Optical proximity correction (OPC) method and method for manufacturing a mask using the OPC method

Abstract

An optical proximity correction method includes: extracting edges of a layout of a pattern on a mask, the edges including at least one edge of the layout that is a curved edge; and generating an optical image of the pattern by applying an edge filter to the extracted edges of the layout, the edge filter including an arbitrary angle filter corresponding to an angle of the curved edge. The arbitrary angle filter may be generated using source sector rotation to correspond to the angle of the curved edge.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the priority of Korean Patent Application No. 10-2019-0110788 filed on September 6, 2019 in the Korean Intellectual Property Office, the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] The inventive concept relates to a method of manufacturing a mask, and more particularly, to an optical proximity correction (OPC) method and a method of manufacturing a mask using the OPC method. Background Art

[0004] In a semiconductor manufacturing process, a photolithography operation using a mask may be performed to form a pattern on a semiconductor substrate such as a wafer. In short, a mask may refer to a pattern transfer material that forms a pattern shape including an opaque material on a transparent substrate. At the beginning of the manufacturing process, a desired circuit may be designed, a layout of the circuit may be designed, and design data obtained using OPC may be transmitted as mask tape-out (MTO) design data. Thereafter, mask data preparation (MDP) may be performed based on the MTO design data, and a front end of line (FEOL) process (e.g., an exposure process) and a back end of line (BEOL) process (e.g., a defect inspection process) may be performed to manufacture a mask. Summary of the invention

[0005] The inventive concept provides an optical proximity correction (OPC) method and a method of manufacturing a mask using the OPC method, which can effectively correct a three-dimensional (3D) effect of a mask on a curved pattern.

[0006] According to one aspect of the inventive concept, an OPC method is provided, comprising: extracting an edge of a layout of a pattern on a mask; and generating an optical image of the pattern by applying an edge filter to the edge of the layout. The edge comprises a curved edge, and the edge filter comprises an arbitrary angle filter corresponding to an angle of the curved edge.

[0007] According to another aspect of the inventive concept, an OPC method is provided, comprising: (i) designing a layout of a pattern on a mask; (ii) extracting an edge of the layout; (iii) generating a first optical image of the pattern using a thin mask approximation; (iv) generating a second optical image of the pattern by applying an edge filter to the edge of the layout; and (v) generating a final optical image of the pattern by combining the first optical image with the second optical image. The edge comprises a curved edge, and the edge filter comprises an arbitrary angle filter corresponding to an angle of the curved edge.

[0008] According to another aspect of the inventive concept, a method for manufacturing a mask is provided. The method includes: extracting an edge of a layout of a pattern on a mask; generating an optical image of the pattern by applying an edge filter to the edge; generating an OPC model based on the optical image; obtaining design data about the mask by performing simulation using the OPC model; sending the design data as mask tape-out (MTO) design data; preparing mask data based on the MTO design data; and performing an exposure process on a substrate for the mask based on the mask data. The edge includes a curved edge, and the edge filter includes an arbitrary angle filter corresponding to an edge of the curved edge. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Embodiments of the present inventive concept will be more clearly understood from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0010] Figure 1 is a flow chart of the operation of an optical proximity correction (OPC) method according to an embodiment;

[0011] Figure 2A is a conceptual diagram showing the shape of a pattern on a mask in an OPC method of the related art;

[0012] Figure 2B is a conceptual diagram showing the shape of a pattern on a mask in an OPC method according to an embodiment;

[0013] FIG. 3A to FIG. 3C is a conceptual diagram for explaining a pattern including orthogonal edges, a pattern including curved edges, and an angle of the curved edges;

[0014] FIG. 4A to FIG. 4E is a conceptual diagram for explaining the principle of using an edge filter to generate an optical image of a pattern on a mask, wherein the pattern on the mask is Figure 3A A pattern including orthogonal edges;

[0015] Figure 5A and Figure 5B It is used to describe the Figure 3B Concept of arbitrary angle filter including patterns of curved edges and concept diagram of rotation of source sectors;

[0016] Fig. 6A is a conceptual diagram showing polarization existing in a transmissive optical system;

[0017] Figure 6B is a conceptual diagram used to illustrate polarization decomposition;

[0018] Fig. 7A and Figure 7B It is used to explain the use of vertical filters, horizontal filters, and about Figure 6BConceptual diagram showing the principle of polarization decomposition by using the weights of arbitrary angle filters;

[0019] Fig. 8A and Figure 8B is a graph showing a near-field image of a mask having an arbitrary angle one-dimensional (1D) pattern to illustrate the difference before and after applying an arbitrary angle filter to the mask having the arbitrary angle 1D pattern;

[0020] Fig. 9 is to show that by using a deep ultraviolet (DUV) light source Figure 3B A graph of an optical image calculated by rigorous simulation (R-SIM) of a mask of a pattern;

[0021] Fig. 10A It is shown that by using a DUV light source Figure 3B A graph of an optical image obtained by applying a prior art edge filter to a mask having a pattern;

[0022] Fig. 10B It is shown that by using a DUV light source Figure 3B a graph of an optical image obtained by applying an edge filter including an arbitrary angle filter to the patterned mask;

[0023] Fig.11A It is shown Fig. 9 The optical image and Fig. 10A a graph of the differences between the optical images of

[0024] Fig. 11B It is shown Fig. 9 The optical image and Fig. 10B a graph of the differences between the optical images of

[0025] Fig.12 is to show that by using an extreme ultraviolet (EUV) light source Figure 3B A graph of an optical image calculated by applying R-SIM to a mask of a pattern;

[0026] Fig.13A It is shown that by using an EUV light source Figure 3B A graph of an optical image obtained by applying a prior art edge filter to a mask having a pattern;

[0027] Fig. 13B It is shown that by using an EUV light source Figure 3B a graph of an optical image obtained by applying an edge filter including an arbitrary angle filter to the patterned mask;

[0028] Fig.14A It is shown Fig.12 The optical image and Fig.13Aa graph of the differences between the optical images of

[0029] Fig. 14B It is shown Fig.12 The optical image and Fig. 13B a graph of the differences between the optical images of

[0030] Fig.15 is a flow chart of operations of a method of manufacturing a mask according to an embodiment. DETAILED DESCRIPTION

[0031] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings are used to refer to the same elements, and their repeated description will be omitted.

[0032] Figure 1 is a flowchart of a process of an optical proximity correction (OPC) method according to an embodiment. Figure 2A is a conceptual diagram showing the shape of a pattern on a mask in the related art OPC method. Figure 2B is a conceptual diagram showing the shape of a pattern on a mask in the OPC method according to the present embodiment. FIG. 3A to FIG. 3C It is a conceptual diagram for explaining a pattern including orthogonal edges, a pattern including curved edges, and the angle of the curved edges.

[0033] Reference Figures 1 to 3C In the OPC method according to the present embodiment, first, the edge of the layout of the pattern on the mask may be extracted (S110). Here, the pattern on the mask may refer to a pattern that is transferred on a substrate such as a wafer using an exposure process to form a target pattern on the substrate. Generally, due to the characteristics of the exposure process, the shape of the target pattern on the substrate may be different from the shape of the pattern on the mask. In addition, since the pattern on the mask is reduced, projected, and transferred on the substrate, the size of the pattern on the mask may be larger than the target pattern on the substrate.

[0034] In addition, the layout of the pattern on the mask may have a one-dimensional (1D) or two-dimensional (2D) shape. Here, a 1D shape may not refer to a line without an area, but to a shape extending in one direction (e.g., a line and space pattern). The edge may correspond to the line surrounding the layout of the pattern. In some embodiments, the edge may correspond to the line of the space between the layout of the pattern. In addition, the edge may correspond to both the line of the layout of the pattern and the line of the space between the layout of the pattern.

[0035] Next, an optical image may be generated by applying an edge filter to the extracted edge (S120). Here, the optical image may be a mask near-field image calculated on the rear surface of the mask immediately after light passes through the mask. The edge filter may be a filter capable of generating an optical image of a portion corresponding to the edge, and the value of the edge filter may vary according to the position and characteristics of the edge. An edge filter corresponding to each edge may be applied to the corresponding edge, and thus, an optical image corresponding to each edge may be generated.

[0036] The optical image of the pattern on the mask will now be described in more detail. In recent years, in order to overcome the composition limitation and ensure the process margin, the shape of the pattern on the mask can be changed from Figure 2A The orthogonal shape shown extends to Figure 2B The curve shape shown. In addition, OPC methods for curved patterns are being studied intensively. Typically, the pattern on the mask may have a thickness that cannot be ignored compared to the exposure wavelength. Therefore, an accurate OPC method may require calculating the optical image while taking into account the mask topography effect (i.e., the mask 3D effect).

[0037] In order to perform OPC on a large area mask, it may be necessary to calculate an optical image in a short time. Therefore, it may not be appropriate to use a rigorous simulation method that takes a long time to calculate the optical image. Here, the rigorous simulation method may refer to, for example, an electromagnetic simulation method such as a rigorous coupled wave analysis (RCWA) method or a finite difference time domain (FDTD) simulation method. In the rigorous simulation method, a mask near-field image that accurately reflects the shape of the pattern on the mask can be obtained, but there is a problem that the calculation takes too much time.

[0038] Therefore, a more compact and computationally efficient simulation model can be used to calculate the optical image of the mask, and the domain decomposition method (DDM) can be a widely used method for calculating the optical image of the mask. The DDM may include dividing the pattern on the mask into various domains and using edge filter approximation to obtain the optical image. For example, the DDM may store the edge filter values ​​previously calculated for each width / space of the pattern using the FDTD simulation method in a filter library, and use the edge filter values ​​corresponding to each edge taken from the filter library to calculate the optical image of each pattern, so that the calculated optical image can be almost similar to the optical image calculated using the rigorous simulation method. Therefore, the DDM can generate an optical image in which the 3D effect of the mask can be effectively corrected.

[0039] Here, the optical image obtained using edge filter approximation may include a first optical image obtained using thin mask approximation (TMA) and a second optical image using an edge filter. In other words, the first optical image may be combined with the second optical image to generate a final optical image of the corresponding pattern. For reference, TMA (also known as Kirchhoff approximation) may refer to a near-field image calculated under the following assumptions: the mask has a very small thickness of substantially zero, the opening portion of the mask completely transmits light, and the blocking portion of the mask completely blocks light. Figure 4B TMA is described in more detail below. FIG. 4A to FIG. 4E The principles of using the TMA and edge filters to generate the final optical image are described in more detail.

[0040] In the OPC method of the present embodiment, the edge filter may also include an arbitrary angle filter as well as a vertical (90°) filter and a horizontal (0°) filter. Here, the vertical filter may refer to an edge filter applied to edges perpendicular to a set direction based on the direction, and the horizontal filter may refer to an edge filter applied to edges parallel to the direction. In addition, the arbitrary angle filter may refer to an edge filter applied to edges at arbitrary angles (neither perpendicular nor parallel to the direction). When non-orthogonal edges exist in the pattern to be formed, the arbitrary angle filter may be used for example for curved edges. The following will refer to Figure 5A and Figure 5B The definition of an arbitrary angle filter and a method of forming an arbitrary angle filter are described in more detail.

[0041] The edge filter will now be described in more detail. Figure 3A The orthogonal polygon shown extracts four edges. Here, the polygon may be referred to as a polygon that includes a combination of the respective vertices of the pattern corresponding thereto. In addition, based on the first direction (x direction), the four edges may be divided into two vertical edges Ve1 and Ve2 perpendicular to the first direction (x direction) and two horizontal edges He1 and He2 parallel to the first direction (x direction). In addition, vertical filters V-f1 and V-f2 may be applied to the corresponding extracted vertical edges Ve1 and Ve2, and horizontal filters H-f1 and H-f2 may be applied to the corresponding extracted horizontal edges He1 and He2, and thus, an optical image corresponding to each edge may be generated. Figure 3A In FIG. 1 , for simplicity, each of the vertical filters V-f1 and V-f2 and the horizontal filters H-f1 and H-f2 is shown in a simple wave form.

[0042] exist Figure 3BIn the curve polygon shown, the edge can be extracted in the form of a curve, and the vertical filter and the horizontal filter may not be applied to the curve edge. According to the OPC method of the prior art, the curve edge of the extreme ultraviolet (EUV) light source is de-angled at 0° and 90°, and the vertical filter and the horizontal filter are applied to the curve edge. Here, the de-angled curve edge of the EUV light source at 0° and 90° may refer to changing the curve edge into a fine orthogonal edge (e.g., a step edge). However, the method of de-angled the curve edge of the EUV light source at 0° and 90° may require too much calculation time and greatly reduce the accuracy of the calculated optical image.

[0043] According to the OPC method of the prior art, an inverse lithography technology (ILT) method may be applied to the curved polygon of the deep ultraviolet (DUV) light source. The ILT method may include: inversely calculating the shape of the pattern on the mask based on the shape of the pattern on the wafer. The ILT method may be disadvantageous in terms of time and cost because the pattern needs to be actually formed on the wafer and often requires several trial and error cycles.

[0044] In the OPC method of the present embodiment, an optical image can be generated by applying an arbitrary angle filter to a curved edge corresponding thereto, and therefore, an optical image that can effectively correct the 3D effect of the mask can be generated. In addition, an OPC model can be generated based on the optical image, and optimal design data about the mask can be obtained by performing simulation using the OPC model. In addition, in the OPC method of the present embodiment, a high-quality mask with reliability can be manufactured based on the optimal design data about the mask obtained using the OPC method. As a result, by using the mask in the process of patterning a product, the OPC method of the present embodiment can greatly contribute to improving the performance and reliability of the product.

[0045] Applying an arbitrary angle filter to a curve edge may not mean applying an arbitrary angle filter that exactly corresponds to the curve to the curve edge, but rather means: (i) extracting an arbitrary angle line representing the curve edge as an arbitrary angle edge, and (ii) applying an arbitrary angle filter corresponding to the linear arbitrary angle edge to the curve edge.

[0046] More specifically, Figure 3C yes Figure 3B An enlarged view of the area A shown as a two-point chain line in FIG. The arbitrary angle edge AAe can be extracted as a line having an average slope relative to the horizontal line within a set range of the curve edge (e.g., between a first position A1 and a second position A2 on the curve edge). Figure 3CAs shown, the arbitrary angle edge AAe may have an angle θ corresponding to the average slope relative to the horizontal line. An optical image of the curved edge may be generated by applying a corresponding arbitrary angle filter AA-f to the extracted arbitrary angle edge AAe. Figure 5A and Figure 5B The definition of an arbitrary angle filter and a method of forming an arbitrary angle filter are described in more detail.

[0047] After the optical image is generated, an OPC model may be generated based on the optical image (S130). Thereafter, design data on the mask may be obtained by performing simulation using the OPC model (S140).

[0048] The above-mentioned process of generating an optical image, generating an OPC model, and obtaining design data may be referred to as an OPC method. However, the OPC method may also include various other processes. An overall description of the OPC method is as follows.

[0049] As the pattern is reduced, an optical proximity effect (OPE) may occur during the exposure process due to the influence between adjacent patterns, and the OPC method may refer to a method of correcting the layout of the pattern to suppress and overcome the OPE. The OPC method can be roughly divided into a rule-based OPC method and a simulation-based or model-based OPC method. For example, the OPC method of the present embodiment can be a model-based OPC method. The model-based OPC method can be advantageous in terms of time and cost because the model-based OPC method only uses the measurement results of representative patterns without having to measure a large number of test patterns.

[0050] The OPC method may include not only a method of modifying the layout of a pattern but also a method of adding a sublithographic feature (which is called a serif) at a corner of the pattern, or a method of adding a sub-resolution assist feature (SRAF) (eg, a scattering strip).

[0051] Performing the OPC method may initially include preparing basic data for OPC. Here, the basic data may include data about the shape of the pattern of the sample, the position of the pattern, the type of measurement (e.g., measurement of the space or line of the pattern), and basic measurement values. In addition, the basic data may include information about the thickness, refractive index, and dielectric constant of the photoresist (PR), and a source map including the shape of the illumination system. However, the basic data is not limited to the above data examples.

[0052] After preparing the basic data, an optical OPC model may be generated. Generating the optical OPC model may include: optimizing the defocus stage (DS) position and the best focus (BF) position in the exposure process. In addition, generating the optical OPC model may include: generating an optical image taking into account the diffraction of light or the optical state of the exposure device. The optical image obtained using the edge filter approximation as described above can be used to generate the optical OPC model. However, generating the optical OPC model is not limited to the above description. For example, generating the optical OPC model may include various contents related to optical phenomena in the exposure process.

[0053] After generating the optical OPC model, an OPC model for PR may be generated. Generating the OPC model for PR may include optimizing a threshold value for PR. Here, the threshold value for PR may refer to a threshold value at which a chemical change occurs during an exposure process. For example, the threshold value may be the intensity of the exposure light. In addition, generating the OPC model for PR may include selecting an appropriate model form from among various PR model forms.

[0054] The combination of the optical OPC model and the OPC model for PR may be generally referred to as an OPC model. Therefore, the combination of the process of generating the optical OPC model and the process of generating the OPC model for PR may be referred to as the process of generating the OPC model (i.e., OPC modeling process). Hereinafter, unless otherwise defined, the OPC model may be interpreted as a combined concept of the optical OPC model and the OPC model for PR.

[0055] After the OPC model is generated, the OPC model may be verified. The verification of the OPC model may be performed by calculating the root mean square (RMS) value of the critical dimension (CD) error and checking the edge placement error (EPE). When the OPC model is included in the set specification, the verification of the OPC model may be completed, and the corresponding OPC model may be selected as the OPC model for simulation. If the OPC model is not included in the set specification, the operation of generating the OPC model (i.e., the operation of generating the optical OPC model and / or the OPC model for PR) may be performed again.

[0056] After verifying the OPC model, simulation can be performed using the corresponding OPC model. By performing simulation using the OPC model, design data of the mask close to actual measurement can be obtained. The design data of the mask obtained using the simulation can be sent to the mask manufacturing team as mask tape-out (MTO) design data to manufacture the mask later.

[0057] FIG. 4A to FIG. 4E is a conceptual diagram for explaining the principle of generating an optical image of a pattern on a mask using an edge filter, wherein the pattern on the mask is Figure 3A A pattern comprising orthogonal edges.

[0058] Figure 4A The shape of the pattern formed on the wafer using the mask is shown. For simplicity, Figure 4A A vertical cross section of a 1D line pattern is shown in FIG. Figure 4A As shown, the 1D line pattern may extend in the second direction (y direction) and have a predetermined width W in the first direction (x direction).

[0059] Figure 4B shows the use of an ideal thin mask TM to form Figure 4A The optical image of the case of the pattern shown. That is, Figure 4B An optical image obtained by approximating TMA with a thin mask (hereinafter referred to as a "thin mask image TMI") is shown. Here, the ideal thin mask TM may be a mask having a very small thickness, the opening portion of the ideal thin mask TM may completely transmit light, and the blocking portion thereof may completely block light. The thin mask image TMI shown below the thin mask TM may have a thickness similar to that of the thin mask TM. Figure 4A The patterns formed on the wafer are substantially similar in shape.

[0060] Figure 4C A real mask RM for forming a pattern is shown. The real mask RM may have a predetermined thickness t, and therefore, an optical image obtained through the real mask RM (hereinafter, referred to as a "real mask image RMI") may be different from a thin mask image TMI. The optical image shown below the real mask RM may be an optical image calculated by R-SIM (e.g., FDTD simulation). The optical image calculated using R-SIM may be substantially the same as the real mask image RMI.

[0061] Reference Figure 4D , showing the real mask image RMI and the thin mask image TMI together. Figure 4D In FIG. 1 , the real mask image RMI is shown by a solid line, and the thin mask image TMI is shown by a dotted line. Figure 4D As shown, the real mask image RMI may be different from the thin mask image TMI. Specifically, it can be confirmed that the portions of the real mask image RMI and the thin mask image TMI corresponding to the edge of the pattern are greatly different from each other. In other words, the farther away from the portion corresponding to the edge of the pattern, the more similar the real mask image RMI and the thin mask image TMI may become.

[0062] Figure 4E The shape of the difference ("subtraction") mask image SMI obtained by subtracting the thin mask image TMI from the real mask image RMI is shown. As described above, it can be seen that the portion of the difference mask image SMI adjacent to each of the left edge LE and the right edge RE has a large amplitude, and the amplitude of the difference mask image SMI decreases as it moves away from each of the left edge LE and the right edge RE. For simplicity, Figure 4EThe difference mask image SMI shown shows an enlarged scale of the amplitude compared to the real mask image RMI and the thin mask image TMI.

[0063] although Figure 4E The case where the portion of the difference mask image SMI corresponding to the left edge LE has the same shape as the portion of the difference mask image SMI corresponding to the right edge RE is shown, but the portions of the difference mask image SMI corresponding to the left edge LE and the right edge RE may exhibit different shapes for various reasons. For example, depending on whether light is incident on the mask obliquely or whether there is a pattern adjacent to the left edge LE and the right edge RE, or depending on the difference in pattern shape, the portion of the different mask image SMI corresponding to the left edge LE may exhibit a different shape from the portion of the different mask image SMI corresponding to the right edge RE.

[0064] The portion of the difference mask image SMI corresponding to the left edge LE and the right edge RE may correspond to an edge filter. Therefore, an edge may be extracted from the layout of the pattern, and an edge filter corresponding to the edge may be applied, and thus, a difference mask image SMI (i.e., a first optical image) may be generated. Thereafter, by adding a thin mask image TMI (i.e., a second optical image) obtained using TMA to the first optical image, an optical image substantially the same as the real mask image RMI may be generated. Here, the real mask image RMI may be substantially the same as the optical image calculated using R-SIM.

[0065] As a result, the edge filter value corresponding to the edge (i.e., the first optical image) may be previously calculated based on the width and space using R-SIM (e.g., FDTD simulation) and stored in the filter library. Thereafter, when calculating the optical image of each pattern, the edge filter value corresponding to each edge may be obtained from the filter library and used. Thus, an optical image substantially the same as the optical image calculated using R-SIM may be calculated.

[0066] Figure 5A and Figure 5B It is used to describe the Figure 3B A conceptual diagram of the concept of an arbitrary angle filter AA-filter including a curved edge pattern and the rotation of the source sector. Figure 5A , an arbitrary angle filter AA-filter can be represented as the sum of a vertical filter V-filter and a horizontal filter H-filter with weight "w". Figure 5AIn the right triangle of , assuming that the bottom portion is the horizontal filter H-filter and the height portion is the vertical filter V-filter, the arbitrary angle filter AA-filter may correspond to the hypotenuse (i.e., the longest side of the right triangle). In addition, the arbitrary angle filter AA-filter may be represented as the sum of the vertical filter V-filter and the horizontal filter H-filter to which appropriate weight w is applied according to the angle θ. For example, the arbitrary angle filter (i.e., AA-filter) may be represented by applying weights to the vertical filter and the horizontal filter according to the following relationship (Formula 1):

[0067] AA-filter=w×V-filter+(1-w)×H-filter,

[0068] Wherein "AA-filter" refers to any angle filter, "V-filter" refers to vertical filter, "H-filter" refers to horizontal filter, and "w" refers to weight. In some embodiments, the weight w can be calculated by sin 2 Therefore, when the angle θ is 0°, the arbitrary angle filter AA-filter may correspond to the horizontal filter H-filter, but when the angle θ is 90°, the arbitrary angle filter AA-filter may correspond to the vertical filter V-filter.

[0069] Although the AA filter has been described simply as the sum of the vertical filter V-filter and the horizontal filter H-filter, it can be referred to as follows Figure 5B The concept of source sector rotation is applied to the generation of arbitrary angle filter AA-filter. Figure 5B As in the compact simulation model described above, edge filter values ​​(e.g., vertical filter values ​​and horizontal filter values) calculated using R-SIM (e.g., FDTD simulation) can be stored in a filter library. Vertical filter values ​​and horizontal filter values ​​can be calculated and stored at each location or source point in the source sector. For example, Figure 5B A light source in the form of a circular source is shown. The circular source may be divided according to sectors or points, and vertical and horizontal filter values ​​may be calculated at the positions of the corresponding sectors and stored in a filter library.

[0070] In an orthogonal polygon, a sector can be specified based on a source, and an edge filter value (i.e., a vertical filter and / or a horizontal filter value) corresponding to the corresponding sector can be directly obtained from a filter library to calculate an optical image. However, in a curved polygon, unlike an orthogonal polygon, an arbitrary angle filter can be defined or generated by rotating a source sector or a source point and applied to the curved polygon to calculate an optical image.

[0071] More specifically, in the curved polygon, although the sector is defined based on the source, an edge filter value obtained by rotating the sector by an angle θ of the corresponding edge instead of the edge filter value corresponding to the corresponding sector may be obtained from the filter library to calculate the optical image. Figure 3C As described above, the angle θ of the corresponding edge can be the angle formed by any angle edge AAe extracted from the curve edge relative to the horizontal line. Figure 5B , assuming that the first point P1(r,0) of the horizontal axis Sx corresponds to an arbitrary angle edge having a first angle θ1, the edge filter value of the corresponding arbitrary angle edge may not be the edge filter value calculated at the first point P1(r,0), but the edge filter value calculated at the second point P2(rcosθ1,rsinθ1) obtained by rotating the first point P1(r,0) by the first angle θ1. As described above, when calculating the edge filter value of an edge filter having an arbitrary angle other than 0° or 90°, calculating the edge filter value at the rotated position of the source point instead of the original position of the source point may be referred to as source sector rotation. In addition, calculating the edge filter value using source sector rotation may include: generating an arbitrary angle filter for a curved edge and calculating the arbitrary angle filter value of the arbitrary angle filter.

[0072] In addition, as described above, the angle θ of the corresponding curve edge (e.g., the arbitrary angle edge AAe) can be determined as the weight "w" using a sine function, and the weight w can be used as in formula (1), so the arbitrary angle filter AA-filter can be expressed as the sum of the vertical filter V-filter and the horizontal filter H-filter.

[0073] However, since a reflective optical system (e.g., an EUV light source) does not need to consider polarization, an arbitrary angle filter can be generated only by source sector rotation. In contrast, since various polarizations (e.g., X, Y, XY, TE, and TM polarizations) exist in a transmissive optical system such as a DUV light source, polarization can be considered together with source sector rotation to generate an arbitrary angle filter. For example, in a DUV light source, the generation of an arbitrary angle filter may also include polarization decomposition as well as source sector rotation.

[0074] Fig. 6A is a conceptual diagram showing polarization existing in a transmission optical system. Figure 6B This is a conceptual diagram for explaining polarization decomposition. Fig. 7A and Figure 7B It is used to explain the use of vertical filters, horizontal filters, and about Figure 6B Conceptual diagram of the principle of polarization decomposition represented by the weights of arbitrary angle filters.

[0075] Now refer to Fig. 6ABriefly describe the polarization existing in the transmission optical system. Transverse magnetic field (TM) polarization TM(p)-pol. can also be called horizontal polarization or p-polarization, and generally refers to an electromagnetic wave whose oscillation direction of the electric field is parallel to the incident plane polarization. Transverse electric field (TE) polarization TE(s)-pol. can also be called vertical polarization or s-polarization, and refers to an electromagnetic wave whose oscillation direction of the electric field is perpendicular to the incident plane polarization. X polarization X-pol., Y polarization Y-pol., and XY polarization XY-pol. can be polarizations determined according to the user's definition. For example, when the polarization in the first direction (x direction) is defined as X polarization X-pol., the polarization in the second direction (y direction) perpendicular to the X polarization X-pol. can be Y polarization Y-pol., and the polarization in which the X polarization X-pol. and the Y polarization Y-pol. are distinguished and combined by sectors can correspond to XY polarization XY-pol.

[0076] Reference Figure 6B , when polarization is applied to each edge, the polarization can be decomposed and replaced by a linear combination of each vertical filter and each horizontal filter. Figure 6B In the above equation, X polarization X pol. can be expressed as the vector sum of TM polarization TM(p) and TE polarization TE(s). Figure 6B As shown, when the source sector rotates clockwise, since the TE polarization TE(s) corresponds to the vertical polarization (i.e., Y polarization Y pol.) and the TM polarization TM(p) corresponds to the horizontal polarization (i.e., X polarization X pol.), the X polarization X pol. after the source sector rotates can be expressed as the vector sum of the X polarization X pol. after the source sector rotates and the Y polarization Y pol.

[0077] For example, the polarization ΔD of an arbitrary angle filter with a 45° diagonal pattern can be converted to x x (r) is expressed as the linear sum of the X polarization X pol. and the Y polarization Y pol. of the vertical filter and the horizontal filter as in equation (2):

[0078] △D x x (r) = sin 2 45°*△V x x (r)+cos 2 45°*△V y y (r)=(△V x x (r)+△V y y (r)) / 2=(△V x x (r)+△Hx x (r)) / 2 (2),

[0079] Where: △(r) represents the signal value or light intensity of the edge filter value of the edge viewed from distance r, at D x x In the above equation, D represents an inclined or diagonal filter at any angle, the subscript represents the input polarization, and the superscript represents the output polarization. x x represents the X polarization of the filter at any angle, V x x Indicates the X polarization of the vertical filter, V y y Indicates the Y polarization of the vertical filter, H x x Indicates the X polarization of the horizontal filter.

[0080] The polarization ΔD of the arbitrary angle filter of the pattern with the edge of arbitrary angle θ can be expressed as in equation (3): x x (r) is expressed as the linear sum of the X polarization X pol. and Y polarization Y pol. of the vertical filter and the horizontal filter:

[0081] △D x x (r) = sin 2 θ*△V x x (r)+cos 2 θ*△V y y (r)=w*△V x x (r)+(1-w)*△V y y (r)=w*△V x x (r)+(1-w)*△H x x (r) (3).

[0082] In the process of deriving equation (2) or equation (3), △V y y (r) can be obtained by △H x x (r) instead, which can refer to Fig. 7A and Figure 7B That is, the polarization θ-pol. of the vertical filter V-filter at any angle θ can be expressed as the vector sum of the X polarization X pol. and the Y polarization Y pol. of the vertical filter V-filter.

[0083] Available from Fig. 7A It can be seen that the Y polarization Y pol. of the vertical filter V-filter can be parallel to the polarization of the vertical filter V-filter, and the X polarization X pol. of the vertical filter V-filter can be perpendicular to the polarization of the vertical filter V-filter. Figure 7B It can be seen that the Y polarization Y pol. of the horizontal filter H-filter can be a polarization perpendicular to the horizontal filter H-filter, and the X polarization X pol. of the horizontal filter H-filter can be a polarization parallel to the horizontal filter H-filter. The Y polarization Y pol. of the vertical filter V-filter can be a polarization substantially the same as the X polarization X pol. of the horizontal filter H-filter, because in terms of the concept of polarization, the polarization is parallel to the direction of the filter. Therefore, △V y y (r) can be obtained by △H x x (r) Replace.

[0084] Fig. 8A and Figure 8B is a graph showing a near-field image of a mask having an arbitrary angle 1D pattern to show the difference before and after applying an arbitrary angle filter to a mask having an arbitrary angle 1D pattern. In each graph, the abscissa represents position, the ordinate represents light intensity, and the values ​​on both the abscissa and the ordinate can be expressed in arbitrary units. Fig. 8A In , "before" means before applying any angle filter used in OPC. Figure 8B In , "after" means after applying the arbitrary angle filter. In addition, REF represents the near-field image calculated using R-SIM, EF-old represents the near-field image calculated using the edge filter of the prior art, and EF-new represents the near-field image calculated using the edge filter including the arbitrary angle filter.

[0085] Available from Fig. 8A and Figure 8B It can be seen that the near-field image calculated using the edge filter including the arbitrary angle filter shows a result level almost similar to that of the near-field image calculated using R-SIM. Therefore, it is predictable that an optical image in which the mask 3D effect is effectively corrected can be generated by applying the arbitrary angle filter to a pattern having a curved edge corresponding to an arbitrary angle. In order to confirm the direct effect on the edge filter including the arbitrary angle filter, two tests can be performed.

[0086] Fig. 9 It is shown that by using a DUV light source Figure 3B A graph of the optical image calculated using R-SIM on a patterned mask. Fig. 10A It is shown that by using a DUV light source Figure 3B A graph of an optical image obtained by applying a prior art edge filter to a patterned mask. Fig. 10B It is shown that by using a DUV light source Figure 3B A graph of an optical image obtained by applying an edge filter including an arbitrary angle filter to a patterned mask. Fig.11A It is shown Fig. 9 The optical image and Fig. 10A A graph of the difference images between the optical images. Fig. 11B It is shown Fig. 9 The optical image and Fig. 10B In each graph, the horizontal and vertical axes represent the position, the vertical shaded band on the right represents the light intensity, and all values ​​can be expressed in arbitrary units.

[0087] Reference Figures 9 to 11B ,exist Fig.11A Shown Fig. 9 and Fig. 10A In the graph of the difference image between the optical images of Fig. 11B Shown Fig. 9 and Fig. 10B In the graph of the difference image between the optical images of , the maximum value of the light intensity is about 0.91. Therefore, in the DUV light source, it can be confirmed that the optical image generated using the edge filter including the arbitrary angle filter is 10% or more better than the optical image generated using the edge filter of the prior art.

[0088] For reference, the higher the light intensity of the difference image, the greater the difference between the optical image calculated using R-SIM and the optical image calculated using the edge filter becomes. In addition, it can be seen that the difference between the optical images occurs due to the mask 3D effect. Therefore, it can be determined that as the light intensity of the difference image decreases, the mask 3D effect is more effectively corrected in the optical image calculated using the edge filter.

[0089] Fig.12 It is shown that by using an EUV light source Figure 3B A graph of the optical image calculated using R-SIM on a patterned mask. Fig.13A It is shown that by using an EUV light source Figure 3B A graph of an optical image obtained by applying a prior art edge filter to a patterned mask. Fig. 13B It is shown that by using an EUV light source Figure 3B A graph of an optical image obtained by applying an edge filter including an arbitrary angle filter to a patterned mask. Fig.14A It is shown Fig.12 The optical image and Fig.13AA graph of the difference images between the optical images. Fig. 14B It is shown Fig.12 The optical image and Fig. 13B In each graph, the horizontal and vertical axes represent positions, the vertical shaded band on the right represents light intensity, and all values ​​can be expressed in arbitrary units.

[0090] Reference Figures 12 to 14B ,exist Fig.14A Shown Fig.12 and Fig.13A In the graph of the difference image between the optical images of Fig. 14B Shown Fig.12 and Fig. 13B In the graph of the difference image between the optical images of , the maximum value of the light intensity is about 5.66. Therefore, in the EUV light source, it can be confirmed that the optical image generated using the edge filter including the arbitrary angle filter is about 20% better than the optical image generated using the edge filter of the prior art.

[0091] Figures 9 to 14B is a diagram showing the first test method (image-based analysis method). That is, by using R-SIM to calculate the Figure 3B The values ​​obtained by optically imaging a mask having a curved pattern (e.g., a peanut pattern) as shown are compared with the values ​​calculated using the edge filter of the prior art and the edge filter including the arbitrary angle filter. Therefore, the accuracy of the optical image in the case of using the edge filter including the arbitrary angle filter can be estimated.

[0092] In the second test method, a pattern may be formed on a wafer due to an exposure process using a mask including an actual curved pattern, and an actual measurement value of the formed pattern may be used. That is, the accuracy of the optical image may be estimated by comparing the actual measurement value of the pattern on the wafer with a calculated value of the pattern formed by calibrating the OPC model according to whether an arbitrary angle filter is used.

[0093] In the second test method, compared with the case of using only the edge filter of the prior art, it can be confirmed that in both the DUV light source and the EUV light source, the results obtained by calculating the OPC model using the edge filter including the arbitrary angle filter are improved. For example, the optical RMS value is improved by at least 25% in the DUV light source, and the optical RMS value is improved by at least 11% in the EUV light source. Here, the optical RMS value may be the RMS value of the difference between the measured value and the value calculated using the OPC model. It can be confirmed that the smaller the optical RMS value becomes, the closer the design data obtained using the OPC model becomes to the design data capable of forming an actual mask.

[0094] Fig.15 is a flowchart of the operation of the method for manufacturing a mask according to an embodiment. Figures 1 to 7B Same description as in . Fig.15 In the method for manufacturing a mask according to the present embodiment, as a start, an OPC method may be performed. For example, the OPC method may include an operation S110 of extracting an edge to an operation S140 of obtaining design data. The operation S110 of extracting an edge to the operation S140 of obtaining design data may be the same as the above-mentioned Figure 1 Operations S110 to S140 of the OPC method are the same.

[0095] After performing the OPC method, the design data may be sent as mask tape-out (MTO) design data (S150). Generally, MTO may refer to requesting the manufacture of a mask by delivering the final mask data obtained using the OPC method to a mask manufacturing team. Therefore, the MTO design data may ultimately correspond to the design data about the mask obtained using the OPC method. The MTO design data may have a graphic data format used in electronic design automation (EDA) software, etc. For example, the MTO design data may have a data format such as Graphic Data System II (GDS2) and Open Artwork System Interchange Standard (OASIS).

[0096] After the MTO design data is transmitted, mask data preparation (MDP) may be performed (S160). For example, MDP may include format conversion called fracturing, barcode enhancement for mechanical reading, standard mask patterns for inspection, and job decks, as well as automatic verification and manual verification. Here, the job deck may refer to a test file that creates a series of commands (e.g., information on the arrangement of multiple mask files, reference metrology, exposure speed or method, etc.).

[0097] Meanwhile, format conversion (or splitting) may refer to the process of splitting the MTO design data of each region and converting the split MTO design data into a format for an electron beam (e-beam) exposure device. For example, the splitting process may include a data manipulation process (e.g., a scaling process), a data resizing process, a data rotation process, a pattern reflection process, and a color inversion process. During the conversion process using the splitting process, data about several system errors that may occur anywhere in the transfer process from the design data to the image on the wafer may be corrected. The process of correcting the data about the system errors may be referred to as mask process correction (MPC), which may include, for example, operations of adjusting the line width (which is referred to as CD adjustment) and increasing the accuracy of the pattern arrangement. Therefore, splitting may help improve the quality of the final mask and is performed before MPC. Here, system errors may occur due to deformations caused by the exposure process, the mask development process, the etching process, and the wafer imaging process.

[0098] Meanwhile, the MDP may include an MPC. As described above, the MPC may refer to a process of correcting an error (i.e., a system error) caused during an exposure process. Here, the exposure process may be a general concept including an electron beam writing process, a developing process, an etching process, a baking process, and the like. In addition, a data processing process may be performed before the exposure process. The data processing process may be a process of preprocessing mask data, and includes performing a syntax check on the mask data and predicting an exposure time.

[0099] After the MDP, the substrate for the mask may be exposed based on the mask data (S170). Here, the exposure process may refer to, for example, an electron beam writing process. For example, the electron beam writing process may be performed in a gray writing manner using a multi-beam mask writer (MBMW). Alternatively, the electron beam writing process may be performed using a variable shape beam (VSB) mask writer.

[0100] After the MDP, an operation of converting the mask data into pixel data may be performed before the exposure process. The pixel data, which is data directly used in the actual exposure process, may include data on the shape to be exposed and data on the dose assigned to each shape. Here, the data on the shape may be bitmap data into which the shape data, which is vector data, is converted due to rasterization.

[0101] After the exposure process, the mask (S180) may be completed by performing a series of processes. The series of processes may include, for example, a development process, an etching process, and a cleaning process. In addition, the series of processes for manufacturing the mask may include a measurement process, a defect inspection process, or a defect repair process. In addition, the series of processes for manufacturing the mask may include a thin film coating process. The thin film coating process may refer to a process in which a thin film is attached to the surface of the mask to protect the mask from subsequent contamination during the transfer of the mask and to protect the usable life of the mask when the absence of contaminants or chemical stains is confirmed by the final cleaning and inspection process.

[0102] In the method for manufacturing a mask according to the present embodiment, the mask may be a mask for DUV or a mask for EUV. However, the mask is not limited to a mask for DUV or a mask for EUV. For example, the mask may be a mask for a wavelength other than DUV or EUV.

[0103] In the method for manufacturing a mask according to the present embodiment, an edge filter including an arbitrary angle filter can be used to generate an optical image. Therefore, an optical image can be generated by effectively correcting a 3D mask effect. In addition, an OPC model can be generated based on the optical image, and optimal mask design data can be obtained by performing simulation using the OPC model. In addition, in the method for manufacturing a mask according to the present embodiment, a mask can be manufactured using an exposure process based on the optimal mask design data, thereby realizing an excellent mask that can effectively form a target pattern on a wafer.

[0104] While the inventive concept has been particularly shown and described with reference to embodiments thereof, it will be understood that various changes in form and details may be made therein without departing from the spirit and scope of the following claims.

Claims

1. An optical proximity correction method, comprising the steps of: extracting edges of a layout of a pattern on a mask, the edges including at least one edge of the layout that is a curved edge; and An optical image of the pattern is generated by applying an edge filter to the extracted edge of the layout, the edge filter being a filter capable of generating an optical image of a portion corresponding to the edge and having a value that can vary according to a position and characteristics of the edge, the edge filter comprising an arbitrary angle filter corresponding to an angle of the curved edge, the arbitrary angle filter being generated using source sector rotation to correspond to the angle of the curved edge.

2. The optical proximity correction method according to claim 1, wherein: The edge filter includes a vertical filter and a horizontal filter applied to orthogonal edges; and wherein the arbitrary angle filter is represented by applying weights to the vertical filter and the horizontal filter according to the following relationship: AA-filter=w×V-filter+(1-w)×H-filter, Wherein "AA-filter" refers to the arbitrary angle filter, "V-filter" refers to the vertical filter, "H-filter" refers to the horizontal filter, and "w" refers to the weight.

3. The optical proximity correction method according to claim 2, wherein: w is equal to sin 2 θ, wherein θ refers to an angle corresponding to an average slope relative to a horizontal line within a set range of the edge of the curve.

4. The optical proximity correction method according to claim 1, wherein: In a reflected light source, the arbitrary angle filter is generated using source sector rotation, and wherein in a transmitted light source, the arbitrary angle filter is generated using source sector rotation and polarization decomposition.

5. The optical proximity correction method according to claim 4, wherein: The polarization decomposition is a method of forming X polarization, Y polarization, XY polarization, TE polarization, or TM polarization by combining X polarization and Y polarization according to the position of a source sector.

6. The optical proximity correction method according to claim 5, wherein: The polarization of the arbitrary angle filter is represented as a linear sum of the X polarization and the Y polarization of the vertical filter and the horizontal filter by applying the polarization decomposition and weights.

7. The optical proximity correction method according to claim 6, wherein: The polarization of the arbitrary angle filter is equal to: △D x x (r)=w×V x x (r)+(1-w)×△H x x (r), Where △(r) is the signal value of the edge filter value of the edge viewed from distance r, D x x is the X polarization of the arbitrary angle filter, V x x is the X polarization of the vertical filter, H x x is the X polarization of the horizontal filter, and "w" refers to the weight, which is equal to sin 2 θ, where θ is the angle of the edge of the curve.

8. The optical proximity correction method according to claim 1, further comprising: generating a first optical image using a thin mask approximation; as well as The first optical image and the optical image generated using the edge filter are combined as a final optical image of the pattern.

9. The optical proximity correction method according to claim 1, wherein: The step of generating an optical image is followed by the following steps: generating an optical proximity correction model based on the optical image; and Design data on the mask is obtained by performing simulation using the optical proximity correction model.

10. An optical proximity correction method comprising the steps of: designing the layout of the pattern on the mask; Extracting edges of the layout, wherein the edges at least include curved edges; generating a first optical image of the pattern using a thin mask approximation; A second optical image of the pattern is generated by applying an edge filter to the edge of the layout, the edge filter being a filter capable of generating an optical image of a portion corresponding to the edge and having a value that can vary according to a position and characteristics of the edge, the edge filter comprising an arbitrary angle filter corresponding to an angle of the curved edge, the arbitrary angle filter being generated using source sector rotation to correspond to the angle of the curved edge.

11. The optical proximity correction method according to claim 10, wherein: The edge filter comprises a vertical filter and a horizontal filter to be applied to orthogonal edges; and wherein the arbitrary angle filter is represented by applying a weight corresponding to the angle of the curved edge to the vertical filter and the horizontal filter.

12. The optical proximity correction method according to claim 10, wherein: The arbitrary angle filter is generated using source sector rotation in a reflected light source; and wherein the arbitrary angle filter is generated using source sector rotation and polarization decomposition in a transmitted light source.

13. The optical proximity correction method according to claim 12, wherein: The polarization decomposition is a method of forming X polarization, Y polarization, XY polarization, TE polarization or TM polarization by combining X polarization and Y polarization according to the position of the source sector; and wherein the polarization of the arbitrary angle filter is represented as a linear sum of the X polarization and the Y polarization of the vertical filter and the horizontal filter by applying the polarization decomposition and weights.

14. The optical proximity correction method according to claim 10, wherein: The step of generating a second optical image is followed by the following steps: generating a final optical image of the pattern by combining the first optical image with the second optical image; generating an optical proximity correction model based on the final optical image; as well as Design data on the mask is obtained by performing simulation using the optical proximity correction model.

15. A method for manufacturing a mask, the method comprising the steps of: Extracting edges of the layout of the pattern on the mask, the edges at least including curved edges; generating an optical image of the pattern by applying an edge filter to the edge, the edge filter being a filter capable of generating an optical image of a portion corresponding to the edge and having a value that can vary according to a position and characteristics of the edge, the edge filter comprising an arbitrary angle filter corresponding to an angle of the curved edge, the arbitrary angle filter being generated using source sector rotation to correspond to the angle of the curved edge; generating an optical proximity correction model based on the optical image; obtaining design data about the mask by performing simulation using the optical proximity correction model; sending the design data as mask tape-out design data; preparing mask data based on the mask tape-out design data; as well as An exposure process is performed on the substrate for the mask based on the mask data.

16. The method according to claim 15, wherein: The edge filter comprises a vertical filter and a horizontal filter to be applied to orthogonal edges; and wherein the arbitrary angle filter is represented by applying a weight corresponding to the angle of the curved edge to the vertical filter and the horizontal filter.

17. The method according to claim 16, wherein: The arbitrary angle filter is represented by applying weights to the vertical filter and the horizontal filter according to the following relationship: AA-filter=w×V-filter+(1-w)×H-filter, Wherein "AA-filter" refers to the arbitrary angle filter, "V-filter" refers to the vertical filter, "H-filter" refers to the horizontal filter, and "w" refers to the weight; and wherein w is equal to sin 2 θ, where θ refers to the angle of the edge of the curve.

18. The method according to claim 15, wherein: The arbitrary angle filter is generated using source sector rotation in a reflected light source; and wherein the arbitrary angle filter is generated using source sector rotation and polarization decomposition in a transmitted light source.

19. The method according to claim 18, wherein: The polarization decomposition is a method of forming X polarization, Y polarization, XY polarization, TE polarization or TM polarization by combining X polarization and Y polarization according to the position of the source sector; and wherein the polarization of the arbitrary angle filter is expressed as a linear sum of the X polarization and the Y polarization of the vertical filter and the horizontal filter by applying the polarization decomposition and weights.

20. The method according to claim 19, wherein: The polarization of the arbitrary angle filter is equal to: △D x x (r)=w×V x x (r)+(1-w)×△H x x (r), Where △(r) is the signal value of the edge filter value of the edge viewed from distance r, D x x is the X polarization of the arbitrary angle filter, V x x is the X polarization of the vertical filter, H x x is the X polarization of the horizontal filter, and "w" refers to the weight, which is equal to sin 2 θ, where θ is the angle of the edge of the curve.

21. An optical proximity correction method, comprising the steps of: An optical image of the pattern is generated by applying edge filters to multiple edges of a layout of the pattern on a mask, the edge filter being a filter capable of generating an optical image of a portion corresponding to the edge and having a value that can vary according to the position and characteristics of the edge, the edge filter including at least one arbitrary angle filter corresponding to an angle of a curved edge within the multiple edges, the arbitrary angle filter being generated using source sector rotation to correspond to the angle of the curved edge.

Images