Computer implemented method for controlling a slice thickness when generating a 3D tomographic image of an inspection volume in a semiconductor wafer

Abstract

The invention relates to a system and method for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer, the method comprising: depositing a marker structure (66) on the wafer surface (32), alternatingly: exposing a cross-section surface (48) in the inspection volume by milling into the inspection volume and the marker structure (66) with a focused ion beam (34) in a focused ion beam system (24), wherein a current position of the focused ion beam (34) is repeatedly derived from a cross-section of the marker structure (66), wherein the cross-section of the marker structure (66) is determined by measuring numbers and / or types of secondary particles released during milling, and wherein the milling is stopped when the current position reaches a desired position; and imaging the cross-section surface (48).

Description

[0001] 25.08.2025 h - se

[0002] - 1 -

[0003] Computer implemented method for controlling a slice thickness when generating a 3D tomographic image of an inspection volume in a semiconductor wafer

[0004] Related Applications

[0005] This application claims benefit of German patent application No. 10 2024 126534.5 filed on September 15th, 2024, which is hereby incorporated by reference in its entirety.

[0006] Field of the Invention

[0007] The invention relates to systems and methods for generating a 3D tomographic image of a wafer with a high accuracy, in particular to systems and methods that in situ control a slice thickness when generating a 3D tomographic image.

[0008] Background of the Invention

[0009] Semiconductor manufacturing involves precise manipulation, e.g., etching, of materials such as silicon or oxide at very fine scales in the range of nm. Therefore, a quality management process is important for ensuring high quality standards of the manufactured wafers.

[0010] A wafer made of a thin slice of silicon serves as the substrate for microelectronic devices containing semiconductor structures built in and upon the wafer. The semiconductor structures are constructed layer by layer using repeated processing steps that involve repeated chemical, mechanical, thermal and optical processes. Dimensions, shapes and placements of the semiconductor structures and patterns are subject to several influences. One of the most crucial steps is the photolithography process.

[0011] Photolithography is a process used to produce patterns on the substrate. The patterns to be printed on the surface of the substrate are generated by computer-aided-design (CAD). From the design, for each layer a photolithography mask is generated, which contains a magnified image of the computer-generated pattern to be etched into the substrate. The photolithography mask can be further adapted, e.g., using optical prox- imity correction techniques. During the printing process an illuminated image projected from the photolithography mask is focused onto a photoresist thin film formed on the substrate.

[0012] Due to the growing integration density in the semiconductor industry, photolithography masks have to image increasingly smaller structures onto wafers. The production process of photolithographic masks and templates for nanoimprint photolithography is, therefore, becoming increasingly more complex and, as a result, more time-consuming and ultimately also more expensive. With the advent of EUV photolithography scanners, the nature of masks changed from transmission-based to reflection-based patterning.

[0013] Each defect in the photolithography mask can lead to unwanted behavior of the produced wafer. Therefore, quality management including defect detection is important.

[0014] In order to analyze a wafer for defects, a 3D tomographic image is often generated of the wafer. Structures are examined in the 3D tomographic image using, for example, image processing or machine learning methods. The structures or properties of the structures are compared to reference structures or norm values, and depending on the deviation defects are detected.

[0015] To generate a 3D tomographic image of a wafer, a focused ion beam system combined with a scanning electron microscope (FIB-SEM) is often used. FIB-SEM alter- natingly exposes a cross-section surface of the inspection volume by milling into the inspection volume using the focused ion beam system (FIB) and then images the cross-section surface using the scanning electron microscope (SEM). However, 3D tomography of semiconductor structures relies on very uniform slicing of the wafer with the focused ion beam system with a slice thickness in the range of nm.

[0016] A known approach to control the slice thickness is disclosed in US 10586680 B2. It deposits a tracking pad on the surface of the wafer in an XY-plane. The tracking pad contains features such as straight-line structures, whose distance in X-direction is used to derive the corresponding Y-position of the FIB beam on the inspection volume from a SEM image. Due to an interference between the FIB and the SEM signals, the FIB and the SEM process are carried out alternatingly. Since the Y-position is derived from a SEM image, the Y-position of the FIB cannot be controlled in situ but only based on the last SEM image and a prediction model for the mill time required to achieve a specific slice thickness. Therefore, the actual Y-position is only approximately known during milling and can be verified only post mill after acquiring the next SEM image.

[0017] Yet, two problems are inherent to this approach that prevent highly homogeneous slicing with nm thickness: a) Pattern matching is required to match features in a first SEM image to corresponding features in a subsequent SEM image. However, the feature contrast in the SEM image depends, among others, on the SEM landing energy, but the SEM landing energy is adjusted for high quality semiconductor imaging and not to optimize the contrast of the features in the tracking pad. The suboptimal contrast in combination with a very delicate wafer sample preparation process to generate buried features in the tracking pad leads to failures in the pattern matching that, consequently, lead to incorrect distance measurements and, ultimately, to incorrect Y-positions of the FIB beam. b) SEM imaging alternating with FIB milling requires a model to predict the mill time to cut a slice with target thickness. But random effects such as external shocks or wafer sample charging induced FIB beam detachments or pull ins cannot be predicted and eventually lead to inhomogeneous slicing for thin slices with a thickness of a few nm.

[0018] Therefore, it is an objective of the invention to obtain a 3D tomography imaging method for wafers that allows for a highly homogeneous slicing, in particular for slices with 4 nm, 2nm, 1 nm, 0.5nm thickness or below. It is another objective of the invention to provide an in situ slice thickness measurement.

[0019] The objectives are achieved by the invention specified in the independent claims. Advantageous embodiments and further developments of the invention are specified in the dependent claims.

[0020] Summary of the invention Embodiments of the invention concern methods and systems for controlling the slice thickness when generating a 3D tomographic image of an inspection volume of a wafer.

[0021] According to a first embodiment, a method for controlling a slice thickness, when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer, comprises: depositing a marker structure on the wafer surface extending in an XY- plane, wherein the marker structure comprises recognizable edges allowing an identification of the corresponding Y-position from a cross-section profile of the marker structure along X-direction; alternatingly carrying out at least the following steps: exposing a cross-section surface in the inspection volume by milling into the inspection volume and the marker structure with a focused ion beam in a focused ion beam system, whose optical axis is orthogonal to the X-direction at a fixed angle with respect to the Y-direction, wherein a current Y-position of the focused ion beam relative to the inspection volume is repeatedly derived from a cross-section profile of the marker structure along X-direction, wherein the cross-section profile of the marker structure is determined by measuring numbers and / or types of secondary particles released when milling along the marker structure, and wherein the milling is stopped when the current Y-position reaches a desired Y-position; and imaging the cross-section surface with the charged particle imaging system to obtain a cross-section image slice of the 3D tomographic image of the inspection volume.

[0022] By using specifically configured marker structures with recognizable edges that can be used to determine the Y-position from a cross-section profile of the marker structure and by measuring numbers and / or types of secondary particles released during milling along the marker structure, the Y-position of the focused ion beam can be controlled in situ during milling. Thus, milling can be stopped as soon as a desired slice thickness is obtained, and SEM imaging can be run on the semiconductor device. The problems inherent to common approaches that use inaccurate models to predict the required mill time to obtain a specific slice thickness and that verify the Y- position in the SEM image only after milling is already completed are, thus, alleviated. Instead, the Y-position can be accurately controlled during milling. By use of the method, the SEM parameters can be solely optimized to obtain high quality SEM images of semiconductor devices without, simultaneously, requiring sufficient contrast of the line features in the tracking pad in the SEM image. In this way, tracking recipes can be transferred to all kinds of wafer samples.

[0023] Furthermore, the improved control of the Y-position of the FIB beam allows the use of lower landing energies during milling of the wafer. On the one hand, lower landing energies reduce amorphization, i.e., the destruction of the wafer surface by implanting ions. This is particularly important for logic wafers. Memory wafers usually have structures that are homogeneous in Z-direction such that their structures can also be examined further away from the wafer surface. However, logic wafers usually do not have homogeneous structures, and examination of all layers is important. Therefore, lower landing energies are beneficial. However, on the other hand, at lower landing energies the focused ion beam is more susceptible to electric or magnetic perturbations and is more easily deflected from a desired position, e.g., due to charges on the wafer surface. Such deflections of the focused ion beam can be prevented by monitoring the Y-position of the FIB beam according to the methods of the invention. Therefore, the methods according to the invention allow the use of lower landing energies.

[0024] The term “slice thickness” refers to a quantity proportional to the distance in Y-direc- tion between consecutive cross-section surfaces e.g., the orthogonal distance of two consecutive milled cross-section surfaces.

[0025] A “cross-section profile” of a marker structure along X-direction refers to a signal indicating the number of secondary particles of a specified type or the different types of secondary particles released when milling along the marker structure. Secondary particles comprise electrons and ions that can be of different types.

[0026] A “marker structure” is a structure that is deposited on the surface of the wafer and is used for identifying the Y-position of the FIB beam relative to the inspection volume. It contains two or more recognizable edges. A “recognizable edge” refers to a transition between two materials that release different numbers and / or different types of secondary particles when milling with the FIB beam. The different numbers and / or different types of secondary particles lead to a different contrast of the two materials in FIB imaging. The two materials can, for example, have a differing FIB mill yield or secondary particle yield. The mill yield refers to the average number of secondary atoms or ions released per incident ion and determines the material removal rate from the inspection volume. The secondary particle yield refers to the average number of any particle species released per incident ion, e.g. atoms, ions or electrons and with a suitable detector any of these species can be detected and counted per time. The two different materials at a recognizable edge produce different numbers of released secondary particles during milling. For example, a recognizable edge exists between the marker structure and air. The marker structure can also be embedded into a different material that releases a different amount and / or different types of secondary particles.

[0027] According to an example, the recognizable edges of the marker structure allow a unique identification of the corresponding Y-position from a cross-section profile of the marker structure along X-direction. For example, the recognizable edges of the marker structure can be configured such that their distance in X-direction strictly monotonously decreases or strictly monotonously increases. In this way, the identification of the Y-position is simplified as no approximate Y-position is required to identify the corresponding Y-position from a cross-section profile of the marker structure.

[0028] In an example, the recognizable edges of the marker structure locally allow a unique identification of the corresponding Y-position from a cross-section profile of the marker structure along X-direction. Since the Y-position can only be locally uniquely identified, the marker structure can be designed more freely, thereby simplifying the process.

[0029] In an example, the marker structure contains multiple recognizable edges such that at least some of the cross-section profiles of the marker structure along X-direction contain multiple recognizable edges, and wherein the Y-position can be determined from the pattern of the recognizable edges along the cross-section profile in X-direc- tion. In this way, the Y-position can be encoded using patterns of recognizable edges that may allow for a simple and more accurate identification of the corresponding Y- position, e.g., using pattern matching or registration.

[0030] According to an aspect of the invention, the marker structure comprises at least two layers extending in the XY-plane, and wherein the recognizable edges are contained in one of the one or more upper layers. In this way, curtaining is prevented and, thus, the quality of the 3D tomographic image is improved.

[0031] According to an example, the recognizable edges are obtained from a top view image of the marker structure taken by the charged particle imaging system. In this way, the recognizable edges of the marker structure can be easily identified, e.g., by use of image processing, and they can be used, e.g., for pattern matching or registration.

[0032] In an example, the current Y-position of the focused ion beam is derived from a crosssection profile of the marker structure along X-direction by computing the distance between the recognizable edges of the marker structure along X-direction. In this way, the identification of the current Y-position is simplified, and the current Y-position can be accurately determined.

[0033] In an example, a coarse Y-position of the focused ion beam is used to derive the current Y-position of the focused ion beam from a cross-section profile of the marker structure along X-direction. By use of the coarse Y-position, a unique identification of the Y-position is possible for a larger class of marker structures. Thus, the marker structures can be designed more freely or with respect to other optimization criteria.

[0034] In an example, the current Y-position of the focused ion beam is derived from a crosssection profile of the marker structure along X-direction by registering locations of recognizable edges in the cross-section profile with edge patterns of the marker structure. In this way, pattern matching or registration techniques can be used that allow for an accurate identification of the Y-position.

[0035] In an example, locations of recognizable edges of the marker structure in the crosssection profile along X-direction are derived from flank points on the flanks of the signal of released secondary particles when milling along the marker structure. In this way, the accuracy of the locations of the recognizable edges of the marker structure is improved.

[0036] According to an aspect of the invention, the method further comprises adjusting the movement of the Y-position of the focused ion beam such that amplitudes of subsequently obtained cross-section profiles of the marker structure along X-direction are approximately equal. In this way, the FIB spot movement in Y-direction can be corrected to ensure matching of the milling progress and the FIB spot movement in Y- direction, thereby, e.g., preventing wafer sample drift induced FIB beam detachment or milling in too fast.

[0037] In an example, the marker structure is of a material with a lower mill yield than the wafer material. The mill yield refers to the average number of atoms removed from the material per incident ion. A low mill yield means that many more incident ions are required to remove the same number of atoms from the material than for a material with a high mill yield. A material with lower mill yield than the wafer material has the advantage that the shape of the milled cross-section surface can be controlled more accurately.

[0038] In an example, the marker structure is deposited next to the inspection volume in X- direction. In this way, curtaining is prevented and, thus, the quality and accuracy of the 3D tomography image is improved.

[0039] In an example, the Y-movement of the focused ion beam is controlled such that the distance between consecutive cross-section surfaces is approximately equal. Due to the homogeneous slices, the quality and accuracy of the 3D tomography image is improved.

[0040] The methods according to the invention can be used for process optimization and / or process control in manufacturing an integrated semiconductor circuit.

[0041] A dual beam system for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer according to a second embodiment of the invention comprises: a charged particle imaging system and a focused ion beam system comprising means for operating a focused ion beam, the charged particle imaging system and the focused ion beam system being configured for performing a slice- and image method that alternatingly exposes a cross-section surface in the inspection volume by milling into the inspection volume and a marker structure deposited on the wafer surface extending in an XY-plane with the focused ion beam, and images the cross-section surface with the charged particle imaging system for obtaining a cross-section image slices of the inspection volume; a detector configured for obtaining cross-section profiles of the marker structure by measuring numbers and / or types of secondary particles released during milling along the marker structure in X-direction; and a control unit in communication with the charged particle imaging system and the focused ion beam system, the control unit being configured for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer according to a method of the first embodiment of the invention described above.

[0042] The invention described by examples and embodiments is not limited to the embodiments and examples but can be implemented by those skilled in the art by various combinations or modifications thereof.

[0043] Brief Description of the Drawings

[0044] Fig. 1 shows an illustration of a wafer inspection or metrology system for 3D volume inspection with a dual beam system;

[0045] Fig. 2 illustrates the slice-and image method of a volume inspection in a wafer;

[0046] Fig. 3 illustrates a tracking pad containing edges deposited on the surface of a wafer for controlling the Y-position of the focused ion beam using a SEM image of the cross-section of the tracking pad according to the state-of-the art;

[0047] Fig. 4 illustrates a cross-section of a wafer including a tracking pad containing edges;

[0048] Fig. 5a), b) illustrate two sections of cross-section images of a tracking pad corresponding to phase A and phase B in Fig. 3;

[0049] Fig. 6 illustrates pattern matching problems due to random effects such as shape changes of the marker profile over Y;

[0050] Fig. 7 shows a flowchart of a method for controlling a slice thickness when generating a 3D tomographic image of an inspection volume in a semiconductor wafer; Fig. 8a) to f) shows top views of different examples of marker structures 66 with recognizable edges 68 that can be used to identify a corresponding Y- position from a cross-section profile along X-direction;

[0051] Fig. 9 illustrates FIB milling of a cross-section surface with a marker structure deposited on top of the wafer surface;

[0052] Fig. 10a), b) illustrate the generation of a cross-section profile as a signal indicating the number of secondary particles of a specified type released during milling along the cross-section surface in X-direction;

[0053] Fig. 11 illustrates cross-section profiles of a marker structure along X-direction of consecutive cross-section surfaces during milling;

[0054] Fig. 12 shows signals of varying amplitude for different distances in Y direction between FIB spot and milled surface; and

[0055] Fig. 13 shows signals of varying amplitude, when the distance in Y direction between FIB spot and milled surface differ.

[0056] Detailed Description

[0057] In the following, advantageous exemplary embodiments of the invention are described and schematically shown in the figures. Throughout the figures and the description, same reference numbers are used to describe same features or components. The coordinate system is selected that the wafer surface 55 coincides with the XY-plane.

[0058] Fig. 1 illustrates a dual beam system 10 for obtaining 3D tomographic images of wafers 12. The dual beam system 10 is configured for a slice and imaging method. For a wafer 12, several inspection volumes 14, 14’ are defined in a location map or inspection list generated from an inspection tool or from design information. The wafer 12 is placed on a wafer support table 16. The wafer support table 16 is mounted on a wafer stage 18 with actuators and position control. Actuators and means for precision control for a wafer stage such as Laser interferometers are known in the art. A control unit 20 configured to control the wafer stage 18 and to adjust an inspection volume 14, 14’ of the wafer 12 at the intersection point 22 of the dual beam system 10.

[0059] The dual beam system 10 is comprising a FIB system 24 with a FIB optical axis 26 and a charged particle beam (CPB) imaging system 28 with optical axis 30. At the intersection point 22 of both optical axes of FIB and CPB imaging system, the wafer surface is arranged at a slant angle GF to the FIB optical axis 26. FIB optical axis 26 and CPB optical axis 30 include an angle GFE, and the CPB optical axis 30 forms an angle GE with the normal of the wafer surface 32. In the coordinate system of Fig. 1 , the normal of the wafer surface 32 corresponds to the Z-axis. The focused ion beam (FIB) 34 is generated by the FIB system 24 and is impinging under angle GF on the surface 32 of the wafer 12. Slanted cross-section surfaces are milled into the wafer by focused ion beam 34 milling at the inspection site 14 under approximately the slant angle GF.

[0060] In the example of Fig. 1 , the slant angle GF is approximately 30°. The actual slant angle of the slanted cross-section surface can deviate from the slant angle GF by up to 1° to 4° due to the beam divergency of the focused ion beam 34, for example a Gallium-lon beam. With the charged particle beam imaging system 28, e.g., a scanning electron microscope (SEM), inclined under angle GE to the wafer normal, images of the milled surfaces are acquired. In the example of Fig. 1 , the angle GE is about 15°. However, other arrangements are possible as well, for example with GE = GF, such that the CPB imaging system optical axis 30 is perpendicular to the FIB optical axis 26, or GE = 0°, such that the CPB imaging system axis 28 is perpendicular to the wafer surface 32.

[0061] During imaging, a beam of charged particles 38 is scanned by a scanning unit of the charged particle beam imaging system 28 along a scan path over a cross-section surface of the wafer at inspection volume 14, and secondary particles as well as scattered particles are generated. Particle detector 40 collects at least some of the secondary particles and scattered particles and communicates the particle count or particle types using a control unit 42. Other detectors for other kinds of interaction products may be present as well. Control unit 42 is in control of the charged particle beam imaging system 28, of the FIB system 24 and is connected to a control unit 20 to control the position of the wafer 12 mounted on the wafer support table 16 via the wafer stage 18. Control unit 42 communicates with operation control unit 46, which triggers placement and alignment, for example, of inspection volume 14 of the wafer 12 at the intersection point 22 via wafer stage movement and triggers repeatedly operations of FIB milling, image acquisition and stage movements.

[0062] Each new intersection surface is milled by the FIB beam 34, and imaged by the charged particle beam 38, which is, for example, a scanning electron beam or a He- lium-lon beam of a Helium ion microscope (HIM).

[0063] Fig. 2 illustrates the slice and imaging method at the example of a 3D-memory stack. Fig. 2 illustrates the situation, when the surface is the new cross-section surface 48 which was milled last by FIB 34. The cross-section surface 48 is scanned for example by charged particle beam 38, which is, in the example of Fig. 2, arranged at normal incidence to the wafer surface 32, and a high-resolution cross-section image slice is generated. The cross-section surfaces 48 are subsequently milled with a FIB beam 34 at an angle GF of approximately 30° to the wafer surface 32, but other angles GF, for example between GF = 20° and GF = 60°, are possible as well.

[0064] The cross-section image slice comprises first cross-section image features, formed by intersections with high aspect ratio (HAR) structures or vias (for example first cross-section image features of HAR-structures 50, 50’, 50”) and second cross-section image features formed by intersections with layers 52, which comprise, for example, SiO2, SiN- or Tungsten lines. Some of the lines are also called “word-lines”. The maximum number M of layers is typically more than 50, for example more than 100 or even more than 200. The HAR-structures 50 and layers 52 extend throughout most of the volume in the wafer 12 but may comprise gaps. The HAR structures 50 typically have diameters below 150nm, for example about 120nm, 100nm, 80nm, or for example 40nm. The cross-section image slices contain, therefore, first cross-section image features as intersections or cross-sections of the HAR structure footprints at different depth (Z) at the respective XY-location. In case of vertical memory HAR structures of a cylindrical shape, the obtained first cross-section image features are circular or elliptical structures at various depths determined by the locations of the structures on the sloped cross-section surface 48. The memory stack extends in the Z-direction perpendicular to the wafer surface 32. The slice thickness 51 refers to the distance in Y-direction between consecutive cross-section surfaces 48. A homogeneous slice thickness is important to obtain accurate 3D tomographic images. The slice thickness 51 between two adjacent cross-section surfaces 48 is adjusted to values typically in the order of a few nm, for example 30nm, 20nm, 10nm, 5nm, 4nm, 2nm or even less. Once a layer of material of predetermined thickness d is removed with FIB, a next cross-section surface 48 is exposed and accessible for imaging with the charged particle imaging beam 38.

[0065] In order to control the slice thickness, i.e., the distance between two consecutive cross-section surfaces, known methods deposit a tracking pad 54 on the wafer surface. The tracking pad 54 can, for example, consist of two layers structured as illustrated in Fig. 3. The lower layer contains line structures 58, 59, 60 in mill direction 56 that are buried in the layer and encode a Y-position.

[0066] Fig. 4 shows a cross-section surface 48 of a wafer with a tracking pad 54 as shown in Fig. 3 deposited on the wafer surface 32. The line structures 58, 59, 60 in the lower layer of the tracking pad are milled away during the FIB-SEM process. Thus, at the beginning of the process phase A may be visible in the cross-section image, whereas later on in the process phase B may be visible in the cross-section image.

[0067] Fig. 5a) illustrates a section of a cross-section image containing phase A of the tracking pad containing the line structures 58, 59, 60 in Fig. 3, whereas Fig. 5b) illustrates a section of a cross-section image containing phase B of the tracking pad containing the line structures 58, 59, 60 in Fig. 3. The line structures 58, 59, 60 are configured such that the Y-position can be derived from their position. In particular, the milling depth information, i.e., the Y-position, is encoded into the distance of the line structures 58 and 60. With mill progress, the distance between the line structures 58, 60 decreases from dAin phase A to dBin phase B. From the known angle of the line structures in Fig. 3 the mill depth, i.e., the Y-position, can be calculated.

[0068] As only the CPB system can view the cross-section phases A and B and not the FIB, the distances dAand dBare measured in consecutive cross-section images acquired by the CPB system, e.g., by the SEM of the FIB-SEM system. In the standard procedure of a slice and imaging method, FIB milling and SEM imaging is alternated, since there is interference between these signals when FIB milling and SEM imaging is carried out simultaneously. The implication is that a location of the FIB in mill direction (Y-direction) can only be estimated from the previously acquired SEM image, and a prediction of the mill progress can only be obtained from a model using the decrease of the distance of the line structures over time. Then the FIB milling is run for the predicted amount of time to achieve, e.g., a 2 nm slice. But the true slice thickness can only be verified after milling from the following SEM image. Thus, the slice will only approximately have the desired thickness.

[0069] The following problems are inherent to this approach:

[0070] First, pattern matching is required to identify the locations of the line structures 58, 60 in phase A and phase B. However, the contrast of the cross-section images is determined by the landing energy of the SEM, and this is adjusted for best semiconductor device imaging. Therefore, the contrast at the line structure locations in the crosssection images is usually suboptimal. In addition, a delicate and elaborate wafer sample preparation to deposit the tracking pad with the buried line structures on the wafer surface can lead to deformations of the line structures in the cross-section images, as illustrated in Fig. 6 for the line structure 60 in phase A and B. Due to the suboptimal contrast and the deformations pattern matching methods easily fail. Therefore, the distance between the line structures 58, 60 and, thus, the Y-position is not correctly estimated or cannot be estimated at all.

[0071] Second, SEM imaging alternating with FIB milling requires a model for the milling progress that predicts the mill time to cut a slice with target thickness. But random effects such as external shocks or wafer sample charging induced FIB beam detachments or pull ins cannot be predicted and lead to inhomogeneous slicing, in particular for very thin slices in the nm range.

[0072] Therefore, the requirement to mill slices in the nm range very homogeneously is often infringed by this technique. For this reason, a method for controlling the slice thickness during milling, if possible in situ, is required.

[0073] Fig. 7 illustrates a flow chart of a method for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer, the method comprising: depositing a marker structure on the wafer surface extending in an XY-plane, wherein the marker structure comprises recognizable edges allowing an identification of the corresponding Y-position from a cross-section profile of the marker structure along X-direction in a step M1 ; alternatingly 64 carrying out at least the following steps M2 and M3: exposing a cross-section surface in the inspection volume by milling into the inspection volume and the marker structure with a focused ion beam in a focused ion beam system, whose optical axis is orthogonal to the X-direction at a fixed angle with respect to the Y-direction, wherein a current Y-position of the focused ion beam relative to the inspection volume is repeatedly derived from a crosssection profile of the marker structure along X-direction, wherein the cross-section profile of the marker structure is determined by measuring numbers and / or types of secondary particles released when milling along the marker structure, and wherein the milling is stopped when the current Y-position reaches a desired Y-position in the inspection volume in a step M2; and imaging the cross-section surface with the charged particle imaging system to obtain a cross-section image slice of the 3D tomographic image of the inspection volume in a step M3.

[0074] Instead of measuring edge distances in the tracking pad based on a SEM image to verify a Y-position of the FIB beam after milling, the Y-position of the FIB beam is measured in situ with the FIB while the FIB is milling. In this way, the current Y-position is regularly checked and milling can be stopped as soon as the desired slice thickness is reached. To this end, a marker structure 66 of known shape and of a material that generates a high contrast to the wafer sample underneath is deposited on top of the wafer surface 32, e.g., of Carbon or Tungsten. The recognizable edges of the marker structure can be measured by monitoring numbers and / or types of secondary particles released during milling. Secondary particles comprise electrons and ions of different types such as Silicon, Nitrogen, Germanium or Oxygen. The type of released secondary particles can be specified (e.g., by a user) with respect to the resulting contrast of the cross-section profiles of the marker structure in FIB imaging. Different numbers of secondary particles can be determined using secondary particle detectors. Different types of ions can, for example, be detected using a detector than can distinguish between ions of different types, or by using a separate detector for each ion type.

[0075] A 3D tomographic image of an inspection volume in a semiconductor wafer refers to a volumetric image of three dimensions of the inspection volume. The 3D tomographic image can, for example, comprise a sequence of 2D cross-section images called slices. These 2D cross-section images are obtained from cross-section surfaces with a homogeneous distance between them, i.e, with a homogeneous slice thickness. Alternatively, the 3D tomographic image can comprise a regular grid of voxels. A voxel value indicates intensity and, thus, for example, different materials in the inspection volume. The marker structure 66 with the recognizable edges can be configured in various ways to allow for an identification of a corresponding Y-position from a cross-section profile along X-direction. The recognizable edges 68 can be created by embedding the marker structure into a material that releases different numbers and / or a different types of secondary particles when milling with the FIB beam. Fig. 8a) to f) shows top views of different examples of recognizable edges 68 of marker structures 66 that can be used to identify a corresponding Y-position from a cross-section profile along X- direction.

[0076] In Fig. 8a) and b), the marker structure 66 is configured such that the recognizable edges of the marker structure allow a unique identification of the corresponding Y- position from a cross-section profile of the marker structure along X-direction. To this end, the recognizable edges of the marker structure can, for example, be configured such that their distance in X-direction strictly monotonously decreases or strictly monotonously increases, e.g., in case of a triangular marker structure as shown in Fig. 8a) or a trapezoidal marker structure as shown in Fig. 8b).

[0077] In Fig. 8c) to e), the marker structure 66 is configured such that the recognizable edges of the marker structure locally allow a unique identification of the corresponding Y-position from a cross-section profile of the marker structure along X-direction. An example is the circular marker structure in Fig. 8c). The recognizable edges of the marker structure can also be configured such that the derivative of their distance in X-direction is non-zero as shown in Fig. 8d). This ensures that the distance between the recognizable edges constantly changes and locally allows a unique identification of the Y-position. Another example are repetitive edge configurations in Y-direction as shown in Fig. 8e). In these cases, if a coarse Y-position is known, the correct Y- position can be determined from the recognizable edges of the cross-section profile along X-direction, e.g., from the edge distance.

[0078] In Fig. 8f) the marker structure 66 contains multiple recognizable edges such that at least some of the cross-section profiles of the marker structure along X-direction contain multiple recognizable edges 68, and such that the Y-position can be, at least locally, determined from the pattern of the recognizable edges 68 along the crosssection profile in X-direction. The number of recognizable edges in the cross-section profile along X-direction can, for example, indicate the Y-position as shown in Fig. 8f). Various other patterns can be used as well. The patterns can be designed to allow for a unique identification of the corresponding Y-position from the pattern of recognizable edges along the cross-section profile in X-direction. Alternatively, the patterns can be designed to allow only locally for a unique identification of the corresponding Y- position from the pattern of recognizable edges along the cross-section profile in X- direction. This is, for example, the case for repetitive patterns in Y-direction.

[0079] In order to derive the corresponding Y-position from a cross-section profile of the marker structure along X-direction, the location of the recognizable edges in the marker structure must be known. To this end, the recognizable edges can be obtained from a top view image of the marker structure taken by the charged particle imaging system, e.g., by the SEM. This top view image can be taken before the first milling operation. Alternatively, it can be loaded from a database. Image processing methods can be used to extract the recognizable edges from the top view image, e.g., edge detection methods, line fitting methods, shape extraction methods such as the Hough transform, pattern matching or machine learning methods.

[0080] In an example, the marker structure comprises at least two layers extending in the XY-plane, and the recognizable edges are contained in one of the one or more upper layers. For example, the base layer can be rectangular, for example of the same shape and size as the wafer surface 32. On top of the base layer, a layer with recognizable edges configured such that the Y-position can be derived from cross-section profiles in X-direction can be arranged. The information on the Y-position can then be derived from the upper layer, while the lower layer prevents curtaining. Curtaining refers to an artefact of vertical lines in SEM images of cross-section surfaces, which is due to a wafer sample having different materials with variable milling rates.

[0081] In an example, the marker structure 66 is deposited next to the inspection volume 14 in X-direction. Next to means that the X-coordinates of the inspection volume 14 and the X-coordinates of the marker structure 66 do not overlap. If the marker structure 66 only contains a single layer curtaining can still occur in this case, but it does not interfere with the inspection volume 14.

[0082] According to an aspect of the invention, the marker structure is of a material with a lower mill yield than the wafer material. A low mill yield means that many more incident ions are needed to remove the same number of atoms from the wafer sample than for a material with a high mill yield. Since the marker structure is, thus, harder than the wafer material, the shape of the cross-section surface milled into the wafer will essentially be tangential to the phase of the marker structure. This is, in particular, the case for shallow wafer samples such as gate layers. The mill yield of a material refers to the average number of atoms removed from the material per incident ion.

[0083] The current Y-position of the FIB beam can be derived from a cross-section profile of the marker structure along X-direction in different ways.

[0084] In an example, the current Y-position of the focused ion beam is derived from a crosssection profile of the marker structure along X-direction by computing the distance d between the recognizable edges of the marker structure along X-direction as illustrated in Fig. 8a) to e).

[0085] In case a unique Y-position of the focused ion beam cannot be derived from a crosssection profile of the marker structure 66 along X-direction, a coarse Y-position of the focused ion beam can be sufficient to identify the corresponding Y-position from a cross-section profile of the marker structure 66 along X-direction, since, locally, the Y-position can still be uniquely determinable.

[0086] In another example, the current Y-position of the focused ion beam is derived from a cross-section profile of the marker structure along X-direction by registering locations of recognizable edges of the cross-section profile with edge patterns of the marker structure. The edge patterns of the marker structure can, for example, be obtained from a top view image of the marker structure and using image processing methods such as edge detectors. Locations of recognizable edges can be identified from the cross-section profile. The patterns of these locations can then be registered with the edge patterns of the marker structure. To this end, pattern matching methods can be used that match the location pattern of the cross-section profile to the edge pattern derived from the top view image. For example, the location pattern can be moved over the edge pattern in Y-direction and the maximum correlation can be found. Alternatively, registration methods can be used.

[0087] Fig. 9 illustrates FIB 34 milling of a cross-section surface 48 with a marker structure 66 deposited on top of the wafer surface 32. The marker structure 66 is triangular such that the Y-position of the FIB beam can be uniquely determined from the distance of the recognizable edges 68 of the marker structure 66.

[0088] The cross-section profile of the marker structure is determined by measuring numbers and / or types of secondary particles released when milling along the marker structure as illustrated in Fig. 10a) and b). In Fig. 10a), during milling, the FIB spot 70 is swiped along X-direction while slowly pulled in Y-direction as illustrated by the FIB trajectory 72. The FIB beam releases secondary particles in the form of secondary electrons (SE) and secondary ions (SI) over cross-section profiles along X-direction for each X- position. Special sensors can be used to monitor the SE or SI signal, e.g., SESI or InLens sensors.

[0089] Fig. 10b) shows a generated cross-section profile in the form of a SE or SI signal 76 indicating the number of SE or SI particles of a specified type released during milling along the cross-section surface 48 in X-direction as shown in Fig. 10a). The amplitude of the signal indicates the difference of the number of SE or SI particles at the transition of the different materials of the marker structure. The width and steepness of the flanks 78 is determined by the effective spot size 74 that is defined by the overlap of the spot with the edge of the milled trench or the marker structure 66 . Locations of recognizable edges 68 of the marker structure 66 in the cross-section profile along X- direction are derived from flank points 80, 84 on the flanks 78 of the signal of released secondary particles when milling along the marker structure 66. Such flank points include, for example, a 50% flank point 84 that indicates 50% of the amplitude of the signal, a 25% flank point that indicates 25% of the amplitude of the signal, a 75% flank point that indicates 75% of the amplitude of the signal, a 90% flank point that indicates 90% of the amplitude of the signal or any other X% flank point that indicates X% of the amplitude of the signal. Such flank points also include infliction points 80 on the flanks 78 of the signal 76. Infliction points 80 can, for example, be obtained by finding zero crossings of the second derivative of the signal 76. Alternatively, filters can be designed for finding such flank points 80, 84. The distance d between these flank points 80, 84 can be used as an estimate of the distance of the recognizable edges 68.

[0090] Even though the spot size broadens the flanks 78 of the signal 76, the evolution of the location of corresponding flank points can be observed in situ and the milling can be stopped as soon as the desired slice thickness is achieved. In this way, the slice thickness can be controlled more accurately compared to blindly milling and verifying the slice thickness only afterwards.

[0091] Since the width and steepness of the flanks 78 depends on the effective spot size 74, it is beneficial to control the effective spot size 74 in order to obtain comparable flank points 80, 84 on the flanks 78 of the signal 76. The effective spot size also determines the amplitude of the signal. To this end, the method can further comprise adjusting the movement of the Y-position of the focused ion beam such that amplitudes 86 of subsequently obtained cross-section profiles 82 of the marker structure 66 along X- direction are approximately equal. Approximately equal refers to a deviation below 50%, preferably below 25%, more preferably below 10%, most preferably below 5%.

[0092] Fig. 11 illustrates cross-section profiles 82 of a marker structure 66 along X-direction of consecutive cross-section surfaces 48 during milling. The first signal 88 is obtained from the first pass 92 of the FIB beam when milling the first cross-section surface 48, the second signal 90 is obtained from the second pass 94 of the FIB beam when milling the second cross-section surface 48’. The Y-position of the FIB beam is controlled such that the amplitudes 86 of the generated first and second signals 88, 90 are approximately equal. The amplitudes 86 of the first and second signals 88, 90 are approximately equal, if the distance v1 and v2 between the spot center and the milled trench surface remains the same, or, equivalently, if the effective spot size 74, 74’, i.e. , the area of the spot of the FIB beam above the wafer surface 32, remains the same. In this way, it is ensured that the FIB spot movement in Y direction matches the milling progress and that the flanks 78 of the signal and, thus, the flank points 80, 84 are comparable between different cross-section surfaces 48, 48’.

[0093] In case of, for example, wafer sample drift 96 as illustrated in Fig. 12 the FIB spot movement in Y-direction does not match the milling progress. The distance v1 and v2 between the spot center and the wafer surface 32 differ as well as the effective spot sizes 74, 74’ of the spot when milling along the cross-sections 48, 48’ in X-direction. Consequently, the amplitudes 86 of the signals are not equal. Instead, the amplitude of the first signal 88 is lower than the amplitude of the second signal 90. By comparing the amplitudes of the signals when milling along consecutive cross-section profiles 82 the FIB spot movement in Y-direction can be corrected until the amplitudes are approximately equal. In case, for example, the FIB spot moves faster than the milling progress, the amplitude of the second signal 90 is higher than the amplitude of the first signal 88, as illustrated in Fig. 13. Here, again the FIB spot movement in Y-direction can be corrected to ensure matching of the milling progress and the FIB spot movement in Y- direction.

[0094] According to an aspect of the invention, the Y-movement of the focused ion beam is controlled such that the distance between consecutive cross-section surfaces is approximately equal, for example, 5 nm, 3 nm or 2nm, 1nm, 0.5nm. In this way, very thin slices can be homogeneously milled leading to a highly accurate 3D tomographic image of the inspection volume.

[0095] The methods described above can be used for process optimization and / or process control in manufacturing an integrated semiconductor circuit. To this end, 3D tomographic images of inspection volumes are generated using the methods above. From the highly accurate 3D tomographic images, the quality of the wafers can be determined, e.g., by detecting defects or by measuring features in the wafer. The measurements can be compared to specifications. Based on detected defects or deviations of measurements from their specifications, the quality of the wafers can be judged and wafers can, for example, be discarded. Furthermore, conclusions can be drawn with respect to the quality of the wafer production process. For example, photolithography masks can be repaired, or wafer production process steps can be optimized.

[0096] A dual beam system for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer, comprises: a focused ion beam system and a charged particle imaging system configured to perform a slice- and image method for alternatingly exposing a cross-section surface in the inspection volume by milling into the inspection volume and a marker structure deposited on the wafer surface extending in an XY-plane with the focused ion beam system, and imaging the cross-section surface with the charged particle imaging system for obtaining a cross-section image slices of the inspection volume; a detector configured for obtaining cross-section profiles of the marker structure by measuring numbers and / or types of secondary particles released during milling along the marker structure in X- direction; and a control unit in communication with the charged particle beam system and the focused ion beam system, the control unit being configured for controlling the Y-position of the focused ion beam according to any of the methods described above. The methods disclosed herein can, for example, be used during research and development of wafers or during high volume manufacturing of wafers, or for process window qualification or enhancement.

[0097] Reference throughout this specification to “an embodiment” or “an example” or “an aspect” means that a particular feature, structure or characteristic described in connection with the embodiment, example or aspect is included in at least one embodiment, example or aspect. Thus, appearances of the phrases “according to an embodiment”, “according to an example” or “according to an aspect” in various places throughout this specification are not necessarily all referring to the same embodiment, example or aspect, but may. Furthermore, the particular features or characteristics may be combined in any suitable manner, as would be apparent to one of ordinary skill in the art from this disclosure, in one or more embodiments.

[0098] Furthermore, while some embodiments, examples or aspects described herein include some but not other features included in other embodiments, examples or aspects combinations of features of different embodiments, examples or aspects are meant to be within the scope of the claims, and form different embodiments, as would be understood by those skilled in the art.

[0099] Embodiments, examples of aspects of the invention may be described by the following clauses:

[0100] 1. A method 62 for controlling a slice thickness 51 when obtaining a 3D tomographic image of an inspection volume 14, 14’ in a semiconductor wafer 12, the method comprising:

[0101] Depositing a marker structure 66 on the wafer surface 32 extending in an XY- plane, wherein the marker structure 66 comprises recognizable edges 68 allowing an identification of the corresponding Y-position from a cross-section profile 82 of the marker structure 66 along X-direction;

[0102] - Alternatingly carrying out at least the following steps: o Exposing a cross-section surface 48, 48’ in the inspection volume 14, 14’ by milling into the inspection volume 14, 14’ and the marker structure 66 with a focused ion beam 24 in a focused ion beam system34, whose optical axis 26 is orthogonal to the X-direction at a fixed angle with respect to the Y-direction, wherein a current Y-position of the focused ion beam 34 relative to the inspection volume 14, 14’ is repeatedly derived from a cross-section profile 82 of the marker structure 66 along X-direction, wherein the cross-section profile 82 of the marker structure 66 is determined by measuring numbers and / or types of secondary particles released when milling along the marker structure 66, and wherein the milling is stopped when the current Y-position reaches a desired Y-position; and o Imaging the cross-section surface 48, 48’ with the charged particle imaging system 28 to obtain a cross-section image slice of the 3D tomographic image of the inspection volume 14, 14’. The method of clause 1 , wherein the recognizable edges 68 of the marker structure 66 allow a unique identification of the corresponding Y-position from a crosssection profile 82 of the marker structure 66 along X-direction. The method of any one of the preceding clauses, wherein the recognizable edges 68 of the marker structure 66 are configured such that their distance in X-direction strictly monotonously decreases or strictly monotonously increases. The method of any one of the preceding clauses, wherein the recognizable edges 68 of the marker structure 66 locally allow a unique identification of the corresponding Y-position from a cross-section profile 82 of the marker structure 66 along X-direction. The method of any one of the preceding clauses, wherein the marker structure 66 contains multiple recognizable edges 68 such that at least some of the crosssection profiles 82 of the marker structure 66 along X-direction contain multiple recognizable edges 68, and wherein the Y-position can be determined from the pattern of the recognizable edges 68 along the cross-section profile 82 in X-di- rection. The method of any one of the preceding clauses, wherein the marker structure 66 comprises at least two layers extending in the XY-plane, and wherein the recognizable edges 68 are contained in one of the one or more upper layers. 7. The method of any one of the preceding clauses, wherein the recognizable edges 68 are obtained from a top view image of the marker structure 66 taken by the charged particle imaging system 28.

[0103] 8. The method of clause 1 , wherein the current Y-position of the focused ion beam 34 is derived from a cross-section profile 82 of the marker structure 66 along X- direction by computing the distance between the recognizable edges 68 of the marker structure 66 along X-direction.

[0104] 9. The method of any one of the preceding clauses, wherein a coarse Y-position of the focused ion beam 34 is used to derive the current Y-position of the focused ion beam 34 from a cross-section profile 82 of the marker structure 66 along X- direction.

[0105] 10. The method of any one of the preceding clauses, wherein the current Y-position of the focused ion beam 34 is derived from a cross-section profile 82 of the marker structure 66 along X-direction by registering locations of recognizable edges 68 in the cross-section profile 82 with edge patterns of the marker structure 66.

[0106] 11 . The method of any one of the preceding clauses, wherein locations of recognizable edges 68 of the marker structure 66 in the cross-section profile 82 along X- direction are derived from flank points 80, 84 on the flanks 78 of the signal 76 of released secondary particles when milling along the marker structure 66.

[0107] 12. The method of any one of the preceding clauses, further comprising adjusting the movement of Y-position of the focused ion beam 34 such that amplitudes 86 of subsequently obtained cross-section profiles 82 of the marker structure 66 along X-direction are approximately equal.

[0108] 13. The method of any one of the preceding clauses, wherein the marker structure 66 is of a material with a lower mill yield than the wafer material.

[0109] 14. The method of any one of the preceding clauses, wherein the marker structure

[0110] 66 is deposited next to the inspection volume 14, 14’ in X-direction. 15. The method of any one of the preceding clauses, wherein the Y-movement of the focused ion beam 34 is controlled such that the distance between consecutive cross-section surfaces 48, 48’ is approximately equal.

[0111] 16. Use of the method according to any one of the preceding clauses for process optimization and / or process control in manufacturing an integrated semiconductor circuit.

[0112] 17. A dual beam system 10 for controlling a slice thickness 51 when obtaining a 3D tomographic image of an inspection volume 14, 14’ in a semiconductor wafer 12, comprising: a charged particle imaging system 28 and a focused ion beam system 24 comprising means for operating a focused ion beam 34 , the charged particle imaging system 28 and the focused ion beam system 24 being configured for performing a slice- and image method that alternatingly exposes a cross-section surface 48, 48’ in the inspection volume 14, 14’ by milling into the inspection volume 14, 14’ and a marker structure 66 deposited on the wafer surface 32 extending in an XY-plane with the focused ion beam 34, and images the cross-section surface 48, 48’ with the charged particle imaging system 28 for obtaining cross-section image slices of the inspection volume 14, 14’; a detector 40 configured for obtaining cross-section profiles 82 of the marker structure 66 by measuring numbers and / or types of secondary particles released during milling along the marker structure 66 in X-direction; a control unit 42 in communication with the charged particle imaging system 28 and the focused ion beam system 24, the control unit 42 being configured for controlling a slice thickness 51 when obtaining a 3D tomographic image of an inspection volume 14, 14’ in a semiconductor wafer 12 according to a method of any one of the preceding clauses.

[0113] In summary, the invention relates to a system and method for controlling a slice thickness when obtaining a 3D tomographic image of an inspection volume in a semiconductor wafer, the method comprising: depositing a marker structure 66 on the wafer surface 32, alternatingly: exposing a cross-section surface 48 in the inspection volume by milling into the inspection volume and the marker structure 66 with a focused ion beam 34 in a focused ion beam system 24, wherein a current position of the focused ion beam 34 is repeatedly derived from a cross-section profile of the marker structure 66, wherein the cross-section profile of the marker structure 66 is determined by measuring numbers and / or types of secondary particles released during milling, and wherein the milling is stopped when the current position reaches a desired position; and imaging the cross-section surface 48.

[0114] Reference number list

[0115] 10 Dual beam system

[0116] 12 Wafer

[0117] 14, 14’ Inspection volume

[0118] 16 Wafer support table

[0119] 18 Wafer stage

[0120] 20 Control unit

[0121] 22 Intersection point

[0122] 24 Focused ion beam system

[0123] 26 FIB optical axis

[0124] 28 Charged particle beam imaging system

[0125] 30 CPB optical axis

[0126] 32 Wafer surface

[0127] 34 Focused ion beam

[0128] 38 Charged particle beam

[0129] 40 Particle detector

[0130] 42 Control unit

[0131] 46 Operation control unit

[0132] 48, 48’ Cross-section surface

[0133] 50, 50’, 50” HAR structure

[0134] 51 Slice thickness

[0135] 52 Layer

[0136] 54 Tracking pad

[0137] 56 Mill direction

[0138] 58, 59, 60 Line structure

[0139] 62 Method

[0140] 64 Alternating

[0141] 66 Marker structure

[0142] 68 Recognizable edge

[0143] 70 FIB spot

[0144] 72 Trajectory

[0145] 74, 74’ Effective spot size

[0146] 76 Signal

[0147] 78 Flank

[0148] 80 Infliction point Cross-section profile 50% flank point Amplitude First signal Second signal First pass Second pass Wafer sample drift

Claims

- 29 -Claims1. A method (62) for controlling a slice thickness (51) when obtaining a 3D tomographic image of an inspection volume (14, 14’) in a semiconductor wafer (12), the method comprising:Depositing a marker structure (66) on the wafer surface (32) extending in an XY-plane, wherein the marker structure (66) comprises recognizable edges (68) allowing an identification of the corresponding Y-position from a crosssection profile (82) of the marker structure (66) along X-direction;- Alternatingly carrying out at least the following steps: o Exposing a cross-section surface (48, 48’) in the inspection volume (14, 14’) by milling into the inspection volume (14, 14’) and the marker structure (66) with a focused ion beam (24) in a focused ion beam system(34), whose optical axis (26) is orthogonal to the X-direction at a fixed angle with respect to the Y-direction, wherein a current Y-po- sition of the focused ion beam (34) relative to the inspection volume (14, 14’) is repeatedly derived from a cross-section profile (82) of the marker structure (66) along X-direction, wherein the cross-section profile (82) of the marker structure (66) is determined by measuring numbers and / or types of secondary particles released when milling along the marker structure (66), and wherein the milling is stopped when the current Y-position reaches a desired Y-position; and o Imaging the cross-section surface (48, 48’) with the charged particle imaging system (28) to obtain a cross-section image slice of the 3D tomographic image of the inspection volume (14, 14’).

2. The method of claim 1 , wherein the recognizable edges (68) of the marker structure (66) allow a unique identification of the corresponding Y-position from a cross-section profile (82) of the marker structure (66) along X-direction.

3. The method of any one of the preceding claims, wherein the recognizable edges (68) of the marker structure (66) are configured such that their distance in X- direction strictly monotonously decreases or strictly monotonously increases.- 30 -4. The method of any one of the preceding claims, wherein the recognizable edges (68) of the marker structure (66) locally allow a unique identification of the corresponding Y-position from a cross-section profile (82) of the marker structure (66) along X-direction.

5. The method of any one of the preceding claims, wherein the marker structure (66) contains multiple recognizable edges (68) such that at least some of the cross-section profiles (82) of the marker structure (66) along X-direction contain multiple recognizable edges (68), and wherein the Y-position can be determined from the pattern of the recognizable edges (68) along the cross-section profile (82) in X-direction.

6. The method of any one of the preceding claims, wherein the marker structure (66) comprises at least two layers extending in the XY-plane, and wherein the recognizable edges (68) are contained in one of the one or more upper layers.

7. The method of any one of the preceding claims, wherein the recognizable edges (68) are obtained from a top view image of the marker structure (66) taken by the charged particle imaging system (28).

8. The method of claim 1 , wherein the current Y-position of the focused ion beam (34) is derived from a cross-section profile (82) of the marker structure (66) along X-direction by computing the distance between the recognizable edges (68) of the marker structure (66) along X-direction.

9. The method of any one of the preceding claims, wherein a coarse Y-position of the focused ion beam (34) is used to derive the current Y-position of the focused ion beam (34) from a cross-section profile (82) of the marker structure (66) along X-direction.

10. The method of any one of the preceding claims, wherein the current Y-position of the focused ion beam (34) is derived from a cross-section profile (82) of the marker structure (66) along X-direction by registering locations of recognizable edges (68) in the cross-section profile (82) with edge patterns of the marker structure (66).11 . The method of any one of the preceding claims, wherein locations of recognizable edges (68) of the marker structure (66) in the cross-section profile (82) along X- direction are derived from flank points (80, 84) on the flanks (78) of the signal (76) of released secondary particles when milling along the marker structure (66).

12. The method of any one of the preceding claims, further comprising adjusting the movement of Y-position of the focused ion beam (34) such that amplitudes (86) of subsequently obtained cross-section profiles (82) of the marker structure (66) along X-direction are approximately equal.

13. The method of any one of the preceding claims, wherein the marker structure (66) is of a material with a lower mill yield than the wafer material.

14. The method of any one of the preceding claims, wherein the marker structure (66) is deposited next to the inspection volume (14, 14’) in X-direction.

15. The method of any one of the preceding claims, wherein the Y-movement of the focused ion beam (34) is controlled such that the distance between consecutive cross-section surfaces (48, 48’) is approximately equal.

16. Use of the method according to any one of the preceding claims for process optimization and / or process control in manufacturing an integrated semiconductor circuit.

17. A dual beam system (10) for controlling a slice thickness (51) when obtaining a 3D tomographic image of an inspection volume (14, 14’) in a semiconductor wafer (12), comprising: a charged particle imaging system (28) and a focused ion beam system (24) comprising means for operating a focused ion beam (34) , the charged particle imaging system (28) and the focused ion beam system (24) being configured for performing a slice- and image method that alternatingly exposes a cross-section surface (48, 48’) in the inspection volume (14, 14’) by milling into the inspection volume (14, 14’) and a marker structure (66) deposited on the wafer surface (32) extending in an XY-plane with the focused ion beam (34), and images the cross-section surface (48, 48’) with the charged particle imaging system (28) for obtaining cross-section image slices of the inspection volume (14, 14’);a detector (40) configured for obtaining cross-section profiles (82) of the marker structure (66) by measuring numbers and / or types of secondary particles released during milling along the marker structure (66) in X-direction; a control unit (42) in communication with the charged particle imaging system (28) and the focused ion beam system (24), the control unit (42) being configured for controlling a slice thickness (51) when obtaining a 3D tomographic image of an inspection volume (14, 14’) in a semiconductor wafer (12) according to a method of any one of the preceding claims.

Images