Exposure method, data processing method, and exposure apparatus

Abstract

To provide a technique of irradiating an object to be exposed with a light beam modulated based on drawing data, and capable of performing smoother pattern drawing by a simple data processing.SOLUTION: An exposure method of irradiating an object to be exposed with a light beam modulated based on drawing data comprises: a data creating process of creating raster data obtained by rasterizing a pattern by a target grid size based on design data showing a pattern to be drawn; and an exposure process of controlling a light modulator on the basis of raster data to modulate a light beam corresponding to a pattern, and irradiating an object to be exposed with the light beam. In the data creating process, design data is rasterized by a temporary grid size which is formed by making a target grid size (1 / N) times (N denotes an integer of 2 or more) to create binary raster data, and one of (N+1) steps of a gradation value is allocated to one pixel of a target grid size corresponding to the total of pixel values of continuous N pixels of the temporary grid size.SELECTED DRAWING: Figure 7

Description

Technical Field

[0001] This invention relates to a technique for exposing a substrate in order to draw a pattern on a substrate such as a semiconductor substrate, a semiconductor package substrate, a printed wiring board, a glass substrate, or the like.

Background Art

[0002] As a technique for forming a pattern such as a wiring pattern on various substrates such as a semiconductor substrate, a semiconductor package substrate, a printed wiring board, and a glass substrate, a light beam modulated according to exposure data is incident on a photosensitive layer formed on the substrate surface to expose the photosensitive layer. For example, in Patent Document 1, when converting CAD data (design data) representing a pattern to be drawn as vector data into run length data suitable for exposure control (that is, rasterization), not only is the image density set to two levels of high density "1" and low density "0", but also based on the arrangement of image elements in the target area and its peripheral area, intermediate tones are realized by making it possible to set it to "0.5", which is an intermediate value between them. As a result, it is possible to reduce the stepped edges (jaggies) that appear in the pattern.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The demand for rendering quality is increasing, with a growing need for higher-resolution and smoother pattern rendering. One way to further improve pattern smoothness is to increase the resolution during rasterization, i.e., to make the grid size finer. However, the resolution used during actual exposure is limited by hardware conditions such as the structure, specifications, and operating speed of the device used.

[0005] Therefore, there are limits to increasing resolution through hardware improvements, and there is a need to improve data processing to enable smoother rendering even with the same grid size. One way to do this is to increase the number of gradation levels during rasterization. However, Patent Document 1 only describes a data processing method specialized for three-level gradation, and it is fundamentally impossible to extend it to represent more gradations.

[0006] This invention has been made in view of the above-mentioned problems, and aims to provide a technology that enables smoother pattern drawing than before with simpler data processing in an exposure technology in which a modulated light beam based on drawing data is irradiated onto an object to be exposed. [Means for solving the problem]

[0007] The exposure method according to this invention is an exposure method in which a light beam modulated based on drawing data is irradiated onto an object to be exposed, and based on design data representing the pattern to be drawn, the pattern is , in a grid of two intersecting grid directions predetermined target grid size Pixel units divided into sections The system includes a data creation step of creating raster data by rasterizing the data, and an exposure step of controlling a light modulator based on the raster data to modulate a light beam emitted from a light source according to the pattern and irradiating the object to be exposed.

[0008] In its first embodiment, In the exposure process, a light modulator capable of modulating the light beam emitted from the light source in multiple tones is controlled to modulate the light beam according to the pattern, and the resulting light beam is irradiated onto the object to be exposed. In the data creation process described above, the target grid size is Along one of the aforementioned grid directionsThe design data is rasterized with a temporary grid size that is (1 / N) times (where N is an integer greater than or equal to 2) to create binary raster data. N consecutive pixels of the temporary grid size are designated as one pixel of the target grid size, and one of the (N+1) grayscale levels is assigned as the pixel value of that one pixel, corresponding to the sum of the pixel values ​​of the N pixels.

[0009] Furthermore, in the second embodiment, when the pattern is a line pattern of a certain width, the data creation step determines the angle between the direction in which the line pattern extends and the grid setting direction during rasterization, based on the design data, and the angle of each pixel after rasterization By multiplying the grayscale value by a correction coefficient determined in advance from the relationship between the angle and the width of the line pattern, The raster data is corrected and created.

[0010] In this configuration, the degree of discrepancy between the edges of pixels constrained by the grid and the edges of the pattern defined by the design data is reflected in the grayscale value of the pixel. Therefore, even with the same grid size, the grayscale value of a pixel will differ depending on how the edges of the pattern appear.

[0011] Specifically, in the first embodiment, binary raster data is created with a temporary grid size that is (1 / N) times the target grid size, thereby dividing one pixel of the target grid size into N pixels. As the grid is subdivided, the discrepancy between the edges of the pattern represented by the design data and the edges of the pattern represented by the raster data becomes smaller. The pixel values ​​of the N subdivided pixels that make up one pixel of the target grid size are then reflected in the grayscale value of the original pixel. In other words, the pixels of the target grid size are represented in multiple grayscale levels according to the state of the subdivided pixels.

[0012] Furthermore, in the second embodiment, based on the inventor's knowledge that a discrepancy occurs between the thickness of the pattern represented by the design data and the thickness of the pattern represented by the raster data due to differences in the grid setting direction, and that the degree of this discrepancy depends on the inclination angle of the line pattern with respect to the grid lines, the grayscale value of the rasterized pixels is corrected according to the inclination angle of the pattern.

[0013] Thus, the present invention specifies a method for reflecting information about the edge portions of a pattern, which would otherwise be lost in the data due to simple rasterization, into the grayscale values ​​of the pixels. By doing so, it is possible to suppress the occurrence of jagged edges and improve the smoothness of the edges.

[0015] Furthermore, the data processing method according to the present invention One The embodiment involves obtaining the line pattern from design data representing a line pattern of a certain width to be drawn. , in a grid of two intersecting grid directions predetermined target grid size Pixel units divided into sections A data processing method for creating raster data, comprising: a step of determining the angle between the direction in which the line pattern extends and the grid setting direction during rasterization, based on the design data; a step of performing rasterization based on the design data; and a step of determining the grayscale value of each pixel after rasterization. By multiplying this by a correction coefficient determined in advance from the relationship between the angle and the width of the line pattern, The process includes a step of correcting and creating the raster data.

[0016] More details will follow later, this This data processing method allows for the degree of discrepancy between the design data's edges and the actual edges to be reflected in the grayscale value of each pixel through relatively simple data processing that does not require a large amount of computation. Therefore, it is applicable as a suitable data processing method for creating raster data as described above.

[0017] Another aspect of this invention is an exposure apparatus comprising: a holding unit for holding an object to be exposed; an exposure unit having a light modulator capable of modulating a light beam emitted from a light source in multiple shades and irradiating the object to be exposed with the modulated light beam; and a control unit that creates raster data by performing one of the above-described data processing methods based on design data representing a pattern to be drawn, and controls the light modulator based on the raster data.

[0018] In the invention configured as described above, according to the above principle, it is possible to form a smooth pattern with few jaggedness by exposure. Therefore, the substantial resolution can exceed the limitation by the grid size defined by hardware. For this reason, it is possible to achieve high definition of the pattern without relying on the improvement of hardware for finer grid size.

Advantages of the Invention

[0019] As described above, according to the present invention, in an exposure technique of irradiating an object to be exposed with a light beam modulated based on drawing data, the pixel value of each pixel after rasterization reflects the degree of deviation between the edge formed by the pixel and the edge on the design data. For this reason, it is possible to perform smoother pattern drawing than before with simple data processing.

Brief Description of the Drawings

[0020] [Figure 1] It is a front view schematically showing a schematic configuration of an exposure apparatus according to the present invention. [Figure 2] It is a block diagram showing an example of an electrical configuration included in the exposure apparatus of FIG. 1. [Figure 3] It is a diagram schematically showing an example of a detailed configuration included in an exposure head. [[ID=2)]] [Figure 4] It is a flowchart showing a process executed by this exposure apparatus. [Figure 5] It is a diagram showing the difference in patterns between vector data and raster data. [Figure 6] It is a diagram explaining the principle of the first embodiment of raster data creation processing. [Figure 7] It is a flowchart showing the first embodiment of raster data creation processing. [Figure 8] It is a diagram exemplifying the effect of the first embodiment. [Figure 9] It is a diagram showing the first modification example of the first embodiment. [Figure 10] It is a diagram showing the second modification example of the first embodiment. [Figure 11] This figure illustrates the principle of the second embodiment of the raster data creation process. [Figure 12] This is a flowchart showing a second embodiment of the raster data creation process. [Figure 13] This figure shows the method for correcting raster data in the second embodiment. [Modes for carrying out the invention]

[0021] <Device Configuration and Operation> Figure 1 is a schematic front view showing the general configuration of the exposure apparatus according to the present invention, and Figure 2 is a block diagram showing an example of the electrical configuration of the exposure apparatus in Figure 1. In Figure 1 and the following figures, the horizontal direction X, the horizontal direction Y which is perpendicular to the X direction, the vertical direction Z, and the rotation direction θ around a rotation axis parallel to the Z direction are shown as appropriate.

[0022] The exposure apparatus 1 draws a pattern on a photosensitive material by irradiating a substrate S (object to be exposed), on which a layer of photosensitive material such as a resist is formed, with laser light of a predetermined pattern. Various substrates can be used as the substrate S, such as printed circuit boards, glass substrates for various display devices, semiconductor substrates, and semiconductor package substrates.

[0023] The exposure apparatus 1 comprises a main body 11, which consists of a main body frame 111 and a cover panel (not shown) attached to the main body frame 111. Various components of the exposure apparatus 1 are arranged both inside and outside the main body 11.

[0024] The interior of the main body 11 of the exposure apparatus 1 is divided into a processing area 112 and a transfer area 113. The processing area 112 mainly houses the stage 2, stage drive mechanism 3, exposure unit 4, and alignment unit 5. Outside the main body 11, an illumination unit 6 is located to supply illumination light to the alignment unit 5. In the transfer area 113, a transfer device 7, such as a transfer robot, is located to transfer substrates S to and from the processing area 112. Furthermore, a control unit 9 is located inside the main body 11, and the control unit 9 is electrically connected to each part of the exposure apparatus 1 to control the operation of each part.

[0025] The transport device 7, located in the transfer area 113 inside the main unit 11, receives unprocessed substrates S from an external transport device or substrate storage device (not shown) and loads them into the processing area 112, and unloads processed substrates S from the processing area 112 and discharges them to the outside. Loading of unprocessed substrates S and unloading of processed substrates S are performed by the transport device 7 in accordance with instructions from the control unit 9.

[0026] Stage 2 has a flat, plate-like shape and holds the substrate S placed on its upper surface in a horizontal position. Multiple suction holes (not shown) are formed on the upper surface of Stage 2, and by applying negative pressure (suction pressure) to these suction holes, the substrate S placed on Stage 2 is fixed to the upper surface of Stage 2. Stage 2 is driven by a stage drive mechanism 3.

[0027] The stage drive mechanism 3 is an XYZ-θ drive mechanism that moves the stage 2 in the Y direction (main scanning direction), X direction (sub-scanning direction), Z direction, and rotational direction θ (yaw direction). The stage drive mechanism 3 includes a Y-axis robot 31, which is a single-axis robot extending in the Y direction; a table 32 driven in the Y direction by the Y-axis robot 31; an X-axis robot 33, which is a single-axis robot extending in the X direction on the upper surface of the table 32; a table 34 driven in the X direction by the X-axis robot 33; and a θ-axis robot 35 that drives the stage 2, which is supported on the upper surface of the table 34, in the rotational direction θ relative to the table 34.

[0028] Therefore, the stage drive mechanism 3 can drive the stage 2 in the Y direction using the Y-axis servo motor of the Y-axis robot 31, drive the stage 2 in the X direction using the X-axis servo motor of the X-axis robot 33, and drive the stage 2 in the rotational direction θ using the θ-axis servo motor of the θ-axis robot 35. These servo motors are not shown in the illustration. The stage drive mechanism 3 can also drive the stage 2 in the Z direction using the Z-axis robot 37, which is not shown in Figure 1. The stage drive mechanism 3 moves the substrate S placed on the stage 2 by operating the Y-axis robot 31, X-axis robot 33, θ-axis robot 35, and Z-axis robot 37 in response to commands from the control unit 9.

[0029] The exposure unit 4 includes an exposure head 41 positioned above the substrate S on the stage 2, and a light irradiation unit 40 which includes a light source drive unit 42, a laser emission unit 43, and an illumination optical system 44, and irradiates the exposure head 41 with laser light. Multiple exposure units 4 may be provided at different positions in the X direction.

[0030] The laser light emitted from the laser emission unit 43 by the operation of the light source drive unit 42 is irradiated onto the exposure head 41 via the illumination optical system 44. The exposure head 41 modulates the laser light emitted from the light irradiation unit using a spatial light modulator and directs it onto the substrate S moving directly below it. By exposing the substrate S with the laser light beam in this way, a pattern is drawn on the substrate S (exposure operation).

[0031] The alignment unit 5 has an alignment camera 51 positioned above the substrate S on the stage 2. This alignment camera 51 has a lens barrel, an objective lens, and a CCD image sensor, and captures alignment marks provided on the upper surface of the substrate S moving directly below it. The CCD image sensor of the alignment camera 51 is composed of, for example, an area image sensor (two-dimensional image sensor).

[0032] The illumination unit 6 is connected to the lens barrel of the alignment camera 51 via an optical fiber 61 and supplies illumination light to the alignment camera 51. The illumination light guided by the optical fiber 61 extending from the illumination unit 6 is guided through the lens barrel of the alignment camera 51 to the upper surface of the substrate S, and the reflected light from the substrate S is incident on the CCD image sensor via the objective lens. As a result, the upper surface of the substrate S is imaged and an image is acquired. The alignment camera 51 is electrically connected to the control unit 9 and acquires the image in response to instructions from the control unit 9 and transmits this image to the control unit 9.

[0033] The control unit 9 acquires the position of the alignment marks indicated by the image captured by the alignment camera 51. The control unit 9 also controls the exposure unit 4 based on the position of the alignment marks to adjust the pattern of the laser light irradiated from the exposure head 41 onto the substrate S during the exposure operation. The control unit 9 then irradiates the substrate S with modulated laser light from the exposure head 41 according to the pattern to be drawn, thereby drawing a pattern on the substrate S.

[0034] The control unit 9 performs various processes by controlling the operation of each of the above-mentioned units. For this purpose, the control unit 9 is equipped with a CPU (Central Processing Unit) 91, memory (RAM) 92, storage 93, input 94, display unit 95, and interface unit 96. The CPU 91 reads and executes a control program 931 that is pre-stored in the storage unit 93, and performs various operations described later. The memory 92 is used for calculation processing by the CPU 91, or for short-term storage of data generated as a result of calculation processing. The storage unit 93 stores various data and control programs for long-term storage. Specifically, the storage unit 93 is a non-volatile storage device such as a flash memory device or a hard disk drive device, and in addition to the control program 931 executed by the CPU 91, it stores, for example, CAD (Computer Aided Design) data 932, which is design data representing the content of the pattern to be drawn.

[0035] The input unit 94 receives user input and, for this purpose, has an appropriate input device such as a keyboard, mouse, or touch panel (not shown). The display unit 95 notifies the user by displaying and outputting various information and, for this purpose, has an appropriate display device such as a liquid crystal display panel. The interface unit 96 manages communication with external devices. For example, the interface unit 96 functions when the exposure apparatus 1 receives a control program 931 and CAD data 932 from an external source. For this purpose, the interface unit 96 may also have a function for reading data from an external recording medium.

[0036] The CPU 91 implements functional blocks such as the exposure data generation unit 911, exposure control unit 912, focus control unit 913, and stage control unit 914 in software by executing the control program 931. Note that at least a portion of each of these functional blocks may be implemented by dedicated hardware.

[0037] The exposure data generation unit 911 generates exposure data 911 for modulating the light beam according to the pattern, based on the CAD data 932 read from the storage 93. If the substrate S is deformed, such as by distortion, the exposure data generation unit 911 modifies the exposure data according to the amount of distortion of the substrate S, enabling drawing that matches the shape of the substrate S. The exposure data is sent to the exposure head 41, and the exposure head 41 modulates the laser light emitted from the light irradiation unit 40 according to the exposure data. The modulated light beam, thus modulated according to the pattern, is irradiated onto the substrate S, and the surface of the substrate S is partially exposed to draw the pattern.

[0038] The exposure control unit 912 controls the light irradiation unit 40 to emit a laser beam having a predetermined power and spot size. The focus control unit 913 controls the projection optical system (described later) provided on the exposure head 41 to focus the laser beam onto the surface of the substrate S.

[0039] The stage control unit 914 controls the stage drive mechanism 3 to move the stage 2 for alignment adjustment and for scanning during exposure. During alignment adjustment, the position of the stage 2 is adjusted in the X, Y, Z, and θ directions so that the relative positional relationship between the substrate S placed on the stage 2 and the exposure head 41 at the start of exposure is a predetermined relationship. On the other hand, during scanning, a main scanning movement is performed by moving the stage 2 in the Y direction at a constant speed to allow the substrate S to pass below the exposure head 41, and a step feed in the X direction at a constant pitch (sub-scanning movement) is performed.

[0040] Figure 3 is a schematic diagram showing an example of the detailed configuration of an exposure head. As shown in Figure 3, the exposure head 41 is provided with a spatial light modulator 410 having a diffractive optical element 411. Specifically, the spatial light modulator 410, which is attached to the upper part of a support column 400 that extends vertically (Z direction) from the exposure head 41, is supported by the support column 400 via a movable stage 412 with the reflective surface of the diffractive optical element 411 facing downwards.

[0041] In the exposure head 41, the diffractive optical element 411 is positioned such that the normal to its reflective surface is inclined with respect to the direction of propagation of the incident light beam L. Light emitted from the illumination optical system 53 passes through the aperture of the support column 400 and enters the mirror 413, is reflected by the mirror 413, and then irradiates the diffractive optical element 411. The state of each channel of the diffractive optical element 411 is switched by the control unit 9 according to the exposure data, thereby modulating the laser light beam L that enters the diffractive optical element 411.

[0042] Then, the laser light reflected from the diffractive optical element 411 as zero-order diffracted light enters the lens of the projection optical system 414, while the laser light reflected from the diffractive optical element 411 as first-order or higher diffracted light does not enter the lens of the projection optical system 414. In other words, basically, only the zero-order diffracted light reflected by the diffractive optical element 411 enters the projection optical system 414. The diffractive optical element 411 is positioned so that the zero-order diffracted light is emitted in the (-Z) direction.

[0043] As the diffractive optical element 411, various elements can be used that can change the intensity of the emitted light in multiple steps for a given amount of incident light, that is, elements that can modulate the incident light at multiple gradation levels before emission. For example, a GLV (Grating Light Valve; "GLV" is a registered trademark of Silicon Light Machines) element, which drives a ribbon-type diffraction grating with a MEMS (Micro Electro Mechanical System), is suitably applicable. This element can emit light with the incident light modulated to up to 1024 levels of intensity.

[0044] Light passing through the lens of the projection optical system 414 is focused by the focusing lens 415 and guided onto the substrate S as a downward exposure beam with the (-Z) direction of travel, i.e., downward, at a predetermined magnification. The projection optical system 414 constitutes a reduction optical system. This focusing lens 415 is attached to the focus drive mechanism 416. Then, in response to a control command from the focus control unit 913 of the control unit 9, the focus drive mechanism 416 raises and lowers the focusing lens 415 along the vertical direction (Z axis direction), thereby adjusting the convergence position of the exposure beam emitted from the focusing lens 415 to the upper surface of the substrate S.

[0045] As shown along the optical path of the laser beam L indicated by the dashed line in Figure 3, the laser beam L guided from the light irradiation unit 40 to the exposure head 41 has a beam spot shape that extends uniformly and elongated in the X direction, with the X direction as the long axis and the Z direction as the short axis. On the other hand, the modulated laser beam Lm after modulation by the optical modulator 410 has the X direction as the long axis and the Y direction as the short axis, and the intensity at each position in the X direction is modulated according to the exposure data. Furthermore, the exposure beam Le emitted from the projection optical system 414 toward the substrate S is a reduced version of the modulated laser beam Lm in the X and Y directions. By irradiating the surface of the substrate S with this narrowed exposure beam Le, a fine pattern can be drawn on the surface of the substrate S.

[0046] By irradiating the substrate S with an exposure beam Le modulated according to the exposure data, and moving the exposure head 41 and the substrate S relative to each other in the Y direction, a band-shaped region of the substrate S that has a width equivalent to the spot size of the exposure beam Le in the X direction and extends in the Y direction can be exposed. By repeatedly performing exposure while sequentially changing the relative position of the exposure head 41 and the substrate S in the X direction, the entire substrate S can eventually be exposed.

[0047] In this manner, by combining scanning movements in the Y direction and scanning movements in the X direction between the exposure head 41 and the substrate S, the entire substrate S can be drawn. In this specification, scanning movements in the Y direction are referred to as "main scanning movements," and the Y direction is referred to as "main scanning direction Dm." On the other hand, scanning movements in the X direction are referred to as "sub-scanning movements," and the Y direction is referred to as "sub-scanning direction Ds." In this embodiment, these scanning movements are achieved by moving the stage 2 that supports the substrate S relative to the fixed exposure head 41.

[0048] Multiple exposure units 4 having the above configuration can be provided at different positions in the X direction. In this embodiment, five sets of exposure units 4 having the same configuration are provided, and these emit exposure beams Le in parallel to perform drawing, thereby improving the throughput of the drawing process. Although these exposure units 4 can operate independently of each other, their scanning movement relative to the substrate S is uniform due to their structure.

[0049] Figure 4 is a flowchart showing the process performed by the exposure apparatus configured as described above. This operation is achieved when the CPU 91 of the control unit 9 executes a control program 931 pre-recorded in the storage 93, causing each part of the apparatus to perform a predetermined operation.

[0050] Once the substrate S to be exposed is set on stage 2 (step S101), alignment adjustment is performed to match the orientation of the substrate S on stage 2 with the position of the drawing pattern (step S102). Since there are many known techniques for alignment adjustment, a detailed explanation is omitted here.

[0051] In parallel with, or before or after, the exposure data generation unit 911 reads the CAD data 932 stored in the storage unit 93 (step S103) and creates raster data based on it (step S104). The CAD data 932, which is design data representing the shape and size of the pattern to be drawn, is generally created as vector data. On the other hand, raster data (bitmap data) that represents the pattern in pixel units divided into a predetermined grid size is suitable for controlling the spatial light modulator 410, or more specifically, the diffractive optical element 411.

[0052] Therefore, the exposure data generation unit 911 performs data processing to convert the vector data and create raster data (step S104). Details of this data processing will be explained later with two specific embodiments as examples.

[0053] Based on the results obtained through alignment adjustment and raster data creation, the exposure data generation unit 911 creates the exposure data that is actually provided to the spatial light modulator 410 (step S105). If the created raster data is applied directly to the control of the spatial light modulator 410 due to positional misalignment during placement on stage 2 or distortion of the substrate S itself, the drawing position on the substrate S may deviate from the intended position. To resolve this problem, the raster data is corrected to reflect the results of the alignment adjustment, thereby creating the final exposure data provided to the spatial light modulator 410 (step S105). Since the correction method for this purpose is well known, its explanation is omitted here.

[0054] Subsequently, Stage 2 is positioned at a predetermined drawing start position (step S106), and the main scanning direction movement of the substrate S relative to the exposure head 41 begins (step S107). The exposure beam Le from the exposure head 41 is then irradiated onto the substrate S to perform drawing (exposure operation). The area exposed in one main scanning movement is referred to here as a "stripe". A series of processes on one substrate S is referred to as one "job".

[0055] The exposure operation continues until exposure of one stripe is complete (NO in step S108). Once exposure of one stripe is complete (YES in step S108), it is determined whether processing for one job is complete (step S109). If processing is not complete (NO in step S109), that is, if there are unexposed areas remaining, the stage S is moved in the sub-scanning direction (X direction) by a predetermined pitch (sub-scanning movement, step S110). Then, the process returns to step S107 and the exposure operation for the next stripe is performed.

[0056] If the processing for one job is complete (YES in step S109), the light irradiation from the exposure head 41 and the movement of the stage 2 are stopped, and the exposed substrate S is discharged (step S111). This completes the processing for one substrate S.

[0057] Next, we will explain the data processing involved in creating raster data from CAD data 932, i.e., the rasterization process. While raster data assigns pixel values ​​to pixel units by dividing the drawing pattern into a predetermined grid size, CAD data 932, which is generally vector data, does not have such grid size constraints. Therefore, as explained below, it is unavoidable that there will be some discrepancy between the original pattern represented in CAD data 932 and the pattern that is approximately represented in raster data.

[0058] Figure 5 schematically illustrates the difference between patterns represented by vector data and raster data. Here, as a general example, raster data is assumed to be created using a grid G ​​whose grid directions correspond to the mutually orthogonal main scanning direction Dm and sub-scanning direction Ds. To distinguish between the coordinate system shown in Figure 1, etc., and the coordinates in the image plane, which are a data processing concept, the coordinate axes in the grid G ​​corresponding to the main scanning direction Dm (Y direction) will be represented by the y-axis, and the coordinate axes in the image plane corresponding to the sub-scanning direction Ds (X direction) will be represented by the x-axis.

[0059] In this exposure apparatus 1, the line beam irradiated onto the substrate S from the light irradiation unit 40 via the spatial light modulator 410 and projection optical system 414 has its longitudinal direction in the X direction, and in the spatial light modulator 410, the diffractive optical element 411 also extends in the X direction. Therefore, in the raster data, the grid size in the x direction is mainly regulated by the resolution based on the lattice structure of the diffractive optical element 411, and the grid size in the y direction is mainly regulated by the operating speed (operating frequency) of the diffractive optical element 411.

[0060] As shown in Figure 5(a), the pattern Pv, ​​represented as vector data, inherently represents the shape of the pattern that should be drawn, independent of the grid G ​​in the raster data. Here, we consider a strip-shaped pattern Pv that is oblique to the grid direction. On the other hand, as shown in Figure 5(b), the pattern Pr on the raster data, represented by pixel units divided by the grid, may contain step-like jaggedness depending on the grid size. In particular, a clear difference appears between the edge shape of the pattern Pv represented as vector data and the edge shape of the pattern Pr represented as raster data. The magnitude of this discrepancy is one of the factors that reduces the smoothness of the pattern.

[0061] While it is possible to reduce this discrepancy and make the pattern smoother by reducing the grid size, as mentioned above, there are limits to how finely the grid size can be reduced due to hardware constraints. Below, we will describe two embodiments that can obtain a smoother pattern without changing the grid size by reflecting the magnitude of the discrepancy between the edges of the pattern Pv, ​​represented by vector data, and the edges of the pattern Pr, represented by raster data, in the gradation value of the pixel containing the edge or its surrounding pixels. These two processes can be selectively executed as the process in step S104 in Figure 4. Basically, by reducing the exposure of pixels containing the edges of the pattern to prevent pattern expansion, and by increasing the exposure of pixels that do not contain edges but are close to them to expand the pattern, the edges can be formed closer to their original positions.

[0062] <First Embodiment> Figure 6 is a diagram illustrating the principle of the first embodiment of the raster data creation process. Figure 7 is a flowchart of this process. In this embodiment, as shown in Figure 6(a), the grid shown by the dotted line is further divided into finer grids. Here, each grid is divided at equal intervals along the y-direction (corresponding to the main scanning direction Dm), that is, using a straight line (dashed line) extending in the y-direction. In this example, one grid is divided into four equal parts, generating four elongated grids in the y-direction. It can also be said that the grid size in the x-direction is divided into four equal parts. The grid division here is virtual in terms of data processing, and as will be described later, the raster data ultimately used for exposure will conform to the original grid size.

[0063] In this specification, the original grid in the raster data used during exposure will be referred to as the "target grid Gt," and its size as the "target grid size." Furthermore, the virtual grid set during the data processing process will be referred to as the "temporary grid Gc," and its size as the "temporary grid size." In this division example, the temporary grid size Sxc in the x-direction is (1 / 4) of the target grid size Sxt, while the grid size Sy in the y-direction remains unchanged.

[0064] The number of divisions from the target grid size to the provisional grid size is arbitrary and is not subject to hardware limitations as long as the processing power of the control unit 9 is sufficient. In general terms, it can be said that the grid is divided into N equal parts using an integer N greater than or equal to 2. The grid size can be said to be (1 / N) times larger due to the division (step S201).

[0065] Next, the grid size after division (temporary grid size) is applied, and the conversion from CAD data 932 to raster data (rasterization) is performed (step S202). Rasterization here represents the pattern in binary. For example, by assigning a pixel value of 1 to pixels that should be exposed and a pixel value of 0 to pixels that should not be exposed, binary raster data is obtained.

[0066] When rasterization is performed on the fine grid Gc after division, a smoother pattern Pr2 is obtained, which is closer to the original pattern Pv compared to pattern Pr in Figure 5(b), as shown in Figure 6(b). However, the raster data obtained in this way cannot be used directly to control the spatial light modulator 410. Therefore, a smooth pattern is achieved by reflecting this result in the pixel value (grayscale value) of each pixel of the target grid size.

[0067] In other words, for each pixel of interest separated by the target grid, a tonal value is assigned that has a weight corresponding to the magnitude of the discrepancy between the original edge of the pattern contained in that pixel of interest and the calculated edge set by rasterization (hereinafter sometimes referred to as "edge discrepancy"). To put it another way, the closer the edge formed when the pixel of interest is exposed with a pixel value of "1" is to the original edge, the higher the tonal value assigned to the pixel of interest.

[0068] More specifically, the grayscale value of a pixel of interest is set according to the sum of the pixel values ​​of four pixels divided by a provisional grid Gc that are contained within a pixel of interest divided by a target grid Gt (step S203). More specifically, the grayscale value of the pixel of interest is set such that the larger this sum, the larger the value. For example, when all four pixels that have been binary rasterized with the provisional grid size are given a pixel value of "1", the pixels containing these are assigned a grayscale value of "100%". The grayscale value of "100%" corresponds to the pixel value of "1" in the binarized data.

[0069] Similarly, when three out of four pixels have a pixel value of "1", a tonal value of "75%" is assigned to the target grid size containing them; when two out of four pixels have a pixel value of "1", a tonal value of "50%" is assigned; and when one out of four pixels has a pixel value of "1", a tonal value of "25%" is assigned. If all four pixels have a pixel value of "0", a tonal value of "0%" is assigned to the target grid size containing them. A tonal value of "0%" corresponds to a pixel value of "0" in the binarized data. In other words, the range between the minimum tonal value of 0% and the maximum tonal value of 100% is divided into four equal parts (or N equal parts if generalized) to set each tonal value. The assignment of these tonal values ​​is then based on the sum of the pixel values ​​within the target pixel, as described above.

[0070] Figure 6(c) shows a visualization of the multi-tone raster data in this way, where the gradation values ​​assigned to each pixel are represented by shades of gray in pattern Pr3. As can be seen, pixels within pattern Pr3 that are contained within the original pattern Pv are given a gradation value of 100%, while pixels containing edges are given a gradation value corresponding to the degree of discrepancy between the edge represented by that pixel and the original edge.

[0071] By performing exposure based on this type of raster data, the exposure amount of each pixel is increased or decreased according to the degree of edge deviation. Pixels given a low grayscale value, that is, pixels with reduced exposure, do not form strong edges and instead act to expand adjacent pixels, thus suppressing the occurrence of jagged edges and making it possible to obtain a smooth pattern.

[0072] Figure 8 illustrates the effects of the first embodiment. Using the exposure apparatus 1 shown in Figure 1, exposure was performed to form a pattern including curves on a substrate S, and the formed pattern was evaluated. Figure 8(a) is a comparative example, showing the resulting pattern and a magnified view of a part thereof when rasterization to the target size was performed directly from CAD data and then exposed. On the other hand, Figure 8(b) is an example in which the data processing of this embodiment is applied, and it is shown that a smoother drawing result can be obtained compared to the pattern in Figure 8(a).

[0073] <Modified form of the first embodiment> Figure 9 shows a first modified example of the first embodiment. As shown in Patent Document 1, raster data is often represented by run-length data that focuses on the continuity of pixels in the main scanning direction. For example, when the result of rasterization with a provisional grid size in step S202 is represented as run-length data, as shown in Figure 9, a series of pixels that are continuous in the y direction are represented as a single run R in the raster data.

[0074] In such cases, the gradation value for each pixel of the target grid size can be determined according to the number of runs (R) passing through that pixel. For example, if there are four runs (R) in a pixel of interest, a gradation value of 100% can be assigned, and if there are three runs (R), a gradation value of 75% can be assigned. By assigning gradation values ​​according to the number of runs within a pixel in this way, it is possible to assign a gradation value to each pixel according to the edge deviation, thereby making it possible to obtain a smooth pattern with reduced jaggedness.

[0075] Figure 10 shows a second modified example of the first embodiment. In the above embodiment, when setting the provisional grid size, each pixel is divided into four equal parts by a straight line in the y direction. However, as shown in Figure 9, a similar effect can be obtained by applying a grid Gc2 subdivided in both the x and y directions. However, when run-lengthening raster data, the unit length of the runs also decreases with grid subdivision, which complicates the process of converting to data of the original target grid size. In this respect, the method of the above embodiment is preferable.

[0076] <Second Embodiment> Figure 11 is a diagram illustrating the principle of the second embodiment of the raster data creation process. Figure 12 is a flowchart of this process. The raster data creation process of the second embodiment is particularly effective when the pattern is a line pattern of a constant width. Jagged edges that occur in the creation of such line patterns result in variations in pattern width. For example, when a pattern is formed as wiring for an electronic device, variations in pattern width can cause instability in the characteristics of the manufactured device.

[0077] In this embodiment, the uniformity of the line pattern width is maintained by grasping the characteristics of the pattern from a macroscopic perspective than in the first embodiment and performing rasterization that reflects this. Specifically, as shown in Figure 11, the angle θ that the edges of the line pattern make with respect to the grid direction is calculated from CAD data 932 that defines the shape of the pattern to be drawn (step S301). Since the CAD data 932, which is vector data, contains information that defines the direction in which the edges extend, it is possible to determine the angle θ using this information. As shown in Figure 11, the angle θ may be defined as either the angle with respect to the x direction or the angle with respect to the y direction, as long as it is consistent as a system. In this embodiment, the angle θ that the pattern makes with respect to the x direction is used.

[0078] Next, rasterization is performed based on the CAD data 932 to create raster data (step S302). In this case, rasterization may be performed at the target grid size, and the resulting number of grayscale levels can also be arbitrary. For example, the processing of the first embodiment may be applied. Then, the pixel value of each pixel represented by the raster data thus obtained is corrected according to the angle θ value (step S303).

[0079] According to the inventor's findings, the variation in pattern width caused by jagged edges resulting from grid coarseness depends on the angle of the pattern with respect to the grid direction. Specifically, when the pattern inclination is small (for example, θ = about 10°), the edges of the pattern tend to be excluded by rasterization (i.e., not reflected in the exposure), and the pattern width becomes narrower than it should be. On the other hand, when the inclination is large (for example, θ = about 80°), the opposite tends to occur, and the pattern width tends to become wider.

[0080] Therefore, in this embodiment, the pixel value of each pixel, which is represented by binary or multi-level rasterization, is corrected according to the angle θ of the pattern. This makes it possible to suppress variations in pattern width caused by the tilt of the pattern. The specific correction method is as follows.

[0081] Figure 13 shows a method for correcting raster data in the second embodiment. As shown in Figure 13(a), in this embodiment, the relationship between the pattern angle θ, the corresponding correction coefficient C, and the final obtained irradiation light amount (dose amount) D is stored in a table in advance in storage 93. In steps S301 and S302 of the processing shown in Figure 12, the pattern angle θ and the pixel value of each pixel after rasterization are obtained, and in step S303, the pixel value is corrected by referring to this table T. Specifically, the corrected pixel value is obtained by multiplying the pixel value of each pixel by the correction coefficient C corresponding to the angle θ.

[0082] For example, the dose D for a pixel with a grayscale value of 100% (or pixel value of 1) is given by the following equation, where angle θ is a variable: D(θ) = C(θ)·D0 This can be expressed as follows: Here, the symbol D0 represents the dose amount when the angle θ is zero, that is, when the pattern is parallel to the x-axis. The value of the dose amount D0 varies depending on the photosensitive material formed on the substrate S. If the gradation value is less than 100%, the dose amount is also reduced accordingly. By achieving halftones in this way, the exposure amount is increased or decreased according to the tilt of the pattern, and fluctuations in the pattern width can be suppressed. This correction may be applied to all pixels included in the pattern, or it may be applied only to pixels near the edge of the pattern.

[0083] As mentioned above, when the angle θ is small, the pattern width tends to be narrower than it should be, and conversely, when the angle θ is large, the pattern width tends to be wider than it should be. Also, when the angle θ is 0° and 90°, there is no increase or decrease in pattern width due to the tilt in principle, and therefore no correction is necessary, and the correction coefficient C is 100%.

[0084] From this, the correction coefficient C shows a tendency to change with respect to the angle θ, for example, as shown in Figure 13(b). However, the actual value of the correction coefficient C for each angle is determined by various complex factors such as the sensitivity of the photosensitive layer, the wavelength of the illumination light, the grid size, and the movement speed in the main scanning movement, and therefore cannot be determined in general terms. For this reason, specific numerical values ​​are omitted. For example, the relationship between the angle θ and the amount of variation in pattern width can be understood through prior experiments, and a table T can be created based on that. A state where the correction coefficient C is 100% regardless of the angle θ corresponds to the conventional technology that does not perform correction according to the tilt.

[0085] Furthermore, it is preferable to provide a function that allows the user to edit Table T retrospectively according to their purpose. For example, this function can be realized by displaying Table T on the display unit 95 and accepting user editing input via the input unit 94. In this case, since the dose amount D is automatically calculated by the above formula, the user only needs to edit the value of the correction coefficient C corresponding to the angle θ. Note that the increment of the angle θ is not limited to those shown in the figure and can be arbitrary, and for example, the angle value itself may be editable by the user.

[0086] As shown in Figure 13(b), the correction coefficient C may exceed 100% in order to compensate for the narrower pattern width. To enable illumination light output exceeding 100%, it is desirable that the gradation value "100%" in this embodiment be set to a point somewhat below the upper limit (for example, corresponding to 90% of the upper limit) within the region where the response of the diffractive optical element 411 to the input gradation is approximately linear. By doing so, a margin can be provided in the region exceeding 100%, and the necessary corrections can be made using this margin.

[0087] In this embodiment as well, the degree of discrepancy between the edges generated by rasterization and the actual pattern edges (edge ​​deviation) is reflected in the weight of the pixel values ​​through a factor called the slope of the pattern (angle θ) when the collection of pixels is viewed macroscopically.

[0088] <Other> As explained above, in the above description, the substrate S corresponds to the "object to be exposed" in the present invention. In the exposure apparatus 1 of this embodiment, the stage 2, exposure head 41, and control unit 9 function as the "holding unit," "exposure unit," and "control unit" of the present invention, respectively. Furthermore, the spatial light modulator 410 functions as the "optical modulator" of the present invention, and of these, the diffractive optical element 411 functions as the "spatial light modulation element" of the present invention.

[0089] It should be noted that the present invention is not limited to the embodiments described above, and various modifications can be made without departing from the spirit of the invention. For example, the drawing pattern Pv used in the above description is just one example, and the present invention is not limited to this and is effective for patterns of various shapes.

[0090] Furthermore, for example, the exposure apparatus 1 in the above embodiment has a control program pre-installed for executing the data processing method and exposure method using the present invention. However, if the optical modulator has the function of modulating the light beam in multiple levels (for example, four or more levels) and has sufficient computing power, it is also possible to execute the present invention in an existing exposure apparatus by implementing the control program in the existing apparatus.

[0091] In this sense, the present invention can also be implemented as a computer program that causes a computer to execute at least one of the data processing method and the exposure method using the present invention, or as a recording medium that non-temporarily stores the computer program.

[0092] As described above with specific embodiments as examples, in the first embodiment of the exposure method according to the present invention, for example, in the exposure step, the light beam is irradiated onto the object to be exposed while moving the exposure unit that emits a modulated light beam and the object to be exposed relatively in the main scanning direction, and in the data creation step, rasterization is performed with the directions corresponding to the main scanning direction and the direction orthogonal thereto as grid directions, and the provisional grid size can be determined by dividing the target grid size into N equal parts along the direction corresponding to the main scanning direction among the grid directions.

[0093] With this configuration, the raster data created with a temporary grid size can be represented as run-length data, which groups together pixels arranged in a single row in the main scanning direction. This makes it possible to reduce the amount of data after rasterization. For the same reason, in the first embodiment of the data processing method according to the present invention, the temporary grid size can be obtained by dividing the target grid size into N equal parts along one grid direction.

[0094] Furthermore, in the exposure method according to the present invention, for example, a correction process may be performed on the raster data to adjust the incident position of the light beam on the object to be exposed, and the optical modulator may be controlled based on the corrected data. Due to the installation conditions or deformation of the object to be exposed, the incident position of the light beam on the object to be exposed may deviate from the intended position. By correcting the raster data in response to these deviations, it is possible to expose the object to the correct position.

[0095] Furthermore, in the first aspect of the data processing method according to the present invention, for example, the larger the sum of the pixel values ​​of N pixels, the higher the grayscale value can be assigned to the pixels of the target grid size that are composed of those pixels. With such a configuration, it is possible to represent the weights of the N pixels by replacing them with grayscale values. For example, (N+1) levels of grayscale values ​​can be obtained by equally dividing the range from the minimum grayscale value to the maximum grayscale value. Since the pixels of the hypothetical grid size are represented in binary, the sum of the values ​​of those N pixels can take on (N+1) levels, and the difference between adjacent values ​​is constant. This means that when representing the sum of image values ​​as grayscale values, it is sufficient to set the grayscale value to (N+1) levels and keep the difference between adjacent grayscale values ​​constant. This can be achieved by dividing the range from the minimum grayscale value to the maximum grayscale value into N equal parts.

[0096] Furthermore, in a second aspect of the data processing method according to the present invention, a table relating angles and correction amounts for pixel values ​​may be prepared in advance. With such a configuration, once the angle of the pattern is determined, the correction amount can be immediately determined by referring to the table. Moreover, the table may be editable by user operation. This makes it possible to adjust the quality of the pattern width according to the user's purpose.

[0097] Furthermore, in the exposure apparatus according to the present invention, a spatial light modulator including a diffractive optical element using a diffraction grating driven by a micro-electromechanical system, for example, can be used as the light modulator. With such a configuration, the gradation level of the modulated light can be changed and set in multiple steps by an electrical signal applied to the diffractive optical element, making it suitable for carrying out the present invention. [Industrial applicability]

[0098] This invention is particularly suitable for the technical field of exposing substrates to form patterns on substrates such as semiconductor substrates, semiconductor package substrates, printed wiring substrates, or glass substrates. [Explanation of Symbols]

[0099] 1. Exposure apparatus 2 stages (holding part) 3-stage moving mechanism 9. Control Unit 41. Exposure head (exposure unit) 410. Spatial Light Modulator (Optical Modulator) 411 Diffractive optical elements (spatial light modulation elements) 932 CAD data (design data) Dm Main scanning direction Ds sub-scanning direction L laser light beam Le exposure beam S Substrate (exposed object) SXC temporary grid size Sxt Target Grid Size

Claims

1. An exposure method in which a light beam modulated based on drawing data is irradiated onto an object to be exposed, A data creation step of creating raster data by rasterizing the pattern, based on design data representing the pattern to be drawn, in units of pixels divided into a predetermined target grid size by two intersecting grid directions, Based on the raster data, an exposure process is performed in which a light modulator capable of modulating a light beam emitted from a light source in multiple shades is controlled to modulate the light beam according to the pattern and irradiate the object to be exposed. Equipped with, In the aforementioned data creation process, The design data is rasterized to create binary raster data by multiplying the target grid size by (1 / N) times (where N is an integer of 2 or more) along one of the grid directions to create a temporary grid size. An exposure method comprising setting N consecutive pixels of the provisional grid size as one pixel of the target grid size, and assigning one of the (N+1) gradation values ​​to the pixel of the one pixel, corresponding to the sum of the pixel values ​​of the N pixels.

2. In the exposure process, the exposure unit that emits the modulated light beam and the object to be exposed are moved relative to each other in the main scanning direction while the light beam is irradiated onto the object to be exposed. In the data creation process, the rasterization is performed with the grid directions being the directions corresponding to the main scanning direction and its orthogonal direction, respectively. The exposure method according to claim 1, wherein the provisional grid size is determined by dividing the target grid size into N equal parts along the grid direction corresponding to the main scanning direction.

3. An exposure method in which a light beam modulated based on drawing data is irradiated onto an object to be exposed, A data creation step of creating raster data by rasterizing the pattern, based on design data representing the pattern to be drawn, in units of pixels divided into a predetermined target grid size by two intersecting grid directions, An exposure process in which a light modulator is controlled based on the raster data to modulate the light beam emitted from the light source according to the pattern, and irradiates the object to be exposed with the light. Equipped with, When the pattern is a line pattern of a certain width, in the data creation process, Based on the design data, the angle between the direction in which the line pattern extends and one of the grid directions is determined. An exposure method for creating raster data by correcting the gradation value of each pixel after rasterization by multiplying it by a correction coefficient that has been determined in advance from the relationship between the angle and the width of the line pattern.

4. The exposure method according to any one of claims 1 to 3, wherein correction processing is performed on the raster data to adjust the incident position of the light beam on the object to be exposed, and the light modulator is controlled based on the corrected data.

5. A data processing method for creating raster data from design data representing a line pattern of a fixed width to be drawn, by rasterizing the line pattern in pixel units divided into a predetermined target grid size by two grids in mutually intersecting grid directions, A step of determining the angle between the direction in which the line pattern extends and one of the grid directions based on the design data, A step of performing rasterization based on the aforementioned design data, The process of creating the raster data involves correcting the gradation value of each pixel after rasterization by multiplying it by a correction coefficient that has been previously determined from the relationship between the angle and the width of the line pattern. A data processing method comprising the following features.

6. The data processing method according to claim 5, wherein a table relating the angle and the correction coefficient is prepared in advance.

7. The data processing method according to claim 6, wherein the table is editable by user operation.

8. A holding part for holding the object to be exposed, An exposure unit having a light modulator capable of modulating a light beam emitted from a light source in multiple steps, and irradiating the object to be exposed with the modulated light beam, A control unit that creates raster data by performing the data processing method described in any one of claims 5 to 7 based on design data representing the pattern to be drawn, and controls the optical modulator based on the raster data. An exposure apparatus equipped with the following features.

9. The exposure apparatus according to claim 8, wherein the optical modulator is a spatial light modulator including a diffractive optical element using a diffraction grating driven by a micro-electromechanical system.

Images