JP3508306B2 - Mask pattern correction method, mask using the same, exposure method and semiconductor device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a simulation method of a transfer light intensity distribution used for obtaining a transfer image of a photomask in manufacturing a semiconductor device, a simulation device, and a transfer image close to a desired design pattern. Method for correcting a mask pattern, a correction apparatus for performing the correction method, a photomask obtained by the correction method, and a photomask having the corrected mask pattern The present invention relates to an exposure method for performing exposure by using a semiconductor device and a semiconductor device manufactured by photolithography using a photomask having the corrected mask pattern.

More specifically, according to the present invention, in the light intensity simulation, by using a photomask pattern in which a shape such as a triangle with a convex corner of an input pattern is deleted and a shape such as a triangle with a concave corner is added, Intensity distribution simulation method in consideration of corner rounding occurring in the mask process, simulation apparatus, mask pattern correction method using the present method, mask created using the same, and exposure method using the mask And semiconductor devices.

[0003]

2. Description of the Related Art In recent years, with the miniaturization of semiconductor devices to be manufactured, it has become essential to improve the resolution even in the photolithography process. Further, in order to realize a high yield, it is necessary to expand the process margin. In order to realize these, it is necessary to find optimum exposure conditions that can obtain high resolution and a wide process margin. In optimizing the exposure method, a means for determining the transferred pattern is required.

The most accurate method for obtaining a transfer pattern is a transfer experiment. However, in the above-mentioned optimization, it is necessary to obtain the transfer pattern under a huge number of conditions, and it is impossible to carry out all of them only by the experiment because the time and cost become huge. In addition, it is impossible to evaluate a process that is expected to be realized in the future, which cannot be tested at present.

Therefore, the light intensity simulator has been used together with experiments in obtaining the transfer pattern. Here, the light intensity simulator is used to input a mask pattern and transfer conditions and calculate the light intensity distribution on or within the resist under the conditions. In the past, in this simulation, the transmittance distribution of the mask was calculated assuming that the mask was created according to the design pattern.

On the other hand, the difference between the mask pattern and the transferred resist pattern has become a problem due to the lack of resolution associated with the miniaturization of semiconductor devices to be produced. This current situation causes problems such as deterioration of device performance due to deformation of the transfer pattern and reduction of yield due to pattern bridging and disconnection. Therefore, in order to obtain a desired resist pattern, a mask pattern in which a plurality of modification patterns are added to the design pattern is formed by cut-and-try, and a transfer experiment obtains the transfer pattern by simulation, and the transfer pattern is closest to the design pattern. It has been practiced to add a modifying pattern to a mask pattern to obtain a transfer pattern. In recent years, an optical proximity effect correction technique has been developed in which a mask pattern is automatically optimized on a computer. In the optical proximity correction, the light intensity distribution is calculated for the input design pattern using light intensity simulation, and the mask pattern deformed so that the transfer image obtained is improved is calculated. It was

[0007]

However, the conventional light intensity simulator has the following problems. In the conventional technique, it is assumed that a mask pattern exactly the same as the design pattern is created, and the light intensity simulation is performed using this pattern. However, in the actual mask formation, due to the scattering of electron beams used during mask formation in the resist, the loading effect in the resist process and etching process, and the side wall of the resist, etc. Will occur. Therefore, the mask pattern used in the simulation is different from the mask pattern in the actual process, and as a result, there is a remarkable drawback that the result of the simulation cannot accurately calculate the result of the actual process.

Further, in the optical proximity correction, the accuracy of the transferred image calculation is insufficient because the actual rounding of the mask pattern is not taken into consideration in the light intensity simulation to be used, and therefore the optical proximity correction is performed. There was a notable flaw that was not done correctly.

The present invention solves the above-mentioned problems of the prior art, and enables the transfer pattern to be simulated at high speed and accurately in consideration of the cornering of the mask, whereby the speed and cost can be reduced accurately. It enables a process evaluation, and thus provides a means for producing a high-performance device with a high yield.

Further, the present invention solves the above-mentioned problems of the prior art and makes it possible to simulate a mask pattern with high accuracy so that a resist pattern close to a design pattern can be obtained, thereby achieving a high yield. It is intended to provide a means for producing high performance devices.

[0011]

[0012]

[0013]

[0014]

[Means for Solving the Problem] How to correct a mask pattern
Method and Correction Device A mask pattern correction method according to the present invention is a mask pattern correction method for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. , along the pattern periphery of the desired design pattern, the evaluation point arranging step of arranging a plurality of evaluation points by using the photomask design pattern wherein the evaluation point is attached, a plurality of preset exposure latitude The exposure amount of
Combination with multiple focal points within the set depth of focus
In simulating the printed image obtained when exposure was performed at transfer conditions of plural kinds, based on the previous
By adding or deleting a figure against the corner of the serial design pattern, to create a real simulation photomask image in consideration of corner rounding on the photomask occurs in the mask making process, the photomask image Using the plurality
By performing a light intensity simulation under the same transfer conditions, a simulation process of calculating a plurality of images after transfer, and a plurality of passes of each of the plurality of simulated transfer images and the design pattern are performed.
The Rino difference, the comparing step of comparing the respective evaluation points, the the average value is the minimum of the difference between the plural kinds of comparison for each evaluation point
And a transformation step of transforming the design pattern so that the difference becomes smaller. According to the present invention
The mask pattern correction method is a photolithography
The mask pattern of the photomask used in
Change so that a transfer image close to the design pattern can be obtained.
In the correction method of the mask pattern to be formed,
Multiple evaluation points along the outer periphery of the design pattern of
And an evaluation point placement step for placing
Pre-set exposure using a photomask of total pattern
Within the range of multiple exposure doses and preset depth of focus
Multiple transfer conditions based on combinations with multiple focal positions
Simulate the transfer image obtained when exposure is performed with
To the corner of the design pattern
By adding or deleting graphics to the actual
The photomask occurring in the mask making process
Simulation considering upper corner rounding
Create a photo mask image for
Light intensity under the above multiple transfer conditions using an image
By performing a degree simulation,
A simulation process for calculating an image, and
Before and after each of the simulated transfer images
For each of the above evaluation points, there are multiple differences from the design pattern.
Comparison process to compare and multiple compared for each evaluation point
Of the street differences, the standard with the largest absolute value and the smallest difference
Use the design pattern to reduce the difference.
And a deforming step of deforming. Mask pattern according to the present invention
The correction method of the screen is used in the photolithography process.
Set the mask pattern of the photomask to the desired design pattern.
Image to obtain a transfer image close to
In the correction method of the mask pattern, the desired design pattern is
Place multiple evaluation points along the perimeter of the pattern
Evaluation point placement step and design pattern with the evaluation points
With multiple preset exposure margins using
Multiple exposures within a range of exposure and preset depth of focus
Exposure is performed under multiple transfer conditions based on the combination with the position.
Simulation of the transfer image obtained when
When designing, draw a figure against the corner of the design pattern.
Actual mask creation by adding or deleting
Corners on the photomask generated in the process
-Simulation photo considering rounding
Create a mask image and add the photomask image
Using the above multiple transfer conditions, the light intensity
Image after transfer
The simulation process to calculate and the multiple simulations
Each of the transferred images and the design pattern
The ratio that compares multiple differences with each evaluation point.
The comparison process and the difference in multiple ways compared for each evaluation point
The difference is reduced by using the criterion that minimizes the root mean square.
So as to deform the design pattern,
It

[0015] In the correction method of a mask pattern according to the present invention, since utilizing simulation method of a light intensity distribution according to the present invention, considering the corner rounding on the photomask which occurs in the actual mask making process Simulation becomes possible. as a result,
Transfer light intensity close to actual exposure can be calculated at high speed and accurately. Therefore, if the photolithography process is performed using the photomask having the corrected mask pattern, a pattern close to the design pattern can be obtained.

Further, according to the mask pattern correcting method of the present invention, the mask pattern is automatically deformed so that a transfer image close to a desired design pattern is automatically obtained regardless of the shape of the mask pattern. Will be possible.
Therefore, it is possible to eliminate the disadvantage of the method of correcting the mask pattern by the trial and error method for each individual mask pattern.

[0017]

[0018]

According to such a mask pattern correction method, since the transfer image when the transfer condition changes within the range of the process margin is taken into consideration (the process margin is taken into consideration), the corrected mask is corrected. Based on the pattern, the process latitude such as exposure latitude and depth of focus will not deteriorate. As a result, if the photolithography is performed using the photomask having this mask pattern, the manufacturing yield is improved.

In the present invention, in the simulation step, the two-dimensional light intensity on the substrate is calculated on the basis of the simulation photomask image and the exposure condition, and a desired focus position on the two-dimensional plane of the substrate is calculated. Calculating and accumulating the influence of the light intensity at the plurality of peripheral positions on the exposure energy at the focused arbitrary position based on the light intensity at the peripheral position and the distance between the focused position and the peripheral position. Thus, the latent image forming intensity at the noted arbitrary position is calculated on the two-dimensional plane of the substrate, and the distribution of the latent image forming intensity on the two-dimensional plane of the substrate is obtained, which corresponds to the exposure amount and the developing condition. The threshold value of the latent image forming intensity is determined, the contour line at the threshold value is obtained for the distribution of the latent image forming intensity, and the contour line is defined by the contour line. A turn may be calculated as a transfer image.

A method of calculating and accumulating the influence of the peripheral position on the exposure energy at the noted arbitrary position is as follows.
The product of a light intensity at the peripheral position and a function that takes the distance between the arbitrary position of interest and the peripheral position as an argument and becomes maximum when the distance is 0 and becomes 0 when the distance is infinite. It is preferable to calculate and accumulate the influence of the light intensity at the plurality of peripheral positions by.

A method of calculating and accumulating the influence of the peripheral position on the exposure energy at the noted arbitrary position is as follows.
With the power of the light intensity at the peripheral position and the distance between the focused arbitrary position and the peripheral position as an argument, the distance is 0.
It is also possible to calculate and accumulate the influence of the light intensity at a plurality of the peripheral positions by the product with a function that becomes maximum when, and becomes 0 when the distance is infinite.

As the function, for example, a Gaussian function can be exemplified. In such a mask pattern correcting method, the threshold value of the simple two-dimensional light intensity distribution is not used as the transferred image. When the threshold of a simple two-dimensional light intensity distribution is used as a transfer image, FIG.
As shown in FIG. 7, when the peak of the light intensity distribution is high, the line width L of the actually formed resist pattern becomes thicker than the line width 1 by the simulation defined by the threshold Eth, and conversely, FIG. As shown in B), when the peak of the light intensity distribution is low, the line width L of the resist pattern actually formed tends to be thin. In view of this point, the present inventor considers that the element contributing to the formation of the resist pattern at the point of interest is not only the light intensity of the arbitrary point of interest but also the light intensity of points around the arbitrary point of interest. We also obtained the knowledge that

Therefore, the present inventor newly creates a concept of latent image forming intensity based on the above-mentioned knowledge, obtains the distribution of the latent image forming intensity, sets a threshold value, and calculates a resist pattern. As a result, they found that this result is very consistent with the actually obtained resist pattern.

Here, the latent image forming intensity is a concept determined in consideration of not only the light intensity of the focused arbitrary position but also the influence of the light intensity of the peripheral position on the exposure energy of the focused arbitrary point. is there. According to the mask pattern correction method of the present invention as described above, it is possible to bring the transfer image transferred by the design pattern (including the deformed design pattern) closer to the transfer image obtained in the actual transfer process. Therefore, the mask pattern can be automatically corrected with high accuracy.

In the present invention, in the deforming step, a mask pattern in the vicinity of the evaluation points is formed in a direction opposite to the difference compared for each evaluation point by a size obtained by multiplying the magnitude of the difference by a constant coefficient. It is preferable to move the boundary line of. It is preferable that the coefficient is greater than 0 and less than 1.

According to such a mask pattern correction method of the present invention, when the mask pattern is deformed, a constant coefficient is applied to the magnitude of the difference in the opposite direction of the difference compared for each evaluation point. The boundary line of the mask pattern in the vicinity of the evaluation point is moved by a size multiplied by. This coefficient is greater than 0 and less than 1.

For example, in photolithography near the resolution limit, a small change in the mask pattern has a great influence on the actual transferred image. In the present invention, the mask pattern is corrected so that the boundary line of the mask pattern near the evaluation point is moved in the direction opposite to the difference from the transferred image.
Therefore, if the mask pattern is corrected with a size equal to or larger than the difference at the evaluation point, the difference between the obtained transfer image and the corrected mask pattern may not be reduced, and an appropriate correction calculation may not be performed. There is.

In such a mask pattern correction method according to the present invention, when the mask pattern is deformed, a constant coefficient (0) is applied to the magnitude of the difference in the opposite direction of the difference compared for each evaluation point. Since the boundary line of the mask pattern in the vicinity of the evaluation point is moved by a size obtained by multiplying by more than 1), the transfer image obtained by the corrected mask pattern gradually approaches the design pattern.

In the present invention, in the evaluation point arranging step, along the pattern outer periphery of the desired design pattern,
A plurality of evaluation points are arranged, and at a predetermined evaluation point, a target point is set separately from the evaluation point, and in the comparison step, at a position where only the evaluation point is set, a simulated transfer image is obtained. , Comparing the difference with the design pattern for each evaluation point, at the position where the target point is set, compare the difference between the target point and the transfer image,
In the deforming step, the design pattern may be deformed so that the difference becomes smaller depending on the difference compared for each evaluation point or each target point.

The target point is set corresponding to, for example, an evaluation point located at the convex corner portion or the concave corner portion of the design pattern, and at the convex corner portion, the target point is determined inside the corner portion. At the concave corner, the target point is determined outside the corner. In the simulation step, it is preferable to use a photomask image for simulation in which a triangle is deleted from a convex corner and a triangle is added to a concave corner at a corner of the design pattern. One side of the triangle is
Greater than 0 to the minimum line width of the design pattern 100%
The following is preferable.

For example, in the case of convex corners or concave corners of the design pattern, when the evaluation points are located on these corners,
If the mask pattern is corrected with the aim of bringing the transferred image closer to the evaluation point itself, the transferred image at positions other than the corners may be separated from the design pattern.

Therefore, in the present invention, for example, a convex corner determines a target point inside the corner, and a concave corner determines a target point outside the corner. By adding a correction to the design pattern so as to approach, the transfer image can be brought close to the design pattern as a whole. As a result, it is possible to favorably prevent bridging or disconnection between patterns.

A mask pattern correction device according to the present invention is a mask pattern correction device for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. along the pattern periphery of the desired design pattern, the evaluation point arranging means for arranging the plurality of evaluation points by using the photomask design pattern wherein the evaluation point is attached, a plurality of the exposure latitude the preset Exposure and in advance
Combination with multiple focal points within the set depth of focus
In simulating the printed image obtained when exposure was performed at transfer conditions of plural kinds, based on the previous
By adding or deleting a figure against the corner of the serial design pattern, to create a real simulation photomask image in consideration of corner rounding on the photomask occurs in the mask making process, the photomask image Using the plurality
By performing a light intensity simulation under different transfer conditions, a simulation means for calculating a plurality of images after transfer, and a plurality of communication between each of the plurality of simulated transfer images and the design pattern are performed.
The Rino difference, the comparing means for comparing the respective evaluation points, the the average value is the minimum of the difference between the plural kinds of comparison for each evaluation point
And a deforming unit that deforms the design pattern so that the difference becomes smaller. According to the present invention
The mask pattern correction device is a photolithography
The mask pattern of the photomask used in
Change so that a transfer image close to the design pattern can be obtained.
In the mask pattern correction device for shaping,
Multiple evaluation points along the outer periphery of the design pattern of
And an evaluation point arrangement means for arranging
Pre-set exposure using a photomask of total pattern
Within the range of multiple exposure doses and preset depth of focus
Multiple transfer conditions based on combinations with multiple focal positions
Simulate the transfer image obtained when exposure is performed with
To the corner of the design pattern
By adding or deleting graphics to the actual
The photomask generated in the mask making process
Simulation considering upper corner rounding
Create a photo mask image for
Light intensity under the above multiple transfer conditions using an image
By performing a degree simulation,
A simulation means for calculating an image;
Before and after each of the simulated transfer images
For each of the above evaluation points, there are multiple differences from the design pattern.
Comparison means for comparison and a plurality of comparisons made for each evaluation point
Of the street differences, the standard with the largest absolute value and the smallest difference
Use the design pattern to reduce the difference.
And a deforming means for deforming. Mask pad according to the present invention
Turn correction device used in photolithography process
Set the mask pattern of the photomask to the desired design pattern.
Transform it to obtain a transfer image close to
In the mask pattern correction device, the desired design pattern is
Place multiple evaluation points along the circumference of the turn pattern
Evaluation point arranging means and design pattern to which the evaluation points are attached
Using a photomask of
Number of exposures and multiple focal points within a preset depth of focus.
Exposure under multiple transfer conditions based on combinations with point positions
Simulate the transfer image obtained when performing
When designing, please refer to the corner of the design pattern
By adding or deleting shapes, the actual mask
Coating on the photomask generated in the formation process
A simulation fog that considers the rounding
Create a photomask image, and use the photomask image
Light intensity simulation under the above multiple transfer conditions.
Image after transfer
And a simulation means for calculating
Each of the transferred images and the design pattern
Compare multiple differences from the turn for each evaluation point
Comparison means and a plurality of differences compared for each evaluation point
Use a criterion that minimizes the root mean square of
So as to deform the design pattern.
To do.

[0035]

[0036]

In the simulation means, the two-dimensional light intensity on the substrate is calculated based on the simulation photomask image and the exposure condition, and the light at the peripheral position of any focused position on the two-dimensional plane of the substrate is calculated. Based on the intensity and the distance between the focused position and the peripheral position, the effects of the light intensities at the peripheral positions on the exposure energy of the focused arbitrary position are calculated and accumulated, thereby The latent image forming intensity at any given position is calculated on the two-dimensional plane of the substrate, the distribution of the latent image forming intensity on the two-dimensional plane of the substrate is calculated, and the latent image forming intensity corresponding to the exposure amount and the developing condition is calculated. Of the latent image forming intensity, the contour line at the threshold value is obtained, and the pattern defined by the contour line is transferred. It is preferable to calculate the over-di.

In the deforming means, the boundary line of the mask pattern in the vicinity of the evaluation point is formed in the opposite direction of the difference compared for each evaluation point by a size obtained by multiplying the difference size by a constant coefficient. It is preferable to move. It is preferable that the coefficient is greater than 0 and less than 1.

The evaluation point arrangement means arranges a plurality of evaluation points along the outer periphery of the pattern of the desired design pattern, and at a predetermined evaluation point, separately from the evaluation points.
A target point is set, and at the position where only the evaluation point is set, the comparison means compares the difference between the simulated transfer image and the design pattern for each of the evaluation points to set the target point. At the determined position, the difference between the target point and the transferred image is compared, and in the deforming means, depending on the difference compared for each evaluation point or for each target point, the difference becomes small, The design pattern may be modified.

The target point is set corresponding to, for example, an evaluation point located at a convex corner portion or a concave corner portion of the design pattern, and at the convex corner portion, the target point is determined inside the corner portion. At the concave corner, the target point is determined outside the corner. In the simulation means, it is preferable to use a photomask image for simulation in which a triangle is deleted from a convex corner and a triangle is added to a concave corner at a corner of the design pattern. One side of the triangle is
Greater than 0 to the minimum line width of the design pattern 100%
The following is preferable.

Each of the above means is an arithmetic circuit or RA.
M, ROM, stored in a storage means such as an optical storage medium,
It is composed of program information processed by the CPU of the computer. Photomask Manufacturing Apparatus The photomask manufacturing apparatus according to the present invention includes any one of the mask pattern correction apparatuses and a drawing unit that draws a photomask of the mask pattern corrected by the mask pattern correction apparatus. . The drawing means includes a printer such as a laser printer, an XY plotter, a fax device, a copying device, and the like. It is preferable that the mask pattern correcting apparatus or the photomask manufacturing apparatus according to the present invention also includes a display unit such as a CRT or a liquid crystal display device for displaying the corrected mask pattern on the screen.

Semiconductor Device Manufacturing Apparatus A semiconductor device manufacturing apparatus according to the present invention exposes using any one of the mask pattern correcting devices and a mask pattern photomask corrected by the mask pattern correcting device. And an exposure unit for performing the exposure. The exposure means is not particularly limited, and may be one that uses modified illumination, one that uses pupil filtering, or one that uses a phase shift mask such as a halftone method or a levelson method.

Exposure Method The exposure method according to the present invention is characterized by performing exposure using a photomask having a mask pattern corrected by the above-described mask pattern correction method.

[0044]

[0045]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below. Of course, the present invention is not limited to the examples described below. Example 1 In this example, the exposure wavelength was 248 nm and the numerical aperture NA was 0.
52, the present invention is applied to the simulation of the transfer light intensity distribution in the capacitor pattern of the DRAM shown in FIG. 11 under the exposure condition that the apparent size σ (Partial coherence) of the light source is 0.6. First, FIG.
Triangles were added to or deleted from all the corners of the design pattern shown in FIG. 12 according to the rule of FIG. 12 to obtain a simulation photomask image (FIG. 13) having corner rounding of the mask.

In the rule of FIG. 12, case 1 (see FIG.
2 (A)), the length of two sides is a
Removed right-angled isosceles triangle. In addition, case 2 (see FIG.
In the concave corner of 2 (B), an isosceles right triangle having two sides of length a is added. If the length of the side adjacent to the corner is less than twice a, the case 3
As shown in (FIG. 12C), the length of each side of the triangle is set to half the length of the adjacent side. Here, the length of a is 0.10 μm in units on the wafer.

Using the simulation photomask image having this corner rounding, under the above exposure conditions, a two-dimensional relative light intensity distribution I at just focus and 0.75 μm defocus was obtained by a light intensity simulator based on the scalar diffraction theory. (X, y)
I asked.

I (x, y) = 0.28 of this light intensity distribution
14 and 15 show contour lines of the obtained light intensity assuming that there is no corner rounding and the contour lines in FIG. FIG. 14 shows the case of just focus, and FIG. 15 shows the case of 0.75 μm defocus. As shown in these drawings, it was confirmed that by using the present Example 1, the actually obtained light intensity distribution can be obtained with higher accuracy than in the case where the corner rounding is not considered (Comparative Example 1). Example 2 In this example, the exposure wavelength is 248 nm, the NA is 0.52, and σ is
The present invention is applied to the simulation of the transfer light intensity distribution in the capacitor pattern of the DRAM shown in FIG. 16 under the exposure condition of 0.6.

First, for all corners of the design pattern shown in FIG. 16, triangles are added or deleted according to the rule of FIG. 12, and a photomask image for simulation having mask corner rounding (FIG. 17).
I asked. Here, the length of a is a unit on the wafer,
It was set to 0.10 μm.

Using this simulation photomask image with corner rounding, under the above exposure conditions, a two-dimensional relative light intensity distribution at just focus and 0.75 μm defocus was obtained by a light intensity simulator based on the scalar diffraction theory. I (x,
y) was calculated.

I (x, y) = 0.28 of this light intensity distribution
18 and 19 show the contour line in FIG. 9 and the contour line of the light intensity obtained assuming that there is no corner rounding. As shown in these drawings, it was confirmed that by using the present Example 2, the actually obtained light intensity distribution can be obtained with higher accuracy than in the case where the corner rounding is not considered (Comparative Example 2). Example 3 In this example, the exposure wavelength is 248 nm, NA is 0.52, and σ is
The present invention is applied to the simulation of the transfer light intensity distribution in the wiring pattern of the gate array (G / A) shown in FIG. 20 under the exposure condition of 0.6.

First, for all the corners of the design pattern shown in FIG. 20, triangles are added or deleted according to the rule of FIG. 12, and a photomask image for simulation having mask corner rounding (FIG. 21).
I asked. Here, the length of a is a unit on the wafer,
It was set to 0.10 μm.

Using this simulation photomask image with corner rounding, under the above exposure conditions, a two-dimensional relative light intensity distribution at just focus and 0.75 μm defocus was obtained by a light intensity simulator based on the scalar diffraction theory. I (x,
y) was calculated.

I (x, y) = 0.28 of this light intensity distribution
22 and 23 show the contour line in FIG. 9 and the contour line of the light intensity obtained assuming that there is no corner rounding. As shown in these drawings, it was confirmed that by using the present Example 3, the actually obtained light intensity distribution can be obtained with higher accuracy than in the case where the corner rounding is not considered (Comparative Example 3). Example 4 In this example, the exposure wavelength was 248 nm, the NA was 0.52, and σ was
The present invention is applied to the simulation of the transfer light intensity distribution in the G / A wiring pattern shown in FIG. 24 under the exposure condition of 0.6.

First, a simulation photomask image with mask corner rounding (FIG. 25) with triangles added or deleted for all the corners of the design pattern shown in FIG. 24 according to the rule of FIG.
I asked. Here, the length of a is a unit on the wafer,
It was set to 0.10 μm.

Using the simulation photomask image having this corner rounding, under the above exposure conditions, a two-dimensional relative light intensity distribution I at just focus and 0.75 μm defocus was obtained by a light intensity simulator based on the scalar diffraction theory. (X, y)
I asked.

I (x, y) = 0.28 of this light intensity distribution
26 and 27 show the contour line in FIG. 9 and the contour line of the light intensity obtained assuming that there is no corner rounding. As shown in these figures, it was confirmed that by using the present Example 4, the actually obtained light intensity distribution could be obtained with higher accuracy than in the case where the corner rounding was not considered (Comparative Example 4). Fifth Embodiment In this embodiment, in addition to the corner rounding correction as shown in the first to fourth embodiments, the mask pattern correction for the following design pattern is performed.

The details will be described below. As shown in FIG. 1, the mask pattern correcting apparatus according to the present embodiment includes an input unit 2
A design pattern storage means 4 and a transfer condition storage means 6
, Evaluation point placement means 8 and simulation means 10
The comparison means 12, the deformation means 14, the correction pattern storage means 16, the repetition means 18, and the output means 20 are included.

The input means 2 is not particularly limited as long as it can input a design pattern, transfer conditions, etc., and a keyboard, a touch panel, etc. can be exemplified. When inputting the design pattern and the transfer conditions in the form of electric signals, the input means 2 is
It may be a wired or wireless input terminal. When inputting design patterns and transfer conditions stored in a recording medium such as a floppy disk, the input means 2 is used.
It consists of a disk drive and so on.

Further, as the output means 20, a CRT or a liquid crystal display device capable of displaying at least the corrected design pattern on the screen can be used. Also,
The output means 20 may be an output means such as a printer or an XY plotter capable of drawing at least the corrected design pattern on paper, film or other substrate.

Other means 4, 6, 10, 1 shown in FIG.
2, 14, 16, and 18 are arithmetic circuits or RAMs,
The program information is stored in a storage unit such as a ROM or an optical storage medium and is processed by the CPU of the computer. The operation of the apparatus shown in FIG. 1 will be described based on the flowchart shown in FIG.

In step S10 shown in FIG.
The design pattern and the transfer condition are stored in the design pattern storage unit 4 and the transfer condition storage unit 6 of the correction apparatus shown in FIG. FIG. 28 shows an example of the design pattern. The transfer conditions include, for example, the wavelength λ of light used for exposure, the numerical aperture NA, the apparent size σ (Partial coherence) of the light source or the transmittance distribution of the light source, the phase / transmittance distribution of the exit pupil, and defocus. Is a condition regarding.

Next, in step S11 shown in FIG. 2, a plurality of evaluation points are created along the outer periphery of the design pattern. The evaluation points are created by the evaluation point placement means 8 based on the design patterns stored in the design pattern storage means 4 shown in FIG. For example, the evaluation points are calculated along the outer circumference of the design pattern 32 shown in FIG.
It is given based on a certain rule at equal intervals or unequal intervals. The arrangement interval of the evaluation points is, for example, about 0.2 μm.

Next, in step S12 shown in FIG. 2, the transfer resist pattern (transfer image) is calculated by the simulation means 10 shown in FIG. In this embodiment, when calculating the transfer resist pattern in step S12 shown in FIG. 2, the design pattern is not directly subjected to the light intensity simulation, but as shown in FIG.
Corner rounding is added to the design pattern (step S17) (step S18). The addition of this corner rounding is the same as in Examples 1 to 4 above, and in all the convex corners, the isosceles triangle whose one side is 0.08 μm in units on the wafer is deleted, and in all the concave corners. , One side is 0.08μ on the wafer
Add an isosceles triangle of m. As a result, the simulation photomask image with the corner rounding obtained in the actual mask making process (see Fig. 2
9) is obtained. A two-dimensional light intensity distribution obtained when the simulation photomask image having this corner rounding is transferred by just focus is obtained (steps S19 to S21 shown in FIG. 3). The light intensity distribution may be obtained using an actual light intensity measuring device.

After the two-dimensional light intensity distribution is obtained, the latent image forming intensity is calculated in step S22 shown in FIG. 3, and the latent image forming intensity distribution is obtained in step S23. Hereinafter, the processing in the latent image formation intensity calculation will be described in detail. In the latent image formation intensity calculation, for example, the latent image formation intensity Mj0 at the point j0 on the wafer plane shown in FIG. 5 is defined as the point jn (where n is 0≤n≤
(An integer that satisfies 24) is determined in consideration of the influence of light intensity. Here, the effect of light intensity at jn point Mj0
Define jn as in the following expression (1).

[0066]

[Equation 1]

In the above equation (1), rn is j0 point and j
The distance to the n point is shown, and f (rn) is shown by the following equation (2).

[0068]

[Equation 2]

However, in the equation (2), the following equation (3) is satisfied. That is, the equation (2) is defined using the Gaussian function.

[0070]

[Equation 3]

In the equation (1), g (I (jn
)) Is defined by the following equation (4).

[0072]

[Equation 4]

That is, the effect M of the light intensity at the point jn
j0 jn is a value obtained by multiplying the distance rn between the points j0 and jn by the light intensity I (jn) at the point jn. In the latent image formation intensity calculation, for example, in the case shown in FIG. 5, the latent image formation intensity Mj0 is obtained by accumulating the influence Mj0jn of the light intensity at the jn point on the exposure energy at the j0 point.

At this time, for example, the size of the wafer is 2
Considering the influence of the light intensity from an infinite number of jn (−∞ ≦ n ≦ ∞) points which are infinite in the dimension direction and are arranged in a predetermined pattern, Mj0 is represented by the following formula (5). Indicated by.

[0075]

[Equation 5]

Substituting equations (2) and (4) into equation (5), Mj0 is defined by equation (6).

[0077]

[Equation 6]

In the latent image formation intensity calculation, Mj0 at the points arranged in a predetermined pattern on the two-dimensional plane on the wafer is calculated in the above-described manner, and the latent image formation on the two-dimensional plane is formed based on the calculation result. Calculate the intensity distribution. Next, in the latent image formation intensity distribution obtained by the above-mentioned procedure, as shown in FIG.
In step S24 shown in step S24, a contour line at which the latent image forming intensity reaches the threshold value is obtained, and the pattern defined by this contour line is set as a resist pattern in step S25. At this time,
The threshold value is determined, for example, according to the exposure amount and the developing condition.

An example of a method of calculating the optimum threshold value Eth and the constant α used in the above simulation will be shown below. A plurality of resist patterns are calculated based on various exposure times and defocus conditions, and the processing shown in FIG. 6 is performed using the calculated resist patterns.

Here, the latent image forming intensity R (x, y) in the latent image forming intensity distribution is defined by the following equation (7), for example. In Expression (7), α is a constant.

[0081]

[Equation 7]

Step S1: In the resist pattern calculated by the simulation method shown in FIG. 3, line widths at a plurality of points are obtained. At this time, the target line has a wide range of line width. Step S2: A transfer experiment is actually performed using the same mask pattern and exposure conditions as in this simulation method, and the line width of the line corresponding to the target line in Step S1 is obtained.

Step S3: With respect to the line widths of the plurality of lines obtained in steps S1 and S2, a difference is obtained by the resist pattern calculation method and the transfer experiment. Step S4: The square value of the difference obtained in step S3 is obtained, and the square value is accumulated for a plurality of lines to obtain the accumulated value.

Step S5: A constant α and a threshold value Eth that minimize the cumulative value obtained in step S4 are calculated.
At this time, the determination of the constant α and the threshold value Eth is performed, for example, by using the predetermined constant α and the threshold value Eth as initial values to perform the simulation shown in FIG. 3 and the processes of steps S1 to 4 of FIG. The cumulative value of step S4 shown in FIG. 6 in the previous processing is compared with the cumulative value of step S4 in the current processing, and the constant α and the threshold value used in the next processing are set so as to reduce the difference between the cumulative values. Determine the value Eth. Then, using the constant α and the threshold value Eth, the simulation of FIG. 3 and the processes of steps S1 to S4 of FIG. 6 are performed again. Then, this procedure is repeated to obtain the constant α and the threshold value Eth when the difference between the cumulative values becomes the minimum.

Step S6: The above-mentioned simulation shown in FIG. 3 is performed using the equation (7) in which the constants α and Eth are the values calculated in step S5. According to the method shown in FIG. 6, the constants α and Eth in the above equation (7) when performing the simulation shown in FIG. 3 can be set appropriately. Therefore, the accuracy of the simulation shown in FIG. 3 can be further improved.

In the above-described embodiment, regarding the line widths at a plurality of corresponding positions, the maximum value of the line width difference between the resist pattern obtained by the resist pattern calculation method and the resist pattern obtained by the experiment is the minimum. The constant α and Eth may be calculated so that

In this embodiment, the Gaussian function is used in the equations (2) and (3) used in the latent image formation intensity calculation (step S22 shown in FIG. 3). However, this function is The function is maximized when the distance rn is 0, and is 0 when the distance rn is infinite.

In the present embodiment, the latent image forming intensity is defined by the product of the light intensity and the distance in the latent image forming intensity calculation in step S22 shown in FIG. It may be defined by the product of the power of and the distance. In this case, the latent image forming intensity Mjo is defined by the following equation (8), for example.

[0089]

[Equation 8]

Further, in the present embodiment, the case where the exposure is performed by using the i-line is illustrated, but the present invention can be applied to the case where the pattern is formed by using the X-ray or the EB (electron beam). . In this embodiment, the latent image forming intensity distribution used when obtaining the resist pattern is determined in consideration of not only the light intensity of the point of interest but also the influence of the light intensity of the peripheral points, so that it is more accurate. A resist pattern (transfer image) can be calculated.

Next, it is shown that a result close to an actual transfer resist pattern can be obtained by performing the simulation shown in FIG. In this example, the constant α in the above equation (6) is set to 0.131 and Eth is set to 197.01.

Further, in this example, in the L / S transfer experiment, i line having a wavelength of 365 nm was used, NA was 0.50, σ
Under the exposure condition of 0.68, the defocus and the exposure time were changed, and the positive resist for i-line of Company A was actually exposed. FIG. 8A shows the line width (S) obtained in the L / S transfer experiment at each defocase and exposure time.
EM), the line width (this method) obtained using the resist pattern calculation method of this example, the line width (this method), and the line width (SE
It is a correspondence table with the difference with M).

FIG. 8B is a graph in which the line width (present method) and line width (SEM) shown in FIG. 8A are plotted on the vertical axis and the exposure time is plotted on the horizontal axis. From the experimental results shown in FIG. 8A, in this example, 3σ = 0.153.
On the other hand, a resist pattern was created using a conventional resist pattern calculation method that does not use latent image formation intensity calculation. At this time, when obtaining the resist pattern from the light intensity distribution, the threshold value Eth was set to 193.54.

Also, in this comparative example, in the L / S transfer experiment, i-line having a wavelength of 365 nm was used, NA was 0.50,
Under the exposure condition of σ of 0.68, the defocus and the exposure time were changed and the i-line positive resist of Company A was actually exposed. FIG. 9A shows the line width (SEM) obtained in the L / S transfer experiment at each defocase and exposure time, the line width obtained by using the resist pattern calculation method of this comparative example (conventional method), and the line width. It is a correspondence table of the difference between the width (conventional method) and the line width (SEM).

FIG. 9B is a graph in which the line width (conventional method) and line width (SEM) shown in FIG. 9A are plotted on the vertical axis and the exposure time is plotted on the horizontal axis. From the experimental results shown in FIG. 9A, in this comparative example, 3σ = 0.0313.

Comparing the result shown in FIG. 8 (this example) with the result shown in FIG. 9 (comparative example), 3σ according to this example is about ½ of that of the comparative example. It was confirmed that the accuracy of the simulation shown in (3) was good. As described above, in step S25 shown in FIG. 3, after obtaining an image of the resist pattern that will actually be formed, then in step S13 shown in FIG. The difference) is calculated for each evaluation point by the comparison means 12 shown in FIG. The measurement direction of the deviation of the resist edge position of the design pattern at this time is a direction vertical to the boundary line (edge) of the design pattern 32, as shown in FIG. The outside of the pattern 32 is the positive direction and the inside is the negative direction. Further, in the corner portion of the design pattern 32, the deviation measuring direction is the direction of the sum of the direction vectors of the two sides forming the corner portion, and similarly, the outside of the pattern is the positive direction.

Next, in step S14 shown in FIG. 2, depending on the deviation (difference) compared for each evaluation point 30, the deforming means 14 shown in FIG. 32 is deformed and corrected. An outline of the deformation correction method is shown in FIG. As shown in FIGS. 4A and 4B, when correcting the deformation of the design pattern 32, a constant coefficient is applied to the magnitude of the difference in the opposite direction of the deviation (difference) compared for each evaluation point 30. The boundary line of the mask pattern in the vicinity of the evaluation point 30 (including not only the evaluation point but also the boundary line in the vicinity thereof) is moved by the size multiplied by. The coefficient is preferably more than 0 and less than 1, and more preferably 0.10.
It is 0.50. If this coefficient is too large, the deformation will be excessively corrected, and the transferred image may be separated from the design pattern instead of coming close to the design pattern even by repeated calculation described later. Note that the coefficient may be constant at all evaluation points, but may be different at specific evaluation points. FIG. 32 shows an example of the design pattern corrected in this way.

The corrected design pattern is stored in the correction pattern storage means 16 shown in FIG. When a good correction pattern is obtained by these series of operations, the corrected mask pattern is obtained in step S15 shown in FIG. The corrected mask pattern is output on the screen or on the paper or film by the output means 20 shown in FIG.

Upon receiving the signal from the repeating means 18 shown in FIG. 1, the simulation means 10, the comparison means 12 and the deformation means 14 are used based on the corrected design pattern stored in the correction pattern storage means 16. It is preferable to repeat the steps S12 to S14 shown in FIG. 2 once or more. At this time, the position of the reference evaluation point 30 is not changed. That is, the transfer image is obtained again based on the corrected design pattern, the shift (difference) between the transfer image and the reference point is obtained, and the corrected design pattern is deformed and corrected again based on the difference. By repeating these operations, the transferred image gradually approaches the original design pattern (position of the evaluation point).

In the correction apparatus and correction method of this embodiment,
Regardless of the design pattern, the mask pattern of the photomask can be automatically deformed so that a transfer image close to a desired design pattern can be obtained. Therefore, if the photolithography process is performed using the photomask having the corrected design pattern obtained in this example, a resist pattern as close as possible to the original design pattern can be obtained, and a bridge or disconnection occurs. Absent. As a result, a semiconductor device having good electric characteristics can be manufactured with a high yield.

Sixth Embodiment In this embodiment, in addition to the corner rounding correction as shown in the first to fourth embodiments, the mask pattern correction for the following design pattern is performed. The details will be described below.

In this embodiment, the design is carried out in the same manner as in the fifth embodiment except that a target point is set in addition to the evaluation point in step S11 shown in FIG. 2 performed by the evaluation point placement means 8 shown in FIG. Correct the pattern. Hereinafter, only parts different from the fifth embodiment will be described.

In the present embodiment, as shown in FIG. 10, the target point 36 is set in correspondence with the convex corner portion or the evaluation point 30 located at the concave corner portion of the design pattern 32, and at the convex corner portion. , Target point 3 inside the corner (for example, -0.08 μm)
6 is determined, and at the concave corner, outside the corner (for example, +
The target point 36 is determined to be 0.08 μm).

In this embodiment, step S13 shown in FIG.
In the comparison step of, the simulated transfer image 34 is generated at the position where only the evaluation point 30 is set.
And the difference a between the target pattern 36 and the design pattern 32 are compared for each evaluation point 30, and the difference b between the target point 36 and the transferred image 34 is compared at the position where the target point 36 is set. And
In the transformation process of step S14 shown in FIG.
The design pattern 32 is deformed based on the evaluation point 30 (not the target point) so that the difference becomes smaller depending on the differences a and b compared for each 0 or for each target point 36.

For example, in the case of the convex corners or the concave corners of the design pattern 32, when the evaluation points 30 are located on these corners, the mask pattern is aimed at that the transfer image approaches the evaluation points 30 themselves. If the correction is performed, the transferred image at a position other than the corner may be separated from the design pattern 32.

In the mask pattern correction method according to this embodiment, for example, in the case of a convex corner, the target point 36 is located inside the corner.
In the concave corner portion, the target point 36 is determined outside the corner portion, and the transfer pattern 34 is corrected so that the transfer image 34 approaches the target point 36. As a result, it is possible to satisfactorily approach the design pattern. As a result, it is possible to favorably prevent bridging or disconnection between patterns.

Embodiment 7 In this embodiment, in addition to the corner rounding correction as shown in the first to fourth embodiments, the mask pattern correction for the following design pattern is performed. The details will be described below.

In the present embodiment, except that the simulation performed in step S12 shown in FIG. 2 using the simulation means 10 shown in FIG. 1 is performed under a plurality of transfer conditions,
The design pattern is corrected in the same manner as in the fifth or sixth embodiment. Only parts different from the above embodiment will be described below.

That is, in the present embodiment, in the simulation process, a plurality of types of transfer based on a combination of a plurality of exposure amounts having a preset exposure allowance and a plurality of focal positions within a preset focal depth range. A plurality of transfer images are obtained by simulating each transfer image under the conditions. Then, in the comparison step of step S13 shown in FIG. 2, the difference between the plurality of transfer images and the design pattern is compared for each evaluation point, and a plurality of differences are calculated for each evaluation point. . Then, in the deforming step of step S14 shown in FIG. 2, the design pattern is deformed so that a plurality of differences for each evaluation point are reduced by a predetermined reference.

The predetermined criterion in the transformation step is, for example, a criterion that minimizes the average value of a plurality of differences for each evaluation point.
In addition, as another predetermined criterion, a criterion that minimizes the difference between the maximum difference and the minimum difference among a plurality of differences for each evaluation point is exemplified.

Furthermore, as another predetermined criterion, a criterion that minimizes the root mean square of a plurality of differences for each evaluation point is exemplified. According to the mask pattern correction method of the present embodiment, since the transfer image when the transfer condition is changed within the range of the process margin is taken into consideration (the process margin is taken into consideration), the corrected mask pattern is corrected. As a result, the process latitude such as exposure latitude and depth of focus will not deteriorate. As a result, if the photolithography is performed using the photomask having this mask pattern, the manufacturing yield is improved.

Example 8 In this example, in addition to the corner rounding correction as shown in Examples 1 to 4, the pattern edge correction shown in Examples 5 to 7 was performed, and the actual pattern was 0.4 μm. The pattern of the polysilicon layer of the memory device of the rule is exposure wavelength 365 nm, NA = 0.50, σ = 0.6
The present invention was applied to the case of exposing a positive novolac resist under the condition of No. 8.

FIG. 28 shows a design pattern applied in this embodiment. First, evaluation points were generated at 0.2 μm pitch on all corners and all sides of the design pattern.
Next, in this design pattern, an isosceles triangle whose one side is 0.08 μm on the wafer is deleted at all convex corners, and one side is 0.08 μm on the wafer at all concave corners. Added an isosceles triangle. As a result, a simulation photomask image (FIG. 29) having corner rounding obtained in the actual mask making process was obtained. The two-dimensional light intensity distribution obtained when the photomask image for simulation with this corner rounding was transferred by just focus was obtained.

This is subjected to convolution integration as shown in, for example, the formula (6) in the fifth embodiment, and the contour line obtained by slicing the convolution integration value at the threshold value Eth is obtained as a resist image (FIG. 30). . However, the threshold value Eth is L in FIG.
Is set to 0.4 μm.

Further, the depth of focus required in the lithography process was set to ± 0.75 μm, the convolution of the light intensity distribution and the Gaussian function at 0.75 μm defocus was obtained, and the contour line sliced at Eth was obtained (FIG. 31). . Further, the exposure image required in the lithography process is ± 10%, and in the above two convolutions, the contour image obtained by slicing at the height Eth ⁻ with the Eth reduced by 10%, the resist image when the exposure amount is increased by + 10% As the resist image, the contour lines sliced at the height Eth ⁺ with Eth increased by 10% were obtained as resist images when the exposure amount was decreased by 10%. As a result, a total of 6 types of resist images were calculated, with the focus position being just focus and 0.75 μm defocus, and the exposure amount being the optimum exposure amount, 10% overdose, and −10% underdose.

Subsequently, for all the evaluation points, the deviation amount of the resist edge position was calculated from the evaluation points for the six types of resist image edges. At this time, the direction to measure the deviation of the edge position is
The direction perpendicular to the edge was set, and the direction outside the pattern was set to the positive direction. At the corner point, the direction is the sum of the direction vectors of the two sides forming the corner, and the outside of the pattern is also the positive direction.

The average value of the amount of edge deviation under the six conditions for each evaluation point thus obtained was obtained. However,
At the corner evaluation points, the target value of the edge shift amount is set to −0.07 μ for the corners protruding outward in order to prevent excessive correction of the pattern to be performed later.
m, and +0.07 μm for the corners concave outward, the difference between these target values and the deviation amount of the evaluation point of the edge position was calculated, and the average value thereof was calculated.

In this way, the mask pattern edge near each evaluation point was moved in the opposite direction with respect to the obtained average value of the deviation amount of the edge. Here, the movement amount of the pattern edge is
The average value of the shift amounts was multiplied by 0.25. Further, these procedures were performed again by keeping the position of the evaluation point as it was and using the pattern obtained by the above procedure as the input pattern.

By repeating this procedure four times, FIG.
I got the design mask pattern. An actual mask pattern in consideration of the corner rounding at this time is as shown in FIG. By this correction, at each evaluation point, there are two focus positions of just focus 0.75 μm defocus, optimum exposure amount, 10% overdose, and −10.
% Underdose in the resist pattern in the combination of three different exposure doses was 3236, which was 0.236 μm in the mask without correction.
It succeeded in reducing to 122 μm. In the designed mask pattern of FIG. 32, a mask having an actual corner rounding (FIG. 33) is obtained, and the resist patterns at just focus and 0.75 μm defocus obtained by this are shown in FIGS. , It is shown that a very good resist pattern is obtained as compared with FIGS.

By using this mask, a semiconductor device having good electric characteristics could be manufactured with a high yield.
The present invention is not limited to the above embodiment. The exposure conditions are not limited to the values described in this embodiment, the photoresist used is not limited to this embodiment, and the mask pattern is not limited to this embodiment.

Further, the modified illumination method or the pupil filtering method may be used as the exposure method, and the mask used may be a phase shift mask such as a halftone method or a Levenson method, which is not limited to the above embodiment. Not a thing.

[0122]

As described above, according to the present invention, the transfer light intensity can be accurately simulated at high speed. As a result, it was possible to provide a means for quick and accurate process evaluation at low cost.

The present invention solves the problems of the prior art and makes it possible to obtain a resist pattern close to the design pattern.
It is possible to correct the mask pattern, thereby providing a means for producing a high-performance device with a high yield.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of a mask pattern correction apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for correcting a mask pattern using the correction device shown in FIG.

FIG. 3 is a flowchart showing a procedure for calculating (simulating) a transfer resist pattern.

FIG. 4 (A) is a schematic diagram showing a method for measuring a shift of a resist edge for each evaluation point, and FIG. 4 (B) is a schematic diagram showing a mask pattern correcting / deforming step.

5 is a diagram showing a concept of part of the procedure shown in FIG. 3;

FIG. 6 is a flowchart showing an example of how to obtain a threshold value Eth and a constant α used for calculation of latent image formation intensity.

7A and 7B are schematic diagrams showing that the two-dimensional light intensity distribution is not necessarily coincident with the actual pattern line width when simply sliced with a threshold value.

FIG. 8 (A) is a line width (SEM) obtained in an L / S transfer experiment at each defocase and exposure time.
And a line width (present method) obtained by using the simulation method according to the embodiment, and a correspondence diagram of the difference between the line width (present method) and the line width (SEM), and FIG. 6 is a graph in which the line width (present method) and line width (SEM) shown in (A) are plotted on the vertical axis and the exposure time is plotted on the horizontal axis.

FIG. 9A is a line width (SEM) obtained in an L / S transfer experiment at each defocase and exposure time.
And the line width (conventional method) obtained using the simulation method according to the comparative example, and the line width (conventional method) and the line width (SE
FIG. 9B is a correspondence chart with the difference between M) and FIG.
6 is a graph in which the line width (conventional method) and line width (SEM) shown in (A) are plotted on the vertical axis and the exposure time is plotted on the horizontal axis.

FIG. 10 is a schematic diagram showing a method of setting a target point according to the embodiment.

FIG. 11 is a plan view showing an initial design pattern according to an embodiment of the present invention.

12A to 12C are schematic diagrams showing corner rounding correction.

FIG. 13 is a schematic diagram showing a pattern after corner rounding correction with respect to the design pattern of FIG. 11.

FIG. 14 is a contour line of the light intensity distribution at just focus at I (x, y) = 0.289,
It is a figure which shows the contour line of the obtained light intensity, assuming that there is no corner rounding.

FIG. 15 is a diagram showing a contour line of the light intensity distribution at 0.75 μm defocus at I (x, y) = 0.289 and a contour line of the light intensity obtained assuming that there is no corner rounding. is there.

FIG. 16 is a schematic diagram showing another example of a design pattern.

FIG. 17 is a schematic diagram showing a pattern after corner rounding correction with respect to the design pattern of FIG. 16;

FIG. 18 is a contour line of the light intensity distribution at just focus at I (x, y) = 0.289,
It is a figure which shows the contour line of the obtained light intensity, assuming that there is no corner rounding.

FIG. 19 is a diagram showing a contour line of I (x, y) = 0.289 of the light intensity distribution at 0.75 μm defocus and a contour line of the light intensity obtained assuming that there is no corner rounding. is there.

FIG. 20 is a schematic diagram showing another example of a design pattern.

21 is a schematic diagram showing a pattern after corner rounding correction with respect to the design pattern of FIG. 20.

22 is a contour line of the light intensity distribution at just focus at I (x, y) = 0.289, and FIG.
It is a figure which shows the contour line of the obtained light intensity, assuming that there is no corner rounding.

FIG. 23 is a diagram showing a contour line of the light intensity distribution at 0.75 μm defocus at I (x, y) = 0.289 and a contour line of the light intensity obtained assuming that there is no corner rounding. is there.

FIG. 24 is a schematic diagram showing another example of a design pattern.

FIG. 25 is a schematic view showing a pattern after corner rounding correction with respect to the design pattern of FIG. 24.

FIG. 26 is a contour line of the light intensity distribution at just focus at I (x, y) = 0.289,
It is a figure which shows the contour line of the obtained light intensity, assuming that there is no corner rounding.

FIG. 27 is a diagram showing contour lines of the light intensity distribution at 0.75 μm defocus at I (x, y) = 0.289 and contour lines of the light intensity obtained assuming that there is no corner rounding. is there.

FIG. 28 is a schematic diagram showing another example of a design pattern.

FIG. 29 is a schematic diagram showing a pattern after corner rounding correction with respect to the design pattern of FIG. 28.

FIG. 30 is a schematic diagram of a resist image obtained when the mask of FIG. 29 is used in just focus.

FIG. 31 is a schematic view of a resist image obtained when the mask of FIG. 29 is used in 0.75 μm defocus.

FIG. 32 is a schematic diagram of a mask pattern on which edge correction has been performed several times.

FIG. 33 is a schematic view of a mask pattern obtained by performing corner rounding correction on the pattern shown in FIG. 32.

FIG. 34 is a schematic view of a resist image obtained when the mask of FIG. 33 is used in just focus.

FIG. 35 is a schematic diagram of a resist image obtained when the mask of FIG. 33 is used in 0.75 μm defocus.

[Explanation of symbols]

2 ... Input means 4 ... Design pattern storage means 6 ... Transfer condition storage means 8 ... Evaluation point arrangement means 10 ... Simulation means 12 ... Comparison means 14 ... Deformation means 16 ... Correction pattern storage means 18 ... Repeating means 20 ... Output means 30 ... Evaluation point 32 ... Design pattern 34 ... Transfer image 36 ... Target point

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI H01L 21/30 516D 519 (56) Reference JP-A-4-179952 (JP, A) JP-A-6-343207 (JP, A ) JP-A 5-343285 (JP, A) JP-A 61-18086 (JP, A) JP-A 6-120101 (JP, A) JP-A 1-107530 (JP, A) JP-A 2- 39152 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G03F 1/00 H01L 21/30

Claims (23)

(57) [Claims] 1. A method for correcting a mask pattern in which a mask pattern of a photomask used in a photolithography process is deformed so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement step of arranging a plurality of evaluation points, using a photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, Consider the corner rounding on the photomask that occurs in the actual mask making process. And a simulation step of calculating a plurality of images after transfer by performing a light intensity simulation under the plurality of transfer conditions using the photomask image for simulation. A plurality of differences between each of the transferred images and the design pattern, a comparison step of comparing each of the evaluation points, a criterion by which the average value of the plurality of differences compared for each of the evaluation points is the minimum And a modification step of modifying the design pattern so as to reduce the difference. 2. A mask pattern correction method for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement step of arranging a plurality of evaluation points, using a photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, Consider the corner rounding on the photomask that occurs in the actual mask making process. And a simulation step of calculating a plurality of images after transfer by performing a light intensity simulation under the plurality of transfer conditions using the photomask image for simulation. A plurality of differences between each of the transferred images and the design pattern, a comparison step of comparing each evaluation point, and among the plurality of differences compared for each evaluation point, the absolute value is the maximum. A method of correcting a mask pattern, which comprises using a criterion that minimizes the difference and deforming the design pattern so that the difference becomes smaller. 3. A mask pattern correction method for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement step of arranging a plurality of evaluation points, using a photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, Consider the corner rounding on the photomask that occurs in the actual mask making process. And a simulation step of calculating a plurality of images after transfer by performing a light intensity simulation under the plurality of transfer conditions using the photomask image for simulation. A plurality of differences between each of the transferred images and the design pattern, a comparison step of comparing each of the evaluation points, and a criterion that minimizes the root mean square of the plurality of differences that are compared for each of the evaluation points. And a modification step of modifying the design pattern so as to reduce the difference. 4. In the simulation step, a two-dimensional light intensity on a substrate is calculated based on the simulation photomask image and exposure conditions, and a peripheral position of an arbitrary focused position on the two-dimensional plane of the substrate is calculated. Based on the light intensity in, and the distance between the focused position and the peripheral position, by calculating and accumulating the influence of the light intensity at the peripheral positions on the exposure energy of the focused arbitrary position, The latent image forming intensity at the focused arbitrary position is calculated on the two-dimensional plane of the substrate, the distribution of the latent image forming intensity on the two-dimensional plane of the substrate is calculated, and the latent image corresponding to the exposure amount and the developing condition is calculated. The threshold of formation intensity is determined, the contour line at the threshold is obtained for the latent image formation intensity distribution, and the pattern defined by the contour line is transferred. Method of correcting the mask pattern according to any one of claims 1 to 3, characterized in that calculated as an image. 5. In the deforming step, a boundary of a mask pattern in the vicinity of the evaluation point is increased in the opposite direction of the difference compared for each evaluation point by a size obtained by multiplying the magnitude of the difference by a constant coefficient. The method of correcting a mask pattern according to claim 1, wherein the line is moved. 6. The mask pattern correction method according to claim 5, wherein the coefficient is greater than 0 and less than 1. 7. In the evaluation point arrangement step, a plurality of evaluation points are arranged along the outer circumference of the pattern of the desired design pattern, and at a predetermined evaluation point, a target point is set separately from the evaluation point. However, in the comparison step, at the position where only the evaluation points are set, the difference between the simulated transfer image and the design pattern is compared for each of the evaluation points, and at the position where the target point is set. Compares the difference between the target point and the transferred image, and in the deforming step, depending on the difference compared for each evaluation point or for each target point, the design pattern is changed so that the difference becomes small. It deforms, The claim 1 characterized by the above-mentioned.
7. The method for correcting a mask pattern according to any one of 6 above. 8. The target point is set corresponding to an evaluation point located at a convex corner portion or a concave corner portion of the design pattern, and the target point is determined inside the corner portion at the convex corner portion. The method of correcting a mask pattern according to claim 7, wherein the target point is determined outside the corner at the concave corner. 9. A photomask image for simulation in which a triangle is deleted at a convex corner and a triangle is added at a concave corner at a corner of a design pattern, and a simulation photomask image is used. A method for correcting a mask pattern as described in. 10. The mask pattern correction method according to claim 9, wherein one side of the triangle is greater than 0 and 100% or less with respect to the minimum line width of the design pattern. 11. An exposure method for performing exposure using a photomask having a mask pattern corrected by the method for correcting a mask pattern according to claim 1. Description: 12. A mask pattern correction device for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement means for arranging a plurality of evaluation points, and using the photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, The corner rounding on the photomask that occurs in the actual mask making process A simulation means for calculating a plurality of post-transfer images by creating a considered photomask image for simulation, and performing light intensity simulation under the plurality of transfer conditions using the photomask image; and the plurality of simulations. Comparing means for comparing a plurality of differences between each of the transferred images and the design pattern for each of the evaluation points, and an average value of the plurality of differences for each of the evaluation points is minimum. A mask pattern correction apparatus, which includes a deforming unit that deforms the design pattern so that the difference is reduced using a reference. 13. A mask pattern correction device for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement means for arranging a plurality of evaluation points, and using the photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, The corner rounding on the photomask that occurs in the actual mask making process A simulation means for calculating a plurality of images after transfer by creating a photomask image for simulation in consideration and performing light intensity simulation under the plurality of transfer conditions using the photomask image; and the plurality of simulations. A plurality of differences between each of the transferred images and the design pattern, comparing means for comparing each of the evaluation points, and among the plurality of differences compared for each of the evaluation points, the absolute value is the maximum. A mask pattern correction apparatus including a deforming unit that deforms the design pattern so that the difference is reduced using a criterion that minimizes the difference. 14. A mask pattern correction device for deforming a mask pattern of a photomask used in a photolithography process so that a transfer image close to a desired design pattern can be obtained. Along with the evaluation point arrangement means for arranging a plurality of evaluation points, and using the photomask of the design pattern with the evaluation points, a plurality of exposure amounts of the preset exposure allowance and the preset depth of focus When simulating a transfer image obtained when exposure is performed under a plurality of transfer conditions based on a combination with a plurality of focus positions within a range, by adding or deleting a figure to a corner of the design pattern, The corner rounding on the photomask that occurs in the actual mask making process A simulation means for calculating a plurality of images after transfer by creating a photomask image for simulation in consideration and performing light intensity simulation under the plurality of transfer conditions using the photomask image; and the plurality of simulations. A plurality of differences between each of the transferred images and the design pattern, comparing means for comparing each of the evaluation points, and a root mean square of the plurality of differences compared for each of the evaluation points is minimized. A mask pattern correction apparatus, which includes a deforming unit that deforms the design pattern so that the difference is reduced using a reference. 15. The simulation means calculates a two-dimensional light intensity on a substrate based on the simulation photomask image and exposure conditions, and determines a peripheral position of an arbitrary focused position on the two-dimensional plane of the substrate. Based on the light intensity in, and the distance between the focused position and the peripheral position, by calculating and accumulating the influence of the light intensity at the peripheral positions on the exposure energy of the focused arbitrary position, The latent image forming intensity at the focused arbitrary position is calculated on the two-dimensional plane of the substrate, the distribution of the latent image forming intensity on the two-dimensional plane of the substrate is calculated, and the latent image corresponding to the exposure amount and the developing condition is calculated. The threshold of the formation intensity is determined, the contour line of the latent image formation intensity is obtained, and the contour line at the threshold value is obtained. Correction apparatus of the mask pattern according to any one of claims 12 to 14, characterized in that calculated as shooting images. 16. The boundary of the mask pattern in the vicinity of the evaluation points, in the deforming means, in the opposite direction of the differences compared for each of the evaluation points, the size of the difference being multiplied by a constant coefficient. claim, characterized in that to move the line 12-15
5. A mask pattern correction device according to any one of 1. 17. The mask pattern correction apparatus according to claim 16 , wherein the coefficient is greater than 0 and less than 1. 18. The evaluation point arrangement means arranges a plurality of evaluation points along the outer periphery of the pattern of the desired design pattern, and at a predetermined evaluation point, sets a target point separately from the evaluation points. However, in the comparison means, at the position where only the evaluation points are set, the difference between the simulated transfer image and the design pattern is compared for each of the evaluation points, and at the position where the target point is set. Compares the difference between the target point and the transferred image, and in the deforming means, depending on the difference compared for each evaluation point or for each target point, the design pattern is changed so that the difference becomes small. claim, characterized in that deformation to 12
18. The mask pattern correction device according to any one of 17 to 27 . 19. The target point is set corresponding to an evaluation point located at a convex corner portion or a concave corner portion of the design pattern, and the target point is determined inside the corner portion at the convex corner portion. 19. The mask pattern correction apparatus according to claim 18 , wherein the target point is determined outside the corner at the concave corner. 20. A simulation photomask image in which a triangle is deleted at a convex corner and a triangle is added at a concave corner at a corner of a design pattern is used, and any one of claims 12 to 19 is used. A device for correcting a mask pattern according to item 1. 21. The mask pattern correction apparatus according to claim 20 , wherein one side of the triangle is greater than 0% and 100% or less with respect to the minimum line width of the design pattern. 22. Manufacturing of a photomask, comprising: the mask pattern correction device according to claim 12 ; and a drawing means for drawing a photomask of the mask pattern corrected by the mask pattern correction device. apparatus. 23. A semiconductor comprising: the mask pattern correction device according to claim 12 ; and an exposure unit that performs exposure using a photomask of the mask pattern corrected by the mask pattern correction device. Equipment manufacturing equipment.