Method and device for analyzing images of microlithographic microstructured samples

Abstract

To provide an improved method and device for analyzing an image of a microlithographic microstructured sample.SOLUTION: A method for analyzing a two-dimensional intensity distribution image of a microstructured sample which comprises at least one first segment and at least one second segment which is raised with respect to the first segment comprises: determining multiple edge candidates (312, 314) of the at least one second segment on the basis of gradients (316, 318) of the two-dimensional intensity distribution; determining a one-dimensional intensity distribution (320) of the image in a direction (R) perpendicular to the multiple edge candidates, where the one-dimensional intensity distribution has a first region (306') with a first mean intensity value (I1) and a second region (308') with a second mean intensity value (I2) greater than the first mean intensity value; and determining, among the multiple edge candidates, the edge candidate closest to the first region as an edge of the at least one second segment.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to a method and device for analyzing images of microlithographic microstructured samples.

[0002] The content of priority application DE 10 2023 113 273.3 is incorporated by reference in its entirety. [Background technology]

[0003] Microlithography is used to manufacture finely structured components, such as integrated circuits. Microlithography processes are carried out using a lithography apparatus comprising an illumination system and a projection system. An image of a mask (reticle), illuminated using the illumination system, is then projected using the projection system onto a substrate, such as a silicon wafer, coated with a photosensitive layer (photoresist) and positioned in the image plane of the projection system, in order to transfer the mask structure into a photosensitive coating on the substrate.

[0004] Driven by the desire for smaller structures in the manufacture of integrated circuits, EUV lithography tools are currently under development that use light having wavelengths in the range of 0.1 nm to 30 nm, specifically 13.5 nm.

[0005] As feature sizes on both the masks used in lithographic processes and the microlithographically structured wafers become smaller, the analysis and processing or repair of these components becomes an increasingly challenging task in practice.

[0006] For the purpose of analyzing microstructured samples, such as microstructured lithography masks and wafers, images captured by a microscope are evaluated to determine the differences between each measured image and a design image containing the intended structure of the sample. In particular, the images captured by a microscope are images captured using an electron or ion beam (e.g., scanning electron microscope images, or SEM images for short). The differences determined based on such images between each measured image and a design image containing the intended structure of the sample are used as a basis for processing and / or repairing the sample. Generally, the image to be analyzed in this case consists of a large number of pixels, each of which is assigned an intensity value as a "grayscale value."

[0007] For example, evaluating an image captured by a microscope of a microstructured sample involves detecting or extracting the contours of the structure of the microstructured sample (detecting or extracting edges). Conventional techniques for determining the edges of structures in a sample are based, for example, on forming the gradient (i.e., first derivative) of the two-dimensional intensity distribution (i.e., of the grayscale value profile) of the image, as described, for example, in DE 10 2021 113 764. Edge detection can be made even more difficult due to the low signal-to-noise ratio of the utilized image of the microstructured sample. Furthermore, mask artifacts (e.g., granulation on the surface) and imaging artifacts (e.g., edge brightening or brightening due to charge) can further complicate edge detection. Summary of the Invention

[0008] Against this background, it is an object of the present invention to provide an improved method and an improved device for analyzing images of microlithographic topographical specimens.

[0009] Therefore, a method for analyzing an image of a microlithographic microstructure sample is proposed. The sample includes at least one first segment and at least one second segment having an edge and rising opposite the first segment. Furthermore, the image includes a plurality of pixels and a two-dimensional intensity distribution according to the pixels. The method includes: a) determining a plurality of edge candidates for an image representation of edges of at least one second segment based on a gradient of the two-dimensional intensity distribution; b) determining a one-dimensional intensity distribution of the image in a direction perpendicular to the plurality of edge candidates, wherein in said direction the one-dimensional intensity distribution includes a first region having a first average intensity value, the plurality of edge candidates, and a second region having a second average intensity value greater than the first average intensity value; and c) determining an edge candidate from the plurality of edge candidates that is closest to the first region of the one-dimensional intensity distribution as an image representation of an edge of the at least one second segment.

[0010] As a result, multiple candidates for the image representation of the edge of at least one second segment of the microstructure sample can be initially determined in the image of the sample. For example, determining image artifacts as edge candidates (e.g., due to a low threshold value applied to the absolute value of the gradient) can be tolerated at this stage. Next, an edge candidate can be selected from the multiple determined edge candidates by evaluating the one-dimensional intensity distribution of the image in a direction perpendicular (orthogonal) to the multiple determined edge candidates, and can thereby be determined as the image representation of the edge of the second segment. In particular, an edge candidate is selected from the multiple determined edge candidates by determining the edge candidate located closest to the first region of the one-dimensional intensity distribution, and therefore closest to the region of the one-dimensional intensity distribution having a lower average intensity value (i.e., a darker region in the image), as the image representation of the edge of the second segment. The darker region in the captured image typically corresponds to a structure located lower in the sample (i.e., at least one first segment of the sample).

[0011] As a result, the orientation or position of the edges of the microstructured specimen can be better detected. Artifacts, in particular, are also better rejected. In particular, the positions of the edges determined by the proposed method are more precise in all details and more accurate in their positioning relative to the geometry of at least one second segment. Furthermore, compared to conventional methods, the positions of edges located closer to the orientation of the lower structures of the specimen are determined.

[0012] For example, a microscope-captured image of a sample is an image captured using a particle beam, such as an electron beam or an ion beam. A microscope-captured image of a sample is, for example, a scanning electron microscope image (SEM image) of a sample.

[0013] For example, at least a portion of the sample is captured in a microscope-captured image of the sample. Further, the microscope-captured image of the sample specifically captures the first and second segments and at least a portion of an edge of the second segment that separates the second segment from the first segment.

[0014] In particular, the image analyzed by the method comprises a plurality of two-dimensionally arranged pixels, each pixel being assigned a respective intensity value, and in particular, the two-dimensionally arranged intensity values ​​form a two-dimensional intensity distribution of the image.

[0015] For example, at least one second segment has a full, closed edge of a circumference parallel to the plane of the main extent of the sample and / or perpendicular to the line of sight of the image recording device. For example, the image representation of the edge of the second segment determined in the method may correspond to a portion of the full edge.

[0016] For example, at least one first and one second segment are connected regions parallel to the plane of the main extent of the sample and / or perpendicular to the line of sight of the image recording device, respectively, such that any two points within such a region can always be connected by a path that lies entirely within this region.

[0017] In particular, the edge of at least one second segment of the specimen is the physical boundary of the second segment that separates the second segment from the first segment.

[0018] For example, in step (a), a plurality of parallel edge candidates for the image representation of the edges of the at least one second segment are determined based on the gradient of the two-dimensional intensity distribution.

[0019] In particular, based on the gradient of the two-dimensional intensity distribution, in step (a), a plurality of edge candidates are determined for one and the same edge of at least one second segment (for example, for one and the same portion of the entire edge).

[0020] For example, in step (a), determining a plurality of edge candidates for the image representation of the edge of at least one second segment based on the gradient of the two-dimensional intensity distribution may be performed based on any suitable known process, such as "Canny," "Laplacian of Gaussian," "Sobel," etc. Two or more of these processes may be applied (e.g., sequentially). Additionally or alternatively, one and the same edge extraction process may be applied multiple times using different parameter settings.

[0021] For example, determining multiple edge candidates for an image representation of at least one second segment edge based on the gradient of a two-dimensional intensity distribution includes determining the gradient of the intensity distribution at each pixel in the image. The gradient of a particular pixel may be determined, for example, based on an evaluation of a predetermined number of pixels surrounding the pixel. For example, this predetermined number may include pixels arranged around a central pixel within a square having a size of 3x3 pixels, 5x5 pixels, 7x7 pixels, 9x9 pixels, and / or 11x11 pixels. In this case, individual pixels may be included in the gradient calculation using different predetermined weights. For example, determining multiple edge candidates for the second segment may include determining a matrix of gradients (e.g., a gradient image) from the original image (i.e., the two-dimensional intensity distribution). The edge of the imaged segment is located at the pixel where the intensity (brightness) of the original image has the greatest change, and therefore the gradient image has the highest intensity. In other words, the edge corresponds to an area of ​​high gradient in the intensity distribution.

[0022] For example, determining a plurality of edge candidates for the image representation of the edge of the second segment in step (a) includes determining image pixels that are candidate edge pixels.

[0023] In particular, in step (c), an edge candidate from among the plurality of edge candidates that is closest to the first region of the one-dimensional intensity distribution, either locally or with respect to its position in an orthogonal direction, is determined as an image representation of an edge of at least one second segment.

[0024] For example, determining a one-dimensional intensity distribution of the image in a direction perpendicular to the plurality of determined edge candidates may include averaging over a plurality of pixels in a direction parallel to the plurality of determined edge candidates to increase the signal-to-noise ratio of the determined one-dimensional intensity distribution.

[0025] For example, when determining a one-dimensional intensity distribution of an image, "perpendicular to a plurality of determined edge candidates" includes being perpendicular to one, some, or all of the plurality of determined edge candidates.

[0026] The one-dimensional intensity distribution around the plurality of determined edge candidates includes first and second regions having different average intensity values. In other words, the plurality of determined edge candidates are flanked by a brighter region (second region) and a darker region (first region). By taking these regions into consideration, the method can better select a suitable edge from the plurality of edge candidates.

[0027] The first and second regions of the one-dimensional intensity distribution correspond specifically to regions of the sample that do not have edges, i.e., regions of the one-dimensional intensity distribution that correspond specifically to regions of the sample where no edge candidates were determined in step (a).

[0028] For example, a microstructured sample has a flat shape with a plane of main area and a height direction arranged perpendicular to the plane of main area. For example, a microscopic image of the sample is recorded using an image recording device, the line of sight of the image recording device being arranged parallel to the height direction of the sample.

[0029] For example, at least one first segment of the sample has a first height relative to the height direction of the sample. For example, at least one second segment of the sample has a second height relative to the height direction of the sample that is greater than the first height. For example, an edge of at least one second segment includes an edge wall. The edge wall may be disposed perpendicular to the plane of the main extent of the sample and parallel to the height direction. However, the edge wall may also be disposed at an angle relative to the plane of the main extent of the sample.

[0030] According to an embodiment, a first region of the one-dimensional intensity distribution of the image is based on an image representation of at least one first segment of the sample, and a second region of the one-dimensional intensity distribution of the image is based on an image representation of at least one second segment of the sample.

[0031] Thus, from among the multiple determined edge candidates, the edge candidate that is located closest to the darker area of ​​the image, corresponding to the lower structure of the sample (first segment) in the orthogonal direction, is selected and is consequently determined as the image representation of the edge of the second segment.

[0032] According to a further embodiment, at least one first segment of the sample comprises a first material and at least one second segment of the sample comprises a second material different from the first material.

[0033] For example, the exposed surface of the at least one first segment and the exposed surface of the at least one second segment may comprise different materials from one another, among other things.

[0034] According to a further embodiment, due to a difference in material between at least one first and second segment of the sample, a second average intensity value within a second region of the one-dimensional intensity distribution of the image is greater than the first average intensity value within the first region of the one-dimensional intensity distribution of the image.

[0035] Thus, different materials of the sample imaged with different brightness (intensity values) in the image in orthogonal directions are located to the left and right of the plurality of determined edge candidates, where the different brightness (intensity values) caused by the different materials in the image are used to select the best edge candidate from the plurality of edge candidates and, in particular, to reject artifacts.

[0036] For example, the material difference between the at least one first and second segment of the sample is the material difference between the (exposed) surfaces of the at least one first and second segment of the sample.

[0037] According to a further embodiment, at least one first and second segment of the sample comprises the same material.

[0038] For example, the exposed surface of the at least one first segment and the exposed surface of the at least one second segment, in particular, comprise the same material.

[0039] According to a further embodiment, due to a shadow formed adjacent to an edge of at least one second segment of the sample, the second average intensity value within the second region of the one-dimensional intensity distribution of the image is smaller than the first average intensity value within the first region of the one-dimensional intensity distribution of the image.

[0040] As a result of the shadow formed by the second segment, the area of ​​the first segment adjacent to the second segment is imaged in the image with a lower brightness (smaller intensity value) than the second segment. The difference in brightness (intensity value) between the imaged second segment and the shadowed area of ​​the first segment adjacent to the second segment caused by the shadow formation is now used to select the most suitable edge candidate from among the multiple edge candidates.

[0041] In other words, due to the shadowing, the average intensity value within the second region of the one-dimensional intensity distribution is greater than the average intensity value within the shadowed region of the one-dimensional intensity distribution.

[0042] According to a further embodiment, when determining a plurality of edge candidates based on the gradient of the two-dimensional intensity distribution, a predefined threshold is applied in such a way that for gradients of the two-dimensional intensity distribution having an absolute value greater than the predefined threshold, a corresponding edge candidate is determined, and for gradients of the two-dimensional intensity distribution having an absolute value less than or equal to the predefined threshold, no edge candidate is determined.

[0043] Setting a low threshold value may capture edges that are weakly imaged in the image, but may increase the number of artifacts among the determined edge candidates. Setting a higher threshold value may reduce the number of artifacts among the determined edge candidates, but may result in edges that are very weakly imaged in the image not being captured.

[0044] According to a further embodiment, before step (a), image processing occurs to reduce noise components in the two-dimensional intensity distribution.

[0045] Within the scope of image processing to reduce noise components, one or more suitable image smoothing processes may be applied. Exemplary suitable processes include binning, Gaussian filtering, low-pass filtering, etc. Merely by way of example, it is herein assumed that a plurality of (e.g., four or possibly more than four or less than three) mutually adjacent pixels can in each case be replaced by a single (e.g., average) pixel, which is then assigned the average intensity value of the plurality of combined pixels.

[0046] According to further embodiments, the microstructure specimen is designed for an operating wavelength of less than 250 nm, less than 200 nm, less than 100 nm, and / or less than 15 nm; and / or Microstructured samples are lithographic masks, in particular EUV or DUV lithographic masks, and / or wafers structured by microlithography.

[0047] For example, a DUV lithography mask is a transmissive photomask in which the pattern imaged during lithography is realized in the form of an absorbent (i.e., opaque or partially opaque) coating (which corresponds to the second segment) on a transparent substrate (which corresponds to the first segment).

[0048] For example, an EUV lithography mask is a reflective photomask in which the imaged pattern is realized in the form of an absorber coating (which corresponds to the second segment) on a reflective substrate (which corresponds to the first segment).

[0049] In particular, lithography masks are used in lithography equipment. For example, the lithography equipment can be an EUV lithography equipment or a DUV lithography equipment. EUV stands for "extreme ultraviolet" and refers to a working light wavelength in the range of 0.1 nm to 30 nm, in particular 13.5 nm. Furthermore, DUV stands for "deep ultraviolet" and refers to a working light wavelength in the range of 30 nm to 250 nm.

[0050] An EUV or DUV lithography apparatus comprises an illumination system and a projection system. In particular, using an EUV or DUV lithography apparatus, an image of a lithography mask (reticle) illuminated by the illumination system is projected by the projection system onto a substrate, e.g., a silicon wafer, coated with a photosensitive layer (photoresist) and positioned in the image plane of the projection system, in order to transfer the mask structure into a photosensitive coating on the substrate.

[0051] According to a further embodiment, at least one first segment of the sample comprises a light-transmitting or light-reflecting material and at least one second segment of the sample comprises a light-absorbing material.

[0052] For example, these materials transmit, reflect, or absorb light at wavelengths in the DUV and / or EUV range of the electromagnetic spectrum.

[0053] For example, if the sample is a DUV lithography mask (transmission photomask, binary mask), at least one first segment of the sample comprises a light-transmitting material. For example, if the sample is an EUV lithography mask (reflective photomask), at least one first segment of the sample comprises a light-reflecting material.

[0054] For example, at least one first segment of the sample includes a substrate. For example, the substrate includes silicon dioxide (SiO2), such as fused quartz. For example, at least one first segment of the sample may include one or more layers (coatings). The one or more layers may include, for example, one or more reflective layers and / or one or more protective layers (e.g., a Ru capping layer).

[0055] For example, at least one second segment of the sample includes an absorber structure, such as chromium, a chromium compound, a tantalum compound, and / or a compound of silicon, nitrogen, oxygen, and / or molybdenum (e.g., molybdenum silicon oxide or molybdenum silicon oxynitride, i.e., silicon oxide or silicon oxynitride (Si3N4) doped with molybdenum (Mo) (e.g., about 5% molybdenum), also referred to as MoSi).

[0056] At least one second segment of the sample may comprise the same material as at least one first segment of the sample. In this case, the corresponding material may be applied to the substrate of the sample with a greater thickness (i.e., a greater height in the height direction of the sample) in the second segment than in the first segment in order to have corresponding light absorption or light transmission / reflection properties. In particular, in this case, a greater thickness (greater height) corresponds to a stronger absorption effect.

[0057] According to a further aspect, a computer program product is proposed comprising instructions, which, upon execution of the program by the at least one computer, cause the at least one computer to carry out the method described above.

[0058] A computer program product, e.g. a computer program medium, may be provided or supplied as, for example, a storage medium, e.g. in the form of a memory card, a USB stick, a CD-ROM, a DVD or a file downloadable from a server in the network. For example, in a wireless communications network, provision of the computer program product may be achieved by transferring a suitable file containing the computer program product or computer program means.

[0059] According to a further aspect, an apparatus for analyzing an image of a microlithographic microstructure sample is proposed. The sample includes at least one first segment and at least one second segment having an edge and raised opposite the first segment. Furthermore, the image includes a plurality of pixels and a two-dimensional intensity distribution according to the pixels. Furthermore, the apparatus includes: a first determining device for determining a plurality of edge candidates for the image representation of the edge of the at least one second segment based on the gradient of the two-dimensional intensity distribution; a second determination device for determining a one-dimensional intensity distribution of the image in a direction perpendicular to the plurality of edge candidates, wherein in said direction the one-dimensional intensity distribution includes a first region having a first average intensity value, the plurality of edge candidates, and a second region having a second average intensity value greater than the first average intensity value; a third determining device for determining an edge candidate among the plurality of edge candidates that is closest to the first region of the one-dimensional intensity distribution as an image representation of an edge of the at least one second segment; Equipped with.

[0060] In particular, the apparatus is configured to carry out the method as described above.

[0061] The above-described method and the above-described apparatus for analyzing images of microlithographic microstructured samples can be applied to edge detection and extraction (contour detection and extraction) in a variety of applications.

[0062] Examples of applications include detecting defects (e.g., defect size, position, (geometric) shape, and contour, as well as multiple segments of the defect in the case of a defect containing multiple segments within the meaning of multiple connected regions) on a sample by calculating the difference between the structure of a defect-free reference and the structure (first and second segments) of the microstructured sample on a recorded microscope image (copy of the pattern). The reference can be selected from the recorded microscope image; the reference can be "empty" so that the segmentation of the defect is identical to the detected defect; the reference can be based on a microscope image simulated by a design file; and / or the reference can be based on changes in the contour of the structure of the sample (e.g., photomask structure) calculated based on a model and the physically manufactured sample to establish the correct exposure behavior of the photomask during wafer exposure, whereby inaccuracies in the contour can be attributed to causes that would otherwise be inaccessible.

[0063] Examples of applications of the aforementioned method include the detection of what are known as opaque defects, i.e., excess absorber material compared to the intended state of a sample (e.g., a lithography mask), and what are known as transparent defects, i.e., the absence of absorber material compared to the intended state of a sample (e.g., a lithography mask). Furthermore, particles (e.g., foreign particles) can also be identified as defects using the proposed method. Furthermore, it is possible to determine repair and / or processing shapes (i.e., geometric shapes, such as two-dimensional geometric shapes, that label areas of the sample that need to be repaired and / or processed). Repair and / or processing shapes include, for example, polishing shapes that label areas of the sample that need to be polished. For example, polishing shapes are used for fine processing of edges or residues. This fine processing also includes what is known as line trimming, which is used to slightly modify the position of the edges of features on the mask. These polishing shapes can be identified and / or created with the aid of this method. Repair and / or processing features include, for example, repair / processing features that label areas where a coating has been deposited on the sample with a halo around the processing site and must be removed again. Repair and / or processing features include, for example, areas of opaque normal structure that need to be etched away so that inaccessible errors in transparent areas can be corrected.

[0064] In applications of the aforementioned method, defect detection can be used as an independent product solution or as a step in a manual or automated workflow procedure. Furthermore, defects can be classified according to type, size, and further parameters. This step can be used as an independent product solution or as a step in a manual or automated workflow procedure (defect classification). In applications of the aforementioned method, defects can be automatically located at a defined position in the image (e.g., the center of the image). This step can be used as an independent product solution or as a step in a manual or automated workflow procedure (defect centering, defect location).

[0065] Further examples of applications of the aforementioned method include the recognition and optional measurement of structures, such as, for example, measuring the edge spacing of segments in a recorded microscope image. This step can be used as an independent product solution or as a step in a manual or automated workflow procedure. Furthermore, the edge spacing of segments in a recorded microscope image (SEM image) can be compared with segments in a reference image. This step can be used as an independent product solution or as a step in a manual or automated workflow procedure. In both cases, it is true that, firstly, an SEM image of any desired position on a photolithography mask can be selected, for example, the SEM image can include defects that have already been processed / repaired or defects that have not yet been processed (e.g., not yet at all), and secondly, the reference image can be a recorded SEM image or an SEM image simulated from a design file.

[0066] Further applications of the aforementioned method include the use of segment detection in SEM images of photolithographic masks for the purpose of modeling the three-dimensional structure of various structures or levels of the photomask. This step can be used as a stand-alone product solution or as a procedural step in a manual or automated workflow.

[0067] Further applications of the aforementioned method include the use of detecting segments in SEM images of photolithographic masks to simulate optical aerial images of the photolithographic masks created in a lithographic process. This step can be used as a stand-alone product solution or as a step in a manual or automated workflow procedure. Furthermore, segments can be detected at various positions of the photolithographic mask in the recorded SEM images to determine the spacing and absolute position of structures. This step can be used as a stand-alone product solution or as a step in a manual or automated workflow procedure. Furthermore, segments can be detected in SEM images of photolithographic masks for comparison with images of the same structures created by different sources and for position comparison (image registration, position comparison, position calibration).

[0068] Further examples of applications of the aforementioned method include detecting segments in SEM images for the purpose of proper placement of drift correction markers under specific boundary conditions (e.g., deposition only on absorber material, minimum distance from defects, minimum distance from the nearest structure edge, maximally symmetric distribution) and for the purpose of automatic drift correction. It is also possible to detect segments in SEM images that are suitable for beam optimization (e.g., focusing, destigmatization, stop alignment). Furthermore, automation can be provided, recognizing whether a defined structure is present in the image field and automatically issuing a warning if this structure disappears from the field of view. A further application is the recognition of structures in SEM images as a search aid for the purpose of finding structures of interest outside the field of view (automatic global alignment). The aforementioned method can also be used to detect segments of hardware attached to the electron column in order to align the electron beam emerging from the electron column with this hardware.

[0069] The aforementioned examples of applications can be used as individual products in equipment for mask repair and / or mask processing.

[0070] In this case, "a" should not necessarily be understood as a limitation to exactly one element. Rather, a plurality of elements, such as two, three, or more, may be provided. Any other number used herein should not be understood to imply that there is a limitation to the exactly stated number of elements. Rather, upward and downward numerical deviations are possible unless specifically indicated.

[0071] Embodiments and features described with respect to the method apply correspondingly to the proposed device and vice versa.

[0072] Further possible implementations of the invention also include implicitly mentioned combinations of features or embodiments described above or below with respect to the example embodiments, in which case a person skilled in the art will also add individual aspects as improvements or supplements to each basic form of the invention.

[0073] Further advantageous configurations and aspects of the invention are the subject of the dependent claims and also of the exemplary embodiments of the invention described below. In the following, the invention will be explained in more detail on the basis of preferred embodiments with reference to the attached drawings. [Brief explanation of the drawings]

[0074] [Figure 1] FIG. 1 illustrates a method flowchart for analyzing an image of a microlithographic microstructure sample, according to an embodiment. [Figure 2] FIG. 1 illustrates a top view of a detail of a microstructure sample, according to an embodiment. [Figure 3] 3 is a cross-sectional view of FIG. 2 taken along line III-III. [Figure 4] FIG. 1 shows an apparatus for recording microscopic images of a microstructured sample, according to an embodiment. [Figure 5] 5 shows an image of a microstructured sample recorded by the apparatus of FIG. 4 in accordance with an embodiment. [Figure 6]FIG. 6 is a diagram showing a one-dimensional intensity distribution of the image shown in FIG. 5. [Figure 7] FIG. 7 shows an enlarged partial detail of the intensity distribution of FIG. 6. [Figure 8] 8 illustrates the gradient of the intensity distribution of FIG. 7, according to an embodiment. [Figure 9] 10A-10C show further microscopic images of a microstructured sample before and after detection of the edges of the segments, according to an embodiment. [Figure 10] 10A-10C show further images of a microstructured sample illustrating the use of shadows to detect edges, according to an embodiment; [Figure 11] FIG. 11 is a diagram showing the one-dimensional intensity distribution of the image shown in FIG. DETAILED DESCRIPTION OF THE INVENTION

[0075] Unless otherwise indicated, identical or functionally identical elements within the figures are marked with the same reference numerals. Furthermore, it should be noted that the figures are not necessarily to scale.

[0076] In the following, FIGS. 1-10 are used to explain a method for analyzing a microlithographic microstructure sample 100, and in particular, an image 300 of the microlithographic microstructure sample 100.

[0077] FIG. 2 shows details of an exemplary microstructured sample 100. FIG. 3 shows details of the sample 100 shown in FIG. 2 in a cross-sectional view taken along line III-III in FIG. 2. The sample 100 includes a microstructure 104. For example, the microstructure 104 includes one or more raised elements 106 (e.g., absorber structures 106) and adjacent lower regions 108. For example, the microstructure 104 includes one or more raised elements 106 with lower regions 108 (e.g., grooves 108) therebetween. The raised elements 106 include edges 110, two of which are marked with reference symbols in FIG. 2. The lower regions 108 include, among other things, one or more first segments 112. Furthermore, the raised elements 106 include one or more second segments 114 that are raised opposite the one or more first segments 112. FIG. 2 shows only two second segments 114 as an example.

[0078] Furthermore, the microstructure sample 100 has, for example, a flat shape including a main plane of the area E (the xy plane in FIGS. 2 and 3). The direction perpendicular to the main plane of the area E is called the height direction z of the sample 100.

[0079] For example, each of the second segments 114 in FIG. 2 is a connected region in a plane parallel to the major plane of the extent E of the sample 100.

[0080] 3, for example, at least one first segment 112 of the sample 100 has a first height H1 in a height direction z of the sample 100. Furthermore, at least one second segment 114 of the sample 100 has a second height H2 in the height direction z that is greater than the first height H1. In particular, the at least one second segment 114 reaches a height ΔH above the surface 116 of the at least one first segment 112. In other words, in particular, the surface 118 of the at least one second segment is disposed at a height ΔH above the surface 116 of the at least one first segment 112.

[0081] The edge 110 of at least one second segment 114 has, in particular, an edge wall 120 (FIG. 3). For example, the corresponding edge wall 120 is arranged parallel to the height direction z of the sample 100 and perpendicular to the plane of the main area E of the sample 100 (FIG. 2) (i.e., at an angle of 90° with respect to the plane of the main area E). In FIG. 3, the edge wall 120 is, for example, located in the yz plane. However, in other examples, the edge wall 120 may be arranged inclined to the main plane of the area E of the sample 100 (i.e., at an angle less than or greater than 90° with respect to the main plane of the area E), or may be located in the xz plane or in any other plane arranged perpendicular to the main plane of the area E.

[0082] For example, the sample 100 includes a substrate 122 (FIG. 3) on which one or more raised elements 106 are disposed that form at least one second segment 114. An exposed surface of the substrate 122 forms, for example, a surface 116 of at least one first segment 112.

[0083] Although not shown, one or more layers (coatings) may be disposed on the substrate 122 of the sample 100. For example, if the sample 100 is an EUV lithography mask, for example, a protective layer (e.g., a Ru capping layer) may be disposed on the substrate 122. When one or more layers are disposed on the substrate 122, the exposed region of the uppermost one of these layers may form the surface 116 of the at least one first segment 112.

[0084] The lower region 108 of the sample 100, e.g., the substrate 122, and the one or more raised elements 106 may comprise different materials from each other or the same material. In other words, the at least one first segment 114 and the at least one second segment 114 may comprise different materials from each other or the same material. For example, the exposed surface 116 of the at least one first segment 114 and the exposed surface 118 of the at least one second segment 114 may comprise different materials from each other or the same material.

[0085] Additionally, at least one first segment 112 of the sample 100 can include a light-transmitting or light-reflecting material, and at least one second segment 114 of the sample 100 can include a light-absorbing material.

[0086] In order to analyze the sample 100 and to process and / or use the sample 100 based on the analysis, it may be necessary to detect the contour of the microstructure 104, i.e., for example, the edge 110 of the second raised segment 114. For example, it may be necessary to determine the position and / or the (e.g., two-dimensional) geometric shape of the edge 110 of the second segment 114. This determination is performed using a method described below that is based on image analysis.

[0087] For example, the microstructured sample 100 analyzed in this way is a lithography mask (reticle), in particular an EUV lithography mask or a DUV lithography mask, but the microstructured sample 100 analyzed in this way can also be, for example, a wafer structured using microlithography, or any other kind of microstructured sample.

[0088] For example, the microstructure specimen 100 analyzed in this manner may be configured for an operating wavelength in the DUV and / or EUV range. For example, the microstructure specimen 100 may be designed for an operating wavelength below 250 nm, below 200 nm, and / or below 15 nm. However, the microstructure specimen 100 analyzed in this manner may be configured for an operating wavelength in other regions of the electromagnetic spectrum or may not be configured for exposure to a working light.

[0089] In a first step S1 of the method, a microscope-captured image 300 (FIG. 5) of the sample 100 (eg, a portion of the sample 100) is provided.

[0090] For example, an image 300 captured by a microscope is recorded by an image recording device 200 (FIG. 4), which creates the image 300 with the assistance of a particle beam, such as an electron beam 202 or an ion beam.

[0091] As an example of an image recording device 200, a scanning electron microscope 200 is shown, by way of example only, in Figure 4. Figure 4 shows a schematic cross section through some components of the apparatus 200 that may be used to image the sample 100.

[0092] Additionally, apparatus 200 can optionally be used for processing and / or repair (e.g., etching, deposition) of electron beam excited processes of specimen 100. For example, apparatus 200 is a repair apparatus (repair tool) for microlithography photomasks, such as for photomasks in DUV or EUV lithography apparatus.

[0093] 4 represents, for example, a modified scanning electron microscope 200. In this case, an electron beam 202 is used to image a sample 100. Most of the apparatus 200 is located within a vacuum enclosure 204. The space enclosed by the vacuum enclosure 204 is maintained at a gas pressure by a vacuum pump 206.

[0094] The sample 100 to be processed is placed on a sample stage 208. For example, the sample stage 208 is configured to position the sample 100 in three mutually orthogonal spatial directions x, y, z, and, for example, additionally in three mutually orthogonal axes of rotation, with an accuracy of a few nanometers.

[0095] The apparatus 200 comprises an electron column 210. The electron column 210 comprises an electron source 212 for providing an electron beam 202. The electron column 210 also comprises electron or beam optics 214. The electron source 212 creates the electron beam 202, and the electron or beam optics 214 focuses the electron beam 202 and directs it at the output of the electron column 210 towards the sample 100. The electron column 210 also comprises a deflection unit 216 (scanning unit 216) configured to guide (scan) the electron beam 202 over the surface of the sample 100. Instead of a deflection unit 216 (scanning unit 216) located within the electron column 210, a deflection unit (scanning unit) located outside the electron column 210 can also be used (not shown).

[0096] Apparatus 200 also includes a detector 218 for detecting secondary electrons and / or backscattered electrons generated in the material of sample 100 by incident electron beam 202. For example, as shown, detector 218 is arranged in electron column 210 in a ring-shaped manner around electron beam 202. As an alternative to and / or in addition to detector 218, apparatus 200 may include other / further detectors for detecting secondary electrons and / or backscattered electrons (not shown in FIG. 4 ).

[0097] The apparatus 200 may optionally include a gas supply unit 220 for supplying a process gas to the surface of the sample 100. For example, the gas supply unit 220 includes a valve 222 and a gas line 224. The electron beam 202 directed by the electron column 210 to a position on the surface of the sample 100 can perform electron-beam induced processing (EBIP) together with a process gas externally supplied by the gas supply unit 220 via the valve 222 and the gas line 224. In particular, the process includes deposition (forming a film) and / or etching of a material.

[0098] The apparatus 200 also includes a computing device 226, such as a computer, that includes a control device 228, a generating device 230, a first determining device 232, a second determining device 234, and a third determining device 236. In the example of Figure 4, the computing device 226 is located outside the evacuated enclosure 204.

[0099] The control device 228 serves, for example, to control the apparatus 200. For example, the control device 228 controls the supply of the electron beam 202 by controlling the electron column 210. In this case, the control device 228 notably controls the guidance of the electron beam 202 on the surface of the sample 100 by controlling the scanning unit 216. The control unit 228 can also control the gas supply unit 220 to supply a process gas.

[0100] The generating device 230 receives measurement data from the detector 218 and / or other detectors of the apparatus 200 and creates from the measurement data images 300, 500 (FIGS. 5, 9) that can be displayed on a monitor (not shown). For example, the spatial resolution of the generated images 300, 500 is on the order of a few nanometers.

[0101] Figure 5 shows an example of a microscope-captured image 300 of a sample similar to sample 100 of Figure 2. For example, image 300 was captured using scanning electron microscope 200 of Figure 4. Thus, image 300 is, for example, a scanning electron microscope image (SEM image).

[0102] For example, image 300 is recorded along a line of sight S (FIG. 3) that is aligned parallel to the height direction z of sample 100 .

[0103] The image 300 includes a number of pixels 302 (n pixels), three of which are referenced in the enlarged partial detail of FIG. 5 by way of example. For example, the image 300 includes n pixels 302, where n is a natural number greater than 1. In particular, the pixels 302 are arranged in a two-dimensional arrangement. The image 300 includes an intensity value I assigned to each ith pixel 302. i ("grayscale values"), where i = 1 to n. i forms the two-dimensional intensity distribution 304 of the image 300.

[0104] By way of example, image 300 of Figure 5 shows brighter and darker regions 306, 308 having different intensities I1, I2. In particular, at least one first segment 112 of sample 100 (Figure 2) is imaged (i.e., imaged darker) within first region 306, i.e., within first region 306 in image 300, for example, having a lower intensity I1. Furthermore, for example, at least one second segment 114 of sample 100 (Figure 2) is imaged (i.e., imaged brighter) within second region 308, i.e., within second region 308 in image 300, for example, having a greater intensity I2. Thus, in the example of Figure 5, I2 is greater than I1 (see also Figures 6 and 7).

[0105] However, in other examples (Figures 10 and 11), it is also possible that at least one first segment 112 and at least one second segment 114 of sample 100 (Figure 2) are imaged at (e.g., approximately) the same average intensities I1', I2' (i.e., I1' = I2' or I1' ≒ I2'), for example, away from shading and / or edge brightening.

[0106] So-called edge brightening 310 can occur when imaging the edge 110 (FIG. 2) of at least one second raised segment 114 in the image 300 (FIG. 5), as can be seen in FIG. 5. In the region of this edge brightening 310, the intensities I3, I3′ of the image 300 are greater than the intensities in each of the first and second regions 306, 308 (see also FIG. 7).

[0107] In an optional second step S2 of the method, image pre-processing is performed to reduce the noise content of the two-dimensional intensity distribution 304 (FIG. 5).

[0108] In a third step S3 of the method, a number of candidates 312, 314 (FIG. 7) for the image representation of the edge 110 (FIG. 2) of the second segment 114 are determined based on the image 300. This determination is performed based on calculating the gradients 316, 318 (i.e., first derivatives 316, 318) of the two-dimensional intensity distribution 304 (FIG. 5) in the image 300.

[0109] Each edge candidate 312, 314 corresponds to a possible image representation of one and the same actual edge 110 of the second segment 114 of the sample 100. In other words, when the gradients 316, 318 of the image 300 of the sample 100 are calculated, for one and the same actual edge 110 of the second segment 114 of the sample 100, two or more candidates for the image representation of the edge 110 are detected.

[0110] For example, a plurality of edge candidates 312, 314 are determined by the first determining device 232 of the apparatus 200 (FIG. 4).

[0111] It should be noted that the figures, particularly in Figure 7, only show the gradients 316, 318 of the one-dimensional intensity distribution 320 (Figure 6) for clarity. Nevertheless, the gradient formation in step S3 is preferably implemented as a determination of gradients within the two-dimensional intensity distribution 304 (Figure 5) of the image 300.

[0112] For example, edge candidates 312, 314 (FIG. 7) may be determined by applying what is known as a Sobel operator and / or any other suitable process to the two-dimensional intensity distribution 304 (FIG. 5) of the image 300. For example, the intensity values ​​I of the corresponding pixel 302 and the pixels 302 surrounding this pixel 302 may be calculated. i Based on this, gradients 316, 318 of the intensity distribution 304 are determined for each pixel 302 in the image 300. Thus, for example, a matrix of gradients 316, 318 is derived from the image 300. Next, candidate edges 312, 314 of at least one second segment 114 captured in the image 300 (FIG. 2) are determined at pixels 302 where the intensity I (i.e., brightness) of the original image 300 changes the most (corresponding to large gradients).

[0113] In a fourth step S4 of the method, a one-dimensional intensity distribution 320 (FIG. 7) of the image 300 is determined in a direction R (FIG. 5) perpendicular to the plurality of edge candidates 312.

[0114] For example, the one-dimensional intensity distribution 320 is determined by the second determining device 234 of the apparatus 200 (FIG. 4).

[0115] Figure 6 shows an example of a one-dimensional intensity distribution 320 in image 300 in a direction R that is perpendicular to the multiple edge candidates 312, 314. In particular, Figure 6 shows the intensity distribution 320 in image 300 along line 322 in Figure 5. In particular, Figure 6 shows a graph of the one-dimensional intensity distribution 320 in image 300 as a function of position x in direction R. In the example shown, the orthogonal direction R is parallel to the x-direction of the image (Figure 5), and therefore the intensity distribution 320 in Figure 6 is shown as a function of the x-coordinate of image 300.

[0116] In the example of Figure 6, the one-dimensional intensity distribution 320 includes a first region 306' having a first average intensity value I1 and corresponding to the darker region 306 of Figure 5, and thus corresponding to the image representation of the first segment 112 of the sample 100. The one-dimensional intensity distribution 320 also includes a second region 308' having a second average intensity value I2 and corresponding to the lighter region 308 of Figure 5, and thus corresponding to the image representation of the second segment 114 of the sample 100. Notably, the second average intensity value I2 is greater than the first average intensity value I1. Furthermore, edge brightening 310 at the edge 110 of the second segment 114 appears as a maximum value 310' in the intensity distribution 320 of Figure 6.

[0117] FIG. 7 shows an enlarged portion of the one-dimensional intensity distribution 320 of FIG.

[0118] An edge brightening 310 (FIG. 5) within a region of the image representation of an edge 110 (FIG. 2) appears as a maximum 310′ in a one-dimensional intensity distribution 320 (FIG. 6 or 7). Because of the edge brightening 310, two gradients 316, 318 with large absolute values ​​are determined on either side of each maximum 310′, i.e., at two different locations x1, x2 within the region of the image representation of each edge 110. As a result, two candidates 312, 314 are determined for each edge 110 image representation. In other words, two different possible locations x1, x2 (FIG. 7) are determined for each edge 110 in direction R (FIG. 5). In an example where the orthogonal direction R is parallel to the x-direction of the image 300, for example, two possible x-locations x1, x2 are determined for the edge 110.

[0119] Optionally, in step S4, a predefined threshold Th may be applied during the determination of multiple (e.g., parallel) edge candidates 312, 314 ( FIG. 7 ) for the image representation of the edge 110 of the second segment 114 based on gradients 316, 318 in the two-dimensional intensity distribution 304. This is illustrated in FIG. 8 for the one-dimensional case. In particular, FIG. 8 shows a diagram of the absolute value of the gradient (absolute value of the first derivative dI / dx) of the intensity distribution I of FIG. 7 as a function of position x. For example, the method may provide only gradients 316, 318 of the two-dimensional intensity distribution 304 having an absolute value greater than the predefined threshold Th to be determined as edge candidates 312, 314 in step S4. In contrast, for example, gradients 324 of the two-dimensional intensity distribution 304 having an absolute value less than the predefined threshold Th ( FIG. 8 ) are not classified as edge candidates 312, 314.

[0120] When imaging the sample 100, a shadow 326 ( FIG. 6 ) may be formed that is visible in the image 300 of the sample 100. In particular, a shadow 326 of the second segment 114 ( FIGS. 2 and 3 ) may be formed in a region of the first segment 112 adjacent to the second segment 114. The shadow 326 results in a shadowed region 328 in the image 300 and in the one-dimensional intensity distribution 320. Note that the shadowed region 328 is visible only in FIGS. 6 and 7 of the drawings. In other words, the shadow 326 causes at least one first segment 112 of the sample 100 ( FIG. 2 ) to be imaged within the shadowed region 328 with a lower average intensity I4 ( FIG. 7 ) than outside the shadowed region 328 (i.e., region 306′ in FIG. 7 ).

[0121] The intensity difference (I2-I1) between the second region 308' and the first region 306' and / or the intensity difference (I2-I4) between the second region 308' and the shadowed region 328 may be used in the next step of the method to select one of the determined candidates 312, 314 as the image representation of the edge 110 of the second segment 114.

[0122] In a fifth step S5 of the method, the edge candidate among the plurality of edge candidates 312, 314 (FIG. 7) that is located closest to the first region 306′ and / or the shadowed region 328 of the one-dimensional intensity distribution 320 is determined as an image representation of the edge 110 (FIG. 2) of the second segment 114.

[0123] For example, in step S5, the best edge candidates 312, 314 for the image representation of the edge 110 are determined by the third determination device 236 of the apparatus 200 (FIG. 4).

[0124] 7, of the two edge candidates 312, 314 determined in step S4, edge candidate 312 is located closest to the first region 306′ of the one-dimensional intensity distribution 320, and as a result is determined as the image representation of edge 110 of second segment 114. For example, position x2 of edge candidate 312 is determined as the position of edge 110 of second segment 114.

[0125] Additionally or alternatively, step S5 may take into account that edge candidate 312 of the two edge candidates 312, 314 determined in step S4 is located closest to the shadowed region 328 in the one-dimensional intensity distribution 320. In this context, the same edge candidate 312 is determined as the image representation of edge 110 of second segment 114 in the example of Figure 7.

[0126] The proposed method allows for better detection of the pose or position of the edge 110 of the microstructured sample 100 (FIG. 2). In particular, the positions of the edge 110 determined by the proposed method are located with greater precision in all details and more accurate in their positioning relative to the geometry of at least one second segment 114. Furthermore, compared to conventional methods, the positions of edges located closer to the direction of the lower-lying structures 108, 112 (FIG. 2) can be determined.

[0127] FIG. 9 illustrates a further example of a method for analyzing a sample 400 similar to sample 100 of FIG. 2. FIG. 9 shows images 500 (e.g., SEM images 500) of sample 400 before (left image) and after (right image) edge detection. Sample 400 includes at least three first segments 412a, 412b, and 412c (similar to first segments 112 of FIG. 2) and at least one second segment 414 (similar to second segment 114 of FIG. 2). In image 500, first segments 412a, 412b, and 412c of sample 400 are imaged as first regions 406a, 406b, and 406c. In this case, first regions 406a and 406b exhibit edge brightenings 410a and 410b, respectively, similar to edge brightening 310 of FIG. 5. However, first region 406c exhibits only very weak or no edge brightening. Additionally, a second segment 414 of sample 400 is imaged in image 500 as second region 408.

[0128] As a result of the above-described method, the edges 110a, 110b, 110c of the first regions 406a, 406b, 406c corresponding to the first segments 412a, 412b, 412c can be accurately detected in terms of their positions, as shown on the right of Figure 9. In particular, the edges 110a, 110b, 110c determined in this way are located closer to the structures 412a, 412b, 412c located at lower positions on the sample 400.

[0129] If a shadow 326, 626 is formed when the sample 100 is imaged (FIGS. 7, 10, and 11), the shadowed region 328, 628 of the one-dimensional intensity distribution 320, 620 may be used in step S5 instead of the first region 306′ of the one-dimensional intensity distribution 320 for the purpose of selecting from the edge candidates 312, 314 (FIG. 7) or 612, 614 (FIG. 11). In other words, in step S5, the edge candidate 312 or 612 of the two determined edge candidates 312, 314 (FIG. 7) or 612, 614 (FIG. 11) located closest to the shadowed region 328, 628 of the one-dimensional intensity distribution 320, 620 may be determined as the image representation of the edge 110 of the second segment 114 (FIG. 2).

[0130] Taking into account the formation of shadow 326 has been found to be advantageous, particularly when at least one of first and second segments 112, 114 comprises the same material and, away from the shadow, is imaged in images 300, 600 with the same average brightness I1', I2'.

[0131] FIG. 10 shows a further image 600 of the topographical sample 100 recorded by the apparatus of FIG. 4, in accordance with an embodiment, illustrating the use of shading for edge detection.

[0132] Figure 11 shows a one-dimensional intensity distribution 620 of the image 600 shown in Figure 10. In this case, at least one first segment 112 and at least one second segment 114 of the sample 100 (Figure 2) are imaged with approximately the same average intensities I1', I2' (Figure 11) in the image 600 (i.e., I1'≈I2'), away from the shadows 626, 628 and edge brightening 610.

[0133] 11 specifically plots the edge candidates 612, 614 and gradients 616, 618 determined in step S4 based on the two-dimensional intensity distribution 604 (FIG. 10) within a one-dimensional intensity distribution 620. A first region 606 having a first average intensity I1′ of the one-dimensional intensity distribution 620 corresponds to the image representation of at least one first segment 112 of the sample 100. A second region 608 having a second average intensity I2′ of the one-dimensional intensity distribution 620 corresponds to the image representation of at least one second segment 114. Additionally, a shaded region 628 having a further average intensity I4 corresponds to the image representation of the shaded region 628 of the at least one first segment 112.

[0134] In particular, due to the formation of shadow 626, the average intensity value I2' in second region 608 of one-dimensional intensity distribution 620 is greater than the average intensity value I4 in shadowed region 628 of one-dimensional intensity distribution 620.

[0135] 10 and 11, the edge candidate among the edge candidates 612, 614 determined in step S4 that is located closest to the shadowed region 628 of the one-dimensional intensity distribution 620 is determined in step S5 of the method as the image representation of the edge 110 (FIG. 2). This edge candidate is the edge candidate 612 at location x2 in the example of FIG. 11.

[0136] Although the present invention has been described with reference to exemplary embodiments, it can be varied in many ways. [Explanation of symbols]

[0137] 100 samples 104 Microstructure 106 elements 108 areas 110 Edge 110a Edge 110b Edge 110c Edge 112 segments 114 segments 116 Surface 118 Surface 120 Wall 122 PCB 200 Image recording device 202 Electron Beam 204 Case 206 Pump 208 Sample Stage 210 Electronic pillar 212 Electron source 214 Electron or beam optical elements 216 Deflection Unit 218 detector 220 Gas Supply Unit 222 Valve 224 Gas Line 226 Computing Devices 228 Control Device 230 Generating Device 232 Decision Device 234 Decision Device 236 Decision Device 300, 300'' statue 302 pixels 304 Intensity distribution 306, 306' area 308, 308', 308'' area 310, 310', 310'' Edge Brightening 312 candidates 314 candidates 316 Slope 318 Slope 320 Intensity distribution 322 line 324 Slope 326 Shadow 328 areas 400 samples 406a area 406b area 406c area 410a Edge Brightening 410b Edge Brightening 412a segment 412b segment 412c segment 414 segments 408 areas 410a, 410b Edge Brightening 500 statues 600 statues 604 Intensity distribution 606 areas 608 areas 610 Edge Brightening 612 candidates 614 segments 616 Slope 618 Slope 620 Intensity distribution 626 Shadow 628 areas dI / dx gradient (first derivative) E plane H Height H2 height H1 Height ΔH height I1, I2 intensity I1', I2' intensity I3, I3 intensity I i strength R direction S Line of Sight Steps S2~S5 of the method Th Threshold x, y, z directions x1, x2 position

Claims

1. 1. A method for analyzing an image (300) of a microlithography microstructure sample (100), the sample (100) comprising at least one first segment (112) and at least one second segment (114) having an edge (110) and raised opposite the first segment (112), the image (300) comprising a plurality of pixels (302) and a two-dimensional intensity distribution (304) responsive to the pixels (302), the method comprising: a) determining (S2) a plurality of edge candidates (312, 314) for an image representation of the edge (110) of the at least one second segment (114) based on gradients (316, 318) of the two-dimensional intensity distribution (304); b) determining (S3) a one-dimensional intensity distribution (320) of the image (300) in a direction (R) perpendicular to the plurality of edge candidates (312, 314), wherein in the direction (R) the one-dimensional intensity distribution (320) has a first mean intensity value (I 1 ), the plurality of edge candidates (312, 314), and the first average intensity value (I 1 ) greater than the second average intensity value (I 2 The step (S3) includes a second region (308′) having and c) determining (S4) the edge candidate among the plurality of edge candidates (312, 314) that is closest to the first region (306') of the one-dimensional intensity distribution (320) as the image representation of the edge (110) of the at least one second segment (114).

2. 2. The method of claim 1, wherein the first region (306') of the one-dimensional intensity distribution (320) of the image (300) is based on an image representation of the at least one first segment (112) of the sample (100), and the second region (308') of the one-dimensional intensity distribution (320) of the image (300) is based on an image representation of the at least one second segment (114) of the sample (100).

3. 3. The method of claim 1, wherein the at least one first segment (112) of the specimen (100) comprises a first material and the at least one second segment (114) of the specimen (100) comprises a second material different from the first material.

4. The second average intensity value (I ) in the second region (308′) of the one-dimensional intensity distribution (320) of the image (300) is increased due to a material difference between the at least one first and second segments (112, 114) of the sample (100). 2 ) is the first average intensity value (I ) in the first region (306′) of the one-dimensional intensity distribution (320) of the image (300). 1 4. The method of claim 3, wherein the .lambda.

5. The method of claim 1 or 2, wherein the at least one first and second segments (112, 114) of the sample (100) comprise the same material.

6. The second average intensity value (I ) in the second region (308′) of the one-dimensional intensity distribution (320) of the image (300) is increased due to a shadow (326) formed adjacent to the edge (110) of the at least one second segment (114) of the sample (100). 2 ) is the first average intensity value (I ) in the first region (306′) of the one-dimensional intensity distribution (320) of the image (300). 1 6. The method of claim 5, wherein the .lambda.

7. 3. The method of claim 1, wherein when determining the plurality of edge candidates (312, 314) based on the gradients (316, 318) of the two-dimensional intensity distribution (304), the predetermined threshold (Th) is applied in such a way that a corresponding edge candidate (316, 318) is determined for a gradient (316, 318) of the two-dimensional intensity distribution (304) having an absolute value greater than the predetermined threshold (Th), and an edge candidate is not determined for a gradient (324) of the two-dimensional intensity distribution (304) having an absolute value equal to or less than the predetermined threshold (Th).

8. 3. The method of claim 1, wherein step (a) is preceded by image processing (S1) to reduce noise components in the two-dimensional intensity distribution (304).

9. the microstructured specimen (100) is designed for an operating wavelength of less than 250 nm, less than 200 nm, less than 100 nm, and / or less than 15 nm; and / or 3. The method according to claim 1 or 2, wherein the microstructured sample (100) is a lithography mask, in particular an EUV or DUV lithography mask, and / or a wafer structured by microlithography.

10. 3. The method of claim 1, wherein the at least one first segment (112) of the sample (100) comprises a light-transmitting or light-reflecting material, and the at least one second segment (114) of the sample (100) comprises a light-absorbing material.

11. A computer program product causing at least one computer to carry out the method according to claim 1 or 2.

12. 1. An apparatus (200) for analyzing an image (300) of a microlithography microstructure sample (100), the sample (100) comprising at least one first segment (112) and at least one second segment (114) having an edge (110) and raised opposite the first segment (112), the image (300) comprising a plurality of pixels (302) and a two-dimensional intensity distribution (304) responsive to the pixels (302), the apparatus (200) comprising: a first determination device (232) for determining a plurality of edge candidates (312, 314) for an image representation of the edge (110) of the at least one second segment (114) based on the gradients (316, 318) of the two-dimensional intensity distribution (304); a second determination device (234) for determining a one-dimensional intensity distribution (320) of the image (300) in a direction (R) perpendicular to the plurality of edge candidates (312, 314), wherein in the direction (R) the one-dimensional intensity distribution (320) has a first average intensity value (I 1 ), the plurality of edge candidates (312, 314), and the first average intensity value (I 1 ) greater than the second average intensity value (I 2 the second determination device (234) including a second region (308') having a a third determination device (236) for determining the edge candidate (312, 314) among the plurality of edge candidates (312, 314) that is closest to the first region (306') of the one-dimensional intensity distribution (320) as the image representation of the edge (110) of the at least one second segment (114); An apparatus (200) comprising:

Images