JP2008216504A - Method for manufacturing mask - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for manufacturing a mask by which the time required for correcting the dimension of a mask pattern can be reduced. <P>SOLUTION: The method for manufacturing a mask is provided, the mask having a mask pattern formed therein to be used for transferring a pattern image of alternately arranged lines and spaces onto a photoreceptor by exposure, and the method includes steps of defining a correction region where the width dimension of the pattern image is shifted with respect to a desired value, and simultaneously correcting the dimension of all patterns in the mask pattern corresponding to the pattern image in the correction region. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a mask manufacturing method.

  For example, in a semiconductor integrated circuit such as a semiconductor memory, a line-and-space repetitive pattern is often used (see, for example, Patent Document 1).

  Such a line and space pattern is exposed and transferred onto a photoconductor such as a resist using a photomask. In recent years, pattern miniaturization has progressed, and the size of the pattern image on the photoconductor approaches the wavelength of exposure light or becomes smaller than the wavelength of light, and the mask pattern can be faithfully exposed and transferred by the optical proximity effect. It has become difficult.

  Therefore, a technique called OPC (Optical Proximity Correction) has been proposed. This is a technique for correcting a pattern on the mask side so that a transferred pattern after exposure becomes as intended.

However, for fine line and space patterns, correction processing that also takes into account light diffraction effects, etc. must be performed, and it takes an enormous amount of time for processing, and it is difficult to find the optimal solution, that is, processing does not converge. ing.
JP 2006-293081 A

  The present invention provides a method for manufacturing a mask that can reduce the time required to correct the dimension of a mask pattern.

  According to one aspect of the present invention, there is provided a method of manufacturing a mask on which a mask pattern for exposing and transferring a pattern image in which lines and spaces are alternately arranged to a photoreceptor is formed, the width dimension of the pattern image Defines a correction area that deviates from a desired value, and performs a dimensional correction of all the patterns in the mask pattern corresponding to the pattern image of the correction area at the same time. .

  According to the present invention, there is provided a mask manufacturing method capable of shortening the time required for mask pattern dimension correction.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 2B is a schematic diagram showing a line-and-space pattern image in which a plurality of lines and spaces are alternately arranged. This pattern image has a periodic structure pattern image 50 composed of a plurality of lines L and spaces S arranged alternately at equal pitches in one direction (X direction), and a space provided at the end of the pattern image 50. S1 and a line L1 provided next to the space S1. The space S1 is wider than the space S, and the line L1 is wider than the line L.

  FIG. 2A shows a mask pattern for exposing and transferring the pattern image of FIG. 2B to a photoreceptor (for example, a photoresist or simply a resist). This mask pattern is formed by, for example, forming a plurality of line-shaped light shielding films (or halftone films) 100 on a quartz substrate.

  After exposing the resist to light that has passed through the mask, the resist is developed, and the resist that has been exposed (or not) is selectively etched away to obtain a resist mask that is patterned as desired. A film to be processed formed on a semiconductor wafer or a liquid crystal substrate is processed using this resist mask.

  In FIG. 2A, a pattern 30 is a pattern for exposing and transferring the pattern image 50 of the periodic structure in FIG. The pattern 20 formed at the end of the pattern 30 has a light shielding film (or halftone film) that is thin enough not to be resolved, and the space 20 is obtained by the pattern 20. Even if the pattern image of the space is relatively wide, by using the pattern 20 provided with a thin light-shielding film (or halftone film) that does not resolve, distortion of the diffracted light can be suppressed and deformation can be prevented. A small pattern image can be obtained. A pattern 10 having, for example, three light shielding films (or halftone films) having a relatively large line width adjacent to each other with a small space is a pattern for exposing and transferring one wide line L1.

  Next, a mask pattern correction method according to this embodiment will be described with reference to the flowchart of FIG.

  First, in step S1, an optical simulation is performed using an originally designed mask pattern. In this process, the exposure conditions are reproduced on a computer, and the intensity distribution of the light transmitted through the mask and the optical system as shown in FIG. 3 is obtained. In FIG. 3, the horizontal axis corresponds to the position in the X direction of the periodic structure pattern 50 in FIG. The vertical axis represents the intensity distribution (or energy distribution) of light transmitted through the mask and the optical system.

  When the light intensity required for the resist to dissolve or not dissolve in the developing solution is Ith, the pattern 50 depends on the distance between the intersections of the straight line representing the level of Ith and the light intensity distribution curve. The width of each line L and each space S is determined. That is, the width dimension in the pattern image 50 is virtually obtained by performing the optical simulation.

  FIG. 4 is a graph showing the width dimension distribution of each line L and each space S in the pattern image 50 obtained by the optical simulation. The horizontal axis corresponds to the position in the X direction in FIG. 2, and the vertical axis represents the dimension (nm) in the width direction (X direction).

  In FIG. 4, the black rhombus represents the width dimension distribution of the pattern image 50 obtained by performing the optical simulation when the mask pattern as originally designed without correction is used. The mask pattern 30 (FIG. 2A) for exposing and transferring the pattern image 50 is designed so that each line-and-space width in the pattern image 50 is, for example, 56 (nm). Due to the proximity effect or the like, a pattern whose width is deviated from a desired value (56 (nm)) is generated. In particular, the pattern width tends to greatly deviate from the desired value at the end of the periodic structure pattern 50 near the irregular patterns L1 and S1 (the left portion in the graph of FIG. 4).

  Therefore, in this embodiment, the dimension correction of the mask pattern 30 is performed as follows.

  As step S2 in FIG. 1, a region where the width dimension in the pattern image 50 obtained by the optical simulation is shifted from a desired value (for example, 56 (nm) in the above example) is defined (recognized) as a correction region. . In the following steps, the pattern width of the mask pattern 30 corresponding to the line and space in the correction area is corrected.

  Specifically, first, the width dimension correction amount for each pattern in the mask pattern 30 corresponding to each line and space in the correction area is calculated (step S3). The width dimension correction amount is determined by the pattern width of the region in which the pattern width in the pattern image 50 is not shifted from the desired value (line and space distributed on the right side in FIG. 4) and the mask pattern 30 corresponding to the line and space. Based on the correlation of the width dimension variation amount with the pattern (correlation such that when the pattern width in the mask pattern 30 is varied by 1 (nm), for example, the pattern width in the pattern image 50 is varied by 4 (nm)). Is calculated. A highly reliable dimensional correction amount can be obtained by calculating the dimensional correction amount based on the correlation between the pattern image in which a desired width dimension is obtained and the corresponding mask pattern.

  As described above, when the correction amount of each pattern whose width dimension is to be corrected in the mask pattern 30 is obtained, the width dimension correction is simultaneously performed for all the mask patterns to be corrected in step S4. That is, width dimension correction is not performed for each line or space, but a plurality of line-and-spaces to be corrected are collectively corrected at once.

  Each mask pattern to be corrected has its width dimension corrected by changing both edges in the width direction in the width direction by the same amount with respect to the center in the width direction. FIG. 5 shows one of, for example, a line-shaped light shielding film (or halftone film) pattern 100 in the mask pattern 30. Both edges 100a in the width direction of the pattern 100 are changed in the width direction by the same amount ΔW with respect to the center C in the width direction, and are corrected to a desired width dimension.

  Next, as step S5, an optical simulation similar to that of step S1 is performed using a mask reflecting the width dimension correction for all patterns to be corrected in the mask pattern 30. As a result of this optical simulation, a width dimension distribution represented by a black square in FIG. 4 is obtained, and a deviation with respect to a desired width (56 (nm)) is smaller than that before correction represented by a black rhombus. If this degree of variation is within the allowable range, the determination in step S6 is “Yes”, and the pattern width correction is completed in step S7.

  If the satisfactory width of the pattern image is not yet obtained in one (previous) correction, “No” is determined in step S6, and the process proceeds to step S8, where each mask pattern 30 to be corrected is to be corrected. The amount of correction of the width dimension of the pattern is appropriately adjusted, and an optical simulation is performed again to converge to a pattern image having a satisfactory width dimension.

  Conventionally, pattern correction is performed one by one, and a pattern image is confirmed by optical simulation each time. That is, first, the width dimension of a certain pattern is corrected, the pattern image is confirmed by optical simulation, and if there is a pattern whose dimension is to be corrected in the pattern image, the second correction is also performed for one pattern. After pattern correction, the pattern image is confirmed by optical simulation, and if there is a pattern to be dimension corrected in this pattern image, the third correction is performed for only one pattern, and the pattern image is verified by optical simulation. Was confirmed, and so on. In this method, since the next correction is performed based on the result of the optical simulation performed after the previous correction, the correction standards are changed one after another, resulting in inefficient correction processing, which is very time consuming. In some cases, the correction process did not converge.

  In contrast, in the present embodiment, first, an optical simulation is performed with a mask pattern as originally designed, and dimensional correction for all patterns to be corrected is performed simultaneously (at a time) using the pattern image obtained as a reference. (Summary) After performing the optical simulation, the pattern image is confirmed. That is, instead of capturing and handling individual patterns to be corrected, correction is performed in consideration of diffraction and interference of the entire light passing through the pattern by capturing and correcting the area. As a result, even a fine-pitch line and space pattern with a prominent proximity effect can be converged to an optimum pattern dimension without taking a long time.

  In addition, when the width of the mask pattern is changed, if the correction is performed so that only one edge is individually changed to obtain the desired width, the center of gravity position of each line-and-space pattern is shifted. May lead to diffraction problems, and the optical image shown in FIG. 3 may be disturbed. When this optical image is seen, the setting margin of the light intensity threshold value Ith that determines the width dimension of the line and space becomes small, and the process conditions become severe.

  On the other hand, in this embodiment, each mask pattern to be corrected has its width dimension corrected by changing both edges in the width direction in the width direction by the same amount with respect to the center in the width direction. The position of the center of gravity of the pattern does not shift, that is, the position of the center of gravity of each pattern is regular at an equal pitch. As a result, it is possible to avoid severe process conditions due to the disturbance of the optical image without imposing a diffraction defect.

The flowchart of the manufacturing method of the mask which concerns on embodiment of this invention. (A) represents an example of a line and space mask pattern, and (b) is a schematic diagram representing a pattern image exposed and transferred by the mask of (a). The schematic diagram which illustrates the simulation result of the light intensity distribution after the optical system transmission in the case of using a line and space mask pattern. The graph which illustrates the width dimension distribution of the pattern image on the resist obtained by optical simulation. The schematic diagram for demonstrating the dimension correction | amendment in embodiment of this invention.

Explanation of symbols

  10, 20, 30 ... Mask pattern, 50 ... Line and space pattern image, 100 ... Line pattern, 100a ... Edge in width direction of line pattern

Claims (4)

A mask manufacturing method in which a mask pattern for exposing and transferring a pattern image in which lines and spaces are alternately arranged to a photoreceptor is formed,
Define a correction region in which the width dimension of the pattern image is shifted from a desired value,
A method for manufacturing a mask, comprising: simultaneously performing dimension correction of all patterns in the mask pattern corresponding to a pattern image in the correction area.   Dimension correction in the mask pattern should be corrected based on the correlation of the dimensional variation between the pattern image in the region where the width dimension in the pattern image is not shifted from the desired value and the mask pattern corresponding to the pattern image. 2. A mask manufacturing method according to claim 1, wherein a correction amount of each pattern is obtained.   3. The size correction is performed by varying both edges in the width direction of each pattern to be dimension-corrected in the mask pattern in the width direction by the same amount with respect to the center in the width direction. The manufacturing method of the mask as described.   In the mask pattern, the dimensional correction amount of each pattern corresponding to the pattern image of the correction area is calculated, and the dimensional correction of all patterns to be corrected is performed simultaneously based on the dimensional correction amount, and this correction is reflected. The method for manufacturing a mask according to claim 1, wherein a width dimension of a pattern image obtained by optical simulation using the mask pattern is confirmed.

Images