JP2003115429A - How to determine the reticle pattern - Google Patents

Abstract

PROBLEM TO BE SOLVED: To find a pattern part causing a problem in dimensional precision of a transcribed pattern caused by distortion of a stencil reticle, and to form this pattern pat into a pattern partitioned into a plurality of reticles. SOLUTION: A numeral 11 is a part where charged particle beam is shielded and corresponds to the part without a hole in the reticle. A numeral 12 is a part where the charged particle beam is not shielded and corresponds to the part with a hole. In a shielding part 13, whose three sides are surrounded by the non-shielding part 12, the largest dimensional error is caused at corners 14 (vertexes A, B) of the top part. Therefore, in the case that the dimensional error at the corners 14 of the top part exceeds an allowable value, the pattern is complementarily partitioned, thus exposure is performed using a plurality of reticles.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention determines a reticle pattern for forming a target pattern on a wafer when a pattern formed on the reticle is transferred to a wafer by using a charged particle beam exposure apparatus. It is about the method.

[0002]

2. Description of the Related Art As the degree of integration of semiconductor devices increases, it becomes impossible to cope with the conventional optical exposure and transfer apparatus, and as an alternative, charging using charged particle beams such as electron beams is used. A particle beam exposure apparatus has been developed.

In the charged particle beam exposure apparatus, it is not possible to expose and transfer a wide area at once due to aberrations and distortions of the charged particle beam optical system. Therefore, for example, the corresponding area of one chip is divided into a plurality of areas called subfields, exposure transfer is performed for each subfield, and the patterns transferred by exposure are connected to obtain the pattern of one chip. The division exposure transfer system has been adopted.

This division projection transfer type exposure apparatus will be described with reference to FIGS. FIG. 9 is a diagram showing a unit of divided exposure. First, a plurality of chips are formed on a transfer body (usually a wafer), and the chips are further divided into stripes, and the stripes are divided into subfields. The transferred material such as a reticle is also divided.

In a charged particle beam exposure apparatus using a divided exposure transfer system, exposure is usually performed by the method shown in FIG. First, the reticle stage and the wafer stage move the centers of the corresponding stripes at a constant speed at a speed according to the reduction ratio. The electron beam illuminates the subfield on the reticle,
The pattern formed on the reticle is projected and exposed on the sample by the projection optical system.

Then, the electron beam is deflected in a direction substantially perpendicular to the traveling direction of the reticle stage, and projection exposure of the subfields arranged in a line is sequentially performed. When the projection exposure of one row of subfields is completed, the projection exposure of the next row of subfields is started. At this time, the electron beam deflection direction is reversed as shown in FIG. Through this, the throughput is increased.

Since the exposure is performed by such a method, compared with the conventional charged particle beam exposure apparatus, the subfield area is collectively exposed, and the reticle has all the patterns to be exposed. Throughput can be improved.

The reticle used in this exposure method is divided into a subfield portion (pattern portion) and a beam portion (hereinafter referred to as a strut) around the subfield portion (pattern portion), which is different from the case of the exposure apparatus using light. The beam is provided to maintain the strength of the reticle itself, and the skirt that can block the electron beam between the strut and the subfield selects only the subfield that the illumination beam should expose. It is provided for the purpose of.

When a pattern such as an LSI is formed on a wafer by using the charged particle beam exposure apparatus using such a divided exposure transfer system, first, the pattern to be formed on the wafer is determined. Then, after that, a reticle pattern for forming the pattern on the wafer is determined.

Reticles are roughly classified into two types. One is a stencil reticle in which the presence or absence of a pattern is the presence or absence of a substance that reduces the transmittance of a charged particle beam on the reticle, and the other is a membrane reticle in which the transmittance of a charged particle beam is high or low. (This transmittance also considers the degree of scattering, and low transmittance does not only mean that the absorption of the charged particle beam is high, but also that the charged particle beam that penetrates the reticle is scattered. This means that the reticle does not reach the wafer by some means.) Stencil reticle does not have these problems in contrast to the membrane reticle that has problems such as heat absorption and chromatic aberration, and is considered to be a promising reticle for high resolution. Has been.

In the stencil reticle, since there is no thin film for supporting the hollow portion in the annular pattern, it is necessary to complementarily divide the pattern into two and realize this by two exposures. Further, a weak structure having a problem in strength of the reticle and a structure having a problematic dimensional error due to distortion of the reticle also need to be similarly divided.

The dimensional error due to the distortion of the reticle occurs when the opening is large in a reticle stretched like a rubber film. That is, in the reticle, it is necessary to apply tension to the whole of the reticle in order to prevent it from slackening. Therefore, if there is a large opening, the shape of the opening will be deformed due to this tension, and There is a problem that it will not fit.

[0013]

Regarding an annular pattern, it is possible to easily recognize that the pattern is annular and requires division by tracing the pattern contour, but a weak structure having a problem in strength. For the part of the pattern that has, or the part of the pattern where the dimensional error due to the distortion of the reticle is a problem, what kind of pattern can be the problem, and what part of the pattern and what size I didn't really know if I had to split.

The present invention has been made in view of the above circumstances, and the design value and the actually manufactured dimension are different due to the distortion of the stencil reticle, which causes a problem in the dimensional accuracy of the transferred pattern. An object of the present invention is to provide a method for determining a reticle pattern including a step of finding a portion and determining it as a pattern divided into a plurality of reticles.

[0015]

A first means for solving the above-mentioned problems is to change the shape of a pattern to be transferred onto a wafer from the shape of a pattern to be transferred onto a wafer in the manufacturing process of a stencil reticle used in a charged particle beam exposure apparatus. A method for determining whether or not a pattern to be formed on a reticle for transfer needs to be divided into two or more reticles, and a shape in which an exposed portion is surrounded on three sides in an unexposed portion on a wafer. When there is a part having a dimensional error E, the dimensional error E of the reticle pattern with respect to the design pattern is calculated from the length L and the width W of this part and the length S of the side to which this part is connected.
In the case of the reticle pattern determining method (claim 1), it is determined that the division is necessary when the number of times exceeds.

As a result of a study by the inventor of the change in the pattern shape caused by the tension applied to the stencil reticle (hereinafter sometimes simply referred to as a reticle), the non-exposed portion on the wafer has three sides in the exposed portion. When there is a part with an enclosed shape, that is, in the reticle,
It has been found that, when a portion (projection portion) surrounded by three sides is formed in the holed portion, the deformation of the pattern due to the tension is particularly large at the tip portion of the projection portion. Further, it has been found that the amount of deformation has a relationship with the length L and width W of this portion and the length S of the side to which this portion connects.

Therefore, when the shape of the pattern to be transferred to the wafer is determined, it is judged whether or not there is a portion having a shape surrounded by three sides in the exposed portion in the unexposed portion on the wafer, When such a portion is found, a dimensional error E with respect to the design pattern of the reticle pattern is obtained from the length L and the width W of this portion and the length S of the side to which this portion is connected by a previously obtained method. By calculating and determining that the pattern needs to be divided when the value exceeds the allowable value T, the accuracy of the pattern formed on the wafer can be kept within the allowable limit.

Needless to say, the step of dividing the pattern into subfields is carried out even if the non-exposed portion on the wafer has a portion surrounded by the exposed portion on three sides. After that, when the pattern in the sub-field is no longer surrounded by the exposed portion on three sides, it is no longer necessary to divide the pattern. Therefore, in such a case, it is judged that it is not necessary to divide by this means.

A second means for solving the above problems is
The reticle pattern determining method (claim 2) is the first means, wherein the error dimension E is calculated by the following equation (1). However, k ₁ , k
₂ and k ₃ are predetermined constants. E = {(k ₁ · L + k ₂ · S) ² + (k ₃ W / 2) ² } ^1/2 (1) The deformation amounts described in the first means are L, S, and W. Value, the material of the reticle, the tension to be applied, the size of the reticle, the exposure reduction ratio, and the reticle manufacturing process. In the present specification, the "reticle size" means the outer diameter size of the reticle (whether it is made from a 300 mm wafer or a 200 mm wafer, etc.), subfield size, membrane size, arrangement of membrane and subfield, It includes the width, number, and arrangement of struts. When the inventor analyzes these relationships by using the finite element method and the regression analysis while changing these conditions and using the dimensional error E with respect to the design pattern of the reticle pattern, these relationships are as simple as the equation (1). We found that it can be approximated by a relational expression.

The constants k ₁ , k ₂ and k ₃ are determined by the material of the reticle, the tension to be applied, the size of the reticle, the exposure reduction magnification, and the reticle manufacturing process. It can be used as a constant value regardless of the shape of the pattern.

A third means for solving the above problems is
In the manufacturing process of a stencil reticle used in a charged particle beam exposure apparatus, it is necessary to divide a pattern to be transferred onto a wafer from a pattern to be transferred onto a reticle into two or more reticles. A method for determining presence / absence of a reticle pattern, wherein when a non-exposed portion on a wafer has a portion having a shape in which the exposed portion is surrounded on two sides, the reticle pattern is determined from the lengths L1 and L2 of the boundaries of the two sides. The reticle pattern determining method (claim 3) is characterized in that the dimensional error E for the design pattern is calculated, and when the dimensional error E exceeds the allowable value T, it is determined that division is necessary.

As a result of further examination by the inventor of the change in the pattern shape caused by the tension applied to the reticle, when there is a portion of the non-exposed portion on the wafer that has a shape surrounded by two sides in the exposed portion. That is, it has been found that, even when a part (projection part) surrounded by two sides is formed in a holed part of the reticle, the deformation of the pattern due to the tension is particularly large at the tip part of the projection part. .
Further, it was also found that the amount of deformation is related to the lengths L1 and L2 of the two boundaries.

Therefore, when the shape of the pattern to be transferred to the wafer is determined, it is determined whether or not the unexposed portion on the wafer has a portion having a shape in which the exposed portion is surrounded on two sides. When such a portion is found, the dimensional error E of the reticle pattern with respect to the design pattern is calculated from the lengths L1 and L2 of the two boundaries by a previously obtained method, and the calculated value is the allowable value T. When it exceeds, it is possible to keep the accuracy of the pattern formed on the wafer within the allowable limit by determining that the pattern needs to be divided.

Needless to say, the step of dividing the pattern into sub-fields is performed even if the non-exposed portion on the wafer has a portion surrounded by two sides of the exposed portion. After that, when the pattern in the sub-field is no longer surrounded by the exposed portion on two sides, it is no longer necessary to divide the pattern. Therefore, in such a case, it is judged that it is not necessary to divide by this means.

A fourth means for solving the above-mentioned problems is as follows.
The third means is a reticle pattern determination method (claim 4), wherein the error dimension E is calculated by the following equation (2). However, k ₄ , k
₅ is a predetermined constant. E = k ₄ · L1 + k ₅ · L2 (2)

The amount of deformation described in the first means is determined by the lengths L1 and L2 of the boundary, the material of the reticle, the tension to be applied, the size of the reticle, the exposure reduction magnification, and the reticle manufacturing process. When the inventor analyzes these relations using the finite element method and the regression analysis by changing these conditions and using the dimensional error E with respect to the design pattern of the reticle pattern, these relations are simple as shown in equation (2). We found that it can be approximated by a relational expression.

The constants k ₄ and k ₅ are determined by the material of the reticle, the tension to be applied, the size of the reticle, the exposure reduction magnification, and the reticle manufacturing process. It can be used as a constant value regardless of the shape.

[0028]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, an example of an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing an LSI pattern for explaining an example of the embodiment of the present invention. In the figure, 11 is a portion where the charged particle beam is shielded (referred to as a shielded portion), which corresponds to a portion where no holes are formed in the reticle. Reference numeral 12 denotes a portion where the charged particle beam is not shielded (referred to as a non-shielded portion) and corresponds to a portion where a hole is formed in the reticle. (In addition, shielding means considering the degree of scattering, and not only means that the charged particle beam is absorbed, but also that the charged particle beam that passes through the reticle does not reach the wafer by some means such as scattering. Is also included.)

The shielded portion 13 has a shape in which the non-shielded portion 12 is surrounded on three sides, and a corner 14 (apex A,
The largest dimensional error occurs in B). In Figure 2,
An example of an exposure pattern that is actually formed when a reticle is manufactured with design values as shown in FIG. 1 is shown by a broken line. The solid line is the design pattern shown in FIG. The corner 14 of the tip actually moves to the corner 14 ', during which a position error (dimension error) of E occurs.

As described above, according to the analysis by the inventors,
This dimensional error E is E = {(k ₁ · L + k ₂ · S) ² + from the length L and width W of the shielded portion 13 and the length S of the side connecting this portion in FIGS. It can be approximated by (k ₃ W / 2) ² } ^1/2 (1). However, k ₁ , k ₂ , and k ₃ are predetermined constants, and the material of the reticle, the tension to be applied,
It is determined by the size of the reticle and the exposure reduction ratio, and in particular, it has a high correlation with the Young's modulus of the reticle and the magnitude of the applied stress.

By dividing the pattern when the calculated dimensional error E exceeds the allowable value T, the dimensional error of the completed reticle pattern can be made lower than the allowable limit. Although there are various methods for dividing, an auxiliary line for cutting is tried from the corner 14 of the protruding shielding portion 13 horizontally and vertically, and a fine pattern which causes a problem in reticle production in the pattern after dividing. Since a method of determining the cutting direction under such conditions that a pattern does not occur or the aspect ratio of the pattern after division does not become extreme is widely known, these known methods (for example, USP 5,1
66,888).

An example of the divided pattern is shown in FIG. That is, when the pattern shown in FIG. 1 requires division, it is divided into two complementary patterns as shown in FIGS. 3 (a) and 3 (b). In FIG. 3, reference numeral 21 denotes a shield portion, 2
2 is a non-shielding portion, 23 is a divided pattern, and 24 shown by a dotted line is a pattern that is actually formed.

FIG. 3 is a diagram showing an LSI pattern for explaining another example of the embodiment of the present invention. In the figure, reference numeral 11 denotes a shielding portion, which corresponds to a portion where no holes are formed on the reticle. Reference numeral 12 is a non-shielding portion, which corresponds to a portion having a hole in the reticle. Shield part 15
Has a shape surrounded on two sides by the non-shielding portion 12, and the largest dimensional error occurs at the corner 16 (vertex B) of the tip portion thereof.

In FIG. 5, an example of an exposure pattern actually formed when the reticle is manufactured with the design values as shown in FIG. 4 is shown by a broken line. The solid line is the design pattern shown in FIG.
Corner 16 actually moves to corner 16 ', during which a position error (dimension error) of E occurs.

As described above, according to the analysis by the inventors,
The dimensional error E is 4, can be approximated by the length L1, L2 from _{E = k 4 · L1 + k} 5 · L2 ... (2) shielding portion 15 in FIG. However, k ₄ and k ₅ are predetermined constants and are determined by the material of the reticle, the tension to be applied, the size of the reticle, and the exposure reduction ratio, and in particular, the Young's modulus of the reticle and the magnitude of the applied stress are highly correlated. .

By dividing the pattern when the calculated dimensional error E exceeds the allowable value T, the dimensional error of the completed reticle pattern can be made lower than the allowable limit. The specific division method is the same as the method in the embodiment shown in FIGS.

An example of the divided pattern is shown in FIG. That is, when the pattern shown in FIG. 3 requires division, it is divided into two complementary patterns as shown in FIGS. 6 (a) and 6 (b). In FIG. 6, 31 is a shielding part, 3
Reference numeral 2 is a non-shielding portion, 33 is a divided pattern, and 34 shown by a dotted line is a pattern to be actually formed.

FIG. 7 shows a flowchart for carrying out the first embodiment of the present invention. First, step S1
In 1, a search is made as to whether or not the unshielded portion has a shape in which the shielded portion is surrounded on three sides. First, it is determined whether or not the shape of the non-shielding portion has a recess. Normally, the non-shielding portion is composed of a polygon formed by combining vertical and horizontal lines, but when the polygon is a convex polygon, such a portion does not exist.

When there is a recess having the shape of the non-shielding portion, it is determined whether the number of vertices forming the recess is two or one. When the recess has two vertices (Fig. 1
In, two vertices A and B form a recess),
The non-shielding portion is surrounded by the shielding portion on three sides by the side connecting the one vertex and the other side of the two vertices.

When such a shape is found, in step S12, the length L, the width W, and the length S of the connecting side of the shape are measured. First, the width W is defined as the distance between the two vertices forming the recess. And the length L of the shape
Is the length of the other side of the two vertices (A in FIG. 1).
Length of D and BC). If the two sides have different lengths, take the average value. The length S is the length of the side (FG in FIG. 1) to which the shielded portion (rectangle ABCD in FIG. 1) surrounded on three sides by the non-shielded portion is connected.

Then, in step S13, (1)
The dimensional error E caused by the distortion of the reticle is obtained using the formula, and it is determined in step S14 whether or not this exceeds the allowable value T. If it exceeds, step S
At 15, the pattern is divided.

FIG. 8 shows a flowchart for carrying out the second embodiment of the present invention. First, step S2
In 1, a search is made as to whether or not the unshielded portion has a shape in which the shielded portion is surrounded on two sides. This is performed by the same method as described in FIG. 7, and it is determined whether or not there is a recess in the shape of the non-shielding portion. When there is a concave portion and the number of vertices forming the concave portion is one (in FIG. 4, one vertex B constitutes the concave portion), both sides of the one vertex (AB in FIG. 4, BC) has a shape in which the non-shielding portion is surrounded on two sides by the shielding portion.

When such a shape is found, the lengths L1 and L2 (lengths of sides BC and AB in FIG. 4) of the two sides where the non-shielding portion surrounds the shielding portion are measured in step S22.

Then, in step S23, (2)
The dimensional error E caused by the distortion of the reticle is obtained using the formula, and it is determined in step S24 whether or not this exceeds the allowable value T. If it exceeds, step S
At 25, the pattern is divided.

[0046]

As described above, according to the present invention,
It is possible to find a pattern part that causes a problem in the dimensional accuracy of the transferred pattern due to the difference between the design value and the actually manufactured size due to the distortion of the stencil reticle, and make it a pattern divided into multiple reticles. Since the reticle pattern determining method including the determining step can be provided, the accuracy of the shape of the exposure pattern can be improved and a semiconductor device or the like having a high pattern density can be manufactured with high yield.

[Brief description of drawings]

FIG. 1 is an L diagram for explaining an example of an embodiment of the present invention.
It is a figure which shows the pattern of SI.

FIG. 2 is a diagram showing an example of an exposure pattern actually formed by the design pattern as shown in FIG.

FIG. 3 is a diagram showing an example of a pattern obtained by dividing the pattern shown in FIG.

FIG. 4 is a diagram showing an LSI pattern for explaining another example of the embodiment of the present invention.

5 is a diagram showing an example of an exposure pattern actually formed by the design pattern as shown in FIG.

6 is a diagram showing an example of a pattern obtained by dividing the pattern shown in FIG.

FIG. 7 is a flowchart for implementing the first embodiment of the present invention.

FIG. 8 is a flowchart for carrying out the first embodiment of the present invention.

FIG. 9 is a diagram showing a unit of divided exposure.

FIG. 10 is a diagram showing exposure by a divided exposure transfer system.

[Explanation of symbols]

Reference numeral 11 ... Shielding portion, 12 ... Non-shielding portion, 13 ... Shielding portion surrounded by non-shielding portions on three sides, 14 ... Corner of tip, 14 '
… Moved tip corners, 15 ... shielded parts surrounded by two non-shielded parts, 16 ... tip corners, 16 '... moved tip corners, 21 ... shielded parts, 22 ... non-shielded parts , 23 ...
Divided pattern, 24 ... Actually formed pattern, 31 ... Shielded portion, 32 ... Non-shielded portion, 33 ... Divided pattern, 34 ... Actually formed pattern

Claims (4)

[Claims] 1. A process for manufacturing a stencil reticle used in a charged particle beam exposure apparatus, wherein a pattern formed on a reticle for transferring the pattern from a pattern shape to be transferred onto a wafer is formed on two or more reticles. A method for determining whether or not there is a need for division, wherein when a non-exposed portion on a wafer has a portion surrounded by three sides, the length L and width W of this portion are defined. Also, the dimensional error E of the reticle pattern with respect to the design pattern is calculated from the length S of the side to which this portion is connected, and this is the allowable value T.
A method for determining a reticle pattern, which is characterized in that it is determined that division is necessary when the value exceeds. 2. The reticle pattern determining method according to claim 1, wherein the error dimension E is calculated by the following equation (1). However, k ₁ , k ₂ and k ₃ are predetermined constants. E = {(k ₁ · L + k ₂ · S) ² + (k ₃ W / 2) ² } ^1/2 (1) 3. A stencil reticle manufacturing process used in a charged particle beam exposure apparatus, wherein a pattern to be transferred onto a wafer is formed on a reticle in order to transfer the shape from two or more reticles. A method for determining whether or not there is a need for division, wherein when there is a portion of the non-exposed portion on the wafer that has a shape in which the exposed portion is surrounded on two sides, the length L1 of the boundary between the two sides, From L2,
A method for determining a reticle pattern, which comprises calculating a dimensional error E of a reticle pattern with respect to a design pattern, and determining that division is necessary when the dimensional error E exceeds an allowable value T. 4. The reticle pattern determining method according to claim 3, wherein the error dimension E is calculated by the following equation (2). However, k ₄ and k ₅ are predetermined constants. E = k ₄ · L1 + k ₅ · L2 (2)