layout pattern achieve the goal, all the members of the first group and the second group of the first layout pattern meet the predetermined requirements. The pattern can be output directly.

  However, in most cases, both the first group and the second group of the first layout pattern cannot achieve the goal. In this case, the method according to the present invention proceeds to step 130, if necessary, and performs a priority selection process on the first group and the second group, both of which have not achieved the goal, and divides it into a priority group and a subordinate group. .

  It should be noted that the sensitivity attributes of the first group and the second group are not known until this stage, so the priority selection process is performed first for the first group and the second group, Can be divided into groups.

  As described above, when the standard pattern on the reticle is transferred to the surface of the wafer by using QUASAR irradiation, the sensitivity of the edge 232/231 to QUASAR irradiation can be used as a criterion for priority selection. Edges 231 and 232 are divided into “sensitive” and “insensitive” edges based on the relative position with respect to QUASAR irradiation, with sensitive (sensitive) priority groups as necessary, and insensitive (insensitive) subordinated Can be a group.

  For example, in FIG. 4, since the first contrast intensity 0.5476 is higher than the second contrast intensity 0.3122, the first contrast intensity 0.5476 is set as the priority group and the second contrast intensity 0.3122 is set as the subordinate group.

  After determining the priority group and the subordinate group, the method proceeds to step 140 and performs a simulation process and a priority group correction process on all members of the priority group, respectively, to obtain a corrected priority group. Since the priority group has been selected in the priority selection process, all members of the priority group should have the same specific attributes. Therefore, if the simulation process is performed for all members of the priority group before the correction process, a more stable simulation result can be obtained under the same conditions.

  If steps 120/130 were not performed, the method proceeds to step 140, where a simulation process and a first group modification process are performed for all members of the first group, respectively, to obtain a modified first group. To do. Since the first group has been selected in the priority selection process, all members of the first group should have the same specific attributes. Therefore, if the simulation process is performed on all members of the first group before the correction process, a more stable simulation result can be obtained under the same conditions.

  For example, as shown in FIG. 2, the edges 232/231 are perpendicular to each other. If the edge 231 is set to “sensitive” and the edge 232 is set to “insensitive” through the priority selection process, the sensitive edge 231 is set as a priority group and the insensitive edge 232 is set according to the above “classification principle”. A subordinate group. The simulation process and the priority group modification process are performed on “all” members 271, 273 (ie, edge 231) of the first group 261. Generally, the simulation process is performed before the priority group modification process.

  In the simulation process, the “adjustment direction” and “correction weight” of each member (ie, members 271 and 273) are simulated. The “adjustment direction” is determined to be outward (inward) or inward (inward) depending on the “boundary” (ambit) of the member. The outward adjustment is to increase the area, and the inward adjustment is to reduce the area. Please refer to FIG. FIG. 5 shows the adjustment direction of the member 271 having the shape 221. For example, the member 271 can be in only two directions such as + Y and -Y. Movement in the + Y direction increases the area of the shape 221 so that it is adjusted outward, and movement in the −Y direction decreases the area of the shape 221 so that it is adjusted inward.

  Once the adjustment direction of each member is determined, the method proceeds to the next stage to evaluate the “correction weight” of each member. The “correction weight” is the amount of movement in the determined adjustment direction. The amount of movement is related to the degree of change of each shape member and the degree of change of the outline (outline) of the corrected geometric shape. It should be noted that the sensitive group changes significantly with minor modifications, and the less sensitive group requires more significant correction to obtain similar geometric profile changes. Here, the “correction weight” after the simulation of the member 271 having the shape 221 shown in FIG.

  After determining the “adjustment direction” and “correction weight” for each member, the method proceeds to the next stage to perform a priority group correction process. In the priority group correction process, the actual “adjustment direction” and “correction weight” of each member are determined based on the “adjustment direction” and “correction weight” of each member proposed by the simulation.

  Although the “adjustment direction” and “correction weight” of each member are proposed by the simulation, the “correction weight” of each member does not necessarily match the “correction scale” as necessary. In the present invention, a value discounted from the “correction weight” can be used as the “correction scale”. This is called correction weight damping. For example, the damping value can be set to 0 to 1 of the correction weight, for example, 90%, 70%, 50%, and 30% of the correction weight depending on the pattern and the OPC model. For example, the “correction scale” of the member 271 having the shape 221 shown in FIG. 5 is “70% A”.

  After execution of the priority group modification process, a modified priority group is obtained and the original contour of the first layout pattern is changed accordingly. In the normal case, the priority group modification process only modifies the members of the priority group. To avoid interference between different groups, it is desirable that none of the subordinate group members modify.

  After the simulation process and the priority group modification process for all members of the priority group have been completed and the modified priority group has been obtained, the method proceeds to step 150 where the simulation process and the Perform a subordinated group correction process to obtain a corrected subordinated group. Since the subordinate group has also been selected by the priority selection process, all members of the subordinate group should have the same specific attributes. Therefore, if the simulation process is performed for all members of the subordinate group before the correction process, a more stable simulation result can be obtained under the same conditions.

  If step 120/130 is not performed, the method proceeds to step 150, where a simulation process and a second group modification process are performed for all members of the second group, respectively, to obtain a modified second group. To do. Since the second group is also selected by the priority selection process, all members of the second group should have the same specific attributes. Therefore, if the simulation process is performed on all members of the second group before the correction process, a more stable simulation result can be obtained under the same conditions.

  For example, as shown in FIG. 2, the edges 231 and 232 are perpendicular to each other. If the edge 231 is set to “sensitive” and the edge 232 is set to “insensitive” through the priority selection process, the sensitive (sensitive) edge 231 is a priority group according to the “classification principle” described above, and the sensitivity is low. (Insensitivity) Edge 232 is a subordinate group. The simulation process and the subordinate group correction process are performed on “all” members 272, 274 (edges 232) of the second group 262 (ie, the subordinate group). Generally, the simulation process is performed before the priority group modification process. Alternatively, in a simple shape as shown in FIG. 3, each side can be a member. The simulation process and the correction process are first executed for either the first group 261 or the second group 262. Preferably, the priority selection process is executed to obtain the priority group, the corresponding simulation process and the correction process are executed for the priority group, and then the corresponding simulation process and the correction process are executed for the subordinate group.

  In the simulation process, the “adjustment direction” and “correction weight” of each member (ie, members 272, 274) are simulated. Refer to FIG. 5 for the detailed operation of the direction adjustment.

  Once the adjustment direction of each member is determined, the method proceeds to the next stage to evaluate the “correction weight” of each member. Please refer to FIG. In the present invention, a value discounted from the “correction weight” can be used as the “correction scale”. This is called correction weight damping. For example, the damping value can be set to 0 to 1 of the correction weight, for example, 90%, 70%, 50%, and 30% of the correction weight depending on the pattern and the OPC model.

  After performing the subordinate group modification process, a modified subordinate group is obtained and the original contour of the first layout pattern is thereby changed again after the modified priority group. In the normal case, the subordinate group correction process only corrects the members of the subordinate group. To avoid interference between different groups, it is desirable that none of the priority group members be modified.

  When executing step 120/130, the simulation process and the correction process are performed on all members of the priority group and the subordinate group to obtain the corrected priority group and the corrected subordinate group, and then the first layout pattern is obtained. The revised preferred group and the modified subordinated group are re-examined to determine whether the goal has been achieved. That is, the present invention proceeds to step 160 and repeatedly verifies them until both the modified priority group and the modified subordinate group achieve the goal. If both the modified priority group of the first layout pattern and the modified subordinate group achieve the goal, all members of the modified first layout pattern's priority group and subordinate group meet the given requirements. Therefore, the corrected first layout pattern can be directly output.

  If step 120/130 is not executed, after performing the simulation process and the correction process for all members of the first group and the second group and obtaining the corrected first group and the corrected second group Then, the modified first group and the modified second group of the first layout pattern are verified again to determine whether the target has been achieved. That is, the present invention proceeds to step 160 and repeatedly verifies them until both the modified first group and the modified second group achieve the goal. If the modified first group of the first layout pattern and the modified second group both achieve the goal, all members of the first group and the second group of the modified first layout pattern will meet the requirements. Since they match, the corrected first layout pattern can be directly output.

However, in most cases, the first and second groups of the first layout pattern are subjected to one simulation process and one correction process, and the corrected first group and the corrected second group. Despite having acquired, both the first group and the second group of the modified first layout pattern cannot achieve the target. Thus, as shown in FIG. 6A, step 160 may include the following steps.
Step 161: The simulation process is again executed for all the members of the modified first group to obtain the first group data.
Step 162: Based on the first group data described above, the first group correction process is executed again on the corrected first group to obtain the corrected first group.
Step 163: Verify whether the modified first group member has achieved the goal.
Step 164: The simulation process is again executed for all members of the modified second group to obtain the second group data.
Step 165: Based on the second group data described above, the second group correction process is performed again on the corrected second group to obtain the corrected second group.
Step 166: Verify whether the modified second group member has achieved the goal.

  In other words, in step 161/164, the simulation process is performed again for all members of the modified first group and the modified second group. Preferably, the independent simulation process is again performed on all the members of the modified first group and the modified second group to obtain corresponding data. Thereafter, in step 162/165, a first group correction process and a second group correction process are sequentially executed based on the data. Preferably, the first group modification process and the second group modification process are independent until all members of the modified first group and modified second group achieve the goal as shown in steps 163/166. And execute.

If the processes are performed separately for the first group and the second group, step 160 can be further considered to include the following steps shown in FIG. 6B.
Step 161 ′: The simulation process and the first group correction process are executed for all the members of the corrected first group, respectively, and the corrected first group is obtained again.
Step 162 ′: The simulation process and the second group correction process are respectively performed on all the members of the modified second group, and the modified second group is obtained again.

That is, step 160 is typically repeated one or more times until both the modified first group and the modified second group achieve the goal. For example, taking the shape exposure intensity of the layout pattern shown in FIG. 4 as an example, the first contrast intensity between the shapes 312 and 313 is 0.42917 after being corrected a plurality of times in step 160, and the shapes 313 and 314 are obtained. The second contrast intensity between 0.3122 and 0.3122 becomes 0.42911. Since the difference between the first contrast intensity and the second contrast intensity is 6 × 10 ⁻⁵ , for example, within a specific range 1 × 10 ⁻⁴ , both the first contrast intensity and the second contrast intensity are considered to achieve the target. It is.

  Steps 161 'and 162' are repeated as necessary as necessary until both the first group and the second group achieve the goal. Therefore, the simulation process and the correction process relating to the first group and the second group are executed a plurality of times until both the corrected first group and the corrected second group finally achieve the target. The result is called the available second correction pattern. Therefore, the original first layout pattern becomes a usable second layout pattern. Thereafter, the method proceeds to step 170 and outputs a second layout pattern including the first and second correction groups that have achieved the goal. As a result, the required usable layout pattern is obtained.

  Those skilled in the art will readily appreciate that numerous modifications and variations of the apparatus and method may be made while maintaining the teachings of the invention.

3 is a flowchart showing a main flow of a method for selectively correcting a layout pattern according to the present invention. It is explanatory drawing which shows the given layout pattern containing a some group and the member which belongs to it. It is explanatory drawing which illustrates the circular pattern used as a contact hole pattern. It is explanatory drawing which shows the exposure intensity | strength of the several shape in a layout pattern. It is explanatory drawing which shows the adjustment direction of the given member of one shape. It is explanatory drawing for demonstrating step 160 further. It is explanatory drawing for demonstrating step 160 further.

Claims (19)

A method for selectively modifying a layout pattern,
Setting a first layout pattern including a plurality of shapes formed from a plurality of edges;
Dividing the plurality of edges included in the plurality of shapes into at least a first group and a second group according to their orientations;
Performing a simulation process and a first group modification process on the edges of the first group, respectively, to obtain a modified first group;
Performing a simulation process and a second group modification process on the edges of the second group, respectively, to obtain a modified second group;
Verifying the modified first group and the modified second group until both the modified first group and the modified second group achieve a goal;
Outputting a second layout pattern that includes the modified first group and the modified second group that have both achieved a goal. The method of claim 1, wherein the second group is less sensitive to asymmetric light source illumination than the first group. The method according to claim 1 or 2 , wherein the edges included in the first group are perpendicular to the edges included in the second group . The method according to any one of claims 1 to 3, wherein the first group and the second group each include a first contrast intensity and a second contrast intensity. The method according to any one of claims 1 to 4, wherein the first layout pattern is selected from the group consisting of a contact hole pattern and a via hole pattern. The method according to any one of claims 1 to 5 , further comprising verifying that both the first group and the second group have achieved a goal. 7. The method of claim 6 , wherein parameters are used to verify that both the first group and the second group have achieved their goals. The method of claim 7 , wherein the parameter includes exposure intensity. The method according to claim 1, wherein the simulation process evaluates a correction direction and a correction weight. The method of claim 9 , wherein the correction of the correction weight is defined in the first group correction process. The method of claim 9 , wherein the second group correction process defines damping of the correction weight. 12. A method according to any one of the preceding claims, wherein the first group modification process is performed for all edges of the first group. 13. A method according to any one of the preceding claims, wherein the first group modification process is not performed for all edges of the second group. 14. A method according to any one of the preceding claims, wherein the second group modification process is performed for all edges of the second group. The step of verifying the modified first group and the modified second group until both the modified first group and the modified second group achieve a goal further comprises:
Performing the simulation process on all edges of the first group respectively to obtain first group data;
Performing the first group modification process on the modified first group based on the first group data until all edges of the first group achieve a target, the modified first group; The stage of obtaining
Performing the simulation process on each of all edges of the second group to obtain second group data;
Performing the second group modification process on the modified second group based on the second group data until all edges of the second group have achieved a target; The method of any one of Claims 1-14 including the step which acquires these. The method according to claim 1, comprising performing optical proximity correction on a preliminary layout pattern to obtain the first layout pattern. Wherein the first group performs the priority selection process for the second group, comprising the step of obtaining the first group and the second group, the method according to any one of claims 1 to 16. The method of claim 17 , wherein the priority selection process is based on motion modification sensitivity. The method according to any one of claims 1 to 18, wherein the first group is a group sensitive to asymmetric light source irradiation and the second group is a group sensitive to asymmetric light source irradiation. .

Images