JP4834784B2 - Manufacturing method of semiconductor device - Google Patents

Description

  The present invention relates to a method of manufacturing a semiconductor device, and is used, for example, for forming a bit line contact connected to a diffusion layer of a select transistor of a NAND cell unit.

  In a semiconductor device, it is important to form a pattern at a high density for high integration. Therefore, for example, in a NAND flash memory, it has been proposed to dispose a plurality of contact holes for bit line contact with each other (see, for example, Patent Document 1).

  However, in the mask pattern for forming the contact hole, the opening pattern is “dense” in an oblique direction. This is because the openings (transparent regions) for forming contact holes are shifted from each other. Therefore, the exposure margin and the depth of focus are reduced, and it is difficult to suppress dimensional errors in the exposure process. In other words, in a NAND flash memory, although it is necessary to form a dense hole pattern in which opening patterns are regularly arranged and the holes are not arranged in an orthogonal lattice pattern, it is conventionally formed with high accuracy. There was a problem that was difficult.

Japanese Patent No. 3441140

  The present invention provides a method for manufacturing a semiconductor device, which is suitable for forming a fine hole pattern, and in particular, can form a fine pattern which is a dense hole pattern and the holes are not arranged in an orthogonal lattice pattern with high accuracy. The purpose is that.

  According to one aspect of the present invention, illumination light from an illumination light source is applied to a photomask including a mask pattern composed of a transparent region and a non-transparent region, and the light from the photomask is projected via a projection optical system. A method of manufacturing a semiconductor device, wherein a diffracted light is projected onto a substrate to form a photoresist pattern corresponding to the mask pattern on the substrate, wherein the mask pattern is fixed in a first direction. A plurality of parallel lines extending in a second direction orthogonal to the first direction and having a center that is constant in the second direction for each of the parallel lines. The plurality of opening patterns that include the plurality of opening patterns that are the transparent regions arranged at intervals and that are arranged on adjacent parallel lines among the plurality of parallel lines, each center is the second pattern. The illumination light source includes three diffracted lights out of the diffracted light from the photomask in the projection optical system. The illumination shape is set so as to pass through the pupils, and the photomask has dimensions of the plurality of aperture patterns and a complex amplitude transmittance of the non-transparent region so that the amplitudes of the three diffracted lights are equal. Is set, a method for manufacturing a semiconductor device is provided.

  According to another aspect of the present invention, illumination light from an illumination light source is applied to a photomask including a mask pattern composed of a transparent region and a non-transparent region, and the photomask is projected via a projection optical system. A method of manufacturing a semiconductor device that forms a photoresist pattern corresponding to the mask pattern on the substrate by projecting diffracted light from the substrate onto the substrate, wherein the mask pattern is constant in a first direction. A plurality of parallel lines having a first interval and extending in a second direction orthogonal to the first direction, each center being constant in the second direction for each of the parallel lines. Each of the plurality of opening patterns arranged on adjacent parallel lines among the plurality of parallel lines has a center at the front. Each of the mask patterns is shifted in the second direction by 1/3 of the second interval, and the illumination light source includes three diffracted lights out of the diffracted lights from the photomask. There is provided a method of manufacturing a semiconductor device, wherein an illumination shape is set so as to pass through a pupil of a projection optical system.

  The above configuration provides a method for manufacturing a semiconductor device that is suitable for forming a fine hole pattern, and in particular, can form a fine pattern that is a dense hole pattern and the holes are not arranged in an orthogonal lattice pattern with high accuracy. it can.

It is a top view which shows an example of the photomask according to the 1st Embodiment of this invention. It is a figure which shows the structural example of the dipole illumination according to 1st Embodiment. It is a top view which shows an example of the contact hole pattern formed in the photoresist according to 1st Embodiment. It is a figure shown in order to demonstrate the (sigma) coordinate system of illumination. It is a figure shown in order to demonstrate the case where an image is not formed using vertical illumination light. It is a figure shown in order to demonstrate the case where an image is formed using diagonal illumination light. It is a top view of a mask pattern shown in order to explain why a dipole illumination is desirable. It is a figure which shows the structural example of the small-sigma illumination which irradiates vertical illumination light, in order to demonstrate the reason for which dipole illumination is desirable. It is a figure which shows distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating with vertical illumination light, in order to explain the reason why dipole illumination is desirable. It is a figure which shows the structural example of the oblique illumination which irradiates oblique illumination light, in order to demonstrate the reason for which dipole illumination is desirable. It is a figure which shows distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to explain the reason why dipole illumination is desirable. In order to explain why dipole illumination is desirable, it is a diagram illustrating an example in which an image is formed on a substrate due to interference of three diffracted lights. FIG. 8 is a diagram illustrating, as an example, a case where a contact hole pattern is formed in a photoresist using the mask pattern of FIG. 7 in order to explain why dipole illumination is desirable. It is a figure shown in order to demonstrate the relationship between the position and intensity | strength of the diffracted light in a projection lens pupil about the reason why dipole illumination is desirable. It is a figure shown in order to demonstrate optimization of an interference wave amplitude. It is a figure which shows the relationship between a mask bias and the complex amplitude transmittance | permeability of a halftone phase shift mask. It is a figure which shows the relationship between a mask bias and a diffracted light amplitude. It is a flowchart for calculating | requiring the diffracted light intensity which shows the case where the Kirchhoff approximation model is not materialized as an example. It is a figure which shows the structural example of quadrupole illumination. It is a figure which shows distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to explain the reason why quadrupole illumination is desirable. It is a figure which compares and shows the exposure margin by quadrupole illumination and the exposure margin by dipole illumination. It is a figure which shows as an example the case where a contact hole pattern is formed in a photoresist, in order to explain the reason why quadrupole illumination is desirable. It is a figure which shows in the example the case where an isolated contact hole pattern is formed in a photoresist, in order to explain the reason why quadrupole illumination is desirable. It is a figure which shows the structural example of hexapole illumination. It is a figure which shows distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to explain the reason why hexapole illumination is desirable. It is a top view which shows an example of the photomask according to the 2nd Embodiment of this invention. FIG. 3 is a diagram showing a triple staggered arrangement of contact holes for bit line contacts, taking a NAND flash memory as an example. It is a figure which shows the structural example of deformation | transformation dipole illumination according to 2nd Embodiment. It is a top view which shows an example of the contact hole pattern formed in the photoresist according to 2nd Embodiment. It is a top view of a mask pattern shown in order to explain why a modified dipole illumination is desirable. It is a figure which shows distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating with vertical illumination light, in order to explain why the modified dipole illumination is desirable. It is a figure which shows the distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to explain the reason why modified dipole illumination is desirable. In order to explain why the modified dipole illumination is desirable, it is a diagram illustrating an example in which an image is formed on a substrate by interference of three diffracted lights. It is a figure shown in order to demonstrate the relationship between the center of a projection lens pupil, and the position of diffracted light about the reason for which modification | reformation dipole illumination is desirable. It is a figure which shows the wave number vector of the diffracted light with respect to the center of a projection lens pupil, in order to demonstrate the reason why modified dipole illumination is desirable. It is a figure shown in order to demonstrate the light intensity of a bright part and a dark part about the reason for which modification | reformation dipole illumination is desirable. It is a figure shown in order to demonstrate the relationship between a diffracted light amplitude and contrast. It is a figure shown in order to demonstrate the relationship between (epsilon) and (DELTA). It is a figure shown in order to demonstrate the relationship between (epsilon) and the diffraction light amplitude A. FIG. It is a figure which shows the structural example of deformation | transformation quadrupole illumination. It is a figure which shows the distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to explain the reason why modified quadrupole illumination is desirable. It is a figure which shows the other distribution of the diffracted light in the surface equivalent to a projection lens pupil at the time of irradiating oblique illumination light, in order to demonstrate the reason for which modified quadrupole illumination is desirable. It is a figure which shows the structural example of deformation | transformation hexapole illumination.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, it should be noted that the drawings are schematic, and the dimensions and ratios of the drawings are different from the actual ones. Moreover, it is a matter of course that the drawings include portions having different dimensional relationships and / or ratios. In particular, some embodiments described below exemplify apparatuses and methods for embodying the technical idea of the present invention, and the technology of the present invention depends on the shape, structure, arrangement, etc. of components. The idea is not specified. Various changes can be made to the technical idea of the present invention without departing from the gist thereof.

[First Embodiment]
FIG. 1 shows an example of a photomask according to the first embodiment of the present invention. In the present embodiment, a contact hole for bit line contact of a NAND flash memory (a so-called dense hole pattern in which holes are not arranged in an orthogonal lattice pattern, for example, a double staggered hole in a NAND-CB layer) ) Will be described as an example.

  In FIG. 1, the photomask includes a main opening (first main opening) 11, a main opening (second main opening) 12, an assist opening (first assist opening) 21, and an assist opening (second assist opening). 22, a plurality of assist openings (third assist openings) 23 and a plurality of assist openings (fourth assist openings) 24 are provided. These openings 11, 12, 21, 22, 23, and 24 are surrounded by a light shielding region (non-transparent region) 31. The light shielding region 31 is, for example, a light shielding region in which a chromium film is formed, or a translucent halftone phase shift region in which, for example, a molybdenum silicide film is formed.

  The main openings 11 and 12 have the same shape and dimensions, and the assist openings 21, 22, 23, and 24 have the same shape and dimensions. The assist openings 21, 22, 23, and 24 are smaller than the main openings 11 and 12.

  The main openings 11 and 12 are opening patterns (transfer patterns) corresponding to the contact hole pattern for bit line contact. After the exposure process and the development process, patterns corresponding to the main openings 11 and 12 are formed in the photoresist. The The assist openings 21, 22, 23, and 24 are auxiliary patterns (non-resolution assist patterns), and the pattern corresponding to the assist openings 21, 22, 23, and 24 is a photoresist even after the exposure process and the development process. Is not formed.

  A plurality of main openings 11 are arranged at a pitch of 2Py (second interval) on a straight line (first straight line) 41 extending in the bit line direction (second direction). That is, the center of each main opening 11 is positioned on the straight line 41. A plurality of main openings 12 adjacent to the main openings 11 are arranged at a pitch of 2Py on a straight line (second straight line) 42 extending in the bit line direction. That is, the center of each main opening 12 is located on the straight line 42.

  The straight line 41 and the straight line 42 are parallel to each other, and the distance between the straight line 41 and the straight line 42 (first distance (first interval) in the first direction (word line direction)) is Px. The main opening 11 and the main opening 12 are arranged so as to be shifted from each other by Py in the bit line direction.

  A plurality of assist openings 21 adjacent to the main opening 11 are arranged at a pitch of 2Py on a straight line (third straight line) 43 extending in the bit line direction. That is, the center of each assist opening 21 is located on the straight line 43. A plurality of assist openings 22 adjacent to the main opening 12 are arranged at a pitch of 2Py on a straight line (fourth straight line) 44 extending in the bit line direction. That is, the center of each assist opening 22 is positioned on the straight line 44. A plurality of assist openings 23 adjacent to the assist openings 21 are arranged at a pitch 2Py on a straight line (fifth straight line) 45 extending in the bit line direction. That is, the center of each assist opening 23 is positioned on the straight line 45. A plurality of assist openings 24 adjacent to the assist openings 22 are arranged at a pitch of 2Py on a straight line (sixth straight line) 46 extending in the bit line direction. That is, the center of each assist opening 24 is positioned on the straight line 46.

  The straight lines 41, 42, 43, 44, 45 and 46 are parallel to each other. The distance (first interval) between the straight line 41 and the straight line 43 is Px, and the distance between the straight line 42 and the straight line 44 is also Px. The distance between the straight line 43 and the straight line 45 is Px, and the distance between the straight line 44 and the straight line 46 is also Px.

  The assist opening 21 is arranged so as to be shifted from the main opening 11 by Py in the bit line direction. Similarly, the assist opening 22 is shifted from the main opening 12 by Py in the bit line direction. Therefore, the assist opening 21 and the assist opening 22 are arranged so as to be shifted from each other by Py in the bit line direction. On the other hand, the assist opening 23 is shifted from the assist opening 21 by Py in the bit line direction. Similarly, the assist opening 24 is shifted from the assist opening 22 by Py in the bit line direction. That is, the assist openings 22 and 23 are arranged at the same pitch (2Py) as the main openings 11 in the bit line direction. The assist openings 21 and 24 are arranged at the same pitch (2Py) as the main openings 12.

  As can be seen from the above description, the assist openings 23, the assist openings 21, the main openings 11, the main openings 12, the assist openings 22, and the assist openings 24 are arranged at the same pitch in the oblique direction. That is, in the photomask shown in FIG. 1, the assist openings 21, 22, 23, and 24 are added to increase the periodicity in the oblique direction (see, for example, JP-A-2008-065586).

Here, assuming that the numerical aperture of the projection lens is NA, the exposure wavelength is λ, and the pitch Px in the word line direction and the pitch Py in the bit line direction of the aperture pattern satisfy the relationship of the following formula (10): For example, a case where NA = 1.3, λ = 193 nm, Px = 80 nm, Py = 90 nm can be considered.

  When a fine hole pattern is to be formed under such conditions (wavelength λ and numerical aperture NA), using conventional illumination (vertical illumination light) as in the past, the contrast of the image formed on the substrate Becomes insufficient, and the conditions for the exposure amount or focus error are fragile. Therefore, the required hole pattern cannot be formed. Further, when the size of the opening pattern on the photomask is a numerical value obtained by dividing the target dimension of the hole pattern on the substrate by the magnification of the projection lens, there is no problem when the pitches Px and Py of the opening pattern are large. However, when the pitches Px and Py are small, the size of the opening pattern becomes a problem.

  The present embodiment is a fine hole pattern (dense hole pattern) suitable for exposure under the condition that the minimum pattern pitch of the opening pattern is λ / NA in the photolithography technique used for exposure of the hole pattern, It is possible to form a fine pattern in which holes are not arranged in an orthogonal lattice pattern.

  FIG. 2 shows a configuration example of illumination in this embodiment. In the case of this embodiment, dipole illumination which is modified illumination is used as illumination.

  As shown in FIG. 2, the dipole illumination has a light emitting region (first light emitting region) 51 and a light emitting region (second light emitting region) 52. These light emitting areas 51 and 52 are surrounded by a non-light emitting area 61.

  The light emitting area 51 and the light emitting area 52 are provided at positions symmetrical to each other with respect to the illumination center 70. That is, the light emitting region 51 and the light emitting region 52 have the same shape and the same size, and the center of the light emitting region 51 and the center of the light emitting region 52 are symmetrical with respect to the illumination center 70. The light emitting area 51 and the light emitting area 52 include a point 71 (first point) and a point 72 (second point), respectively. The points 71 and 72 are symmetric with respect to the illumination center 70. A point 71 and a point 72 pass through the illumination center 70 and extend in the word line direction (x direction, first direction) perpendicular to the bit line direction (y direction, second direction). Are symmetrical to each other. That is, the distance (dy) between the illumination center 70 and the point 71 and the distance (dy) between the illumination center 70 and the point 72 are equal to each other.

  Ideally, it is desirable that the center of the light emitting region 51 and the point 71 coincide, and the center of the light emitting region 52 and the point 72 coincide. In this case, the light emitting region 51 and the light emitting region 52 are symmetric with respect to the straight line 81.

When the wavelength of the illumination light is λ and the numerical aperture of the projection lens through which the illumination light passes is NA, the distance dy between the illumination center 70 and the point 71 and the distance dy between the illumination center 70 and the point 72 are: In the σ coordinate system, it is desirable that the relationship of the following expression (11) is satisfied. The σ coordinate system will be described later.

  By irradiating the photoresist with the oblique illumination light from the above-described modified illumination through the above-described photomask (see FIG. 1), a highly accurate contact hole pattern in which dimensional errors are suppressed is formed on the photoresist. Can be formed.

  FIG. 3 shows an example of a contact hole pattern formed in the photoresist after the exposure and development steps.

  As shown in FIG. 3, contact hole patterns 91 and 92 are formed in the photoresist 90. That is, patterns corresponding to the main openings 11 and 12 shown in FIG. 1 are formed in the photoresist 90 as contact hole patterns 91 and 92. Further, the pattern corresponding to the assist openings 21, 22, 23, and 24 shown in FIG.

  Here, the above-described σ coordinate system will be described with reference to FIG.

  In FIG. 4, 111 is an illumination optical system, 112 is a photomask, 113 is a projection optical system (projection lens), 114 is a substrate (semiconductor wafer), and 115 is an optical axis. The exit optical aperture of the illumination optical system 111 is sin (θ1), the incident-side numerical aperture of the projection optical system 113 is sin (θ2), and the σ value is defined as sin (θ1) / sin (θ2).

In modified illumination such as dipole illumination, it is common to use the σ coordinate system by extending the definition of the σ value. The σ coordinate system is a coordinate system in which the optical axis is the origin, and the numerical aperture on the incident side of the projection optical system is normalized to “1”. Therefore, the illumination position at point T in FIG.
(Σx, σy) = (sin (θ1) / sin (θ2), 0)
It becomes.

  The reason why a highly accurate contact hole pattern with reduced dimensional errors can be formed by the exposure method using the above-described photomask (see FIG. 1) and modified illumination (see FIG. 2) will be described below.

  When the pattern interval is smaller than λ / NA in the dimension on the substrate, the diffraction angle other than the 0th-order diffracted light does not reach the substrate when the vertical illumination light is used because the diffraction angle is large. Therefore, for example, as shown in FIG. 5, no light interference occurs and no image is formed. When oblique illumination light is used, an image can be formed by interference between 0th-order diffracted light and 1st-order diffracted light, for example, as shown in FIG.

  When the oblique illumination light is used, the periodic pattern has a greater depth of focus than the isolated pattern. Therefore, in the present embodiment, assist openings 21, 22, 23, and 24 as shown in FIG. 1 are added so that the entire pattern has periodicity. That is, since the main opening 11 and the main opening 12 shown in FIG. 1 are arranged in an oblique direction, the assist openings 21, 22, 23, and 24 are added to improve the periodicity in the oblique direction. .

  Next, the reason why the dipole illumination shown in FIG. 2 is desirable will be described. In the following, for simplification of description, description will be made assuming that the mask pattern (photomask) shown in FIG. 7 is used instead of the photomask shown in FIG.

  The photomask shown in FIG. 1 can be considered to generate diffracted light in the same direction as the mask pattern shown in FIG. 7 as a diffraction grating. In FIG. 7, 121 indicates a light shielding region, and 122 indicates an opening.

  Consider a case in which the mask pattern shown in FIG. 7 is irradiated with vertical illumination light from illumination (small σ illumination) as shown in FIG. That is, in the illumination of FIG. 8, the light emitting region 131 is provided at the center of the illumination. In this case, the diffracted light on the surface corresponding to the projection lens pupil shows a distribution as shown in FIG. The coordinate system of FIG. 9 is a σ coordinate system in which the radius (σ value) of the projection lens pupil is normalized to “1”. That is, FIG. 9 shows the distribution of diffracted light on the projection lens pupil plane when the mask pattern shown in FIG. 7 is Fourier transformed.

In FIG. 9, 141g is 0th-order diffracted light, and 141f is 1st-order diffracted light. The coordinate positions of the four first-order diffracted beams 141f are respectively
(+ Qx, + Qy)
(+ Qx, -Qy)
(-Qx, + Qy)
(-Qx, -Qy)
It becomes. However,
Qx = λ / (2Py × NA)
Qy = λ / (2Px × NA)
It is. Note that λ is the wavelength of the illumination light, and NA is the numerical aperture of the projection lens (projection optical system). In FIG. 7, the pitch of the openings 122 in the x direction is Px, and the pitch of the openings 122 in the y direction is Py. Further, reference numeral 142 in FIG. 9 denotes an effective area (unit circle) of the projection lens pupil, and only the diffracted light in the effective area 142 reaches the substrate. Accordingly, in the case of FIG. 9, since only one diffracted light (0th order diffracted light) 141g reaches the substrate, no light interference occurs and no image is formed on the substrate.

  Consider a case where the mask pattern shown in FIG. 7 is irradiated with oblique illumination light from modified illumination (oblique illumination) as shown in FIG. By appropriately shifting the position of the oblique illumination light (light emitting area 132) in the y-axis direction (shift amount σs), for example, as shown in FIG. 11, three diffracted lights 141a, 141b, 141c can be positioned. Accordingly, since the three diffracted beams 141a, 141b, and 141c pass through the projection lens and reach the substrate, light interference occurs and an image can be formed on the substrate.

  FIG. 12 shows an example in which an image (see FIG. 13) corresponding to the opening 122 in FIG. 7 is formed on the substrate due to the interference of the three diffracted beams 141a, 141b, and 141c shown in FIG. is there.

  As shown in FIG. 12, a one-dimensional interference fringe 151 is generated on the substrate due to the interference of the diffracted light 141a and the diffracted light 141b. Similarly, the interference fringes 152 are generated on the substrate due to the interference between the diffracted light 141b and the diffracted light 141c, and the interference fringes 153 are generated on the substrate due to the interference between the diffracted light 141c and the diffracted light 141a. The solid line indicates the bright part peak of the interference fringes, and the broken line indicates the dark part peak of the interference fringes. The light intensity is particularly high at the portion 155 where the bright portions of the three interference fringes 151, 152, 153 overlap. Therefore, as shown in FIG. 13, when a positive photoresist 90a is used, a contact hole pattern 93 is formed at a position corresponding to the portion 155.

  FIG. 13 shows an example in which the mask pattern shown in FIG. 7 is used. When the photomask as shown in FIG. 1 is used, an image is formed on the substrate with an image intensity corresponding to the sizes of the main openings 11 and 12 and the assist openings 21, 22, 23, and 24, thereby allowing the main Only the contact hole patterns 91 and 92 corresponding to the openings 11 and 12 can be formed in the photoresist 90 (see FIG. 3).

  Here, the relationship between the position and intensity of the diffracted light in the projection lens pupil will be described.

In FIG. 14, diffracted light 141a is light traveling straight through the mask, that is, zero-order diffracted light. The diffracted lights 141b and 141c are first-order diffracted lights. In a normally used binary mask or halftone phase shift mask, the amplitude and phase of the diffracted light beams 141b and 141c are the same. When the mask pattern is formed by a binary mask or a halftone phase shift mask, the complex amplitude transmittance of the mask light shielding body (light shielding region) is expressed as γ (a phase shift mask when negative, a binary mask when 0). Then, the amplitude (intensity) A of the diffracted light 141a and the amplitudes B and C of the diffracted lights 141b and 141c at the projection lens pupil are expressed by the following equations (12) and (13).

<Optimization of lighting position>
The illumination shift amount σs satisfies the following expression (14).

  In this case, the distances a, b, c from the pupil center of the three diffracted beams 141a, 141b, 141c are equal. As a result, there is no defocus dependence in the interference fringes on the substrate to be calculated later. That is, the depth of focus becomes sufficiently large.

  In addition, the illumination should be axisymmetric to eliminate image misregistration with respect to the focus. Therefore, the dipole illumination of FIG. 2 in which the illumination areas are arranged in axial contrast is desirable.

<Optimization of interference wave amplitude>
The intensity distribution I (x, y) of the interference wave formed on the substrate is expressed by the following equation (15).

When the above formulas (12) and (13) are substituted into the formula (15) and expanded, the following formula (16) is obtained.

  Here, θ is an angle formed by the traveling direction of each diffracted light and the normal line of the substrate surface. The first term on the right side of the above formula (16) is a uniform component, and the second and third terms are generated by interference between the diffracted light 141a and the diffracted light 141b and interference between the diffracted light 141a and the diffracted light 141c, respectively. It is an interference wave. The fourth term is an interference wave generated by the interference between the diffracted light 141b and the diffracted light 141c.

  Here, since illumination is given so as to satisfy the condition of the above equation (14), no component depending on z appears in the above equation (16). This indicates that the interference fringes are not affected by defocusing in the vicinity of the best focus.

In order to consider the contrast of the interference wave, the light intensity of the bright portion 156 and the two types of dark portions 157 and 158 shown in FIG. 15 will be described. The light intensities of the bright portion 156 and the two types of dark portions 157 and 158 are given by the following equations (17), (18), and (19), respectively. Although there are a plurality of bright parts and dark parts other than the illustrated part, their intensities are common due to symmetry. Each intensity | strength is calculated | required from said Formula (16).

In the case of FIG. 15, two types of contrast can be defined. That is, as shown in the following Expressions (20) and (21), the contrast C ₁ along the bright part 156 -dark part 157 -the bright part 156 and the bright part 156 -dark part 158-along the bright part 156 Contrast C ₂ .

Consider a condition in which the respective contrasts C ₁ and C ₂ are maximized. First, the maximum value of the contrast C ₁ along the bright part 156 to the dark part 157 to the bright part 156 is a condition in which the intensity of the dark part 157 is 0, that is, the case of “A = 0” from the above equation (18). At this time, “C ₁ = 1” but “C ₂ = 0”, so that a desired image cannot be obtained. On the other hand, the maximum value of the contrast C ₂ along the bright portion 156 to the dark portion 158 to the bright portion 156 is In this case, the condition where the intensity of the dark portion 158 is “0”, that is, the case of “A = 2B” from the above equation (19), where “C ₁ = 0.6, C ₂ = 1”. That is, the intensity of contrast C ₁ is not sufficient.

Therefore, assuming that the condition of “C ₁ = C ₂ ” is the optimum condition, this is a case where the intensities of the dark part 157 and the dark part 158 are equal, so that “A = B” from the above equations (18) and (19). It turns out that. At this time, “C ₁ = C ₂ = 0.8”.

  From the above, it has been clarified that a desirable imaging state can be obtained when the amplitudes A, B and C of the three diffracted beams 141a, 14ab and 14ac used for imaging are equal.

<Optimization of mask bias and light transmittance amplitude>
Since the desirable imaging state is when “A = B”, it can be rephrased as the case where the following expression (22) is established from the above expressions (12), (13), and (14).

When this equation (22) is transformed, the following equation (23) is obtained.

  That is, a desirable imaging state is obtained when the mask bias (ε) and the complex amplitude transmittance (γ) of the halftone phase shift mask satisfy the predetermined relationship represented by the above equation (23).

Here, for simplification, assuming that “εx = εy = ε”, the above equation (23) is expressed by the following equation (24).

  FIG. 16 shows the relationship between the mask bias (ε) and the complex amplitude transmittance (γ) of the halftone phase shift mask. “Γ” being negative represents a state in which the phase of light passing through the halftone area (mask light shielding body) is 180 degrees shifted from the light passing through the light transmission area. In all the states on this curve, the condition that the diffracted light intensity is equal is satisfied.

  FIG. 17 shows the relationship between the mask bias (the dimension ε of the opening on the mask) and the diffracted light amplitude A. From this figure, it can be seen that the image becomes bright when the mask bias is large and the complex amplitude transmittance of the mask light shield is large on the negative side. The image brightness needs to be set according to the desired throughput, resist sensitivity, laser luminance stability, etc., but can be set by appropriately selecting a combination of mask bias and complex amplitude transmittance.

  As described above, according to the present embodiment, the photomask having the assist openings 21, 22, 23, and 24 as shown in FIG. 1 is used, and the light emitting regions 51 and 52 as shown in FIG. By using the dipole illumination, even when the pattern is miniaturized, it is possible to form the high-precision contact hole patterns 91 and 92 with reduced dimensional errors as shown in FIG.

  Therefore, by applying the above-described photomask and modified illumination to the manufacture of a semiconductor device (hole pattern exposure), in a NAND flash memory, for example, a bit connected to a diffusion layer of a select transistor of a NAND cell unit Contact holes for line contacts can be formed with high accuracy.

  In the case of the first embodiment described above, the diffracted light amplitude expressed by the equations (12) and (13) in Formula 5 is a model in which the mask pattern is composed of an infinitely thin film, ie, Kirchhoff approximation. It is derived based on the model. In recent years, it has been known that when the minimum dimension of a mask pattern is about a wavelength or less, the Kirchhoff approximation model does not hold due to the influence of the mask thickness. In this case, the diffracted light intensity cannot be expressed by a simple expression such as Expressions (12) and (13) in Formula 5, but can be obtained by solving the Maxwell equation numerically. That is, the condition for equalizing the intensity of the three diffracted lights can be obtained by repeatedly calculating while changing the complex amplitude transmittance of the mask and the dimension of the contact hole pattern, as shown in FIG. 18, for example.

  In particular, when it is not necessary to perform a calculation in consideration of the influence of the mask thickness, the calculation shown in FIG. 18 may be executed by replacing the optical constant of the halftone film (HT) with the HT transmittance.

  In the above-described embodiment, the case where four rows of assist openings are provided has been described as an example. However, the present invention is not limited to this. For example, six rows or more assist openings may be provided.

  Further, the shape of the main opening and the assist opening is not limited to a square, and may be, for example, a rectangle, a circle, or an ellipse.

  Further, the shape of the light emitting region of the modified illumination is not limited to a circle, and may be an ellipse.

  Further, the modified illumination is not limited to the above-described dipole illumination, and for example, quadrupole illumination can be used.

  FIG. 19 shows a configuration example of quadrupole illumination suitable for forming a contact hole for bit line contact of a NAND flash memory (for example, a double staggered hole in a NAND-CB layer). .

  As shown in FIG. 19, the quadrupole illumination which is a modified illumination includes a light emitting region (first light emitting region) 251, a light emitting region (second light emitting region) 252, a light emitting region (third light emitting region) 253, In addition, a light emitting region (fourth light emitting region) 254 is provided. These light emitting areas 251, 252, 253 and 254 are surrounded by a non-light emitting area 261.

The light emitting regions 251, 252, 253, and 254 are provided at respective positions (regions) in the X-shape that are substantially symmetric in the x direction and the y direction with respect to the illumination center 270. That is, the light emitting regions 251, 252, 253, and 254 have the same shape and the same size. For example, the distance (σ) between the illumination center 270 and the point that the light emitting region 251 includes is expressed by the following equation (25): , (26), (27).

  Where NA is the numerical aperture of the projection lens, λ is the exposure wavelength, Px is the pitch of the aperture pattern in the word line direction, and Py is the pitch of the aperture pattern in the bit line direction.

  Incidentally, the distance (σ) between the center 270 of the illumination and the point included in the light emitting region 251 is σx, σy, and the distance (σ) between the center 270 of the illumination and the point included in the light emitting region 252 is σx, − The distance (σ) between the center 270 of the illumination and the point included in the light emitting region 253 is determined by σy, and the distance (σ) between the center 270 of the illumination and the point included in the light emitting region 254 is −− are given by σx and −σy, respectively.

  Ideally, the center of the light emitting region 251 matches the included point, the center of the light emitting region 252 matches the included point, and the center of the light emitting region 253 matches the included point. It is desirable that the center of the region 254 coincides with the included point. In this case, the light emitting regions 251, 252, 253, and 254 have the same distance (σ) from the illumination center 270.

  That is, in the quadrupole illumination having the light emitting regions 251, 252, 253, and 254 in the directions defined by the x direction and the y direction, three diffracted lights out of the diffracted light from the photomask are projected lens pupils. The illumination shape is set so as to pass through. In the case of the mask pattern shown in FIG. 7, according to this quadrupole illumination, for example, as shown in FIG. 20, there are three diffracted lights 241a and 241b in the effective areas 242a, 242b, 242c and 242d of the projection lens pupil, respectively. , 241c can be positioned. Therefore, since the three diffracted beams 241a, 241b, and 241c pass through the projection lens and reach the substrate, for example, light interference as shown in FIG. 15 (light peak and dark peak of interference fringes) occurs. An image can be formed on the substrate.

  FIG. 21 shows an exposure margin by quadrupole illumination having the light emitting regions 251, 252, 253 and 254 compared with an exposure margin by dipole illumination having the light emitting regions 51 and 52 (see FIG. 2). is there. The required margin shown in the figure is a contact hole actually formed with reference to a case where a contact hole pattern formed when a predetermined mask pattern is exposed with a certain exposure amount and focus has a desired dimension. This is a measure of how much the pattern size is deviated and cannot be accepted as an exposure amount or defocus.

  As is clear from this figure, the exposure margin (exposure amount fluctuation margin (EL) and depth of focus (DOF)) in the case of dipole illumination (shown ■) is almost the same as that in the case of quadrupole illumination (shown □). )) Is obtained.

  As a result, when the photomask of FIG. 1 is used, contact hole patterns 91 and 92 are formed in the photoresist 90 as shown in FIG.

  Further, by adopting quadrupole illumination, isolated contact hole patterns 190 arranged at random, for example, can be formed in the photoresist 90 separately from the dense hole patterns having periodicity such as the contact hole patterns 91 and 92. It becomes possible.

  In the case of the quadrupole illumination having the light emitting regions 251, 252, 253, and 254, the symmetry when rotated 90 degrees is superior to the case of the dipole illumination having the light emitting regions 51 and 52. Therefore, for example, as shown in FIG. 23A, by slightly correcting the photomask opening 222 to be slightly horizontally long, the isolated contact hole pattern 190 is formed into a good circular shape as shown in FIG. 23B, for example. Can be formed. Reference numeral 221 denotes a light shielding region of the photomask.

  Therefore, when forming a contact hole for a bit line contact in a NAND flash memory, by using a photomask including an opening pattern for forming an isolated contact hole, not only a contact hole for a bit line contact. For example, it becomes possible to simultaneously form isolated contact holes for peripheral circuits having a different period from the contact holes for bit line contacts.

  Further, the illumination is not limited to the above-described dipole illumination and quadrupole illumination. For example, as shown in FIG. 24, the light emitting regions (fifth and sixth light emitting regions) 51 and 52 and the light emitting regions (first Or a fourth light emitting region) 251, 252, 253, 254. These light emitting areas 51, 52, 251, 252, 253, and 254 have the same shape and the same dimensions as each other, and are surrounded by the non-light emitting area 261.

The light emitting areas 51, 52, 251, 252, 253, and 254 are positions of interest in the y direction with respect to the center of illumination, and positions (areas) in the X-shape that are substantially symmetric in the x and y directions. ) And. For example, the distance (σ) between the center of the illumination and the point included in the light emitting region 251 is expressed by the distance (σ) between the center of the illumination and the point included in the light emitting region 51 according to the following equations (28) and (29). ) Is given by the following Expression 15 (30).

  Where NA is the numerical aperture of the projection lens, λ is the exposure wavelength, Px is the pitch of the aperture pattern in the word line direction, and Py is the pitch of the aperture pattern in the bit line direction.

  Incidentally, the distance (σ) between the center of the illumination and the point included in the light emitting region 251 is σx and σy, and the distance (σ) between the center of the illumination and the point included in the light emitting region 252 is σx and −σy. The distance (σ) between the center of the illumination and the point included in the light emitting region 253 is −σx, σy, and the distance (σ) between the center of the illumination and the point included in the light emitting region 254 is −σx, −σy. Thus, the distance (σ) between the center of the illumination and the point included in the light emitting region 51 is given by σy, and the distance (σ) between the center of the illumination and the point included in the light emitting region 52 is given by −σy, respectively. .

  Ideally, the center of the light emitting region 51 matches the included point, the center of the light emitting region 52 matches the included point, and the center of the light emitting region 251 matches the included point. It is desirable that the center of the region 252 matches the included point, the center of the light emitting region 253 matches the included point, and the center of the light emitting region 254 matches the included point. In this case, the light emitting areas 51, 52, 251, 252, 253, and 254 have the same distance (σ) from the center of illumination.

  That is, in the hexapole illumination having the light emitting regions 51, 52, 251, 252, 253, and 254, the illumination shape is set so that three diffracted lights from the diffracted light from the photomask pass through the projection lens pupil. Has been. In the case of the mask pattern shown in FIG. 7, according to this hexapole illumination, as shown in FIG. 25, for example, there are three diffraction patterns in the effective areas 242a, 242b, 242c, 242d, 242e, and 242f of the projection lens pupil. Lights 241a, 241b, 241c can be positioned. Accordingly, since the three diffracted beams 241a, 241b, and 241c pass through the projection lens and reach the substrate, for example, light interference as shown in FIG. 15 occurs, and an image can be formed on the substrate. That is, as in the case of quadrupole illumination, when forming a contact hole for a bit line contact in a NAND flash memory, a photomask including an opening pattern for forming an isolated contact hole is used. In addition to the contact hole for bit line contact, for example, it is possible to simultaneously form isolated contact holes for peripheral circuits having a period different from that of the contact hole for bit line contact.

[Second Embodiment]
FIG. 26 shows an example of a photomask according to the second embodiment of the present invention. In the present embodiment, a contact hole for bit line contact of a NAND flash memory (a so-called dense hole pattern in which the holes are not arranged in an orthogonal lattice pattern, for example, a triple staggered hole in a NAND-CB layer) ) Will be described as an example.

  That is, in the present embodiment, as shown in FIG. 27, for example, in the NAND flash memory, the contact holes CB for bit line contacts respectively connected to the bit lines BL having a half pitch (HPnm) width shift the position. However, they are arranged in three rows (triple staggered arrangement). In this case, the contact holes CB are arranged at the same positions on the bit lines BL separated by 6 HP nm. As a result, in the NAND flash memory in which the bit lines BL are thinned and the pitch between the bit lines BL is narrowed, the contact holes CB for bit line contacts can be arranged (formed) with high accuracy. .

  In FIG. 26, the photomask includes a main opening (first main opening) 311, a main opening (second main opening) 312, a main opening (third main opening) 313, and an assist opening (first assist opening). 321, assist opening (second assist opening) 322, assist opening (third assist opening) 323, assist opening (fourth assist opening) 324, assist opening (fifth assist opening) 325, and assist opening A plurality of (sixth assist openings) 326 are provided. These openings 311, 312, 313, 321, 322, 323, 324, 325, 326 are surrounded by a light shielding region (non-transparent region) 331. The light shielding region 331 is, for example, a light shielding region in which a chromium film is formed, or a translucent halftone phase shift region in which, for example, a molybdenum silicide film is formed.

  The main openings 311, 312, and 313 have the same shape and dimensions, and the assist openings 321, 322, 323, 324, 325, and 326 have the same shape and dimensions. The assist openings 321, 322, 323, 324, 325, and 326 are smaller than the main openings 311, 312, and 313.

  The main openings 311, 312, and 313 are opening patterns (transfer patterns) corresponding to the contact hole pattern for bit line contact, and the pattern corresponding to the main openings 311, 312, and 313 is a photo pattern after the exposure process and the development process. Formed on resist. The assist openings 321, 322, 323, 324, 325, and 326 are auxiliary patterns (non-resolution assist patterns), and the assist openings 321, 322, 323, 324, 325, even after the exposure process and the development process. A pattern corresponding to 326 is not formed in the photoresist.

  A plurality of main openings 311 are arranged at a pitch Px (second interval) on a straight line (first straight line) 341 extending in the bit line direction (second direction). That is, the center of each main opening 311 is located on the straight line 341. A plurality of main openings 312 adjacent to the main openings 311 are arranged at a pitch Px on a straight line (second straight line) 342 extending in the bit line direction. That is, the center of each main opening 312 is located on the straight line 342. A plurality of main openings 313 adjacent to the main openings 312 are arranged at a pitch Px on a straight line (third straight line) 343 extending in the bit line direction. That is, the center of each main opening 313 is located on the straight line 343.

  The straight line 341, the straight line 342, and the straight line 343 are parallel to each other, and the distance between the straight line 341, the straight line 342, and the straight line 343 (first distance (first distance) in the first direction (word line direction)) is Py. It is. The main opening 311, the main opening 312, and the main opening 313 are disposed so as to be shifted from each other by Px / 3 (2 HPnm) in the bit line direction.

  A plurality of assist openings 321 adjacent to the main openings 311 are arranged at a pitch Px on a straight line (fourth straight line) 344 extending in the bit line direction. That is, the center of each assist opening 321 is positioned on the straight line 344. A plurality of assist openings 322 adjacent to the main openings 313 are arranged at a pitch Px on a straight line (fifth straight line) 345 extending in the bit line direction. That is, the center of each assist opening 322 is positioned on the straight line 345. A plurality of assist openings 323 adjacent to the assist openings 321 are arranged at a pitch Px on a straight line (sixth straight line) 346 extending in the bit line direction. That is, the center of each assist opening 323 is positioned on the straight line 346. A plurality of assist openings 324 adjacent to the assist openings 322 are arranged at a pitch Px on a straight line (seventh straight line) 347 extending in the bit line direction. That is, the center of each assist opening 324 is positioned on the straight line 347. A plurality of assist openings 325 adjacent to the assist openings 323 are arranged at a pitch Px on a straight line (eighth straight line) 348 extending in the bit line direction. That is, the center of each assist opening 325 is positioned on the straight line 348. A plurality of assist openings 326 adjacent to the assist openings 324 are arranged at a pitch Px on a straight line (ninth straight line) 349 extending in the bit line direction. That is, the center of each assist opening 326 is located on the straight line 349.

  Straight lines 341, 342, 343, 344, 345, 346, 347, 348 and 349 are parallel to each other. The distance (first interval) between the straight line 341 and the straight line 344 is Py, and the distance between the straight line 343 and the straight line 345 is also Py. The distance between the straight line 344 and the straight line 346 is Py, and the distance between the straight line 345 and the straight line 347 is also Py. The distance between the straight line 346 and the straight line 348 is Py, and the distance between the straight line 347 and the straight line 349 is also Py.

  Note that the assist openings 322 and 325 are arranged at the same pitch (Px) as the main openings 311 in the bit line direction. The assist openings 323 and 324 are arranged at the same pitch (Px) as the main openings 312. The assist openings 321 and 326 are arranged at the same pitch (Px) as the main openings 313. That is, the assist openings 321 and 326, the assist openings 323 and 324, and the assist openings 322 and 325 are arranged so as to be shifted by Px / 3 in the bit line direction.

  As can be seen from the above description, the assist opening 325, the assist opening 323, the assist opening 321, the main opening 311, the main opening 312, the main opening 313, the assist opening 322, the assist opening 324, and the assist opening 326 are inclined. They are arranged at the same pitch. That is, in the photomask shown in FIG. 26, assist periodicities 321, 322, 323, 324, 325, and 326 are added to increase the periodicity in the oblique direction.

Here, assuming that the numerical aperture of the projection lens is NA, the exposure wavelength is λ, and the pitch Py in the word line direction and the pitch Px in the bit line direction of the aperture pattern satisfy the relationship of Expression (31) of the following Expression 16, For example, a case where NA = 1.3, λ = 193 nm, Px = 110 nm, Py = 110 nm is conceivable.

  When a fine hole pattern is to be formed under such conditions (wavelength λ and numerical aperture NA), using conventional illumination (vertical illumination light) as in the past, the contrast of the image formed on the substrate Becomes insufficient, and the conditions for the exposure amount or focus error are fragile. Therefore, the required hole pattern cannot be formed. Further, when the size of the opening pattern on the photomask is a numerical value obtained by dividing the target dimension of the hole pattern on the substrate by the magnification of the projection lens, there is no problem when the pitches Px and Py of the opening pattern are large. However, when the pitches Px and Py are small, the size of the opening pattern becomes a problem.

  The present embodiment is a fine hole pattern (dense hole pattern) suitable for exposure under the condition that the minimum pattern pitch of the opening pattern is λ / NA in the photolithography technique used for exposure of the hole pattern, It is possible to form a fine pattern in which holes are not arranged in an orthogonal lattice pattern.

  FIG. 28 shows a configuration example of illumination in this embodiment. In the case of this embodiment, modified dipole illumination that is modified illumination is used as illumination.

  As shown in FIG. 28, the modified dipole illumination has a light emitting region (first light emitting region) 451 and a light emitting region (second light emitting region) 452. These light emitting regions 451 and 452 are surrounded by a non-light emitting region 461.

  The light emitting region 451 and the light emitting region 452 are provided at symmetrical positions defined by the x direction and the y direction with respect to the illumination center 470. That is, the light emitting region 451 and the light emitting region 452 have the same shape and the same size, and the center of the light emitting region 451 and the center of the light emitting region 452 are symmetrical with respect to the illumination center 470. In this case, the distance (σ) between the center of illumination 470 and the center of the light emitting region 451 and the distance between the center of illumination 470 and the center of the light emitting region 452 are equal to each other. Ideally, it is desirable that the center of the light emitting region 451 coincides with the contained point, and the center of the light emitting region 452 coincides with the contained point.

When the wavelength of the illumination light is λ and the numerical aperture of the projection lens through which the illumination light passes is NA, the distance σ between the illumination center 470 and the enclosing point is expressed by the following equations (32) and (33). Given by.

  By irradiating the photoresist with the oblique illumination light from the above-described modified illumination through the above-described photomask (see FIG. 26), a highly accurate contact hole pattern in which dimensional errors are suppressed is formed on the photoresist. Can be formed.

  FIG. 29 shows an example of a contact hole pattern formed in the photoresist after the exposure and development steps.

  As shown in FIG. 29, contact hole patterns 491, 492, 493 are formed in the photoresist 490. That is, patterns corresponding to the main openings 311, 312, and 313 shown in FIG. 26 are formed in the photoresist 490 as contact hole patterns 491, 492, and 493. In addition, the pattern corresponding to the assist openings 321, 322, 323, 324, 325, and 326 shown in FIG. 26 is not formed in the photoresist 490.

  The reason why a highly accurate contact hole pattern with reduced dimensional errors can be formed by the exposure method using the above-described photomask (see FIG. 26) and modified illumination (see FIG. 28) will be described below.

  When the pattern interval is smaller than λ / NA in terms of the dimension on the substrate, when vertical illumination light such as small σ illumination is used, the diffraction angle is large, so that diffracted light other than the 0th-order diffracted light does not reach the substrate. Therefore, for example, as shown in FIG. 5, no light interference occurs and no image is formed. When oblique illumination light such as modified dipole illumination is used, an image can be formed by interference between 0th-order diffracted light and 1st-order diffracted light, for example, as shown in FIG.

  When the oblique illumination light is used, the periodic pattern has a greater depth of focus than the isolated pattern. Therefore, in this embodiment, assist openings 321, 322, 323, 324, 325, and 326 as shown in FIG. 26 are added so that the entire pattern has periodicity. That is, since the main openings 311, 312, and 313 shown in FIG. 26 are arranged in an oblique direction, the periodicity in the oblique direction is improved by adding the assist openings 321, 322, 323, 324, 325, and 326. I am doing so.

  Next, the reason why the modified dipole illumination shown in FIG. 28 is desirable will be described. In the following, for simplification of description, description will be made assuming that the mask pattern (photomask) shown in FIG. 30 is used instead of the photomask shown in FIG.

  The photomask shown in FIG. 26 can be considered to generate diffracted light in the same direction as the mask pattern shown in FIG. 30 as a diffraction grating. In FIG. 30, reference numeral 521 denotes a light shielding region and 522 denotes an opening.

  Consider a case in which the mask pattern shown in FIG. 30 is irradiated with vertical illumination light from illumination (small σ illumination) as shown in FIG. That is, in the illumination of FIG. 8, the light emitting region 131 is provided at the center of the illumination. In this case, the diffracted light on the surface corresponding to the projection lens pupil shows a distribution as shown in FIG. The coordinate system of FIG. 31 is a coordinate system normalized by the numerical aperture NA of the projection lens. That is, FIG. 31 shows the distribution of diffracted light on the projection lens pupil plane when the mask pattern shown in FIG. 30 is Fourier transformed.

  In FIG. 31, 541g is 0th-order diffracted light, and 541f is 1st-order diffracted light. In FIG. 30, the pitch in the x direction of the openings 522 is Px, and the pitch in the y direction of the openings 522 is Py. Further, 542 in FIG. 31 is an effective area (unit circle) of the projection lens pupil, and only the diffracted light in the effective area 542 reaches the substrate. Therefore, in the case of FIG. 31, since only one diffracted light (0th order diffracted light) 541g reaches the substrate, no light interference occurs and no image is formed on the substrate.

  Consider a case where the mask pattern shown in FIG. 30 is irradiated with oblique illumination light from modified illumination (oblique illumination) as shown in FIG. By appropriately shifting the position of the oblique illumination light (light emitting region 132) in the x-axis direction and the y-axis direction (shift amounts σx, σy), for example, within the effective region 542 of the projection lens pupil as shown in FIG. Three diffracted beams 541a, 541b, and 541c can be positioned. Accordingly, since the three diffracted beams 541a, 541b, and 541c pass through the projection lens and reach the substrate, light interference occurs and an image can be formed on the substrate.

  FIG. 33 shows an example in which an image corresponding to the opening 522 in FIG. 30 is formed on the substrate due to interference of the three diffracted beams 541a, 541b, and 541c shown in FIG.

  As shown in FIG. 33, one-dimensional interference fringes 551 are generated on the substrate due to the interference of the diffracted light 541a (A) and the diffracted light 541b (B). Similarly, interference fringes 552 are generated on the substrate due to the interference between the diffracted light 541b (B) and the diffracted light 541c (C), and the interference on the substrate due to the interference between the diffracted light 541c (C) and the diffracted light 541a (A). Interference fringes 553 are generated. The solid line indicates the bright part peak of the interference fringes, and the broken line indicates the dark part peak of the interference fringes. The light intensity is particularly high at the portion where the bright portions 555 of the three interference fringes 551, 552, and 553 overlap. Therefore, when a positive photoresist is used, a contact hole pattern is formed at a position corresponding to the portion where the bright portion 555 overlaps.

  Therefore, when a photomask as shown in FIG. 26 is used, the image has an image intensity corresponding to the size of the main openings 311, 312, 313 and the assist openings 321, 322, 323, 324, 325, 326 on the substrate. By forming an image, only the contact hole patterns 491, 492, 493 corresponding to the main openings 311, 312, 313 can be formed in the photoresist 490 (see FIG. 29).

  Here, the relationship between the position and intensity of the diffracted light in the projection lens pupil will be described.

In FIG. 32, diffracted light 541a is light traveling straight through the mask, that is, zero-order diffracted light. The diffracted lights 541b, 541c, and 541f are first-order diffracted lights. In a normally used binary mask or halftone phase shift mask, the diffracted light 541b, 541c, and 541f have the same amplitude and phase of light. When the mask pattern is formed by a binary mask or a halftone phase shift mask, the complex amplitude transmittance of the mask light shielding body (light shielding region) is expressed as γ (a phase shift mask when negative, a binary mask when 0). Then, the amplitude (intensity) A of the diffracted light 541a and the amplitudes B, C, and D of the diffracted lights 541b, 541c, and 541f at the projection lens pupil are expressed by the following equations (34), (35), (36). , (37).

<Optimization of lighting position>
The shift amount σ of the modified dipole illumination is shifted by σx and σy with respect to the illumination center 470 so as to satisfy the above equations (32) and (33).

    In this case, as shown in FIG. 34, the distances r1, r2, r3 from the pupil center of the three diffracted beams 541a, 541b, 541c are equal. As a result, the defocus dependence is eliminated in the interference fringes on the substrate to be calculated later. That is, the depth of focus becomes sufficiently large.

<Optimization of interference wave amplitude>
The intensity distribution I (x, y, z) of the interference wave formed on the substrate by the three diffracted beams 541a, 541b, 541c is expressed by the following equation (38).

However, A, B, and C are the amplitudes of the three diffracted beams 541a, 541b, and 541c shown in FIG. 32, x is a position vector on the substrate, and ka, kb, and kc are wave number vectors shown in FIG. is there. The wave vector ka, kb, kc is expressed by the following equations (39), (40), (41).

  However, Sx and Sy are amounts representing the shift amount σ (σx, σy) of the modified dipole illumination, and when these values satisfy the above equations (32) and (33), the three diffracted beams 541a and 541b , 541c are incident on the substrate at the same angle. Therefore, kz is a common z component of the wave number vectors ka, kb, and kc.

When Expressions (39), (40), and (41) are substituted into the above Expression (38) and expanded, Expression (42) of the following Expression 21 is obtained.

  However, the first term on the right side of the formula (42) is a uniform component, and the second and third terms are generated by interference between the diffracted light 541a and the diffracted light 541b and interference between the diffracted light 541a and the diffracted light 541c, respectively. It is an interference wave. The fourth term is an interference wave generated by the interference between the diffracted light 541b and the diffracted light 541c.

  Here, since illumination is given so as to satisfy the condition of the above equation (35), a component depending on z does not appear in the above equation (42). This indicates that the interference fringes are not affected by defocusing in the vicinity of the best focus. That is, it can be said that the shift amount σ given by the above equations (32) and (33) is the optimum position for arranging the illumination area.

In order to consider the contrast of the interference wave, the light intensity of the bright portion 555 and the three types of dark portions 556 (dark portion 1), 557 (dark portion 2), and 558 (dark portion 3) shown in FIG. 33 will be described. As shown in FIG. 36, the light intensity of the bright portion 555 and the three types of dark portions 556, 557, and 558 is expressed by the following equations (43), (44), (45) and (46).

In the case of FIG. 33, three types of contrasts C ₁ , C ₂ , and C ₃ can be defined as shown in the following equations (47), (48), and (49).

Consider a condition in which the respective contrasts C ₁ , C ₂ and C ₃ are maximized. That is, it can be said that the condition in which the minimum value is the largest among the contrasts C ₁ , C ₂ , and C ₃ is the optimum values of the amplitudes A, B, and C of the three diffracted lights 541a, 541b, and 541c.

Therefore, assuming that B = pA and C = qA, the above equations (47), (48), and (49) become the following equations (50), (51), and (52) of the following equation (24).

FIG. 37 shows the relationship between the minimum values of contrasts C ₁ , C ₂ , and C ₃ and p and q. From this figure, it can be seen that when p = 1 and q = 1, the minimum values of the contrasts C ₁ , C ₂ and C ₃ are the largest. At this time, “A = B = C”, and high contrast (C ₁ = C ₂ = C ₃ = 0.8) can be obtained.

  From the above, when the diffracted light is taken as shown in FIG. 34, the desired image is formed when the amplitudes A, B, and C of the three diffracted lights 541a, 541b, and 541c used for image formation are equal. It became clear that the condition was obtained.

<Optimization of mask bias and light transmittance amplitude>
A desirable imaging state is when “A = B = C = D”. Consider the case of “B = C”. From the above equations (35) and (36), the following equation (53) is obtained.

When this equation (53) is developed, the following equation (54) is obtained.

However, it is clear that the above equation (54) is the following equation (55).

  That is, it can be said that “B = C” is not possible. Therefore, it is impossible to establish “A = B = C = D”.

Therefore, in order to give the state closest to “A = B = C = D”, it is defined that the optimum state is to minimize Δ (delta) given by the following equation (56).

  FIG. 38 shows the relationship between ε and Δ when the mask bias ε = εx = εy. From this figure, it can be seen that as the absolute value of γ (the complex amplitude transmittance of the halftone phase shift mask) increases, ε that minimizes Δ increases.

  FIG. 39 shows the relationship between ε and the amplitude A of the diffracted light 541a. From this figure, it can be seen that the image becomes brighter as ε increases. The image brightness needs to be set according to the desired throughput, resist sensitivity, laser luminance stability, etc., but can be set by appropriately selecting a combination of mask bias and complex amplitude transmittance.

  As described above, according to the present embodiment, a photomask having assist openings 321, 322, 323, 324, 325, and 326 as shown in FIG. 26 is used, and a light emitting region 451 as shown in FIG. , 452, even if the pattern is miniaturized (for example, the minimum pattern pitch is λ / NA or less), the dimensional error is suppressed as shown in FIG. The contact hole patterns 491, 492, and 493 can be formed.

  Therefore, by applying the above-described photomask and modified illumination to the manufacture of the semiconductor device (hole pattern exposure), in the NAND flash memory, for example, as shown in FIG. The contact hole CB for bit line contact can be formed with high accuracy.

  In the case of the second embodiment described above, the diffracted light amplitude expressed by the equations (34), (35), (36), and (37) is such that the mask pattern is composed of an infinitely thin film. However, if the Kirchhoff approximation model does not hold, for example, as shown in FIG. 18, it is repeated while changing the complex amplitude transmittance of the mask and the dimension of the contact hole pattern. It is obtained by calculating.

  In particular, when it is not necessary to perform a calculation in consideration of the influence of the mask thickness, the calculation shown in FIG. 18 may be executed by replacing the optical constant of the halftone film (HT) with the HT transmittance.

  In the embodiment described above, the case where six rows of assist openings are provided has been described as an example. However, the present invention is not limited to this, and for example, eight rows or more assist openings may be provided.

  Further, the shape of the main opening and the assist opening is not limited to a square, and may be, for example, a rectangle, a circle, or an ellipse.

  Further, the shape of the light emitting region of the modified illumination is not limited to a circle, and may be an ellipse.

  Further, the modified illumination is not limited to the above-described modified dipole illumination, and for example, modified quadrupole illumination can be used.

  FIG. 40 shows a configuration example of modified quadrupole illumination suitable for forming a contact hole for bit line contact of a NAND flash memory (for example, a triple staggered hole in the NAND-CB layer). is there.

  As shown in FIG. 40, the modified quadrupole illumination includes a light emitting region (first light emitting region) 651, a light emitting region (second light emitting region) 652, a light emitting region (third light emitting region) 653, and light emission. A region (fourth light emitting region) 654 is provided. These light emitting regions 651, 652, 653, 654 are surrounded by a non-light emitting region 661.

  The light emitting region 651, the light emitting region 654, the light emitting region 652, and the light emitting region 653 are provided at symmetrical positions defined by the x direction and the y direction with respect to the illumination center 670, respectively. The center of the light emitting region 651 and the center of the light emitting region 654 are symmetrical to each other with respect to the illumination center 670, and the center of the light emitting region 652 and the center of the light emitting region 653 are symmetrical to each other with respect to the illumination center 670. In position.

  Incidentally, the distance (σ) between the illumination center 670 and the point included in the light emitting region 651 is σx and σy, and the distance (σ) between the illumination center 670 and the point included in the light emitting region 652 is σx, − The distance (σ) between the center 670 of the illumination and the point included in the light emitting region 653 is determined by σy, and the distance (σ) between the center 670 of the illumination and the point included in the light emitting region 654 is −− They are given by σx and −σy, respectively.

  Ideally, the center of the light emitting region 651 matches the included point, the center of the light emitting region 652 matches the included point, the center of the light emitting region 653 matches the included point, and the light emitting region 654. It is desirable that the center and the included point coincide. In this case, the light emitting regions 651, 652, 653, and 654 have the same distance (σ) from the illumination center 670.

When the wavelength of the illumination light is λ and the numerical aperture of the projection lens through which the illumination light passes is NA, the distance (σ) between the illumination center 670 and the point included in the light emitting region 651 is expressed by the following equation (57) ), (58). Here, Px is the pitch of the opening pattern in the bit line direction, and Py is the pitch of the opening pattern in the word line direction.

  In this case, the three diffracted beams 541a, 541b, and 541c can take a state as shown in FIG.

In FIG. 40, for example, the distance (σ) between the illumination center 670 and the point included in the light emitting region 652 is given by the following equations (59) and (60).

  In this case, the three diffracted lights 541a, 541b, and 541c can take a state as shown in FIG.

  As described above, the modified quadrupole illumination having the light emitting regions 651, 652, 653, and 654 in the respective directions defined by the x direction and the y direction has three diffracted lights among the diffracted lights from the photomask. The illumination shape is set so that 541a, 541b, and 541c pass through the projection lens pupil. Therefore, since the three diffracted beams 541a, 541b, and 541c pass through the projection lens and reach the substrate, for example, light interference as shown in FIG. 33 (light peak and dark peak of interference fringes) occurs. An image can be formed on the substrate. As a result, when the photomask of FIG. 26 is used, contact hole patterns 491, 492, and 493 are formed in the photoresist 490, for example, as shown in FIG.

  In addition, by adopting modified quadrupole illumination, isolated contact hole patterns (for example, randomly arranged) can be formed on the photoresist 490 separately from the dense hole patterns having periodicity such as the contact hole patterns 491, 492, and 493. (Not shown) can be formed.

  Furthermore, the modified illumination is not limited to the above-described modified dipole illumination and modified quadrupole illumination. For example, as shown in FIG. 43, the light emitting regions (fifth and sixth light emitting regions) 451 and 452 and the light emitting regions are used. Modified hexapole illumination having (first to fourth light emitting regions) 651, 652, 653, 654 can also be used. These light emitting regions 451, 452, 651, 652, 653, 654 have the same shape and the same dimensions as each other, and are surrounded by the non-light emitting region 661.

  For example, the distance (σ) between the center 670 of the illumination and the points included in the light emitting regions 451 and 452 is given by the above equations (32) and (33). Further, the distances (σ) between the illumination center 670 and the points included in the light emitting regions 651 and 654 are given by the above equations (57) and (58). Further, the distance (σ) between the center 670 of the illumination and the point included in each of the light emitting regions 652 and 653 is given by the above formulas (59) and (60).

  As described above, in the modified hexapole illumination having the light emitting regions 451, 452, 651, 652, 653, and 654, three diffracted lights 541a, 541b, and 541c among the diffracted lights from the photomask are projected lens pupils. The illumination shape is set so as to pass through. Accordingly, since the three diffracted lights 541a, 541b, and 541c pass through the projection lens and reach the substrate, for example, light interference as shown in FIG. 33 occurs, and an image can be formed on the substrate. That is, as in the case of the modified quadrupole illumination, when forming a contact hole for a bit line contact in a NAND flash memory, use a photomask including an opening pattern for forming an isolated contact hole. Thus, not only a contact hole for bit line contact but also, for example, an isolated contact hole for a peripheral circuit having a different period from that of the contact hole for bit line contact can be formed at the same time.

  In the first and second embodiments described above, the case where the contact hole for the bit line contact of the NAND flash memory is formed has been described as an example. However, the present invention is not limited to this. For example, the present invention can be applied to the case where a pattern groove for wiring is formed in various semiconductor devices.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  11, 12, 311, 312, 313 ... main opening, 21, 22, 23, 24, 321, 322, 323, 324, 325, 326 ... assist opening, 31, 331 ... light shielding area, 51, 52, 251, 252 , 253, 254, 451, 452, 651, 652, 653, 654 ... light emitting region, 91, 92, 190, 491, 492, 493 ... contact hole pattern, 111 ... illumination optical system, 112 ... photomask, 113 ... projection Optical system, 114, substrate, 141a, 141b, 141c, 541a, 541b, 541c, diffracted light, 142, 542, effective area of projection lens pupil.

Claims (8)

Irradiating illumination light from an illumination light source to a photomask including a mask pattern composed of a transparent region and a non-transparent region, and projecting diffracted light from the photomask onto a substrate via a projection optical system A method for manufacturing a semiconductor device, wherein a photoresist pattern corresponding to the mask pattern is formed on the substrate,
The mask pattern has a fixed first interval in a first direction, and a plurality of parallel lines extending in a second direction orthogonal to the first direction, each center being individually parallel to the parallel line. Each of the lines includes a plurality of opening patterns which are the transparent regions arranged at a constant second interval in the second direction, and is arranged on an adjacent parallel line among the plurality of parallel lines. Each of the plurality of opening patterns is a mask pattern in which each center is arranged so as to be shifted by 1 / n (n is an integer of 3 or more) of the second interval in the second direction.
The illumination light source has an illumination shape set such that three of the diffracted lights from the photomask pass through the pupil of the projection optical system,
The illumination light source is a dipole illumination having a first light-emitting region and a second light-emitting region respectively arranged at symmetrical positions shifted from the center in the first direction and the second direction. There,
When the exposure wavelength is λ, the numerical aperture of the projection lens is NA, the second interval is Px, and the first interval is Py, the first light emitting region and the second light emitting region are the illumination light source. A point shifted from the center by a distance represented by the following formula (8) in the second direction and a distance represented by the following formula (9) in the first direction: Production method.

The plurality of opening patterns irradiated with illumination light from the dipole illumination are for forming contact holes for bit line contacts of a NAND flash memory,
The plurality of parallel lines are three,
The method of manufacturing a semiconductor device according to claim 1. Irradiating illumination light from an illumination light source to a photomask including a mask pattern composed of a transparent region and a non-transparent region, and projecting diffracted light from the photomask onto a substrate via a projection optical system A method for manufacturing a semiconductor device, wherein a photoresist pattern corresponding to the mask pattern is formed on the substrate,
The mask pattern has a fixed first interval in a first direction, and a plurality of parallel lines extending in a second direction orthogonal to the first direction, each center being individually parallel to the parallel line. Each of the lines includes a plurality of opening patterns which are the transparent regions arranged at a constant second interval in the second direction, and is arranged on an adjacent parallel line among the plurality of parallel lines. Each of the plurality of opening patterns is a mask pattern in which each center is arranged so as to be shifted by 1 / n (n is an integer of 3 or more) of the second interval in the second direction.
The illumination light source has an illumination shape set such that three of the diffracted lights from the photomask pass through the pupil of the projection optical system,
The illumination light source includes a first light emitting region, a second light emitting region, a third light emitting region, and a fourth light emitting device disposed at symmetrical positions shifted from the center in the first direction and the second direction, respectively. A quadrupole illumination having a light emitting area of
When the exposure wavelength is λ, the numerical aperture of the projection lens is NA, the second interval is Px, and the first interval is Py, the first light emission region to the fourth light emission region are the illumination light source. From the center of the illumination light source by the distance indicated by the following formula (10) in the second direction, by the distance indicated by the following formula (11) in the first direction, and from the center of the illumination light source 12) includes a point shifted in the second direction by a distance indicated by 12) and a point shifted in the first direction by a distance indicated by the following formula (13), respectively.
A method for manufacturing a semiconductor device.

The plurality of opening patterns irradiated with illumination light from the quadrupole illumination are for forming contact holes for bit line contacts of a NAND flash memory,
The plurality of parallel lines are three,
The semiconductor according to claim 3 , wherein the mask pattern further includes an opening pattern for forming an isolated contact hole for a peripheral circuit having a period different from that of the contact hole for the bit line contact. Device manufacturing method. Irradiating illumination light from an illumination light source to a photomask including a mask pattern composed of a transparent region and a non-transparent region, and projecting diffracted light from the photomask onto a substrate via a projection optical system A method for manufacturing a semiconductor device, wherein a photoresist pattern corresponding to the mask pattern is formed on the substrate,
The mask pattern has a fixed first interval in a first direction, and a plurality of parallel lines extending in a second direction orthogonal to the first direction, each center being individually parallel to the parallel line. Each of the lines includes a plurality of opening patterns which are the transparent regions arranged at a constant second interval in the second direction, and is arranged on an adjacent parallel line among the plurality of parallel lines. Each of the plurality of opening patterns is a mask pattern in which each center is arranged so as to be shifted by 1 / n (n is an integer of 3 or more) of the second interval in the second direction.
The illumination light source has an illumination shape set such that three of the diffracted lights from the photomask pass through the pupil of the projection optical system,
The illumination light source includes a first light emitting region, a second light emitting region, a third light emitting region, and a fourth light emitting device disposed at symmetrical positions shifted from the center in the first direction and the second direction, respectively. A hexapole illumination having a light emitting region, a fifth light emitting region, and a sixth light emitting region,
When the exposure wavelength is λ, the numerical aperture of the projection lens is NA, the second interval is Px, and the first interval is Py, the first light emission region to the sixth light emission region are the illumination light source. From the center of the illumination light source by the distance indicated by the following formula (14) in the second direction, by the distance indicated by the following formula (15) in the first direction, and from the center of the illumination light source by the following formula ( 16) the distance indicated by the following formula (18) from the center of the illumination light source and the point shifted in the first direction by the distance indicated by the following formula (17) by the distance indicated by 16) Only in the second direction, the point shifted in the first direction by a distance represented by the following formula (19), respectively,
A method for manufacturing a semiconductor device.

The plurality of opening patterns irradiated with illumination light from the hexapole illumination is for forming a contact hole for a bit line contact of a NAND flash memory,
The plurality of parallel lines are three,
The mask pattern further includes an opening pattern for forming an isolated contact hole for a peripheral circuit having a period different from that of the contact hole for the bit line contact.
6. A method of manufacturing a semiconductor device according to claim 5 , wherein: The semiconductor device manufacturing method according to claim 1 , wherein the mask pattern further includes a plurality of opening patterns that are different in dimension from the plurality of opening patterns and are not formed as the photoresist pattern. Method. The size of the plurality of aperture patterns and the complex amplitude transmittance of the non-transparent region are set in the photomask so that the difference in amplitude between the three diffracted lights is minimized.
The complex amplitude transmittance of the non-transparent region is γ, the second interval is Px, the interval of 1 is Py, the dimension of the opening pattern in the second direction and the dimension of the first direction are wx, When wy, εx = wx / Px, εy = wy / Py, so that Δ expressed by the following formula (20) is minimized, 6. The method of manufacturing a semiconductor device according to claim 1, wherein a dimension wx in the direction of the first dimension and a dimension wy in the first direction are set.

Images