Method for cross-sectional imaging of inspection volume in wafer

Abstract

To provide a slice imaging method applicable to 3D volume imaging of an inspection volume in a wafer without removing a sample, provide a slice imaging method applicable to 3D volume imaging of a volume in a wafer with high resolution in depth without removing a sample, provide a wafer defect inspection apparatus and method capable of inspecting defects in an inspection volume of a wafer without removing a sample from the wafer, provide a wafer inspection tool and method for 3D volume imaging including imaging of deep structures without destroying the wafer, provide a wafer inspection tool and method for generating cross sections through HAR structures in a single layer by 3D volume imaging without removing a sample from the wafer, provide a dual beam apparatus and method for inspecting an inspection volume with a large extension in depth inside a wafer by 3D volume imaging, and provide 3D volume inspection of semiconductor devices fabricated in a wafer with high throughput.SOLUTION: The present invention relates to a dual beam apparatus and a three-dimensional circuit pattern inspection technique by cross-sectioning an inspection volume extending significantly to a depth of more than 1 μm below the surface of a semiconductor wafer, and more particularly to a method, computer program product, and apparatus for generating 3D volume image data of an inspection volume deep inside a wafer without removing a sample from the wafer. The present invention further relates to a 3D volume image generation and cross-sectional image alignment method using a dual beam apparatus for three-dimensional circuit pattern inspection.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The present invention relates to a three-dimensional circuit pattern inspection and measurement technique using cross-sectioning of integrated circuits. More particularly, the present invention relates to a three-dimensional circuit pattern inspection technique using cross-sectioning of an inspection volume at a measurement site on a semiconductor wafer containing integrated circuits. Even more particularly, the present invention relates to a method, computer program product, and corresponding semiconductor inspection apparatus for obtaining a 3D volumetric image of the inspection volume at the measurement site on the semiconductor wafer. The method involves cutting a cross-sectional surface at an oblique angle relative to the inspection volume of the wafer, and then imaging the oblique cross-sectional surface with a charged particle imaging microscope. The method, computer program product, and apparatus can be used for quantitative metrology, defect detection, process monitoring, defect review, and inspection of integrated circuits in semiconductor wafers. [Background technology]

[0002] Semiconductor structures are among the finest man-made structures, and defects are rare. These rare defects are the expected signature of defect detection, defect probing, or quantitative metrology tools. Fabricated semiconductor structures are based on prior knowledge. They are manufactured from a series of layers parallel to the substrate. For example, in logic-type samples, metal lines run parallel in the metal layers or HAR (High Aspect Ratio) structures, and metal vias run perpendicular to the metal layers. The angles between metal lines in different layers are either 0° or 90°. On the other hand, in the case of VNAND-type structures, it is known that their cross sections are, on average, circular.

[0003] Feature sizes are shrinking in integrated circuit fabrication. Current minimum feature sizes, or critical dimensions, are below 10 nm (e.g., 7 nm or 5 nm) and are expected to fall below 3 nm in the near future. This has led to challenges in measuring pattern edge shapes and accurately determining feature dimensions or line edge roughness. Pattern edge shape or line roughness is influenced by multiple factors. In general, the line or pattern edge shape can be determined by the properties of the materials themselves, the lithography exposure, or any other associated process steps, such as etching, deposition, or implantation. The measurement resolution of charged particle systems is typically limited by the sampling raster of the dwell time per individual image point or pixel on the sample and the charged particle beam diameter. The sampling raster resolution is configurable within the imaging system and adaptable to the charged particle beam diameter on the sample. Typical raster resolution is 2 nm or less, but the raster resolution limit can be lowered if there are no physical limitations. The charged particle beam diameter is limited in size depending on the operating conditions of the charged particle beam and the lenses. Beam resolution is limited to approximately half the beam diameter. Resolutions of less than 2 nm are possible, for example even less than 1 nm.

[0004] As feature sizes of semiconductor integrated circuits decrease and the resolution requirements of charged particle imaging systems increase, the inspection and 3D analysis of semiconductor integrated circuits in wafers becomes increasingly difficult. A semiconductor wafer is 300 mm in diameter and consists of multiple parts (called dies), each containing at least one integrated circuit pattern, such as a memory chip or a processor chip. A semiconductor wafer undergoes approximately 1000 process steps, within which approximately 100 or more parallel layers (including transistor layers, intermediate layers, and interconnect layers) are formed, resulting in a three-dimensional array of memory cells in a memory device.

[0005] A common method for generating nm-scale 3D tomographic data from semiconductor samples is the so-called slice and image technique, implemented, for example, by a dual-beam instrument. In such instruments, two particle optics are arranged at an angle. The first particle optic can be a scanning electron microscope (SEM). The second particle optic can be a focused ion beam (FIB) optic using, for example, gallium (Ga) ions. Using the focused ion beam (FIB) of gallium ions, layers are removed from the edge of the semiconductor sample, one by one, and each cross section is imaged by the scanning electron microscope (SEM). The two particle optics can be oriented perpendicularly or at an angle of 45°–90°. Figure 1 shows a schematic diagram of the slice and image technique. Using an FIB optical column 50, a thin layer is removed from a cross section through a semiconductor sample 10 by scanning a focused ion beam 51 in the z-direction in the yz-plane, revealing a new front surface 52 as the cross-section surface 52. In the next step, the SEM 40 is used to perform scan imaging of, for example, the front surface of the cross-sectional surface 52. In this example, the SEM optical axis 42 is oriented parallel to the x-direction, and the image is generated by scanning the electron beam 44 along a scan imaging line 46 in the yz-plane. Performing a raster scan of the cross-sectional surface 52 produces a cross-sectional image slice 100.1. Repeating this FIB milling and SEM imaging technique through, for example, the front cross-sectional surfaces 53 and 54 produces cross-sectional image slices 100.2 and 100.3 at a distance d. Finally, a series of 2D cross-sectional images 1000 at various depths through the sample is obtained. The distance d between two subsequent image slices can be from 1 nm to several tens of nanometers (e.g., 30 nm). From this series of 2D cross-sectional images 1000, a 3D image of the semiconductor integrated structure can be reconstructed.

[0006] FIG. 1 illustrates a slice imaging technique for an example of a block-shaped semiconductor integrated sample 10 extracted from a semiconductor wafer by known techniques. Fiducials are formed on the block sample. It is common practice to derive the lateral position of each slice as well as the distance between layers by using so-called fiducials. U.S. Pat. No. 9,633,819 discloses an alignment method based on guide structures ("fiducials") exposed on the sample. U.S. Pat. No. 7,348,556 describes alignment marks on a surface for determining three-dimensional surface roughness from a series of successive image slices.

[0007] Typical slice imaging techniques in the prior art cannot be applied to inspection volumes inside a wafer. Typical slice imaging techniques require the removal or extraction of a sample from a semiconductor wafer before the slice imaging technique can be performed to obtain a 3D volume image of the sample. Therefore, it is a challenge to provide a slice imaging technique that can be applied to generate a 3D volume image of an inspection volume in a wafer without removing the sample.

[0008] U.S. Patent No. 7,438,556 describes a method for determining line edge or surface roughness using a dual-beam FIB / SEM tool. A series of cross-sectional surfaces is generated by FIB cutting perpendicular to the sample's surface. A series of cross-sections is also generated at slice distances of approximately 10 nm or greater. A fiducial is applied to the top surface to determine the lateral position of the cross-section. A SEM is tilted relative to the normal to the sample's surface, and a sequence of cross-section images is generated. From each cross-section image, the critical dimension (CD) of the feature parallel to the wafer surface is determined. The line edge roughness of the feature is determined from the CD, and in each cross-section image, the feature edge is determined by referencing the fiducial. The edge position perpendicular to the sample surface is determined by referencing the fiducial and a known angle of the SEM. Therefore, accurate determination of the feature depth is not possible. Therefore, the proposed method is limited to 1D measurements of semiconductor features extending parallel to the sample's surface. In particular, the proposed method does not provide a means for accurately determining the depth of semiconductor features below the wafer surface.

[0009] Slicing and imaging using a single wedge cut with a small tilt angle of the FIB beam for milling has been proposed for analyzing deep, repetitive semiconductor features, such as HAR structures in memory devices. For example, U.S. Patent No. 9,466,537 describes a method for inspecting semiconductor devices having a mold layer. Milling the mold layer at a certain tilt angle creates a tilted cross-sectional surface through the mold layer. To achieve the desired tilt angle, the stage holding the sample under investigation is tilted between the milling and imaging steps. The resulting image of the milled or cross-sectional surface is analyzed to derive, for example, the center position of the semiconductor feature relative to a selected semiconductor feature that serves as a reference. It is claimed that this analysis can derive process deviations during the fabrication process. However, analysis of a single milled or cross-sectional surface using the feature of interest as a reference has limited accuracy and has proven unable to provide the information necessary for monitoring the fabrication process. Furthermore, the need to tilt the stage between milling and imaging is impractical for high-throughput inspection tasks.

[0010] U.S. Patent No. 10,026,590 discloses a similar method of inspecting a feature of interest by cutting a single cross-sectional surface at a glancing angle relative to a sample and constructing a virtual feature from various cross-sections of the feature of interest at various depths. The depth is determined by the lateral distance of the feature cross-section to the trench edge. The depth resolution of the virtual feature can be improved by additional cutting operations. Alignment fiducials perpendicular to the FIB have been proposed to monitor the cutting operation. However, it has been found that cutting at very small glancing angles is either impossible or results in incorrect cutting operations, and the derivation of virtual features from a few cross-sections has limited accuracy and cannot provide the information necessary for monitoring the fabrication process. Furthermore, determining the depth from the trench or cutting edge is inaccurate, and applying fiducials on surfaces perpendicular to the FIB beam is a difficult and time-consuming process.

[0011] As the thickness of individual layers in a semiconductor device's multilayer stack decreases, the tilt angle required to cut a cross-sectional surface relative to the wafer's inspection volume becomes increasingly smaller, e.g., less than 5° or even 3°. Such techniques are described in U.S. Pat. No. 9,941,096, BB. However, it is impossible to achieve such small tilt angles in practical setups. Cutting a cross-sectional surface at a very small angle, e.g., less than 15°, for deep semiconductor structures with a multilayer depth of approximately 5 μm or more requires the FIB to cut a very large surface extending over a large distance in the direction of the cutting beam, exceeding 40 μm or even 100 μm, toward the deeper structure. Cutting such a large surface is time-consuming, and the surface quality significantly deteriorates with increasing depth. Furthermore, large cross-sectional surfaces exceed the 10 μm to 20 μm field of view of typical high-resolution charged particle imaging devices, and imaging large cross-sectional surfaces requires image stitching. U.S. Patent No. 10,184,790 proposes a method for forming a 2D image of an oblique surface by image stitching, in which a series of SEM images are obtained by lateral shifting of the sample, and a single 2D image of the oblique surface is formed by stitching these images together. Depth is measured by referencing the edge of the trench seen in one of the SEM images. Therefore, the depth determination is not accurate. The obtained 2D information and limited accuracy are insufficient for modern requirements. Therefore, the object of the present invention is to provide a wafer inspection tool and method for generating 3D volume images, including imaging of deep structures with high quality.

[0012] U.S. Patent Application Publication No. 2009 / 296073 describes a method for analyzing semiconductor features by cutting a surface substantially parallel to the wafer surface, but it is unclear how FIB cutting can be performed with sufficient precision to create a surface that is parallel to the wafer surface or at a very small angle of less than 5° or 10° without removing the sample from the wafer.

[0013] Recent developments have required greater precision at significantly higher throughput, including accurate determination of the depth of semiconductor features below the wafer surface. Therefore, it is an object of the present invention to provide a slice imaging method applicable to generating a 3D volumetric image of a volume under inspection in a wafer without removing a sample. The depth of features in a volume under inspection in a wafer must be determined with high accuracy. Therefore, it is an object of the present invention to provide a slice imaging method applicable to generating a 3D volumetric image of a volume under inspection in a wafer with high resolution in the depth direction without removing a sample. It is a further object of the present invention to provide a wafer defect inspection apparatus and method capable of inspecting a volume under inspection in a wafer for defects without removing a sample from the wafer.

[0014] In recent years, the trend toward higher integration of semiconductor circuits has resulted in ever-higher stacks of alternating layers formed in or on silicon wafers. Current memory chips contain up to 100 or more (e.g., 92) different layers. Stacks of approximately 100 layers exceed 6 μm in thickness (height), with current and future stack heights reaching 10 μm. As stack heights increase, imaging of deep structures within the wafer inspection volume becomes increasingly difficult. Therefore, it is a challenge to provide wafer inspection tools and methods for 3D volumetric imaging, including imaging of deep structures, without destroying the wafer.

[0015] In addition to increasing depth, the thickness of each layer becomes smaller and smaller. For inspection operations, it is desirable to obtain cross-sectional images through multiple HAR structures in a single layer (e.g., word lines or insulating layers). Therefore, it is a challenge to provide wafer inspection tools and methods for generating cross-sections through HAR structures in a single layer by 3D volumetric imaging without removing a sample from the wafer.

[0016] As the depth increases, it becomes increasingly difficult to image cross-sectional surfaces of inspection volumes with large depth extensions using a charged particle imaging beam. Therefore, it is a challenge to provide a dual beam apparatus and method for inspecting inspection volumes with large depth extensions inside a wafer using 3D volume imaging.

[0017] Conventional wafer inspection for in-line testing requires very high throughput. Therefore, another object of the present invention is to provide 3D volume inspection of semiconductor devices fabricated in wafers with high throughput.

[0018] The above-mentioned problems are solved by the invention described in the examples shown in the embodiments of the present invention. Summary of the Invention

[0019] An object of the present invention is to provide a method for 3D inspection of an inspection volume in a wafer using a dual beam device without the need to extract or remove the inspection volume from the wafer, and a dual beam device configured to inspect the inspection volume in a wafer. Another object of the present invention is to provide a computer program product including program code for executing the method for 3D inspection of an inspection volume in a wafer using a dual beam device. In accordance with an embodiment, the 3D inspection of the inspection volume in a wafer is performed using a wedge cut geometry. The wedge cut geometry allows for the examination of inspection volumes extending approximately 10 μm to 15 μm laterally within the wafer and up to 10 μm or more in depth without destroying the wafer or extracting a sample volume from the wafer. The wedge cut slicing imaging method allows for the generation of a 3D representation or 3D volumetric image of the entire inspection volume of the wafer with a high lateral resolution of less than 5 nm, preferably less than 2 nm, and more preferably less than 1 nm. A 3D volumetric image for 3D inspection is provided for the entire inspection volume.

[0020] The present invention provides a dual beam apparatus and a 3D measurement method for cross-sectioning an integrated circuit and generating 3D volumetric image data without removing a sample from the wafer. In particular, a dual beam apparatus and a method are provided for high-resolution wafer inspection of an inspection volume extending to a significant depth below the wafer's surface. The present invention provides a 3D measurement method for inspecting a 3D circuit pattern in an inspection volume inside a wafer without removing a sample from the wafer. More specifically, the present invention relates to a 3D circuit pattern inspection technique for cross-sectioning an inspection volume at a measurement site on a semiconductor wafer containing an integrated circuit. Even more particularly, the present invention relates to a method, computer program product, and corresponding semiconductor inspection apparatus for obtaining a 3D volumetric image of an inspection volume at a measurement site on a semiconductor wafer without removing a sample from the wafer. In this method, a second cross-section image feature is utilized to determine the depth of a first cross-section image feature. The plurality of second cross-section image features correspond to structures in a layer in the integrated circuit, typically corresponding to structures at a known or predetermined depth. In one example, the determination of the depth of the first cross-section is a relative depth determination with respect to the plurality of second cross-section image features. After acquiring and aligning at least one cross-sectional image slice and determining the depth of the first cross-sectional image feature, the inspection volume of the measurement site on the wafer is evaluated, e.g., with respect to wafer fabrication errors. The fabrication errors are analyzed, e.g., to perform a failure analysis of the fabricated wafer. In one example, the fabrication errors are analyzed, e.g., to improve a particular fabrication process step that creates the wafer. The method, computer program product, and apparatus can be used for quantitative metrology, defect detection, process monitoring, defect review, and inspection of integrated circuits in semiconductor wafers.

[0021] According to one embodiment of the present invention, a method for wafer inspection of at least a first inspection volume in a wafer using a first dual beam device includes loading the wafer onto a wafer support table in a dual beam device including at least a FIB column and a charged particle imager, wherein a first optical axis of the FIB column forms a tilt angle GF with a surface of the wafer support table and a second optical axis of the charged particle imager forms an angle GE with a normal to the surface of the wafer support table, the first and second optical axes forming an intersection point. The method for wafer inspection further includes moving the wafer support table to align a first measurement location on the wafer with the intersection point of the dual beam device, and cutting a first cross-sectional surface in the first inspection volume at the tilt angle GF using the FIB column. The method for wafer inspection further includes generating a first cross-sectional image slice of the first cross-sectional surface using the charged particle imager. The method for wafer inspection further includes obtaining performance indicators of a plurality of first semiconductor features in the first inspection volume, and analyzing the at least one first cross-sectional image slice using preliminary information regarding the plurality of first semiconductor features. In one example, the first semiconductor feature is one of a via, a HAR structure, or a HAR channel, and the analyzing step includes extracting, by image processing, a plurality of first cross-sectional image features representing cross sections of the plurality of first semiconductor features at a tilt angle GF, the image processing including at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation. In one example, obtaining the performance index further includes calculating at least one descriptive parameter of the first semiconductor feature from at least one of the plurality of first cross-sectional image features, the at least one descriptive parameter being one of a dimension, a diameter, an angle, an area, a shape, or a volume. In one example, obtaining the performance index further includes calculating an average or statistical deviation of the at least one descriptive parameter of the plurality of first semiconductor features. In one example, the analyzing step further includes generating a depth map Z(x,y) of the first cross-sectional image slice.In one example, generating the depth map further includes determining at least two second cross-sectional image features in the first cross-sectional image slice and determining a depth map Z(x,y) from lateral positions of the at least two second cross-sectional image features. The at least two second cross-sectional image features may represent semiconductor integrated structures or features at different depths within the inspection volume. The plurality of first semiconductor features may extend in a direction perpendicular to the wafer surface, and the second cross-sectional image features may include cross sections of the semiconductor structures extending in a direction parallel to the wafer surface. In one example, the second cross-sectional image features include cross sections of at least one of an insulating line or insulating layer, a metal line or metal layer, and a semiconductor line or semiconductor layer. Obtaining the performance index may further include deriving tilt angle deviations of the plurality of first semiconductor features, which are angles of the first semiconductor features relative to an axis perpendicular to the wafer surface, from the lateral positions of the plurality of first cross-sectional image features.

[0022] In one example, the method of wafer inspection further includes deriving a 3D representation of the first plurality of semiconductor features from the first plurality of cross-sectional image features arranged at the tilt angle GF and the depth map Z(x,y).

[0023] In one example, the analyzing step further includes comparing the first cross-sectional image slice to a 2D digital image slice through the inspection volume of a reference wafer or die. The 2D digital image slice can be a cross-sectional image slice acquired in a previous measurement of the reference wafer or die and stored in memory. In one example, the 2D digital image slice is a virtual cross-sectional image slice, and the virtual cross-sectional image slice is generated from 3D volumetric image data stored in memory. In one example, the 3D volumetric image data is acquired in a previous slice image measurement of the inspection volume of the reference wafer or die and stored in memory. In one example, the previous slice image measurement is performed by a second dual beam device including at least a second FIB column and a second charged particle imaging device. In one example, the previous slice image measurement is performed by the first dual beam device.

[0024] In one example, the method for wafer inspection includes: moving the wafer support table to align a second measurement location on the wafer with an intersection point of the first dual beam device; cutting a second cross-sectional surface in a second inspection volume with the FIB column at a tilt angle GF; generating a second cross-sectional image slice of the second cross-sectional surface with a charged particle imaging device; obtaining performance metrics for the plurality of first semiconductor features in the first and second inspection volumes by analyzing the first and second cross-sectional image slices with prior information about the plurality of first semiconductor features; Further includes:

[0025] In one example, the wafer support table rotates about an axis perpendicular to the wafer support surface between cutting the first cross-sectional surface in the first inspection volume and cutting the second cross-sectional surface in the second inspection volume, and the step of deriving tilt angle deviations of the plurality of first semiconductor features may include analyzing the first and second cross-sectional image slices.

[0026] The present invention provides a wafer defect inspection apparatus configured to inspect an inspection volume in a wafer without extraction from the wafer, comprising: a focused ion beam (FIB) column configured to cut and expose a first cross-sectional surface through at least a first inspection volume in the wafer at a tilt angle GF; a charged particle imaging device configured to image at least a first cross-sectional surface to form a first cross-sectional image slice; an image processing unit having installed thereon software code configured to determine, in at least a first cross-sectional image slice, a plurality of cross-sectional image features that are cross-sections of the semiconductor structure at a tilt angle GF inside the inspection volume, and to determine depths of the plurality of cross-sectional features within the inspection volume; a defect detection unit configured to determine, from the plurality of cross-sectional image features, deviations of the semiconductor structure inside the inspection volume from a predetermined characteristic; The present invention further provides a wafer defect inspection apparatus comprising: a control unit configured to: (a) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (b) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (c) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (d) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (e) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (f) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (g) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (h ...i) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (ii) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (iii) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (iv) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (v) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor structure; (vi) process a plurality of first cross-sectional image features arranged at a tilt angle GF for detecting a first cross-sectional image of a semiconductor

[0027] According to one embodiment of the present invention, a method for inspecting an inspection volume in a wafer using a dual beam tool includes loading the wafer onto a wafer support table in a dual charged particle beam tool and moving the wafer support table to align a first measurement location on the wafer with an intersection point of optical axes of an FIB and a charged particle imager. The method further includes acquiring a series or a plurality of N cross-sectional image slices in the inspection volume, the series or a plurality of N cross-sectional image slices including at least a first cross-sectional image slice and a second cross-sectional image slice. The number N of cross-sectional image slices is at least N=10, preferably N>100 (e.g., N is approximately 300 or greater (e.g., N>1000)). The plurality of cross-sectional image slices is obtained by subsequently exposing a plurality of N cross-sectional surfaces in the inspection volume by cutting the inspection volume at approximately angle GF with the FIB column to obtain the plurality of cross-sectional image slices, and imaging each of the plurality of N cross-sectional surfaces with the charged particle imager. The plurality of cross-sectional surfaces includes at least a first and a second cross-sectional surface. The plurality of cross-sectional image slices includes at least a first and a second cross-sectional image slice. The method further includes determining at least one first cross-sectional image feature in the first and second cross-sectional image slices and determining at least one second cross-sectional image feature in the first and second cross-sectional image slices, and determining a depth of the at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the first cross-sectional image slice. In one example, the wafer does not move during the step of acquiring the series of N cross-sectional image slices through the inspection volume.

[0028] Additionally, this method of inspecting an inspection volume in a wafer optionally further includes performing a mutual lateral alignment of the first and second cross-sectional image slices with at least one common cross-sectional image feature. In one example, the common cross-sectional image feature is a first cross-sectional image feature present in the first and second cross-sectional image slices. In another example, the common cross-sectional image feature is an alignment feature provided near the inspection volume or an image segment of an alignment feature present near the inspection volume. The step of lateral mutual image alignment is optionally further improved by subtraction or numerical compensation of image distortion deviations between at least the first and second cross-sectional image slices.

[0029] According to one embodiment of a method for inspecting an inspection volume in a wafer, a depth of a first cross-sectional image feature in a first cross-sectional image slice is determined, and a 3D volumetric image of the inspection volume is derived by an algorithm employing multiple steps. The first step is cross-sectional image feature detection and classification, where the cross-sectional image features are detected and classified into first and second cross-sectional image features. The second step is generation of a depth map from the second cross-sectional image features in the multiple cross-sectional image slices. The third step is determination of the depth of each of the first cross-sectional image features based on the depth map. The fourth step is generation of a 3D volumetric image of the semiconductor structure of interest in the inspection volume. The fifth step is derivation of an integrated circuit feature or characteristic of the semiconductor structure of interest in the 3D volumetric image. The sixth step is derivation of a defect in the integrated circuit feature or characteristic of the semiconductor structure of interest in the 3D volumetric image. In one example of a method for inspecting an inspection volume in a wafer, a tilt angle GF of the FIB beam cutting the multiple cross-sectional surfaces is adjustable. As the tilt angle GF changes, the depth coverage of the inspection volume also changes. In one example method for inspection of an inspection volume in a wafer, the distance between subsequent cross-sectional surfaces is adjustable. In one example, the distance between multiple cross-sectional surfaces is adjusted to be different for at least some of the distances of the cross-sectional image surfaces. Adjusting the distance globally or locally adjusts the throughput and resolution of the 3D volumetric image of the inspection volume to suit the needs of the wafer inspection operation.

[0030] According to one embodiment of a method for inspecting an inspection volume in a wafer, determining the depth map or depth includes determining a lateral position of a second cross-sectional image feature. The depth map or depth of the first cross-sectional image feature in the first cross-sectional image slice is determined from a lateral difference between a first position of the second cross-sectional image feature in the first cross-sectional image slice and a second position of the second cross-sectional image feature in the second cross-sectional image slice. The depth determination using the second cross-sectional image feature does not utilize the first cross-sectional image feature, which may represent, for example, a HAR channel of a memory device. The first cross-sectional image feature (e.g., the entire three-dimensional HAR structure in the inspection volume) and errors during fabrication of the semiconductor structure represented by the first cross-sectional image feature are determined with high precision. In an example method for inspecting an inspection volume in a wafer, tilt or wobble relative to the wafer surface, alignment errors during fabrication of the HAR structure, or orientation of the memory stack are determined with high precision and accuracy. In one example method for inspecting an inspection volume in a wafer, HAR structures are examined and compared at multiple depths throughout the inspection volume of the wafer.

[0031] According to one embodiment of a method for inspection of an inspection volume in a wafer, positions of two second cross-sectional image features are determined in a first cross-sectional image slice, each of the second cross-sectional image features representing a semiconductor integrated structure at a predetermined depth within the inspection volume. In one example, determining the depth of the first cross-sectional image feature in the first cross-sectional image slice or a depth map of the first cross-sectional image slice includes determining lateral positions of the two second cross-sectional image features.

[0032] In one embodiment of a method for inspection of an inspection volume in a wafer, the method further includes forming at least one alignment feature. The at least one alignment feature is formed or exposed near the inspection volume. The alignment feature is configured to laterally align multiple cross-sectional image slices, including the first and second cross-sectional image slices. In one embodiment, the alignment feature is fabricated above the inspection volume and configured to determine the location of an edge formed by an intersection of the cross-sectional surface with the wafer surface. In one example, an alignment feature is provided or exposed at a particular depth or depths of the additional trench or trenches to enable alignment at various imaging depths in the inspection volume beyond the wafer surface. Additionally, the method for inspection of an inspection volume in a wafer optionally further includes lateral aligning the first and second cross-sectional image slices with at least one common cross-sectional image feature. In one example, the common cross-sectional image feature is a first cross-sectional image feature present in the first and second cross-sectional image slices. In another example, the common cross-sectional image feature is an alignment feature provided near the examination volume or an image segment of an alignment feature present near the examination volume. The step of lateral mutual image alignment is optionally further improved by subtraction or numerical compensation of image distortion deviations between at least the first and second cross-sectional image slices.

[0033] According to one embodiment of the present invention, the series of N cross-sectional image slices includes at least a first cross-sectional image slice and a second cross-sectional image slice of the inspection volume, and the first cross-sectional image surface is cut to extend longer in a direction perpendicular to the FIB beam than the second cross-sectional surface such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface. At least one alignment feature can be formed on the parallel surface segment of the first cross-sectional surface and near the second cross-sectional surface for first lateral mutual alignment of the first and second cross-sectional image slices.

[0034] According to one embodiment of the present invention, precise alignment of multiple cross-sectional image slices of an inspection volume below a wafer surface is obtained. For example, a first coarse alignment of first and second cross-sectional image slices is performed using additional alignment features formed near the inspection volume. The first alignment provides a mapping of cross-sectional image features in the first and second cross-sectional image slices. A second precision alignment of the first and second cross-sectional image slices utilizes cross-sectional image features of semiconductor structures within the inspection volume below the wafer surface, thereby achieving a mutual position accuracy of less than 5 nm, less than 3 nm, or less than 2 nm. In one example, the second fine alignment includes a first displacement ΔY′ of the first cross-sectional image feature between the first and second cross-sectional image slices. Ch and calculating a second displacement ΔY′ of the second cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice. WL , determining a distance d between the first and second cross-sectional image slices, and determining a mutual lateral displacement vector ΔY′ between the first and second cross-sectional image slices. Therefore, such a two-step registration process can avoid registration errors and achieve accurate registration.

[0035] In one example method for inspecting an examination volume in a wafer, acquiring a series of N cross-sectional image slices includes acquiring a plurality of lateral displacement image segments and stitching the plurality of lateral displacement image segments to form the cross-sectional image slice. In one example, acquiring the plurality of image segments with a charged particle imaging microscope includes changing a focal position of a charged particle imager for at least a subset of the plurality of image segments, thereby maintaining high resolution imaging over long elongations of the examination volume in a direction perpendicular to the wafer surface.

[0036] In one embodiment of a method for inspecting an examination volume in a wafer, a fan beam tomography technique is used. In this embodiment, a plurality or series of N cross-sectional image slices in the examination volume are obtained by scanning a focused ion beam of a FIB column in a first direction by a scanning unit to expose a first cross-sectional surface in the examination volume by FIB cutting, tilting the focused ion beam in a second direction perpendicular to the first direction by the scanning unit, and scanning the focused ion beam in the first direction by the scanning unit to expose a second cross-sectional surface in the examination volume by FIB cutting, such that the first and second cross-sectional surfaces form different angles with the wafer surface by approximately a tilt angle GF. In the fan beam tomography technique, the multiple cross-sectional surfaces form different angles with the wafer surface with an angular spread GZ centered on the tilt angle GF. In the fan beam tomography technique, the wafer is not moved during the step of acquiring the multiple cross-sectional image slices in the examination volume.

[0037] In one embodiment, a cross-section imaging method in a wedge-cut geometry and a method for inspecting an inspection volume in a wafer include distortion compensation. Image distortion in the cross-sectional image slice occurs, for example, due to a charged particle imaging beam with an angle GE deviating from 0°, for example, when the beam angle is 10° or greater. Other sources of image distortion include errors in the image scanning unit of the charged particle imager, dynamic changes in the focal position of the charged particle imager, or deviations from the planar shape of the cross-sectional surface. Image distortion in the cross-sectional image slice is determined from a predetermined angle GE or a second cross-sectional image feature, for example, by comparison with prior knowledge of the semiconductor design, and the image distortion is digitally compensated for.

[0038] One embodiment of the present invention includes an algorithm and method for generating a 3D volumetric image of an inspection volume. In a first step, cross-sectional image features are detected in multiple cross-sectional image slices, e.g., by object detection methods known in the art. The cross-sectional image features are further classified into first and second cross-sectional image features in a feature classification. In a second step, a depth map is generated for each cross-sectional image slice from the second cross-sectional image features, which represent features at a known or reference depth, as described above. The depth map can be an absolute depth map in units of nm, or a relative depth map scaled relative to a reference given by the relative depth of a second semiconductor integrated feature (e.g., a feature in a particular layer or planar layers that make up an integrated circuit). In a third step, the depth of each first cross-sectional image feature is determined based on the depth map. In a fourth step, a 3D volumetric image of the inspection volume is generated, which includes depth information for multiple first cross-sectional image features from multiple depth maps for each of the multiple cross-sectional image slices. A 3D volume image of the inspection volume is generated from the depth map, for example, by projection and interpolation of the virtual cross-sectional image slices, in conjunction with the multiple first cross-sectional image features of the multiple cross-sectional image slices. A fifth step is to derive characteristics of the integrated circuit structure of interest in the 3D volume image, such as the tilt angle or wobble, for example, of the HAR structure. A sixth step is to derive defects in the integrated circuit feature or characteristic in the 3D volume image. In one example, integrated circuit features deviate from their respective lateral design positions or depth positions beyond the wafer surface, for example, corresponding to fabrication errors of the HAR structure. Of particular interest in wafer defect inspection is the detection of fabrication errors. By utilizing the method of determining the depth of the first cross-sectional image features using the second cross-sectional image features, fabrication errors of the HAR structures can be determined with high accuracy, including global offset errors, such as the global lateral offset of all HAR structures.

[0039] In one embodiment, a method is provided for acquiring a virtual cross-sectional image slice or a series of virtual cross-sectional image slices from a set of cross-sectional image slices, each of the virtual cross-sectional image slices comprising a plurality of virtual cross-sectional image pixels. The method includes acquiring a series of N cross-sectional image slices by alternately imaging and cutting a series of N cross-sectional surfaces at an inclination angle GF relative to an inspection volume inside a wafer, and determining a first orientation direction of first semiconductor features constituting a first plurality of first cross-sectional image features in the series of N cross-sectional image slices. The method for acquiring the virtual cross-sectional image or series of virtual cross-sectional images further includes computing a virtual cross-sectional image perpendicular to the first orientation direction, wherein a pixel value is computed for each virtual cross-sectional image pixel by projecting a subset of at least one cross-sectional image slice of the series of N cross-sectional image slices in the first orientation direction and interpolating the pixel values ​​from the projections of the subset of at least one cross-sectional image slice. In the method for acquiring at least one virtual cross-sectional image, the number N of cross-sectional image slices is at least N=10, preferably N>100 (e.g., N is approximately 1000 or greater).

[0040] In one example of a method for acquiring at least one virtual cross-sectional image, a subset of at least one cross-sectional image slice is selected by evaluating, for each virtual cross-sectional image pixel, a distance from each cross-sectional image slice of the series of N cross-sectional image slices in a first orientation direction to the virtual cross-sectional image pixel and selecting at least a first cross-sectional image slice having the shortest distance. In one example, a second cross-sectional image slice of the subset of at least one cross-sectional image slice is therefore selected as the cross-sectional image slice having a second shortest distance. In another example, another cross-sectional image slice of the subset of at least one cross-sectional image slice is selected in order of increasing distance to the virtual cross-sectional image pixel in the first orientation direction.

[0041] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the steps of projecting a subset of at least one cross-sectional image slice and interpolating pixel values ​​from the projection of the subset of at least one cross-sectional image slice include projecting and interpolating at least a subset of the first plurality of first cross-sectional image features to form a third plurality of first cross-sectional image features in the virtual image slice. In one example, the steps of projecting and interpolating are combined with at least one of feature extraction, thresholding, contour interpolation, and model-based interpolation. The feature extraction, thresholding, contour interpolation, or model-based interpolation detects the first cross-sectional image features in the subset of the at least one cross-sectional image slice and accurately interpolates the third plurality of first cross-sectional image features in the virtual image slice.

[0042] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the method further includes generating a depth map Z(x,y;n) (index n=1...N) for each of the series of N cross-sectional image slices. In one example, the depth map Z(x,y;n) (index n=1...N) for each of the N cross-sectional image slices is generated from a plurality of second cross-sectional image features representing a cross section through a second semiconductor feature oriented in a second orientation direction perpendicular to the first orientation direction. In one example, the depth map Z(x,y;n) (index n=1...N) for each of the cross-sectional image slices is generated by determining a depth of a first cross-sectional image feature in the cross-sectional image slice from lateral positions of at least two second cross-sectional image features in the cross-sectional image slice.

[0043] In one example, a virtual cross-sectional image depth map ZV(x,y) is generated. A subset of at least one cross-sectional image slice is selected from the depth map ZV(x,y) and a plurality of depth maps Z(x,y;n) (index n=1...N) by evaluating, for each of the N cross-sectional image slices, a distance from each depth map Z(x,y;n) for each cross-sectional image slice in a first orientation direction to the depth map ZV(x,y) of the virtual cross-sectional image pixel, and selecting at least a first cross-sectional image slice having the shortest distance.

[0044] In one example, a second cross-sectional image slice of the subset of at least one cross-sectional image slice is therefore selected as the cross-sectional image slice having the second shortest distance, hi another example, another cross-sectional image slice of the subset of at least one cross-sectional image slice is selected in order of increasing distance to the virtual cross-sectional image pixel in the first orientation direction.

[0045] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the first semiconductor feature comprises at least one of a via, a HAR structure, or a HAR channel of a semiconductor integrated circuit inside the inspection volume of the wafer. The second semiconductor feature comprises at least one of an insulating line or insulating layer, a metal line, a metal layer, or a metal word line, or a semiconductor line or semiconductor layer of a semiconductor integrated circuit inside the inspection volume of the wafer. The method further comprises determining at least one first cross-sectional image feature in each of the series of N cross-sectional image slices and determining at least one second cross-sectional image feature in each of the series of N cross-sectional image slices.

[0046] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the first orientation direction is a z-direction perpendicular to the wafer surface, and the virtual cross-sectional image slices are computed in a plane parallel to the wafer surface at a depth ZV below the wafer surface. In one embodiment of the method for acquiring at least one virtual cross-sectional image, the subset of at least one cross-sectional image slice is determined by selecting, for each virtual cross-sectional image pixel coordinate (x,y), at least m cross-sectional image slices having the shortest distance to the depth ZV, such that the distance ZV(m) = Z(x,y;m) - ZV is the minimum value of all depth maps Z(x,y;n) (index n = 1...N). In one example, a second, different cross-sectional image slice of the subset of at least one cross-sectional image slice is selected in order of increasing distance ZV(n) to the virtual cross-sectional image pixel in the z-direction. In one example, the depth ZV is adjusted according to the depth of layers parallel to the wafer surface, the layers being formed by second semiconductor features oriented in a second orientation direction parallel to the wafer surface. In one example, a first virtual cross-sectional image slice is computed at a depth ZV1 in an insulating layer between two adjacent metal layers or word lines, and in one example, a second virtual cross-sectional image slice is computed at a depth ZV2 inside the metal layers or word lines.

[0047] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the optical axis of a charged particle beam imaging system acquiring the series of N cross-sectional image slices is oriented perpendicular to the wafer surface such that the angle G between the optical axis and a z-axis normal to the wafer surface is G = 0°. In one example, the charged particle imaging device is a helium ion microscope (HIM).

[0048] In one embodiment of the method for acquiring at least one virtual cross-sectional image, the method further includes forming at least one alignment feature near the examination volume configured to form at least one common cross-sectional image feature for lateral mutual registration, and performing lateral mutual registration of each of the series of N cross-sectional image slices by the at least one common cross-sectional image feature. In one example, the lateral mutual image registration step includes subtraction of image distortion deviations.

[0049] The method for acquiring at least one virtual cross-sectional image includes a dual beam apparatus including at least a FIB column and a charged particle imager, wherein a first optical axis of the FIB column forms a tilt angle GF with a surface of the wafer support table and a second optical axis of the charged particle imager forms an angle GE with a normal to the surface of the wafer support table, the first and second optical axes forming an intersection point, and further includes loading a wafer onto the wafer support table and moving the wafer support table to align a first measurement location on the wafer with the intersection point of the dual beam apparatus. In one example, the step of acquiring the series of N cross-sectional image slices through the inspection volume does not involve movement of the wafer.

[0050] According to one embodiment, a method for obtaining at least one 2D virtual cross-sectional image or a set of 2D virtual cross-sectional images from a set of cross-sectional image slices comprises: obtaining a series of N cross-sectional image slices by alternating imaging and cutting a series of N cross-sectional surfaces at a tilt angle GF relative to an inner inspection volume of the wafer; determining a first orientation direction of a first semiconductor structure or feature of interest comprising a first plurality of first cross-sectional image features in the series of N cross-sectional image slices; computing virtual cross-sectional images perpendicular to the first orientation direction, each virtual image slice including a plurality of virtual cross-sectional image pixels; Including, For each virtual cross-sectional image pixel, a pixel value is calculated by projecting a subset of at least one cross-sectional image slice of the series of N cross-sectional image slices in a first orientation direction and interpolating the pixel values ​​from the projections of the subset of cross-sectional image slices. For each virtual cross-sectional image pixel, a subset of at least one cross-sectional image slice is determined by calculating a distance from each of the series of N cross-sectional image slices in the first orientation direction to the virtual cross-sectional image pixel and selecting at least a first cross-sectional image slice having the shortest distance. In one example, a second cross-sectional image slice of the subset of at least one cross-sectional image slice is therefore selected as the cross-sectional image slice having the second shortest distance. Therefore, other cross-sectional image slices of the subset of at least one cross-sectional image slice can be selected in order of increasing distance to the virtual cross-sectional image pixel in the first orientation direction.

[0051] In one example, a method for obtaining at least one 2D virtual cross-sectional image or a set of 2D virtual cross-sectional images includes: determining at least one first cross-sectional image feature in each of the series of N cross-sectional image slices; determining at least one second cross-sectional image feature in each of the series of N cross-sectional image slices and deriving a depth of the at least one first cross-sectional image feature relative to a depth of the at least one second cross-sectional image feature; The depth of the at least one second cross-sectional image feature is known in advance or is used as a reference. Further details of the depth map generation method will be described in the examples and embodiments.

[0052] In one embodiment of a wafer defect inspection system and method for inspection of an inspection volume in a wafer, the charged particle imager is a scanning helium ion microscope. The cross-sectional surface is exposed at approximately a tilt angle GF and extends through the entire inspection volume to the wafer surface at approximately a tilt angle GF. The scanning helium ion microscope provides the necessary depth of focus to enable high-resolution imaging of the cross-sectional surface in a single image scan without having to change the focus of the charged particle imager, even if the inspection volume extends more than 2 μm in a direction perpendicular to the wafer surface.

[0053] According to one embodiment of the present invention, a wafer inspection method is disclosed for 3D inspection of an inspection volume in a wafer, the inspection volume being not extracted or removed from the wafer and having a depth range of greater than 1 μm, preferably greater than 2 μm, greater than 6 μm, or even 10 μm. The method includes generating an image by 3D inspection with a lateral resolution of less than 2 nm, preferably less than 1 nm, or even less than 0.5 nm, where the 3D inspection of the inspection volume is performed on the wafer, and the image is a 3D image. The method further includes cutting at least one cross-sectional surface through the depth range of the inspection volume with a FIB column. The image is obtained by acquiring at least one image of the at least one cross-sectional surface through the inspection volume in a single image scan with a helium ion microscope. The method further includes image processing to perform at least one of feature extraction, edge detection, pattern recognition, and pixel interpolation. The method further includes computing one of a depth map, a 3D volume image, and a virtual cross-sectional image. The method further includes calculating one of a dimension, an area, a diameter, an angle, and a shape of the semiconductor feature. The method further includes calculating one of an average, a statistical deviation of a plurality of semiconductor features.

[0054] In one embodiment of the present invention, a method for inspecting a wafer using a dual beam apparatus is provided. The dual beam apparatus includes a focused ion beam (FIB) column with its optical axis disposed at an angle GF of 30° to 45° relative to the support surface of a wafer support table, and a helium ion microscope (HIM) with its optical axis disposed perpendicular to the support surface, and the optical axes of the FIB column and HIM form an intersection. This wafer inspection method includes: positioning a first measurement site on the wafer by a wafer stage having a wafer support table configured to hold the wafer at the intersection; cutting a first cross-sectional surface through a first inspection volume at approximately angle GF with the FIB column, the first cross-sectional surface having a depth extension LZ below the wafer surface of greater than 1 μm; imaging a first cross-sectional surface with a single image scan using a helium ion microscope to form a high resolution cross-sectional image slice; Further includes:

[0055] In one example, the first inspection volume has a depth range of greater than 2 μm, greater than 6 μm, or even 10 μm, and the HIM is configured to generate images with a lateral resolution of less than 2 nm, preferably less than 1 nm, or even less than 0.5 nm in a single image scan of a cross-sectional surface at an angle GF through the inspection volume.

[0056] In one embodiment of the present invention, a method for inspecting an inspection volume below the wafer surface of a wafer with a dual beam apparatus is disclosed, the dual beam apparatus comprising a focused ion beam (FIB) column with its optical axis oriented at an angle GF of greater than 30° relative to a support surface of a wafer support table, and a helium ion microscope (HIM) with its optical axis oriented perpendicular to the support surface, the optical axes of the FIB column and the HIM forming an intersection point. The method comprises: loading a wafer onto a wafer support table of a dual beam apparatus; positioning a first measurement site on the wafer by a wafer stage having a wafer support table configured to hold the wafer at the intersection; cutting, with the FIB column, a plurality of N alternating cross-sectional surfaces in the inspection volume at approximately angle GF, the plurality of N alternating cross-sectional surfaces having a depth extension LZ greater than 1 μm below the wafer surface; imaging each cross-sectional surface with a single image scan by a helium ion microscope to form a plurality of N high resolution cross-sectional image slices; Includes.

[0057] In one example, the number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is greater than or equal to about 1000. In an alternative example, the number N of cutting planes is less than 50, preferably less than 20, and the examination volume is separated into a number B of blocks arranged diagonally through the examination volume.

[0058] In one example, the plurality of N cross-sectional surfaces are cut to a depth extension LZ below the wafer surface of greater than 2 μm, preferably greater than 6 μm, e.g., 10 μm. In one example, the method further includes computing at least one virtual cross-sectional image parallel to the wafer surface from the plurality of N cross-sectional image slices. For example, a first virtual cross-sectional image is computed at a conductive layer or word line, and a second virtual cross-sectional image is computed at an insulating layer.

[0059] In one example, the method further includes performing lateral co-registration of each of the series of N cross-sectional image slices with at least one common cross-sectional image feature. In one example, the method includes forming at least one alignment feature near the examination volume configured to form the at least one common cross-sectional image feature, and performing lateral co-registration of the series of N cross-sectional image slices with the at least one common cross-sectional image feature. The method includes: generating a depth map Z(x,y;n) for each of a series of N cross-sectional image slices (index n=1...N for each of the N cross-sectional image slices); determining, in each of the series of N cross-sectional image slices, at least one second cross-sectional image feature representing a cross section through a second semiconductor feature oriented parallel to the wafer surface, wherein a depth map Z(x,y;n), index n=1 N, for each cross-sectional image slice is generated from the lateral positions of the at least one second cross-sectional image feature; Further includes:

[0060] In one embodiment, a method for cross-sectional imaging in a wedge cut geometry for a dual beam device includes providing additional trenches adjacent to the inspection volume for debris suppression. The additional trenches are created by FIB cutting in a first direction and include a first proximal trench and a second distal trench. After creating the additional trenches, the wafer is rotated 90° and the inspection volume is inspected using the cross-sectional imaging in a wedge cut geometry technique. Debris generated during cutting of the cross-sectional surface is collected in the distal additional trench.

[0061] According to one embodiment, a wafer inspection apparatus for inspection of an inspection volume in an inspection site of a wedge-cut shaped wafer includes a six-axis wafer stage with a wafer support table configured to hold the wafer by the wafer surface on a support surface. The wafer inspection apparatus is configured to inspect the inspection volume in the wafer without extracting it from the wafer and includes a dual beam apparatus. The dual beam apparatus includes a focused ion beam (FIB) column configured to cut and expose at least a first cross-sectional surface in a first inspection volume inside the wafer at approximately a tilt angle GF relative to the surface of the wafer, and a charged particle imager configured to perform imaging. The charged particle imager is, for example, a scanning electron microscope (SEM) or a helium ion microscope (HIM), wherein a first optical axis of the FIB column forms a tilt angle GF with the surface of the wafer support table and a second optical axis of the charged particle imager forms an angle GE with a normal to the surface of the wafer support table. The first and second optical axes of the dual beam apparatus columns form an intersection point. The wafer inspection apparatus further includes a control unit configured to perform and control one of the methods for inspection of the inspection volume in the wafer. The wafer inspection apparatus further comprises an image processing unit, the control unit and the image processing unit comprising a processor and a memory having installed thereon software code configured to carry out the method of the present invention, including the image processing and depth map generating steps described above.

[0062] In one embodiment of the present invention, the wafer defect inspection apparatus is configured to cut and expose a plurality of cross-sectional surfaces in a wafer, including at least first and second cross-sectional surfaces. The wafer defect inspection apparatus is further configured to image the at least first and second cross-sectional surfaces to form first and second cross-sectional image slices. The wafer defect inspection apparatus further includes an image processing unit having installed thereon software code configured to determine cross-sectional image features in the at least first and second cross-sectional image slices and determine depths in the inspection volume of the cross-sectional image features, which are cross-sections of the semiconductor structure within the inspection volume. The wafer defect inspection apparatus further includes a defect detection unit configured to determine, from the cross-sectional image features, deviations of the semiconductor structure within the inspection volume from a predetermined characteristic.

[0063] In one example, a wafer defect inspection apparatus includes a wafer stage for holding a wafer, a focused ion beam (FIB) column configured to cut and expose a series of N cross-sectional surfaces in an inspection volume inside the wafer at a depth extension of greater than 2 μm and at a tilt angle GF relative to the surface of the wafer, and a charged particle imager configured to image the plurality of N cross-sectional surfaces to form a plurality of N cross-sectional image slices in a single image scan. The wafer defect inspection apparatus further includes an image processing unit having installed thereon software code configured to determine a virtual cross-sectional image oriented parallel to the surface of the wafer from the series of N cross-sectional image slices.

[0064] In one example, the wafer defect inspection apparatus further comprises a focused ion beam (FIB) column configured, in use, to cut and expose a series of N cross-sectional surfaces in an inspection volume inside the wafer at an inclination angle GF relative to the surface of the wafer, a charged particle imager configured, in use, to image the series of N cross-sectional surfaces to form a series of N cross-sectional image slices, and an image processing unit having installed thereon software code configured, in use, to determine a virtual cross-sectional image oriented parallel to the surface of the wafer from the series of N cross-sectional image slices. The wafer defect inspection apparatus with a dual beam device comprises an operational unit and an image processor configured to perform a method for obtaining a virtual cross-sectional image from a set of cross-sectional image slices.

[0065] In some embodiments or examples of the present invention, the charged particle imaging beam is a scanning electron beam. In another example, the charged particle imaging beam is an ion beam of an ion microscope, such as a helium ion beam of a helium ion microscope (HIM). The resolution achieved by HIM through slice imaging is less than 1 nm, preferably less than 0.5 nm, and more preferably 0.25 nm. HIM also provides higher material contrast. HIM also provides a large depth of focus of up to 10 micrometers, which is significantly greater than that of a scanning electron beam, and particularly greater than the depth range (1 μm, greater than 2 μm, and up to 10 μm) of a cross-sectional surface cut relative to the wafer surface at a cutting angle GF of preferably 30° to 80°. Typically, the inspection volume has a depth extension LZ below the wafer surface greater than 1 μm, preferably greater than 2 μm, and the HIM is configured with a depth of focus (DOF) greater than the depth extension LZ so that a single image scan with the HIM beam can obtain cross-sectional image slices without image stitching. Other examples of imaging ion beam microscopes include other noble gases as ions, such as neon or a mixture of helium and neon.

[0066] In one embodiment, a dual beam apparatus for wafer inspection includes: (a) a wafer stage supporting a wafer; (b) a gallium FIB column positioned at an angle GF relative to a support surface of the sample support stage and configured to cut at least one cross-sectional surface at the angle GF relative to the wafer; (c) a helium ion beam column positioned perpendicular to the support surface of the wafer stage and configured to generate a first cross-sectional image slice of the cross-sectional surface with a single image scan; (d) a secondary electron detector configured to collect multiple secondary electrons during the single image scan; and (e) an operational unit configured to control the operation of the dual beam apparatus in use. The FIB with an angle GF of 30°-40° is configured to cut the cross-sectional surface to a depth range below the wafer surface extending over more than 1 μm, more than 2 μm, or more than 6 μm. The operational unit further includes (f) an image processing unit performing at least one of feature extraction, edge detection, pattern recognition, and pixel interpolation. The image processing unit is further configured to compute one of a depth map, a 3D volume image, and a virtual cross-sectional image. The operational units further include (g) a defect detection unit configured to calculate one of a dimension, an area, a diameter, an angle, and a shape of the semiconductor features, the defect detection unit further configured to calculate one of an average, a statistical deviation of the plurality of semiconductor features.

[0067] According to one embodiment, a wafer inspection apparatus for inspecting an inspection volume at an inspection site on a wafer in a wedge cut geometry includes a control unit configured to control a FIB column and a HIM to alternately cut a plurality of N cross-sectional surfaces in the inspection volume at approximately an angle GF with the FIB beam and image each cross-sectional surface with a single image scan with the HIM beam to form a plurality of N high-resolution cross-sectional image slices with a resolution better than 2 nm, preferably 1 nm, where the inspection volume has a depth extension LZ of greater than 1 μm, greater than 3 μm (e.g., 6 μm or 10 μm) in the direction of an optical axis of the HIM. The imaging of each cross-sectional surface to form the plurality of N high-resolution cross-sectional image slices is achieved by the HIM beam having a depth of focus (DOF) greater than the depth extension LZ. In use, the control unit is configured to scan the HIM beam with a single image scan over an area greater than a lateral dimension LX or LY of the inspection volume (approximately 10 μm (LX or LY)) and time-sequentially collect a plurality of secondary electrons with the secondary electron detector.

[0068] In one example of the present invention, there is provided a dual beam apparatus for inspection of an inspection volume below a wafer surface of a wafer, comprising: a wafer stage including a wafer support table configured, in use, to hold a wafer on a support surface; a focused ion beam (FIB) column having an optical axis oriented at an angle GF greater than 30° relative to a support surface of a wafer support table; a helium ion microscope (HIM) positioned with its optical axis perpendicular to the support surface, the HIM forming an intersection between the FIB column and the optical axis of the HIM; a stage control unit configured, in use, to position a first measurement site on the wafer at the intersection; a control unit configured to control the FIB column and the HIM to alternately cut a plurality of N cross-sectional surfaces in the inspection volume at approximately an angle GF with the FIB beam and image each cross-sectional surface by scanning with the HIM beam to form a plurality of N high-resolution cross-sectional image slices, wherein the inspection volume has a depth extension LZ below the wafer surface of more than 1 μm, preferably more than 2 μm; A dual-beam apparatus is disclosed, comprising: a FIB beam having a depth extension LZ greater than 5 μm (e.g., up to 10 μm) in the direction of the HIM beam; and a HIM beam having a depth extension LZ greater than 5 μm (e.g., up to 10 μm). Cutting a plurality of N cross-sectional surfaces in the inspection volume at approximately the angle GF with the FIB beam means that the actual angle of the cross-sectional surfaces may deviate from the cutting angle GF by a few degrees (e.g., 1° to 4°) due to divergence of the FIB beam. The control unit is further configured, in use, to scan the HIM beam in a single scan across an area greater than a lateral dimension LX or LY of the inspection volume (approximately 5 μm to 10 μm (LX or LY)) and time-sequentially collect a plurality of secondary electrons with the secondary electron detector. The HIM has a depth of focus (DOF) greater than the depth extension LZ such that a single image scan with the HIM beam can obtain a plurality of N high-resolution cross-sectional image slices each with a high resolution of approximately less than 2 nm (e.g., 1 nm or 0.5 nm). In one example, the dual beam apparatus according to the embodiment further comprises an image processing unit having installed thereon software code configured to determine at least a first virtual cross-sectional image oriented parallel to the surface of the wafer from the plurality of N cross-sectional image slices.

[0069] The optical axis of the FIB column is positioned at a tilt angle GF relative to the wafer support table, and the tilt angle GF can be between 30° and 80° (e.g., GF is approximately 30° and 45°). In alternative embodiments, the tilt angle GF between the first optical axis of the FIB column and the surface of the wafer support table is in the range of 45° and 80°. In alternative embodiments, the tilt angle GF between the first optical axis of the FIB column and the surface of the wafer support table is in the range of 8° and 30° (e.g., 8° and 15°).

[0070] The optical axis of the charged particle imager can be positioned at an angle (GE=0°) perpendicular to the wafer support table. In one example, the optical axis of the charged particle imager is positioned approximately perpendicular to the wafer, or such that the angle GE between the optical axis of the charged particle imager and the normal to the surface of the wafer support table is less than 10°.

[0071] Another method of the present invention is a method for generating a metrology recipe for inspection of a representative inspection volume in a wafer, the method comprising: generating a first set of critical design performance indicators D1 including dimensions of critical design features in the CAD image data and the inspection volume; generating a second set of key fabrication performance indicators D2 that include 3D volumetric image data obtained during fabrication process development; deriving a measurement recipe R that includes obtaining at least one of a third set of performance indicators D3; Including, Generating the second set of key fabrication performance indicators D2 and the measurement recipe R includes acquiring at least a first cross-sectional image slice of at least a first cross-sectional surface through a representative inspection volume having a depth extension LZ>1 μm below the wafer surface. At least one of acquiring at least one of the performance indicators D1-D3 includes extracting a plurality of first cross-sectional image features representing cross sections of the plurality of first semiconductor features at the tilt angle GF by image processing, where the image processing includes at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation. At least one of acquiring at least one of the performance indicators D1-D3 further includes calculating at least one descriptive parameter from at least one of the plurality of first cross-sectional image features, where the at least one descriptive parameter of the at least one first semiconductor feature is one of a dimension, a diameter, an angle, an area, a shape, or a volume. At least one of acquiring at least one of the performance indicators D1-D3 further includes calculating an average or statistical deviation of the at least one descriptive parameter of the plurality of first semiconductor features.

[0072] The present invention further provides a computer program product including program code for performing any of the methods of the present invention. For example, a computer program product including program code for performing a method of inspection of an inspection volume in a wafer is provided. For example, a computer program product including program code for performing any of the method steps of a method of obtaining at least one virtual cross-sectional image is provided. For example, the computer program product includes an element for controlling generation of a set of N cross-sectional image slices, an element for selecting a first orientation direction, and an element for computing at least one virtual cross-sectional image from the set of N cross-sectional image slices by projection and interpolation in the first orientation direction.

[0073] According to an embodiment, there is provided a dual beam device comprising a control unit and an image processing unit configured to perform any of the method steps of the above methods.

[0074] Throughout the embodiments, the first cross-sectional image feature includes a cross-section of the semiconductor structure extending in a direction parallel to at least an axis perpendicular to the wafer surface, and the second cross-sectional image feature includes a cross-section of the semiconductor structure extending in a direction parallel to at least the wafer surface. For example, the first common cross-sectional image feature includes a cross-section of at least one of a via, a HAR structure, or a HAR channel of a semiconductor integrated circuit. For example, the second cross-sectional image feature includes a cross-section of at least one of an insulating line or insulating layer, a metal line, a metal layer, or a metal word line, or a semiconductor line or semiconductor layer of a semiconductor integrated circuit.

[0075] The above-described embodiments or examples may be combined in whole or in part with one another, as may examples or embodiments describing different aspects of the present invention.

[0076] The invention may be better understood with reference to the following drawings. [Brief explanation of the drawings]

[0077] [Figure 1] FIG. 1 illustrates a cross-sectional imaging technique on a sample extracted from a wafer. [Figure 2] FIG. 1 is a diagram of a dual beam apparatus configured for cross-sectional imaging techniques in a wedge cut geometry applied to a wafer. [Figure 3] FIG. 10 shows details of cross-sectional imaging technique in wedge cut geometry. [Figure 4] 10A-10C illustrate another aspect of the cross-sectional imaging technique in a wedge cut geometry. [Figure 5] FIG. 10 is an illustration of a method for determining the depth of a cross-sectional image feature in the x-direction. [Figure 6] FIG. 10 is an illustration of a method for determining the depth of a cross-sectional image feature in an example of two cross-sectional image slices. [Figure 7] FIG. 10 is a diagram of a cross-sectional imaging technique in a wedge cut geometry utilizing additional alignment features. [Figure 8]FIG. 1 is a diagram of a cross-sectional imaging technique in a wedge cut geometry using stitching of image segments. [Figure 9] FIG. 10 is an illustration of a cross-sectional imaging technique in a wedge cut geometry utilizing additional alignment features at different depths. [Figure 10] FIG. 1 is a diagram of a cross-sectional imaging technique in a wedge cut geometry using a fan beam tomography technique. [Figure 11] FIG. 10 is a diagram of a cross-sectional imaging technique in a wedge cut geometry with an additional trench for debris suppression. [Figure 12] FIG. 1 is a diagram of a cross-sectional imaging technique in a wedge cut geometry with distortion compensation. [Figure 13] FIG. 10 is a diagram of a cross-sectional imaging technique in a wedge cut geometry where additional alignment features are created above the examination volume. [Figure 14] 1A-1C are diagrams of method steps for cross-sectional imaging techniques in wedge cut geometry. [Figure 15] FIG. 1 shows the operation and control unit. [Figure 16] FIG. 10 is a diagram of the interpolation of a virtual image slice from a series of cross-sectional image slices. [Figure 17] FIG. 10 is another illustration of the interpolation of a virtual image slice from a series of cross-sectional image slices from a curved cross-sectional surface. [Figure 18a] FIG. 1 is a diagram of a dual beam setup using a helium ion microscope. [Figure 18b] FIG. 1 is a diagram of a dual beam setup using a helium ion microscope. [Figure 19-1] 19. Diagram of the slice image technique using the dual beam device according to Figure 18 and the results of single scan image acquisition and virtual cross section image calculation on alternating layers of the memory device. [Figure 19-2] 18. b)-d) Slice image technique using the dual beam device according to FIG. 18 and the results of single scan image acquisition and virtual cross section image calculation in alternating layers of the memory device. [Figure 20] FIG. 1 is a diagram of a statistical evaluation of HAR features of a memory device. [Figure 21] FIG. 1 is a diagram of a time-efficient method for inspecting volumes in a wafer. [Figure 22] Illustrates inspection of a deep inspection volume with a large cutting angle GF. [Figure 23] FIG. 1 is a diagram of a process flow for the generation of a monitoring recipe for in-line wafer inspection. [Figure 24] FIG. 10 is a diagram of an analysis of an examination volume based on preliminary information. [Figure 25] FIG. 10 is a diagram illustrating an example of a monitoring recipe for analyzing the gradient of a HAR structure. [Figure 26] FIG. 10 is a diagram of a cross-sectional imaging technique in a wedge cut geometry utilizing additional alignment features on the longer elongated past or first cross-sectional surface. [Figure 27] 1 is a diagram of a method for determining the depth of a cross-sectional image feature. DETAILED DESCRIPTION OF THE INVENTION

[0078] FIG. 1 is a schematic diagram of a common cross-sectional imaging technique for acquiring a 3D volumetric image of a semiconductor integrated sample. The cross-sectional imaging technique, also known as the slice imaging technique, achieves three-dimensional (3D) volumetric imaging in a "step-and-repeat" manner. First, the semiconductor integrated sample is prepared for the common cross-sectional imaging technique by methods known in the art. Throughout this disclosure, the terms "cross-sectional imaging" and "slice" are used synonymously. In one step, a thin surface layer, or "slice," of material is removed. This slice of material may be removed by several methods known in the art, such as using focused ion beam (FIB) cutting or polishing at a viewing angle with a focused ion beam (FIB) column 50. For example, a focused ion beam 51 propagates approximately parallel to the z-axis and scans in the y-direction, cutting the entire top surface of the sample 10 and exposing a new cross-sectional surface 52 in the yz-plane. As a result, the newly exposed cross-sectional surface 52 is accessible for imaging. In a subsequent step, a charged particle beam (CPB) imaging system 40, such as a scanning electron microscope (SEM) or a second FIB, raster-scans the cross-sectional surface layer 52 to acquire a cross-sectional image slice 100.1. The optical axis 42 of the charged particle imaging system 40 can be aligned parallel to the x-direction or tilted at an angle relative to the x-direction. A detector (not shown) collects secondary and backscattered electrons to reveal material contrast within the semiconductor integrated sample, which is visualized as different gray levels in the cross-sectional image slice 100.1. Metal structures produce brighter measurements. A three-dimensional (3D) data set is constructed by repeating the surface layer removal and cross-sectional imaging process for equidistant cross-sectional surfaces 53, 54, and other cross-sectional surfaces to acquire a series of 2D cross-sectional image slices 1000, including, for example, N cross-sectional image slices 100.2, 100.3, ... 100.N, at different depths. A representative cross-sectional image slice 100.1 is acquired by measuring a commercially available 14 nm technology Intel processor semiconductor integrated chip.

[0079] The method then involves cutting cross-sectional surfaces into the semiconductor integrated sample with the focused ion beam to expose or make accessible a series of cross-sectional surfaces for imaging, and imaging each cross-sectional surface of the semiconductor integrated sample with the charged particle beam imaging system 40 to generate at least first and second cross-sectional image slices. A 3D image of the semiconductor integrated structure is then reconstructed from the series of N 2D cross-sectional image slices 1000. The distance d between the cross-sectional image slices 100.1, 100.2, 100.3 can be controlled by the FIB cutting or polishing process and can be between 1 nm and 30 nm.

[0080] In the above example, as shown in FIG. 1, the cross-sectional image plane is oriented perpendicular to the top surface 55 of the semiconductor integrated wafer, and the normal to the wafer top surface 55 is oriented parallel to the z-axis. This results in a 2D cross-sectional image slice oriented parallel to the yz-plane. In other words, the cross-sectional image plane contains the z-axis or wafer normal axis, and the imaging direction x is parallel to the wafer surface. Therefore, conventional slice imaging methods in this conventional geometry are only applicable to samples extracted from the wafer.

[0081] The present invention provides a slice imaging method applicable to an inspection volume inside a wafer. In a first embodiment of the present invention, a 3D volume image generation method utilizes a slice imaging method applied to an inspection volume inside a wafer in a so-called "wedge cut" technique or wedge cut shape without the need to remove a sample from the wafer. The slice imaging method is applied to inspection volumes with dimensions of a few microns (e.g., 5-10 μm lateral extension on a 200 mm or 300 mm wafer) without removing a sample from the wafer. Cutting a groove or edge on the top surface of a semiconductor integrated wafer allows access to the cross-sectional surface at an angle relative to the top surface. A 3D volume image of the inspection volume is acquired at a limited number of measurement sites (e.g., representative sites on a die, sites identified by a process control monitor (PCM), or other inspection tool). Because the slice imaging method only locally destroys the wafer, it is still possible to use other dies or use the wafer for further processing.

[0082] An example of a first embodiment of the present invention and a wedge cut geometry is shown in FIG. 2. A wafer 8 is provided, including multiple measurement sites 6.1 and 6.2 defined in a position map generated, for example, from an inspection tool or design information. The wafer 8 is mounted on a wafer support table 15. The measurement site 6.1 of the wafer 8 is aligned with a five-axis wafer stage (not shown) at an inspection point 43 of a dual-beam apparatus including an FIB column 50 having an FIB optical axis 48 and a charged particle beam (CPB) imaging system 40 having an optical axis 42. At the inspection point 43 of both the optical axes of the FIB and CPB imaging systems, the wafer surface is disposed at an angle GF relative to the FIB axis 48. The FIB axis 48 and the CPB imaging system axis 42 form an angle GFE, and the CPB imaging system axis forms an angle GE with a z-axis normal to the wafer plane. An FIB 51 impinges on the surface of the wafer 8 at angle GF, causing an ion beam cut at inspection site 6.1 at approximately the tilt angle GF, resulting in a tilted cross-sectional surface being cut into the wafer. In the example of FIG. 2, the tilt angle GF is approximately 30°. The actual tilt angle of the tilted cross-sectional surface can deviate from the tilt angle GF by up to 1° to 4° due to beam divergence of a focused ion beam (e.g., a gallium ion beam). An image of the cut surface is acquired by a charged particle beam imaging system 40 tilted at an angle GE with respect to the wafer normal. In the example of FIG. 2, the angle GE is less than 15°. During imaging, a scanning unit of the charged particle beam imaging system 40 scans a beam of charged particles along a scanning path on the cross-sectional surface of the wafer at the measurement site 6.1, generating secondary particles and scattered particles. A particle detector 17 collects at least a portion of the secondary particles and scattered particles and transmits the particle count to a control unit 19. The control unit 19 controls the charged particle beam imaging system 40 and the FIB 50 and is connected to a control unit 16 that controls the position of a wafer mounted on a wafer support table via a wafer stage (not shown). The control unit 19 communicates with an operating unit 2 which triggers the placement and alignment of the measurement site 6.1 of the wafer 8 at the inspection point 43, for example by movement of the wafer stage, and which repeatedly triggers the operations of FIB cutting, image acquisition and stage movement.

[0083] Each intersecting surface is imaged by an essentially vertical charged particle imaging beam 44 (eg, a scanning electron beam or any other charged particle beam microscope such as a helium ion microscope (HIM)).

[0084] In a second embodiment, a slice imaging method in a wedge cut geometry is provided. Repeated slicing and imaging in the wedge cut geometry generates a stack of N cross-sectional image slices, including image slices of cross-sectional surfaces 52, 53, 54 (see FIGS. 1 and 3), and generates a 3D volume image of the inspection volume of the wafer 8 at the measurement site 6.1. FIG. 3 illustrates the wedge cut geometry in an example 3D memory stack. The FIB beam 51 cuts the wafer surface at an angle GF less than 80° relative to the wafer surface (e.g., a 30° angle GF), although other angles less than 45° (e.g., 40° or 36°) are also possible. In one example, a small cutting angle GF between 8° and 45° is preferred because it enables excellent imaging of the first cross-sectional image features, such as HAR structures, at high resolution. In another example, a steeper cutting angle GF (e.g., greater than 30° or greater than 45°) is preferred. A steep cutting angle GF between 45° and 80° allows for deep cutting of the inspection volume with small lateral dimensions. The steep cutting angle allows for greater depths and enables 3D volumetric imaging of deep semiconductor stacks, e.g., at depths greater than 6 μm (e.g., 10 μm or 15 μm for thick or deep semiconductor stacks). The steep angle maintains a lateral extension of the cross-sectional surface, e.g., below 30 μm, preferably 20 μm, allowing for acquisition of cross-sectional image slices without image stitching at high throughput.

[0085] Depending on the coordinate system chosen, wafer surface 55 coincides with the XY plane. The memory stack extends in the z direction, perpendicular to wafer surface 55. FIB beam 51 creates a new cross-sectional surface 52 in the wafer, tilted approximately at angle GF relative to the wafer surface. For example, in the example of FIG. 3, SEM beam 44, positioned at normal incidence to wafer surface 55, scans cross-sectional surface 52 to create a high-resolution image. The cross-sectional image slice includes first cross-sectional image features formed by intersections with high-aspect-ratio (HAR) structures or vias (e.g., first cross-sectional image features 4.1, 4.2, and 4.3 of HAR structures) and second cross-sectional image features formed by intersections with layers L1...LM, e.g., including SiO2, SiN, or tungsten lines. Some of these lines are also referred to as "word lines." The maximum number of layers M is typically greater than 50 (e.g., greater than 100 or greater than 200). The HAR structures and layers extend throughout most of the wafer volume, but may contain gaps. The HAR structures typically have diameters of less than 100 nm (e.g., approximately 80 nm or 40 nm). Thus, the cross-sectional image slice contains first cross-sectional image features as intersections or cross sections of the HAR channel footprint at different depths (z) for each xy position. For a cylindrically shaped vertical memory channel, the resulting first cross-sectional image features are circular or elliptical structures at various depths depending on the location of the structures on the inclined cross-sectional surface 52.

[0086] The selection of the slice distance d is a balance between the required sampling and throughput. The thickness d or the shortest distance d between two adjacent cross-sectional image slices is typically adjusted to a value on the order of a few nm (e.g., 30 nm, 20 nm, 10 nm, 5 nm, 4 nm, or less). When a material layer of a predetermined thickness d is removed by the FIB, a new cross-sectional surface 53 is exposed and becomes accessible for imaging by the nearly perpendicular imaging beam 44. The plurality of N cross-sectional image slices thus acquired covers the inspection volume of the wafer 8 at the measurement location 6.1 and is used to form a 3D volume image with a high 3D resolution, e.g., less than 10 nm, preferably less than 5 nm, and to reconstruct the characteristics of the semiconductor structure of interest (e.g., a memory stack inside the inspection volume). The plurality of N cross-sectional image slices can include from a few images (e.g., N=10 or N=20) to hundreds of cross-sectional image slices (e.g., N=1000 or more). The inspection volume typically has a lateral extension LX = LY = 5 μm to 15 μm in the xy plane and a depth LZ below the wafer surface of 2 μm to 15 μm, although the lateral extension of the inspection volume can be significantly larger, reaching approximately 1 mm. The lateral extension of a single cross-sectional surface is preferably less than 30 μm (e.g., less than 20 μm). In some instances, for memory devices including memory HAR structures, it is sufficient to provide at least three cross-sectional image slices per HAR structure. Therefore, the preferred slice distance d is less than 30 nm.

[0087] In one example, the features and 3D positions of the semiconductor structure of interest (e.g., the positions of the HAR channels) are detected by image processing methods, e.g., from the HAR centroid. 3D volume image generation, including image processing methods and feature-based registration, is described in detail in U.S. Provisional Patent Application No. 62 / 858,470 and German Patent Application No. 10 2019 006645.6, the entire contents of which are incorporated herein by reference. It should be noted that the layers and HAR structures need not extend throughout the entire measurement volume.

[0088] In the following, a third embodiment of the present invention will be described, which describes a method for recovering information about the relative position, depth, and orientation of cross-sectional image features of a HAR structure in an inspection volume. The angle GF between the FIB direction and the wafer surface 55 (xy plane) controls the depth LZ, and scanning in the x and y directions controls the field of view (FOV) of the imaging beam 44 and the total lateral area LX and LY available for scanning imaging by the imaging beam 44. Proper alignment of image slices with respect to each other is often referred to as alignment, and involves, in a first step, registering 2D image features and, in a second step, recovering the correct distance d between subsequent cross-sectional image slices. A first alignment method is shown in FIG. 4. A columnar HAR structure (e.g., a channel or channel hole), an example of which is referenced by the numeral 75, extends in the z direction through the volume of a wafer (e.g., a memory chip). The HAR structure is oriented perpendicular to the wafer surface and is visualized in the cross-sectional image slices acquired by the imaging beam 44 (see FIG. 3). As in the previous figure, the cross-sectional surfaces are approximately oriented at a predetermined angle GF relative to the HAR structure. In the example of Figure 4, the tilt angle GF is 25°. Figure 4b shows two example slices of two consecutive 2D cross-sectional images with indices n and n+1, with the first cross-sectional image feature of the HAR channel shown at 77.1-77.5. In this example, cross-sectional image slices n and n+1 perpendicular to the cross-sectional image surfaces n and n+1 are acquired with the imaging beam 44 positioned at an angle GFE of approximately 90° relative to the ion beam 51, and the image coordinate system X', Y', Z' is rotated about the x-axis by an angle GE relative to the wafer coordinate system (x, y, z). Cross-sectional surfaces n and n+1 each form an edge with the top surface of the wafer 55; two example edges are shown at 76.1 and 76.2. In this example, the first HAR structure intersects with a first cross-sectional image feature 77.1 in a first cross-sectional image slice with index n and with a first cross-sectional image feature 77.2 in a second image slice with index n+1. The first cross-sectional image features 77.1 and 77.2 are common cross-sectional image features for both image slices and are displaced by a lateral displacement vector dy'.Due to the lateral shift, some of the cross-sectional image features of the HAR structure disappear, such as cross-sectional image feature 77.4 in the nth image nearly disappearing in the n+1 cross-sectional image slice (only part of cross-sectional image feature 77.5 is visible). Generally, the apparent lateral displacement dy' in the image coordinate system is given by dy' = d·sin(GE) / cos(GF), where GF is the tilt angle of the tilted cross-sectional surface, d is the distance between the image slices, and GE is the angle of the imaging beam 44. The difference in lateral coordinates corresponds to the difference in depth dz beyond the wafer surface, dz = (dy') / sin(GE). In one example, the change in depth dz of corresponding first cross-sectional image features 77.1, 77.2 perpendicular to the wafer surface 55 is derived from the lateral displacement dy' of the first cross-sectional image features 77.1, 77.2 in at least two images n and n+1 of a plurality of images generated by cutting the FIB beam at a predetermined angle GF and imaging at a predetermined angle GE.

[0089] In the example depth determination method illustrated in FIG. 4 , the charged particle imaging beam 44 (see FIG. 3 ) is oriented at a predetermined angle GE with respect to a normal to the wafer surface 55, which corresponds to the z-axis in a selected coordinate system. This method utilizes the apparent lateral displacement of a first cross-sectional image feature in multiple cross-sectional images, where the first cross-sectional image feature is a cross-section of a vertically oriented structure in the semiconductor wafer, such as a via or HAR structure. The vertically oriented via or HAR structure is oriented perpendicular to the wafer surface 55. However, in the example method illustrated in FIG. 4 , the lateral displacement error of the first cross-sectional image feature affects the depth determination of the first cross-sectional image feature. Therefore, the third embodiment method illustrated in FIG. 4 utilizes the first cross-sectional image feature for depth determination, but has limited accuracy, especially for deep structures. For example, the actual cross-sectional image surface deviates from a planar shape, and this deviation increases with increasing depth. Furthermore, the HAR structure is not perfectly perpendicular to the wafer surface.

[0090] Therefore, another object of the present invention is to provide a method for determining the position of a first cross-sectional image feature with even greater accuracy. A solution is provided by a second example of the depth determination method of the present invention. This method is described with the predetermined angle GE selected to be approximately GE=0° by orienting the charged particle imaging beam 44 parallel to the normal to the wafer surface 55. Figure 5 illustrates this setup. However, this method is not limited to a charged particle imaging beam 44 oriented perpendicular to the wafer surface 55; angles GE deviating from 0° are possible as well.

[0091] Two cross-sectional surfaces 52 and 53 are shown tilted at a predetermined angle GF relative to wafer surface 55. In FIG. 2, the tilt angle GF is approximately 26°, but other tilt angles (e.g., 30° or greater) are possible as well. Two HAR structures 75.1 and 75.2 and a set of layers L1-L4 are shown. Because the charged particle imaging beam 44 is perpendicular to wafer surface 55, the first cross-sectional image features 77.1 and 77.2 of a theoretically perfect HAR structure 75.1 appear in the nth and n+1th images at the same y-coordinate, making depth determination from the apparent lateral displacement of the cross-sectional image features 77.1 and 77.2 in the nth and n+1th images impossible. If the first cross-sectional image features 77.1 and 77.2 of an actual HAR structure 75.1 appeared in the nth and n+1th images at different y-coordinates, any depth determination from the different y-coordinates would be erroneous. The depth determination according to the second example utilizes second cross-sectional image features of multiple layers including layers L1 to L4 in multiple N cross-sectional image slices including cross-sectional image slices n and n+1. For example, the second cross-sectional image feature (see reference numbers 73.1 and 73.2) of layer L4, whose depth below the wafer surface of top surface 78 is Z41, has a y' coordinate of Y41' in the nth cross-sectional image. n , Y41' in the n+1th cross-sectional image n+1 For example, the second cross-sectional image feature 73.3 of the lower surface 72 of layer L1, which is at a depth Z12 below the wafer surface, has a y' coordinate of Y12' in the nth cross-sectional image. nBy knowing the depths of at least two edges of two layers (e.g., depth Z41 and depth Z12), the y coordinates in the cross-sectional image slices can be converted to z coordinates to derive a depth map Z(x,y). For example, the depth z of the centroid 79.1 of the first cross-sectional image feature 77.1 in the nth image is derived from the y coordinate of the second cross-sectional image feature 73.3 and the y coordinate y' of the centroid 79.1: z=Z12+(y'-Y12') n )·tan(GF). In an alternative embodiment, the depth z of the centroid 79.1 of the first cross-sectional image feature 77.1 in the nth image is derived from the y-coordinates of at least two second cross-sectional image features 73.1 and 73.3, and the depth map Z(x,y) is calculated as z=Z12+(y-Y12' n )·(Z41-Z21) / (Y41' n -Y12' n ) for example, as a reference for lateral alignment of depth maps of two consecutive cross-sectional image slices n and n+1, or as a verification of the local cutting angle GF or the local slice distance dz, the local displacement of the position of the second cross-sectional image feature (e.g., the y-coordinate Y41' in two consecutive cross-sectional image slices n and n+1) n and Y41' n+1The local displacement D2) of the upper boundary surface of layer L4 at the position of is used. In the latter method, it is not necessary to accurately know the cutting angle GF, and a depth map Z(x, y; n) can be derived for each cross-sectional image slice with index n. In this example, since the cross-sectional image slices are assumed to be parallel to the x direction, the depth of each x coordinate is the same for a particular y coordinate, and therefore the depth map Z(x, y; n) is determined only by y, regardless of x. In another example, the depth Z(x, y; n) is different for each x coordinate, and therefore a depth map Z(x, y; n) is generated separately for each x coordinate in a similar manner. In the method of the third embodiment, a depth map Z(x, y) is generated for each of multiple N cross-sectional image slices, thereby enabling accurate depth and position determination of cross-sectional image features in multiple cross-sectional image slices obtained with a wedge cut shape. In one example of the method of the third embodiment, a depth map Z(x,y) of the cross-sectional image slice is generated from the lateral position of the second cross-sectional image feature within the cross-sectional image slice, thereby enabling accurate depth and position determination of the first cross-sectional image feature relative to the cross-sectional image slice obtained with the wedge cut shape.

[0092] FIG. 6 shows the nth and n+1th cross-sectional images obtained by cutting with the FIB beam 51 at a predetermined angle GF and imaging with the charged particle beam 44 (see FIG. 5) oriented in the z-direction, i.e., perpendicular to the wafer surface 55, at an angle GE=0°. Vertical HAR structures appear in the cross-sectional image slices as first cross-sectional image features (e.g., first cross-sectional image features 77.1, 77.2, and 77.3). Because the charged particle imaging beam 44 is oriented parallel to the HAR structures, the first cross-sectional image features representing, for example, ideal HAR structures will appear at the same y-coordinate. Deviations from the fixed position correspond to fabrication errors (e.g., tilt or "wobble" of the HAR structures). For example, the first cross-sectional image features of ideal HAR structures 77.1 and 77.2 are centered on line 80 at the same y-coordinate in the nth and n+1th image slices. Therefore, a first cross-sectional image feature (e.g., 77.1-77.3) of the HAR structure not only does not allow for the determination of the depth of the cross-sectional image feature within the wafer, but also the determination of the slice distance d and the slice angle GF. The cross-sectional image slice further includes multiple second cross-sectional image features of multiple layers, including, for example, layers L1-L5 (e.g., second cross-sectional image features 73.1 and 73.2 of layer L4). The layer structure appears in the cross-sectional image slice as high-contrast stripe segments along the x-direction. However, the positions of these stripes, i.e., the second cross-sectional image features of multiple layers (here, layers L1-L5 as shown), change with each cross-sectional image slice relative to the first cross-sectional image feature. As layers intersect the image plane with increasing depth, the positions of the second cross-sectional image features change in a predetermined manner from image slice n to image slice n+1. The top surface of layer L4, denoted by reference numerals 78.1 and 78.2, is displaced in the y direction by a distance D2, from which the slice distance d or slice angle GF is calculated, for example, by the formula d=D2·sin(GF). From the determination of the position of the second cross-sectional image feature (e.g., 78.1 and 78.2), two depth maps Z(x,y) of two consecutive cross-sectional image slices can be aligned in terms of the Z coordinate, and multiple depth maps Z(x,y) can be generated for multiple cross-sectional image slices. N (x,y) can be derived.

[0093] Therefore, by feature extraction of the second cross-sectional image feature, such as by edge detection or centroid calculation and image analysis, and by knowing the depth of the second cross-sectional image feature from prior information (e.g., design information), the lateral position and depth of the first cross-sectional image feature in the cross-sectional image slice can be determined with high accuracy. The depth of any boundary or surface of each layer (e.g., layers L1-L5) extending parallel to the wafer surface is typically known with very high accuracy and is constant over a larger area of ​​the wafer due to the planar fabrication techniques involved in wafer fabrication. In one example, the depth determination method is applied when the depth of the second cross-sectional image feature (e.g., a layer or top or bottom surface of a layer) is not known. Even if the exact depth value in nm is not known, a depth specification on a scale relative to the layer boundary or surface can be performed with high accuracy. In this example, the depth of the first cross-sectional image feature in the cross-sectional image slice is given relative to the depth of the boundary of the second cross-sectional image feature, such as the top and bottom surfaces of multiple layers. In other words, the depth of the first cross-sectional image feature is given relative to the depth of the second cross-sectional image feature outside of, at the same depth as, or on the depth scale of, the plurality of second cross-sectional image features L1-LM of the M layers.

[0094] In a fourth embodiment of the present invention, a method and structure for aligning cross-sectional image slices are provided. FIG. 7 illustrates several other aspects of a method for 3D volume inspection in a wafer. The inspection volume 13 of the measurement site 6.1 is marked with, for example, an alignment mark 24. After the wafer including the measurement site 6.1 is positioned at the intersection of the FIB beam 51 and the charged particle imaging beam 44, a wedge above the first cross-sectional surface 52 is first removed, and the charged particle imaging beam 44 is scanned to acquire a first cross-sectional image slice. After imaging the first cross-sectional surface 52, the FIB cut is repeated at an angle GF relative to the wafer surface 55, and multiple cross-sectional surfaces, including cross-sectional surfaces 52, 53, and 54, are imaged by the charged particle imaging beam 44, and multiple cross-sectional image slices are generated using the slice imaging technique described above. In one example, the wedge-cut slice imaging technique includes additional alignment features used to align the multiple cross-sectional images with each other. Prior to cutting the cross-sectional surface, additional alignment features, such as alignment marks 20.1, 20.2, and 20.3, are created on wafer surface 55 by deposition of layers and structuring with alignment patterns. An example of an alignment pattern is a cross, as shown in alignment feature 20.1, but any alignment pattern known in the art is equally applicable. The alignment feature can also be an alignment trench or edge 22 cut relative to the boundary of the wedge cut, for example, in a portion of measurement site 6.1 that is not altered by the FIB during cutting of a slice through inspection volume 13. The alignment feature 22 is not limited to an edge and can also be, for example, a tooth-like feature. The alignment feature can also be an integrated circuit feature exposed on the wafer surface near the inspection volume, such as integrated circuit structure 25. The alignment feature described above forms a common cross-sectional image feature in multiple cross-sectional image slices, enabling an additional method for lateral alignment of the cross-sectional image slices.The coordinates of the first cross-sectional image features representative of the semiconductor structure of interest are obtained with higher accuracy, and the depth map Z(x,y) is derived with higher accuracy from the Y coordinates of the representative second cross-sectional image features.

[0095] The alignment features are present in all cross-sectional image slices as common cross-sectional image features and can be analyzed for each individual cross-sectional image slice using conventional edge detection techniques (e.g., gradient-based contour extraction, etc.). It is known that edge positions can be determined most accurately in the direction perpendicular to the edge. In one example, the accuracy of edge localization is improved by additional features (e.g., at least two alignment features). In one example, additional alignment features 20.1, 20.2, and 20.3 are used for lateral image alignment of the cross-sectional image slices. The alignment features 20 and 22 enable a 3D volume to be reconstructed from multiple tilted cross-sectional image slices without relying on the shape of the semiconductor inside the inspection volume. In one example, the above-mentioned additional alignment features are configured for high-contrast imaging with a charged particle imaging beam. For example, high contrast and good visibility of the alignment pattern or edge can be achieved by locally coating the wafer surface at the measurement site with a material that exhibits high material contrast with the silicon material of the wafer sample prior to patterning or cutting a wedge onto the wafer surface. Suitable coating materials are, for example, platinum (Pt), carbon (C), or a combination thereof.

[0096] Alternatively or additionally to creating alignment marks, visible integrated circuit features present on the surface of the wafer can be used as additional alignment marks. For example, HAR structures present on the wafer surface near the inspection volume form visible structures 25, which are visible in cross-sectional image slices similar to alignment marks 20.1, 20.3, and 24. In one example, visible integrated circuit structures 25 are used as additional alignment features for lateral image alignment of cross-sectional image slices. For example, processing the surface of the wafer near the inspection volume exposes integrated circuit structures 25, making the exposed integrated circuit structures 25 visible to a charged particle imaging column and available for alignment using common cross-sectional image features.

[0097] The fifth embodiment describes another aspect of the present invention. The cross-sectional surface cut by the FIB beam 51 is imaged by the charged particle imaging beam 44 through scan imaging. In one example, the cross-sectional surface may not be perpendicular to the charged particle imaging beam 44, and the field of view (FOV) of a single scan may be too small to capture an image in a single scan, or the cross-sectional surface may be outside the depth of focus of the charged particle imaging beam 44. In such cases, at least two scan image segments are obtained by the charged particle imaging beam 44, and these at least two image segments are stitched together. An example is shown in FIG. 8. After cutting and exposing the surface 53 with the FIB beam (not shown), a first image segment 26.1 is acquired at a first position through scan imaging with the charged particle imaging beam 44. The first image segment 26.1 includes, for example, an alignment feature 20.1 on the wafer surface 55. Then, as the wafer mounted on a wafer stage (not shown) is laterally displaced, image segments (e.g., image segments 26.2 and 26.3) are acquired by scanning the charged particle beam 44.1 at the second or third position. The image segments 26.1, 26.2, 26.3, etc., overlap each other, allowing mutual alignment of at least each pair of image segments through the use of cross-sectional image features of structures present in the overlapping areas of adjacent image segments. "Top-down" imaging of tilted cross-sectional surfaces requires a constant depth of focus (DOF) of the vertical charged particle imaging beam. If the depth of focus of the imaging beam 44 is insufficient (e.g., if the DOF of the imaging beam 44 is smaller than the LZ), the FOV of the scanning imaging beam 44 is divided into smaller subfields of view, which are sequentially imaged by progressive focus adjustment. Generally, the depth of focus of the charged particle imaging beam is constrained by resolution requirements. High-resolution imaging with a resolution of less than 2 nm (e.g., 1 nm) results in a shallow depth of focus, e.g., in a scanning electron microscope (SEM), the depth of focus at 1 nm resolution is less than 50 nm.As will be explained in more detail below, for inspection volumes with large depth extensions LZ, a helium ion microscope is preferably applied as an alternative, which can provide a much larger depth of focus of up to approximately 10 μm with a high resolution of approximately 1 nm.

[0098] In one example of an LZ where the required resolution exceeds the DOF of the charged particle imaging beam by less than 2 nm (e.g., 1 nm), the imaging plane of the charged particle imaging beam 44.2 is shifted deeper into the wafer to acquire image segments 28. The extension of the image segments 28 is selected so that the depth extension 30 of the image segments 28 relative to the cross-sectional image surface 53 is less than the depth of focus (DOF) of the charged particle imaging beam 44 at the desired resolution. Note that, for example, SEMs also allow dynamic focus adjustment during an image scan, but this is limited to a z-position range of approximately 1 μm to 2 μm, requiring the focal position to be changed, e.g., by moving the wafer stage, such that the depth extension 30 of the image segments 28 is less than the dynamic focusing range of the charged particle imaging beam 44. By adjusting the z-position, i.e., the focal position, e.g., by moving the wafer stage in the z-direction, the desired resolution can be maintained throughout the imaging of the tilted cross-sectional surface 53. It is also possible to create a second, smaller trench near the inspection volume and create at least one additional alignment feature at the bottom of the second trench. This provides multiple alignment marks at a given depth, overcoming the DOF limitation of the imaging beam 44. In one example, as shown in FIG. 9, additional alignment features 32.1 and 32.2 are located at three depth levels. For example, alignment feature 32.2 is fabricated at depth level DLZ (reference number 34) to align image segments at depth level DLZ. To this end, a small trench at the desired depth DLZ is cut relative to the wafer surface, and an alignment pattern is structured at depth level DLZ. In one example, alignment features 32.1 or 32.2 at different depth levels are integrated semiconductor integrated circuit features near the inspection volume. Alignment features 32.1 or 32.2 at different depth levels form common image features in a series of cross-sectional image slices, thereby enabling highly accurate 3D volume imaging through accurate alignment of the series of cross-sectional image slices.In one example, the image stitching method and the image alignment method using additional alignment features at different depths are applied in cooperation with a dual beam apparatus in which the FIB columns are positioned at an angle of 8° to 45° and an SEM as a charged particle beam imaging system. In one example, the image alignment method using additional alignment features at different depths is applied to the inspection of a deep inspection volume extending below the wafer surface to a depth of more than 5 μm (e.g., 6 μm or 10 μm or more).

[0099] FIG. 26 shows another example of additional alignment features according to the fourth embodiment. In this example, the angle GFE between the charged particle imaging beam 44 and the FIB 51, which slices multiple cross-sectional surfaces, is GFE=90°. The angle GFE is selected to be approximately 30°, but can generally be between 25° and 60°. The plane formed by the FIB 51 and the charged particle imaging column 44 is perpendicular to the wafer surface 55. As a result, the imaging beam is perpendicular to the imaging cross-sectional surfaces (slices). For simplicity, only a single cross-sectional surface 52 is shown. This setup allows the imaging beam 44 to remain "focused" across the entire cross-sectional surface. This is especially important for electron beams as the charged particle imaging column, which have a limited depth of focus of approximately 200 nm. The large GFE angle GFE=90° between the FIB 51 and the imaging column 44 reduces the possibility of volume collisions between the two columns.

[0100] For particularly deep samples (>10 μm), a steep cutting angle (e.g., >60°) can cover the entire depth range, keeping the depth of the cross-section surface 52 within the field of view (FOV) of a typical single-beam charged particle microscope at approximately 10 μm to 15 μm. In contrast, shallower angles (e.g., angles GF below 20°) significantly expand the field of view, necessitating image stitching. Thus, in one example, a tilt angle GF between 25° and 60° is preferred.

[0101] Alignment marks 10.1 or 20.2 can be used to initially align the cross-sectional image slices in a direction perpendicular to the charged particle imaging beam (i.e., parallel to the slices). In the example of FIG. 26, alignment marks 20.1 and 20.2 are formed on a first cross-section 96 formed by FIB 44, such as cross-section 52, extending in the x-direction longer than multiple second cross-sectional image slices below the first cross-section 96. This configuration results in at least one parallel surface segment 23.1 as part of the first cross-section surface 96 remaining on one side of at least the second cross-section surface (such as cross-section surface 52). Alignment feature 20.1 can be formed in parallel surface segment 23.1, for example, by deposition and etching. A second parallel surface segment 23.2 can also be formed, including another alignment feature 20.2.

[0102] Due to various factors, such as image distortion or focal plane changes from first cross-sectional surface 96 to second cross-sectional surface 52, the alignment of multiple cross-sectional surfaces using additional alignment features such as features 20.1 and 20.2 according to the above-described alignment method is typically too coarse for 3D volume image reconstruction with a lateral resolution of less than a few nm (e.g., less than 5 nm, 3 nm, or 2 nm). Therefore, according to a fourth embodiment of the present invention, an additional method of precise alignment based on wafer structures or semiconductor features inside the inspection volume is applied. This precise alignment method also includes calculating the actual slice thickness d. The first coarse alignment step avoids alignment errors of repeating patterns and provides a mapping or alignment of cross-sectional image features in the cross-sectional image slices. The second fine alignment step achieves a mutual position accuracy of each pair of cross-sectional image slices to an accuracy of less than 5 nm, 3 nm, or 2 nm.

[0103] Integrated circuit structures inside a wafer typically include vertical structures (such as 3D memory channels or vias 4.1 and 4.2 in FIG. 26) and horizontal structures (such as "word lines" or metal lines L1-Ln). As mentioned above, both types of structures can be used for fine alignment and slice thickness determination. An example is shown in FIG. 27, which shows a cross section of a wafer according to the wedge cut method. The horizontal and vertical structures are indicated by a horizontal edge 80 of a layer and a vertical edge 78, for example, of a HAR structure (e.g., element 4.1 or 4.2 in FIG. 26). The Y' components of the positions of the horizontal edge 80 and vertical edge 78 in the cross-sectional image slice with index n+1 are compared with the Y' components of the positions of the horizontal edge 80 and vertical edge 78 in the cross-sectional image slice with index n, and the difference is called ΔY'. WL ("WL" = "Word-Line") and ΔY' Ch ("Ch" = "Channel"), where the Y' axis lies in the plane of the cross-sectional image slice. Imperfect registration will result in an additional displacement ΔY', which is equal for all structures in the cross-sectional image slice. Therefore, the actual displacements of the horizontal and vertical structures mentioned above are:

[0104]

number

number

[0105] In this way, the measured ΔY' WL and ΔY' Ch and the inclination angle GF,

number

number

[0106] The multiple cross-sectional image slices typically do not form perfectly parallel planes, but have some topography, as shown in FIG. 17, for example. To take this topography into account, the above equation can be solved locally at multiple locations of the cross-sectional image slices. Thus, the slice thickness d can be calculated separately for multiple locations of each pair of cross-sectional image slices. Also, the lateral displacements ΔX and ΔY′ can be calculated locally, and possible image distortions can be taken into account.

[0107] The coarse and fine registration methods described above result in accurate registration of multiple cross-sectional image slices, enabling high-precision 3D volumetric image reconstruction to within 5 nm, 3 nm, or even 2 nm.

[0108] In a method for generating a 3D volume image of an inspection volume at a measurement location 6.1 on a wafer, a plurality of subsequent cross-sectional surfaces are cut at a distance d by the FIB beam 51, which is set at a predetermined angle GF relative to the wafer surface. In the coordinate system of the above example, the FIB beam 51 is scanned in the x direction, tilting the cross-sectional surfaces at an angle approximately equal to GF. "Approximately equal" means that due to FIB beam divergence, the actual angle of the cross-sectional surfaces may deviate from the cutting angle GF by a few degrees, between 1° and 4°. In one example, generating multiple subsequent cross-sectional surfaces at a distance d less than 15 nm, less than 5 nm, or even less can be achieved by lateral displacement of the wafer stage in the y or z direction. In a sixth embodiment of the present invention, multiple subsequent cross-sectional surfaces are cut by the FIB beam 51 at a distance d of approximately 5 nm without lateral or vertical movement of the wafer stage. As shown in FIG. 10 , the FIB beam is scanned in the x- and z-directions around the FIB beam coincidence point 58, while the wafer is not displaced by the wafer stage. The fan-shaped tomography technique applied to this wedge-cut geometry allows for precise and fast adjustment of the slice distance d of subsequent cross-sectional image slices. The cutting angle is not constant at a given angle GF, but varies by the cutting angle difference 64 between the minimum and maximum cutting angles around the given angle GF (the angular spread GZ is approximately 1 milliradian). As a result, the distance between adjacent image slices also varies. As an example, the cutting can be set so that the slice distance 60.1 between two adjacent cross-sectional image surfaces (e.g., cross-sectional image surface 53 and cross-sectional image surface 54) is constant at the wafer surface 55. The slice distance 62.1 at the bottom of the inspection volume varies from slice to slice depending on the different slice angles. For example, a first cross-sectional image surface 53 is cut by the FIB beam 51.1 at a first cutting angle 66 and a second cross-sectional image surface 54 is cut by the FIB beam 51.2 at a second cutting angle 68 so that the distance between two adjacent cross-sectional image surfaces in the wafer plane is dy1 (reference number 60.1) and the distance dy2 (reference number 62.1) between two adjacent cross-sectional image surfaces at a depth LZ below the wafer plane is greater than dy1.This allows for maintaining precise control of the mutual distance of the cross-sectional surfaces with nm precision, although the resulting cross-sectional slices are not parallel to each other, and the alignment and depth determination methods according to the embodiments of the present invention described above are still equally applicable.

[0109] In one example, the cutting angle GF is adjusted by utilizing a method for tilting the FIB beam with a scanning unit of the FIB column. In one example, the distance between subsequent cross-sectional surfaces is adjustable. In one example, the distance between multiple cross-sectional surfaces is adjusted to be different for at least a portion of the distance of the cross-sectional image surfaces. Adjusting the distance and angle GF globally or locally adjusts the throughput and resolution of the 3D volumetric image of the inspection volume to suit the needs of the wafer inspection operation.

[0110] One challenge is achieving uniform cutting across a large area of ​​multiple cross-sectional image surfaces. In a seventh embodiment, a configuration and method are disclosed that avoids the accumulation of large amounts of debris on tilted cross-sectional surfaces. As shown in FIG. 11 , at measurement site 6.1, a first proximal trench 92 is first cut into the wafer surface adjacent to inspection volume 13, followed by a second distal trench 94 into the wafer surface adjacent to inspection volume 13, such that inspection volume 13 is located between the first and second trenches 92 and 94. The wafer is then rotated 90°, and multiple cross-sectional image surfaces, including cross-sectional image surfaces 53 and 54, are cut into inspection volume 13 by FIB beam 51. For each cross-sectional surface, the FIB beam is scanned in the x-direction with an angular spread GX (reference numeral 90) and tilted in the z-direction within an angular spread DZ (reference numeral 64) by the scanning mechanism of the FIB column. A charged particle imaging beam (not shown) acquires multiple cross-sectional images in the vertical direction, optionally with imaging of alignment features (e.g., alignment feature 20.1). Debris generated during FIB cutting accumulates in a distal trench 94 located on the opposite side of the inspection volume 13 with respect to the FIB beam. The separation of the two trenches 92 and 94 is configured to accommodate the inspection volume 13. The width and depth of the proximal trench 92 are adjusted to achieve a slice angle approximately equal to the angle GF described above. The depth of the distal trench 94 is adjusted to achieve a slice angle approximately equal to the angle GF described above and maintain debris accumulation in the distal trench 94. This maintains high accuracy in slice imaging of the wedge cut geometry with minimal debris. In one example, three or more trenches with different orientations and separations are configured to enable the generation of multiple sub-volumes of desired size and orientation for 3D tomography. Alignment features, such as alignment feature 20.1, are located near the inspection volume 13. Similar to the example shown in FIG. 9, at least one alignment feature 20.1 can be created at depth DLZ.

[0111] In an eighth embodiment, imaging by a charged particle imaging device is degraded by image distortion. The method steps of the above-described embodiments and examples are equally applicable to charged particle imaging device configurations with an angle GE>0°. In one example, the optical axis of the charged particle imaging device is positioned at an angle GE relative to the normal to the wafer surface. In this case, the imaging coordinate system (X', Y') rotates by angle GE relative to the wafer surface, distorting the cross-sectional image in the y direction. The cross-sectional image slices are digitally converted by anamorphic image transformation, in which a first image magnification in the x direction differs from a second image magnification in the y direction. In one example, the scanning unit of the charged particle imaging beam device is degraded by distortion (e.g., keystone). The cross-sectional image slices are digitally converted by distortion compensation, which compensates for the image distortion. The results of the anamorphic image transformation and distortion compensation are shown in FIG. 12. The left side of FIG. 12 shows a cross-sectional image slice including image distortion (example keystone distortion) resulting from image acquisition at angle GE. The right side of FIG. 12 shows the cross-sectional image slice after distortion compensation, with the reference numbers from FIG.

[0112] In a ninth embodiment, an additional alignment feature is provided above the inspection volume. FIG. 13 illustrates determining the distance between a subsequent cross-sectional image slice tilted approximately with respect to a predetermined principal cutting angle GF and the top surface of a semiconductor integrated sample. The cross-sectional image slice is tilted at approximately angle GF relative to the top surface. The additional alignment feature can also be an edge of the cross-sectional surface with wafer surface 55, such as edge 76.1 (see FIG. 7) between surface 52 and wafer surface 55. Edge 76.1 is removed in the subsequent FIB cutting of a subsequent cross-sectional image surface, such as surface 53. However, other edges 76.3 and 76.4 provide a measure of the slice distance d. In one example, prior to cutting, alignment features 38.1 and 38.2 are provided on the top surface of the wafer at the location of inspection volume 13. Alignment features 38.1 and 38.2 are not parallel but are inclined relative to one another at an angle GV, and have a different distance in the x-direction at each edge of each cross-sectional image surface (e.g., distance 36.4 at edge 76.4 of cross-sectional image surface 54). From the predetermined shape and position of alignment features 38.1 and 38.2 and distance 36.4, the y-position of edge 76.4 can be derived with high precision. In another example, from a first cross-sectional image surface 53 to an adjacent second cross-sectional image surface 54, the distance between the two cross-sections made by alignment features 38.1 and 38.2 changes by a value dx from distance 36.3 to distance 36.4. At each cross-sectional surface, the distance between alignment features 38.1 and 38.2 is measured at the cross-sectional surface's corresponding edge with wafer surface 55. With the change in distance dx determined and angle GV known, the y-direction distance dy between subsequent edges of the cross-sectional images can be calculated as follows: dy=dx / (2 tanGV / 2)

[0113] To determine the distance between slices, consider the cutting angle GF: d=dy·sin(GF)=(dx·sin(GF)) / (2·tanGV / 2)

[0114] In the wedge-cut geometry, any error in dx itself or in the measurement of dx is limited by the factor sin(GF). Therefore, the slice thickness d can be derived using additional criteria 38.1 and 38.2, and the positional change in y-position between second cross-sectional image features in adjacent cross-sectional image slices can be derived with high accuracy. As a result, the depth of the second cross-sectional image feature and the depth map for each cross-sectional image slice can be derived with high accuracy. Furthermore, the depth of the first cross-sectional image feature can be derived with high accuracy from the depth or depth map of the second cross-sectional image feature.

[0115] A tenth embodiment of a method for 3D inspection of a 3D inspection volume in a wafer at least at a measurement site includes the steps shown in Figure 14. In step S1, a wafer is loaded onto a wafer support table on a wafer stage in a vacuum chamber of a dual beam apparatus. The dual beam apparatus includes a focused ion beam (FIB) column having a first optical axis disposed at an angle GF relative to a surface of the wafer support table or the wafer surface when loaded onto the wafer support table. The measurement tool further includes a charged particle imaging device (e.g., an SEM or a helium ion microscope) having a second optical axis that forms an angle GE with a normal to the surface of the wafer support table and that is disposed at an angle GFE relative to the FIB column, where the first and second optical axes form an intersection point.

[0116] The position of the inspection volume at a first measurement site on the wafer is determined, and the wafer is positioned by moving a wafer stage so that the first measurement site is located at an intersection of the FIB column and the charged particle imaging column. Optionally, the orientation of the wafer is further adjusted by rotating a wafer support table of the wafer stage.

[0117] In operation step S3, at least one alignment feature, such as an alignment mark, is fabricated near the measurement site. In optional step S3, a coating is deposited locally on the wafer surface at the measurement site to cover the inspection volume, create an alignment mark with high imaging contrast, and achieve high contrast between the sample surface and the cross-section that is not covered during cutting. The additional alignment features in FIG. 7 (e.g., alignment mark 20.1 or alignment mark 22) are configured so that at least one of the alignment features is visible in at least two cross-sectional image slices to form a common cross-sectional image feature for mutual alignment of the at least two cross-sectional image slices. In one example, the additional alignment features in FIG. 7 (e.g., alignment mark 20.1 or alignment mark 22) are configured so that at least one of the alignment features is visible in all cross-sectional image slices for mutual alignment of all cross-sectional image slices. In one example, alignment features are fabricated above the inspection volume and configured to determine the positions of multiple edges (including first and second edges) formed by the intersections of multiple cross-sectional surfaces (including first and second cross-sectional surfaces) with the wafer surface.

[0118] In step S5, a first wedge is removed from the top surface of the wafer by FIB cutting to expose a first cross-sectional surface for cross-sectional imaging, which is monitored, for example, by a charged particle imaging beam.

[0119] In step S7, a slice imaging method is performed in the wedge cut geometry to generate multiple cross-sectional images. The FIB cutting of new cross-sectional surfaces is monitored, for example, by a charged particle imaging beam, including controlling the distance d between subsequent image slices by monitoring the alignment feature created in step S3. A series or multiple N cross-sectional image slices, including at least a first cross-sectional image slice and a second cross-sectional image slice, are acquired in the inspection volume. Acquiring the first and second cross-sectional image slices includes subsequently exposing at least the first and second cross-sectional surfaces in the inspection volume by cutting the FIB column through the inspection volume at approximately angle GF, and imaging the at least the first and second cross-sectional surfaces with a charged particle imaging device to acquire the at least first and second cross-sectional image slices. Typically, the number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is approximately 1000 or greater.

[0120] In one example, the slice imaging method is performed in a fan tomography technique, which includes scanning a focused ion beam of a FIB column in a first direction by a scanning unit to expose a first cross-sectional surface within an examination volume; tilting the focused ion beam in a second direction perpendicular to the first direction by the scanning unit; and scanning the focused ion beam in the first direction by the scanning unit to expose a second cross-sectional surface within the examination volume, such that the multiple cross-sectional surfaces, including the first and second cross-sectional surfaces, form different angles with the wafer surface by approximately a tilt angle GF.

[0121] In step S8, the cross-sectional image features are detected and classified such that the cross-sectional image features are classified into first cross-sectional image features and second cross-sectional image features, and optionally into alignment features forming a common cross-sectional image feature. For detecting and classifying the cross-sectional image features, methods known in the art can be applied, such as edge detection, image comparison, feature extraction, object detection, or combinations thereof, also known as machine learning algorithms. Thereby, optionally, common cross-sectional image features are detected and classified in the first and second cross-sectional image slices.

[0122] In step S9, the multiple cross-sectional images are aligned with each other so that multiple first cross-sectional image features of vertical structures such as HAR structures or vias appear at approximately the same lateral coordinates (e.g., x- and y-coordinates). In one example, the multiple cross-sectional images are aligned using at least one additional alignment feature or a common cross-sectional image feature. The multiple cross-sectional images are aligned with each other using at least one of the previously generated alignment features so that multiple first cross-sectional image features of vertical structures such as HAR structures or vias appear at similar lateral coordinates. By using the at least one additional alignment feature, local lateral position errors of the first cross-sectional image features, such as cross-sections of individual HAR structures, are detected. By determining the cross-sectional position of the HAR structure relative to the alignment mark, a global wobble error of the HAR structure at a depth z perpendicular to the wafer surface is detected.

[0123] In one example, the lateral mutual image alignment includes subtraction or digital compensation of image distortion deviations between at least the first and second cross-sectional image slices.

[0124] In step S10, a depth map Z(x,y) is generated from the positions (x,y) of the second cross-sectional features in each of the cross-sectional image slices. In one example, the depth z of any (x,y) point in the cross-sectional image slice is derived from the lateral displacement of the second cross-sectional image feature. The second cross-sectional image feature corresponds to, for example, a layer structure of a semiconductor device oriented parallel to the wafer surface. In one example, the depth z of any (x,y) point in the cross-sectional image slice is extrapolated from the lateral positions of multiple second cross-sectional image features formed by multiple layers of a semiconductor integrated device whose layer surfaces are parallel to the wafer surface.

[0125] As a result, in step S10, a plurality of N depth maps Z(x, y; n) (index n=1···N) are generated for a plurality of N cross-sectional image slices.

[0126] In one example, determining depth and generating a depth map using second cross-sectional image features is combined with alignment features, such as additional alignment marks or fiducial-based alignment. A semiconductor integrated sample may include repetitive features, such as gates in a gate layer, that can result in ambiguities in image alignment of the second cross-sectional image features. Generally, coarse alignment using alignment features or fiducials formed on the semiconductor integrated sample can reduce ambiguities and increase the speed of fine image alignment of the second cross-sectional image features used for depth determination according to any of the embodiments of the present invention.

[0127] In step S11, a depth of the first cross-sectional image feature in at least one cross-sectional image slice is derived from the depth map. In one example, steps S10 and S11 constitute a single depth determination step of determining the depth of at least one first cross-sectional image feature from at least one second cross-sectional image feature.

[0128] In one example, the step of determining the depth of the first cross-sectional image feature includes determining a lateral difference between a first position of the second cross-sectional image feature in the cross-sectional image slice and a second position of the second cross-sectional image feature in the second cross-sectional image slice.

[0129] In one example, at least two second cross-sectional image features are determined in the first cross-sectional image slice, each of the second cross-sectional image features representing the semiconductor integrated structure at a different depth within the inspection volume, and determining the depth of the first cross-sectional image features includes determining lateral positions of the at least two second cross-sectional image features.

[0130] The depth determination of the first cross-sectional image feature by the second cross-sectional image feature involves statistical averaging and is therefore more robust to, for example, image noise. Due to the typically large number of layers (e.g., 5, 10, or up to 100), the cross-sectional image includes multiple second cross-sectional image features, and depth determination by these multiple second cross-sectional image features improves the statistical accuracy of the depth determination. Since defects occur rarely, they do not affect the overall quality of the depth determination method. Furthermore, potential defects can be detected as outliers from the expected depth according to statistical evaluation.

[0131] In the above-described embodiments and examples, the cross-sectional surfaces are considered to be flat surfaces exposed by FIB cutting and polishing. In another example, at least one cross-sectional surface is curved due to errors or imperfections in the cutting and polishing process. As a result, the corresponding cross-sectional image slices contain distortions and artifacts due to being images of a curved surface. In one example, cross-sectional image features are used to determine the curvature of the cross-sectional surface and apply the determined surface curvature to correct for distortions in the cross-sectional image slices.

[0132] In step S12, a 3D volumetric image of the inspection volume is generated, including 3D position information of the first cross-sectional image features. In step S12 of generating the 3D volumetric image of the inspection volume, depth information of the multiple first cross-sectional image features from the multiple depth maps of the multiple cross-sectional image slices is combined. A 3D volumetric image of the inspection volume is constructed from the depth maps in conjunction with the multiple first cross-sectional image features of the multiple cross-sectional image slices. Also, integrated circuit features corresponding to the first cross-sectional image features are derived to form the 3D volumetric image of the corresponding integrated circuit features. In another step, defects or deviations of the corresponding integrated circuit features or characteristics of the corresponding integrated circuit features in the 3D volumetric image are derived.

[0133] In one example, the second cross-sectional image feature is analyzed, and the slope of the local cross-sectional surface can be determined from local variations in the width of the second cross-sectional image feature, which represents the structure in the layer. In one example, the local cross-sectional surface information is used for iterative cutting in a feedback loop correction. In one example, the digital image transformation is achieved by transferring a distorted image slice to an undistorted image slice obtained from the planar cross-sectional surface. In one example, the cross-sectional surface is curved in the x and y directions. As with the above embodiment, the position and edges of the second cross-sectional image feature are used for distortion correction and extraction of the corresponding z coordinate (including the z position of the first cross-sectional image feature) of each pixel of the corresponding cross-sectional image slice.

[0134] In one aspect of the present invention, a comparison of detection results at different measurement locations on a wafer is performed, e.g., by slice imaging in a wedge cut shape at different measurement locations to derive integrated circuit features and characteristics in 3D. The integrated circuit features and characteristics can be, for example, tilt and wobble errors of a HAR structure. At different measurement locations, the above-described method utilizes a second cross-sectional image feature to determine the depth of the integrated circuit feature represented by the first cross-sectional image feature, enabling accurate comparison, including comparison of the depth of the integrated circuit features and characteristics. The second cross-sectional image feature representing a line or layer parallel to the wafer surface constitutes a common reference for determining depth at different measurement locations. In this way, the tilt and wobble of a vertical HAR structure at different positions or measurement locations on the wafer can be determined in 3D without the need for additional common alignment references.

[0135] In an eleventh embodiment, a wafer defect inspection apparatus is provided that is configured to inspect an inspection volume in a wafer without extracting a sample from the wafer. Some aspects of the wafer defect inspection apparatus have been described with reference to FIG. 2 . The wafer defect inspection apparatus includes a focused ion beam column 50 configured to cut and expose at least first and second cross-sectional image surfaces in the inspection volume in the wafer 8, and a charged particle imager 40 configured to image the cross-sectional surfaces, including at least the first and second cross-sectional surfaces, to form a plurality of cross-sectional image slices, including the first and second cross-sectional image slices. The wafer defect inspection apparatus further includes an operational unit 2 that triggers positioning and alignment of a measurement feature 6.1 of the wafer 8 at the inspection point, for example, by moving a wafer stage via the control unit 19 and the stage control unit 16. The operational unit 2 has software operating code installed therein, and during operation, triggers inspection of the inspection volume by the slice imaging method in a wedge cut shape, including repeated FIB cutting, image acquisition, and stage movement operations. The operation unit is a control unit 19 that controls the charged particle beam imaging column 40 and the FIB 50, and communicates with the control unit 19 connected to a control unit 16 that controls the position of a wafer mounted on a wafer support table via a wafer stage (not shown). Another embodiment of a wafer defect inspection apparatus is shown in FIG. 15. In one example, the control unit 19 does not communicate directly with the charged particle beam column, but communicates with a charged particle beam control unit 86 that controls the operation of the charged particle beam column, such as the charged particle imager 40, or the FIB 50. The operation unit 2 includes an image processing unit 82 installed with a computer program and software program code that receives image information from the charged particle detector 17 via the control unit 19 and is synchronized with the scanning operation of the charged particle imaging beam scanned by a scanning unit (not shown) of the charged particle imager 40.The image processing unit 82 and the software program code installed therein are configured, in use, to determine first and second cross-sectional image features in at least the first and second cross-sectional image slices by feature detection and classification as described above in step S8. The image processing unit 82 and the software program code installed therein are further configured, in use, to determine depths of the first and second cross-sectional image features within the examination volume, the first and second cross-sectional image features being cross-sections of a semiconductor integrated structure inside the examination volume. The image processing unit 82 and the software code implement image processing methods as described above, such as corner or edge detection, thresholding, morphological operations, or similar operations, which are well known in the art. Image processing has recently been improved by increasing the speed of operations, for example, through the use of computer clusters with hundreds of processors in the image processing unit 82. Additionally, image processing methods for extracting features or structures of a semiconductor integrated sample can include or be replaced by machine learning algorithms.

[0136] In one example, the semiconductor integrated structure is a commonly known structure. The use of design information or 3D CAD information can improve the edge extraction of metal lines and HAR channels, the extraction of layer surface positions, and depth determination with high precision. For example, the use of CAD information can identify the end positions of metal lines so that they are not visible in the cross-sectional image. This reduces outliers in the image processing method. Furthermore, the depth of the second cross-sectional image feature is typically known with high precision from the 3D CAD information. The image processing unit 82 and the software program code installed in the image processing unit 82 can, during use, determine the depth of the second cross-sectional image feature within the inspection volume by comparison with the 3D CAD information, and derive the depth of the first cross-sectional image feature, for example, according to the method described above.

[0137] The operational unit 2 further comprises a defect detection unit 84 configured, in use, to determine deviations from predetermined 3D characteristics, for example by comparing the first cross-sectional image features against 3D CAD data or library data of predetermined 3D characteristics of the semiconductor structure inside the inspection volume, and is capable of detecting defect candidates as outliers from a statistical evaluation.

[0138] The operation unit 2 further includes an interface and communication unit 88 in communication with an external input / output device (for example, a user operation terminal, a tablet, a database, a wafer handler, or a fabrication operation system).

[0139] To obtain information about an inspection volume inside the wafer, such as defects or deviations from the desired shape of semiconductor structures, a series of oblique cross-sectional surfaces are sequentially cut and imaged "slice by slice." After a layer of material of adjustable thickness is removed by FIB cutting, a new cross-sectional image slice is acquired with a vertical charged particle imaging beam. A number N of cross-sectional image slices are acquired in this manner to cover the inspection volume inside the wafer, generating 3D volumetric image data with a lateral pixel size and distance dz between successive image slices. The 3D volumetric image data is analyzed and used to reconstruct, for example, characteristics of a semiconductor memory stack inside the inspection volume. The above-described embodiments of the present invention result in reconstructions that include recovery of information about the relative positions and orientations, including depths, of individual cross-sectional image features of interest inside the inspection volume.

[0140] In a twelfth embodiment, 3D volumetric image data is generated by resampling or interpolating a series of N cross-sectional image slices from oblique cross-sectional surfaces relative to a conventional raster. The interpolation method described below reduces interpolation artifacts in generating a 3D volumetric image from a series of N cross-sectional image slices. An embodiment of the 3D volumetric image data generation includes a method for acquiring at least one 2D virtual cross-sectional image from a set of cross-sectional image slices. Integrated semiconductor elements in a semiconductor device typically have a predetermined shape and orientation. They are typically located in layers parallel to the wafer surface or extend perpendicular to the wafer surface. Examples of such elements are memory channel or HAR structures, as well as layers containing metal lines, as described above and shown in FIG. 3. In a proposed method for high-speed 3D inspection, virtual cross-sectional image slices are generated in a plane perpendicular to the orientation of the semiconductor features of interest. An example is shown in FIG. 16. FIG. 16 shows three representative cross-sectional surfaces 52, 53, and 54 sequentially formed by FIB milling within an inspection volume (not shown) below the wafer surface 55 at approximately a glancing angle GF. From cross-sectional surfaces 52, 53, and 54, respectively, cross-sectional image slices with indices n-1, n, and n+1 are generated by a charged particle imaging microscope (e.g., scanning electron microscope beam 44 or HIM). The cross-sectional surfaces are cut by an ion beam (not shown) at an angle GF at a relative distance dz in the z direction, the z distance being controlled, for example, by lateral movement of a wafer stage (not shown) or by a scanning offset of the FIB in the z direction, as described above with respect to the fan-shaped tomography technique (see FIG. 10). In the present example shown in FIG. 16, a first distance dz in the z direction between first cross-sectional surface 52 and second cross-sectional surface 53 is n is the second distance dz in the z direction between the second cross-sectional surface 53 and the third cross-sectional surface 54 n+1 The distance dz is in the range of a few nm (e.g., 5 nm, 6 nm, or even more (e.g., 10 nm)). For example, dz n is approximately 6 nm, and dz n+1is approximately 7 nm. Each cross-sectional image is scanned by a lateral pixel raster in the xy plane, with a pixel raster of approximately 2 nm, 1 nm, or even less (e.g., 0.5 nm). Projections of the pixel raster in the y direction onto cross-sectional surfaces 52, 53, and 54 are shown by dashed line 123 (index 123 is shown only for surface 52), which extend in the x direction. Also shown is an example of a virtual cross-sectional image slice 121 perpendicular to the z axis. Virtual image slice 121 is parallel to wafer surface 55 at depth ZV. The pixel value of point C in virtual cross-sectional image slice 121 is obtained by projecting pixel values ​​of adjacent cross-sectional image slices in the direction of the structure of interest. In this example, the structure of interest is a HAR structure oriented perpendicular to the wafer surface. Therefore, the projection or interpolation direction is parallel to the z axis. Several example interpolation directions are indicated by arrows 125.1, 125.2, 125.3, and 125.4. In one example, the shortest distance between point C in the virtual image slice and an adjacent cross-sectional image surface is determined in the projection or interpolation direction. In this example, the first cross-sectional surface 52 is closest to point C, and the pixel value of pixel location C in the virtual cross-sectional image slice 121 is selected to be the same as the pixel value of pixel A in the first cross-sectional image slice of the first cross-sectional surface 52. Pixel A in the cross-sectional image n-1 of the first cross-sectional image surface 52 and pixel C in the virtual cross-sectional image slice 121 have the same x- and y-coordinates, and the distance AC in the z-direction is the shortest distance of all distances in the z-direction between the virtual image slice 121 and each of the multiple cross-sectional surfaces, including cross-sectional surfaces 52, 53, and 54.

[0141] In another example, the pixel value of point C in the virtual cross-sectional image slice 121 is interpolated using two adjacent cross-sectional surfaces (here, the first cross-sectional surface 52 and the second cross-sectional surface 53). A first pixel value A of the first cross-sectional image slice with index n-1 and a second pixel value B of the second cross-sectional image slice with index n are at the same absolute lateral coordinates of the first and second cross-sectional surfaces 52 and 53 and are used to calculate the pixel value of pixel C in the virtual cross-sectional image slice. The interpolation can be, for example, linear or weighted interpolation, and the probability of the pixel value of pixel C is calculated. Since the pixel value of pixel C is usually a binary result of the pixel value of the first pixel A or the second pixel B and is expected to represent the measurement result of a particular material, the interpolation can be combined with a thresholding operation.

[0142] In one example, three or more pixel values ​​of three or more cross-sectional image slices of adjacent corresponding cross-sectional surfaces can be used for interpolation (e.g., high-order polynomial interpolation). In this example, the pixel size on the virtual cross-sectional image slice 121 is selected to be equal to the pixel size of cross-sectional image slices n-1, n, and n+1 obtained from the multiple cross-sectional surfaces 52, 53, 54. For example, scanning the third cross-sectional surface 54 by the charged particle imaging microscope 44 (e.g., electron beam 44) is adjusted in the y direction to generate a third cross-sectional image slice with index n+1 relative to a common transverse coordinate system having the same absolute pixel coordinates as the first and second cross-sectional image slices with index n-1 and n. In an alternative example, the transverse pixel raster of at least one of the multiple cross-sectional image slices is digitally adjusted and resampled so that the virtual cross-section 121 is realized from at least two cross-sectional image slices by projection or interpolation along the z direction. In an alternative example, the direction of projection or interpolation between a first pixel location A on a first cross-sectional surface and a second pixel location B on a second cross-sectional surface is tilted relative to the projection direction (here, the z-direction), and the pixel value of pixel location C in the virtual image slice is interpolated accordingly.

[0143] The method of computing at least virtual cross-sectional image slices by interpolation in a predetermined projection direction (e.g., a z-direction perpendicular to the wafer surface 55 and parallel to the orientation of the first set of semiconductor features) suppresses interpolation artifacts and the virtual cross-sectional image accurately represents a cross-section through the first set of semiconductor features. The method of interpolation in a projection direction parallel to the orientation of the first semiconductor features is particularly interesting for analyzing HAR structures in a wafer, where cross-sectional images perpendicular to multiple HAR structures can be obtained by FIB milling.

[0144] For each virtual cross-sectional image pixel, a subset of at least one cross-sectional image slice is determined by calculating a distance from each of the series of N cross-sectional image slices in the first orientation direction to the virtual cross-sectional image pixel and selecting at least a first cross-sectional image slice having the shortest distance. In one example, a second cross-sectional image slice of the subset of at least one cross-sectional image slice is therefore selected as the cross-sectional image slice having the second shortest distance. Thus, other cross-sectional image slices of the subset of at least one cross-sectional image slice can be selected in order of increasing distance to the virtual cross-sectional image pixel in the first orientation direction.

[0145] In one example, a depth map ZV(x,y) of each virtual cross-sectional image is generated. The depth map is generated using a geometric configuration such that the virtual cross-sectional image is oriented perpendicular to the first orientation direction. For each pixel value of the virtual cross-sectional image, a subset of at least one cross-sectional image slice is selected by evaluating the distance Zrv(n)=Z(x,y;n)-ZV(x,y) and determining the slice index m by the minimum value for the distance Zrv(m). Thus, second and subsequent cross-sectional image slices of the subset of at least one cross-sectional image slice can be selected in order of increasing distance Zry(n) to the virtual cross-sectional image pixel in the Z direction.

[0146] In one example of a method for acquiring at least one 2D virtual cross-sectional image, a first orientation direction is the z direction perpendicular to the wafer surface, and virtual cross-sectional image slices are computed in a plane parallel to the wafer surface at a certain depth ZV below the wafer surface. For each virtual cross-sectional image pixel having a lateral coordinate (x, y), a subset of at least one m-th cross-sectional image slice is determined by selecting at least one cross-sectional image slice having a shortest distance to the plane of depth ZV such that distance Zrv(m) = Z(x, y; m) - ZV is a minimum value over all depth maps Z(x, y; n) (index n = 1...N). Thus, second and subsequent cross-sectional image slices of the subset of at least one cross-sectional image slice can be selected in order of increasing distance Zry(n) to the virtual cross-sectional image pixel in the Z direction.

[0147] In one example, a depth map ZV(x,y) of each virtual cross-sectional image is generated. The depth map is generated using a geometric configuration such that the virtual cross-sectional image is oriented perpendicular to the first orientation direction. For each pixel value of the virtual cross-sectional image, a subset of at least one cross-sectional image slice is selected by evaluating the distance Zrv(n)=Z(x,y;n)-ZV(x,y) and determining the slice index m by the minimum value for the distance Zrv(m). Thus, second and subsequent cross-sectional image slices of the subset of at least one cross-sectional image slice can be selected in order of increasing distance Zry(n) to the virtual cross-sectional image pixel in the Z direction.

[0148] In one example, pixel values ​​of the virtual cross-sectional image are interpolated in a first orientation direction from a subset of at least one cross-sectional image slice by geometric projection. An angle ω between the virtual cross-section and a cross-section corresponding to the multiplication of the m-th cross-sectional image slice is determined, and a projected lateral coordinate in the plane of the m-th cross-sectional image slice is calculated from the pixel coordinate of the virtual image slice by division by cos(ω). Then, pixel values ​​of pixels of the virtual image slice are calculated by interpolating the projected lateral coordinate from adjacent pixel values ​​of the m-th cross-sectional image slice.

[0149] In one example, the optical axis of a charged particle imaging system acquiring a series of N cross-sectional image slices is oriented perpendicular to the wafer surface such that the angle GE between the optical axis and a z-axis perpendicular to the wafer surface is GE=0°. In this example, the lateral coordinates (x, y) of at least one virtual cross-sectional image and the lateral coordinates of the series of N cross-sectional image slices can be identical, and the angle ω is approximately equal to the angle GF. This example is shown in FIG. 16. A method for acquiring at least one 2D virtual cross-sectional image may include forming at least one alignment feature near the inspection volume configured to form at least one common cross-sectional image feature for lateral mutual alignment of each of the cross-sectional image slices. The lateral mutual image alignment step may include subtracting image distortion deviations of each cross-sectional image slice.

[0150] In one example of a method for acquiring at least one 2D virtual cross-sectional image, the interpolation includes at least one of feature extraction, thresholding, contour interpolation, and model-based interpolation, which allows for highly accurate interpolation of the first semiconductor structure or feature of interest, and a third cross-sectional image feature representing the first semiconductor structure or feature oriented in the first orientation direction in the virtual cross-sectional image is interpolated with high accuracy and low interpolation artifacts using the first cross-sectional image features from a subset of at least one cross-sectional image slice.

[0151] The method for acquiring at least one 2D virtual cross-sectional image or a set of 2D virtual cross-sectional images computes a virtual cross-sectional image from a set of N cross-sectional image slices at an arbitrary depth ZV inside the inspection volume, where, for example, the depth ZV is selected according to the depth of a layer parallel to the wafer surface formed by a second semiconductor feature oriented in a second orientation direction parallel to the wafer surface. For example, a first virtual cross-sectional image slice is computed at a depth ZV1 between two adjacent identical layers such as word lines (e.g., layers L2 and L3 in FIG. 5), and a second virtual cross-sectional image slice is computed at a depth ZV2 inside the layer or word line (e.g., layer L4 in FIG. 5).

[0152] A depth map Z(x,y;n) (index n=1...N) for each cross-sectional image slice is computed from a second cross-sectional image feature in the cross-sectional image slice, the second cross-sectional image feature representing a cross-sectional image through a second semiconductor feature oriented in a second orientation direction parallel to the wafer surface. The projection or interpolation method relies on actual depth information for each pixel of the multiple cross-sectional image slices. The pixel-by-pixel depth information or depth map Z(x,y;n) for the cross-sectional image slice with index n is derived by the method described in the above embodiments and examples. In the example of analyzing a HAR structure, the depth Z(x,y;n) of each pixel location in the first cross-sectional image corresponding to the HAR structure is derived from a second cross-sectional image feature representing an elongated structure parallel to the wafer surface, such as multiple layers or word lines of a 3D memory device, as described above.

[0153] A set of virtual image slices, such as slice 121, reliably represents a cross-sectional image of a semiconductor structure or feature under investigation, such as a HAR structure, and can be used for statistical or defect analysis. One example of a method for acquiring a set of 2D virtual cross-sectional images provides high-speed data acquisition by selecting a distance dz in the z-direction, or a distance d perpendicular to each cross-sectional image slice, to be substantially larger than the pixel size of the corresponding cross-sectional image. The pixel size is typically less than 2 nm (e.g., 1 nm, 0.5 nm, or 0.25 nm). For high-speed data acquisition, the slice distance dz perpendicular to the wafer surface is selected or adjusted to be greater than 6 nm (e.g., 8 nm, 10 nm, or more). By selecting a slice thickness or distance dz substantially larger than the image pixel size, the number of FIB cutting operations is significantly reduced, enabling high-speed 3D image acquisition of an inspection volume within the wafer. The voxels of the 3D volume data have substantially different extensions along different axes, and the lateral pixel size or resolution differs from the depth pixel size or resolution. The method realizes a plurality of virtual cross-sectional image slices at a predetermined distance in the z-direction adapted to the lateral pixel spacing in the x- and y-directions, generating isotropic 3D voxels of a 3D volumetric image of the examination volume. Interpolation is performed in a first direction of the semiconductor feature of interest, such as a HAR structure, optionally combined with a thresholding operation or other contour interpolation method, thereby suppressing interpolation artifacts. The method computes a first cross-sectional image of the semiconductor element in the virtual image slice from first cross-sectional images of a subset of the plurality of cross-sectional image slices having a depth or z-direction distance greater than the lateral pixel spacing.

[0154] In one example, in a reduced subset of a plane orthogonal to the direction of the HAR structure, a predetermined set of virtual cross-sectional images such as a virtual cross-sectional image 121 including a first cross-sectional image of the HAR structure is derived. The virtual cross-sectional image slices are calculated for a small number of z-planes parallel to the wafer surface at a z-direction interval dz larger than, for example, a 1 nm lateral pixel raster. In one example, since the z-direction interval of a large number of virtual cross-sectional images is selected according to the number of layers parallel to the wafer surface, the processing time of the calculation is significantly reduced. Therefore, the z-direction distance between the reduced set of virtual cross-sectional surfaces is selected at, for example, 5 nm to 50 nm (for example, 10 nm or 25 nm). In one example, the z-direction distance can be variable according to the depth of the layer extending parallel to the wafer surface.

[0155] In one example of a method for obtaining a set of 2D virtual cross-sectional images, a plurality of N cross-sectional image slices are generated by scanning a plurality of N cross-sectional surfaces with a charged particle imaging microscope, and a set of F virtual cross-sectional image slices are calculated, where F < N. In one example, the maximum number of layers of the semiconductor memory stack is M (see FIG. 3), and F is selected to be equal to M such that virtual cross-sectional image slices are generated within each layer. In another example, F is selected to be equal to the number of word lines such that virtual cross-sectional image slices are generated within each word line layer and all insulating layers between the layers. In each example, the number of pixels in the lateral direction far exceeds the number N of cross-sectional image slices and the number F of virtual cross-sectional image slices.

[0156] High-precision milling with an FIB, including subsequent polishing of the multiple cross-sectional surfaces, generates multiple highly planar cross-sectional surfaces. However, precise milling, including polishing, is time-consuming. For high-throughput applications, such as wafer inspection, it is desirable to speed up 3D volumetric image acquisition of the inspection volume inside the wafer. In a thirteenth embodiment of the present invention, a high-speed milling operation is applied. After high-speed milling, the surface shape of each cross-sectional surface may deviate from a perfect plane due to, for example, FIB divergence, milling artifacts caused by imperfect FIB control, or the interaction between the FIB and the milled material inside the inspection volume. One example of such an effect is the well-known curtain effect. An improved example of the above-described interpolation method for generating virtual slices is shown in FIG. 17. The pixel value of point C in virtual slice 121 is obtained by interpolating the pixel values ​​of points A and / or B in cross-sectional image slices obtained from curved cross-sections 52, 53, and 54, and the z-coordinate of pixel Z(x,y;n) in the nth cross-sectional image slice is derived by the above-described method of the present invention. The depth or z-coordinate of an image pixel of a first cross-sectional image, e.g., representing a HAR structure, is derived from a known depth or relative to an image pixel of a second cross-sectional image, e.g., representing a word line oriented perpendicular to the HAR structure. In one example, a method for acquiring a virtual cross-sectional image or a series of virtual cross-sectional images, each including a plurality of virtual cross-sectional image pixels, from a set of cross-sectional image slices includes acquiring a series of N cross-sectional image slices by alternating imaging and cutting a series of N cross-sectional surfaces at a tilt angle GF relative to an inspection volume inside the wafer, wherein a pixel value for each virtual cross-sectional image pixel is calculated by projecting a subset of at least one cross-sectional image slice of the series of N cross-sectional image slices in a first orientation direction and interpolating the pixel values ​​from the projections of the subset of at least one cross-sectional image slice. The number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is about 1000 or greater.

[0157] In the above-described embodiments, the requirements for inspection of an inspection volume with a high resolution below 1 nm and a large depth extension LZ in the wafer are achieved by a slice imaging method with a wedge-cut geometry with an angle GF greater than 30°. As described in the fifth embodiment with reference to FIG. 8, the high-resolution requirement requires the SEM to acquire a large series of images at different focus positions. As shown in FIG. 12 with reference to the eighth embodiment, an example embodiment utilizing an SEM for imaging requires additional image processing, such as stitching and distortion compensation. In another example, the optical axis of the SEM is adjusted by an angle GE so that the angle GFE between the SEM and the FIB is approximately 90° (see FIGS. 2 and 4b). However, in embodiments and examples using an SEM, image acquisition and lateral alignment of multiple N cross-sectional image slices of the inspection volume with a large depth extension, e.g., greater than 1 μm, is time-consuming, requiring additional time for focus adjustment and image processing. Furthermore, accuracy is reduced. In a fourteenth embodiment, the charged particle imaging device is a helium ion microscope (HIM) with a depth of focus of up to 10 μm. The depth of focus (DOF) of a high-resolution SEM is approximately 10–20 nm. Generally, the DOF of a charged particle imaging microscope is limited by the required resolution. HIM maintains a resolution of less than 1 nm for large depths of field exceeding 6 μm and reaching 10 μm or more. Another advantage of HIM is that imaging is strongly influenced by secondary electrons and little influenced by backscattered charged particles. Therefore, imaging contrast is less affected by topography. Furthermore, with helium ions, wafer charging is always positive, minimizing contrast variations due to charging effects on the wafer surface.

[0158] Thus, the dual beam apparatus 1 with HIM according to FIG. 18 provides a preferred embodiment for inspection volumes with large depth extensions. In this fourteenth embodiment, a wafer inspection apparatus and method for inspecting three-dimensional circuit patterns in an inspection volume inside a wafer without removing a sample from the wafer are provided, which are configured and applicable to thick or deep layer stacks, for example, with a depth extension or height below the surface of the semiconductor wafer being processed of more than approximately 6 μm (e.g., 10 μm or 15 μm). As described above, the wedge-cut slice imaging technique avoids both wafer destruction and removal of the inspection volume. In the fourteenth embodiment, by setting the tilt angle GF of the FIB column for cutting relative to the wafer surface 55 to be greater than 20°, preferably 30°, and more preferably 35°, a cross-sectional surface lateral extension of less than 25 μm (e.g., a lateral dimension LY of approximately 20 μm or less) is achieved, and image acquisition is performed very quickly and efficiently without image stitching. The challenges of wafer inspection operations are high resolution requirements, e.g., 2 nm, 1 nm, or even 0.5 nm or less, and a required depth range within the wafer down to 15 μm directly below the wafer's top surface 55. According to a fourteenth embodiment, the high resolution and required depth of focus of the charged particle imaging beam are achieved by using a helium ion microscope (HIM). The helium ion microscope provides the required resolution of approximately 1 nm or less at the required large depth of focus, and is capable of capturing images of each cross-sectional surface of the wedge cut shape at the required resolution in a single image. "Single image scan" refers to an image capture scan without changing the focal plane, difficult focus control over a large depth range, or image stitching. In a single image scan, the wafer does not move, and the focal plane of the charged particle imaging beam does not change.

[0159] Dual-beam apparatus for wafer inspection using HIM provide high-resolution (sub-1 nm resolution) and high-throughput 3D volume inspection of inspection volumes, particularly deep inspection volumes with depths greater than 100 nm (e.g., LZ depths greater than 1 μm). In one example, inspection of memory devices with HAR structures extending approximately 6 μm in depth below the wafer's top surface is performed without destroying the wafer or removing the inspection volume from the wafer. HIM in certain configurations of dual-beam apparatus provides high-resolution imaging of multiple N cross-sectional surfaces in a depth range greater than 5 μm (e.g., 6 μm) and generation of multiple N cross-sectional image slices with sub-1 nm resolution.

[0160] Alternating between slicing multiple cross-sectional surfaces with the Ga-FIB 50 at approximately 36° and imaging with the HIM, slicing imaging is achieved down to the bottom of the HAR channel of the memory device. Details of the fourteenth embodiment are described in Figures 18a and 18b. Both figures show different aspects of a dual-beam apparatus 1 for inspecting deep volumes in a wafer. A helium ion microscope (HIM) 140 having an optical axis 142 is aligned perpendicular to a wafer support table (see Figure 18b) that holds the wafer with its top surface 55 perpendicular to the optical axis 142 of the HIM 140. The gallium FIB column 50 is oriented at a tilt angle GF of approximately 36° relative to the wafer top surface 55. Within an inspection volume with lateral extensions LX × LY × LZ of approximately 10 μm × 10 μm × 6 μm, the first and second cross-sectional surfaces of the plurality of N cross-sectional surfaces are denoted by indices n and n+1. The cross-sectional surface represents a cross-sectional feature, and in this example, the first cross-sectional feature of the HAR structure is oriented perpendicular to the wafer top surface 55. The first HAR cross-section 177.1 is shallow, close to the wafer surface 55. The second HAR cross-section 177.2 is deeper, e.g., with a depth LZ of approximately 6 μm or more. Both surface details are within the depth of field or depth of focus (DOF) of the HIM 140 (approximately 10 μm or more). Because the tilt angle is 36°, a depth LZ of 6 μm results in a lateral extension of the cross-sectional surface LY of only approximately 10 μm. In another example, a depth LZ of 15 μm results in a lateral extension LY of 20 μm. For such a deep structure with an LZ of approximately 15 μm, a large tilt angle, e.g., 45°, allows for a lateral image extension of, e.g., 15 μm. The slice distance is, for example, approximately 10 nm, or is selected according to sampling requirements; smaller slice distances can be selected for larger tilt angles. In either case, each cross-sectional image surface can be acquired by a single image scan with the HIM, without the need for focus changes, distortion compensation, or image stitching.

[0161] For alignment of the N number of cross-sectional image slices, alignment marks, such as alignment mark 148, are provided on wafer top surface 55. Both the alignment mark and the tilted cross-sectional surface of wafer top surface 55 are well within the DOF of the HIM, allowing for highly accurate lateral alignment of the N number of cross-sectional image slices. For example, the deeper second HAR cross section 177.2 and alignment feature 148 on wafer surface 55 are both intersected by fast scan line 152 of HIM 140, and both are within the depth of focus or depth of field (DOF) of HIM 140. An inherent advantage of the large depth of field of HIM 140 is that, together with alignment features or fiducials 148 fabricated on wafer surface 55, structures or features at different depths within a depth range of more than 1 μm (e.g., 2 μm, 5 μm, or even 10 μm) can be imaged.

[0162] A dual beam apparatus 1 (see FIG. 18A) for inspection of an inspection volume at an inspection site of a wedge-cut shaped wafer includes a six-axis wafer stage 155 having a wafer support table 15 configured to hold a wafer with a wafer surface 55 on a support surface 152, an FIB column 50 positioned at an angle GF relative to the support surface 152 of the wafer support table 15, and a helium ion microscope (HIM) 140 having an optical axis 142 positioned perpendicular to the support surface 152, the optical axes of the FIB column 50 and the HIM 140 forming an intersection 43, and a first beam stage 155 for holding a wafer at an angle GF relative to the support surface 152 of the wafer support table 15, the HIM 140 being configured to hold a wafer at an angle GF relative to the support surface 152 of the wafer support table 15, the first beam stage 155 being configured to hold a wafer at an angle GF relative to the support surface 152 of the wafer support table 15, the first beam stage 155 being configured to hold a wafer at an angle GF relative to the support surface 152 of the wafer support table 15, and a first beam stage 155 for holding a wafer at an angle GF relative to the support surface 152 of the wafer support table 15, the first beam stage 155 being configured to hold ... and a helium ion microscope (HIM) 140 having an optical axis 142 positioned perpendicular to the support surface 152, the optical axes of the FIB column 50 and the HIM 140 The system includes a stage control unit 16 configured to position the measurement site 6 at the intersection 43, and a control unit 19 for controlling the FIB column 50 and the HIM 140, and is configured to form a plurality of N high-resolution cross-sectional image slices with a resolution of less than 2 nm, preferably less than 1 nm, by alternating cutting a plurality of N cross-sectional surfaces in the inspection volume at approximately the angle GF with the FIB beam 51 and imaging each cross-sectional surface with the HIM beam 144, wherein the depth extension LZ of the inspection volume in the direction of the optical axis 142 of the HIM 140 is greater than 1 μm (e.g., 3 μm). In one example, the depth extension LZ in the direction of the optical axis 142 of the HIM 140 is greater than 5 μm (e.g., 6 μm or 10 μm). The imaging of each cross-sectional surface to form the plurality of N high-resolution cross-sectional image slices is achieved by the HIM beam 144 having a depth of focus (DOF) that exceeds the depth extension LZ. In use, the control unit 19 is configured to scan the HIM beam 144 over an area exceeding the lateral size LX or LY of the examination volume (at least approximately 5 μm to 10 μm (LX or LY)) and to collect multiple secondary electrons in a time series manner using the secondary electron detector 17.

[0163] In one example, the angle GF between the FIB beam 51 and the support surface 152 of the wafer support table 15 is between 30° and 60° (e.g., 36° or 45°). In one example, the dual beam apparatus 1 further comprises an operation unit 2 configured to mutually align the plurality of N cross-sectional image slices with respect to at least one common cross-sectional image feature. In one example, the control unit 19 is further configured to generate at least one alignment mark 148 on the wafer surface 55, which, in use, constitutes the at least one common cross-sectional image feature for mutual alignment of the plurality of N cross-sectional image slices. In one example, the operation unit 2 further comprises an image processing unit 82 (see FIG. 15 ) that, in use, computes at least one virtual cross-sectional image slice in a plane parallel to the wafer surface 55 of the inspection volume.

[0164] The large depth of focus and telecentric configuration of the HIM 140 allows cross-sectional image slices to be imaged with high precision and without distortion, without the need to change the focal plane of the charged particle imaging microscope. The large depth extension LZ (e.g., 6 μm or even up to 10 μm) of more than 1 μm allows efficient and rapid acquisition of 3D information of the inspection volume inside the wafer. For example, for memory devices with multiple HAR structures, virtual image slices can be generated for each alternating layer, such as a layer of word lines and an insulating layer between the word lines. HIM perpendicular to the wafer surface or parallel to the HAR structure allows cross-sectional images to be obtained from tilted cross-sectional surfaces, thereby reducing image processing time and minimizing interpolation artifacts. Optional alignment structures 148 formed on the wafer surface 55 are well within the depth of focus of the HIM and can be used to mutually align the cross-sectional image slices.

[0165] FIG. 19 illustrates some aspects of the fourteenth embodiment for an inspection volume with a large depth extension LZ. The inspection volume 160 has a large z-direction extension LZ of 2 μm to 10 μm. In the inspection volume for a memory device, multiple alternating layers 162 are arranged parallel to the wafer surface 55. Also, multiple HAR structures are arranged perpendicular to the wafer surface 55. Multiple N cross-sectional surfaces are formed in the inspection volume 160. Here, cross-sectional surfaces with indices n-1, n, and n+1 are shown. Each cross-sectional surface is cut with respect to the inspection volume at an approximate tilt angle GF of 30° to 60° (e.g., 36°). The distance between consecutive cross-sectional surfaces is selected to be less than 12 nm. Controlling the slice distance (e.g., reducing the distance and increasing the number of slices) can further improve the accuracy of 3D volumetric image acquisition and reduce image processing errors. Alternating with the cutting, each cross-sectional surface is imaged by the HIM 144 to form a plurality of N cross-sectional image slices (e.g., cross-sectional image slice index n in FIG. 19b). The cross-sectional image slices include a plurality of cross-sectional features of the HAR structure, such as 177.1, and a plurality of layer cross-sectional features, such as conductive layer cross-section 175. Each of the plurality of N cross-sectional image slices includes an image detail 150 representing an additional alignment feature 148 and is laterally aligned with the image detail 150 representing the additional alignment feature 148, which is not modified by subsequent cutting of the cross-sectional image surface. Computing the virtual cross-sections forms a virtual cross-sectional image 166.1 in the first conductive layer or word line and a virtual cross-sectional image 166.2 in the second insulating layer. In this manner, a plurality of virtual cross-sections are computed (FIGS. 19c and 19d).

[0166] As shown in Figure 20, multiple cross-sectional images or virtual cross sections are each analyzed by image processing to derive statistical characteristics of the HAR structure. Figure 20 shows two example analysis results. The left side shows the distribution of radii and mean radii for multiple HAR channels over a depth range of 0 μm to 4.5 μm. The right side shows the distribution of ellipticities and mean ellipticities for multiple HAR channels 164 over a depth range of 0 μm to 4.5 μm. Other parameters for multiple HAR channels 164 can be similarly obtained and illustrated.

[0167] FIG. 21 illustrates a fifteenth embodiment of the inspection of an inspection volume with a large depth extension. In this embodiment, the inspection time for an inspection volume with a large depth extension is reduced by dividing the inspection volume into multiple B blocks 181.1, 181.b, 181.B. For simplicity, three blocks are shown. The multiple blocks 181 are arranged diagonally along the cutting direction of the FIB beam. This reduces the number of cross-sectional surfaces to be cut and imaged, thereby shortening the inspection time. In each block 181, multiple (e.g., 70) HAR structures 164 are analyzed in a specific depth range lz1 to lzB. The block size is adjusted according to the cutting angle GF of the cross-sectional surface, the size of the HAR features, and the number of HAR cross sections 177 required for statistical analysis in each depth range. Therefore, the preferred slice distance d is less than 30 nm. Multiple virtual cross sections are calculated in different depth ranges of multiple layers 162. This reduces the number of cutting operations by more than a factor of 3, and therefore the operation time required for the inspection volume by more than a factor of 3. In one example, for a memory device including a memory HAR structure, it is sufficient to provide at least three cross-sectional image slices per HAR structure.

[0168] 22 shows a sixteenth embodiment of a dual beam apparatus for inspection of a deep, small inspection volume 191 with a large depth extension. In this example, the angle GF is selected to be a large angle between 45° and 80° (e.g., a cutting angle of approximately 76°) to slice the HAR features of the memory device at a large angle DF through the entire depth range. The axis of the charged particle microscope 195 is oriented at an angle GE (e.g., an angle of 40° or more) with respect to the normal to the wafer surface 55. In this example, wafer destruction is minimized.

[0169] Typically, wafer inspection for monitoring and controlling high-volume manufacturing (HVM) of semiconductor features on or in semiconductor wafers requires ultra-high speed and high throughput. In a seventeenth embodiment, development of a monitoring recipe using a dual-beam tool according to any of the other embodiments of the present invention is described. The seventeenth embodiment is illustrated in FIG. 23. Semiconductor device development proceeds through three major phases, M1 to M3. In the research and development (RnD) phase M1, a functional prototype of a new semiconductor device is developed. In step M1.1, a semiconductor device design is developed. In step M1.2, a functional demonstrator of the new semiconductor device or new design features is fabricated, and their functionality is tested in step M1.3. In phase M1, new design features and new process steps are tested. A 3D volume analysis of the demonstrator is performed, for example, according to any of the embodiments described above or by a method using sample extraction as described in the prior art. This determines and provides a list of critical design parameters and key design performance indicators D1.

[0170] In Phase M2, the fabrication process is enhanced. Step M2.1 involves the development of a new fabrication process or the improvement of an existing one. Step M2.2 involves comprehensive 3D analysis or 3D volume imaging, for example, by any of the methods or apparatus described in the above embodiments. For example, a new set of representative defect patterns is obtained, and from these representative defect patterns, a list of critical fabrication performance indicators D2 is generated. If fabrication in Phase M2 is found to be critical, the process can return to Phase M1 and design changes can be made.

[0171] In both phases M1 and M2, comprehensive 3D measurements are performed to understand the process and identify possible defects throughout the depth direction. Step M2.2 monitors multiple fabrication steps and performs multiple 3D measurements, with the measurement time in step M2.2 advantageously being shorter than, for example, step M1.3. A preferred method for 3D inspection in phases M1 or M2 is 3D volume imaging in a wedge cut geometry, as described in one of the above embodiments. In the "wedge cut" approach, a 3D memory stack is cut by FIB at an angle of less than 80° relative to the wafer surface, as shown in FIG. 3. In the example of a memory device, the sloped cross-sectional surface 52 of the trench includes both the footprints of the memory channel and the word line. The cross-sectional surface is imaged "from above" by using a vertical charged particle imaging beam (e.g., helium ion or electron beam). The corresponding cross-sectional image slice includes a projection of the channel footprint at each depth (Z) on the XY plane. For a cylindrically shaped vertical memory channel, the cross-sectional image slices contain circular footprints of varying depths determined by the footprint's location on the inclined surface 52 of the trench, as shown, for example, in FIG. 6 . The angle GF between the FIB beam 51 and the wafer surface (xy plane) controls the depth range included in the imaging beam's field of view and the total lateral area available for imaging. "Top-down" imaging of an inclined cross-sectional surface requires a constant depth of focus (DOF) of the vertical beam. If the DOF is insufficient, the FOV can be divided into smaller sub-fields of view, which can be sequentially imaged using progressive focus adjustment, as described for the fifth embodiment in the example of FIG. 8 . Alternatively, the HIM according to the fourteenth embodiment of FIG. 18 can be employed. From the measurement results, multiple performance parameters and statistical analyses can be performed, for example, virtual cross-sectional image slices can be generated, and evaluations can be performed as described in FIG. 20 .

[0172] Phase M3 is the high-volume manufacturing (HVM) of semiconductor devices. In the HVM phase, measurement time is reduced to maximize throughput. In step M3.1, the fabrication process developed in phase M2 is applied, and in step M3.2, a limited number of metrics or monitoring measurements are routinely performed. Based on the key design performance indicators D1 and key fabrication performance indicators D2 derived in the RnD phase M1 and phase M2 described above, a measurement or monitoring recipe R is defined, and then routinely executed to monitor the fabrication process in phase M3.

[0173] From the key design performance indicators D1 and key fabrication performance indicators D2, a measurement or monitoring recipe R for high-speed, high-throughput monitoring of the HVM phase M3 is generated. The key design performance indicators D1 and key fabrication performance indicators D2 may include a large amount of data, including 2D and 3D image data, 2D virtual image slices, and dimensional and statistical analysis of the above data for semiconductor devices fabricated with ideal performance and semiconductor devices fabricated with inferior performance due to typical fabrication errors. The key design performance indicators D1 and key fabrication performance indicators D2 may also include analysis data of other semiconductor devices provided from simulations (e.g., CAD simulations) or databases. Representative defect patterns in the inspection volume at specific inspection sites and specific measurement operations are derived and stored in the measurement recipe R. For example, volume effects in the inspection volume can be represented, typically by one or two representative cross-sectional image slices through the inspection volume. From the 3D volume image acquired by the slice imaging method described above, a preferred angle and orientation of at least one representative cross-sectional image slice is determined, thereby determining a preferred angle and orientation of a cross-sectional surface at a specific inspection site on the wafer. The preferred angle and orientation of the cross-sectional surface at a particular inspection site is an example of prior information that utilizes prior knowledge about the semiconductor structures and features in the inspection volume.

[0174] The HVM measurement recipe R is derived, for example, by machine learning or deep learning applied to key design performance indicators D1 and key fabrication performance indicators D2, and includes a set of monitoring or HVM performance indicators D3 in the monitoring recipe R to define a set of representative reduced measurement results for the HVM in phase M3. The measurement or monitoring recipe R is applied to the HVM in step M3 to generate multiple actual values ​​of the monitoring indicators D3. The actual values ​​of the HVM performance indicators D3 may include a large amount of data, including 2D and 3D image data, as well as measurement dimensions and statistical analysis of the above data for semiconductor devices fabricated by HVM with ideal performance and semiconductor devices fabricated with inferior performance due to HVM fabrication errors. The multiple actual values ​​of the HVM performance indicators D3 can be fed back into the measurement or monitoring recipe generation in step R, which can be confirmed or realized at regular time intervals.

[0175] Thus, a seventeenth embodiment of a metrology or monitoring recipe generation method includes generating a first set of critical design performance indicators (D1) including CAD image data and dimensions of critical design features in an inspection volume; generating a second set of critical fabrication performance indicators (D2) including 3D volumetric image data acquired during development of the fabrication process; and deriving a metrology recipe (R) including acquiring at least one of a third set of performance indicators (D3), wherein generating the second set of critical fabrication performance indicators (D2) or the monitoring recipe (R) includes acquiring at least a first cross-sectional image slice of at least a first cross-sectional surface through a representative predetermined inspection volume having a depth extension (LZ) > 1 μm below the wafer surface. At least one of the steps of acquiring at least one of the performance indicators (D1-D3) includes extracting, by image processing, a plurality of first cross-sectional image features representing cross sections of a plurality of first semiconductor features at a tilt angle (GF), wherein the image processing includes at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation. At least one of the steps of acquiring at least one of the performance indicators D1-D3 further includes calculating at least one descriptive parameter of the at least one first semiconductor feature from at least one of the plurality of first cross-sectional image features, the descriptive parameter being one of a dimension, a diameter, an angle, an area, a shape, or a volume. At least one of the steps of acquiring at least one of the performance indicators D1-D3 further includes calculating an average or a statistical deviation of the at least one descriptive parameter of the plurality of first semiconductor features. The measurement recipe R may include measurement results according to any of the above-described embodiments, preferably the method described in the eighteenth embodiment. In the HVM phase M3, full 3D tomography may consume an unacceptable amount of time, while a single wedge cross-section is possible. With the above-described monitoring recipe generation, a single wedge cut may provide the information necessary for process monitoring. In the eighteenth embodiment, a method is provided for extracting 3D information from an analysis of a single wedge cut through a 3D memory stack.

[0176] The single wedge cut monitoring method according to the eighteenth embodiment provides information about a single tilted plane 51 in a 3D memory stack. The information provided by a single cut is insufficient for full 3D reconstruction of the stack geometry, compared to, for example, 3D tomography using a stack of parallel slices cut and imaged sequentially. However, the advantage of the single wedge cut approach over cutting and imaging multiple slices is that the cutting and image acquisition time is significantly reduced. Therefore, the single wedge cut monitoring method is advantageously applied in HVM phase M3. The orientation of the FIB and imaging beam allows cutting a localized trench without additional damage to the wafer (without "taking out" the wafer sample), which is important for wafer in-line inspection. Figure 24 illustrates the single wedge cut monitoring method for an example of multiple HAR structures in a memory device. The FIB 55 cuts a single interface 51 at an angle GF with respect to the wafer surface (not shown) parallel to the y-axis. A charged particle imager (not shown) is positioned in the z direction and generates a digital image of surface 51, including multiple cross-sectional image features 203.1-203.4 of HAR structures 205.1-205.4 oriented perpendicular to the wafer surface. The combination of footprints 203.1-203.4 of the cross-sectional image features of individual memories or HAR channels 205.1-205.4 results in a reconstruction of a representative vertical HAR profile 201 of the average HAR channel. To reconstruct the vertical profile of the representative vertical HAR profile 201 (the diameter, or more generally, the shape, of the representative vertical HAR profile as a function of Z), the combination of footprints 203.1-203.4 of the individual channels imaged at tilt angles corresponding to angle GF can form the representative vertical HAR profile 201 shown in FIG. 24. The Z coordinates of the image pixels of the digital image can be reconstructed from the XY coordinates on the top-down image using the method described above, for example, using the cross-sectional image features of the word lines at the inclined plane 51 as shown in Figures 5, 6, 12, and 19b.This method applies the prior information generated in Phase M2 (e.g., that all channels in the field of view ideally have the same vertical profile or a predetermined variation of the vertical profile). In a single cross-sectional surface, cross sections of the HAR structure are located at different depths, and Phase M2 allows for training or learning of the characteristics of the HAR channels as a function of depth through 3D volumetric inspection. In a single example, it is assumed that the diameter of the HAR channels is constant in the z direction and that the axes of all HAR channels are parallel to each other. If a more complex vertical profile of the average channel is predicted from the 3D volumetric inspection in Phase M2, the application of the assumed profile can compensate for critical dimensions and proximity measured at different depths.

[0177] Also, from the wedge cut data, it is possible to reconstruct the lateral displacement of the average channel 201 as a function of Z (so-called "tilt" or "wobble") if the channels are assumed to form a regular lattice in the XY plane. In one example, the memory channels are designed in step M1 to form a hexagonal lattice in the lateral plane. For each channel footprint, including footprints in the image or cross-sectional image segments 203.1-203.4, the centroid (x i ,y i ) is measured (index i corresponds to a number of i = 1...N cross-sectional image segments of the HAR channel). As mentioned above, the z-coordinate of the footprint centroid z i is the (x i ,y i ) can be determined for all channels in the FOV. If all channels in the FOV have the same shape and orientation, then the channel footprint centroid (x i ,y i The lateral position of

[0178]

number

number

number

number

number

number

number

number

number

[0179]

number

number

number

number

[0180] From the above, the digital images can be used to reconstruct the 3D shape of the 3D memory stack under certain statistical assumptions or prior information obtained during the 3D volume inspection in phases M1 and M2. For example, prior information or machine learning methods can be used to identify the type or characteristics of defects from the digital image of a single surface 51. In phases M1 and M2, a set of representative digital images of a single surface is obtained from the 3D volume images of typical defects and used to train a machine learning algorithm. The type of defect in the 3D volume is determined from the digital image of the single inspection surface 51 by analyzing the single surface using prior information or a machine learning algorithm. A simple example is shown in FIG. 25. FIG. 25a shows multiple HAR structures (three of which are designated by reference numeral 205) in the inspection volume 160 with lateral extensions LX and LY and corresponding depth ranges LZ. The multiple HAR structures are tilted with respect to the z-axis. FIG. 25b shows a first wedge cut by cutting a first cross-sectional surface 51.1 with the FIB 50 oriented in the local wafer coordinate system in the yz plane. The right side of Figure 25b shows a first digital image 207.1 obtained by the charged particle imaging device 40 or 140, which includes multiple cross-sectional image features 203.1 representing cross sections of multiple HAR structures 205. The spacing between multiple lines 209.1 passing through parallel rows of HAR structures parallel to the x-direction varies across the digital image of the cross-sectional surface 51.1, which allows the deriving of the slopes of the multiple HAR channels 205, along with preliminary information regarding the design distance of the HAR channels.

[0181] For separate analysis of the channel slope, a second cross-sectional surface 51.2 is generated and imaged, but the orientation of the wedge cut changes between the first and second cross-sectional surfaces. For example, two cross-sectional surfaces 51.1 and 51.2 are etched and imaged by rotating the wafer about the wafer normal or z-axis between the first cross-sectional surface generation and the first digital image generation, and between the second cross-sectional surface generation and the second digital image generation. This results in two cross-sectional surfaces, but the first cross-sectional surface is related to the second cross-sectional surface by, for example, a 90° rotation angle. The results of the second cross-sectional surface generation and second digital image generation are shown in Figure 25c in the local wafer coordinate system. For illustrative purposes, the wafer is not rotated, but the orientation of the imaging device is rotated so that the FIB column 50 is oriented in the xy plane of the local wafer coordinate system. The second cross-sectional surface 51.2 is again generated at an angle GF, but is rotated relative to the first cross-sectional surface 51.1 by a predetermined angle about the z-axis. In this example, the predetermined angle is 90°, but other angles are possible. A corresponding second digital image 207.2 of the second cross-sectional surface 51.2 is again acquired by the charged particle imaging device 40 or 140, which includes multiple cross-sectional image features 203.2 representing cross sections through the multiple HAR structures at different depths. Here, the line 209.2 passing through the parallel rows of HAR structures is tilted with respect to the x-axis, which derives the tilt of the multiple HAR channels 205, along with preliminary information regarding the design distance of the HAR channels. By performing the first and second wedge-cut inspections, the tilt angles and orientations of the multiple HAR structures in the inspection volume 160 can be comprehensively determined from only the two cross-sectional measurements.

[0182] In one example, the tilt of the HAR channel 205 can be the result of a misaligned etching process, where the tilt angle of the HAR structures depends on their location on the wafer (e.g., the radial tilt angle increases with increasing distance to the wafer center). Another example is a systematic error in alignment, where all HAR structures are still parallel with equal tilt angles in one direction. By generating several representative cross-sectional surfaces cut and imaged at given locations on the wafer, the characteristics or origin of the defects can be derived.

[0183] In the above example, the HAR structures in the examination volume are parallel and tilted, so that cross sections of the HAR structures exhibit different pitches in the x and y directions. In other examples, for example, when the HAR structures are tilted relative to a common center of all HAR axes or when the tilt is random, all tilt characteristics can be evaluated with at least two cross-sectional surfaces generated at different orientations, for example, by tilting the wafer between the first and second cross-sectional image generation.

[0184] The channel footprints detected in the top-down image of the wedge-cut cross section can be used to investigate lateral variations in channel diameter (CD uniformity) and proximity of adjacent channels, as described in German Patent Application No. 10 2019 006645.6, filed September 20, 2019, and PCT Application PCT / EP2020 / 000101, filed May 15, 2020, both of which are incorporated herein by reference.

[0185] Thus, a method for wafer inspection of at least a first inspection volume in a wafer according to a monitoring recipe includes the steps of: loading the wafer onto a wafer support table in a dual beam apparatus; moving the wafer support table to align a first measurement site on the wafer with an intersection point of the dual beam apparatus; cutting a first cross-sectional surface in the first inspection volume at a tilt angle GF with the FIB column; generating a first cross-sectional image slice of the first cross-sectional surface with a charged particle imager; and obtaining performance indicators of a plurality of first semiconductor features in the first inspection volume, the performance indicators including analyzing the first cross-sectional image slice with preliminary information about the plurality of first semiconductor features. The analyzing step may include extracting a plurality of first cross-sectional image features representing cross-sections of the plurality of first semiconductor features at the tilt angle GF with image processing, the image processing including at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation. The obtaining of the performance index may further include calculating at least one descriptive parameter from at least one of the plurality of first cross-sectional image features, wherein the at least one descriptive parameter of the first semiconductor feature is one of a dimension, a diameter, an angle, an area, a shape, or a volume. The analyzing may also include generating a mean or a statistical deviation of the at least one descriptive parameter of the plurality of first semiconductor features. The analyzing may also include generating a depth map Z(x,y) of the first cross-sectional image slice according to a third embodiment. To generate the depth map, the method may further include: (1) determining at least two second cross-sectional image features in the first cross-sectional image slice; and (2) determining the depth map Z(x,y) from lateral positions of the at least two second cross-sectional image features, wherein the at least two second cross-sectional image features may represent semiconductor integrated structures at different depths within the inspection volume. In one example, the step of obtaining the performance index includes deriving tilt angle deviations of the plurality of first semiconductor features from the lateral positions of the plurality of first cross-sectional image features, the tilt angle deviations being angles of the first semiconductor features relative to an axis perpendicular to the wafer surface.A 3D representation of the first plurality of semiconductor features can be generated from the plurality of first cross-sectional image features and the depth map Z(x,y) arranged at the tilt angle GF. The first cross-sectional image slice can be compared to a 2D digital image slice through the inspection volume of the reference wafer or die, the 2D digital image slice being a cross-sectional image slice or a virtual cross-sectional image slice acquired in a previous measurement of the reference wafer or die and stored in memory, the virtual cross-sectional image slice being generated from 3D volumetric image data stored in memory. The 3D volumetric image data was acquired in a previous slice image measurement of the inspection volume of the reference wafer or die (e.g., in phase M2) and stored in memory. The previous slice image measurement can be performed by a second dual beam device or the same first dual beam device including at least a second FIB column and a second charged particle imaging device.

[0186] In one example of a monitoring recipe R, the recipe includes: moving a wafer support table to align a second measurement location on the wafer with an intersection point of the first dual beam device; cutting a second cross-sectional surface in a second inspection volume at a tilt angle GF; and generating a second cross-sectional image slice of the second cross-sectional surface with a charged particle imager. In one example, moving the wafer includes rotating the wafer support table about an axis perpendicular to the wafer support surface. Performance indicators of the plurality of first semiconductor features in the first and second inspection volumes are obtained by analyzing the first and second cross-sectional image slices with prior information about the plurality of first semiconductor features. In one example, deriving the tilt angle deviation of the plurality of first semiconductor features includes analyzing the first and second cross-sectional image slices.

[0187] The wafer defect inspection apparatus includes a focused ion beam (FIB) column configured to cut and expose at least a first cross-sectional surface through a first inspection volume of the wafer at a tilt angle GF, and a charged particle imager configured to image the at least first cross-sectional surface to form a first cross-sectional image slice. The wafer defect inspection apparatus also includes an image processing unit having installed thereon software code configured to determine, in the at least first cross-sectional image slice, a plurality of cross-sectional image features that are cross-sections of the semiconductor structure at the tilt angle GF inside the inspection volume and to determine depths of the plurality of cross-sectional features within the inspection volume. The wafer defect inspection apparatus is configured to inspect the inspection volume in the wafer without extracting it from the wafer. The wafer defect inspection apparatus includes a memory that stores preliminary information.

[0188] In some examples, the FIB for cutting the inclined cross-sectional surface of the wafer is described as a gallium FIB, which is a common FIB used for cutting. Other examples of cutting FIBs can utilize other materials, such as gases generated in a gas field ionization source (GFIS), such as hydrogen, helium, or neon. In one example, a dual beam device of the present invention includes a cutting FIB and an imaging charged particle beam, both of which utilize noble gases, such as He and Ne, for cutting and imaging.

[0189] The above-described embodiments and examples are intended to be merely illustrative of the present invention. While the above embodiments have been described in the context of semiconductor structures as probes, the methods and apparatus of at least some of these embodiments are equally applicable to materials or probes of equivalent structures whose depths can be determined from second cross-sectional image features representing layers in the sample, or whose depths are known. Those skilled in the art will be able to make modifications, improvements, variations, and combinations to the specific embodiments and examples without departing from the scope defined by the provisions attached hereto.

[0190] The invention is further described below by the following sets of clauses.

[0191] Clause 1: A method of wafer inspection of at least a first inspection volume in a wafer by a first dual beam apparatus, comprising: In a dual beam apparatus comprising at least an FIB column and a charged particle imaging device, a first optical axis of the FIB column forms a tilt angle GF with a surface of the wafer support table, a second optical axis of the charged particle imaging device forms an angle GE with a normal to the surface of the wafer support table, and the first and second optical axes form an intersection point, moving the wafer support table to align a first measurement location on the wafer with an intersection point of the dual beam device; cutting a first cross-sectional surface in a first inspection volume with a FIB column at a tilt angle GF; generating a first cross-sectional image slice of the first cross-sectional surface with a charged particle imaging device; obtaining performance metrics for a plurality of first semiconductor features in a first inspection volume, the performance metrics including analyzing a first cross-sectional image slice with preliminary information about the plurality of first semiconductor features; A method comprising:

[0192] Clause 2: The method of clause 1, wherein the first semiconductor feature is one of a via, a HAR structure, or a HAR channel.

[0193] Clause 3: The method described in clause 1 or 2, wherein the analyzing step includes extracting a plurality of first cross-sectional image features representing cross sections of a plurality of first semiconductor features having a tilt angle GF by image processing, and the image processing includes at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation.

[0194] Clause 4: The method of clause 3, wherein the step of obtaining a performance index further includes a step of calculating at least one descriptive parameter of the first semiconductor feature from at least one of the plurality of first cross-sectional image features, and the at least one descriptive parameter is one of a dimension, a diameter, an angle, an area, a shape, or a volume.

[0195] Clause 5: The method of clause 4, wherein obtaining a figure of merit further comprises calculating a mean or statistical deviation of at least one descriptive parameter of the plurality of first semiconductor features.

[0196] Clause 6: The method of any one of clauses 3 to 5, wherein the analyzing step further comprises generating a depth map Z(x,y) of the first cross-sectional image slice.

[0197] Clause 7: The step of generating a depth map comprises: determining at least two second cross-sectional image features in the first cross-sectional image slice; determining a depth map Z(x,y) from the lateral positions of the at least two second cross-sectional image features; 7. The method of clause 6, further comprising:

[0198] Clause 8: The method of clause 7, wherein at least two second cross-sectional image features each represent a semiconductor integrated structure at a different depth within the inspection volume.

[0199] Clause 9: The method of clause 7 or 8, wherein each of the plurality of first semiconductor features extends in a direction perpendicular to the wafer surface, and the second cross-sectional image feature comprises a cross-section of the semiconductor structure extending in a direction parallel to the wafer surface.

[0200] Clause 10: The method of any one of clauses 7 to 9, wherein the second cross-sectional image feature comprises a cross-section of at least one of an insulating line or insulating layer, a metal line or metal layer, and a semiconductor line or semiconductor layer.

[0201] Clause 11: A method according to any one of clauses 6 to 10, wherein the step of obtaining a performance index includes deriving a tilt angle deviation of a plurality of first semiconductor features from the lateral positions of the plurality of first cross-sectional image features, the tilt angle deviation being the angle of the first semiconductor features relative to an axis perpendicular to the wafer surface.

[0202] Clause 12: The method of any one of clauses 4 to 11, further comprising deriving a 3D representation of the first plurality of semiconductor features from the first plurality of cross-sectional image features and the depth map Z(x,y) arranged at a tilt angle GF.

[0203] Clause 13: A method according to any one of clauses 1 to 12, wherein the analyzing step further comprises comparing the first cross-sectional image slice to a 2D digital image slice through the inspection volume of a reference wafer or die.

[0204] Clause 14: The method of clause 13, wherein the 2D digital image slice is a cross-sectional image slice acquired in a previous measurement of a reference wafer or die and stored in memory.

[0205] Clause 15: The method of clause 13, wherein the 2D digital image slices are virtual cross-sectional image slices, and the virtual cross-sectional image slices are generated from 3D volumetric image data stored in a memory.

[0206] Clause 16: The method of clause 15, wherein 3D volumetric image data is acquired in a past slice image measurement of an inspection volume of a reference wafer or die and stored in memory.

[0207] Clause 17: The method of clause 16, wherein the past slice image measurement is performed by a second dual beam device comprising at least a second FIB column and a second charged particle imaging device.

[0208] Clause 18: The method of clause 16 or 17, wherein the past slice image measurement is performed by a first dual beam device.

[0209] Clause 19: moving the wafer support table to align a second measurement location on the wafer with an intersection point of the first dual beam device; cutting a second cross-sectional surface in a second inspection volume with the FIB column at a tilt angle GF; generating a second cross-sectional image slice of the second cross-sectional surface with a charged particle imaging device; obtaining performance metrics for the plurality of first semiconductor features in the first and second inspection volumes by analyzing the first and second cross-sectional image slices with prior information about the plurality of first semiconductor features; 19. The method of any one of clauses 1 to 18, further comprising:

[0210] Clause 20: The method described in clause 19, wherein the wafer support table is rotated about an axis perpendicular to the wafer support surface between cutting the first cross-sectional surface in the first inspection volume and cutting the second cross-sectional surface in the second inspection volume.

[0211] Clause 21: The method of clause 19 or 20, wherein the step of deriving tilt angle deviations of the plurality of first semiconductor features includes analysis of first and second cross-sectional image slices.

[0212] Clause 22: The method according to any one of clauses 1 to 21, wherein the tilt angle GF of the FIB column is between 30° and 80°, preferably between 30° and 45°.

[0213] Clause 23: The method of any one of clauses 1 to 22, wherein the charged particle imaging device is oriented perpendicular to the wafer surface with GE=0°.

[0214] Clause 24: The method of any one of clauses 1 to 23, wherein the charged particle imaging device is a helium ion microscope (HIM).

[0215] Clause 25: The method of clause 24, wherein the first cross-sectional image slice is obtained by acquiring an image of a first cross-sectional surface through the examination volume in a single image scan.

[0216] Clause 26: A method for inspecting a wafer by a dual beam apparatus, the dual beam apparatus comprising: a focused ion beam (FIB) column having an optical axis disposed at an angle GF of 30° to 45° with respect to a support surface of a wafer support table; and a helium ion microscope (HIM) having an optical axis disposed perpendicular to the support surface, wherein the optical axes of the FIB column and the HIM form an intersection; positioning a first measurement site on the wafer by a wafer stage having a wafer support table configured to hold the wafer at the intersection; cutting a first cross-sectional surface through a first inspection volume at approximately angle GF with the FIB column, the first cross-sectional surface having a depth extension LZ below the wafer surface of greater than 1 μm; imaging a first cross-sectional surface with a single image scan using a helium ion microscope to form a high resolution cross-sectional image slice; A method comprising:

[0217] Clause 27: The method of clause 26, wherein the first inspection volume has a depth range of greater than 2 μm, greater than 6 μm, or greater than 10 μm.

[0218] Clause 28: The method of clause 26 or 27, wherein in the imaging step, the HIM is configured to generate images with a lateral resolution of less than 2 nm, preferably less than 1 nm, or less than 0.5 nm.

[0219] Clause 29: A computer program product comprising program code for performing the method according to any one of clauses 1 to 28.

[0220] Clause 30: A dual beam apparatus configured to carry out the method according to any one of clauses 1 to 28.

[0221] Clause 31: A focused ion beam (FIB) column configured to cut and expose a first cross-sectional surface through at least a first inspection volume in the wafer at a tilt angle GF; a charged particle imaging device configured to image at least a first cross-sectional surface to form a first cross-sectional image slice; an image processing unit having installed thereon software code configured to determine, in at least a first cross-sectional image slice, a plurality of cross-sectional image features that are cross-sections of the semiconductor structure at a tilt angle GF inside the inspection volume, and to determine depths of the plurality of cross-sectional features within the inspection volume; a defect detection unit configured to determine, from the plurality of cross-sectional image features, deviations of the semiconductor structure inside the inspection volume from a predetermined characteristic; Equipped with A wafer defect inspection apparatus configured to inspect an inspection volume in a wafer without extraction from the wafer.

[0222] Clause 32: The apparatus described in Clause 31, wherein the image processing unit having software code installed therein is further configured to calculate a 3D representation of a plurality of first semiconductor structures from a plurality of first cross-sectional image features arranged at a tilt angle GF.

[0223] Clause 33: The apparatus of clause 31 or 32, further comprising a memory for storing preliminary information.

[0224] Clause 34: The apparatus of any one of clauses 31 to 33, wherein the charged particle imaging apparatus further comprises a wafer support table.

[0225] Clause 35: An apparatus according to clause 34, wherein the optical axis of the FIB column is arranged at a tilt angle GF relative to the wafer support table of between 30° and 80°, preferably between 30° and 45°.

[0226] Clause 36: An apparatus according to clause 34 or 35, wherein the optical axis of the charged particle imaging device is positioned at an angle perpendicular to the wafer support table.

[0227] Clause 37: An apparatus described in any one of clauses 31 to 36, wherein the charged particle imaging device is a helium ion microscope (HIM).

[0228] Clause 38: An apparatus as described in Clause 37, wherein the inspection volume has a depth extension LZ below the wafer surface of more than 1 μm, preferably more than 2 μm, and the HIM is configured with a depth of focus (DOF) that exceeds the depth extension LZ so that a cross-sectional image slice is obtained by a single image scan with the HIM beam.

[0229] Clause 39: An apparatus according to any one of clauses 31 to 38, comprising a control unit having installed therein software code for carrying out the method according to any one of clauses 1 to 28.

[0230] Clause 40: A method for generating a metrology recipe for inspection of a representative inspection volume in a wafer, comprising: generating a first set of critical design performance indicators D1 including dimensions of critical design features in the CAD image data and the inspection volume; generating a second set of key fabrication performance indicators D2 that include 3D volumetric image data obtained during fabrication process development; deriving a measurement recipe R, the measurement recipe R comprising obtaining at least one of a third set of performance indicators D3; Including, A method, wherein generating a second set of key fabrication performance indicators D2 and a process recipe R includes acquiring at least a first cross-sectional image slice of at least a first cross-sectional surface through a representative inspection volume having a depth extension LZ>1 μm below the wafer surface.

[0231] Clause 41: The method described in Clause 41, wherein at least one of the steps of obtaining at least one of the performance indicators D1 to D3 includes extracting a plurality of first cross-sectional image features representing cross sections of a plurality of first semiconductor features having a tilt angle GF by image processing, and the image processing includes at least one of feature extraction, edge detection, pattern recognition, or pixel interpolation.

[0232] Clause 42: The method of clause 41, wherein at least one of the steps of obtaining at least one of the performance indicators D1 to D3 further includes a step of calculating at least one descriptive parameter from at least one of the plurality of first cross-sectional image features, and the at least one descriptive parameter of the at least one first semiconductor feature is one of a dimension, a diameter, an angle, an area, a shape, or a volume.

[0233] Clause 43: The method described in Clause 42, wherein at least one of the steps of obtaining at least one of the performance indicators D1 to D3 further comprises a step of calculating the average or statistical deviation of at least one descriptive parameter of a plurality of first semiconductor features.

[0234] Clause 44: A method as described in any one of clauses 41 to 43, wherein at least one of the steps of obtaining at least one of the performance indicators D1 to D3 includes a method step as described in any one of clauses 1 to 28.

[0235] Clause 45: A method for inspection of an inspection volume in a wafer by a dual beam apparatus, the dual beam apparatus comprising at least a FIB column and a charged particle imager, a first optical axis of the FIB column forming an inclination angle GF with a surface of the wafer support table, a second optical axis of the charged particle imager forming an angle GE with a normal to the surface of the wafer support table, and the first and second optical axes forming an intersection point; loading a wafer onto a wafer support table of a dual beam apparatus; moving the wafer support table to align a first measurement site on the wafer with the intersection point; acquiring a series of N cross-sectional image slices including at least a first cross-sectional image slice and a second cross-sectional image slice in the examination volume, wherein acquiring the first and second cross-sectional image slices comprises: cutting into the examination volume at approximately angle GF with the FIB column to subsequently expose at least first and second cross-sectional surfaces in the examination volume; and imaging the at least first and second cross-sectional surfaces with a charged particle imager to acquire the at least first and second cross-sectional image slices; determining at least one first cross-sectional image feature in the first and second cross-sectional image slices; determining at least one second cross-sectional image feature in the first and second cross-sectional image slices; determining a depth of at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of at least one second cross-sectional image feature in the first cross-sectional image slice; A method comprising:

[0236] Clause 46: The method of clause 45, further comprising performing lateral co-registration of the first and second cross-sectional image slices with at least one common cross-sectional image feature.

[0237] Clause 47: The method of clause 46, wherein the step of lateral mutual image alignment includes subtraction of image distortion deviations between at least the first and second cross-sectional image slices.

[0238] Clause 48: A method according to any one of clauses 45 to 47, wherein the step of determining the depth of at least one first cross-sectional image feature includes determining at least one lateral difference between at least a first position of at least one second cross-sectional image feature in the first cross-sectional image slice and at least a second position of the second cross-sectional image feature in the second cross-sectional image slice.

[0239] Clause 49: A method according to any one of clauses 45 to 48, wherein in the step of determining at least one second cross-sectional image feature, at least two second cross-sectional image features are determined in the first cross-sectional image slice, each of the second cross-sectional image features representing a semiconductor integrated structure at a different depth within the inspection volume.

[0240] Clause 50: The method described in Clause 49, wherein the step of determining the depth of at least one first cross-sectional image feature includes determining the depth of at least one first cross-sectional image feature in the first cross-sectional image slice from the lateral positions of at least two second cross-sectional image features.

[0241] Clause 51: An apparatus according to any one of clauses 45 to 50, wherein the optical axis of the charged particle imaging device is arranged perpendicular to the wafer.

[0242] Clause 52: A method according to any one of clauses 45 to 51, wherein the charged particle imaging device is a scanning helium ion microscope.

[0243] Clause 53: A method according to any one of clauses 45 to 52, further comprising forming at least one alignment feature near the inspection volume configured to laterally align the first and second cross-sectional image slices with each other.

[0244] Clause 54: The method described in Clause 53, wherein alignment features are fabricated above the inspection volume and configured to determine the positions of first and second edges formed by the intersections of the first and second cross-sectional surfaces with the wafer surface.

[0245] Clause 55: A method as described in Clause 53, wherein a first cross-sectional surface is formed in the wafer extending longer in an x-direction perpendicular to the FIB beam than a second cross-sectional surface, such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface, and at least one alignment feature is formed on the remaining parallel surface segment of the first surface segment.

[0246] Clause 56: A method according to any one of clauses 45 to 55, wherein the tilt angle GF between the first optical axis of the FIB column and the surface of the wafer support table is in the range of 8° to 45°, preferably in the range of 8° to 30°, more preferably in the range of 8° to 15°.

[0247] Clause 57: A method according to any one of clauses 45 to 55, wherein the tilt angle GF between the first optical axis of the FIB column and the surface of the wafer support table is in the range of 25° to 60°, preferably 25° to 45°, more preferably 30° to 40°.

[0248] Clause 58: A method according to any one of clauses 45 to 57, wherein the step of acquiring first and second cross-sectional image slices in the inspection volume includes: exposing a first cross-sectional surface in the inspection volume by scanning a focused ion beam of the FIB column in a first direction by the scanning unit; tilting the focused ion beam in a second direction perpendicular to the first direction by the scanning unit; and exposing a second cross-sectional surface in the inspection volume by scanning the focused ion beam in the first direction by the scanning unit so that the first and second cross-sectional surfaces form different angles with the wafer surface by approximately the tilt angle GF.

[0249] Clause 59: The method of claim 59, wherein the wafer does not move during the steps of acquiring at least first and second cross-sectional image slices in the inspection volume.

[0250] Clause 60: A method according to any one of clauses 45 to 59, wherein the first cross-sectional image feature comprises a cross-section of the semiconductor structure extending in a direction parallel to an axis perpendicular to at least the wafer surface, and the second cross-sectional image feature comprises a cross-section of the semiconductor structure extending in a direction parallel to at least the wafer surface.

[0251] Clause 61: A method according to any one of clauses 45 to 60, wherein the first common cross-sectional image feature comprises a cross-section of at least one of a via, a HAR structure, or a HAR channel of a semiconductor integrated circuit.

[0252] Clause 62: A method according to any one of clauses 45 to 61, wherein the second common cross-sectional image feature comprises a cross-section of at least one of an insulating line or insulating layer, a metal line or metal layer, or a semiconductor line or semiconductor layer of a semiconductor integrated circuit.

[0253] Clause 63: The method of any one of clauses 45 to 62, wherein the number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is about 1000 or more.

[0254] Clause 64: A computer program product comprising program code for performing the method according to any one of clauses 45 to 63.

[0255] Clause 65: A dual beam device configured to carry out the method according to any one of clauses 45 to 63.

[0256] Clause 66: A wafer defect inspection apparatus configured to perform the method according to any one of clauses 45 to 63, a focused ion beam column configured to cut and expose at least first and second cross-sectional image surfaces in an inspection volume in the wafer; a charged particle imaging device configured to image at least first and second cross-sectional surfaces to form first and second cross-sectional image slices; an image processing unit having software code installed thereon configured to determine, in at least the first and second cross-sectional image slices, cross-sectional image features that are cross-sections of the semiconductor structure inside the inspection volume, and to determine a depth of the cross-sectional features within the inspection volume; a defect detection unit configured to determine, from the cross-sectional image features, deviations of the semiconductor structure inside the inspection volume from a predetermined characteristic; Equipped with A wafer defect inspection apparatus configured to inspect an inspection volume in a wafer without extraction from the wafer.

[0257] Clause 67: A dual beam device (1) for inspection of an inspection volume below a wafer surface (55) of a wafer (8), comprising: a wafer stage (155) comprising a wafer support table (15) configured, in use, to hold a wafer (8) on a support surface (152); a focused ion beam (FIB) column (50) having an optical axis oriented at an angle GF greater than 30° relative to a support surface (152) of a wafer support table (15); a helium ion microscope (HIM) (140) arranged with its optical axis perpendicular to a support surface (152), the optical axes of the FIB column (50) and the HIM (140) forming an intersection point (43); a stage control unit (16) configured, in use, to position a first measurement site (6.1) of the wafer (8) at the intersection (43); a control unit (19) configured to control the FIB column (50) and the HIM (140) to alternately cut a plurality of N cross-sectional surfaces in the inspection volume with the FIB beam (51) at approximately angle GF and image each cross-sectional surface by scanning with the HIM beam (144) to form a plurality of N high-resolution cross-sectional image slices, the inspection volume having a depth extension LZ below the wafer surface (55) of more than 1 μm, preferably more than 2 μm; A dual beam device (1) comprising:

[0258] Clause 68: A dual beam device (1) according to clause 67, wherein the depth extension LZ in the above direction is greater than 5 μm (for example, 6 μm or 10 μm).

[0259] Clause 69: A dual beam device (1) as described in clause 67 or 68, wherein the control unit (19) is further configured, in use, to scan the HIM beam (144) over an area exceeding the lateral size LX or LY of the inspection volume (approximately 5 μm to 10 μm (LX or LY)) and to collect multiple secondary electrons in a time series manner by the secondary electron detector (17).

[0260] Clause 70: A dual beam device (1) described in any one of clauses 67 to 69, wherein the HIM (140) has a depth of focus (DOF) that exceeds the depth extension LZ so that multiple N high-resolution cross-sectional image slices can each be obtained by a single image scan with the HIM beam (144).

[0261] Clause 71: A dual beam device (1) described in any one of clauses 67 to 70, further comprising an image processing unit having installed thereon software code configured to determine a first virtual cross-sectional image oriented parallel to the surface (55) of the wafer (8) from a plurality of N cross-sectional image slices.

[0262] Clause 72: A method for inspecting an inspection volume below a wafer surface (55) of a wafer (8) by a dual beam apparatus, the dual beam apparatus comprising: a focused ion beam (FIB) column (50) having an optical axis arranged at an angle GF of more than 30° with respect to a support surface (152) of a wafer support table (15); and a helium ion microscope (HIM) (140) having an optical axis (142) arranged perpendicular to the support surface (152), the optical axes of the FIB column (50) and the HIM (140) forming an intersection point (43); positioning a first measurement site (6.1) of the wafer (8) by a wafer stage (155) having a wafer support table (15) configured to hold the wafer (8) at the intersection (43); cutting, with the FIB column (50), a plurality of N alternating cross-sectional surfaces in the inspection volume at approximately an angle GF, the plurality of N cross-sectional surfaces having a depth extension LZ greater than 1 μm below the wafer surface (55); imaging each cross-sectional surface with a single image scan by a helium ion microscope (140) to form a plurality of N high-resolution cross-sectional image slices; A method comprising:

[0263] Clause 73: The method of clause 72, wherein the cutting of the plurality of N cross-sectional surfaces has a depth extension LZ below the wafer surface (55) of more than 2 μm, preferably more than 6 μm, more preferably more than 10 μm.

[0264] Clause 74: The method of clause 72 or 73, further comprising computing at least one virtual cross-sectional image parallel to the wafer surface from the plurality of N cross-sectional image slices.

[0265] Clause 75: The method of clause 74, wherein the first virtual cross-sectional image is computed at the conductive layer or word line.

[0266] Clause 76: The method of clause 74 or 75, wherein a second virtual cross-sectional image is computed at the insulating layer.

[0267] Clause 77: A method according to any one of clauses 72 to 76, further comprising the step of performing lateral co-registration of each of the series of N cross-sectional image slices with at least one common cross-sectional image feature.

[0268] Clause 78: forming at least one alignment feature near the inspection volume configured to form at least one common cross-sectional image feature; performing lateral co-registration of the series of N cross-sectional image slices with at least one common cross-sectional image feature; 78. The method of any one of clauses 72 to 77, further comprising:

[0269] Clause 79: The method of any one of clauses 72 to 78, further comprising loading the wafer onto a wafer support table of the dual beam apparatus.

[0270] Clause 80: A method according to any one of clauses 72 to 79, further comprising a step of generating a depth map Z(x,y;n) (index n=1···N for each of the N cross-sectional image slices) for each of a series of N cross-sectional image slices.

[0271] Clause 81: The method described in Clause 80, further comprising determining, in each of the series of N cross-sectional image slices, at least one second cross-sectional image feature representing a cross section through a second semiconductor feature oriented parallel to the wafer surface 55.

[0272] Clause 82: The method of clause 81, wherein a depth map Z(x,y;n) (index n=1···N) for each cross-sectional image slice is generated from the lateral position of at least one second cross-sectional image feature.

[0273] Clause 83: The method of any one of clauses 72 to 82, wherein the number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is about 1000 or more.

[0274] Clause 84: A method according to any one of clauses 72 to 82, wherein the number N of cutting surfaces is less than 50, preferably less than 20, and the inspection volume is separated into a plurality B of blocks 181.1...181.B arranged diagonally through the inspection volume.

[0275] Clause 85: A method of obtaining at least one virtual cross-sectional image from a set of cross-sectional image slices, comprising: obtaining a series of N cross-sectional image slices by alternating imaging and cutting a series of N cross-sectional surfaces at a tilt angle GF relative to an inner inspection volume of the wafer; determining a first orientation direction of a first semiconductor structure comprising a first plurality of first cross-sectional image features in the series of N cross-sectional image slices; computing a virtual cross-sectional image including a plurality of virtual cross-sectional image pixels, the virtual cross-sectional image being perpendicular to the first orientation direction; Including, The method, wherein for each virtual cross-sectional image pixel, a pixel value is calculated by projecting a subset of at least one cross-sectional image slice of the series of N cross-sectional image slices in a first orientation direction and interpolating the pixel values ​​from the projections of the subset of at least one cross-sectional image slice.

[0276] Clause 86: A method as described in Clause 85, wherein for each virtual cross-sectional image pixel, a subset of at least one cross-sectional image slice is selected by evaluating the distance from each of a series of N cross-sectional image slices in a first orientation direction to the virtual cross-sectional image pixel and selecting at least the first cross-sectional image slice having the shortest distance.

[0277] Clause 87: The method according to clause 86, wherein a second cross-sectional image slice of the subset of at least one cross-sectional image slice is selected as the cross-sectional image slice having the second shortest distance.

[0278] Clause 88: The method of clause 86 or 87, wherein another cross-sectional image slice of the subset of at least one cross-sectional image slice is selected in order of increasing distance to the virtual cross-sectional image pixel in the first orientation direction.

[0279] Clause 89: A method according to any one of clauses 85 to 88, wherein the step of projecting a subset of at least one cross-sectional image slice and interpolating pixel values ​​by projecting a subset of at least one cross-sectional image slice includes forming a third plurality of first cross-sectional image features in the virtual image slice by projecting and interpolating at least a subset of the first plurality of first cross-sectional image features.

[0280] Clause 90: The method of any one of clauses 85 to 89, wherein the steps of projection and interpolation are combined with at least one of feature extraction, thresholding, contour interpolation, and model-based interpolation.

[0281] Clause 91: A method according to any one of clauses 85 to 90, further comprising a step of generating a depth map Z(x,y;n) (index n=1···N for each of the N cross-sectional image slices) for each of a series of N cross-sectional image slices.

[0282] Clause 92: The method described in Clause 91, wherein the depth map Z(x,y;n) (index n=1···N) of each cross-sectional image slice is generated from a plurality of second cross-sectional image features representing a cross section through a second semiconductor feature oriented in a second orientation direction perpendicular to the first orientation direction.

[0283] Clause 93: A method as described in Clause 91 or 92, wherein a depth map Z(x,y;n) (index n=1···N) for each cross-sectional image slice is generated by determining the depth of a first cross-sectional image feature in the cross-sectional image slice from the lateral positions of at least two second cross-sectional image features.

[0284] Clause 94: A method according to any one of clauses 85 to 93, wherein the first semiconductor feature comprises at least one of a via, a HAR structure, or a HAR channel of a semiconductor integrated circuit inside an inspection volume of the wafer.

[0285] Clause 95: The method of clause 94, wherein the second semiconductor feature comprises at least one of an insulating line or insulating layer, a metal line or metal layer, or a semiconductor line or semiconductor layer of a semiconductor integrated circuit inside the inspection volume of the wafer.

[0286] Clause 96: determining at least one first cross-sectional image feature in each of a series of N cross-sectional image slices; determining at least one second cross-sectional image feature in each of the series of N cross-sectional image slices; 96. The method of any one of clauses 93 to 95, further comprising:

[0287] Clause 97: A method according to any one of clauses 85 to 96, wherein the first orientation direction is the z direction perpendicular to the wafer surface, and the virtual cross-sectional image slice is calculated in a plane parallel to the wafer surface at a depth ZV below the wafer surface.

[0288] Clause 98: A method as described in Clause 97, wherein for each virtual cross-sectional image pixel coordinate (x,y), a subset of at least one cross-sectional image slice is determined by selecting at least the mth cross-sectional image slice having the shortest distance to depth ZV such that distance Zrv(m) = Z(x,y;m) - ZV is the minimum value of all depth maps Z(x,y;n) (index n = 1···N).

[0289] Clause 99: The method of clause 98, wherein the second and subsequent cross-sectional image slices of the subset of at least one cross-sectional image slice are selected in order of increasing distance Zrv(n) to the virtual cross-sectional image pixel in the z direction.

[0290] Clause 100: A method according to any one of clauses 97 to 99, wherein the depth ZV is adjusted according to the depth of layers parallel to the wafer surface, these layers being formed by second semiconductor features oriented in a second orientation direction parallel to the wafer surface.

[0291] Clause 101: The method of any one of clauses 97 to 100, wherein a first virtual cross-sectional image slice is computed at a depth ZV1 in an insulating layer between two adjacent metal layers or word lines.

[0292] Clause 102: The method of any one of clauses 97 to 101, wherein a second virtual cross-sectional image is calculated at a depth ZV2 inside the metal layer of the word line.

[0293] Clause 103: A method according to any one of clauses 85 to 102, wherein the optical axis of a charged particle beam imaging system for acquiring a series of N cross-sectional image slices is oriented perpendicular to the wafer surface such that the angle GE between the optical axis and a z-axis perpendicular to the wafer surface is GE=0°.

[0294] Clause 104: The method of any one of clauses 85 to 103, further comprising performing lateral co-registration of each of the series of N cross-sectional image slices with at least one common cross-sectional image feature.

[0295] Clause 105: A method according to any one of clauses 85 to 104, further comprising forming at least one alignment feature near the inspection volume configured to form at least one common cross-sectional image feature for lateral mutual alignment.

[0296] Clause 106: The method of clause 104 or 105, wherein the step of lateral mutual image alignment includes subtraction of image distortion deviations.

[0297] Clause 107: The method of any one of clauses 85 to 106, wherein the number N of cross-sectional image slices is at least N=10, preferably N>100, and more preferably N is about 1000 or more.

[0298] Clause 108: In a dual beam apparatus comprising at least an FIB column and a charged particle imaging device, a first optical axis of the FIB column forms an inclination angle GF with a surface of the wafer support table, a second optical axis of the charged particle imaging device forms an angle GE with a normal to the surface of the wafer support table, and the first and second optical axes form an intersection point, the step of loading a wafer on the wafer support table; moving the wafer support table to align a first measurement location on the wafer with an intersection point of the dual beam device; 108. The method of any one of clauses 85 to 107, further comprising:

[0299] Clause 109: The method of clause 108, wherein the wafer does not move during the step of acquiring the series of N cross-sectional image slices through the inspection volume.

[0300] Clause 110: A computer program product comprising program code for performing the method according to any one of clauses 85 to 109.

[0301] Clause 111: A dual beam device configured to carry out the method according to any one of clauses 85 to 109.

[0302] Clause 112: A wafer defect inspection apparatus configured to perform the method according to any one of clauses 85 to 109, a wafer stage for holding the wafer; a focused ion beam (FIB) column configured to cut and expose a series of N cross-sectional surfaces in an inspection volume inside the wafer at a tilt angle GF relative to the surface of the wafer; a charged particle imaging device configured to image a series of N cross-sectional surfaces to form a series of N cross-sectional image slices; an image processing unit having installed thereon software code configured to determine a virtual cross-sectional image oriented parallel to the surface of the wafer from the series of N cross-sectional image slices; A wafer defect inspection device comprising:

[0303] Clause 113: A wafer defect inspection device according to clause 112, wherein the tilt angle GF of the FIB column is between 30° and 80°, preferably between 30° and 45°.

[0304] Clause 114: A wafer defect inspection apparatus according to clause 112 or 113, wherein the charged particle imaging device is oriented perpendicular to the wafer surface.

[0305] Clause 115: A wafer defect inspection device according to any one of clauses 112 to 114, wherein the charged particle imaging device is a helium ion microscope (HIM).

[0306] Clause 116: A dual beam apparatus for wafer inspection, comprising: a wafer stage for supporting the wafer; a gallium FIB column positioned at an angle GF relative to a support surface of the sample support stage and configured to cut at least one cross-sectional surface at the angle GF relative to the wafer; a helium ion beam column positioned perpendicular to the support surface of the wafer stage and configured to generate a cross-sectional image slice of the cross-sectional surface with a single image scan; a secondary electron detector configured to collect a plurality of secondary electrons during a single image scan; an operating unit configured, in use, to provide operational control of the dual beam device; A dual beam device comprising:

[0307] Clause 117: A dual beam apparatus as described in clause 116, wherein the FIB is configured to cut a cross-sectional surface to a depth range below the wafer surface extending over more than 1 μm, more than 2 μm, or more than 6 μm.

[0308] Clause 118: A dual beam device as described in clause 116 or 17, wherein the operational unit further comprises an image processing unit that performs at least one of feature extraction, edge detection, pattern recognition, and pixel interpolation.

[0309] Clause 119: A dual beam device as described in Clause 118, wherein the image processing unit is further configured to compute one of a depth map, a 3D volume image, and a virtual cross-sectional image.

[0310] Clause 120: A dual beam apparatus as described in clause 118 or 19, wherein the operational unit further comprises a defect detection unit that calculates one of the dimensions, area, diameter, angle, and shape of the semiconductor feature.

[0311] Clause 121: The dual beam apparatus of clause 120, wherein the defect detection unit is further configured to calculate one of an average and a statistical deviation of the plurality of semiconductor features.

[0312] Clause 122: A wafer inspection method for 3D inspection of an inspection volume in a wafer, wherein the depth range of the inspection volume is greater than 1 μm, preferably greater than 2 μm, greater than 6 μm, or greater than 10 μm; A method in which 3D inspection is performed on the wafer to generate an image with a lateral resolution of less than 2 nm, preferably less than 1 nm, or even less than 0.5 nm.

[0313] Clause 123: The method of clause 122, wherein the image is a 3D image.

[0314] Clause 124: The method of clause 122 or 123, further comprising cutting at least one cross-sectional surface through a depth range of the inspection volume with the FIB column.

[0315] Clause 125: A method according to clause 124, wherein the image is obtained by acquiring at least one image of at least one cross-sectional surface through the examination volume in a single image scan with a helium ion microscope.

[0316] Clause 126: The method of clause 122 or 125, further comprising image processing to perform at least one of feature extraction, edge detection, pattern recognition, and pixel interpolation.

[0317] Clause 127: The method of clause 122 or 126, further comprising the step of computing one of a depth map, a 3D volume image, and a virtual cross-sectional image.

[0318] Clause 128: The method of clause 122 or 127, further comprising calculating one of a dimension, an area, a diameter, an angle, and a shape of the semiconductor feature.

[0319] Clause 129: The method of clause 122 or 128, further comprising calculating one of an average and a statistical deviation of a plurality of semiconductor features.

[0320] Clause 130: A method for forming a 3D volume image from a series of cross-sectional image slices, comprising the step of acquiring a series of N cross-sectional image slices including at least a first cross-sectional image slice and a second cross-sectional image slice of an inspection volume, wherein acquiring the first and second cross-sectional image slices includes subsequently exposing at least first and second cross-sectional surfaces in the inspection volume by cutting into the inspection volume at approximately an angle GF with a FIB column, and acquiring at least the first and second cross-sectional image slices by imaging at least the first and second cross-sectional surfaces with a charged particle imaging device, wherein the first cross-sectional image surface is cut extending longer in a direction perpendicular to the FIB beam than the second cross-sectional surface such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface.

[0321] Clause 131: The method described in Clause 130, further comprising forming at least one alignment feature on a parallel surface segment of the first cross-sectional surface for first lateral mutual alignment of the first and second cross-sectional image slices.

[0322] Clause 132: determining at least one first cross-sectional image feature in the first and second cross-sectional image slices; performing a second lateral co-registration of the first and second cross-sectional image slices based on the first lateral co-registration to achieve a predetermined co-position accuracy of the first and second cross-sectional image slices; 133. The method of claim 132, further comprising:

[0323] Clause 133: The method of clause 133, wherein the first cross-sectional image feature is a cross-section of a via, a HAR structure, or a HAR channel.

[0324] Clause 134: determining at least one second cross-sectional image feature in the first and second cross-sectional image slices; determining a depth of at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of at least one second cross-sectional image feature in the first cross-sectional image slice; determining a depth of the at least one first cross-sectional image feature in the second cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the second cross-sectional image slice; achieving a mutual position accuracy of at least the first and second cross-sectional image slices of less than 5 nm, less than 3 nm, or less than 2 nm by taking depth into account in the second lateral mutual registration; 134. The method of claim 132 or 133, further comprising:

[0325] Clause 135: The method of clause 134, wherein at least one second cross-sectional image feature is formed by a cross-section of one of an insulating line or insulating layer, a metal line or metal layer, or a semiconductor line or semiconductor layer.

[0326] Clause 136: A first displacement ΔY′ of a first cross-sectional image feature between a first cross-sectional image slice and a second cross-sectional image slice Ch and a second displacement ΔY′ of the second cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; WL and determining a distance d between the first and second cross-sectional image slices; determining a lateral mutual displacement vector ΔY′ between the first and second cross-sectional image slices; 136. The method of claim 134 or 135, further comprising:

[0327] Clause 137: The method according to any one of clauses 130 to 136, wherein the tilt angle GF is adjusted to 25° to 45° or 30° to 36°.

[0328] Clause 138: A method for configuring precise alignment of multiple cross-sectional image slices acquired from within an inspection volume below a wafer surface, comprising: acquiring a series of cross-sectional image slices, including first and second cross-sectional image slices, from a series of cross-sectional image surfaces formed by successively cutting a series of cross-sectional surfaces with a FIB column positioned at an oblique angle GF relative to the wafer surface and imaging each of the cross-sectional surfaces with a charged particle imaging column positioned at an angle GE relative to a normal to the wafer surface to acquire the series of cross-sectional image slices; performing a first coarse alignment of the first and second cross-sectional image slices to obtain a mapping of cross-sectional image features in the first and second cross-sectional image slices; performing a second fine alignment of the first and second cross-sectional image slices to achieve a mutual position accuracy of less than 5 nm, less than 3 nm, or less than 2 nm; A method comprising:

[0329] Clause 139: forming a first cross-sectional surface by cutting in a direction perpendicular to the FIB beam longer than the second cross-sectional surface, such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface; forming at least one alignment feature on a parallel surface segment of the first cross-sectional surface for first lateral mutual alignment of the first and second cross-sectional image slices; 139. The method of claim 138, further comprising:

[0330] Clause 140: determining at least one first cross-sectional image feature in the first and second cross-sectional image slices; determining at least one second cross-sectional image feature in the first and second cross-sectional image slices; determining a depth of at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of at least one second cross-sectional image feature in the first cross-sectional image slice; determining a depth of the at least one first cross-sectional image feature in the second cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the second cross-sectional image slice; taking depth into account in the second fine registration; 139. The method of claim 138 or 139, further comprising:

[0331] Clause 141: A first displacement ΔY′ of a first cross-sectional image feature between a first cross-sectional image slice and a second cross-sectional image slice Ch and a second displacement ΔY′ of the second cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; WL and determining a distance d between the first and second cross-sectional image slices; determining a lateral mutual displacement vector ΔY′ between the first and second cross-sectional image slices; 141. The method of claim 140, further comprising:

[0332] Clause 142: The method of clause 140 or 141, wherein the at least one first cross-sectional image feature is a cross-section of a via, a HAR structure, or a HAR channel.

[0333] Clause 143: A method according to any one of clauses 140 to 142, wherein at least one second cross-sectional image feature is formed by a cross-section of one of an insulating line or insulating layer, a metal line or metal layer, or a semiconductor line or semiconductor layer.

[0334] Clause 144: The method according to any one of clauses 138 to 143, wherein the tilt angle GF is adjusted to 25° to 45° or 30° to 36°.

[0335] Clause 145: The method according to any one of clauses 140 to 144, wherein the angle GE is adjusted to 0°.

[0336] Clause 146: A computer program product comprising program code for performing the method according to any one of clauses 130 to 146.

[0337] Clause 147: A dual beam device configured to carry out the method according to any one of clauses 130 to 146. [Explanation of symbols]

[0338] 1 Dual beam device 2. Operating Unit 4.1, 4.2, 4.3 First cross-sectional image feature 6.1, 6.2 Measurement site 8 wafers 10 Semiconductor samples 11 Cross section surface 13 Examination volume 15 Wafer support table 16 Stage Control Unit 17 Secondary electron detector 19 Control Unit 20 Alignment marks 22 Alignment trench or edge 23 Parallel surface segments of the first cross-sectional surface 24 Alignment marks 25 Integrated Circuit Features 26.1, 26.2, 26.3 Image segments 28 statue segments 30 Depth extension 32.1, 32.2, 32.3 Alignment features 34 Depth Level DLZ 36 Alignment feature distance 38 Alignment Features 40 Charged Particle Beam (CPB) Imaging System 42 Optical axis of the imaging system 43 inspection points 44 Electron Beam 46 Scanning Imaging Lines 48 FIB optical axis 50 FIB columns 51 Focused ion beam 52 Cross section surface 53 Cross section surface 54 Cross section surface 55 Wafer top surface 58 FIB beam coincidence points 60 Slice distance on wafer surface 62 Slice distance at the bottom of the examination volume 64 Angle Spread 66 First cutting angle 68 Second Cutting Angle 72 Bottom surface of layer L1 73.1, 73.2, 73.3 Second cross-sectional image feature 74 Upper boundary of layer L4 75 Pillar HAR structure 76.1, 76.2 Top edge of cross-sectional surface 77.1, 77.2, 77.3 Cross-sectional image segments of HAR channels 78 Vertical edge of HAR structure 79.1 Center of gravity of cross-sectional image features 80 layer horizontal edge 82 Image Processing Unit 84 Defect Detection Unit 86 Charged particle column control unit 88 Interface Unit Angular divergence of FIB in 90 x direction 92 Proximal trench 94 Distal trench 96 First cross section surface 100.1, 100.2, 100.3 2D cross-sectional image 1000 series of 2D cross-sectional images 121 Virtual cross-sectional image slices 123 pixel raster in y direction 125 Interpolation direction of HAR structure 140 Helium Ion Microscope (HIM) 142 HIM optical axis 146 Examination Area 148 Alignment Features 150 Alignment feature 148 image 152 single scan lines in the x direction 155 wafer stage 160 test volumes 162 Multiple Alternating Layers 164 Multiple HAR Structures 166.1, 166.2 Virtual cross section 175 Cross section of conductive layer or word line 177.1, 177.2 Cross section of HAR structure 179.1, 179.2 Virtual cross section of HAR structure 181.1, 181.b Block of inspection volume 191 Deep Examination Volume 201 Typical vertical HAR profile 203 Cross-sectional image features 205 HAR structure 207 Digital Image

Claims

1. 1. A method for generating a 3D volumetric image from a series of cross-sectional image slices, comprising: acquiring a series of N cross-sectional image slices including at least a first cross-sectional image slice and a second cross-sectional image slice of an examination volume, wherein acquiring the first and second cross-sectional image slices includes subsequently exposing at least first and second cross-sectional surfaces in the examination volume by cutting into the examination volume at approximately angle GF with an FIB column; and imaging the at least first and second cross-sectional surfaces with a charged particle imager to acquire the at least first and second cross-sectional image slices, wherein the first cross-sectional image surface is cut to extend longer in a direction perpendicular to the FIB beam than the second cross-sectional surface such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface. method.

2. 2. The method of claim 1, further comprising forming at least one alignment feature on the parallel surface segments of the first cross-sectional surface for first lateral mutual alignment of the first and second cross-sectional image slices.

3. determining at least one first cross-sectional image feature in the first and second cross-sectional image slices; performing a second lateral co-registration of the first and second cross-sectional image slices based on the first lateral co-registration to achieve a predetermined co-position accuracy of the first and second cross-sectional image slices; The method of claim 2 further comprising:

4. The method of claim 3 , wherein the first cross-sectional image feature is a cross-section of a via, a HAR structure, or a HAR channel.

5. determining at least one second cross-sectional image feature in the first and second cross-sectional image slices; determining a depth of the at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the first cross-sectional image slice; determining a depth of the at least one first cross-sectional image feature in the second cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the second cross-sectional image slice; taking the depth into account in the second lateral mutual registration to achieve a mutual position accuracy of the at least first and second cross-sectional image slices of less than 5 nm, less than 3 nm, or less than 2 nm; 5. The method of claim 3 or 4, further comprising:

6. The method of claim 5 , wherein the at least one second cross-sectional image feature is formed by a cross-section of one of an insulating line or layer, a metal line or layer, or a semiconductor line or layer.

7. a first displacement ΔY′ of the first cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; Ch and a second displacement ΔY′ of the second cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; WL and determining a distance d between the first and second cross-sectional image slices; determining a lateral mutual displacement vector ΔY′ between the first and second cross-sectional image slices; The method of claim 5 or 6, further comprising:

8. The method according to any one of claims 1 to 7, wherein the tilt angle GF is adjusted to between 25° and 45° or between 30° and 36°.

9. 1. A method for configuring precise alignment of multiple cross-sectional image slices acquired from within an inspection volume below a wafer surface, comprising: acquiring a series of cross-sectional image slices, including first and second cross-sectional image slices, from a series of cross-sectional image surfaces formed by successively cutting a series of cross-sectional surfaces with a FIB column positioned at an oblique angle GF relative to a wafer surface and imaging each of the cross-sectional surfaces with a charged particle imaging column positioned at an angle GE relative to a normal to the wafer surface to acquire a series of cross-sectional image slices; performing a first coarse alignment of the first and second cross-sectional image slices to obtain a mapping of cross-sectional image features in the first and second cross-sectional image slices; performing a second fine alignment of the first and second cross-sectional image slices to achieve a mutual position accuracy of less than 5 nm, less than 3 nm, or less than 2 nm; A method comprising:

10. forming the first cross-sectional surface by cutting longer than the second cross-sectional surface in a direction perpendicular to the FIB beam, such that a parallel surface segment of the first cross-sectional surface remains after forming the second cross-sectional surface; forming at least one alignment feature on the parallel surface segment of the first cross-sectional surface for first lateral mutual alignment of the first and second cross-sectional image slices; 10. The method of claim 9, further comprising:

11. determining at least one first cross-sectional image feature in the first and second cross-sectional image slices; determining at least one second cross-sectional image feature in the first and second cross-sectional image slices; determining a depth of the at least one first cross-sectional image feature in the first cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the first cross-sectional image slice; determining a depth of the at least one first cross-sectional image feature in the second cross-sectional image slice from a lateral position of the at least one second cross-sectional image feature in the second cross-sectional image slice; taking the depth into account in the second fine registration; 11. The method of claim 9 or 10, further comprising:

12. a first displacement ΔY′ of the first cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; Ch and a second displacement ΔY′ of the second cross-sectional image feature between the first cross-sectional image slice and the second cross-sectional image slice; WL and determining a distance d between the first and second cross-sectional image slices; determining a lateral mutual displacement vector ΔY′ between the first and second cross-sectional image slices; The method of claim 11 further comprising:

13. The method of claim 11 or 12, wherein the at least one first cross-sectional image feature is a cross-section of a via, a HAR structure, or a HAR channel.

14. 14. The method of any one of claims 11 to 13, wherein the at least one second cross-sectional image feature is formed by a cross-section of one of an insulating line or layer, a metal line or layer, or a semiconductor line or layer.

15. The method according to any one of claims 9 to 14, wherein the tilt angle GF is adjusted to between 25° and 45° or between 30° and 36°.

16. The method according to any one of claims 9 to 15, wherein the angle GE is adjusted to 0°.

17. A computer program product comprising program code for performing the method according to any one of claims 1 to 16.

18. A dual beam device configured to carry out the method according to any one of claims 1 to 16.

Images