Method for recording an image using a particle microscope

Abstract

A method of recording images using a particle microscope, the method comprising: recording a plurality of images of an object, wherein the intensity values of the image data are associated with positions in the coordinate system of the recorded images and positions on the object; determining regions of interest associated with the positions; determining image regions that include the regions of interest; determining displacement vectors associated with these image regions, the displacement vectors being determined by correlating the image data associated with a given image region with the image data of other image regions that are also associated with the given region of interest; and generating a resulting image based on the image data of the recorded images, wherein the positions in the coordinate system of the resulting image are determined based on the displacement vectors associated with these image regions.

Description

[0001] Cross - Reference to Related Applications

[0002] This application claims priority to German Application No. 10 2018 124 401.0, filed Oct. 2, 2018, and U.S. Provisional Application No. 62 / 888,866, filed Aug. 19, 2019. The entire disclosures of these applications are incorporated herein by reference. Technical Field

[0003] The present invention relates to a method for recording an image using a particle microscope. Background Art

[0004] A particle microscope generally includes an object mount for mounting an object to be imaged, and a particle beam column that generates one or more particle beams and directs the one or more particle beams onto the object. Signals generated by the particles incident on the object are detected such that the intensity values of the detected signals can be associated with the positions on the object where the signals are generated. The collected intensity values and the positions on the object associated with the intensity values represent an image of the object, which can be displayed on a screen, for example, analyzed, or stored for other purposes.

[0005] Since the detected signals are affected by signal noise, it is necessary to collect a sufficient amount of signals from the region of the object to be imaged in order to obtain an image with a desired image resolution and image quality. The collection of signals takes a large amount of time. In practice, it may occur that the object mount holding the object moves relative to the particle beam column while signals for generating an image are being collected. Such movement may be caused by mechanical instability, thermal drift, or other reasons. Such movement results in the association between the detected signals and the positions on the object not being defined well enough, and results in the resulting image being blurred or distorted. In addition, the particles incident on the object and the particles that leave the object and form the detected signals generate surface charges on the object. The surface charges may cause image artifacts to appear in the generated image.

[0006] According to a conventional method, a plurality of images of an object are sequentially recorded, wherein the image quality of each of the recorded images is lower than the resulting image calculated based on the combination or superposition of the recorded images. When the individually recorded images are combined or superimposed to form the resulting image, this method takes into account the movement of the object holder relative to the particle beam column during the recording of subsequent images. Herein, computer-aided image processing can be used to determine the displacement between the recorded images. For example, such determination may include a method of correlating the recorded images. An example of such a method is shown in US 7,034,296 B2.

[0007] Thus, computer-aided image processing can generate a higher-quality image based on a plurality of images with relatively low recorded image quality. The higher-quality image can exhibit, for example, a higher contrast-to-noise ratio. For example, compared to an image obtained by conventional superposition, the higher-quality image can exhibit lower image blur. Image blur can affect the higher-quality image if, for example, drift is involved during the recording of the lower-quality images and this drift is not properly taken into account when combining the lower-quality images to form the higher-quality image. Additionally, compared to each of the lower-quality images, the higher-quality image can exhibit a higher contrast-to-noise ratio.

[0008] Furthermore, the question is raised as to whether it is possible to reduce image artifacts due to surface charges in a particle-optical image by using computer-aided image processing. Summary of the Invention

[0009] The present invention has been made in view of the above considerations.

[0010] The object of the present invention is to provide a method for recording an image using a particle microscope, wherein computer-aided image processing provides improved image quality.

[0011] Embodiments of the present invention provide a method for recording an image using a particle microscope, wherein the method includes using the particle microscope to record a plurality of images of an object.

[0012] The recorded images are represented by data records. The data records can include metadata such as time, position, type of the object being measured, magnification, coordinates of the imaged object area in the coordinate system of the object or in the coordinate system of the particle microscope and other items. The recorded images further include image data, wherein the image data includes a plurality of intensity values, and each intensity value is associated with a position in the coordinate system of the recorded image and a position on the object.

[0013] Each intensity value can represent, for example, the number of detected signals detected during a given duration when the particle beam of the particle microscope is directed to the same position on the object (dwell time). The detected signals can be detection events resulting from the detection of secondary electrons, backscattered electrons, x-rays, light, and other events generated when particles in the particle beam hit the object.

[0014] According to an exemplary embodiment, the position on the object associated with a given intensity value is the position on the object where the particle beam is actually directed while collecting the signal for determining the given intensity value.

[0015] According to an exemplary embodiment, the position in the coordinate system of the recorded image associated with a given intensity value corresponds to the position on the object that is also associated with the given intensity value. The association can be, for example, a correspondence represented by a coordinate transformation that allows the position on the object in the coordinate system of the object to be calculated based on the position in the coordinate system of the image. A suitable coordinate transformation can be determined, for example, based on the measured position of the object relative to the particle microscope, the magnification of the microscope, and other data.

[0016] Thus, a simplified assumption can be used to transform the position in the coordinate system of the recorded image into the corresponding position in the coordinate system of the object. According to one simplified assumption, the position of the object relative to the particle microscope is known or at least does not change during the recording of the image. In practice, the object may move relative to the particle microscope, or the position of the particle beam relative to the particle microscope may unexpectedly change during the recording of the image, such that the position where the particle beam impinges on the object is different from the expected position, because the amount of displacement may be unknown. In such a case, the position in the coordinate system of the recorded image represents a position of reduced accuracy in the coordinate system of the object. The position where the beam impinges on the object is known only with limited accuracy, and since the beam impinges at this position on the object, this position of the object is associated with an intensity value.

[0017] The position in the coordinate system of the recorded image associated with a given intensity value can be determined based on, for example, the excitation applied to the beam deflector of the particle microscope while the signal for determining the given intensity value is detected. The deflector is excited to direct the particle beam to the desired position on the object.

[0018] The intensity values representing the image can be stored, for example, as a two-dimensional matrix with integer indices representing the positions in the coordinate system of the recorded image. However, each recorded intensity value can also be stored together with two additional values representing the position in the coordinate system of the recorded image. For example, these values can be represented by integers or floating-point numbers.

[0019] According to an exemplary embodiment, a method of recording an image can include determining a plurality of regions of interest, wherein at least one position on the object is associated with a given region of interest.

[0020] For example, a region of interest can be predefined in the coordinate system of an object. For example, the region of interest can be a region of the object that is expected to contain important features. For example, if the object is an integrated circuit, the region of interest can be determined based on the available design data of the integrated circuit, which provides information about the location in the circuit where the selected features of the integrated circuit can be found. The selected circuit elements are preferably circuit elements that can be identified and located in a particle-optical image with high contrast. According to a further example, the location of the region of interest is determined based on an analysis performed on the recorded image itself. For example, one or more recorded images can be analyzed to determine where in the recorded image the region to which image data representing an image portion including features with high contrast has been assigned is located. These determined regions of interest can be associated with the object and include at least one location of the object and typically include multiple locations of the object.

[0021] According to a further embodiment, a method of recording an image includes determining a plurality of image regions in each recorded image, wherein each of these image regions is associated with one of a plurality of regions of interest. The plurality of image regions are determined such that each image region includes intensity values of the recorded image at locations within a neighborhood of a location assigned to the object, the location of the object also being associated with the region of interest located in that image region.

[0022] An image region can be an extended region in the coordinate system of the recorded image. In the context of the present application, an image region "includes" or "contains" image data. This means that the image data is considered to be the image data of the image region when the image data of the recorded image associated with the locations in the coordinate system of the recorded image is located within the extended region.

[0023] When the region of interest of the object is, for example, an element of an electronic circuit, each image region of a given recorded image can be selected such that the image data of the recorded image is associated with an image region including intensity values at locations on the object within a neighborhood of the location assigned to the object associated with the corresponding region of interest. For example, this can be achieved when each of the plurality of image regions of the recorded image at least partially overlaps with the region of interest of the object in the representation of the recorded image. However, it is not required that the image regions are always located at the corresponding positions in the coordinate system of the recorded image. In particular, this can occur when the region of interest is determined based on an analysis of the recorded image or when the object is moved relative to the particle microscope while recording multiple images. If the region of interest of the object is predefined, for example, in the coordinate system of the object, coordinate transformation can be used considering the position, magnification, and other data of the object relative to the particle microscope to determine the image regions of the recorded image assigned to the region of interest.

[0024] According to a further exemplary embodiment, a method of recording an image includes determining displacement vectors associated with at least some of a plurality of image regions. Herein, a displacement vector associated with a given image region is determined based on correlating the image data of the given image region with the image data of another image region or a plurality of other image regions. In particular, the image regions that are correlated with each other may be associated with the same region of interest among a plurality of regions of interest. For example, the image regions that are correlated with each other may be image regions of different recorded images.

[0025] Correlating the image data of one image region with the image data of another image region includes calculations that can be represented, for example, by a mathematical function. The mathematical function may receive the image data of the image regions to be correlated as input parameters and may calculate an output based on these input parameters. For example, the output may be a scalar value, a tuple of scalar values (such as a vector), and image data. An example of such a function is a function that determines the displacement vector between two image regions by performing a convolution using a two-dimensional Fourier transform. Examples of such calculation methods are shown in Section A.1 of the following article: Lauterborn et al., Optik - Grundlagen für Physiker und Ingenieure, Berlin Heidelberg 1993.

[0026] For example, for each region of interest, the following steps may be performed: First, determine these image regions of the recorded images associated with the region of interest of the object. This means that the features of the object located in the region of interest are also visible in the image representation generated based on the image data contained in the determined image regions. In the representation of the image region, the features of the object are not always located at the same position in the image region. This is because, for example, due to the displacement of the object relative to the particle microscope, the correspondence between the region of interest of the object and the corresponding image region of the recorded image is known only with limited precision. The displacement vector determined based on the correlation of the image data represents this displacement. In addition, the displacement vector enables the position of the intensity values of the image region in the coordinate system assigned to the recorded image to be shifted by the corresponding displacement vector, such that after the shift, the intensity values of the image regions of different recorded images are associated with the same position on the object.

[0027] Thus, by taking into account the displacement vectors, the image data of the image regions (including the image data representing an image of a given quality) can be combined to form an image having image data representing an image of a higher quality.

[0028] According to an exemplary embodiment, the method of recording an image further includes generating the resultant image by assigning the image data of the recorded image to the resultant image, wherein one or more intensity values among the plurality of intensity values of each image region are associated with a position in the coordinate system of the resultant image, and wherein the position in the coordinate system of the resultant image is calculated based on a displacement vector associated with the image region.

[0029] In a given recorded image having image regions associated with different displacement vectors, the image data contributing to the resultant image is shifted by different amounts depending on its position within the recorded image. At least some intensity values associated with positions within a predetermined image region in the coordinate system of the recorded image are shifted by a vector that is equal to the displacement vector associated with the image region or is calculated based on the displacement vector. A vector for shifting the positions of other intensity values can be determined, for example, based on the entire set of displacement vectors associated with the image regions of the recorded image. According to another example, a vector for shifting the positions of other intensity values can be determined by interpolation and / or extrapolation between such displacement vectors. According to some examples, a displacement vector for shifting the positions of other intensity values can be determined by fitting the parameters of a selected set of two-dimensional basis functions to the displacement vectors determined based on respective regions of interest. A parameterized set of basis functions can be applied to a given position to calculate the displacement at that given position. The basis functions can be selected to describe the characteristic distortion effects of a microscope, such as image rotation and quadratic or higher-order aberrations.

[0030] The method of recording an image shown can be advantageously used in cases where, during the recording of a plurality of images, the positions at which a particle beam impinges on an object deviate from the expected positions at which the particle beam impinges on the object, and these deviations are not the same for all positions of the recorded image, such that the deviations in the coordinate system of the recorded image are locally different. For example, this can occur when localized surface charges on the object produce deflections of the incident particle beam that are not the same for all positions of the object. In such cases, the method shown provides a flexible way to shift the image data of each recorded image based on different displacement vectors so that a resultant image with high quality and low image blur can be generated.

[0031] Multiple images of an object can be recorded using a particle microscope that generates a single particle beam or a multi-beam particle microscope that generates an array of multiple particle beams. The methods presented may be particularly useful for particle microscopes that generate multiple particle beams because each particle beam may experience drift or other deflections relative to the other particle beams, which may depend particularly on time. Image regions in each recorded image can then be determined such that the image data for each image region is generated by only one particle beam. In other words, a displacement vector can be associated with each particle beam, where the displacement vector can even account for the time variation of the deflections of each particle beam relative to the other particle beams in order to allow the superposition of the individual recorded images to form a resulting image with high quality.

[0032] According to a further exemplary embodiment, a method of recording an image using a multi-beam particle microscope includes scanning an array of multiple particle beams over a surface of an object and detecting signals generated by particles in the particle beams incident on the object, where adjacent particle beams are scanned over adjacent object regions and where the adjacent object regions overlap. The method may further include generating multiple images based on the detected signals such that each image is based on the detected signals generated by a single one of the multiple particle beams, where each image is an image of a corresponding surface region. The method may further include determining multiple image regions in each image such that an object portion of the object imaged into a given image region of a first image is also imaged into an image region of a second image, where the first image and the second image are generated based on detection signals generated by adjacent particle beams. The method may further include correlating the image data of multiple pairs of image regions of the images, where each pair of image regions is selected such that the same object portion of the object is imaged into both image regions of the pair of image regions. The method may further include: determining an image distortion of the multiple images based on correlating the image data, and generating a combined resulting image based on the multiple images and the determined image distortion. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Embodiments of the present invention will be illustrated with reference to the accompanying drawings, in which:

[0034] Figure 1 is a schematic representation showing a method of recording an image using a particle microscope;

[0035] Figure 2 is showing Figure 1 another schematic representation of the method of;

[0036] Figure 3 is a schematic representation showing another method of recording an image using a particle microscope;

[0037] Figure 4 is a front view of an object surface, showing an area of the object scanned by an array of particle beams in a multi-beam particle microscope;

[0038] Figure 5 is similar to Figure 4 a front view of an object surface, showing a method of imaging a larger portion of the object surface;

[0039] Figure 6A is Figure 4 a schematic illustration of details of;

[0040] Figure 6B is related to Figure 6A a diagram of an image of an exemplary object corresponding to the area of the object shown;

[0041] Figure 7A is Figure 6B a magnified view of an overlapping portion in the upper image shown;

[0042] Figure 7B is Figure 6B a magnified view of an overlapping portion of the lower image shown;

[0043] Figure 7C shows Figure 7A and Figure 7B a difference image of the images shown and a magnified portion of the difference image;

[0044] Figure 7D is a representation of displacement vectors determined based on Figure 7A and Figure 7B the images shown;

[0045] Figure 8A , Figure 8B , Figure 8C and Figure 8D show illustrative representations of image distortion that can be determined based on displacement vectors as shown in Figure 7D ;

[0046] Figure 9 is a flowchart showing an embodiment of a method of using a multi-beam particle microscope to record an image; and

[0047] Figure 10 is a flowchart showing a further embodiment of a method of using a multi-beam particle microscope to record an image. DETAILED DESCRIPTION

[0048] An embodiment of a method of using a particle microscope to record an image will be described below with reference to Figure 1 and Figure 2 . Figure 1is a schematic representation showing details of an image recorded using a particle microscope, and Figure 2 is a schematic representation showing details of a resulting image obtained from the Figure 1 recorded image.

[0049] In Figure 1 , the rectangle 11 drawn with a solid line represents a first recorded image recorded using a particle microscope, and in Figure 1 , the rectangle 11' drawn with a dashed line represents a second recorded image recorded using a particle microscope after the first recorded image 11 is recorded. Each of the recorded images 11, 11' includes image data containing a plurality of intensity values, where each intensity value is associated with a position in the coordinate system of the recorded image and a position on the object.

[0050] The intensity values are obtained by measuring when a particle beam is directed onto an object and a signal generated by the particle beam incident on the object is detected. The intensity values represent the intensity of these detected signals. For example, the particle beam can be systematically scanned over the object, and the intensity value can be determined for each scan position. In this case, it may be advantageous to store the intensity values as a two-dimensional array, where the array content is determined by the intensity values and the array indices represent the positions associated with the corresponding intensity values in the coordinate system of the recorded image. The array content is then generally referred to as the "pixels" of the recorded image. However, each intensity value can also be stored together with two additional values representing the coordinates of the intensity value in the recorded image. This may be advantageous in cases where the particle beam of the particle microscope is sequentially directed onto positions that are not arranged in a regular array of rows and columns, such as in a raster scanning method.

[0051] In this illustrative example, it is assumed that the object from which the recorded images 11 and 11' are obtained has a simplified uniform structure that has three significant features represented as triangles and diamonds in Figure 1 . The first of these three features is represented by a triangle 13 represented by a solid line in the first recorded image 11. The same feature of the object is represented by a triangle 13' represented by a dashed line in the second recorded image 11'. Clearly, even though the significant features 13 and 13' in the recorded images 11 and 11' are obtained by measuring the same feature of the object, they do not coincide. In Figure 1 's representation, the elements 13 and 13' in the recorded images 11 and 11' each have corresponding shapes such that the feature 13 coincides with the feature 13' when translated by the translation vector V1.

[0052] Another notable feature of the object is represented in the first recorded image 11 as a triangle 15 shown by a solid line, while in the second recorded image 11' it is represented as a triangle 15' shown by a dashed line. Similarly, even though the features 15 and 15' in the recorded images 11 and 11' are obtained from the measured intensity values associated with the same position on the object, they do not coincide with each other. Figure 1 The vector V2 in [description] represents the displacement for shifting the feature 15 of the recorded image 11 so that this feature coincides with the feature 15' of the recorded image 11'.

[0053] Similarly, Figure 1 Another feature of the object is shown, which is represented in the first recorded image 11 as a rhombus 14 drawn by a solid line, and in the second recorded image 11' as a rhombus 14' drawn by a dashed line. The displacement vector V3 indicates the displacement for shifting the feature 14 so that this feature coincides with the feature 14'. In Figure 1 In the illustrative example shown, even the outer boundaries represented by the rectangles 11 and 11' of the recorded images do not coincide. The vector V4 represents the displacement for shifting the rectangle 11 so that this rectangle coincides with the rectangle 11'. The displacement V4 can be caused, for example, by the movement of the object relative to the particle microscope that occurs between the start of the recording of the first recorded image 11 and the start or end of the recording of the second recorded image 11'. Such movement may be caused, for example, by inevitable drift (such as drift caused by temperature changes), or by intentional displacement that occurs when the object holder holding the object moves relative to the particle microscope.

[0054] From Figure 1 It is also clear that the displacement vectors V1, V2, and V3 are different in terms of their magnitude and direction. This may be caused, for example, by localized surface charges that the incident particle beam may generate on the object. Here, the amount and distribution of the localized surface charges on the object may be time-dependent, depending on the scanning strategy used for the particle beam, and the time-dependent localized surface charges may deflect the incident particle beam so that the position where the incident particle beam impinges on the object deviates from the expected position where the particle beam should impinge based on the current operating conditions of the particle microscope in the absence of localized surface charges. Therefore, it may not be possible to represent the coordinate transformation between the coordinate system of the recorded image and the coordinate system of the object by an affine transformation.

[0055] Obviously, the differences between the displacement vectors V1, V2, and V3 are caused by measurement. However, the displacement vector V4 can be arbitrarily selected. A change in the displacement vector V4 results in corresponding changes in the displacement vectors V1, V2, and V3, and in this case, the differences between the displacement vectors V1, V2, and V3 are maintained. In Figure 1In the example, the displacement vector V4 is selected such that the sum of the magnitudes of the vectors V1 and V2 is minimized. This is based on the assumption that the elements 13, 13′, 15, and 15′ result from the measurement of features on an object that do not move relative to each other when the recorded images 11 and 11′ are recorded.

[0056] Figure 2 is a schematic representation of the resulting image 21 obtained from the recorded images 11 and 11′. The resulting image 21 is obtained by combining the image data of the recorded images 11 and 11′.

[0057] If the recorded images 11 and 11′ are combined using a conventional method, the position of the intensity value assigned to the first recorded image 11 in the coordinate system of the first recorded image 11 would be shifted by the displacement vector V4 in order to provide the position of the intensity value assigned to the first recorded image 11 in the coordinate system of the resulting image 21. The position associated with its intensity value in the coordinate system of the second recorded image 11′ can be used unchanged without further shifting for the position of these intensity values in the resulting image 21.

[0058] Obviously, the intensity values originating from the feature 13 in the first recorded image 11 would then be inconsistent with the intensity values originating from the feature 13′ in the second recorded image 11′. As a result, the features of the object represented as the feature 13 and the feature 13′ in the recorded images 11 and 11′ respectively appear blurred in the resulting image. The same applies to the features of the object represented as the features 15 and 14, and the features 15′ and 14′ respectively in the recorded images 11 and 11′.

[0059] The exemplary method shown for combining the image data of the recorded images 11 and 11′ to form the resulting image 21 takes into account the above problems and allows for providing a resulting image 21 in which the representation of the features of the object appears clearer and less blurred.

[0060] For this purpose, the region of interest of the object is determined. The region of interest of the object can be a region of the object that contains suitable elements. Suitable elements can be features of the object that can be easily recognized, for example, in a particle micrograph. Suitable features are, for example, features that can be recognized in the recorded image, where the position of these features in the image can be determined with high precision. Aperiodic features having a sufficient size and that can be represented with high contrast in the recorded image are examples of suitable features. In the above reference Figure 1In the illustrative example discussed, two regions of interest are determined. These two regions of interest contain the features of the objects represented as triangles 13, 13', 15, and 15' in the recorded images 11 and 11'. At least one position of the object is assigned to each region of interest. In practice, since the lateral extent of the region of interest on the object is greater than zero, a large number of positions can be assigned to the region of interest. For example, the region of interest can be selected such that a suitable feature with a lateral extent on the object is completely contained within the region of interest and is accordingly regarded as a feature of interest.

[0061] After determining a plurality of regions of interest of the object, a plurality of image regions are determined in the recorded images 11, 11'. The image regions determined in the recorded image 11 are Figure 1 represented as rectangles 23 and 25 shown by solid lines, and the image regions determined in the recorded image 11' are Figure 1 represented as rectangles 23' and 25' shown by dashed lines. Each of the image regions 23, 23', 25, and 25' is associated with exactly one region of interest of the object.

[0062] Each of the image regions in the image regions is determined such that, in the recorded image, one of the regions of interest of the object is contained within the image region. For example, the image region 23 in the recorded image 11 is determined such that one of the regions of interest of the object is located within the image region 23. In this example, the region of interest of the object located within the image region 23 is Figure 1 the region of interest shown as triangle 13. The recorded image 11 contains such intensity values that have been selected for illustrative purposes at the vertices of triangle 13 and are associated with the positions having coordinates represented by the Figure 1 vector O1 in the coordinate system of the recorded image 11. This intensity value is also associated with the position on the object located at the vertex of the triangular feature.

[0063] The above description assumes that such intensity values have been actually measured that are associated with exactly that position on the object which is also associated with the region of interest of the object. Applied to Figure 1 the example, this means that when at least one position on the object associated with the region of interest is the lower vertex of the triangular feature on the object, the intensity value associated with exactly that position must also be obtained by directing the particle beam to exactly that position. In practice, this is not necessarily the case because the particle beam can be directed to one or more positions in the vicinity of that position. Such a vicinity is Figure 1 represented by the circle 27 in

[0064] In summary, the image region 23 of the recorded image 11 is determined such that it is associated with a region of interest of the object and that it includes intensity values associated with positions in the neighborhood 27 of the position O1 of the object, where the position is also associated with the region of interest associated with the image region 23.

[0065] The other image regions 23′, 25, and 25′ are determined in a corresponding manner.

[0066] After determining the image regions 23, 23′, 25, and 25′, a displacement vector is associated with each image region. In Figure 1 the example, the displacement vector V1 is associated with the image region 23, a displacement vector of zero length is associated with the image region 23′, the displacement vector V2 is associated with the image region 25, and a displacement vector of zero length is associated with the image region 25′.

[0067] The displacement vector V1 is determined by correlating the image data of the image region 23 with the image data of the image region 23′, with the constraint that the displacement vector of the image region 23′ has a zero length. Similarly, the displacement vector V2 is determined by correlating the image data of the image region 25 with the image data of the image region 25′. In this context, the displacement vectors can be determined according to various other methods. For example, the displacement vectors assigned to the image regions 23 and 23′ can be calculated such that a displacement vector of zero length is associated with the image region 23, while a displacement vector having the same length but opposite in orientation to the Figure 1 displacement vector V1 shown is associated with the image region 23′. It is also possible to determine the displacement vectors associated with the image regions 23 and 23′ such that they have the same length but opposite directions. Other methods of determining the displacement vectors are possible and can be used equivalently.

[0068] After determining the displacement vectors associated with the image regions, the resulting image 21 is generated from the image data of the recorded images 11 and 11′, where the displacement vectors V1 and V2 are taken into account.

[0069] Specifically, the intensity values of the recorded image 11 associated with the position O1 at the vertex of the triangle 13 and located within the region 23 are also associated with the position of the coordinates represented by the Figure 2 vector B1 in the coordinate system of the resulting image 21. The position B1 is calculated based on the displacement vector V1 associated with the image region 23. This also applies to the position at the vertex of the triangle 13′ within the image region 23′, and as a result, the intensity values associated with different positions in the coordinate systems of the recorded images 11 and 11′ are associated with the same position in the coordinate system of the resulting image 21.

[0070] This calculation based on the displacement vector associated with the image region is performed for at least one intensity value of the corresponding image region. It may be advantageous to also perform this calculation for intensity values associated with positions in the neighborhood of such positions. For example, all intensity values belonging to triangle 13 can be shifted by the displacement vector V1 in order to obtain a representation of triangle 13″ in the representation of the resulting image 21 with high sharpness and low blurriness.

[0071] Similarly, the displacement vector V2 associated with the image region 25 can be used to shift all intensity values belonging to triangle 15 in order to obtain a correct superposition with triangle 15′, such that Figure 2 the representation of triangle 15″ in shows a high contrast-to-noise ratio and low blurriness.

[0072] Obviously, the representation of the features of the object in the resulting image can be very sharp at least for such features located in the closer neighborhood of those regions of interest having associated image regions, the displacement vectors of which are determined by performing the correlation. In some cases, it may not be possible to determine a sufficient number of regions of interest of the object such that the regions of interest are spaced apart from each other by a small enough distance so that all intensity values of the recorded image lie within the image regions associated with the regions of interest. For example, this may occur when there are only a small number of features on the object, which can be distinguished in the recorded image and have a high enough contrast so that they are suitable for performing the correlation. In addition, in view of the higher performance of the method, the number of regions of interest of the object can also be reduced, since the fewer the number of regions of interest, the fewer the number of correlations that need to be calculated.

[0073] However, for the image regions of the recorded image that are not included in the image regions 23, 23′, 25, 25′, a representation can still be obtained that shows high sharpness and low blurriness in the resulting image. For example, those image regions that are not included in the image regions 23, 23′, 25, 25′ can also be associated with displacement vectors and these displacement vectors can be used when combining the corresponding image data to form the resulting image 21. This will be illustrated below with reference to the features of the object, the features of which are represented as rhombus 14 in the recorded image 11 and rhombus 14′ in the recorded image 11′ in Figure 1 The displacement vector V3 is associated with the image region in the recorded image 11 that contains the rhombus 14. The displacement vector V3 is not calculated based on the correlation of the image data associated with the image region containing the rhombus 14, but is calculated by performing interpolation between the displacement vectors associated with the image regions of the recorded images 11 and 11′. According to a further example, the displacement vector V3 is calculated based on a selected set of basis functions parameterized according to the correlation of the image data.

[0074] The rhombus 14 is located between the image regions 23 and 25 in the coordinate system of the recorded image 11. It can be assumed that the length and direction of the displacement vector V3, which can be associated with the image region containing the rhombus 14, are also between the lengths and directions of the displacement vectors V1 and V2 of the image regions 23 and 25, respectively. Therefore, the displacement vector V3 for the image region containing the rhombus 14 can be determined by interpolating between the displacement vectors V1 and V2. Various methods can be used to perform the interpolation. For example, the inverse distance between the position of the displacement vector calculated by interpolation and the corresponding image region can be used as a weight in the interpolation.

[0075] In practice, determining the displacement vector by interpolation can be accurate enough but still not precise enough. This is represented as Figure 2 in Figure 1 the rhombus 14 with the shifted displacement vector V3 and Figure 1 the rhombus 14' do not exactly coincide. Therefore, the resulting rhombus 14'' in the resulting image 21 is a slightly blurred rhombus. Still, compared with the conventional method of combining the images 11 and 11', the display of the resulting image is less blurred.

[0076] The image region 23 of the recorded image 11 is the first image region of the recorded image 11 and is associated with the first displacement vector V1. In the resulting image 21, at least one intensity value is associated with the position B1 in the coordinate system of the resulting image 21, where, in the recorded image 11, this intensity value is associated with the first position O1 in the first image region 23 in the coordinate system of the recorded image 11.

[0077] The image region 25 of the recorded image 11 is the second image region of the recorded image 11 and is associated with the second displacement vector V2. The first displacement vector V1 and the second displacement vector V2 are different from each other. In the resulting image 21, at least one intensity value associated with the second position O2 in the second image region 25 in the coordinate system of the recorded image 11 is associated with the second position B2 in the coordinate system of the resulting image 21.

[0078] These determinations are performed to satisfy the following relationship:

[0079] B2 - B1 = O2 + V2 - O1 - V1

[0080] In the example shown in Figure 1 the rhombus 14 is approximately located between the elements 13 and 15. In addition, the rhombus 14 is outside each of the image regions 23 and 25. Figure 1 The position of the rhombus 14 in Figure 1 is represented by the vector O3. Since the rhombus 14 is approximately located between the elements 13 and 15, the following relationship holds:

[0081] |(O2 - O1)i| > |(O3 - O1)i|.

[0082] In this text, (O2 - O1)i and (O3 - O1)i represent the ith component of the vector differences (O2 - O1) and (O3 - O1). The ith component of the vector Xi can be calculated as Xi = X * ei, where ei represents the ith unit vector in the coordinate system of the recorded image, and "*" indicates the scalar product.

[0083] The intensity value of the recorded image 11 assigned to the position O3 outside the image regions 23 and 25 is assigned to the position in the resulting image 21 with the coordinate vector B3 calculated based on the displacement vector V3 as shown. The displacement vector V3 can be calculated by interpolation such that the following relations hold: Figure 2 B3 - B1 = O3 + V3 - O1 - V1 and

[0084] |(V2 - V1)i| > |(V3 - V1)i|,

[0085] where (V2 - V1)i and (V3 - V1)i are the ith components of the vector differences (V2 - V1) and (V3 - V1), respectively.

[0086] The displacement vectors for all the image data of the recorded image can be determined based on the displacement vectors V1 and V2 associated with the image regions, similar to the exemplary displacement vector V3 shown above. Many interpolation and extrapolation methods (such as the spline method) can be used for this purpose. Although the number of regions of interest in the above illustrative example is two, a larger number of regions of interest can be used.

[0087] A further embodiment of a method for recording an image using a particle microscope will be described below with reference to

[0088] FIG. Figure 3 illustrates a method for recording an image of the surface 31 using a particle microscope. In this example, the particle microscope is a multi-beam particle microscope that simultaneously directs a plurality of particle beams onto the surface of the object. Each particle beam can scan the object region 33 associated with that particle beam. The object regions are square-shaped and are arranged adjacent to each other. The adjacent object regions 33 overlap each other. This means that the positions of the objects located near the edge of one of the object regions 33 are also located near the edge of the object region arranged adjacent to that object region. Thus, such positions on the object surface are scanned by two or three particle beams and are included in two or three of the recorded images. Figure 3 is a schematic elevation view of the surface of the object 31. A method for recording an image of the surface 31 using a particle microscope is provided. In this example, the particle microscope is a multi-beam particle microscope that simultaneously directs a plurality of particle beams onto the surface of the object. Each particle beam can scan the object region 33 associated with that particle beam. The object regions are square-shaped and are arranged adjacent to each other. The adjacent object regions 33 overlap each other. This means that the positions of the objects located near the edge of one of the object regions 33 are also located near the edge of the object region arranged adjacent to that object region. Thus, such positions on the object surface are scanned by two or three particle beams and are included in two or three of the recorded images. Figure 3An exemplary object region 330 is shown as a rectangle drawn in solid lines, while four object regions 331, 332, 333, and 334 arranged adjacent to the exemplary object region 330 are shown as rectangles drawn in dashed lines.

[0089] In the example shown, the object regions 33 all have a square shape, and the object regions are arranged in a checkerboard pattern. Other configurations are possible. For example, the object regions can have a rectangular shape such that their edges, for example, have different lengths. Additionally, the object regions can have, for example, a hexagonal shape, where the object regions are arranged in a honeycomb pattern.

[0090] The particle beam that scans the object region 33 generates signals such as secondary electrons, which can be detected and assigned to the corresponding particle beam that generated the detected signal. An intensity value can be determined based on the detected signal, where the intensity value is associated with the position in the coordinate system of the object on which the particle beam impinged when generating the detected signal. The sequence of intensity values obtained from the sequence of detected signals provides the image data of the recorded image. Thus, the recorded image can be associated with each object region 33.

[0091] Figure 3 The vectors S0, S1, S2, S3, and S4 in represent position vectors in the coordinate system of the object, and these position vectors respectively represent the positions of the object regions 330, 331, 332, 333, and 334 in the coordinate system of the object. These vectors can be obtained, for example, through a suitable calibration performed by a particle microscope. These position vectors can be used to determine the correspondence between the position in the coordinate system of the recorded image and the position in the coordinate system of the object.

[0092] The object has a plurality of significant features 35 distinguishable in the particle micrograph. These features can be arranged on the object in a regular pattern. A subset of these features located close to the object region 330 is schematically shown as diamonds in Figure 3 These features form a region of interest for the method of recording an image using a particle microscope as shown. Then, the recorded images of the individual object regions 33 will contain a plurality of image regions, where the plurality of image regions can be determined in each recorded image, where each image region is associated with a region of interest corresponding to the significant feature 35, and where the image region includes intensity values associated with positions within and in the neighborhood of the corresponding region of interest. A part of the region of interest corresponding to the significant feature 35 shown in Figure 3 is located close to the edge of the object region 330 in the part that overlaps with the adjacent object regions 331, 332, 333, and 334.

[0093] As described above, displacement vectors can be associated with image regions of different images, where the different image regions are associated with the same region of interest of an object, and where the displacement vectors are determined based on a correlation performed on the image regions. For illustrative purposes, such displacement vectors are represented in Figure 3 by vector 37 with an exaggerated length. The displacement vector 37 is associated with an image region of the object region 330 of the recorded image, where the image region exists in the images of adjacent object regions 331, 332, 333, and 334 that are associated with the same region of interest.

[0094] The arrangement of the displacement vector 37 indicates that the imaging of the object region 330 into the recorded image includes barrel image distortion. On the other hand, in the case where the central object region 330 is substantially free of distortion and where the adjacent object regions 331, 332, 333, and 334 each have pincushion distortion, the arrangement of the displacement vectors shown can also be explained. Figure 3 This ambiguity can be resolved by considering all the overlaps between all pairs of adjacent object regions and minimizing the total amount of the determined displacements.

[0095] By performing interpolation or extrapolation based on the displacement vector 37, the displacement vector can be associated with each position in the coordinate system of the recorded image. Thus, the positions associated with the intensity values of the recorded image can be corrected based on these displacement vectors to obtain an image of the object region 330 with less distortion.

[0096] This process can be applied to all the overlaps between all pairs of adjacent object regions 33 of the object in order to determine the correction of the distortion in the image recorded by multiple particle beams. These determined distortion corrections can be used to correct the distortion in the image recorded by multiple particle beams. For example, the determined distortion corrections can be used to combine the image data of the individual object regions and form a high-quality image of the overall object. However, it is not necessary to generate a combined image. For example, in an application where it is desired to evaluate whether there are defects in a semiconductor wafer, it is sufficient to use the determined distortion corrections to determine the positions of the defects on the wafer based on a small portion of the image, without explicitly calculating the combined image of the overall object.

[0097] In Figure 3 the illustrated example, the displacement vectors 37 are evenly distributed in all directions. The arrangement of the displacement vectors 37 indicates that the image distortion disappears at the center of the image. In other examples, it may occur that the displacement vectors determined by interpolation do not disappear at the center of the image region.

[0098] The types of algorithms for performing interpolation and extrapolation can vary widely. For example, it may be advantageous to determine the displacement vectors assigned to image regions such that one component of the displacement vector has a predetermined value, such as zero. For example, when scanning a particle beam over an object along a horizontal scan line, the coordinate system of the recorded image can be selected such that one unit vector of the coordinate system is parallel to the scan line. The displacement vectors can then be determined such that the values of those components corresponding to the unit vectors not oriented parallel to the scan line are zero. This can be achieved, for example, by projecting the displacement vectors determined by two-dimensional correlation onto the unit vector oriented parallel to the scan line. Thereafter, the displacement vectors determined based on the correlation of the image data are then oriented parallel to the scan line. Similarly, the displacement vectors determined by interpolation or extrapolation from these displacement vectors of the image regions are oriented in that direction. Specifically, the displacement vectors associated with the image regions and the displacement vectors determined by interpolation or extrapolation will be parallel to each other. Such a method is particularly advantageous for compensating for scanning distortions that occur in image recording methods in which a particle beam is scanned over an object.

[0099] Figure 4 is a front view of the object surface, showing the object region scanned by an array of particle beams in a multi-beam particle microscope.

[0100] A multi-beam particle microscope generates an array of multiple particle beams. These particle beams are scanned over the object such that each individual beam is scanned over a rectangular object region 51. In Figure 4 the example shown, each object region has a height of 10 μm and a width of 12 μm on the object surface 53. Adjacent beams in the array of multiple particle beams are scanned over adjacent object regions 51 on the object surface 53. Here, the object regions 51 scanned by adjacent particle beams overlap by a certain amount, as will be further described below.

[0101] Images associated with each object region 51 can be generated by scanning an array of particle beams over the multiple object regions 51. By causing the bundled particle beams in the particle beam array to deflect jointly so as to scan the object regions 51 in parallel, multiple images can be generated simultaneously.

[0102] The object regions 51 that can be scanned simultaneously using a multi-beam particle microscope are arranged in a pattern corresponding to the pattern of the particle beam array. In the example shown, the multi-beam particle microscope uses 91 particle beams, and the object regions 51 are arranged in multiple rows, where the top row 55 includes six object regions 51. The number of object regions 51 in each row increases by one row by row until the center row 57 includes eleven object regions 51. From here on, the number of object regions 51 in each row decreases by one until the bottom row 59 reverts to having six object regions 51.

[0103] The portion 61 of the object surface 53 that can be scanned simultaneously with 91 beams has a generally hexagonal shape with a stepped boundary.

[0104] Figure 5 It is an illustration of an extended portion of the object surface 53 that is scanned by translating the object surface 53 relative to the electron microscope between each recording of 91 images of the corresponding object region 51. Clearly, the object surface 53 can be translated relative to the particle microscope such that the hexagonal portion 61 covers the entire surface 53 of the object, and thus an image of the extended portion of the object surface 53 can be obtained by the process shown.

[0105] Background information related to a multi-beam particle microscope is shown in US 9,536,702 B2, and other documents are cited therein.

[0106] In the example shown, the size of the object region 51 scanned by a single beam is 10 μm x 12 μm. The pixel size used for imaging on the object surface 53 is 2 nm. Thus, the recorded image of one object region has 13,000,000 pixels. In the example, the color depth for storing image data is 8 bits, such that the image data associated with one image occupies 30 MB. Since the number of beams is 91, the image data obtained from one simultaneous scan of one object part 61 occupies 2.73 GB. The duration of one simultaneous scan is one second, such that the time required to scan 1 mm 2 of the object surface 53 is 2.35 hours, and the amount of the generated image data is 22.75 TB.

[0107] Clearly, processing such amounts of data within a given time is challenging. Here, it should be noted that before the images of the individual object regions 51 can be stitched together to form an overall image of the extended object region as Figure 5 shown, the image data must pass through an intensive image processing pipeline.

[0108] If, for example, only defects in a semiconductor are to be detected, there are cases where it is not necessary to display the overall image. However, also in such cases, the distortion of the individual images can be determined and corrected. In semiconductor applications, the distortion mainly results in an offset of features from their expected positions based on the circuit design. The distortion correction must restore the original position with a precision better than, for example, one-fifth of the feature size. If the feature is a 40-nm contact hole, this precision corresponds to better than 8 nm, corresponding to 4 pixels (assuming a pixel size of 2 nm).

[0109] Figure 6A is a front view of the object surface 53, showing Figure 4Details. The adjacent object regions 51 overlap by a certain amount such that each pair of adjacent object regions 51 scanned by a corresponding pair of adjacent particle beams share an overlapping portion 63 on the object surface 53.

[0110] Figure 6B Two image portions 65 and 66 of an exemplary object recorded by a particle microscope are shown.

[0111] Figure 6B The upper image 65 shown is the lower portion of the image obtained from the surface portion 51 labeled “ID3” in Figure 6A and the lower image 66 shown is the upper portion of the image recorded from the object region 51 labeled “ID1” in Figure 6B Each of the images 65 and 66 includes an image portion 67 that is an image of the overlapping portion 63 between adjacent object regions 51 on the object surface 53. Figure 6A

[0112] Figure 7A is Figure 6B a magnified view of the image portion 67 of the upper image 65 shown, and Figure 7B is Figure 6B a magnified view of the image portion 67 of the lower image 66 shown.

[0113] Figure 7C shows the difference image 71 obtained by subtracting the intensity values of the image portion 67 shown in Figure 7B from the intensity values of the image portion 67 shown in Figure 7A

[0114] A plurality of overlapping object portions 73 can be defined and selected within the overlapping portion 63 on the object surface 53.

[0115] Figure 7C An exemplary number of three overlapping object portions 73 and magnified views 75 of these overlapping object portions 73 are shown. Each magnified view 75 includes a representation of a displacement vector 77 that represents the shift performed between the overlapping object portions 73 between the two images 65 and 66. The displacement vector 77 can be determined based on the correlation performed between the image data associated with the overlapping object portions 73 in the two images 65 and 66.

[0116] Figure 7D is a graphical representation of a plurality of displacement vectors 77 determined by correlating the plurality of overlapping object portions 73 of the overlapping portion 63 between adjacent object regions 51. Here, the displacement vectors obtained from the three overlapping object portions 73 shown in Figure 7C are represented by arrows drawn with solid lines, while from Figure 7C ​​The displacement vectors obtained for additional overlapping object portions not shown in the figure are drawn by dashed lines. In the example shown, the number of displacement vectors 77 obtained for each overlapping object region is nine. A higher or lower number of overlapping object portions 73 may be used to determine a higher or lower number of displacement vectors 77.

[0117] Each object region 51 includes a perimeter of the overlap portion 63, and a plurality of displacement vectors 77 distributed around the perimeter may be determined as described above. The set of displacement vectors 77 distributed around the perimeter of the image of the object region 51 indicates image distortion resulting from imaging the object region 51 into the image associated with the object region 51. In other words, the displacement vectors 77, or data derived from or representing the displacement vectors distributed around the perimeter of the image, represent a "fingerprint" of image distortion included in the overlapped image.

[0118] Figure 8A , Figure 8B , Figure 8C and Figure 8D An illustrative representation of exemplary image distortion that may be determined based on displacement vectors 77 distributed around the perimeter of the image is shown. Figure 8A shows the image distortion resulting from the scale error, Figure 8B shows the image distortion arising from rotation error, and Figure 8C The image distortion resulting from shearing errors that occur in imaging is shown. In addition, Figure 8D A quadratic distortion is shown which cannot be represented by an affine transformation. This distortion is an example of distortion arising from the inaccuracy of the beam scanning system.

[0119] A set of appropriately selected basis functions may be used to correct the distortion of an image. Herein, it may be considered that all image distortions must be calculated in one optimization process, since the displacement vectors 77 originate from common overlapping portions of the images. The distortion correction method may adapt the weights of the basis functions for each image across all images until the resulting displacement vector fingerprint 77 substantially disappears or is minimized. Typically, there will be residual displacement vectors remaining, which may be eliminated by extending the basis function set to add more correction degrees of freedom, or may be acceptable if the width of the distribution of the residual displacement vectors, their standard deviation, is below a threshold of, for example, 2 pixels.

[0120] Obviously, image distortions may be determined for each image of each object region 51. When generating a combined resulting image based on the images of the object regions 51, these image distortions may be taken into account.

[0121] Figure 9A flowchart showing an embodiment of a method of using a multi-beam particle microscope to record images is shown. Specifically, in the main process 102, an array of particle beams of a particle microscope is scanned simultaneously over the object regions 51 in order to obtain images of these object regions 51. The control system 101 stores the recorded images in the image database 103. The stored images can be analyzed or further processed at a later time. Such further processing of the images can include stitching the images together to form a larger combined resultant image. As described above, it may be advantageous to generate the combined resultant image based on the determined image distortion of the individual images. The determination of the image distortion includes computationally expensive calculations such as determining displacement vectors 77 based on correlations made between pairs of multiple image portions. These calculations require a large amount of computational time, which can be of the same or similar order of magnitude as the time required to record images using the particle microscope under some operating conditions.

[0122] Assuming that each surface portion 51 is scanned using the same scanning strategy, such as a top-down line scan, it is obvious that some of the multiple overlapping object portions 77 are scanned earlier than other overlapping object portions 77.

[0123] In Figure 9 In the method shown, the calculation step 105 determines whether the next overlapping object portion 77 has been scanned by two adjacent particle beams. If the answer is no, the process waits in step 107. If the answer is yes, in step 109 the overlapping object portion 77 that has become available so far is extracted from the recorded image data. In step 111, the image portion corresponding to the overlapping object portion 77 is created as a "patch". In step 113, these image portions are analyzed to determine whether they are suitable for determining the displacement vector 77. For example, an image portion that shows no structure at all will be useless. Thus, step 113 determines whether the image portion includes a structure originating from a feature on the object and whether these structures have an orientation that allows the determination of a displacement vector within a desired range of orientations.

[0124] In step 115, the displacement vectors 77 corresponding to the useful overlapping object portions 75 are determined and in step 119 these displacement vectors are stored in the database 117. The displacement vectors 77 indicate the image distortion of the individual images and can be used when analyzing the images stored in the image database 103 at a later time.

[0125] Figure 10This process is shown in more detail in the flowchart shown. Based on the displacement vectors 77 stored in the database 117, the image distortion of each image of the object region 51 can be determined in step 121. The image distortion relates to the set of all object parts 51 that can be scanned simultaneously using the particle beam array. The images of the hexagonal surface parts 61 can be stitched together based on this information in step 123, and the corresponding combined resulting image can be output in step 125.

[0126] Assuming that the image distortion associated with each particle beam remains sufficiently constant over a certain period of time, the calculation of the displacement vectors for some of the recorded hexagonal surface parts 61 can be omitted and the previously determined image distortion can be used when generating the combined resulting image. Therefore, the image distortion associated with each particle beam can be determined at regular intervals (such as whenever the next, for example, 10, 20, 50, or 100 hexagons are scanned), thereby saving a significant amount of computing time.

Claims

1. A method of using a multi-beam particle microscope to record an image, wherein, The method includes: recording a plurality of images of an object by scanning a plurality of particle beams on the object and detecting signals generated by the particle beams, wherein each of the recorded images is associated with image data, wherein each of the image data among the image data includes a plurality of intensity values, wherein the intensity value of each of the plurality of recorded images is determined based on the detected signal generated by one of the plurality of particle beams, and wherein each intensity value is associated with a position in the coordinate system of the recorded image and a position on the object; determining a plurality of regions of interest, wherein each region of interest is associated with at least one position on the object; determining a plurality of image regions in each of the recorded images, wherein each of the plurality of image regions is associated with one of the plurality of regions of interest, and wherein each image region includes intensity values associated with positions in the neighborhood of the position of the object, and the position of the object is also associated with the region of interest associated with the image region; determining a plurality of displacement vectors, wherein each of the plurality of displacement vectors is associated with a given image region among the plurality of image regions, wherein the given image region is associated with a given region of interest among the plurality of regions of interest, wherein the displacement vector associated with the given image region is determined by correlating the image data associated with the given image region with the image data of other image regions also associated with the given region of interest; and determining image distortion based on the image data of the recorded images and the determined displacement vectors.

2. The method according to claim 1, further comprising generating a resulting image based on the image data of the recorded images and the determined image distortion.

3. The method according to claim 2, wherein, When generating the resulting image, at least one intensity value of a given image region of a given recorded image is associated with a position in the coordinate system of the resulting image, wherein the position in the coordinate system of the resulting image is determined based on the displacement vector associated with the given image region.

4. The method according to any one of claims 1 to 3, wherein, The displacement vectors assigned to the image regions of the recorded image are different from each other.

5. The method according to claim 2 or 3, Among them, a first displacement vector is associated with a first image region of a given recorded image; wherein at least one intensity value associated with a first position in the first image region of the recorded image is associated with a first position in the resulting image in the coordinate system of the resulting image; wherein a second displacement vector is associated with a second image region of the given recorded image; wherein the second displacement vector is different from the first displacement vector; wherein at least one intensity value associated with a second position in the second image region of the recorded image is associated with a second position in the resulting image in the coordinate system of the resulting image; wherein the intensity value is assigned to the second position in the second image region of the recorded image; and wherein the following relationship is satisfied: B2 - B1 = O2 + V2 - O1 - V1; wherein, B1 represents the coordinate vector of the first position in the coordinate system of the resulting image; B2 represents the coordinate vector of the second position in the coordinate system of the resulting image; O1 represents the coordinate vector of the first position in the coordinate system of the given recorded image; O2 represents the coordinate vector of the second position in the coordinate system of the given recorded image; V1 represents the first displacement vector; and V2 represents the second displacement vector.

6. The method according to claim 5, Among them, a third displacement vector is assigned to a third image region of the given recorded image; wherein the third image region is located between the first image region and the second image region; wherein at least one intensity value is associated with a third position in the resulting image in the coordinate system of the resulting image; wherein the intensity value is associated with a third position in the third image region of the recorded image; wherein the following relationship is satisfied: |(O2 - O1)i| > |(O3 - O1)i|, |(V2 - V1)i| > |(V3 - V1)i|, wherein, O3 represents the coordinate vector of the third position in the coordinate system of the given recorded image; V3 represents the third displacement vector; (O2 - O1)i represents the i-th component of the difference vector (O2 - O1); (O3 - O1)i represents the i-th component of the difference vector (O3 - O1); (V2 - V1)i represents the i-th component of the difference vector (V2 - V1); and (V3 - V1)i represents the i-th component of the difference vector (V3 - V1).

7. The method according to any one of claims 1 to 3, wherein The displacement vectors associated with these image regions are oriented parallel to each other.

8. The method according to any one of claims 1 to 3, wherein Positions in the neighborhood of the same position of the object are arranged to be at a distance less than 0.1 times the maximum diameter of the recorded image in the coordinate system of the recorded image from the position of the object in the coordinate system of the recorded image.

9. The method according to any one of claims 1 to 3, wherein, The lateral dimension of the image region of each recorded image measured in any direction in the coordinate system of the recorded image is less than 0.5 times the extent of the recorded image in the corresponding direction.

10. The method according to any one of claims 1 to 3, wherein The positions of these regions of interest in the coordinate system of the object are predetermined.

11. The method according to any one of claims 1 to 3, wherein The positions of these regions of interest are determined based on an analysis of these recorded images.

12. The method according to any one of claims 1 to 3, further comprising holding the object in a fixed position relative to the particle microscope while recording the plurality of images.

13. The method according to any one of claims 1 to 3, further comprising shifting the object relative to the particle microscope while recording the plurality of images.

14. The method according to any one of claims 1 to 3, further comprising shifting the object relative to the particle microscope in a direction parallel to the surface plane of the object while recording the plurality of images.

15. The method according to any one of claims 1 to 3, wherein The plurality of images are recorded one after another.

16. The method according to any one of claims 1 to 3, wherein Recording of the given image is started only after recording of the image preceding the given image has been completed.

17. The method according to any one of claims 1 to 3, Among them, recording the image includes scanning at least one particle beam on the object and detecting a signal generated by the at least one particle beam; wherein the intensity value of the recorded image is determined based on the detected signal; and The position in the coordinate system of the recorded image that is associated with these intensity values is determined based on the scanning position of the at least one particle beam that occurs during scanning.

18. The method according to any one of claims 1 to 3, Among them, adjacent particle beams are scanned over adjacent object regions; wherein the adjacent object regions partially overlap; and wherein a plurality of regions of interest are associated with positions on the object that are located within at least two adjacent object regions.

19. A method of using a multi-beam particle microscope to record an image, wherein, The method includes: scanning an array of a plurality of particle beams on the surface of an object and detecting signals generated by particles in the particle beams incident on the object, wherein adjacent particle beams are scanned over adjacent object regions, and wherein the adjacent object regions overlap; generating a plurality of images based on the detected signals such that each image is based on the detected signals generated by a single one of the plurality of particle beams, wherein each image is an image of a corresponding surface region; determining a plurality of image regions in each image such that an object portion of the object imaged into a given image region of a first image is also imaged into an image region of a second image, wherein the first image and the second image are generated based on detection signals generated by adjacent particle beams; correlating the image data of a plurality of pairs of image regions of the images, wherein each pair of image regions is selected such that the same object portion of the object is imaged into both image regions of the pair of image regions; and determining image distortion of the plurality of images based on the correlation of the image data.

20. The method according to claim 19, further comprising generating a combined resulting image based on the plurality of images and the determined image distortion.

21. The method according to claim 20, further comprising determining positions on the object corresponding to the image data of at least one of the plurality of generated images based on the plurality of images and the determined image distortion.

22. The method according to claim 21, wherein The object is a semiconductor circuit, and wherein the positions on the object include defects of the semiconductor circuit.

23. The method according to any one of claims 19 to 22, wherein, The overlapping portions of the pairs of adjacent object regions include a plurality of overlapping object portions imaged into a plurality of pairs of image regions.

24. The method according to claim 23, wherein, The number of overlapping object portions scanned by a pair of adjacent particle beams and imaged into a plurality of pairs of image regions is greater than five.

25. The method according to claim 24, wherein, The number is greater than ten.

26. The method according to one of claims 19 to 22, wherein, The scanning of the first object portion imaged into a first pair of image regions by the pair of adjacent particle beams is performed before the scanning of the second object portion imaged into a second pair of image regions by the same pair of adjacent particle beams, wherein the correlation of the image data of the first pair of image regions is started before the scanning of the object portion imaged into the second pair of image regions by the same pair of adjacent particle beams.

27. The method according to claim 26, wherein, The correlation of the image data of the first pair of image regions is completed before the scanning of the object portion imaged into the second pair of image regions by the same pair of adjacent particle beams is completed.

Images