JP2003124097A - Mask and manufacturing method thereof, projection aligner, and manufacturing method of semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a mask that can transfer an angled pattern without forming an angle at the opening section of the mask, and a manufacturing method of the mask, and to provide a projection aligner and a manufacturing method of a semiconductor device. SOLUTION: In a mask 10, the manufacturing method of the mask 10, a projection aligner, and the manufacturing method of a semiconductor device; the mask 10 has a first mask 21 for shielding a charged particle beam (preferably, an electron beam), a first opening 23 that does not have an angle for transmitting the charged particle beam formed in the first mask 21, a second mask 22 that is used while being overlapped to the first mask 21 and shields the charged particle beam, and a second opening 24 that is formed in the second mask so that it has a portion that is overlapped to the first opening and does not have any corners for transmitting the charged particle beam.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a mask used in lithography, a method of manufacturing the mask, an exposure apparatus, and a method of manufacturing a semiconductor device.

[0002]

2. Description of the Related Art One of the manufacturing processes for semiconductor devices such as IC and LSI is a lithography process for forming a fine circuit pattern on a semiconductor substrate. The performance of a semiconductor device is largely determined by how many circuits are provided in the device. The degree of integration of a semiconductor device greatly depends on the size of a circuit pattern formed on a substrate. In recent years, miniaturization and high integration of semiconductor devices have been significantly advanced.

As a method of forming an integrated circuit pattern on a semiconductor substrate, a photolithography method using ultraviolet light has been generally used so far. However, as the circuit pattern becomes finer, the resolution limit of light begins to become a concern, and charged particle beams such as electron beams and ion beams and X
Higher resolution lithographic techniques using lines are being investigated.

For example, according to the exposure technique using a charged particle beam, the beam diameter can be narrowed down to the nm order, and a fine pattern of 100 nm or less can be easily formed. Above all, electron beam drawing technology has been put into practical use for a relatively long time. However, it takes an enormous amount of time to form a large area or a large pattern by the direct writing method in which an electron beam that is extremely narrowed down is scanned and written. Therefore, there is a problem that the processing amount (throughput) per unit time is low.

Therefore, as a lithography method in manufacturing a semiconductor integrated circuit, a photolithography method using ultraviolet light as a light source is still often used. The electron beam direct writing method is limited in that low throughput is relatively less problematic, such as in manufacturing a reticle (mask) for photolithography and a prototype of a next-generation device in which it is difficult to meet design rules in photolithography. Was used for different purposes.

In order to solve the problem of low throughput, there is a method of drawing a predetermined pattern using a variable shaped electron beam in the 1980s, instead of directly drawing with a Gaussian shaped electron beam as before. Appeared. Furthermore, in the 1990s, a lithography technology for reducing the partial batch pattern and writing on the wafer, called block exposure or cell projection method, appeared (Science Forum published in November 1994, ULS).
I Lithography Technology Innovation ”p.177).

Due to these technological advances, the throughput of electron beam direct writing has been dramatically improved. In addition, SCALPEL (sc
attering with angular limitation in projection ele
ctron-beam lithography ／ ST Stanton etc. Proceeding
s of SPIE 3676 p.194 (1999)), and PREVAIL (projection e developed by IBM in collaboration with Nikon).
xposure with variable axis immersion lenses / A High
-Throughput Electron Beam Approach to'Suboptical '
Lithography, Hans C. Pfeiffer, JJAP Vol. 34 (199
5) Through electron beam reduction drawing (electron beam lithography) such as p.6658-6662), it is possible to further increase the throughput.

However, in these electron beam reduction drawing, it is necessary to increase the energy of the electron beam in order to converge the electron beam and form a clear image. The energy of the electron beam in the block exposure and cell projection method is generally 50 keV,
Energy of electron beam is 100k in electron beam reduction drawing
It becomes eV. With such high energy, the mechanism for controlling the electron beam optical system becomes large. Therefore,
The cost of the device also increases.

Further, in the case of a high-energy electron beam, electrons pass through the resist without releasing energy in the resist, so that the resist sensitivity per the number of electrons becomes small. Therefore, when resists having the same sensitivity are used, the higher the electron beam energy, the larger the required electron beam current amount and the higher the electron density in the beam.

When the electron density in the beam becomes higher, the beam is out of focus and the pattern resolution is lowered. Further, in the drawing using the electron beam, there is a problem that the formed pattern is distorted as a result of backscattering from the lower substrate to the resist (proximity effect). The larger the electron beam current amount, the larger the influence of the proximity effect.

Further, the higher the electron beam current amount, the easier the mask, resist and substrate are heated, and the greater the distortion of the formed pattern. Therefore, in order to maintain the precision required for the pattern, it is necessary to limit the amount of electron beam current, which lowers the throughput.

In order to suppress the proximity effect without affecting the throughput, an exposure method has been developed in which a pattern is formed by a low energy electron beam. It has been reported that the proximity effect is substantially reduced by using a low energy electron beam ('Low voltage alternative for e
lectron beam lithography 'J. Vac. Sci. Technol.B 1
0 (6), Nov / Dec (1992) p. 3094-3098).

As a lithography technique using a low energy electron beam, LEEPL (low energy E-beam pr) utilizing the technique disclosed in Japanese Patent No. 2951947 is used.
Oximity projection lithography) is under development. The energy of the electron beam used in LEEPL is about 1 to 4 keV, preferably about 2 keV. LEE
In PL, the mask is placed about 50 μm away from the resist coated wafer.

LEEPL is an equal magnification proximity exposure, and in order to form a fine pattern having a line width of 100 nm or less on a wafer, it is necessary to form a pattern of 100 nm or less on a mask. In the case of lithography using a high-energy electron beam, it is also possible to provide a heavy metal portion for scattering the electron beam on a part of a thin film (membrane) and transfer the pattern by the electron beam passing through the membrane.

However, in the case of LEEPL, since the energy of the electron beam is low and electrons do not pass through the membrane, a stencil mask having holes in the membrane is used. The electron beam transmits only the hole portion, and the pattern is transferred. In order to form a fine pattern of, for example, 100 nm or less on the LEEPL stencil mask with high accuracy, it is preferable that the aspect of the hole is low. Therefore, it is required to make the membrane thin.

For example, in the cell projection system, 50
Electron beam writer (HL900D) using keV electron beam
In the case of, the membrane thickness of the mask is generally 10 μm. On the other hand, in LEEPL, the membrane thickness of the mask is about 1/10 or less and about 500 nm.

[0017]

In a thin mask such as the above-described LEEPL mask, the stress in the membrane changes due to the formation of holes, and the mask is easily deformed. When the mask is deformed, the circuit pattern transferred onto the wafer is deformed or displaced. In particular, when the corners of the opening pattern in the mask are corners, stress concentrates on the corners and the opening pattern is deformed.

As a result, the circuit pattern formed on the wafer is deformed from the original desired pattern, and the performance and reliability of the finally obtained semiconductor device deteriorate. In some cases, the corners of the opening pattern in the mask may crack, rendering the mask unusable. Further, when the opening pattern of the mask is a long line, the pattern may be deformed due to the non-uniformity of stress in the membrane, and the line width may change, for example, near the center of the line.

The present invention has been made in view of the above problems. Therefore, the present invention does not form a corner that causes deformation or cracks of the opening pattern in the opening of the mask, but has a circuit having a corner. An object of the present invention is to provide a mask capable of transferring a pattern onto a wafer and a manufacturing method thereof. Also,
An object of the present invention is to provide an exposure apparatus that can transfer a fine pattern having corners with high accuracy. A further object of the present invention is to provide a method of manufacturing a semiconductor device, which can form a fine pattern having corners with high accuracy.

[0020]

In order to achieve the above-mentioned object, the mask of the present invention has a first mask for blocking a charged particle beam and a charged particle beam formed on the first mask. The first opening having no corners, the second mask used to overlap the first mask and blocking a charged particle beam, and the second mask having the second opening It is characterized in that it has a second opening which is formed so as to have an overlapping portion with the first opening and through which the charged particle beam passes, and which has no corners.

[0021] Preferably, the overlapping portions have corners sandwiched by straight lines. Suitably, the overlapping portion is rectangular. In the first and second masks of the mask of the present invention, a third opening having corners may be formed in addition to the first or second opening. Suitably, the charged particle beam comprises an electron beam.

Thus, a pattern having corners can be formed by lithography without providing a mask with a corner opening. Therefore, local stress concentration at the corners of the opening of the mask is suppressed, and deformation of the opening and generation of cracks at the corners of the opening are prevented.

In order to achieve the above object, the method of manufacturing a mask of the present invention predicts a local stress concentration at a corner of an opening corresponding to a mask pattern, and the stress concentration exceeds the resistance of the mask. And adding an additional pattern that includes curved portions at both ends in one direction of the extracted opening and the shape of the opening has no corners to create a first opening pattern. Step and at both ends of the extracted opening in the other direction,
A step of creating a second opening pattern by adding an additional pattern that includes a curved portion and the shape of the opening has no corners; an opening having the first opening pattern; Forming a first mask having a second opening pattern, an opening having the second opening pattern, and an opening not extracted,
And a step of forming a second mask to be used by being overlapped with the above mask.

Preferably, when the extracted opening has a reentrant shape, the opening is reentrant before the first opening pattern and the second opening pattern are formed. The method further includes the step of dividing into a plurality of openings that are not present. Preferably, the method further comprises the step of dividing the opening having no reentrant angle into a plurality of openings before forming the first opening pattern and the second opening pattern. This makes it possible to form an opening that is difficult to deform in the mask. Also, the corners of the opening can be eliminated,
It is possible to prevent damage to the mask due to stress concentration at the corners.

To achieve the above object, the exposure apparatus of the present invention comprises a charged particle beam generating means, a charged particle beam deflecting means,
A first mask for blocking the charged particle beam; a first opening formed in the first mask for transmitting the charged particle beam; the first opening having no corners; A second mask which is used in combination with the above-mentioned mask to block a charged particle beam, and a second mask which is formed in the second mask so as to have a portion overlapping with the first opening and through which the charged particle beam passes. 2 openings, the second opening having no corners. This makes it possible to expose a cornered pattern using two masks having a cornerless opening.

In order to achieve the above object, the semiconductor device manufacturing method of the present invention comprises a first mask for blocking the charged particle beam except at the opening and a second mask for blocking the charged particle beam at other than the opening. A method of manufacturing a semiconductor device, comprising the step of irradiating a charged particle beam onto a substrate through a mask, the method comprising:
The opening formed in the mask of 1) includes a first opening having no corners, and the opening formed in the second mask has no corners, and the first opening overlaps the first opening. The substrate is irradiated with a charged particle beam including a second opening having a portion and passing through the first overlapping portion.

In the method for manufacturing a semiconductor device of the present invention, preferably, a third mask for blocking the charged particle beam except the opening and a fourth mask for blocking the charged particle beam other than the opening are provided. And irradiating the substrate with a charged particle beam, the opening formed in the third mask includes a third opening without corners, and the opening is formed in the fourth mask. The opening has no corners and includes a fourth opening having a second overlapping portion with the third opening, and the charged particle is transmitted through the second overlapping portion different from the first overlapping portion. A line is applied to the substrate.

As a result, it becomes possible to form a fine pattern with high accuracy without necessarily increasing the number of exposures as compared with the conventional method for manufacturing a semiconductor device. This is because pattern distortion and displacement of the openings of the mask are prevented. Among lithographic masks, a stencil mask having openings has a donut-shaped pattern, for example, and it is not possible to form all the patterns with one continuous mask. Multiple exposure is performed by sequentially using at least two masks (complementary masks) on which different patterns are formed, without using them simultaneously.

According to the method of manufacturing a semiconductor device of the present invention,
The first overlapping portion is exposed by using two sheets of the first and second masks at the same time. If you divide the opening into multiple parts,
Further, the second overlapping portion is exposed by using two sheets of the third and fourth masks at the same time. Therefore, it is not necessary to increase the number of exposures as compared with the conventional case where multiple exposure is performed using a complementary mask.

[0030]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of a mask, a method of manufacturing the same, an exposure apparatus and a method of manufacturing a semiconductor device according to the present invention will be described below with reference to the drawings. Hereinafter, an example in which the present invention is applied to LEEPL, which is one of lithography techniques, will be described.

FIG. 1 is a schematic diagram of an exposure system used for LEEPL. In the exposure system 1, the electron beam 2 is emitted from the electron gun 3. The exposure system 1 has, in addition to the electron gun 3, a condenser lens 4, an aperture 5, a pair of main deflectors 6 and 7, and a pair of fine adjustment deflectors 8 and 9.

The condenser lens 4 makes the electron beam 2 a parallel beam. Aperture 5 is stencil mask 1
Limit the electron beam 2 towards 0. The main deflectors 6 and 7 deflect the electron beam 2 so that the electron beam 2 is incident parallel to the stencil mask 10 while remaining parallel.

The electron beam 2 is incident on the stencil mask 10 in either the raster or vector scanning mode, and in either case, the main deflectors 6 and 7 are used to deflect the electron beam 2. Fine adjustment deflector 8,
Reference numeral 9 further finely adjusts the electron beam 2 deflected by the main deflectors 6 and 7.

The thickness d ₁ of the stencil mask 10 used for LEEPL is about 500 nm, but it can be changed according to the mask material. As the material of the stencil mask 10, for example, diamond or single crystal silicon is used. The distance d ₂ between the stencil mask 10 and the wafer 11 is about 50 μm. The pattern formed on the stencil mask 10 is transferred onto the wafer 11 at the same size.

FIG. 2 is an enlarged sectional view of the stencil mask 10 of FIG. The stencil mask 10 has a first mask 21 arranged on the electron gun side and a second mask 22 arranged on the wafer side. FIG. 2 shows the first mask 21.
An example is shown in which the second mask 22 and the second mask 22 are in contact with each other. When the electron beam parallelism, emittance, and positional accuracy are good, the interval between the first mask 21 and the second mask 22 is set to 1
It is also possible to set the thickness to 00 μm or more and to a maximum of about 300 μm. The first mask 21 and the second mask 22 have openings 23 and 24 in different patterns at substantially the same position.

FIG. 3A is a schematic view of the stencil mask of this embodiment. As shown in FIG. 3A, the wafer 1
A first mask 21 and a second mask 22 are arranged on top of one another. The openings 23 of the first mask 21 and the openings 24 of the second mask 22 partially overlap each other. Figure 3 (b)
Is the first mask 21 and the second mask 22 of FIG.
Also, the distance between the wafers 11 is shown enlarged.
The overlapping portion of the openings 23 and 24 corresponds to the desired rectangular pattern 25 on the wafer 11.

FIGS. 4A and 4B are views showing patterns of openings formed in the stencil mask of FIG. FIG. 4C shows a pattern transferred onto the wafer by combining the patterns of FIGS. 4A and 4B. 3 and 4 show an example of transferring a rectangular pattern onto a wafer.

As shown in FIGS. 3 and 4A, the first
An opening 23 is formed in the mask 21. The opening 23 is formed with additional patterns 26 at both ends in the longitudinal direction with respect to a desired rectangular pattern 25 shown by hatching. The additional pattern 26 has a round shape, and the opening 23 does not have a corner.

On the other hand, as shown in FIGS. 3 and 4 (b), an opening 24 is formed in the second mask 22 and is in a direction orthogonal to the longitudinal direction with respect to a desired rectangular pattern 25 shown by hatching. Additional patterns 27 are formed on both ends.
The additional pattern 27 is also round like the additional pattern 26, and the opening 24 does not have a corner.

The overlapping portion of the opening 23 including the additional pattern 26 and the opening 24 including the additional pattern 27 coincides with the desired rectangular pattern 25, and a corner is formed. A resist is applied on the wafer 11 in FIG. 3 in advance, and a desired rectangular pattern 25 is formed by the electron beam transmitted through both the first mask 21 and the second mask 22 (see FIG. 4C).
Is transcribed.

Next, a mask pattern for forming a pattern other than a rectangle will be described. In the case of a stencil mask, it is not possible to form a topologically donut-shaped pattern 31 as shown in FIG. 5, for example. One stencil mask needs to be all continuous except for the opening.

Therefore, when forming a pattern as shown in FIG. 5, for example, as shown in FIG.
1 is divided into at least two division patterns 32a and 32b. Alternatively, as shown in FIG. 6B, the pattern 31 is divided into at least two division patterns 33a and 33b.

Division patterns 32a, 32b (or 3)
3a, 33b) are formed on different stencil masks and these masks are used as complementary masks. That is, after performing exposure using one stencil mask, multiple exposure is performed using the other stencil mask.

For example, when the pattern 31 (see FIG. 5) is divided as shown in FIG. 6A, L as shown in FIG.
The V-shaped pattern 34 is formed on the mask 35. However, the concave corner portion (the portion surrounded by the dotted line in FIG. 7) cannot be formed by overlapping the openings having no corners as shown in FIGS. 3 and 4. Therefore, when forming the L-shaped pattern 34 as shown in FIG. 7, the pattern 34 is further divided until the concave corner portions are eliminated.

FIG. 8 shows an example of a method of dividing the L-shaped pattern 34. As shown in FIG. 8, by dividing the pattern 34 into, for example, division patterns 36 and 37,
The concave corner can be eliminated. Dividing pattern 36, 3
7 is a rectangular pattern, and as shown in FIGS. 3 and 4, it can be formed by superposing two openings having no corners.

Solid lines in FIGS. 9A and 9B show patterns transferred by exposure. The solid line in FIG. 9A is shown in FIG.
Corresponding to the division pattern 36 of. The solid line in FIG. 9B corresponds to the division pattern 37 in FIG. 10 and 11
In, the dotted line indicates the desired L-shaped pattern 34 (see FIG. 8). The solid line in FIG. 10C is one mask for forming the divided pattern 36 in FIG.
The pattern of the opening formed in the first mask 21) of FIG. The solid line in FIG. 10D shows the pattern of the opening formed in one mask for forming the division pattern 37 in FIG. 9B.

The solid line in FIG. 11E shows the pattern of the opening formed in the mask (for example, the second mask 22 in FIG. 2) used in superposition with the mask in which the pattern in FIG. 10C is formed. Show. The solid line in FIG. 11 (f) is shown in FIG.
The pattern of the opening formed in the mask used by overlapping with the mask in which the pattern of (d) is formed is shown.

The first exposure is carried out by superposing the patterns shown in FIGS. 10C and 11E, and the pattern shown in FIG. 9A is transferred, and then the patterns shown in FIGS. Second exposure is performed by superimposing the pattern shown in f), and the pattern shown in FIG. 9B is transferred. With such double exposure,
The L-shaped pattern 34 is transferred.

The order of the first exposure and the second exposure can be exchanged. Either of the masks on which the patterns of FIG. 10C and FIG. 11E are formed may be the first mask (mask on the electron gun side). Similarly, FIG.
Either the mask in which the patterns of (d) and FIG. 11 (f) are formed may be the first mask.

FIG. 12 shows an L-shaped pattern 3 shown in FIG.
Another example of a method of dividing 4 will be shown. Pattern 34 is 1
It is possible to divide into two rectangular patterns by the separating line 38 of the book. However, when the divided pattern has a line shape that is long in one direction, the stress of the mask differs between the longitudinal direction of the pattern and the direction (width direction) orthogonal to the longitudinal direction. From the stress analysis, it is known that in the linear pattern, the opening easily expands in the width direction near the center in the longitudinal direction. Further, cracks are likely to occur in the longitudinal direction at the corners of the line pattern.

Therefore, even if the divided pattern does not have concave corners, if the pattern becomes a long line pattern in one direction, it is desirable to further divide the pattern so that the aspect ratio of the pattern approaches 1. . Thereby, the stress of the mask can be made more uniform around the divided pattern, and the distortion of the pattern and the damage of the mask at the corners of the pattern can be prevented.

FIG. 13 is a flow chart showing the steps of designing an opening in the mask manufacturing method of this embodiment. Step 1 (ST1): Start designing the opening. For example, as shown in FIG. 5, when one mask has a pattern that is not continuous (topologically donut-shaped pattern), or as shown in FIG. 7, when a pattern includes concave corners, Before step 1, it is divided into rectangular patterns in advance.

Step 2 (ST2): In the circuit pattern of the semiconductor device, a pattern in which stress concentration at corners is a problem is extracted. The extraction of the pattern is performed based on the structural analysis using the finite element method or the like and the knowledge obtained by the preliminary experiment. Even if a corner is formed,
If the pattern is such that stress concentration at the corners does not matter, the steps after step 3 are unnecessary, and the opening can be formed in the mask without changing the pattern.

Step 3 (ST3): With respect to the pattern extracted in step 2, it is judged whether an additional pattern having a prescribed width can be provided at a corner where stress concentration becomes a problem. At this time, it is determined whether an additional pattern having a specified width can be formed in both a specific direction including a corner where stress concentration becomes a problem and a direction different from this direction.

Here, the prescribed width is determined in consideration of the resolution limit of the drawing device (for example, the variable shaping type electron beam drawing device) for forming the opening of the mask, the interval between adjacent patterns, and the like. To be done. Also, in determining the prescribed width, by forming an additional pattern in the subsequent steps,
Consider also whether stress concentration at the corners is relaxed to an acceptable range.

For example, when the rectangular patterns 41 and 42 which are separated from each other are arranged close to each other as shown in FIG. 14A, the respective patterns are defined as shown in FIG. 14B. When the widths 41a and 42a (dotted line portions) are added, an overlapping portion 43 (hatched portion) occurs. In such a case, No is determined in step 3.

Step 4 (ST4): When it is judged in Step 3 that there is no room for providing an additional pattern having a specified width in the portion in contact with the opening, the pattern is formed so that the specified width can be added. Further divide. Figure 1
As shown in FIG. 4, when the specified width cannot be added to the pattern, the rectangular pattern is further divided into two pattern groups as in the case shown in FIG. As a result, the pattern interval is widened, and it becomes possible to add a prescribed width to the pattern.

Step 5 (ST5): Yes in Step 3
A first pattern is created by adding a specified width to both sides of the pattern determined as s or the pattern divided in step 4 in a specific one direction. Further, a second pattern having a specified width is formed on both sides in a direction different from the above direction.

Step 6 (ST6): The corners of the portion added in step 5 are rounded. Step 7 (ST7): The first pattern is used as the opening pattern of the first mask, and the second pattern is used as the opening pattern of the second mask. Step 8 (ST8): The design of the opening ends.

Through the above steps, the opening of the mask of this embodiment is designed. After that, an opening is formed in the first mask with the first pattern, and a second opening is formed with the second pattern. A mask having an opening can be manufactured according to a conventional mask manufacturing method. The stencil mask of this embodiment is formed by combining the first and second masks.

In the method of manufacturing the semiconductor device of this embodiment, lithography is performed using the above stencil mask,
The step of transferring the pattern to the resist on the wafer is included.
As a result, pattern distortion and displacement are prevented, and a fine pattern can be transferred with high accuracy. A fine pattern can be formed on a semiconductor device by using the resist to which the pattern is transferred in a semiconductor device manufacturing process such as etching or ion implantation.

According to the mask and its manufacturing method of the embodiment of the present invention described above, local stress concentration at the corners of the opening of the mask is suppressed, and damage to the mask and distortion of the pattern of the opening are prevented. . The embodiments of the mask, the method for manufacturing the mask, the exposure apparatus, and the method for manufacturing the semiconductor device according to the present invention are not limited to the above description. For example, the mask of the present invention is not limited to the LEEPL mask, and may be a charged particle beam lithography mask other than LEEPL. The present invention can also be applied to, for example, a mask for X-ray lithography in which a heavy metal film having an opening is formed on a thin film. Besides, various modifications can be made without departing from the scope of the present invention.

[0063]

According to the mask of the present invention, a circuit pattern having a corner can be transferred without forming a corner in the opening of the mask which causes deformation or crack of the opening pattern. According to the mask manufacturing method of the present invention, it is not necessary to form a corner in an opening for forming a circuit pattern having a corner, and deformation of the opening pattern and damage to the mask are prevented. According to the exposure apparatus of the present invention, it is possible to expose a circuit pattern having a corner by using a mask having no corner in the opening pattern. According to the method for manufacturing a semiconductor device of the present invention, a fine pattern having corners can be formed with high accuracy.

[Brief description of drawings]

FIG. 1 is a schematic view of an exposure apparatus of the present invention including a mask of the present invention.

FIG. 2 is a sectional view of a mask of the present invention.

FIG. 3 is a schematic diagram of a mask of the present invention.

4 (a) and 4 (b) show patterns formed on the mask of the present invention, and FIG. 4 (c) shows patterns transferred by the mask of the present invention.

FIG. 5 is a diagram schematically showing an example of a circuit pattern.

6A and 6B show examples of dividing the pattern of FIG.

FIG. 7 is a diagram schematically showing an example of a circuit pattern.

FIG. 8 is a diagram showing an example of a pattern formed on the mask of the present invention.

FIG. 9A and FIG. 9B show examples of patterns transferred by the mask of the present invention.

10 (c) shows a pattern formed on a mask for transferring the pattern of FIG. 9 (a), and FIG.
9D shows a pattern formed on a mask for transferring the pattern of FIG. 9B.

11 (e) shows a pattern formed on a mask for transferring the pattern of FIG. 9 (a), and FIG.
9F shows a pattern formed on a mask for transferring the pattern of FIG. 9B.

FIG. 12 is a diagram showing an example of a pattern formed on the mask of the present invention.

FIG. 13 is a flowchart showing a process of designing an opening of a mask of the present invention.

FIG. 14 shows an example in which the pattern needs to be divided in ST4 of FIG.

[Explanation of symbols]

1 ... Exposure system, 2 ... Electron beam, 3 ... Electron gun, 4 ...
Condenser lens, 5 ... Aperture, 6, 7 ... Main deflector, 8, 9 ... Fine adjustment deflector, 10 ...
Stencil mask, 11 ... Wafer, 21 ... First mask,
22 ... 2nd mask, 23, 24 ... Opening part, 25 ... Desired rectangular pattern, 26, 27 ... Additional pattern, 31, 3
4 ... pattern, 32a, 32b, 33a, 33b ... division pattern, 35 ... mask, 36, 37 ... division pattern,
38. Separation line.

Claims (14)

[Claims] 1. A first mask for blocking a charged particle beam, and a first opening for transmitting a charged particle beam formed in the first mask, the first opening having no corners. A second mask that is used to overlap with the first mask to block a charged particle beam; and charged particles formed in the second mask so as to have a portion overlapping the first opening. A mask having a second opening through which a line is transmitted, the second opening having no corners. 2. The shape of the first opening includes a straight line and a curved line, the shape of the second opening includes a straight line and a curved line, and the overlapping portion has an angle between the straight lines. The listed mask. 3. The overlapping portion has a rectangular shape, the first opening portion has a shape in which a portion including a curved line is added outside a pair of opposite sides of the rectangle, and the second opening portion has the shape. The mask according to claim 2, which has a shape in which a portion including a curved line is added to the outside of another pair of opposite sides of the rectangle. 4. The mask according to claim 1, wherein the first mask is a third opening through which a charged particle beam passes, and further has the third opening having a corner. 5. The mask according to claim 1, wherein the second mask has a fourth opening through which a charged particle beam is transmitted, the fourth opening having an angle. 6. The mask according to claim 1, wherein the charged particle beam includes an electron beam. 7. A step of predicting a local stress concentration at a corner of an opening corresponding to a mask pattern and extracting the opening whose stress concentration exceeds the resistance of the mask, and one direction of the extracted opening. A step of creating a first opening pattern by adding an additional pattern including curved portions to both ends of the opening and eliminating the corners of the shape of the opening, and both ends of the extracted opening in the other direction. A step of creating a second opening pattern by adding an additional pattern including a curved portion and having no corners in the shape of the opening, and an opening having the first opening pattern; Forming a first mask having unopened openings, opening having the second opening pattern, and unextracted openings, overlapping the first mask Used And a step of forming a second mask, the method for manufacturing a mask. 8. When the extracted opening has a reentrant shape, a plurality of the openings having no reentrant angle are formed before the first opening pattern and the second opening pattern are created. 8. The method according to claim 7, further comprising the step of dividing the openings into
A method for manufacturing the described mask. 9. The first opening pattern and the second opening pattern.
8. The method for manufacturing a mask according to claim 7, further comprising the step of dividing the opening having no reentrant angle into a plurality of openings before forming the opening pattern of. 10. A charged particle beam generating means, a charged particle beam deflecting means, a first mask for blocking the charged particle beam, and a first opening for transmitting the charged particle beam formed in the first mask. A first opening having no corners, a second mask used to overlap with the first mask to block charged particle beams, and the first opening to the second mask. An exposure apparatus having a second opening which is formed so as to have an overlapping portion and through which the charged particle beam is transmitted and which has no corners. 11. A first means for blocking a charged particle beam except at an opening.
And a second mask that shields the charged particle beam at a portion other than the opening, the method including the steps of: irradiating the substrate with the charged particle beam; The formed opening has a first corner
The opening formed in the second mask has no corners and includes a second opening having a first overlapping portion with the first opening. A method for manufacturing a semiconductor device, wherein the substrate is irradiated with a charged particle beam that has passed through an overlapping portion. 12. The shape of the first opening includes straight lines and curved lines, the shape of the second opening includes straight lines and curved lines, and the charged particle beam has corners sandwiched by straight lines. First
12. The method of manufacturing a semiconductor device according to claim 11, wherein the overlapping portion of the light is transmitted. 13. A third means for blocking a charged particle beam except at the opening.
The method further comprises the step of irradiating the substrate with the charged particle beam through the mask of No. 4 and a fourth mask that blocks the charged particle beam except at the opening, and the opening formed in the third mask is , Third without horns
The opening formed in the fourth mask has no corners and includes a fourth opening having a second overlapping portion with the third opening, and the first opening The charged particle beam that has passed through the second overlapping portion different from the overlapping portion is irradiated onto the substrate.
A method for manufacturing a semiconductor device as described above. 14. The method of manufacturing a semiconductor device according to claim 11, wherein the charged particle beam includes an electron beam.