Optical proximity correction method maintaining consistency and method for manufacturing a mask using the same

By dividing the semiconductor device layout into multiple regions and calculating representative values, the layout deformation problem caused by mask image proximity in semiconductor device manufacturing is solved, and more efficient optical proximity correction and consistency are achieved, thereby improving device reliability.

CN112668272BActive Publication Date: 2024-12-06SAMSUNG ELECTRONICS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010716289.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-11-06
Filing Date
2020-07-23
Publication Date
2024-12-06
Estimated Expiration
2040-07-23

AI Technical Summary

Technical Problem

In semiconductor device manufacturing, as process integration increases, the distance between mask image patterns becomes very close, resulting in interference and diffraction of light, resulting in layout deformation, existing optical proximity corrections are difficult to maintain consistency, and the calculation bias is calculated The value setting process takes a long time.

Method used

By dividing the layout of the semiconductor device into multiple zones and calculating hash values ​​and bias values ​​for each zone, the representative values ​​are calculated based on these values, applied to segments with the same hash values ​​for more uniform optical Proximity correction and consistency.

Benefits of technology

It effectively reduces layout deformation, improves the reliability of semiconductor devices, and shortens the bias value calculation time through the optimization calculation process, improving the efficiency and consistency of optical proximity correction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112668272B_ABST
    Figure CN112668272B_ABST
Patent Text Reader

Abstract

The present application discloses a method for maintaining consistent optical proximity correction and a method for manufacturing a mask using the same. According to the present application, a computer-readable medium includes program code, which, when executed by a processing circuit, causes the processing circuit to divide the layout of a semiconductor chip into a plurality of slices, generate a plurality of segments from the layout of each of the plurality of slices, wherein a first slice of the plurality of slices includes a first segment and a second slice of the plurality of slices includes a second segment, calculate hash values ​​corresponding to the first segment and the second segment respectively by using a hash function, calculate an offset value of a segment having a first hash value among the first segments, calculate a representative value based on the offset value, and apply the representative value to the segment having the first hash value among the first segments.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to U.S. Patent Provisional Application No. 62 / 915,109, filed on October 15, 2019, and Korean Patent Application No. 10-2019-0141095, filed on November 6, 2019, in the Korean Intellectual Property Office, the disclosure of each of which is incorporated herein by reference. Technical Field

[0003] Example embodiments of the inventive concepts described herein relate to a semiconductor process, and more particularly, to a method of optical proximity correction in which consistency is maintained and / or a method of manufacturing a mask by using the same. Background Art

[0004] With the rapid development of the electronics industry, there is a growing demand for electronic devices with higher performance, higher reliability and / or smaller size. Electronic devices are implemented using, for example, semiconductor devices manufactured by semiconductor manufacturing processes. Therefore, in order to meet the requirements / demands, the structure of semiconductor devices may gradually become complex and / or may be highly integrated.

[0005] At least one photolithography process may be used in the manufacture of semiconductor devices. For example, a layout of one of the layers including various patterns may be printed on a semiconductor substrate such as a semiconductor wafer by the photolithography process. However, as the integration of semiconductor processes increases, the distance between the image patterns of the mask becomes very close. Due to this "closeness", interference and / or diffraction of light may occur, and a deformed layout different from the target layout may be printed on the substrate.

[0006] In order to prevent or reduce the amount of deformation of the layout, a resolution enhancement technique such as optical proximity correction (OPC) can be used, and the mask can include OPC features such as serifs to reduce the amount of deformation on the layout. However, in the case of performing OPC on patterns having the same shape and the same surrounding conditions, it is very important to maintain consistency in terms of the reliability of the semiconductor device. Alternatively or additionally, the simulation process for generating bias values ​​takes a lot of time. Therefore, it is very important to effectively perform optical proximity correction while maintaining consistency. Summary of the invention

[0007] Some example embodiments of the inventive concepts provide an optical proximity correction method that makes it possible to maintain consistency with respect to patterns having the same environmental conditions.

[0008] Alternatively or additionally, some example embodiments of the inventive concepts provide a method for manufacturing a mask generated by using optical proximity correction, which may enable consistency.

[0009] According to some example embodiments, a non-transitory computer-readable medium includes program code that, when executed by at least one processing circuit, causes the at least one processing circuit to divide a layout of a semiconductor chip into a plurality of patches, generate a plurality of segments from the divided layout of each of the plurality of patches, wherein a first patch of the plurality of patches includes a first segment, and a second patch of the plurality of patches includes a second segment, calculate a hash value corresponding to the first segment and a hash value corresponding to the second segment, calculate the hash value by using a hash function, calculate an offset value of a subset of the first segments having the first hash value, calculate a representative value based on the offset value, and apply the representative value to a subset of the segments having the first hash value among the hash values ​​of the first segments. The hash function depends on at least one of a first characteristic value of each of the plurality of segments, a second characteristic value of at least one segment adjacent to each segment, or a third characteristic value between each segment and at least one segment.

[0010] According to some example embodiments, a computer-readable medium includes non-transitory program code that, when executed by at least one processing circuit, causes the at least one processing circuit to generate a plurality of segments from a layout of a semiconductor device, calculate a hash value for each of the plurality of segments by using a hash function that depends on at least one of a first characteristic value of each segment, a second characteristic value of at least one segment adjacent to each segment, or a third characteristic value between each segment and at least one segment, calculate bias values ​​for the plurality of segments, respectively, calculate representative values ​​from the calculated hash values ​​based on the bias values ​​of a subset of the plurality of segments having the same hash value, and apply the representative values ​​to the plurality of segments having the same hash value.

[0011] According to some example embodiments, a method for manufacturing a semiconductor device includes generating a plurality of segments from a layout of the semiconductor device, calculating a hash value of each of the plurality of segments by using a hash function depending on a first characteristic value of each segment, a second characteristic value of at least one segment adjacent to each segment, and a third characteristic value between each segment and the at least one segment, respectively, for bias values ​​of the plurality of segments, calculating a representative value from the calculated hash values ​​based on the bias values ​​of the plurality of segments having the same hash value, generating a bias mask depending on the representative value, and forming a pattern on a substrate by using the bias mask. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other objects and features of the present inventive concept will become apparent by describing in detail exemplary embodiments thereof with reference to the accompanying drawings.

[0013] Figure 1 is a flowchart illustrating a method for designing and manufacturing a semiconductor device according to some example embodiments of the inventive concept.

[0014] Figure 2 A photolithography system for manufacturing a mask according to some example embodiments of the inventive concepts is conceptually illustrated.

[0015] Figure 3 A layout formed on a substrate is conceptually shown.

[0016] Figure 4 A process of dividing an outline of a layout into multiple segments in optical proximity correction is conceptually illustrated.

[0017] Figure 5 A design layout updated by optical proximity correction is conceptually shown.

[0018] Figure 6 A system for performing optical proximity correction according to some example embodiments of the inventive concepts is illustrated.

[0019] Figure 7 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0020] Figure 8 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0021] Fig. 9 Elements / factors that determine a hash value of a segment according to some example embodiments of the inventive concept are conceptually illustrated.

[0022] Fig.10 Elements / factors that determine a hash value of a segment according to some example embodiments of the inventive concept are conceptually illustrated.

[0023] Fig.11 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0024] Fig.12 It is conceptually shown how representative values ​​are calculated from offset values ​​for segments of multiple patches.

[0025] Fig.13 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0026] Fig.14 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0027] Fig.15 The tasks assigned by the master device are conceptually shown.

[0028] Fig.16 is a block diagram illustrating an apparatus for generating a mask generated by optical proximity correction of the present inventive concept. DETAILED DESCRIPTION

[0029] Below, embodiments of the inventive concept may be described in detail and clearly to such an extent that a person having ordinary skill in the art can easily realize the inventive concept.

[0030] Components described in the detailed description with reference to the terms "unit", "module", "block", "device", "device" or "machine", etc. and the functional blocks shown in the drawings, for example, as described below, the master device 110 and each of the at least one slave device 121 to 12n, and the mask generation device 2000 can be implemented with processing circuits including software, hardware, or a combination thereof. For example, the software can be or can include machine code, firmware, embedded code, and / or application software. For example, the hardware can be or can include a circuit, an electronic circuit, a processor, a computer, an integrated circuit, a central processing unit (CPU), an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a system on a chip (SoC), a programmable logic unit, a microprocessor, an application specific integrated circuit (ASIC), an integrated circuit core, a pressure sensor, an inertial sensor, a microelectromechanical system (MEMS), a passive component, or a combination thereof.

[0031] Figure 1 is a flowchart illustrating a method for designing and processing / manufacturing a semiconductor device according to some example embodiments of the inventive concept.

[0032] refer to Figure 1 , in operation S11, a high-level design of a semiconductor integrated circuit may be performed. High-level design may mean and / or correspond to a description of a design-targeted integrated circuit with a high-level computer language. For example, a high-level language such as C language may be used. Alternatively or additionally, other hardware description languages ​​such as VHDL and / or Verilog may be used in the high-level language; however, the example embodiments are not limited thereto. The circuit designed by the high-level design may be expressed in more detail by register transfer level (RTL) coding and / or simulation. In addition, the code generated by the RTL coding may be converted into a netlist so as to be synthesized as a part of a semiconductor device or the entire semiconductor device. The synthesized schematic circuit may be verified by a simulation tool, and adjustment processing may be performed depending on the verification result.

[0033] In operation S12, for example, a layout of a single level (layer) or multiple levels (layers) for implementing a semiconductor integrated circuit on a silicon substrate may be designed. For example, the layout design may be performed with reference to a schematic circuit synthesized in a high-level design and / or a netlist corresponding to the schematic circuit. The layout design may include a process for placing various standard cells provided from a cell library, and a wiring process for connecting the placed cells. The layout design may be repeated for each level / layer associated with the processing / manufacturing of the semiconductor integrated circuit.

[0034] The cell library for layout design can include information about the operation, speed and / or power consumption of standard cells. In most layout design tools, a cell library for expressing a specific gate-level circuit using a layout can be defined. Layout can be a process that defines the size and / or shape of patterns associated with electrical components such as transistors, polysilicon and / or metal lines that will actually be formed on a silicon substrate and / or interconnections such as contacts and vias. Additional components such as interlayer insulating materials can be associated with the layout of the layer. For example, in order to manufacture an inverter circuit on a silicon substrate, layout patterns such as PMOS transistors, NMOS transistors, N wells, gate electrodes and / or metal lines arranged thereon can be appropriately arranged. For this reason, first, one of the inverters defined as a priori in the cell library can be found and selected.

[0035] In addition, routing can be performed on the selected and placed standard cells. For example, routing using upper layer lines can be performed on the selected and placed standard cells. The routing process can interconnect standard cells suitable for the design. Most of the above series of processes can be automatically performed by a layout design tool, or can be at least partially performed manually. For example, the placement and routing of standard cells can be automatically performed by using a separate placement and routing (P&R) tool. The P&R tool can be implemented in hardware and / or software.

[0036] After routing, the design can be verified to determine whether there are any parts that violate the design rules. Verification can include: design rule check (DRC) to verify whether the components of the layout are properly placed and / or sized for the given design rules; electrical rule check (ERC) to verify whether the components of the layout are connected without internal electrical disconnects; layout vs schematic (LVS) to determine whether the layout matches the gate-level netlist, etc. Each of the DRC, ERC, and LVS tools, or at least some of them, can be implemented in hardware.

[0037] In operation S13, optical proximity correction (OPC) may be performed. The layout pattern obtained by the layout design may be implemented on a silicon substrate by using a photolithography process. In this case, OPC may refer to a technique for correcting or reducing deformation caused by the photolithography process. For example, optical proximity correction may be used to correct or improve deformation such as refraction caused by the characteristics of light in an exposure process using the designed layout pattern. When performing optical proximity correction, the shape and / or position of the designed layout pattern may be finely corrected (or biased). The optical proximity correction conceived by the present invention will be described in detail later.

[0038] In operation S14, a photomask may be manufactured based on the layout changed by the optical proximity correction. Typically, the photomask may be manufactured in a manner of depicting a layout pattern by using a chrome thin layer applied to a glass substrate. Typically, the photomask may correspond to a level (a layer) associated in the manufacture of a semiconductor device; however, example embodiments are not limited thereto.

[0039] In operation S15, the semiconductor device may be processed (e.g., manufactured) by using a photomask. In the process of manufacturing the semiconductor device by using a photomask, various processes such as exposure and etching process and / or exposure and implantation process may be repeated. Through such processes, the shape of the pattern determined when designing the layout may be sequentially formed on the silicon substrate.

[0040] Figure 2 A photolithography system for manufacturing a mask according to some example embodiments of the inventive concept is conceptually illustrated. Figure 2 , the lithography system 1000 may include a light source 1200, a mask 1400, a reduction projection device 1600, and a wafer stage 1800. However, the lithography system 1000 may further include Figure 2 For example, the lithography system 1000 may further include a sensor for measuring the height and tilt of a surface of a substrate, such as a wafer WF.

[0041] The light source 1200 may generate light. The light irradiated from the light source 1200 may be emitted to the mask 1400. For example, the light source 1200 may include an ultraviolet light source (e.g., a KrF light source having a wavelength of 234 nm, an ArF light source having a wavelength of 193 nm, etc.). For example, the light source 1200 may further include a collimator (not shown). The collimator may convert the ultraviolet light into collimated light or parallel light. The collimated light may be provided to the mask 1400. For example, the collimator may include a dipole aperture or a quadruple aperture for making the focusing depth of the ultraviolet light higher.

[0042] The mask 1400 may include an image pattern for printing a layout on the wafer WF. The image pattern may be formed by a transparent area and an opaque area. The transparent area may be formed by etching a metal layer on the mask 1400. The transparent area may transmit light irradiated from the light source 1200. On the contrary, the opaque area may not transmit light. According to some example embodiments of the inventive concept, the mask 1400 may be manufactured based on optical proximity correction. The optical proximity correction of the inventive concept will be described in detail later.

[0043] The reduced projection device 1600 may receive light passing through the transparent area of ​​the mask 1400. The reduced projection device 1600 may be a lens or include a lens. The reduced projection device 1600 may match a circuit pattern of a layout to be printed on a substrate (e.g., a wafer WF) with an image pattern of the mask 1400. The wafer stage 1800 may support the substrate / wafer WF.

[0044] The transparent area included in the image pattern of the mask 1400 can transmit the light irradiated from the light source 1200. The light passing through the mask 1400 can be transmitted to the substrate / wafer WF through the reduced projection device 1600. The substrate / wafer WF can have a photosensitive material on the surface, such as a photoresist material, and the photosensitive material can be exposed to the light generated by the light source 1200. In addition, other layers / wafer WF on the component / substrate can be removed, for example, can be etched, or can be added, for example, can be placed. As a result, a layout including a circuit pattern corresponding to the image pattern of the mask 1400 can be patterned / printed on the substrate / wafer WF.

[0045] However, as the integration of semiconductor processes increases, the distance between the image patterns of the mask 1400 (e.g., between adjacent image patterns of the mask 1400) gradually becomes very close, and the width of the transparent area gradually becomes narrower. Due to this "closeness", interference and diffraction of light occur, and a deformed layout different from the target layout may be printed on the substrate / wafer WF. When the deformed layout is printed on the substrate / wafer WF, the designed circuit may work abnormally.

[0046] Resolution enhancement technology is used to prevent layout deformation and / or reduce the amount of deformation. Optical proximity correction is an example of resolution enhancement technology. Based on optical proximity correction, the degree of deformation such as interference and diffraction of light is estimated in advance. Alternatively or additionally, based on the estimation result, the image pattern to be formed on the mask 1400 can be pre-biased. In this way, the target layout can be printed on the substrate / wafer WF.

[0047] In the optical proximity correction of some example embodiments of the present invention, a hash value (level of a semiconductor device) of each segment constituting a layer is calculated based on the characteristics of the corresponding segment itself and the surrounding conditions. The hash value may be or correspond to a numerical value, such as a 256-bit numerical value. A representative value is calculated based on the bias value of the segment with the same hash value. The representative value may be assigned to the segment with the same hash value. Based on the representative value, the segment with the same hash value may be biased in the same or similar manner. In this way, optical proximity correction may be performed more uniformly on a pattern with the same shape and the same surrounding conditions, and a mask 1400 that is very similar to the intended mask may be provided.

[0048] Below, an example embodiment of the inventive concept will be described. In the following description, it is assumed that a model-based optical proximity correction using computer simulation is performed. However, this assumption is not intended to limit the inventive concept. A person of ordinary skill in the art may apply the example embodiment to other types of resolution enhancement techniques.

[0049] Figure 3 A layout formed on a substrate such as a wafer is conceptually illustrated. In some example embodiments, the layout L01 may include first to fourth circuit patterns R1 to R4. Figure 3 The shape of the layout L01 shown in FIG. 1 is an example for better understanding of the present inventive concept and is not intended to limit the present inventive concept.

[0050] The area defined by the solid lines may indicate a target layout to be printed / patterned on the substrate / wafer WF. For example, the designer of the layout L01 may intend to print the layout L01 of the first circuit pattern R1 onto the substrate / wafer WF. Figure 3 The solid line defines a fourth circuit pattern R4. Figure 3The solid line in indicates a layout to be printed / patterned as a target layout. The target layout is provided as an initial design layout.

[0051] In contrast, the area defined by the dotted lines may indicate the layout to be actually printed on the substrate / wafer WF. Deformations such as interference and diffraction of light may actually be caused by Figure 2 The mask 1400 is different from the designer's intention, due to deformation, Figure 3 The first to fourth circuit patterns R1 to R4 defined by the dotted lines may be printed on the substrate / wafer WF. When the deformed layout is printed on the substrate / wafer WF, the designed circuit may function abnormally.

[0052] In order to prevent layout deformation or reduce the amount of deformation, optical proximity correction can be performed. In optical proximity correction, the design layout can be biased to reduce the error between the actual layout to be actually printed and the target layout based on the design layout. Figure 4 and 5 An example of optical proximity correction is described.

[0053] Figure 4 Conceptually illustrates the process of dividing the outline of a layout into multiple segments in optical proximity correction. In some example embodiments, the process of dividing the outline of a layout into multiple segments is described. Figure 3 The solid lines correspond to the process of dividing the outline of the design layout L01 into multiple segments.

[0054] A plurality of division points may be set on the outline of the design layout L01. For example, a first division point PD_1 and a second division point PD_2 may be set on the outline of the design layout L01. For example, a segment may be obtained based on the first division point PD_1 and the second division point PD_2. Other segments, such as segment SEG1, segment SEG2, and segment SEG3, may be obtained from other division points. As described above, the outline of the design layout L01 may be divided into a plurality of segments based on a plurality of division points. A segment may represent or correspond to a small unit, such as a minimum unit for biasing.

[0055] Here, the term "partition" is used, but partition does not mean physical division. Figure 4 , an example of physically dividing a plurality of segments is shown, but the example is provided conceptually for better understanding of the inventive concept, and the division may correspond to a logical division of the outline of the design layout L01.

[0056] In optical proximity correction, each divided segment can be or correspond to a bias target. The divided segments can be biased independently of each other. For example, a segment SEG can be biased independently of the remaining segments along one of a first direction (e.g., an outward direction corresponding to each circuit pattern of the plurality of segments) and a second direction (e.g., an inward direction corresponding to each circuit pattern of the plurality of segments). Each divided segment can be biased to reduce the error between the actual layout and the target layout. For example, the bias can include adding and / or removing serif components (or dog-ear components) to each segment.

[0057] The process of calculating the bias value is known to those of ordinary skill in the art, and therefore, additional description will be omitted to avoid repetition. Each segment can be biased based on the calculated bias value. Figure 5 An example of an updated design layout obtained based on biased segmentation is described.

[0058] Figure 5 The design layout updated by optical proximity correction is conceptually shown. In order to better understand the concept of the present invention, the following description will be given. Figure 3 The first circuit pattern R1 is updated with a first new circuit pattern R1'. The description corresponding to the second circuit pattern R2 to the fourth circuit pattern R4 is omitted.

[0059] Figure 5 The solid line may indicate the first bias pattern R1' included in the updated design layout. Figure 4 The processing described, Figure 3 The outline of the first circuit pattern R1 may be divided into a plurality of segments, and each or at least one of the divided segments may be biased. Figure 5 As shown, each segment may be biased in one of a first direction (eg, outward direction) and a second direction (eg, inward direction). For example, the biasing may include adding serifs / corners to the segment. In this way, a first biasing pattern R1' may be obtained.

[0060] Each partitioned segment can be biased to reduce the error between the actual layout and the target layout. Figure 5 The dashed line may indicate the actual layout to be actually printed based on the updated design layout. The error between the actual layout and the target layout may be reduced by biasing each divided segment. For example, the error corresponding to Figure 3 The actual layout of the region R1 around the solid line and Figure 5 The dashed line in the figure corresponds to the root mean square error between the actual layouts.

[0061] However, reference Figure 3 and 4, the left top segment, the left bottom segment, and the right top segment of the first circuit pattern R1 have the same surrounding conditions. Therefore, it is desirable that the left top segment, the left bottom segment, and the right top segment of the first circuit pattern R1 have the same bias value. However, referring to Figure 5 It can be understood that the left top segment, the left bottom segment, and the right top segment of the first biasing pattern R1 ′ are asymmetrically biased.

[0062] certainly, Figures 2 to 4 This is just an example. However, in actual optical proximity correction, there is a possibility that segments with the same surrounding conditions are biased differently due to errors in numerical calculation. Specifically, when the process of calculating the bias values ​​corresponding to the divided segments is repeated, errors may accumulate. When the design layout deformed due to the accumulated errors is printed on the substrate / wafer WF, the designed circuit may operate abnormally, which is different from the designer's intention.

[0063] Therefore, in some example embodiments of the present inventive concept, a hash value of a segment is calculated taking into account the characteristics of a specific segment and the characteristics of surrounding segments. With respect to two segments, in the case where the characteristics of the two segments are the same and the characteristics of the segments adjacent to the two segments are the same, the hash values ​​of the two segments may be the same. A representative value may be calculated from the bias values ​​of the segments having the same hash value. Therefore, optical proximity correction may be performed uniformly or more uniformly.

[0064] Figure 6 A system for performing optical proximity correction according to some example embodiments of the inventive concepts is illustrated. Figure 7 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are shown. For example, Figure 7 It can be shown in Figure 1 A detailed embodiment of the optical proximity correction performed in operation S13 of FIG. Figure 8 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated.

[0065] In some example embodiments, Figure 6 The system 100 may include a master device 110 and a plurality of slave devices 121 to 12n. For example, the master device 110 and the slave devices 121 to 12n may be implemented with separate computing devices. Alternatively or additionally, each of the master devices 110 and at least one of the slave devices 121 to 12n may be implemented by a plurality of processor cores.

[0066] Although not shown in the drawings, the system 100 may further include a storage medium in which an OPC tool for running optical proximity correction is stored. When the OPC tool is run, the system 100 may further include a memory on which the OPC tool is loaded.

[0067] In some example embodiments, the OPC tool for performing the optical proximity correction of the inventive concept may be or correspond to a computer program product including a computer-readable program code, and / or may be a computer program product including a non-transitory computer-readable medium in which the computer-readable program code is stored. Alternatively or additionally, the OPC tool for performing the optical proximity correction of the inventive concept may be a product that can be downloaded via the Internet.

[0068] Reference together Figure 6 and 7 A method for optical proximity correction of the inventive concept is described. When the host device 110 drives the OPC tool, the optical proximity correction may be run on the layout of at least one layer on the semiconductor chip "A". The host device 110 may be provided with the layout of at least one layer of the semiconductor chip "A". For example, the layout of the semiconductor chip "A" may be Figure 1 The layout design operation S12 may include a plurality of circuit patterns.

[0069] The master device 110 may divide the layout of the semiconductor chip "A" into a plurality of slices PA1 to PAn. For example, a slice may be the smallest unit processed by a slave device. The master device 110 may allocate the slices PA1 to PAn of the semiconductor chip "A" to the slave devices 121 to 12n, respectively. Each slave device may process the allocated slices. In some example embodiments, the processing of the first slice PA1 by the first slave device 121 is described.

[0070] The first slave device 121 may generate segments based on the layout included in the first parcel area PA1 (S110). Figure 4 Describes the process of generating segments.

[0071] The first slave device 121 may calculate hash values ​​of the segments of the first area PA1 (S120). For example, the hash value may reflect the characteristics of a particular segment and the characteristics of at least one segment adjacent to the segment. For example, segments having the same characteristics and the same surrounding conditions may have the same hash value, for example, may be hashed to the same hash value. Fig. 9 and 10 A more complete description of the factors that determine the hash value.

[0072] The first slave device 121 may calculate the bias value of the segment constituting the first area PA1 or associated with the first area PA1 (S130). The bias value of the segment may be obtained based on the result of simulating a grid as the minimum simulation unit of the OPC tool. For example, as the size of the grid increases, the time taken to obtain the bias value may become shorter, but fewer fine bias values ​​may be obtained. In more detail, the consistency of the segments with the same surrounding conditions may weaken. On the contrary, as the size of the grid decreases, the time taken to obtain the bias value may become longer, but a finer bias value may be obtained. In more detail, the consistency of the segments with the same surrounding conditions may be enhanced. Meanwhile, the process of calculating the bias value, such as the size of the added or deleted serif, is well known to those of ordinary skill in the art, and therefore, additional description will be omitted to avoid repetition.

[0073] The master device 110 may collect the hash value and the bias value calculated by the first slave device 121. The master device 110 may calculate a representative value based on the collected hash value and the bias value (S140). Figure 5 As described, the bias values ​​of the segments may be different even if the segments have the same surrounding conditions. Therefore, for consistency, the master device 110 may calculate a representative value from the bias values ​​of the segments having the same hash value. For example, the representative value may be, but is not limited to, the average value of the bias values ​​of the segments having the same hash value, such as some kind of central tendency measure.

[0074] The master device 110 may update the representative value corresponding to the segment having the same hash value in the library (S150). For example, the library may include hash values ​​and bias values ​​corresponding to the hash values, respectively. The master device 110 may update the bias value of the segment having a unique hash value (not matching any other hash value) in the library. The master device 110 may generate a bias pattern based on the library.

[0075] The same bias value (e.g., representative value) is applied to the segments having the same hash value among the segments of the corresponding layer of semiconductor chip "A". Therefore, the consistency or higher consistency between the patterns having the same surrounding conditions in multiple slices of semiconductor chip "A" can be maintained. However, the inventive concept is not limited to the example of maintaining consistency between the same chip or multiple slices of the same layer in the same chip. For example, the consistency between semiconductor chips can be maintained. For example, the above processing can be similarly applied to semiconductor chip "B" to semiconductor chip "Z", and thus, the consistency between semiconductor chips can be maintained. Reference will be made to Figure 6 and 8 Describe this.

[0076] refer to Figure 6 and 8, the library may be updated according to the optical proximity correction performed on the semiconductor chip 'A' (S150).

[0077] Thereafter, in order to perform optical proximity correction on the semiconductor chip "B", the layout of the semiconductor chip "B" may be provided to the master device 110. The master device 110 may divide the layout of the semiconductor chip "B" into a plurality of tiles PA1 to PAn. The master device 110 may allocate the tiles PA1 to PAn of the semiconductor chip "B" to the slave devices 121 to 12n, respectively.

[0078] The slave devices 121 to 12n may process the slices PA1 to PAn of the semiconductor chip "B", respectively. For the sake of brevity, a description is given below of how the first slave device 121 processes the first slice PA1. The first slave device 121 may generate a segment based on the layout included in the first slice PA1 (S210). The first slave device 121 may calculate a hash value of the segment of the first slice PA1 (S220).

[0079] The library may be searched based on the hash value calculated in operation S220 (S230). In some example embodiments, the first slave device 121 may send the calculated hash value to the master device 110. The master device 110 may compare the hash value received from the first slave device 121 with the hash value stored in the library. Based on the comparison result, the master device 110 may determine whether the hash value is stored in the library.

[0080] When the search result indicates that the hash value is pre-stored in the library, there is no need or reason to recalculate the bias value of the segment corresponding to the hash value stored in the library. Therefore, when generating a bias layout for semiconductor chip "B", the bias value stored in the library can be used without modification (S260).

[0081] On the contrary, when the search result indicates that the hash value is not pre-stored in the library, it is necessary or reasonable to recalculate the bias value of the segment corresponding to the hash value not stored in the library. For example, the master device 110 can command / guide the first slave device 121 to calculate the bias value of the corresponding segment, and the first slave device 121 can calculate the bias value (S240). Operation S240 can be performed by each of the slave devices 121 to 12n processing the slices PA1 to PAn of the semiconductor chip "B". The slave devices 121 to 12n can send the calculated bias value and the hash value corresponding to the calculated bias value to the master device 110.

[0082] The master device 110 may calculate a representative value based on the hash value and the bias value (S250). The master device 110 may update the representative value corresponding to the segment having the same hash value in the library (S270). When there is a bias layout for the generation of semiconductor chip "A" and semiconductor chip "C" to semiconductor chip "Z" and semiconductor chip "B", the updated representative value (or bias value) may be applied. For example, consistency between semiconductor chips may also be maintained.

[0083] At the same time, Figure 7 Compared to the example embodiment of Figure 8 The example embodiment of can have a relatively short library generation time. The reason may be that the process of calculating the bias value of the segment having the same hash value as one of the hash values ​​pre-stored in the library is omitted. Therefore, when calculating the bias value of the segment of the semiconductor chip "B", the size of the grid as the minimum simulation unit can be made smaller. By reducing the size of the grid, even if the simulation time increases, a trade-off between the security library generation time and the simulation time is possible.

[0084] Fig. 9 Conceptually illustrates the elements / factors that determine the hash value of a segment according to some example embodiments of the present invention. In some example embodiments, Fig. 9 The segments shown in may correspond to Figure 4 The first segment SEG1 may correspond to Figure 3 The first circuit pattern R1, the second segment SEG2 and the third segment SEG3 may correspond to Figure 3 The third circuit pattern R3.

[0085] A hash function depending on multiple elements / factors can be used to calculate the hash value of the first segment SEG1. Multiple elements / factors can be associated with the value of a specific segment and the conditions associated with the segments around the specific segment. For example, the elements / factors that affect the hash value may include a first characteristic value associated with the characteristic of the specific segment itself. The elements / factors that affect the hash value may include a second characteristic value associated with the characteristic of an adjacent segment adjacent to the specific segment. Moreover, the elements / factors that affect the hash value may include a third characteristic value associated with the characteristic between the specific segment and the adjacent segment. As a result, segments with the same surrounding conditions and the same attributes may have the same hash value.

[0086] The length L1 of the first segment SEG1 may be used as an element for determining the hash value of the first segment SEG1. The length L1 of the first segment SEG1 includes the length in the first direction D1 and / or the length in the second direction D2. Fig. 9As shown, in the case where the first segment SEG1 is composed of a component in the first direction D1 adjacent to a component in the second direction D2, each of the component in the first direction D1 and the component in the second direction D2 can independently serve as an element or factor for determining the hash value. Therefore, the direction in which each component extends can also be used as an element. Therefore, even if there are segments with the same total length, when the lengths in the first direction D1 are different or the lengths in the second direction D2 are different, the hash values ​​of these segments may be different.

[0087] The bias direction ① of the first segment SEG1 can be used as an element or factor to determine the hash value of the first segment SEG1. The bias direction can represent the order of the segments to be simulated by the OPC tool in units of grids. For example, a circuit pattern (e.g., Figure 3 The four segments of R1) can have the same bias direction. Fig. 9 Some example embodiments are shown as forming or corresponding to circuit patterns (e.g., Figure 3 The segments of R1) are sequentially biased in a clockwise direction. For example, the bias direction when the segment is biased in a clockwise direction may have a first value, and the bias direction when the segment is biased in a counterclockwise direction may have a second value. The bias direction ② of the adjacent segment may be used as an element or factor to determine the temporary hash values ​​of the second segment SEG2 and the third segment SEG3.

[0088] The distances D12 and D13 between the first segment SEG1 and the adjacent segments may be used as elements (e.g., factors) for determining the hash value of the first segment SEG1. Here, the adjacent segments may refer to segments arranged within a distance that has an influence on the first segment SEG1 associated with the optical proximity correction. In addition, the adjacent segments may refer to circuit patterns (e.g., Figure 3 For example, the adjacent segments may be referred to as the adjacent circuit patterns to which the first segment SEG1 belongs (eg, Figure 3 R1) adjacent circuit pattern (for example, Figure 3 At least a portion of a segment of R3) (e.g., SEG2 and SEG3).

[0089] Each of the lengths L2 and L3 of the adjacent segments SEG2 and SEG3 may be used as another element or factor to determine the hash value of the first segment SEG1. The lengths L2 and L3 of the adjacent segments SEG2 and SEG3 include the length in the first direction D1 and / or the length in the second direction D2. Fig. 9As shown, when the surrounding segment SEG2 or SEG3 is composed of a component in the first direction D1 and a component in the second direction D2, the component in the first direction D1 and the component in the second direction D2 can respectively and independently serve as elements or factors that determine the hash value.

[0090] The hash value of the first segment SEG1 may be calculated with the above elements / factors comprehensively considered. For example, the slave device may calculate the hash value of the first segment SEG1 through a hash function using the above elements / factors as arguments.

[0091] Fig.10 Conceptually illustrates the elements / factors that determine the hash value of a segment according to some example embodiments of the inventive concept. Fig. 9 In the embodiment of Fig.10 In an embodiment, only elements or factors associated with the segment itself are considered to calculate the temporary hash value. Afterwards, the final hash value is calculated based on the temporary hash value and the distance between the segments.

[0092] The bias direction ① and length L1 of the first segment SEG1 can be used as elements or factors to determine the temporary hash value of the first segment SEG1. In the length L1, the components in the first direction D1 and the components in the second direction D2 can independently affect the hash value. Figure 6 One of 121 to 12n) can calculate the temporary hash value of the first segment SEG1 by considering the length of the first segment SEG1 in the first direction D1 and / or the length of the first segment SEG1 in the second direction D2, and the order of the first segments SEG1 to be simulated in grid units by the OPC tool (for example, clockwise).

[0093] The bias direction ② and the length L2 of the second segment SEG2 may be used as elements or factors to determine the temporary hash value of the second segment SEG2. The slave device may calculate the temporary hash value of the second segment SEG2 by considering the length of the second segment SEG2 in the first direction D1 and / or the length of the second segment SEG2 in the second direction D2, and the order of the second segment SEG2 to be simulated by the OPC tool in grid units (e.g., counterclockwise).

[0094] The bias direction ③ and the length L3 of the third segment SEG3 may be used as elements or factors to determine the temporary hash value of the third segment SEG3. The slave device may calculate the temporary hash value of the third segment SEG3 by taking into account the length of the third segment SEG3 in the first direction D1 and / or the length of the third segment SEG3 in the second direction D2, and / or the order of the third segment SEG3 to be simulated by the OPC tool in units of grids (e.g., counterclockwise).

[0095] The final hash value of the first segment SEG1 may be calculated by considering the hash values ​​of the segments SEG1, SEG2, and SEG3, the distance D12 between the first segment SEG1 and the second segment SEG2, and the distance D13 between the first segment SEG1 and the third segment SEG3. For example, the slave device may calculate the final hash value of the first segment SEG1 by considering at least one of the temporary hash value of the second segment SEG2, the temporary hash value of the third segment SEG3, or the distances D12 and D13 to the temporary hash value of the first segment SEG1.

[0096] However, when calculating the final hash value of the first segment SEG1, in the case where the distance between the first segment SEG1 and the adjacent segment SEG2 or SEG3 exceeds a threshold, the corresponding hash value may not be applied to calculate the final hash value of the first segment SEG1. Here, the threshold may be or may correspond to a distance that has no effect on the simulation using the OPC tool.

[0097] Fig.11 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated. Fig.11 Some example embodiments show Figure 7 The processing performed by semiconductor chip "A" is shown in detail.

[0098] In operation S310, the master device 110 may divide the layout of the semiconductor chip "A" into a plurality of slices PA1 to PAn. Figure 4 As described, the master device 110 may allocate the divided tiles PA1 to PAn to the plurality of slave devices 121 to 12n, respectively. In operations S330 to S350, each slave device may process the allocated tiles.

[0099] In some example embodiments, the first slave device 121 may generate a plurality of segments from the first slice PA1 (S331). The first slave device 121 may respectively calculate hash values ​​of the generated segments (S341). For example, the first slave device 121 may be as described in reference Fig. 9 The hash value of each segment is calculated in various ways as described in or 10. The above operation can be similarly performed by each of the second slave device 122 to the nth slave device 12n.

[0100] In operation S351, the first slave device 121 may calculate an offset value for each segment. The offset value may be obtained based on the result that each of the simulation grids is or corresponds to the minimum simulation unit of the OPC tool. Each of the second slave device 122 to the nth slave device 12n may calculate an offset value for each segment of the allocated tile.

[0101] In operation S360 , the master device 110 may collect hash values ​​and offset values ​​calculated by the slave devices 121 to 12 n .

[0102] In operation S370, the master device 110 may generate a representative value for the segments having the same hash value. For example, the representative value may be obtained from the bias values ​​of the segments having the same hash value. For example, the representative value may be or correspond to, but is not limited to, an average value, as some measure of the central tendency of the bias values ​​of the segments having the same hash value. For example, the representative value may be or correspond to, but is not limited to, the mean of the bias values ​​of the segments having the same hash value. For example, the representative value may be or correspond to, but is not limited to, the median or mode of the bias values ​​of the segments having the same hash value.

[0103] In operation S380, the library may be updated. For example, representative values ​​of offset values ​​for segments having the same hash value may be stored in the library. Offset values ​​for segments having hash values ​​that do not match any other hash values, such as offset values ​​for segments having unique hash values, may be stored in the library.

[0104] Fig.12 It is conceptually shown how to calculate the representative value from the offset values ​​of the segments of multiple patches. For ease of description, reference will be made to Figure 6 Give a description together.

[0105] The first slave device 121 generates segments SEG1 to SEGi from the first area PA1, and calculates hash values ​​HV1 to HVi corresponding to each segment in the segments SEG1 to SEGi, respectively. The second slave device 122 generates segments SEG1 to SEGj from the second area PA2, and calculates hash values ​​HV1 to HVj corresponding to each segment in the segments SEG1 to SEGj, respectively. As described above, the nth slave device 12n generates segments SEG1 to SEGk from the nth area PAn, and calculates hash values ​​HV1 to HVk corresponding to each segment in the segments SEG1 to SEGk, respectively.

[0106] The first slave device 121 may calculate bias values ​​BV1 to BVi corresponding to the segments SEG1 to SEGi, respectively. The second slave device 122 may calculate bias values ​​BV1 to BVj corresponding to the segments SEG1 to SEGj, respectively. As described above, the nth slave device 12n may calculate bias values ​​BV1 to BVk corresponding to the segments SEG1 to SEGk, respectively.

[0107] The master device 110 may collect the hash value and the bias value of each slice, and may calculate the representative value from the bias value of the segment having the same hash value. In some example embodiments, it may be assumed that the hash value corresponds to the hash value HV1 of the first segment SEG1 of the first slice PA1, the hash value HV2 of the second segment SEG2 of the second slice PA2, and the hash value HVk of the kth segment SEGk of the nth slice PAn. Here, the representative value may be or may correspond to, but is not limited to, a mean value, a median, or a mode. The calculated representative value may be updated at the library, and / or may be commonly applied to the segments having the same hash value.

[0108] according to Fig. 9 and 10 In an example embodiment, the first segment SEG1 of the first area PA1, the second segment SEG2 of the second area PA2, and the kth segment SEGk of the nth area PAn having the same hash value can be predicted to have the same surrounding environment. Therefore, the consistency between the areas can be maintained by applying the same bias value to the segments having the same surrounding conditions, so the designed circuit can be operated normally depending on the designer's intention, and / or the efficiency of OPC can be improved. For example, segments having the same hash value can be resized in a similar manner and / or the same or similar serifs can be applied (e.g., added thereto and / or removed therefrom) in an amount corresponding to the representative value, and the consistency can be improved.

[0109] Fig.13 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated. Fig.13 Some example embodiments show Figure 8 The semiconductor chip "B" is shown to perform the processing.

[0110] In operation S410, the master device 110 may divide the layout of the semiconductor chip "B" into a plurality of slices PA1 to PAn. Figure 4 As described, the master device 110 may allocate the divided tiles PA1 to PAn to the plurality of slave devices 121 to 12n, respectively. In operations S430 to S440, each slave device may process the allocated tiles.

[0111] In some example embodiments, the first slave device 121 may generate a plurality of segments from the first slice PA1 (S431). The first slave device 121 may respectively calculate hash values ​​of the generated segments (S441). For example, the first slave device 121 may be as described in reference Fig. 9The hash value of each segment is calculated in various ways as described in or 10. The above operations can be similarly performed by the second slave device 122 to the nth slave device 12n.

[0112] In operation S450, the master device 110 may receive hash values ​​from the slave devices 121 to 12n and may search for a library based on the hash values. Here, the library may include a library associated with the semiconductor chip "A" (e.g., Fig.11 The master device 110 may compare the hash values ​​received from the slave devices 121 to 12n with the hash values ​​stored in the library, and may determine whether optical proximity correction is required for each segment of the semiconductor chip "B".

[0113] When the search result indicates that there is a value matching the hash value received from the device 121 to 12n in the library, it is not necessary / desirable to calculate the bias value of the segment corresponding to the matched hash value. Conversely, when the search result indicates that the value matching the hash value received from the device 121 to 12n does not exist in the library, it is necessary / desirable to calculate the bias value of the segment corresponding to the corresponding hash value.

[0114] In operation S460, the master device 110 may send the search result in operation S450 to the slave devices 121 to 12n. For example, referring to the search result, the first slave device 121 may calculate the bias value of the segment not stored in the library among the segments of the first area PA1 (S461). As described above, the remaining slave devices 122 to 12n may calculate the bias value of the segment (S462 to S46n).

[0115] In operation S470 , the master device 110 may collect the bias values ​​calculated by the slave devices 121 to 12 n .

[0116] In operation S480, the master device 110 may generate a representative value based on the segments having the same hash value. For example, the representative value may be obtained from the bias value of the segments having the same hash value. For example, the representative value may be or may correspond to or may be based on at least one of the mean, median, or mode of the bias value of the segments.

[0117] In operation S490, the library may be updated. For example, representative values ​​of offset values ​​of segments having the same hash value may be stored in the library, and offset values ​​of segments having hash values ​​(e.g., unique hash values, such as values ​​that do not match any other hash values) may be stored in the library.

[0118] Summarize Fig.12 and 13In an example embodiment, a library is generated by calculating bias values ​​(including representative values) for segments of semiconductor chip "A". Here, consistency between slices of semiconductor chip "A" is maintained by applying a common representative value to segments having the same hash value. Alternatively or additionally, consistency between semiconductor chips is maintained by applying the representative values ​​stored in the library in common to semiconductor chips "B" to "Z". For example, segments having the same hash value may be resized in a similar manner and / or may have the same or similar serifs applied (e.g., added thereto and / or removed therefrom) applied thereto in an amount corresponding to the representative value, and consistency may be improved.

[0119] Meanwhile, since each slave device calculates the offset values ​​of all segments of the corresponding patch, the time taken to perform optical proximity correction may be excessively long. Therefore, another method for reducing the time taken to calculate the offset values ​​may be adopted. Fig.14 and 15 Describe this.

[0120] Fig.14 Methods for performing optical proximity correction according to some example embodiments of the inventive concept are illustrated. Fig.15 The tasks assigned by the master device are conceptually shown.

[0121] refer to Fig.14 , the master device 110 may divide the layout of the semiconductor chip "A" into a plurality of slices PA1 to PAn (S510), and may respectively allocate the divided slices PA1 to PAn to each of the plurality of slave devices 121 to 12n (S520). Each slave device may generate a segment from the corresponding slice (S531 to S53n), and may calculate a hash value of the generated segment (S541 to S54n). The slave devices 121 to 12n may send the calculated hash value to the master device 110.

[0122] The master device 110 may assign a task to each slave device based on the hash value (S560). Here, the task may mean obtaining an offset value of a segment having the same hash value (eg, a hash value that matches any other hash value). Fig.15 How the master device 110 allocates tasks is described in detail together.

[0123] In some example embodiments, the first slave device 121 generates segments SEG1 to SEGi from the first area PA1, and calculates hash values ​​HV1 to HVi corresponding to the segments SEG1 to SEGi, respectively. The second slave device 122 generates multiple segments SEG1 to SEGj from the second area PA2, and calculates hash values ​​HV1 to HVj corresponding to the segments SEG1 to SEGj, respectively. As described above, the nth slave device 12n generates multiple segments SEG1 to SEGk from the nth area PAn, and calculates hash values ​​HV1 to HVk corresponding to the segments SEG1 to SEGk, respectively.

[0124] Here, it is assumed that the hash values ​​HV1 and HV3 of the first area PA1, the hash value HV2 of the second area PA2, and the hash value HVk of the nth area PAn are the same (hereinafter referred to as the "first hash value"). It is further assumed that the hash value HV2 of the first area PA1, the hash values ​​HV1 and HVj of the second area PA2, and the hash value HV3 of the nth area PAn are the same (hereinafter referred to as the "second hash value"). It is further assumed that the hash value HVi of the first area PA1, the hash value HV3 of the second area PA2, and the hash values ​​HV1 and HV2 of the nth area PAn are the same (hereinafter referred to as the "third hash value").

[0125] according to Fig.11 In the exemplary embodiment of FIG. 1 , since each slave device calculates the offset values ​​of all segments, the time taken to perform optical proximity correction may be longer. Fig.15 In an example embodiment, the master device 110 assigns a task to a slave device responsible for a slice where the number of segments having the same hash value is the largest. Here, the task may mean calculating the offset value of the segments having the same hash value and calculating the representative value based on the calculated offset value.

[0126] For example, since the slice with the largest number of segments having the first hash value is the first slice PA1, the master device 110 may assign the task "task A" so that the first slave device 121 responsible for the first slice PA1 calculates the offset values ​​of the segments SEG1 and SEG3 of the first slice PA1. The master device 110 may not assign the task of calculating the offset value of the second segment SEG2 of the second slice PA2 to the second slave device 122. As described above, the master device 110 may not assign the task of calculating the offset value of the kth segment SEGk of the nth slice PAn to the nth slave device 12n.

[0127] As described above, because the slice with the largest number of segments having the second hash value is the second slice PA2, the master device 110 can assign the task "task B" so that the second slave device 122 calculates the offset values ​​of the segments SEG1 and SEGj of the second slice PA2. As described above, the master device 110 can process the segments having the third hash value.

[0128] return Fig.14 , in operation S570, the slave devices 121 to 12n may calculate the bias value of the segment. For example, the calculation of the bias value may include calculating the bias value associated with the task assigned by the master device 110. That is, the task assigned from the master device 110 may include calculating the bias value of multiple segments having the same hash value. Alternatively or additionally, the calculation of the bias value may include calculating the bias value of the segment having a unique hash value (not matching any other hash value) that is not associated with the task assigned from the master device 110. However, the slave device may not be able to calculate the bias value associated with the task assigned to another slave device.

[0129] refer to Fig.15 For example, the first slave device 121 may calculate the bias value of the segment with the first hash value. However, the first slave device 121 may not calculate the bias value associated with the task assigned to the second slave device 122 (i.e., calculate the bias value of the segment with the second hash value). The first slave device 121 may not calculate the bias value associated with the task assigned to the nth slave device 12n (i.e., calculate the bias value of the segment with the third hash value). That is, the first slave device 121 does not calculate the bias values ​​of the segments SEG2 and SEGi. Of course, the first slave device 121 can calculate the bias value associated with the task not assigned to any other slave device.

[0130] As described above, the second slave device 122 may calculate the hash values ​​of the segments SEG1 and SEGj having the second hash value, and may not calculate the hash values ​​of the segments SEG2 and SEG3 assigned to any other slave device. The nth slave device 12n may calculate the hash values ​​of the segments SEG1 and SEG2 having the third hash value, and may not calculate the hash values ​​of the segments SEG3 and SEGk assigned to any other slave device.

[0131] The master device 110 may collect the bias values ​​calculated by the slave devices 121 to 12 n ( S580 ), and may calculate a representative value based on the collected bias values ​​( S590 ).

[0132] For example, the master device 110 may calculate a representative value to be commonly applied to segments having a first hash value based on a bias value received from the first slave device 121. The master device 110 may calculate a representative value to be commonly applied to segments having a second hash value based on a bias value received from the second slave device 122. As described above, the master device 110 may calculate a representative value to be commonly applied to segments having a third hash value based on a bias value received from the nth slave device 12n.

[0133] The master device 110 may update the calculated representative value in the library. In addition, the master device 110 may update the bias value of the segment having a unique hash value (e.g., a hash value that does not match any other hash value) in the library. The representative value and the bias value of the updated library may be commonly applied to any other semiconductor chips and the slices PA1 to PAn. As a result, the consistency of the biased mask may be maintained.

[0134] Fig.16 is a block diagram showing an apparatus for generating a mask generated by optical proximity correction of the present inventive concept. Fig.16 , the mask generation device 2000 may include a processor 2100, a memory / storage 2200, and a user interface 2300. The mask generation device 2000 may be used to generate a mask according to a reference Figures 2 to 15 The described embodiments of the inventive concepts produce the mask 1400 .

[0135] The processor 2100 may include at least one of a general-purpose processor and a dedicated processor such as a workstation processor. The processor 2100 may perform various arithmetic operations and / or logical operations to perform reference Figures 3 to 15 The operations described above include layout partitioning, segment generation, hash value calculation, offset value calculation, representative value calculation, and updating. To this end, the processor 2100 may include one or more processor cores. For example, the processor core of the processor 2100 may include a dedicated logic circuit, such as a field programmable gate array (FPGA) or an application specific integrated chip (ASIC).

[0136] The memory / storage 2200 may temporarily or semi-permanently store data processed or to be processed by the processor 2100. To this end, the memory / storage 2200 may include at least one of a volatile memory such as a dynamic random access memory (DRAM), a static RAM (SRAM), a synchronous DRAM (SDRAM), etc., and / or a non-volatile memory such as a flash memory, a phase change RAM (PRAM), a magnetoresistive RAM (MRAM), a resistive RAM (ReRAM), a ferroelectric RAM (FRAM), etc.

[0137] Reference may be implemented depending on the operations of the processor 2100 and the memory / storage 2200. Figures 3 to 15 The mask generating apparatus 2000 may be used to generate the mask 1400 depending on the operations of the processor 2100 and the memory / storage 2200 .

[0138] The mask generating device 2000 may execute software depending on the operation of the processor 2100 and the memory / storage 2200. For example, the software may include an operating system (OS) and / or one or more application programs. The operating system may provide one or more services to the application programs and may operate as an intermediary between components of the mask generating device 2000. For example, according to reference Figures 3 to 15 According to the described exemplary embodiments of the inventive concept, the application program may include a program for designing a layout.

[0139] The user interface 2300 may provide the results obtained by the operation of the processor 2100 and the memory / storage 2200 to the user 10, such as a physical design engineer (PHY engineer). In addition, the user interface 2300 may be used to receive various data (e.g., data associated with the design layout) from the user 10. For example, the user 10 may be a designer of the mask 1400 and the layout. For example, the user interface 2300 may include an input / output interface, such as a display device, a speaker, a keyboard, or a mouse.

[0140] The mask generating device 2000 can be based on the reference Figures 3 to 15 The described embodiments of the inventive concept output a final updated design layout FU. The mask generation device 2000 may generate a mask 1400 based on the final updated design layout FU. The mask 1400 may be generated to include an image pattern corresponding to the final updated design layout FU. The mask generation device 2000 may be or may include an electron beam (e-beam) writer that forms a pattern of a photomask (such as a blank photomask). For example, the mask generation device 2000 may etch chrome-on-glass to form a pattern on the photomask.

[0141] Although according to Figure 8 and 9Example embodiments may improve OPC and consistency of chip-to-chip OPC designs across multiple chips, however example embodiments are not limited thereto. For example, according to some example embodiments, OPC processing for semiconductor chips having similar designs (such as semiconductor chip "A" to semiconductor chip "Z") may be improved, and / or consistency may be improved. In addition to semiconductor chips having similar designs, individual layers within a particular semiconductor chip may have similar elements / features, and according to some example embodiments, OPC may be improved. For example, semiconductor chip "A" may have two, three, four, or more metal layers stacked on top of each other. Certain metal layers, such as a first metal layer and a second metal layer, may have segments with similar lengths and / or distances, as shown in reference to FIG. Fig. 9 and 10 Thus, in addition to example embodiments that enable inter-chip OPC to be performed more consistently, example embodiments may also enable intra-chip OPC to be performed more consistently.

[0142] According to some example embodiments, with respect to each segment generated from the layout, a hash value is calculated taking into account the characteristics of the segment itself and the surrounding segments. In addition, a representative value of the offset value of the segment having the same hash value is calculated. The calculated representative value is commonly applied to the segment having the same hash value. Therefore, consistency between slices or semiconductor chips is maintained, and / or the efficiency of the OPC process is improved.

[0143] While the inventive concept has been described with reference to exemplary embodiments thereof, it will be apparent to those skilled in the art that various changes and modifications may be made thereto without departing from the spirit and scope of the inventive concept as set forth in the appended claims.

Claims

1. A non-transitory computer readable medium comprising program code, which, when executed by at least one processing circuit, causes the at least one processing circuit to perform the following operations: Dividing the layout of the semiconductor chip into a plurality of slices; Generating a plurality of segments from the layout of each of the divided plurality of patches, in, A first patch of the plurality of patches includes a first segment, and a second patch of the plurality of patches includes a second segment; calculating a hash value corresponding to the first segment and a hash value corresponding to the second segment, by calculating the hash values ​​using a hash function; calculating a bias value for a subset of the first segments having the first hash value; calculating a representative value based on the bias value; as well as applying the representative value to a subset of the segments having the first hash value among the hash values ​​of the first segment, The hash function depends on at least one of a first characteristic value of each of the multiple segments, a second characteristic value of at least one segment adjacent to each segment, or a third characteristic value between each segment and the at least one segment.

2. The computer-readable medium according to claim 1, in, The representative value is based on at least one of a mean, a median, or a mode, and the mean, the median, and the mode are based on bias values ​​of the segments having the first hash value.

3. The computer-readable medium according to claim 1, in, The program code, when executed by at least one processing circuit, further causes the at least one processing circuit to: Update the representative value in the library.

4. The computer-readable medium according to claim 1, in, The program code, when executed by at least one processing circuit, further causes the at least one processing circuit to: An offset value of another segment having another hash value among the hash values ​​of the first segment that does not match any other hash value among the hash values ​​of the first segment is calculated.

5. The computer-readable medium according to claim 4, in, The program code, when executed by at least one processing circuit, further causes the at least one processing circuit to: The offset value of the another segment in the library is updated.

6. The computer-readable medium of claim 1, in, The program code, when executed by at least one processing circuit, further causes the at least one processing circuit to: The representative value is applied to a segment having the first hash value among the second segments.

7. The computer-readable medium of claim 1, in, At least one of the following: (a) the first characteristic value comprises at least one of the length or the offset direction of each segment, (b) the second characteristic value comprises the length of the at least one segment, or (c) The third characteristic value includes a distance between each of the segments and the at least one segment.

8. The computer-readable medium of claim 7, in, In response to the length of each segment extending in the first direction and the second direction, The hash function is independently based on each of a component of a first length of each segment in a first direction and a component of a second length of each segment in a second direction.

9. The computer-readable medium of claim 7, in, In response to a length of the at least one segment extending in the first direction and the second direction, The hash function is independently based on each of a component of the at least one segment in a first direction and a component of the at least one segment in a second direction.

10. The computer-readable medium of claim 7, in, When calculating the bias value of each segment, the bias direction indicates the order to be simulated in units of grids.

11. A computer readable medium comprising non-transitory program code, which, when executed by at least one processing circuit, causes the at least one processing circuit to: generating a plurality of segments from a layout of a semiconductor device; calculating a hash value of each of the plurality of segments by using a hash function, the hash function depending on at least one of a first characteristic value of each segment, a second characteristic value of at least one segment adjacent to each segment, or a third characteristic value between each segment and the at least one segment; Calculate bias values ​​for multiple segments respectively; calculating a representative value based on the bias value of a subset of the plurality of segments, the subset of the plurality of segments having the same hash value among the calculated hash values; as well as Apply a representative value to multiple segments with the same hash value.

12. The computer readable medium of claim 11, in, The first characteristic value includes at least one of the length or the offset direction of each segment, The second characteristic value comprises the length of the at least one segment, and The third characteristic value includes a distance between each of the segments and the at least one segment.

13. The computer readable medium of claim 12, in, In response to the length of each segment extending in the first direction and the second direction, The hash function is independently based on each of a component of a first length of each segment in a first direction and a component of a second length of each segment in a second direction.

14. The computer readable medium of claim 12, in, In response to a length of the at least one segment extending in the first direction and the second direction, The hash function is independently based on each of a component of the at least one segment in a first direction and a component of the at least one segment in a second direction.

15. The computer readable medium of claim 12, in, When calculating the bias value of each segment, the bias direction indicates the order to be simulated in units of grids.

16. The computer readable medium of claim 11, in, Calculating the representative value includes: Calculate the average of the bias values ​​for multiple segments with the same hash value.

17. A method for manufacturing a semiconductor device, the method include: generating a plurality of segments from a layout of a semiconductor device; calculating a hash value of each of the plurality of segments by using a hash function, the hash function depending on a first characteristic value of each segment, a second characteristic value of at least one segment adjacent to each segment, or a third characteristic value between each segment and the at least one segment; Calculate bias values ​​for multiple segments respectively; calculating a representative value based on bias values ​​of a plurality of segments having the same hash value among the calculated hash values; Generate a biased mask depending on the representative value; as well as A pattern is formed on the substrate by using a biased mask.

18. The method according to claim 17, in, The first characteristic value includes at least one of the length or the offset direction of each segment, The second characteristic value includes the length of the at least one segment, and The third characteristic value includes the distance between each segment and the at least one segment.

19. The method according to claim 18, in, When the length of each segment extends in the first direction and the second direction, each of a component of a first length of each segment in the first direction and a component of a second length of each segment in the second direction independently affects the hash function.

20. The method according to claim 17, in, Calculating the representative value includes calculating an average value of bias values ​​of a plurality of segments having the same hash value.