Exposure control in direct photolithography for manufacturing printed circuit boards or circuits

Abstract

This invention relates to an apparatus for exposure control when employing a direct photolithography method for two-dimensional structures in a photosensitive coating, and a method for converting registration data into direct exposure data. The solution of this invention to achieve improved exposure control when employing a direct photolithography method for two-dimensional structures in a photosensitive layer, allowing registration of a target mark independent of its defined position, is that a plurality of proximal cameras for forming a gapless linear scanning region (23) within a predetermined width of the substrate (2) are arranged in a registration unit (1) in a one-dimensional motion linear alignment transverse to the substrate (2), wherein the image angles of adjacent proximal cameras have an overlapping region along the linear scanning region (23), redundant image records of the substrate (2) of the adjacent cameras can be detected in the overlapping region, and the calculation unit (5) has a component for calculating the position of the target mark based on the redundant image records of the adjacent proximal cameras, supplementarily using the height position of the target mark determined by triangulation of the distance to the substrate surface (21).

Description

[0001] This invention relates to an apparatus for exposure control in photolithographic direct exposure of two-dimensional structures in photosensitive coatings on preferred printed circuit boards, display substrates, or wafers, and a method for converting registration data into direct exposure data, particularly adapting the registration to an uneven substrate and converting it into "dynamic" registration. "Dynamic" registration here refers to sequentially acquiring position data of the target marker and thus the position and unevenness of the substrate to be processed during continuous substrate movement, and aligning the exposure pattern with the substrate by adapting it to the exposure data of the subsequent direct exposure.

[0002] The invention is particularly applicable to the manufacture of printed circuit boards, displays, and chips in the electronics and semiconductor industries.

[0003] Existing technologies disclose exposure systems for disc-shaped or plate-shaped workpieces, which expose the object using electromagnetic radiation, primarily in the visible or ultraviolet spectral range, by means of a laser beam or a beam of electrons or particles with a predetermined pattern. In this case, exposure occurs only after a correct positional relationship is established between the object to be exposed, bearing a mark (target mark or target), and the predetermined pattern stored in the exposure apparatus. For this purpose, one or more cameras are used to detect the target mark on the object and to align the object and the pattern to be exposed before or within the exposure zone.

[0004] For fabricating conductive paths or minimal electronic structures on plate-like substrates (such as printed circuit boards, display substrates, or wafers), the exposure process performed with high spatial accuracy, and the processing and alignment time of the plate-like workpiece required for this, are limiting factors for increasing throughput. Therefore, when processing and exposure steps need to be performed overlappingly or simultaneously, and when it is desired to expose both the front and back sides in the same apparatus, it is necessary to reduce the auxiliary processing time of the exposure process. Such solutions are disclosed, for example, in EP 0 951 054 A1, EP0 722 123B1, US 6 806 945B2, and JP 2010-181519 A.

[0005] Another challenge when using the direct exposure method is to achieve registration in a way that is independent of the type, number, and location of the target markers (targets) on the exposed object, as well as the height deviation of the object surface.

[0006] A solution for detecting surface topography is known from WO 2016 / 115536 ​​A2, in which a known two-dimensional graphic is projected or otherwise applied onto the surface, recorded in two dimensions along with the surface, and the three-dimensional shape of the surface is determined by combining the distortion of the graphic with the surface roughness. However, due to limited resolution, this method is not suitable for detecting roughness with small differences.

[0007] EP 0 954 768 B1 describes an apparatus for focusing on the surface of a semiconductor wafer, wherein, for the semiconductor wafer, the actual surface topography is recorded before exposure in an exposure apparatus. For this purpose, this surface is planarly detected as height information using a distance sensor, and periodic height changes are subsequently measured and stored. Based on the measured periodic height changes, the optimal focal position of the exposure optics is determined for the exposed sections of the wafer surface, and the wafer is aligned accordingly. The disadvantages are the relatively time-consuming height information acquisition using a distance sensor and the subsequent wafer alignment.

[0008] WO 03 / 094582A2 discloses another method for registration control during direct laser exposure of substrates, particularly on multilayer PCBs (Printed Circuit Boards). This method involves modifying the circuit pattern in a non-uniform manner to generate a digitally controlled image, enabling precise matching of the circuit pattern printed on the substrate using this digitally controlled image with existing circuit components. For this purpose, selected reference marks are registered on the existing actual structure, where the target structure to be exposed is corrected in the spatial direction based on the deviation between the actual and target positions, so that exposure can be performed using a modified scanning grating. However, since a camera located in the exposure head is used for registration, approaching each reference mark requires relative movement between the substrate and the scanning head, which is disadvantageous.

[0009] DE 10 2018 132 001 A1 discloses an apparatus for direct exposure of printed circuit boards (PCBs) for processing plate-shaped workpieces with high throughput. In this apparatus, to detect target marks on the PCB with known target mark positions, a registration unit is equipped with two to three area-scanning cameras that can move laterally to the PCB. These cameras are alternately arranged parallel to the processing path to register the PCB using two stages moving on the same track system parallel to the processing path, thereby minimizing the cycle time of PCB processing by reducing processing and auxiliary processing times. In this case, the cameras are regularly arranged at the edge regions of the PCB or above the circuitry of the PCB panel, where the target marks are known to be predictable. For PCBs and wafers, there is an increasing need to detect arbitrarily positioned target marks, but due to the required camera displacement, detection is only possible with reduced throughput, and changes in substrate surface height cannot be detected at all.

[0010] EP 2 775 349 A1 describes a method for determining the correct focal position in an optical inspection system, wherein the difference between the focal position of the inspection system and the position of the object to be inspected is measured, and an image of the object is acquired in a manner independent of the correct focal position. The algorithm of the inspection system can infer the magnitude and direction of the difference between the focal position and the object position based on features of the acquired image, specifically by adapting the object's position to the focal position of the inspection system based on the difference and direction. In the case of a method described for a sample to be inspected by a microscope, different feature objects that may have different spatial extensions and therefore considerable height differences within the observed surface may be found, requiring individual adaptation of the focal position to enable the definitive detection of different objects. In this case, the height profile across the entire surface is not detected.

[0011] Patent specifications US 6,245,585 B1 and US 6,449,029 B1 describe a method and apparatus for adapting a focal position in a photolithography process on a semiconductor wafer. Before exposure, the surface height along the z-direction is measured in each individual section of the wafer to be exposed. These sections are rectangular grids arranged in a grid pattern, and these rectangles are exposed sequentially. Reflections from five obliquely incident laser beams are detected on the surface of each section, with one laser beam aligned to the center and four laser beams aligned to the corresponding corners of the sections. Based on the position of the reflections, the offset of the height position of each section from a reference height can be determined, and the surface angle of each section can be calculated. In this case, before exposing each section, the wafer is aligned according to stored values, wherein alignment is performed by individually controllable actuators, by means of which the height and angle are adapted. However, for each section, only an average correction can be set across the entire section.

[0012] An apparatus for introducing patterns onto a wound continuous substrate by radiation is described in DE 10 2019 128 198.9, which was not disclosed in advance. This apparatus involves target registration and pattern exposure on a continuous substrate that is tensioned and guided on a processing roller. If the registration unit and the processing unit are radially opposed on the roller, the height change of the substrate due to the roller curvature can be used to focus the camera of the registration unit and the processing beam by longitudinally moving the roller relative to the direction of substrate movement. However, measuring the magnitude of defocus requires additional auxiliary components to determine the absolute position of the substrate surface.

[0013] US 2004 / 0223129 A1 discloses an exposure apparatus for planar exposure of photosensitive materials, which moves in a plane relative to an exposure apparatus having multiple similar exposure heads arranged in a matrix. In each exposure head, light from a light source is modulated into a two-dimensional pattern and projected onto the material surface for exposure using telecentric optics. To compensate for unevenness at different locations on the material surface, each exposure head has a pair of wedge prisms in the optical path, which are used to adjust the optical length between the lens and the material surface when a distance sensor based on reflected laser radiation detects unevenness caused by changes in the material surface position. A disadvantage in this case is that tracking each exposure head and its image field as a single image segment can lead to abrupt changes in the imaging ratio of adjacent image segments.

[0014] The objective of this invention is to find a novel approach to improve exposure control when employing direct exposure methods for two-dimensional structures in photosensitive layers on printed circuit boards or wafers. This approach allows for "dynamic" registration of target marks independent of their defined locations and avoids the use of costly telecentric lenses. Another objective is to achieve flexible adaptation of the exposure pattern to determined substrate unevenness.

[0015] The solution of the present invention to achieve the above-mentioned objective is an apparatus for exposure control during direct exposure of a two-dimensional structure in a photosensitive coating on a substrate by photolithography. The apparatus includes a registration unit for registering target marks located on the substrate surface, a movable stage system for supporting and one-dimensionally moving the substrate below the registration unit in a defined manner, a processing unit with a controllable linear processing path for photolithographic processing of the substrate to print a two-dimensional structure using a processing beam, and a calculation unit for controlling the alignment between the processing path and the substrate based on the substrate position determined by the registered target marks through local adaptation of the photolithographic processing. Specifically, the apparatus includes... Multiple proximal cameras for forming a gapless linear scanning area within a preset width of the substrate are arranged in the registration unit in a manner that is linearly aligned with the substrate in one-dimensional motion and have extended image angles in the direction of the linear scanning area, wherein the image angles of adjacent proximal cameras have an overlapping area along the linear scanning area so as to detect redundant image records of the substrate of adjacent cameras in the overlapping area, and the calculation unit has a component for calculating the position of the target mark based on the redundant image records in the overlapping area (13) of adjacent proximal cameras, in addition to using the height position of the target mark determined by triangulation of the distance to the substrate surface.

[0016] Advantageously, the registration unit is configured in a certain way with a plurality of pericentric cameras for generating linearly continuous, gapless sensor regions, such that the image angles of adjacent cameras have an overlapping region, the overlapping region being at least half the size of the image angle, wherein the computing unit is configured in a certain way to determine a target mark arbitrarily positioned within the width of the substrate by triangulating the distance at any location on the substrate surface in a manner independent of the position of the target mark in the gapless, continuous overlapping region of the image angles of adjacent pericentric cameras.

[0017] The camera is preferably a line scan camera, so that a linear scan area can be formed over the entire width of the substrate in a narrow, gapless and overlapping manner using a camera with a large scan length.

[0018] Advantageously, the cameras are aligned with the substrate surface by means of optical axes that are parallel to each other, wherein the overlapping area of ​​the image angles of all cameras is the same size.

[0019] In another advantageous embodiment, two adjacent cameras are aligned with the substrate surface by means of optical axes that are tilted relative to each other, wherein the overlapping area of ​​the image angles of the cameras that are tilted relative to each other is set in such a way that the image angles of the two cameras completely overlap on the substrate surface.

[0020] In this case, the overlapping area formed by pairs of cameras tilted relative to each other is connected to another overlapping area at least without gaps, until the extension of these overlapping areas is at least equivalent to the width of the substrate, wherein there may be overlap between the pairs of cameras tilted relative to each other, so as to ensure a gapless scanning area of ​​the registration unit for all permissible height variations Δz on the substrate surface.

[0021] These pairs of cameras, tilted relative to each other, are preferably arranged in a way that subject them to the Scheimpflug condition.

[0022] In an advantageous embodiment of the invention, the computing unit further includes a control device for rapidly focusing the processing unit along the processing path based on the height variation Δz of the substrate surface, comprising triangulation of a target mark or any imaging structure on the substrate surface based on redundant image recordings in the overlapping areas of adjacent pericentric cameras.

[0023] In another preferred variant, the registration unit for target mark detection has only two centripetal cameras (transverse to the direction of substrate movement) arranged on the scan line. When the substrate is guided as a flexible, continuous substrate on a roller conveyor system in a taut and height-variable manner Δz, the overlap region of the centripetal cameras is one-hundredth to one-third of the image angle of the cameras. The overlap region of the image angles of the two cameras is designed in such a way that triangulation can be applied on the roller conveyor system to accurately determine the thickness of the substrate in the overlap region of the image angles, and this thickness can be assumed to be constant over the entire width of the roller conveyor system.

[0024] Furthermore, the registration unit is advantageously equipped with light sources for illuminating the linear, gapless scanning area, these light sources being arranged in a uniformly distributed manner within the housing to achieve uniform illumination of the scan lines by dark-field or bright-field illumination.

[0025] In this case, these light sources are suitable for continuous illumination and have means for controlling at least one of the characteristics of brightness, angle of incidence, or spectral range.

[0026] Furthermore, these light sources can be configured for continuous illumination so that image recording can be achieved by controlling the integration time of the sensor line using the principle of electronic shutter.

[0027] In another preferred embodiment of the invention, focus tracking for the height change Δz of the substrate is integrated into the processing unit for real-time adaptation of the height change Δz, which is determined by the computing unit through triangulation based on images redundantly recorded by two adjacent cameras with the aid of the registration unit, by means of rapid focus correction for each pixel of the registration unit, wherein focus tracking can be controlled based on changes in the position or curvature of the lens or mirror.

[0028] In this case, focus tracking can advantageously be controlled based on the change in the mirror curvature at least along the lateral x-direction of motion relative to the substrate in the direction y. Preferably, this focus tracking can be controlled separately based on the changes in the mirror curvature along the substrate in the direction y and the changes in the mirror curvature along the lateral x-direction.

[0029] In one advantageous implementation, focus tracking can be controlled by using piezoelectric elements to change the curvature of the mirror.

[0030] Advantageously, focus tracking based on changes in the position or curvature of a lens or mirror can also be used to correct imaging-related focus deviations caused by optical design or optical manufacturing of upstream focusing optics or other upstream optical components.

[0031] The solution of the present invention to achieve the above-mentioned objective also lies in a method for exposure control during the direct photolithographic exposure of a two-dimensional structure in a photosensitive coating on a substrate, which includes the following steps:

[0032] In the registration unit, multiple centripetal cameras are arranged in a seamless linear scanning area transverse to the direction of movement of the substrate to detect target marks located on the substrate. Each centripetal camera has an extended image angle along the linear scanning area, including an overlapping region formed by adjacent centripetal cameras, to obtain redundant image recordings of the substrate based on adjacent cameras within the overlapping region.

[0033] -Move the substrate below the registration unit with a defined one-dimensional motion on the movable stage system.

[0034] - Provides a machining unit for photolithographically producing two-dimensional structures using a machining beam that can be controlled along a linear machining path.

[0035] - When the substrate passes through the linear scanning area of ​​the registration unit in a single pass, the spatial position of the target mark arbitrarily distributed within a predetermined width of the substrate is detected, taking into account its length, width, and height.

[0036] -In addition to using triangulation of the distance to the substrate surface to determine the height position of the target mark based on redundant image records from adjacent pericentric cameras, the position of the target mark arbitrarily distributed within the width of the substrate is determined based on redundant image records in the overlapping area of ​​the adjacent pericentric cameras.

[0037] - Calculate data for alignment and local adaptation of the processing unit for processing a substrate with a two-dimensional structure, in order to control the processing beam along a linear processing path transverse to the direction of movement of the substrate, and

[0038] - Control the alignment between the processing path and the substrate and locally adapt the photolithography process according to the substrate position determined by the registered target mark.

[0039] In a preferred variation of the method, the calculation of the spatial position of the target mark arbitrarily distributed within the width of the substrate is extended based on the height position as the substrate passes through to the triangulation of other detectable structures of the substrate in redundant images recorded in the overlapping region, and the focus of the processing beam is rapidly adjusted along the processing path by means of focus tracking based on the control of the position or curvature of the lens or mirror.

[0040] In this case, it is advantageous to rapidly focus and adjust the focal point of the processing beam at a frequency at least two to thirty times higher than the conventional scanning frequency of the processing beam.

[0041] This invention is based on the following fundamental considerations: For so-called registration systems, one or more cameras with two-dimensional sensors (such as CCD or CMOS cameras) are used. These cameras are typically equipped with telecentric lenses and arranged precisely perpendicularly within a defined section of the substrate. Therefore, even if the focal position of the lens shifts relative to the substrate due to variations in thickness or topography, the measured position remains constant within the usable depth of field of the telecentric camera. Telecentric lens structures are consequently relatively expensive and bulky; therefore, the mechanical extension of the lens must always be greater than the image field to be recorded. Thus, telecentric lenses do not allow for gapless image recording using multiple cameras arranged in a straight line; instead, they must be arranged offset along several parallel lines. Furthermore, it is impossible to inherently measure the height deviation of the substrate, which is crucial for highly accurate determination of target marker positions and for precise alignment and adaptation of exposed images.

[0042] The present invention addresses the aforementioned problems by combining a linear image scan (hereinafter referred to as FPSS – Full Panel Scan System) performed across the entire substrate width by cameras with proximal lenses during continuous relative motion between a substrate and a linear scanning area. The image angles of these cameras overlap to a certain extent so that height differences can be triangulated for each substrate location within the linear scanning area based on multiple camera images from different cameras. Alternatively, it may employ a multi-camera layout defined by cameras positioned tilted to the substrate, which, under the Scheimpflug condition, allows complete overlap of the image angles of two adjacent cameras, thereby enabling triangulation for each substrate location within the linear scanning area based solely on two camera images.

[0043] Within the scope of this invention, the image angle adopts the definition used in the field of photography. In this context, the image angle refers to the angle in object space (the space in which the object is located) limited by the edge of the recording format of the camera (here, the camera of the registration unit). Therefore, the image angle is determined here by the height and width of the recording format (unlike the commonly used method of giving the diagonal of the maximum image angle in a manner independent of the aspect ratio of the actual recording format used). The recording format is preset by the sensor format, thereby defining the object space spanned by the horizontal and vertical image angles through the lens image.

[0044] Aside from the image format—the height H and width B of the recording format—the image angle is essentially determined solely by the lens's current focal length f. However, the focal length f can only be directly used to define the image angle when the lens is set to "infinity" (object-side telecentric lens). For images of objects with finite widths (shorter object distances), the image width b becomes greater than the focal length f and replaces it, thus deriving the width B of the recording format for the horizontal image angle.

[0045] α = 2·arctan [B / (2·b)] (1).

[0046] When using a line scan camera, since the sensor format is almost linear, according to equation (1), the horizontal image angle is the decisive image angle of the camera, and can therefore be used alone to define the linear field of view (FOV) on the object side.

[0047] The height deviation of the substrate along the scan line of the pericentric camera, measured by local triangulation, can be used not only for precise two-dimensional registration of target marks, but also for highly accurate tracking of the focal position of the processing beam moving linearly along the processing path. Based on the measured height change, the tracking of the processing beam along the processing path can be supplemented by an adjustment mechanism for rapid focal change of the processing beam to achieve routine alignment between the exposure pattern and the registered target mark position. This requires focal change at least two to ten times the routine scanning frequency of the processing beam (0.5 to 1 kHz for polygon scanners), and therefore must be controlled by simple linear position changes or by altering the radius of curvature of the lens or mirror.

[0048] This invention provides a novel approach to improve exposure control when using direct exposure methods for two-dimensional structures in photosensitive layers on printed circuit boards or wafers. It allows for "dynamic" registration of target marks using a linear scanning area independent of a fixed position preset value, avoiding the use of expensive telecentric lenses, and enabling flexible alignment and adaptation of the exposure pattern to the determined substrate unevenness.

[0049] The present invention will now be described in detail with reference to the embodiments and accompanying drawings. Wherein:

[0050] Figure 1 : A schematic diagram of a registration unit for target marker detection. This registration unit is constructed as a multi-camera configuration, which includes a linear layout of pericentric cameras with overlapping image regions for achieving linear scanning regions across the entire object width laterally to the object's motion.

[0051] Figure 2 : A schematic diagram of a registration unit configured for multiple cameras, wherein the linear arrangement of the cameras consists of pairs of centripetal cameras tilted relative to each other, which satisfy the Scheimpflug condition and have completely overlapping image regions in pairs.

[0052] Figure 3 A schematic diagram illustrating the alignment of a centripetal camera under Scheimpflug conditions;

[0053] Figure 4 A perspective view of two selected line scans performed sequentially in time by the registration unit, and a diagram showing the result of the height change of the object along the two selected line scans;

[0054] Figure 5 : A schematic diagram of triangulation used to measure height in the overlapping area of ​​two adjacent pericentric cameras;

[0055] Figure 6The present invention is shown in a side view transverse to the direction of object movement, with an enlarged view showing the height change of the substrate surface. The registration unit is schematically connected to the processing unit through the calculation unit to convert the height measurement achieved by triangulation into precise control of the height focused within the processing beam scanned perpendicular to the drawing plane.

[0056] Figure 7 The perspective view of the present invention includes a registration unit and a processing unit transverse to the substrate movement direction on a continuous substrate, which is tensioned and guided on a roller conveyor system.

[0057] Figure 8 A schematic diagram illustrating focus control achieved by tracking the current focal point using a movable objective lens;

[0058] Figure 9 A schematic diagram illustrating focus control achieved by tracking the current focal point using a fixed-angle reflector and a movable retroreflector.

[0059] Figure 10 : A perspective view of a flexible, bendable mirror element that uses a linear actuator to change the curvature of the mirror.

[0060] Figure 11 : Figure 10 A schematic diagram of a flexible mirror element;

[0061] Figure 12 : with the help of Figure 10 The diagram shows a flexible mirror element used for independent mirror control along the direction of movement and laterally of the substrate to achieve focus tracking.

[0062] According to Figure 1 In an advantageous basic variant, the registration unit 1 according to the invention comprises a plurality of cameras 11 arranged in a linear configuration such that their optical axes 111 are aligned with a line (scan line 23) in the plane of the substrate 2 transverse to the direction of movement of the substrate 2. Figure 2 , Figure 4 and Figure 6 As shown in the figure, the width of the substrate is completely covered by the partially overlapping image angle 112 or the field of view (FOV) constructed by the camera 11. This is possible based on a proximal rather than telecentric lens 15, in which the camera 11, as a line scan camera, is also equipped with one or more parallel rows of sensors 114.

[0063] Camera 11 (a so-called line scan camera, not in line) with a single sensor line 114 Figure 1The structure shown is scalable, meaning that each desired scan width can be achieved by combining multiple cameras 11. Since the accurate position of the target mark 22 can only be determined within the overlapping region 13 of the image angles 112 of two cameras 11, and the non-overlapping region cannot be accurately measured, if the overlapping region 13 is limited to half of the image angle 112, only half of the image angle 112 of the cameras 11 in the edge region of the substrate 2 can be used. Therefore, regarding the cameras 11 in the edge region of the substrate 2, Figure 1 Only half of the image angle 112 is shown.

[0064] The registration unit 1, with its elongated housing 12, is lateral to the target mark 22 to be scanned (only when...). Figure 2 and Figure 7 The substrate 2 is oriented according to the direction of movement of the substrate 2 (shown in the figure), wherein the substrate 2 is advantageously guided to the underside of the stage system 3 (shown only as a support surface). Figure 1 As shown in the side view, the registration unit 1 has densely packed illumination components spaced at small distances from the substrate surface 21. These illumination components are constructed in a way that serves as a light source 17, such that its illumination light only illuminates the substrate 2 and does not emit direct light toward the camera 11, and can illuminate with different spectral colors depending on the situation at different angles of incidence (dark-field illumination). Alternatively, the light from the light source 17 can be directly coupled into the recording optical path of the camera 11 as bright-field illumination (not shown).

[0065] Figure 1 The selected camera configuration consists of five cameras 11, each with an optical axis 111 aligned parallel to each other and a relatively large overlap area 13, which corresponds to half the image angle 112 of each camera 11. This enables gapless, near-linear scanning across the entire width of the substrate 2, allowing for smooth lateral movement of the substrate 2. Figure 1 As indicated by the arrows in the right-side side view, all target marks 22 along the width of the substrate can be detected in a single scan without mechanically moving the camera. A near-linear scan is achieved by a registration unit 1 equipped with a camera 11 in the form of a line scan camera. To generate a linear scan area transverse to the direction of movement of the substrate 2, the proximal camera 11, having an optical axis 111 in a plane (not shown), must be aligned in a certain way with the scan line 23 on the substrate surface 21 (only when...). Figure 2 (as shown in the image), so that its image angle 112 overlaps there and forms an overlapping region 13 along the scan line 23.

[0066] With a standard substrate width of 500-635 mm, for the currently required resolution (8-12 μm / pixel on substrate 2), a pericentric camera 11 (in the form of a line scan camera) with a scan length of 330 mm can be used in the focused area. This allows the registration unit 1 equipped with five cameras 11 to completely span the entire substrate width with a distance of 165 mm from the optical axis 111 of the cameras 11 in the direction of the substrate width (lateral extension of substrate 2). Each position of the scan line 23 is recorded simultaneously and redundantly in two different camera images of adjacent cameras 11. In this example, a total scan length of 660 mm is achieved by using five cameras 11 aligned with the scan lines 23. This allows for further close proximity of these cameras, even with positioning or mounting tolerances of the cameras 11 and / or variations in the height Δz of substrate 2, to ensure gapless double scanning. Since this total scan length exceeds the current maximum substrate width (635 mm) by 25 mm, an additional small overlap 14 of the overlap region 13 and reliable scanning across the edge region of substrate 2 can be achieved.

[0067] With the help of Figure 1 The layout of the camera 11 shown can scan a linear scanning area along the scan lines 23, with an aspect ratio exceeding 2000. Generally, the registration unit 1 can have scan lines 23 with an aspect ratio of several thousand to hundreds of thousands, wherein the scan width along the movement direction y of the substrate 2 can be adapted by electronically controlling (synchronizing) the readout mode and speed of the camera 11 by the computing unit 5.

[0068] according to Figure 1 Cameras 11, arranged along scan lines 23 using optical axes 111 that are parallel to each other, are positioned along scan lines 23 (only in...). Figure 2 , Figure 4 and Figure 6 (As shown in the figure) are arranged close to each other in such a way that adjacent cameras 11 form an overlap area 13 of at least half of the image angle 112. If, in the case of three or more cameras 11, even in the presence of height variation Δz of substrate 2 and mechanical adjustment errors, it should be ensured that the substrate surface 21 is covered without gaps by means of the overlap area 13, then there is another small overlap 14 with the image angle 112 of the next camera 11.

[0069] At the edge of the substrate 2 or on the scan line 23 formed by the camera 11, only half of the image angle 112 can be used based on the optical axis 111 of the camera 11, which is oriented perpendicular to the substrate 2. Therefore, an overlap region 13 always exists in the edge region of the substrate 2. This overlap region is necessary because, when using a pericentric camera 11, the position determination of the target mark 22 located on the substrate 2—the farther away it is from the optical axis 111—in the camera image is very sensitive to the distance of the substrate surface 21 from the focal plane F of each camera 11. n (only in) Figure 5 The distance is related to the target mark 22 (shown in the figure). Therefore, in order to determine the distance of the target mark 22 arbitrarily positioned on the surface 21 of the substrate 2, triangulation is performed on the images of two adjacent cameras 11 along the scan line 23. Triangulation can also be performed on any other point of the scan line 23, as long as there is an analyzable structure on the substrate 2 except for the target mark 22.

[0070] The accurate measurement of the height variation Δz of the substrate surface 21 along the z-direction is limited to the region with detectable points (analyzable structures). Therefore, the height profile of the substrate 2 must be supplemented by interpolation.

[0071] The triangulation of the distance from the substrate surface 21 is crucial for determining the scanning position of the target mark 22, because any height fluctuation Δz of the substrate surface 21 can lead to considerable measurement errors when determining the x and y coordinates of the target mark 22, especially as the target mark is closer to the edge of the image angle 112 of the camera 11.

[0072] The light source 17, positioned in the housing 12 of the registration unit 1 at a small distance from the substrate 2 and arranged with different angles of incidence, is adapted to selectively illuminate the scan line 23 defined by the incident point of the optical axis 111 of the camera 11.

[0073] As mentioned above, in accordance with Figure 1 In the camera layout, a corresponding half of the image recording area (image angle 112) cannot be used by the external camera 11. Figure 2 The alternative system structure shown, based on Scheimpflug's law, avoids the aforementioned drawbacks and optimally utilizes the scanning area within the entire image angle 112 of each of the adjacent cameras 11 that are tilted towards each other. However, the requirements for alignment and adjustment of the lens 15 and the camera 11 also increase.

[0074] Although Figure 1 The first gapless registration principle shown is designed for a total detection width of >635 mm (>25 inches) equipped with five cameras 11, but in the case of using... Figure 2Under the layout principle shown, the same detection width of the registration unit 1 can be scanned by only four cameras 11.

[0075] The detection width of all cameras 11 is set in a certain way: such that it is larger than the maximum processing area of ​​processing unit 4 (only when...). Figure 6 and Figure 7 (as shown in the diagram), that is, the maximum length greater than the processing path 41 on the substrate 2, for example, generated by the processing beam 45 scanned by a polygon scanner (only when...). Figure 6 and Figure 7 (as shown in the image).

[0076] The second condition regarding the number and layout of cameras 11 is the required object-side optical resolution, which is typically set to about 10 micrometers per pixel, depending on the size of the target mark to be detected, and is implemented as about 11 micrometers per pixel in the proposed example. The third boundary condition relates to the scanning speed, which should be adapted to a substrate speed of 1000 mm / s to 1800 mm / s for the desired throughput of the printed circuit board.

[0077] As a compromise between high readout speed, maximizing scan width based on large line length, and reasonable price of sensor line 114, camera 11 primarily employs a line scan camera with sensor line 114 having more than 3000 pixels and a side length of 11 micrometers x 11 micrometers. The required width of substrate 2 is adapted to the resolution of line scan camera by utilizing the imaging ratio of lens 15 of camera 11.

[0078] and Figure 1 They are different. Figure 2 The diagram shows a configuration consisting of four cameras 11, which have optical axes 111 tilted relative to each other in the same camera plane along scan line 23, so as to enable... Figure 1 The target marker 22 is registered without gaps within the same scanning area (635 mm). In this case, the two cameras 11 form a pair of cameras 11 positioned under Scheimpflug conditions, with their image angles 112 having a large overlap area 13, which preferably corresponds to the complete overlap of the image angles 112 of the two cameras 11.

[0079] If, viewed from the image angle 112 of two adjacent cameras 11, the width of the substrate 2 does not exceed the scan line 23, then the triangulation of each substrate point along the scan line 23 can be calculated based on two camera scans of only two adjacent cameras 11 tilted relative to each other. Otherwise, if the substrate width is large, other pairs of cameras 11 tilted relative to each other can be arranged together along the desired scan line 23 until the overlapping area 13 of the corresponding two cameras 11 covers the entire width of the substrate 2. In this case, the overlapping area 13 must at least make contact, but due to the possible height variation Δz of the substrate 2 and the mechanical mounting tolerances or adjustment tolerances of the cameras 11, there should also be a small overlap 14, which ensures, for the maximum height variation Δz and the mounting tolerances of the cameras 11, that the scan line 23 on the substrate 2 is always scanned without gaps by the additional overlap 14 of the overlapping area 13 formed by the complete overlap of the image angles 112 of each pair of cameras 11. Figure 2 The advantage of the camera configuration shown is that, in the simplest case, exactly two cameras 11 "see" the same area of ​​scan line 23 while following the Scheimpflug condition, and the image angle 112 of camera 11 does not have any unused partitions. Figure 1 In contrast, with the same length of scan line 23, i.e., with the same substrate width, one camera 11 can be saved.

[0080] Meanwhile, if there are enough target markers 22 or other scannable structures in the overlapping region 13, each pair of cameras 11 arranged under the Scheimpflug conditions can perform additional height triangulation to determine the height deviation Δz of the substrate surface 21.

[0081] Figure 3 One of two adjacent cameras 11 tilted relative to each other is shown, wherein the imaging of the object plane and the image plane is corrected by setting different tilt angles (or tilt of the lens relative to the camera, when the camera is considered to be separated from the lens 15) for the lens 15 and the sensor chip 113, and these tilt angles satisfy the Scheimpflug condition.

[0082] Figure 4 The sensor row 114 is schematically shown as a representative of the camera 11 of the registration unit 1, so as to illustrate the problem of the height of the fluctuation of the substrate surface 21 by taking substrate waviness as an example. Figure 4 The upper right portion of the diagram schematically illustrates a stage system 3 that moves along the y-direction, on which a substrate 2 is placed. Above the substrate 2 is a registration unit 1 reduced to sensor row 114, which aligns the target mark 22 (only along scan line 23) with the sensor row 114. Figure 2 and Figure 7 (As shown in the figure) Registration is performed. Regarding the substrate 2 to be scanned relative to the target mark 22, it is assumed or known that the substrate is unstable or flat with a wave pattern, which readily causes a height variation Δz on the substrate surface 21. As the sensor row 114 continues to move relative to the target mark 22 along the y-direction, different height distributions are recorded along the scan line 23 (along the x-direction), depending on the position of the target mark 22 in the image angle 112 of the pericentric camera 11. These height distributions reduce the accuracy of the xy position determination of the target mark 22.

[0083] Due to the feed of the worktable system 3, sensor line 114 first scans the dotted line scan line 23', resulting in the dashed outline in the figure below. After several readout steps, sensor line 114 detects the solid line scan line 23 and records the solid outline that deviates significantly from the dashed outline. This is done before passing through the machining unit 4 (only in...). Figure 6 During processing (as shown in the diagram), these locally varying height changes Δz can cause significant deviations in the exposure structure, which are caused by defocusing (broadening) of the processing beam 45. Defocusing can only be remedied by refocusing if the position of the height change Δz is accurately measured and the processing beam 45 is tracked along the processing line 41 that matches the position of the target mark 22 and the height change Δz.

[0084] Figure 5 This illustrates an example of determining the height change Δz of substrate surface 21 of substrate 2 using triangulation, wherein two adjacent cameras 11 have parallel optical axes 111 and two focal planes F of sensor chip 113 are in the same plane. n and F n+1 (Here, n = 0) and an overlap region 13 (i.e., a linear scan region) is generated for the image angles 112 of the two cameras 11. This overlap region is necessary because, when imaging with the pericentric lens 15, the position of the target marker detected in the XY plane on the substrate surface 21 is sensitively related to the object position along the z-direction. Therefore, the target marker 22 to be registered is detected using camera images from a pair of adjacent cameras 11, and the target marker positions x and z are calculated according to triangulation, where, in the z-direction, the height variation is referenced to two reference planes, labeled as substrate surface 21 in the lower plane and substrate surface 21' in the upper plane to represent the surface variation. Finally, the obtained x-position and the obtained height z are determined based on the positions of the different calibration values ​​Z1 and Z0 detected in the two cameras 11 relative to the calibration planes of substrate surfaces 21 and 21' in the following manner.

[0085]

[0086]

[0087] This indicates that, in addition to accurately determining the x-position of the target marker, the z-position can also be determined relative to the calibration planes Z1 and Z0. This enables (absolute) height measurement.

[0088] Figure 6 To and Figure 1 The right-side sectional view is similar to the side view of registration unit 1, and the target marker 22 is schematically shown (only when...). Figure 2 and Figure 7 The measured position data (shown in the figure) is correlated with the processing unit 4 through the calculation unit 5. The problem of uneven substrate 2 is shown in a non-realistic manner and magnified for the waveform substrate surface 21, wherein the stage system 3 is assumed to be a precision stage.

[0089] As the stage system 3 moves along the y-direction, the registration unit 1 detects the x and y positions of any target mark 22 arbitrarily arranged on the substrate 2 in the linear scanning area (scan lines 23 composed of multiple line-scan cameras perpendicular to the drawing plane) through sequential line scanning. Therefore, based on the image angle 112 formed by the camera 11 (only when...), the x and y positions of the target mark 22 are determined. Figure 1 and Figure 2 The overlapping area 13 (visible in the image) is double-scanned at each substrate point by two adjacent cameras 11, so that not only can the precise target mark position be determined by triangulation, but also the height change Δz of the substrate surface 21 can be calculated. Then, based on the measured value of the actual substrate height z at each point detectable by the imaging structure of the scan line 23, in addition to the conventional alignment of the two-dimensional structure of the processed pattern relative to the actual position of the target mark 22, the processing focus FP of the processing beam 45 (only when the processing focus FP is located) is also tracked along the processing path 41 in the calculation unit 5 by focus tracking 43. Figure 8 and 9 (As shown in the figure) Adapts to the height variation Δz point of the substrate surface 21.

[0090] In the present invention Figure 7 In the exemplary embodiment shown, substrate 2 is a continuous substrate, which is tensioned and guided from one roller to another (not shown) by a roller conveyor system 31. The diameter of the rollers in the roller conveyor system 31 can be between 200 and 500 mm.

[0091] Based on the tensioned material guidance of the continuous substrate 2, the scan line 23 of the camera 11 in the registration unit 1 (in) Figure 7 (Not visible in the middle) can be arranged in front of the contact line between the substrate 2 and the roller system 31.

[0092] In this case, the registration unit 1 with two centripetal cameras 11 is designed in such a way that the scanning areas of the two cameras 11 form a scan line 23 (invisible), which also extends across the edge of the substrate 2 so as to be able to detect the calibration mark 32 on the roller system 31 together. For the necessity and processing of the calibration mark 32, see DE 10 2019 128 198.9, which was not disclosed in advance.

[0093] In other cases, such as for Figure 1 and Figure 2 As described in the planar substrate 2, the registration unit 1 detects all target marks 22 located on the substrate 2 that are guided by the roller system 31, regardless of their location on the substrate 2.

[0094] Because the substrate 2, as a continuous substrate, is guided on the roller system 31 in a tensioned manner and thus does not experience any height change Δz due to the waviness of the substrate 2, in this embodiment of the invention, local point-by-point height measurements can be omitted, and the height measurement of the substrate surface 21 can be limited to a small overlap region 13 of the image angles 112 of the two cameras 11. In this overlap region 13, the overlap can be much smaller than half the image angles 112 of the two cameras 11, but at least 1 / 50 (>5 mm) of the detection area of ​​the registration unit 1 (i.e., the substrate width together with the edge area of ​​the roller system 31), preferably between 1 / 40 and 1 / 10, and particularly preferably between 1 / 35 and 1 / 25 (approximately 10-15 mm), to match... Figure 5 The triangulation calculation is performed in a similar manner to the description above. In this example, the thickness of substrate 2 can be determined based on this triangulation, and the thickness variation over the total length of the continuous substrate can be measured as appropriate. However, it is usually not necessary to perform triangulation calculations as described above. Figure 6 As described, the focus is rapidly changed along the processing path 41 using focus tracking 43. However, thickness information, i.e., the z-measurement of the height of the substrate surface 21, is essential for calculating the position of the target mark 22 along the x and y directions. But after the initial single measurement (and individual measurements repeated from time to time, as needed), this thickness information can be retained to calculate the position of all target marks 22 throughout the continuous substrate. The processing unit 4, aligned with the substrate surface 21 in another radial plane of the roller system 31, emits a scanning processing beam 45 to imprint a two-dimensional structure on the processing path 41 of the substrate surface 21. However, this processing unit can also be arranged on the roller system 31 in a manner completely opposite to that of the registration unit 1 in a common plane (e.g., the axial plane of the roller system 31) (not shown).

[0095] Just like Figure 4 and Figure 6Regarding the height variation Δz caused by the unevenness of the substrate 2 or substrate surface 21, according to the present invention, it is necessary to perform rapid point-focus tracking 43 on the processing beam 45 along the linear processing path 41, so that the processing beam degenerates into a processing line with variable height.

[0096] Conventional autofocus systems are not suitable for rapid focus changes along the z-direction of a processing beam 45 that scans along the x-direction, which is a laser beam used in direct photolithography. The basis for rapidly achieving local focus changes along the processing path 41 according to the present invention lies in detecting the height change Δz of the substrate 2 through redundant double image recording using a pericentric camera 11 with a gapless overlapping region 13 along a linear scanning area (scan line 23) having its image angle 112 while the target mark 22 is registered. Through height measurement of each image point of the scan line 23, calculated by triangulation and located upstream of the processing unit 4 in time and space, in addition to the conventional adaptation of the data decomposed into the processing path 41 of the two-dimensional structural pattern with the position of the substrate 2 detected by the measured target mark 22, the point-like change of the processing focus FP along the processing path 41 of the processing beam 45 can be calculated using the calculation unit 5 based on the height change Δz measured along the scan line 23 of the registration unit 1 and calculated in the calculation unit 5.

[0097] For rapid focus changes along the processing path 41, additional components are needed to supplement the commonly used focusing optics 44. These components must possess a quality such that their scanning frequency exceeds that of the processing beam 45 in the lateral x-direction of motion relative to the substrate 2 by at least twice. The frequency of focus changes is preferably between two and thirty times, particularly preferably between five and twenty times, the scanning frequency of the processing beam 45; when using a polygon scanner, this frequency is in the range of 0.5 to 1 kHz.

[0098] to this end, Figure 8 This illustrates a first scheme for achieving focus tracking 43 using a movable lens 431.

[0099] according to Figure 9 Another implementation variant of the focus tracking 43 sets a fixed corner mirror 432 and a movable retroreflector 433, wherein the corner mirror 432, arranged in the focused beam, couples the focused beam out to the retroreflector 433 and couples it back in after reflecting it, and the retroreflector 433 moves toward or away from the corner mirror 432 in order to move the focus FP in the z direction.

[0100] Figures 10 to 12 Another mirror-based implementation of focus tracking 43 is described. To this end, Figure 10 and Figure 11A mirror assembly with a variable curvature mirror is shown, which is referred to below as a flexible bendable mirror 434. For this purpose, Figure 11 The working principle is shown as an equivalent circuit diagram of the machine, in which a flexible reflector 434 is movably hinged to a reflector support 436 of the substrate in the edge region and contacts a linear actuator in the form of a piezoelectric stack 435 in the central region, which is supported on the same substrate as the reflector support 436. Figure 10 The diagram illustrates the physical realization of the mirror assembly as a nearly monolithic component, where the mirror support 436 acts as a cuboid supporting the elastically bendable mirror 434, which is a thinned, bulging metal plate elastically supported on two parallel linear solid bearings. Thus, the bendable mirror 434 is a convex or concave cylindrical mirror whose curvature can only be changed in one spatial direction.

[0101] according to Figure 12 In one embodiment, two intersecting (e.g., along the x and y directions) oriented, resiliently bendable reflectors 434 are arranged for the processing beam 45 (only in... Figure 7 and Figure 8 Focusing control is achieved in the folded optical path of a multi-element focusing optics 44 (shown in the figure). In this case, fast focus tracking 43 is integrated into the conventional focusing optics 44, wherein a point-like generation of a two-dimensional structure (not shown) that is locally adapted to the position of the target mark 22 and suitable for processing is preset in the intermediate image plane 441. This structure is then transferred in the image plane 442 to an advantageously used polygon scanner (not shown) to guide the processing beam 45 onto the substrate 2 along the height-controlled processing line 41. The advantage of this fast focus control technique with two effective one-dimensional focus tracking 43s is that it can operate with different focuses along the x-direction (scanning direction of the processing beam 45) and the y-direction (substrate movement direction). Therefore, when using astigmatism or other aspherical optics, different focus variations can be operated along the scanning direction of the processing beam 45 and the cross-scanning direction (if needed).

[0102] With the present invention, a gapless and nearly one-dimensional scan line 23 can be achieved on a substrate 2 with height variations or curvature by using a centripetal camera 11 for target mark registration. This allows not only the detection of target marks 22 but also the measurement of the height and height variation Δz of the substrate 2. Based on a double scan along the scan line 23 of the registration unit 1 using two cameras 11 positioned in different ways, triangulation and calculation can be performed in the overlapping region 13 by a centripetal line scan camera whose image angles 112 overlap, allowing for height measurement. By point-based height measurement, the height variation Δz of the substrate surface 21 can be taken into account by rapid focus control in the processing unit 4, which also operates using a linear processing path 41, and these height variations can be corrected by focus tracking 43 or converted into height-matched processing lines point-by-point along the processing path 41 during the control of the processing beam 45.

[0103] Furthermore, the dynamic focus adaptation along the processing path 41 can also correct for known focus deviations (caused by optical design and optical production) related to the imaging of the processing beam 45 scanned through the processing path 41.

[0104] Appendix Label Table

[0105] 1. Registration Unit

[0106] 11 cameras

[0107] 111 optical axis

[0108] 112 Image Angle

[0109] 113 Sensor Chip

[0110] 114 Sensor Line

[0111] 12 (Registration unit) Housing

[0112] 13 (like angles) overlapping areas

[0113] 14 (like horns) slightly overlapping

[0114] 15 (camera 11) shots

[0115] 151 Lens Adapter

[0116] 16 (Image planes of registration unit 1)

[0117] 161 Camera Adapter

[0118] 162 (surface normal of sensor chip 113)

[0119] 17 (Light source for illuminating the scanned area)

[0120] 2 Substrate

[0121] 21, 21' substrate surface

[0122] 22. Target Marking (Target)

[0123] 23, 23' scan lines

[0124] 3. Workbench System

[0125] 31 (Roller system for continuous substrates)

[0126] 32 Calibration Marks

[0127] 4 processing units

[0128] 41 Processing Path

[0129] 43 Focus Tracking

[0130] 431 Movable Lens

[0131] 432 (fixed) corner reflector

[0132] 433 (movable) retroreflector

[0133] 434 Flexible reflector

[0134] 435 Piezoelectric Stack (Linear Actuator)

[0135] 436 Reflector Support

[0136] 437 Solid Connector

[0137] 44 Focusing Optical Devices

[0138] 441 Intermediate Image

[0139] 442 Image Plane

[0140] 5. Calculation Unit

[0141] F n ,F n+1 focal plane

[0142] FP Focus (Tracking)

[0143] x Lateral direction (across the width of substrate 2)

[0144] y: Direction of motion (along the length of substrate 2)

[0145] z Focusing direction (at the height of substrate 2)

[0146] Δz (height variation of substrate surface 21).

Claims

1. An apparatus for exposure control when directly exposing two-dimensional structures in a light-sensitive coating on a substrate in a lithographic manner, the apparatus comprising a registration unit for registering a target mark located on the substrate surface, a movable table system for supporting and moving the substrate one-dimensionally in a defined manner below the registration unit, a machining unit with a controllable linear machining path for lithographically machining the substrate by means of a machining beam to print the two-dimensional structures, and a calculation unit for controlling the alignment between machining path and substrate by means of a local adaptation of the lithographic machining depending on the position of the substrate determined by the registered target mark, characterized in that - a plurality of close-up cameras (11) for forming gapless linear scan regions within a predetermined width of the substrate (2) are arranged in the registration unit (1) in linear alignment transverse to the one-dimensional movement of the substrate (2) and have an extended image angle (112) in the direction of the gapless linear scan regions, wherein the image angles (112) of adjacent close-up cameras (11) have an overlap region (13) along the gapless linear scan regions in order to detect redundant image recordings of the substrate (2) of the adjacent close-up cameras (11) in the overlap region (13), and - the calculation unit (5) has means for calculating the position of the target mark (22) from the redundant image recordings in the overlap region (13) of the adjacent close-up cameras (11) in the case of a supplementary use of the height position of the target mark (22) determined by means of a triangulation of the distance to the substrate surface (21).

2. The apparatus according to claim 1, characterized in that - the registration unit (1) is equipped in such a way with a plurality of close-up cameras (11) for generating linearly continuous gapless sensor regions that the image angles (112) of adjacent cameras (11) have an overlap region (13) which is at least as large as half of the image angle (112), wherein the calculation unit (5) is arranged in such a way that an arbitrary target mark (22) positioned within the width of the substrate (2) is determined independently of its position in the gapless continuous overlap region (13) of the image angles (112) of adjacent close-up cameras (11) by triangulating the distance at an arbitrary position of the substrate surface (21).

3. The apparatus according to claim 1 or 2, characterized in that - the cameras (11) are line scan cameras in order to form the gapless linear scan regions in a narrow and overlapping manner over the entire width of the substrate (2) by means of cameras (11) having a large scan length.

4. The apparatus according to claim 1 or 2, characterized in that - the cameras (11) are aligned to the substrate surface (21) by means of optical axes (111) which are parallel to one another, wherein the overlap regions (13) of the image angles (112) of all cameras (11) are of the same size. ​ ​ ​ ​ ​ 5. The device according to claim 1 or 2, characterized in that two adjacent cameras (11) are aligned to the substrate surface (21) by means of optical axes (111) which are tilted relative to one another, wherein the overlapping region (13) of the image angles (112) of the cameras (11) which are tilted relative to one another is designed in such a way that the image angles (112) of both cameras (11) overlap completely on the substrate surface (21).

6. The device according to claim 5, characterized in that the overlapping region (13) formed by cameras (11) which are tilted relative to one another in pairs is connected without gaps to another overlapping region (13) at least up to the extent of the overlapping region (13) corresponding to the width of the substrate (2), wherein an overlap (14) is provided between pairs of cameras (11) which are tilted relative to one another in order to ensure a gapless linear scanning region of the registration unit (1) for all permissible height variations (Δz) of the substrate surface (21).

7. The device according to claim 5, characterized in that the cameras (11) which are tilted relative to one another in pairs are arranged in such a way that they are subject to the Scheimpflug condition.

8. The device according to claim 1 or 2, characterized in that the calculation unit (5) also has a control device for rapid focus tracking (43) of the processing unit (4) along the processing path (41) depending on the height variation Δz of the substrate surface (21), which comprises a triangulation of target markings (22) or any imaging structure of the substrate surface (21) based on redundant image recordings in the overlapping region (13) of adjacent telecentric cameras (11).

9. The device according to claim 1 or 2, characterized in that the registration unit (1) has only two telecentric cameras arranged transversely to the direction of movement of the substrate (2) on the gapless linear scanning region for target marking detection, the overlapping region (13) of which is one to three times the image angle (112) of the cameras (11) when the substrate (2) is guided on a roller system (31) as a flexible continuous substrate in a tensioned and height-variation-free manner, wherein the overlapping region (13) of the image angles (112) of both cameras (11) is designed in such a way that the triangulation is applied on the roller system (31) in order to accurately determine the thickness of the substrate (2) in the overlapping region (13) of the image angles (112) and the thickness is assumed to be constant for the entire width of the roller system (31).

10. The device according to claim 1 or 2, characterized in that the registration unit (1) is equipped with a light source (17) for illuminating the gapless linear scanning region, which is arranged in a uniformly distributed manner in a housing (12) in order to achieve a uniformly illuminated gapless linear scanning region by means of dark-field or bright-field illumination.

11. The apparatus of claim 10, wherein, The light source (17) is adapted for continuous illumination and has means for controlling at least one of the properties of brightness, angle of incidence or spectral range.

12. The apparatus of claim 10, wherein, The light source (17) is adapted for continuous illumination in order to realize the image recording by controlling the integration time of the sensor rows (114) of the camera (11) by means of the electronic shutter principle.

13. The apparatus according to claim 1 or 2, characterized in that A focus tracking (43) for height variations (Δz) of the substrate (2) is integrated into the processing unit (4) for real-time adaptation of the height variations (Δz) determined by the calculation unit (5) by triangulation from redundant image recordings by two adjacent cameras (11) of the registration unit (1) by fast focus correction of each pixel of the registration unit (1), wherein the focus tracking (43) is controlled based on variations of lens or mirror position or curvature.

14. The apparatus according to claim 13, characterized in that The focus tracking (43) is controlled based on variations of mirror curvature at least in the lateral direction x relative to the direction of movement y of the substrate (2).

15. The apparatus according to claim 14, characterized in that The focus tracking (43) is controlled based on variations of mirror curvature in the direction of movement y of the substrate (2) and variations of mirror curvature in the lateral direction x separately.

16. The apparatus according to claim 13, characterized in that The focus tracking (43) is controlled by changing the mirror curvature by means of piezoelectric elements (435).

17. The apparatus of claim 13, wherein, The focus tracking (43) based on variations of lens or mirror position or curvature is used to correct imaging-related focus deviations of upstream focusing optics (44) or other upstream optical elements due to optical design or optical production.

18. A method for exposure control in the direct exposure of a two-dimensional structure in a light-sensitive coating on a substrate, comprising the following steps: - arranging a plurality of near-to-center cameras (11) in a registration unit (1) in a gapless linear scan field in a manner transverse to the direction of movement of the substrate (2) for detecting target marks (22) located on the substrate (2), wherein the near-to-center cameras (11) have an extended angle of view (112) along the gapless linear scan field including an overlap region (13) formed by adjacent near-to-center cameras (11) in order to obtain redundant image recordings of the substrate (2) in the overlap region (13) by adjacent cameras (11), - moving the substrate (2) under the registration unit (1) in a defined one-dimensional movement on a movable table system (3), - providing a processing unit (4) for lithographically producing the two-dimensional structure by means of a processing beam (45) which can be controlled along a linear processing path (41), - providing a processing unit (4) for lithographically producing the two-dimensional structure by means of a processing beam (45) which can be controlled along a linear processing path (41), - detecting the spatial position of target marks (22) arbitrarily distributed over the width of the substrate (2) for length, width and height position while the substrate (2) is being passed through the gapless linear scanning area of the registration unit (1) in a single pass, - determining the position of target marks (22) arbitrarily distributed over the width of the substrate (2) from redundant image recordings in the overlap area (13) of the adjacent close-up cameras (11) in the case of a supplementary use of the height position of the target marks (22) determined from the triangulation of the distance to the substrate surface (21) from the redundant image recordings of the adjacent close-up cameras (11), - calculating data for the processing unit (4) for aligning and locally adapting the processing of substrates (2) having a two-dimensional structure for controlling the processing beam (45) along a linear processing path (41) transverse to the direction of movement of the substrate (2), and - controlling the alignment between the processing path (41) and the substrate (2) and locally adapting the lithographic processing in accordance with the position of the substrate (2) determined by the registered target marks (22).

19. The method according to claim 18, wherein the calculation of the spatial position of target marks (22) arbitrarily distributed over the width of the substrate (2) is extended to the triangulation of other detectable structures of the substrate (2) in the redundant images recorded in the overlap area (13) on the basis of the height position at the time of passage of the substrate (2), and the rapid focus adjustment of the focus point (FP) of the processing beam (45) along the processing path (41) is carried out by means of focus tracking (43) on the basis of the control of the lens or mirror position or the mirror curvature.

20. The method according to claim 19, wherein the rapid focus adjustment of the focus point (FP) of the processing beam (45) is carried out at a frequency which is two to thirty times higher than the conventional scanning frequency of the processing beam (45).

21. The method according to claim 20, wherein the rapid focus adjustment of the focus point (FP) of the processing beam (45) is carried out at a frequency which is two to thirty times higher than the conventional scanning frequency of the processing beam (45).

Images