WO2025237825A1

WO2025237825A1 - Calibration method for a particle imaging apparatus, and method for operating a calibrated particle imaging apparatus

Info

Publication number: WO2025237825A1
Application number: PCT/EP2025/062705
Authority: WO
Inventors: Daniel Schwarz; Thomas Korb
Original assignee: Carl Zeiss SMT GmbH
Current assignee: Carl Zeiss SMT GmbH
Priority date: 2024-05-14
Filing date: 2025-05-09
Publication date: 2025-11-20
Anticipated expiration: 2026-11-14
Also published as: DE102024113386A1; DE102024113386B4

Abstract

Multiple calibration methods for particle imaging apparatuses are proposed, which allow a lateral focal shift to be determined depending on a focusing setting change or a refocusing of a particle imaging apparatus Distortions in particle-optical images created by correspondingly calibrated particle imaging apparatuses can be reduced in this way, and metrological examinations in the particle-optical images and in particular in 3D tomographic images may be implemented with greater accuracy. Using correspondingly calibrated particle imaging apparatuses, it is possible in particular to measure 3D samples such as deep channels in semiconductor samples more accurately. It is also possible to calibrate and hence correct distortion effects that arise on account of sample geometries in electrostatic immersion fields.

Description

Calibration method for a particle imaging apparatus, and method for operating a calibrated particle imaging apparatus

Field of the invention

The invention relates to particle imaging apparatuses for creating particle-optical images. Specifically, the invention relates to a calibration method for a particle imaging apparatus and a method for operating a calibrated particle imaging apparatus.

Prior art

Particle beam systems comprising particle beam columns, such as for example electron beam columns or ion beam columns, are known from the prior art. One example is a scanning electron microscope, in which a focused electron beam scans a region to be imaged of an object to be examined, and secondary electrons or backscattered electrons created by the incident electron beam on the object are detected depending on the deflection of the focused particle beam, in order to create or compute an electron-microscopic image of the scanned region of the object.

Basically, the primary particle beam is created by a beam generator having a particle source, passes through beam-shaping elements such as for example a condenser lens, a stigmator or other beam-shaping elements and is then focused onto the object to be examined by an objective lens. In order to achieve a high resolution of the particle beam column or of the particle beam microscope, the particle beam on the object must be focused as well as possible, i.e. a region illuminated by the focused particle beam on the surface of the object ("beam spot") ought to be as small and round as possible. For this purpose, the particle beam microscope with its particle-optical components is adjusted with the aim that an image plane into which the particle source is imaged by the optical system coincides with the surface of the object. In the case of an arrangement of the object at a given distance from the objective lens, this can be achieved by changing the focusing setting of the particle beam microscope until the beam spot on the surface of the object is as small as possible. By way of example, the focusing setting of the particle beam microscope can be changed by changing the excitation of the objective lens and/or by changing the kinetic energy of the particles of the particle beam when passing through the objective lens. Moreover, a region illuminated by the focused particle beam on the surface of the object ought to be as round as possible. For this purpose, it is often necessary to correct imaging aberrations of the particle beam system by means of stigmators. By way of example, electrostatic multipole electrodes or magnetic multipoles can be used for such an astigmatism correction.

After the focusing setting of the particle beam microscope has been set in this way, even further measures are necessary, as a rule, to improve the quality of the beam focus on the surface of the object. This includes adjusting the particle beam such that it passes substantially centrally through the objective lens. That is based on the consideration that lens aberrations of a particle-optical lens become ever more noticeable the further away the beam is from the particle-optical axis of the lens when passing through the latter. Since the objective lens typically provides a majority of the refractive power required for imaging the particle source on the object surface, it is advantageous in view of reducing the imaging aberrations involved during this imaging to adjust the beam relative to the objective lens such that it passes through the objective lens as centrally as possible. For this adjustment, the particle beam system or the particle beam column according to the prior art comprises for example one or more deflection devices for displacing the particle beam within the objective lens and which is or are arranged in the beam path between the particle beam source and the objective lens. By changing the excitation of this deflection device or these deflection devices, it is possible to displace the location within a principal plane of the objective lens at which the centre of the particle beam passes through the principal plane. This also applies analogously to objective lens systems, for example to a system with a combination of a magnetic objective lens and an electrostatic objective lens.

It is not easy to adjust the particle beam correctly. There are already a few methods in the prior art for this purpose.

By way of example, it is known to use the recording of particle-microscopic images with the particle beam microscope in at least two different focusing settings for the adjustment in order to set the optimum excitation of the deflection device based on an analysis of these recorded images or computed images. That is based on the following consideration: If the particle beam passes centrally through the objective lens and is focused on the surface of the object, the recorded particle-microscopic image is substantially in focus. If the focusing setting is slightly changed proceeding from this setting and if a particle-microscopic image is recorded for this altered setting, this image is only slightly less sharp in comparison with the previously recorded particle-microscopic image and is otherwise substantially the same as the latter. However, if the two images recorded with different focusing settings are recorded using a particle beam that does not pass through the objective lens centrally, the two images differ not only in respect of image sharpness but also in respect of the relative position thereof. The change in focusing setting leads to the second image being displaced or offset relative to the first image. Therefore, in the prior art, what is known as a "wobble method" is performed for the purpose of adjusting the particle beam, within the scope of which the focusing setting is changed periodically while images are recorded continuously. A user observes the recorded images, which move back and forth in the case of an improperly adjusted beam, and changes the excitation of the deflection device or deflection devices until the resultant images are all essentially stationary. This is a manual process that requires a certain amount of experience and is therefore also time-consuming. There are also automated adjustment methods.

Particle beam systems and, in particular, particle imaging apparatuses are also used for process control, for example in semiconductor manufacturing. With the ongoing development of ever smaller and ever more complex microstructures such as semiconductor components, there is a need to further develop and optimize planar production techniques and inspection systems for producing and inspecting small dimensions of the microstructures. For instance, the development and production of the semiconductor components require monitoring of the design of test wafers, and the planar production techniques require process optimization for reliable production with high throughput. Moreover, there have been recent demands for an analysis of semiconductor wafers for reverse engineering and for a customized, individual configuration of semiconductor components. Therefore, there is a need for inspection means which can be used with high throughput to examine the microstructures on wafers with high accuracy.

Typical silicon wafers used in the production of semiconductor components have diameters of e.g. 300 mm. Each wafer is subdivided into approximately 30 to 60 repeating regions (“dies”) with a size of up to 800 mm², and these each contain at least one integrated circuit pattern, e.g. for a memory chip or for a processor chip. A semiconductor apparatus comprises multiple semiconductor structures, which are produced in layers on a surface of the wafer by planar integration techniques. Semiconductor wafers typically have a plane surface on account of the production processes. The structure size of the integrated semiconductor structures in this case extends from a few pm to the critical dimensions (CDs) of 5 nm, and the structure sizes will become even smaller in the near future; in future, structure sizes or critical dimensions (CDs) are expected to be less than 3 nm, for example 2 nm, or even less than 1 nm. In the case of the aforementioned small structure sizes, defects of the order of the critical dimensions must be identified quickly over a very large area. For multiple applications, the specification requirement regarding the accuracy of a measurement provided by inspection equipment is even higher, for example by a factor of two or one order of magnitude. For instance, a width of a semiconductor feature must be measured with an accuracy better than 1 nm, for example 0.3 nm or even less, and a relative position of semiconductor structures must be determined with an overlay accuracy better than 1 nm, for example 0.3 nm or even less.

Semiconductor structures are among the most delicate structures made by man and are afflicted by various defects. Systems for quantitative 3D metrology, defect detection or error testing search for these errors. Manufactured semiconductor structures are based on prior knowledge. The semiconductor structures are produced from a sequence of layers that extend parallel to a substrate. For example, given a sample of an integrated circuit pattern, the metal lines in the metal layers extend parallel to one another, or what are known as HAR structures (high aspect ratio structures) and metal feedthroughs extend perpendicular to the metal layers. The angle between the metal lines in various layers is either 0° or 90°. Structures of the VNAND type, by contrast, are known to have a circular cross section on average.

During production, the semiconductor disks run through approximately 1000 process steps, and approximately 100 or more parallel layers are formed within the semiconductor disk, the layers comprising the transistor layers, the layers of the line centre and the connection layers and, in the case of memory components, a plurality of 3D arrays of memory cells. The dimensions, shapes and arrangements of the semiconductor structures and patterns are subject to various influences. Currently, etching and deposition are the critical processes during the production of 3D memory components. Other involved process steps such as lithographic exposure or implantation also influence the properties of the elements in the integrated circuits.

The aspect ratio and the number of layers of integrated circuits is increasing continuously, and the structures grow in the third (vertical) dimension. The current height of memory stacks already exceeds a dozen micrometres and counting. By contrast, the size of the features is getting ever smaller. The minimal structure size or critical dimension is below 10 nm, e.g. 7 nm or 5 nm, and will approach structure sizes below 3 nm in the near future. While the complexity and the dimensions of the semiconductor structures grow in the third dimension, the lateral dimensions of the integrated semiconductor structures are becoming ever smaller. Hence, it is becoming ever more difficult to measure the shape, the dimensions and the alignment of the features and patterns in 3D and the overlay thereof with great precision.

As the demands on the resolution of particle imaging systems in three dimensions become ever more stringent, the inspection and 3D analysis of integrated semiconductor circuits on wafers is becoming ever more sophisticated. As a rule, the lateral measurement resolution of charged particle systems is limited by the scanning raster of the individual pixels or the dwell times per pixel on the sample and by the diameter of the charged particle beam. The resolution of the scanning raster may be set within the imaging system and matched to the diameter of the charged particle beam on the sample or the object. The typical raster-scan resolution is 2 nm or better, but the raster-scan resolution limit may in principle be reduced without physical limitation. The diameter of the charged particle beam has a limited size, which depends on the operating conditions of the charged particle beam and the utilized lens. The beam resolution is limited by approximately half the beam diameter or by the full width at half maximum of the beam. The resolution may be better than 2 nm, e.g. even better than 1 nm.

A conventional method for creating 3D tomography data of semiconductor samples on the nm scale is the so-called slice and image approach, which for example is realized using a dual particle beam system. Dual particle beam systems having a charged particle beam column for imaging and having an FIB column for milling are operated at the so-called coincidence point, i.e. the optical axis of the charged particle beam, the optical axis of the FIB and the sample surface meet at a single point (are coincident). This is advantageous since this ensures that imaging is in fact implemented at the sample position that is affected by the milling.

For example, a slice and image method is described in WO 2020/244795 A1. According to the method in WO 2020/244795 A1 , a 3D volume inspection is performed on an inspection sample taken from a semiconductor wafer. This method is disadvantageous in that a wafer must be destroyed in order to obtain an inspection sample in block form. Then again, the method is advantageous in that the working distance between a charged particle imaging apparatus such as a scanning electron microscope (SEM), more precisely the column thereof, and a sample to be imaged or a wafer may be comparatively small, especially if the two columns are arranged perpendicular to one another and the optical axis of the SEM is perpendicular to the sample surface. In general, the resolution of a particle imaging apparatus improves with shorter working distance.

An alternative method for creating 3D tomography data of semiconductor samples on the nm scale is described in WO 2021/180600 A1. According to that document, the disadvantage that a wafer must be destroyed is overcome by a different geometric arrangement of a milling or FIB column on the one hand and an imaging column with charged particles on the other hand. As a result of this arrangement, the coincidence point may be navigated over the entire wafer without needing to destroy the wafer. The so-called wedge cut approach uses the slice and image method with an oblique cut angle into the surface of a semiconductor wafer. However, the working distance between the imaging column and an imaging sample or a wafer increases in comparison with the approach described in WO 2020/244795 A1 on account of geometric restrictions when arranging the two columns with respect to one another. This leads to a slightly lower image resolution.

The demand for ever larger memory chip capacities given substantially the same chip area leads to the design of 3D memory architectures. In many cases, these 3D architectures consist of a large number (>50) of insulating and conductive layers in alternation, which stack up to form stacks with a height of 5-10 pm or even more. Deep channels that are ultimately filled with insulating and conductive material are etched into these layers. In some chip architectures, such stacks are produced and then stacked one on top of the other in order to obtain even higher memory chip capacities on the same chip area.

A tricky step in the production of such structures lies in the monitoring of the etching of these deep channels: Dry etching chambers are designed for unidirectional etching of the entire wafer, but slight deviations from the desired etching direction have ever more serious results as the etching depth increases. In the case of 3D memory architectures, a slight deviation from the desired etching direction already leads to unwanted channel shifts on the wafer; and these must be controlled e.g. in order to ensure the overlay of the memory cells on the chip structures located therebelow. This problem becomes ever more serious as the etching depth increases. Therefore, semiconductor manufacturers have interest in measuring this channel inclination over the wafer.

When measuring the channel inclination according to a slice and image method, for example using a dual system comprising an electron beam column for imaging and an ion beam column for material removal (FIB-SEM system), a control of the image distortions is required over a height range of 5-10 pm or even higher. The lateral position of the imaging particle beam must be known to an accuracy of a few nm over the entire height range. While the position of the imaging or inspecting particle beam is known on the surface of a sample to be inspected by way of position marks or what are known as fiducials, this at least does not apply automatically to the lateral position of the charged particle beam when focusing into the sample.

US 2004/0173746 A1 discloses an examination of a sample in a so-called “tilt mode”. Tilting a sample surface relative to a scanning electron beam can be carried out in two different ways: firstly, by a mechanical tilt and secondly, by an electronic tilt. A basic idea of US 2004/0173746 A1 is that a combination of the two tilting types offers advantages for increasing resolution and minimizing aberrations. Problems with a lateral focus shift when changing the focus in a depth direction (z-direction) are not addressed in the cited reference.

US 5,894,124 discloses an inspection of a sample, wherein the surface of the sample is tilted with respect to the inspecting beam. The tilt of the sample surface generates a lateral component of the electric field. This lateral component of the electric field is corrected by a deflecting electric field that is generated using deflection electrodes.

US 4,728,799 discloses further background art related to a height measurement and correction method for an electron beam lithography system.

Description of the invention

The problem addressed by the present invention is therefore that of providing a method with which inclinations, for example channel inclinations in a semiconductor sample or in a sample of an integrated circuit, may be determined more precisely. In particular, accurate knowledge of the position of an imaging charged particle beam should also be possible in the depth of a sample, and this position should be controllable.

The problem is solved by the subject matter of the independent claims. Advantageous embodiments of the invention are evident from the dependent patent claims.

The present patent application claims the priority of German patent application No. 10 2024 113386.4 filed on 14 May 2024, the disclosure of which in the full scope thereof is incorporated in the present patent application by reference.

The invention is based on the discovery that a lateral beam offset may occur when focusing settings change in some particle imaging apparatuses, and this beam offset may be so large that distortions or deformations occur in particle-optical images and in 3D tomography images in particular, preventing a highly accurate determination of a structure inclination or channel inclination. Therefore, the inventors have examined the causes of these distortions or deformations in more detail. Causes include, inter alia, a lateral beam offset when changing the focus on account of focusing particle-optical lenses that are not aligned exactly to one another at a minimum and effects occurring in an electrostatic immersion field in conjunction with the geometry of the sample to be inspected or the wedge-like cutout. Thus, according to the invention, multiple calibration methods for particle imaging apparatuses are proposed, which allow a lateral focal shift to be determined depending on a focusing setting change or refocusing of a particle imaging apparatus. Distortions in particle-optical images created by correspondingly calibrated particle imaging apparatuses can be reduced in this way, and metrological examinations in the particle-optical images may be implemented with greater accuracy. It is also possible to calibrate and hence correct distortion effects that arise on account of sample geometries in electrostatic immersion fields.

According to a first aspect of the invention, the latter relates to a calibration method for a particle imaging apparatus. A particle imaging apparatus that operates with at least one charged particle beam is provided in a first method step (1a). For example, the charged particles can be electrons, positrons, muons or ions or other charged particles. In principle, the particle imaging apparatus may take different forms, for example a particle microscope, an SEM, a TEM, an STEM, an SEM-STEM, a dual beam system, a multi-beam particle microscope or a mask repair system, or another form.

A test object is provided in a method step (1b), the test object having a test structure on its surface. This test structure has geometric dimensions that are accurately known, for example a special pattern made of mutually orthogonal line structures. The test structures of the test object may be produced with great accuracy and therefore serve as a reference object. In principle, the test object is a 2D test object.

The test object is arranged in an object plane of the particle imaging apparatus in the particle- optical beam path of the particle imaging apparatus in method step (1c), in such a way that a surface normal of the test object makes an angle 01 with the particle-optical axis of the particle imaging apparatus. This angle 01 may be 0° but need not be 0°. Given an angle 01 of 0°, the at least one charged particle beam would be incident on the test structure of the test object at right angles or in telecentric fashion.

A first focusing setting z1 of the particle imaging apparatus is set in a method step (1d), in such a way that the at least one particle beam of the particle imaging apparatus is focused onto the surface of the test object, at least at points. If the angle 01=0°, then the first focusing setting z1 is set such that the surface of the test object is in focus not only at one or a few points but at multiple points and in particular at all points on the test object that should be scanned at a later stage. Raster-scanning or scanning over the test structure with the particle beam with the first (fixed) focusing setting z1 is implemented in a method step (1e), and a first particle-optical image of the test structure is created. In the context of this application, a particle-optical image is understood to mean an image which is created by means of the detection of particle radiation. A particle detector is used to detect the particle radiation, and it is possible to combine a particle detector with an optical detector. The detected particles can be for example electrons or other charged particles. It is possible for example to detect secondary electrons or backscattered electrons or mirror electrons or mirror ions for the creation of the particle-optical image. The particle-optical images are images in which the intensity of the detected particle radiation is represented as a function of an image position for a pixel.

A first position P1 of the test structure in the first particle-optical image is determined in a method step (1f). This first position P1 of the test structure may denote the overall position of the test structure. However, it is also possible that the first position P1 of the test structure denotes only one position of multiple positions of a part of the test structure. Moreover, it is of course possible that multiple first positions P1 of the test structure are determined, from which it is then for example also possible to derive the overall position P1 of the test structure.

The test object is arranged in the particle-optical beam path of the particle imaging apparatus in a method step (1g), in such a way that a surface normal of the test object makes a second angle 02 with the particle-optical axis of the particle imaging apparatus, wherein this second angle 02 differs from the first angle 01 , i.e. 02 01. By preference, the second angle 02 is an angle that differs from 0°. Thus, this preferably is a true oblique position of the surface of the test object with respect to the particle-optical axis of the particle imaging apparatus.

The focusing setting is changed and a second focusing setting z2 of the particle imaging apparatus is set in a method step (1 h), in such a way that a particle beam of the particle imaging apparatus is focused onto the surface of the test object, at least at points, wherein this second focusing setting z2 differs from the first focusing setting z1 , i.e. z2 z1 applies.

Raster-scanning over the test structure with the particle beam with the second fixed focusing setting z2 is implemented in a method step (1 i), and a second particle-optical image of the test structure is created.

A second position P2 of the test structure in the second particle-optical image is determined in a method step (1 j) . Thus, the test structure is imaged multiple times overall, to be precise with different focusing settings and angles. In so doing, the statements already made as regards the determination of the first position P1 of the test structure in the first particle-optical image in principle apply to the determination of the second position P2 of the test structure in the second particle-optical image.

A positional shift d21 between the second position P2 of the test structure in the second particle-optical image and the position P1 of the test structure in the first particle-optical image is determined in a method step (1 k). Naturally, it is once again possible that multiple positional shifts d21 are determined overall for different partial structures or different characteristic points of the test structure. However, there is at least one determination of a positional shift d21.

A focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is determined in a method step (11) depending on the focusing setting change AZ =Z2 - z1 , with the determination of the focal shift fs being implemented on the basis of the determined lateral positional shift d21 of the test structure and on the basis of the angular difference A0 = 02 - 01 . Thus, the focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is a lateral focal shift. This lateral focal shift is the cause of distortions that occur in the event of a focusing setting change. Since the test structure or test structures of the test sample, which represents a reference sample, are accurately known, the lateral focal shift fs can be inferred from the measured positional shifts d21. According to a preferred embodiment of the invention, the focal shift fs is determined on the basis of a discrepancy between an expected positional shift d21' of the test structure and the actual positional shift d21 of the test structure that was ascertained in step (1 k). The angular difference A0 = 02 - 01 is known, allowing inference about the expected positional shift d21' of the test structure. Deviations from this can be ascribed to the lateral focal shift fs.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying. The two directions x1 and x2 are preferably orthogonal to one another, and the directions may also be referred to as x-direction and y-direction in a Cartesian coordinate system. However, in the context of this patent application, a y-direction denotes the direction in which the sample is tilted, especially for a tilted test sample or sample. In this respect, reference is initially made to two directions x1 and x2, in which the focal shift fs is ascertained in each case.

According to a preferred embodiment of the invention, the first angle is 01 = 0°. This allows particularly simple determination of the focal shift fs. According to a preferred embodiment of the invention, method steps (1 h) to (11) are carried out repeatedly for a further focusing setting or for multiple further focusing settings. Thus, the corresponding method steps are repeated for further focusing settings z3, z4, z5 and so on, and corresponding positional shifts d31 , d41 , d51 and so on are ascertained. This allows a further determination of the lateral focal shift fs depending on different focusing setting changes or focusing settings. In this way, the particle imaging apparatus may be calibrated for large changes in the focusing setting, which are required during an inspection of 3D features, for example channel structures in semiconductor samples. A lateral focal shift fs generally does not grow linearly with z for relatively large changes in the focusing setting Az, and so an exact determination of the lateral focal shift fs and the subsequent correction thereof are required in each case for multiple focusing settings z3, z4, z5, etc.

According to a preferred embodiment of the invention, the focusing setting is modified incrementally by a constant value dz.

According to a preferred embodiment of the invention, the following relation applies to the constant value dz: 1 pm < dz < 50 pm. Depending on the focal ranges or focal variations over which a particle imaging apparatus should be settable overall, the incremental value dz may be chosen in matched fashion. An overall focal variation range for example comprises approximately 100 pm, e.g. in each case approx. +/-50 pm about a fixed coincidence point for dual particle beam systems.

According to a preferred embodiment of the invention, the test object is arranged on a mini stage with a settable tilt angle, and the method moreover comprises tilting of the mini stage, in particular for setting the angle 02. The angle 01 may for example be 0° for a non-tilted mini stage.

According to a further preferred embodiment of the invention, the calibration method according to the invention is carried out multiple times for different angles 0. This may be achieved particularly easily using the described mini stage. For example, the method may be carried out for 3, 4, 5, 10 or even more angles 0i. As a result, the accuracy of the calibration method may be increased overall.

During the calibration method according to the first aspect of the invention, the test structure is raster-scanned or scanned with a fixedly set, i.e. constant, focusing setting in each case. However, there are also particle imaging apparatuses which, given obliquely cut test objects or test objects arranged at a tilt, operate with a variable focusing setting that is adapted to the bevel or the tilt such that there is in principle sharp focusing at each point on the surface of the test object (adaptive focusing), even in the event of obliquely arranged samples or samples arranged at a tilt.

According to a second aspect of the invention, a calibration method is provided for such a particle imaging apparatus, the calibration method including the following steps:

A particle imaging apparatus that operates with at least one charged particle beam is provided in a first method step (2a). Incidentally, statements already made in conjunction with the first aspect of the invention apply to the particle imaging apparatus.

A test object is provided in a method step (2b), the test object having a test structure on its surface. As regards the details of a test object or a test structure, reference is made to the statements already made in conjunction with the first aspect of the invention.

The test object is arranged in an object plane of the particle imaging apparatus in the particle- optical beam path of the particle imaging apparatus in a method step (2c), in such a way that a surface normal of the test object makes a first angle 01 = 0° with the particle-optical axis of the particle imaging apparatus.

A first focusing setting z1 of the particle imaging apparatus is set in a method step (2d), in such a way that the particle beam of the particle imaging apparatus is focused onto the surface of the test object.

Raster-scanning over the test structure with the at least one particle beam with the fixed first focusing setting z1 is implemented in a method step (2e), and a first particle-optical image of the test structure is recorded. Up until this point, the calibration method according to the second aspect of the invention may correspond to method steps (1a) to (1 e) according to the first aspect of the invention, provided the angle 01 is 0° there.

The test object is arranged in the particle-optical beam path of the particle imaging apparatus in a method step (2f), in such a way that a surface normal of the test object makes a second angle 02, which differs from the first angle 01 , with the particle-optical axis of the particle imaging apparatus, i.e. wherein 02 01 applies.

Raster-scanning over the test structure with the particle beam with a variable focusing setting AZ that is adapted to the respective raster-scan position Rij of the particle beam is implemented in a method step (2g), and a second particle-optical image of the test structure is recorded. In other words, the variable focusing setting AZ is matched to the tilt of the test object such that (at least without lateral focal shift fs) the raster-scanning particle beam would in theory be incident on the surface of the test object in focus at each raster-scan position Rij.

A distortion of the test structure in the second particle-optical image relative to the first particle- optical image is determined in a method step (2h). Once again, the distortion of the test structure may be determined on the basis of determining positional changes of the test structure or of parts thereof, or else on the basis of deviations from the regularity of the test structure. In principle, there are a number of options available to this end, and these are known to a person skilled in the art.

A focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is determined in a method step (2i) depending on a focusing setting change Z in the direction of the particle-optical axis Z, or in the depth direction, with the determination of the focal shift fs being implemented on the basis of the determined distortion of the test structure and on the basis of the angle 02. In this case the focusing setting change AZ corresponds to the variable focusing setting AZ that in each case is known or has been adaptively set for a respective raster-scan position Rij of the particle beam.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying. What may apply in particular in this case is that the two directions x1 and x2 are in turn orthogonal to one another. One of the directions may correspond to a y-direction, into which the test object has been tilted.

According to a preferred embodiment of the invention, the test object is arranged on a mini stage with a settable tilt angle, and the method moreover comprises tilting of the mini stage for setting the angle 02.

According to a preferred embodiment of the invention, the method is carried out multiple times for different angles 0 or 02.

According to a preferred embodiment of the invention, the following relation applies to the edge length L of a field of view FOV raster-scanned with the particle beam: L > 10 pm, preferably L > 20 pm and most preferably L > 50 pm. In this way, a sufficiently large focusing setting change AZ can be represented during the raster-scanning process. Moreover, a sufficiently large region of the test object is represented, allowing a sufficiently accurate determination of the arising distortion.

According to a third aspect of the invention, the latter relates to a calibration method for a particle imaging apparatus, the method including the following steps:

A particle imaging apparatus is provided in a method step (3a). In this context, the particle imaging apparatus comprises a particle source for creating a particle beam with charged particles, with an acceleration voltage EHT being applied to the particle source during operation. Moreover, the particle imaging apparatus comprises at least two focusing lenses that are traversed by the particle beam and comprise a magnetic objective lens on the one hand and an electrostatic lens on the other hand, with the electrostatic lens being arranged downstream of the magnetic objective lens in the direction of the particle-optical beam path, the magnetic objective lens and the electrostatic lens focusing the particle beam onto an object at a distance z from the magnetic objective lens. The particle imaging apparatus furthermore comprises a deflector unit that is configured to deflect the particle beam towards the centre of the electrostatic lens. Moreover, the particle imaging apparatus comprises an object stage or object holder that is configured to hold an object at a working distance WD from the magnetic objective lens. Moreover, the particle imaging apparatus comprises a detection unit for detecting interaction particles that emanate from the object, for example for detecting secondary electrons. Moreover, the particle imaging apparatus comprises a controller for controlling the particle imaging apparatus. In this case, the controller is configured to control the particle source for the provision of the acceleration voltage EHT. The controller is moreover configured to control the magnetic objective lens or for example to vary the current in a coil of the magnetic objective lens. Moreover, the controller is configured to control the deflector unit, whereby the particle beam is deflectable exactly onto the centre of the electrostatic lens, and this serves the adjustment of the particle imaging apparatus. Moreover, the particle imaging apparatus is adjusted into such an energy-independent state that a change in the acceleration voltage EHT has no influence on the beam position upon incidence on the object. In the case of such an adjustment, a change in the acceleration voltage EHT only has the change of a beam diameter on the object as a consequence, i.e. the image sharpness of the recorded particle-optical image is thus modified but not the beam position.

In an alternative, it is possible in principle to reverse the order of the arrangement of electrostatic lens and magnetic objective lens. In that case, the particle beam would be adjusted in advance, in such a way that it passes centrally through the magnetic objective lens, and subsequently a focusing setting is only still modified by the electrostatic lens. However, in practice this approach is accompanied by slightly poorer imaging quality.

Thereupon, an object is provided at the distance z in a method step (3b). Once again, this object may be a test object or a test sample with regular structures; however, it may also be a different object with a characteristic feature, for example with a position mark on a sample surface.

A reference feature of the object is located in a step (3c).

A first particle-optical image is recorded by means of the focusing setting z in a step (3d). Thus, the focusing setting z in this case corresponds exactly to the working distance WD, which is why the recorded particle-optical image is in principle recorded in focus.

A position P1 of the reference feature in the first particle-optical image is determined in a method step (3e).

The acceleration voltage EHT is modified by an offset AV in a method step (3f). Which offset AV causes which change in the focusing setting Az may be known at least approximately in advance in this case, e.g. from theoretical calculations or from a preceding calibration in this regard. Thus, the value of the offset AV may be chosen meaningfully.

The particle imaging apparatus is refocused by AZ in a method step (3g) by modifying the control of the magnetic objective lens. As described, the consequence of modifying the acceleration voltage EHT by the offset AV is that the particle beam is no longer imaged on the object in focus but in blurred fashion. However, the position of the particle beam upon incidence on the object was left unchanged in the process. By contrast, there may be a change in the position of the particle beam on the object as a result of refocusing the particle imaging apparatus or as a result of refocusing the particle beam by AZ by modifying the control of the magnetic objective lens if the magnetic objective lens and the electrostatic lens are not aligned exactly to one another; the latter is practically never the case with greatest accuracy. By contrast, the control of the electrostatic lens is preferably not modified when refocusing the particle imaging apparatus by AZ by modifying the control of the magnetic objective lens. Possible lateral shifts or focal shifts fs can then be traced back exclusively to the changed control of the magnetic objective lens, enabling a calibration. A second particle-optical image is recorded in method step (3h) in the setting of the particle imaging apparatus that has been refocused by AZ.

A position P2 of the reference feature in the second particle-optical image is determined in method step (3i).

A positional shift d21 between the second position P2 of the reference feature and the first position P1 of the reference feature is determined in method step (3j). Naturally, it is also possible to determine multiple positional shifts of multiple reference features.

A focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is determined in method step (3k) depending on the refocusing AZ by means of the magnetic objective lens, with the determination of the focal shift fs being implemented on the basis of the determined positional shift d21 of the reference feature. In particular, the positional shift d21 may correspond exactly to the focal shift fs in the process.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis, with x1 x2 applying. In particular, the two directions x1 and x2 may once again be orthogonal to one another.

According to a preferred embodiment of the invention, method steps (3f) to (3k) are performed repeatedly for in each case different offsets AV in the acceleration voltage EHT. This enables a more accurate calibration and a calibration over a range AZ that is as large as possible.

According to a preferred embodiment of the invention, the offset AV is in each case modified incrementally by a constant value dV. As a result, the calibration method can be performed particularly easily in automated fashion.

According to a preferred embodiment of the invention, the electrostatic lens is an electrostatic objective lens. What is possible here in particular is that this electrostatic lens is provided by the provision of a potential at an end region of a beam tube within the magnetic objective lens (a so-called "Gemini lens"). In this case, each pole shoe of the magnetic objective lens is at earth potential; the object to be examined is preferably likewise at earth potential. Thus, work in this case is not carried out with an electrostatic immersion field. According to an alternative embodiment of the invention, the electrostatic lens develops its lens effect by means of a potential difference between the object on the one hand and the magnetic objective lens on the other hand. For example, this may be achieved by virtue of the object being at a potential that differs from zero while the magnetic objective lens is at earth potential. For example, a sample voltage of -4 kV may be applied to a wafer should the charged particles of the particle beam be electrons.

However, it is also possible that the electrostatic lens according to the third aspect of the invention has a different embodiment. It is also possible to combine the two above-described configurations of the electrostatic lens.

According to a fourth aspect of the invention, the latter relates to a calibration method for a particle imaging apparatus, the method including the following steps:

A particle imaging apparatus that operates with at least one charged particle beam and that may operate in a first mode of operation and in a second mode of operation is provided in a first method step (4a). In this case, the particle imaging apparatus works with a first depth of field TS1 and a first resolution A1 in the first mode of operation, and the particle imaging apparatus works with a second depth of field TS2 and a second resolution A2 in the second mode of operation. In this case, the first depth of field TS1 is greater than the second depth of field TS2. Moreover, the first resolution A1 is lower than the second resolution A2. In this case, the depth of field TS is in each case defined as the defocus df at which the diameter of a beam waist increases by 10% (proceeding from the smallest, best-possible beam diameter in the focal plane). In general, there is an optimal opening angle, at which a beam diameter is minimal, and the resolution is optimal. The beam diameter increases both for larger opening angles (deterioration as a result of lens aberrations) and for smaller opening angles (deterioration on account of diffraction effects). Within the scope of the invention, a smaller opening angle that benefits a better depth of field but has an attendant increase in the beam waist and hence a non-optimal resolution for image recording is chosen at least intermittently.

A test object with a wedge-like cutout and 3D structures is provided in method step (4b), wherein the cutout is defined by means of an angle GF with respect to the surface of the test object. For example, the 3D structures may be channel structures that extend into the depth of the test object over a large distance, for example over more than 10 pm, more than 50 pm or more than 100 pm. The test object with a wedge-like cutout might have already been created before method step (4b) but might likewise only be created during method step (4b). According to a preferred embodiment of the invention, the provision of the test object with a wedge-like cutout and the 3D structures comprises a wedge-like ablation of the test object with the 3D structures by means of a focused ion beam (FIB).

The test object is arranged relative to the particle imaging apparatus in method step (4c) such that the particle-optical axis Z of the particle imaging apparatus and the uncut surface of the test object are orthogonal to one another, with the 3D structures being exposed at different depths z of the test object in the direction of the particle-optical axis Z of the particle imaging apparatus. Thus, ideally, the direction of the 3D structures corresponds or substantially corresponds to the particle-optical axis Z.

Raster-scanning the test object with a wedge-like cutout in the first mode of operation and with a fixed setting z1=z1ij for the focusing setting is implemented in method step (4d), and a first particle-optical image of the test object with the 3D structures is created. Imaging is thus implemented with a greater depth of field TS1 , wherein the different arrangement of the 3D structures as regards the height or depth z is rather irrelevant to the imaging. The image representation of the 3D structures in the particle optical image has no distortions on account of a possible change in focus; it only varies slightly as regards the sharpness of the depicted 3D structures.

Positions P1 of the 3D structures in the first particle-optical image are determined in method step (4e).

Raster-scanning the test object with a wedge-like cutout in the second mode of operation and with an adaptive focusing setting z2ij that is adapted to the respective raster-scan position Rij of the particle beam is implemented in method step (4f), and a second particle-optical image of the test object with the 3D structures is created. In this case, the adaptive focusing setting z2ij describes a focusing setting in which raster-scan positions of the test object are focused on, in each case in a manner that is sharp or focused at points, i.e. with a minimal beam waist. Hence, the raster-scanning and creation of the second particle-optical image is implemented with a smaller depth of field TS2 but in return with a better resolution A2 or, expressed differently, with a greater numerical aperture NA2.

Positions P2 of the 3D structures in the second particle-optical image are determined in method step (4g).

The positions P1 and P2 are compared to one another in method step (4h), and respective shifts d21=P2-P1 of the 3D structures are ascertained. These shifts d21 are the result of a lateral focal shift fs of the particle beam during the respective focusing setting change Z = z2ij

- z1.

A lateral focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is determined in method step (4i) depending on a focusing setting change AZ = z2ij - z1 , with the determination of the focal shift fs being implemented on the basis of the determined shifts d21 of the 3D structures.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

According to a preferred embodiment of the invention, method steps (4b) to (4i) are repeated, in particular repeated multiple times, to be precise for different angles GF. Thus, the method is repeated in that case for a plurality of different test objects with a wedge-like cutout and 3D structures. In particular, it is possible to yet again ablate a first wedge-like cutout such that this gives rise to a different test object with a wedge-like cutout that has a different angle GF. For example, a focused ion beam FIB can be used to this end, in particular according to the "slice and image method".

According to a preferred embodiment of the invention, the following relation applies to a ratio of first depth of field TS1 to second depth of field TS2: 5 < TS1/TS2 < 30. In addition to that or in an alternative, the following relation applies to a ratio of first resolution A1 to second resolution A2: 2 < A1/A2 < 6.

According to a fifth aspect of the invention, the latter relates to a calibration method for a particle imaging apparatus. To this end, a particle imaging apparatus that operates with at least one charged particle beam and that may operate in a first mode of operation and in a second mode of operation is provided in a first method step (5a). An object to be imaged is not arranged in an electrostatic immersion field in the first mode of operation, while the object to be imaged is arranged in an electrostatic immersion field in the second mode of operation. For example, this may be realized by virtue of a controller of the particle imaging apparatus being configured to apply an electric potential to the object to be imaged or to a sample stage holding the object, or to switch on or else switch off said electric potential. A first test object with a wedge-like cutout and 3D structures is provided in method step (5b), wherein the cutout is defined by means of a first angle GF1 with respect to the surface of the test object.

The test object is arranged relative to the particle imaging apparatus in method step (5c) such that the particle-optical axis Z of the particle imaging apparatus and the uncut surface of the test object are in particular substantially orthogonal to one another, with the 3D structures being exposed at different depths Z of the test object in the direction of the particle-optical axis of the particle imaging apparatus.

Raster-scanning the first test object with a wedge-like cutout with a calibrated particle beam in the first mode of operation without electrostatic immersion field is implemented in method step (5d), and a first particle-optical image of the test object with the 3D structures is created.

Positions P1 of the 3D structures in the first particle-optical image are determined in method step (5e).

Raster-scanning the first test object with a wedge-like cutout with a calibrated particle beam in the second mode of operation with a first electrostatic immersion field is implemented in method step (5f), and a second particle-optical image of the test object with the 3D structures is created.

Positions P2 of the 3D structures in the second particle-optical image are determined in method step (5g).

The positions P1 and P2 are compared to one another in method step (5h), and respective shifts d21 = P2 - P1 of the 3D structures are ascertained.

A lateral focal shift fs in a direction orthogonal to the particle-optical axis Z is determined in method step (5i) depending on a strength of the electrostatic immersion field, with the determination of the focal shift fs being implemented on the basis of the determined shifts d21 of the 3D structures.

In this embodiment of the invention, the lateral focal shift fs is based on small bends in equipotential lines of the electrostatic field at the surface with a wedge-like cutout of the test object or results from the geometry of the test object with a wedge-like cutout. In this case, the lateral focal shift can be traced back exclusively to this effect. A lateral focal shift on account of a non-exact alignment of successive particle-optical lenses has already been corrected or calibrated for in advance. In this respect, raster-scanning in steps (5d) and (5f) is implemented in each case with an already calibrated particle beam or with a particle imaging apparatus calibrated in this respect.

According to a preferred embodiment of the invention, the lateral focal shift fs is in each case determined for a plurality of raster-scan positions Rij. This is meaningful since the shape of the test object with a wedge-like cutout is defined along the respective raster-scan positions. In this case, the raster-scan may be chosen with different degrees of fineness. It must be chosen to be fine enough such that the geometric bends of the potential lines at the surface of the test sample with a wedge-like cutout can be reconstructed accurately enough during the rasterscanning process.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 once again applying.

According to a preferred embodiment of the invention, method steps (5f) to (5i) are carried out repeatedly, in particular in a manner repeated multiple times, to be precise with a different strength of the electrostatic immersion field in each case. In this case it is possible that the strength of the electrostatic immersion field is modified incrementally, for example over a range from approximately 500 V/mm to approximately 4000 V/mm.

According to a preferred embodiment of the invention, method steps (5b) to (5i) are carried out repeatedly, in particular in a manner repeated multiple times, to be precise for test objects with a wedge-like cutout that have different geometries in each case. In that case, determining the focal shift fs in a direction orthogonal to the particle-optical axis Z is implemented depending on the geometry of the test object with a wedge-like cutout.

According to a preferred embodiment of the invention, the angle GF is in each case varied for the test objects with a wedge-like cutout that have different geometries. In addition to that or in an alternative, the maximum cut depth T is varied for the test objects with a wedge-like cutout that have different geometries. In addition to that or in an alternative, the maximum cut width B is varied for the different test objects with a wedge-like cutout that have different geometries. In this way it is possible to ascertain the separate influence of different parameters of the geometry on the focal shift fs. The calibration can thus be implemented very finely. According to an advantageous embodiment of the invention, the various parameters are varied separately and successively, and their influence on the focal shift fs is ascertained separately.

The statements made below apply to all calibration methods of the first to fifth aspect of the invention:

According to a preferred embodiment of the invention, the particle imaging apparatus is an apparatus from the following list of particle imaging apparatuses: a particle microscope, an SEM, a TEM, an STEM, an SEM-STEM, a dual particle beam system, a multi-beam particle microscope, a mask repair system. In this context, this list should not be construed as exhaustive. The described lateral focal shift may occur in all of the aforementioned particle imaging apparatuses, be it on account of an alignment of successive particle-optical lenses that is not 100% exact or else on account of electrostatic or magnetic immersion fields at an object or sample surface.

According to a preferred embodiment of the invention, the particle imaging apparatus comprises a magnetic objective lens and an electrostatic objective lens. In particular, this objective lens system may operate according to the so-called "Gemini principle".

In this context, in particular an end region of a beam tube that projects into the magnetic objective lens may be at an electric potential, for example at a potential of several kV.

According to a preferred embodiment of the invention, the electrostatic objective lens is arranged or formed within the pole shoes of the magnetic objective lens.

According to a sixth aspect of the invention, the latter relates to a method for operating a particle imaging apparatus. In this context, the particle imaging apparatus may be a particle imaging apparatus as described above in multiple embodiment variants.

A particle imaging apparatus that operates with at least one charged particle beam is provided in a first method step (6a).

The particle imaging apparatus is calibrated in a method step (6b) in respect of a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on a focusing setting change AZ or a refocusing Z of the particle imaging apparatus; in particular, the calibration is implemented according to one of the calibration methods as described above according to the first to fourth aspect of the invention. In addition to that or in an alternative, the particle imaging apparatus is calibrated in respect of a focal shift fs of the particle beam on the basis of the strength of an electrostatic immersion field at an object to be imaged and/or on the basis of a geometry of an object to be imaged in an electrostatic immersion field, wherein this calibration may be implemented in particular according to a calibration method that was described in conjunction with the fifth aspect of the invention, although this need not be the case.

At least one particle-optical image of an object is created by means of the particle imaging apparatus in a method step (6c). As a result, a distortion in the particle optical image on account of a lateral focal shift does not occur in the first place, offering advantages in metrological applications in particular.

According to a preferred embodiment of the invention, the calibrated particle imaging apparatus is used to create a plurality of slice images, aligned parallel to one another, through a 3D structure, and the method moreover includes the following steps:

(6d) creating a 3D volume image from the slice images; and

(6e) determining a tilt of the 3D structure on the basis of the 3D volume image.

By preference, the calibrated particle imaging apparatus may be a dual particle beam system which for example operates by means of a system made of a particle microscope and a focused ion beam FIB, for example using the slice and image method.

According to a preferred embodiment of the invention, the 3D structure comprises a NAND structure.

It is possible to combine different embodiments of the invention and even different aspects of the invention with one another, provided that no technical contradictions arise as a result.

The invention will be understood even better with reference to the accompanying figures, in which:

Fig. 1 : schematically shows a particle imaging apparatus using the example of a particle microscope;

Fig. 2: schematically shows a portion of a particle imaging apparatus having an objective lens system with a magnetic objective lens and an electrostatic objective lens; Fig. 3: schematically shows a portion of a particle imaging apparatus with an electromagnetic immersion field at the object;

Fig. 4: schematically illustrates a focal shift caused by particle-optical lenses not exactly aligned with respect to one another;

Fig. 5: schematically illustrates the occurrence of a lateral focal shift in the event of a modification of the focusing setting in the z-direction;

Fig. 6: schematically illustrates a focal shift due to an electrostatic immersion field at a sample with a wedge-like cutout;

Fig. 7: schematically illustrates problems when determining the inclination of a 3D structure as a result of the presence of a focal shift;

Fig. 8: schematically shows various arrangements of a sample stage and a mini stage;

Fig. 9: schematically shows raster-scanning of a sample for different arrangements of the sample;

Fig. 10: schematically shows a flowchart for a calibration method according to the invention;

Fig. 11: schematically shows raster-scanning of a sample for different arrangements of the sample with inclination-matched focusing settings;

Fig. 12: schematically shows a flowchart for a calibration method according to the invention;

Fig. 13: schematically shows a focal shift during a refocusing;

Fig. 14: schematically shows a flowchart for a calibration method according to the invention;

Fig. 15: schematically shows a dual particle beam system and 3D volume images created therewith;

Fig. 16: schematically shows a dual particle beam system which operates according to the slice and image method with a wedge-like cutout;

Fig. 17: schematically illustrates various depths of field for an imaging particle beam;

Fig. 18: schematically shows a flowchart for a calibration method according to the invention;

Fig. 19: schematically illustrates positional shifts of 3D structures with oblique cutouts in two different recording modes, in a perspective illustration and in a plan view;

Fig. 20: schematically shows a flowchart for a calibration method according to the invention;

Fig. 21 : schematically shows various test objects with a wedge-shaped cutout; and

Fig. 22: schematically shows a flowchart for a method for operating the particle imaging apparatus. Fig. 1 schematically shows a particle imaging apparatus 100 using the example of a particle microscope in the form of a scanning electron microscope. The scanning electron microscope 100 comprises a particle source 1 , which creates an electron beam 122, wherein the electron beam 122 passes through a condenser lens 4, an xy-stigmator 24, a beam deflector 25 and an objective lens 11 , such that the electron beam 122 is focused onto the surface of an object 8 or a sample 8, the sample or the object 8 being held by a sample stage 15. The working distance WD between the lower end of the objective lens 11 and the surface of the sample 8 is likewise depicted.

In the example shown, the condenser lens 4 is a magnetic condenser lens comprising a pole shoe 21 and a coil 23, the excitation of which is effected by means of the controller 20.

The xy-stigmator 24 is an electrostatic stigmator comprising a plurality of eight electrodes, for example, which are arranged around the electron beam 122, the electrodes being controlled or excited by means of the controller 20. In the example shown, four of these electrodes form an x-stigmator and four other electrodes of this total of eight electrodes form a y-stigmator. Both the x-stigmator and the y-stigmator can create a quadrupole field by means of their associated four electrodes. Alternatively, the xy-stigmator may comprise eight coils for generating a magnetic field, each of which is controlled by means of the controller 20 in order to create a quadrupole field in each case for the x-stigmator or the y-stigmator. The xy- stigmator 24 thus provides the functions of both the x-stigmator and a y-stigmator combined in one component, in order to influence or to set or to correct an astigmatism of the particle beam 122.

The beam deflector 25 may be a magnetic or an electrostatic beam deflector, which is in turn controlled by means of the controller 20 in order to scan the incidence location of the particle beam 122 across the surface of the object 8. In the example shown, an electron detector 17 is provided in order to detect secondary electrons and backscattered electrons emanating from the incidence point of the electron beam 122 on the object surface 8. The controller 20 is configured to assign measured electron intensities measured by means of the electron detector 17 with the incidence locations of the particle beam 122 on the object surface 8 according to the state of the beam deflector 25, in order to record a particle-optical image of the object 8 or the object surface thereof.

The objective lens 11 comprises a pole shoe 29 and a coil 31 , which can be controlled or excited by means of the controller 20. Additionally, the particle beam column of the example shown comprises an electrostatic objective lens, the lens effect of which is created by applying a voltage to an end region of the beam tube 10 (only parts of the beam tube are depicted in Fig. 1) in the interior of the magnetic objective lens 11. The applied voltage is controllable by the controller 20. The pole shoes 29 of the objective lens 11 are preferably at a different potential, in particular at earth potential. The condenser lens 4 and the objective lenses 10, 11 focus the particle beam 122 in such a way that the minimum beam diameter, i.e. therefore the beam focus, is imaged in a manner focused at the working distance WD from the objective lens 11 , provided that the x-stigmator and the y-stigmator are adapted such that the beam cross section has a circular cross section in the region of the beam focus. A sharp particle- optical image may be recorded precisely when the surface of the object 13 coincides with the beam focus, i.e. the object 13 is at the correct distance from the objective lens 11 .

The sample stage 14 comprises an actuator (not illustrated), which can likewise be controlled by means of the controller 20, the actuator being configured to set the position of the object 13 or of the object surface in the z-direction or in the direction of the particle beam 122. Therefore, the controller 20 can vary firstly the excitation of the condenser lens 4 and of the objective lens 11 or the objective lenses 10, 11 or else the position of the object 8 in the z-position by means of the actuator in order to create the beam focus exactly on the object surface.

Figure 2 schematically shows a portion of a particle imaging apparatus 100 having an objective lens system, which comprises a magnetic objective lens 11 and an electrostatic objective lens 10. Moreover, the optical analogue of the electrostatic objective lens is depicted by reference sign 10' in Figure 2, and the optical analogue to the magnetic objective lens is depicted by reference sign 1T. The effect of the magnetic lens may be controlled by means of the controller 20, to be precise by an appropriate excitation of the coil 31 . The magnetic field emerges from the pole shoes 29 at the opening of said pole shoes, which are earthed, and develops its particle-optical effect, indicated by 1T. The electrostatic lens is formed by a high potential applied to the end region of the beam tube 10, while earth potential, i.e. II = 0 V, is applied to the sample or the object 8, for example a wafer, in the example shown. In the example shown, the voltage applied to the end region of the beam tube 10 is 8 kV; although it may also be more than this, or less, it is typically in the kV range.

In the example shown, the magnetic field of the magnetic objective lens 11 and the electrostatic field of the electrostatic objective lens 10 are each arranged within the objective as a physical object. During operation, a charged particle beam 122 successively passes through the magnetic objective lens 1T first and then the electrostatic objective lens 10'. For constructional reasons, these two particle-optical lenses are never exactly 100% aligned to one another. Slight shifts of the particle-optical lenses to one another may now contribute to a lateral focal shift fs when the focusing setting of the magnetic objective lens 11 is modified. For setting purposes there is a periodic variation when controlling the electrostatic objective lens 10'. Should a particle beam pass centrally through the electrostatic lens 10', this periodic modification to the electrostatic field changes only the sharpness of the imaging but not the position of the particle beam on the object 8. The magnetic field strength of the magnetic objective lens 11 , 1T is normally not modified during this setting process. By contrast, the strength of the magnetic field of the magnetic objective lens 11 is modified at a later time, when a different working distance WD or a different focusing setting z is intended to be set by the particle imaging apparatus 100. In this context, a lateral focal shift fs may arise in the context of an adjustment or alignment of the particle-optical lenses 10', 1T of the objective lens system that is not 100% exact, and said lateral focal shift may lead to problems or distortions in the particle-optical images or in a 3D volume image, especially in metrological applications and especially when examining 3D objects. In principle, the arrangement of electrostatic objective lens 10' and magnetic objective lens 1T may also be reversed. That is to say, the particle beam would be adjusted such that it passes centrally through the magnetic objective lens 1T. However, the arrangement shown in Fig. 2 is slightly better since image aberrations tend to depend more strongly on the electrostatic objective lens 10' than on the magnetic objective lens 11', and so the particle beam is optimally adjusted in relation to the electrostatic objective lens.

Figure s schematically shows a portion of a particle imaging apparatus 100 with an electromagnetic immersion field at the object 8. In the depicted exemplary embodiment, the magnetic objective lens 11 is the only objective lens. In the example shown, the pole shoes 29a, 29b of the magnetic objective lens 11 are earthed or are at earth potential, II = 0 V. However, a negative voltage is applied to the sample 8 in the example shown, e.g. Us = -4 kV. Thus, an electrostatic lens, the optical analogue of which is depicted by reference sign 10' in Figure 3, is created between the lower pole shoe 29b and the sample surface 8. Thus, two particle-optical lenses are also provided in succession in this embodiment variant with an electrostatic immersion field, specifically the magnetic objective lens 11 or its optical analogue 1T first and, following this, the electrostatic lens or its optical analogue 10'. Naturally, the case that the two particle-optical lenses 11', 10' are not aligned exactly to one another, i.e. their lens centers or axes thus do not exactly correspond to one another, may also occur here. It is also possible to combine the embodiments of Figures 2 and 3. Thus, e.g. 8 kV would then be applied to the end region of the beam tube 10, and e.g. a voltage of Us = -4 kV would then be applied to the sample. For the examples depicted in Figures 2 and 3, Figure 4 shows the focal shift fs, which may occur in each case: First of all, the axis A1 is depicted; it runs centrally through the middle of the magnetic objective lens 11 or its optical analogue 1T. Moreover, the associated axis of the electrostatic lens or its optical analogue 10' is depicted and denoted by A2. In the example shown, these two axes are shifted parallel to one another. The axis A2 also corresponds to the particle-optical axis Z of the particle imaging apparatus 100. A particle beam 122 moving along the axis A2 does not pass centrally through the middle of either lens, which is why it is deflected. Following the traversal of the lens system 11 ', 10', this offset is expressed by the lateral focal shift fs.

Figure 5 schematically illustrates the occurrence of the lateral focal shift fs in the event of a modification of the focusing setting in the z-direction: Figure 5a depicts the optimal case in which there is no focal shift fs when changing the focusing setting z by +/-dz: Instead, only the foci f1, f2 and f3 are displaced on the z-axis in the event of a modified focusing setting. Something else happens in the real case that is depicted in Figure 5b: When the focusing setting z is modified by +/-dz, the foci f 1 , f2 and f3 are shifted not only along the z-direction but also laterally in a direction y. In this case, the position in the y-direction depends on the respective focusing setting z. In this context, attention should be drawn to the fact that the lateral focal shift fs occurs within a plane that is orthogonal to the particle-optical axis Z, i.e. the entire focal shift fs coincides only with the y-direction in Figure 5b per chance or by definition.

Figure 6 schematically illustrates a focal shift fs due to an electrostatic immersion field at a sample 8 with a wedge-like cutout. For example, the wedge-like cutout can be created using a focused ion beam FIB. The cut angle GF vis-a-vis the planar surface of the sample 8 is likewise plotted in the example shown. The overall cut depth is T, and the width of the cutout B is likewise depicted. In the example shown, a voltage of -1000 V is applied to the sample 8. The equipotential lines 123.1 , 123.2, 123.3 and 123.4 are depicted above the sample surface. They are slightly pulled into the wedge-like cutout or do not extend parallel at said location to the otherwise flat outer surface of the object 8. Accordingly, charged particle beams 122 are accordingly deflected to the side when approaching the angled sample surface. While the particle beam 122.1 depicted by way of example is still incident on the sample 8 perpendicularly to all intents and purposes, the charged particle beam 122.2 is deflected to the left in the example shown; its incidence location is displaced slightly in the lateral direction. This lateral focal shift fs2 is also plotted in Figure 6. By contrast, the third individual particle beam 122.3 is deflected in the other direction just above the sample surface; this gives rise to a lateral focal shift fs3 to the right in the example shown. The depicted lateral focal shift fs thus is exclusively the result of the electrostatic immersion field present at the sample. In this context, the effect itself is dependent on the geometry of the cutout in the sample 8 and the voltage applied to the sample.

Figure 7 schematically illustrates problems when determining the inclination of a 3D structure on account of a lateral focal shift fs present. Figure 7a depicts a first 3D structure 60, which extends in the depth direction or z-direction. A first particle beam 122.1 now experiences a lateral focal shift fs during a focusing setting change to a greater depth. Without taking account of this lateral focal shift fs, the 3D structure 60 would be assumed to be inclined from associated particle-optical images or a volume image of the 3D structure 60. However, it is not inclined in reality. Figure 7b shows the opposite case: In this case, too, a change in the depth focusing setting z is also accompanied by a lateral focal shift fs for the charged particle beam 122.2 depicted by way of example. Although the 3D structure 61 is inclined in reality, the inference drawn without knowledge of the lateral focal shift fs is that the 3D structure 61 is straight or arranged without inclination. By way of example, this shows the importance of a calibration of a lateral focal shift fs depending on a focusing setting or focusing setting change dz.

Figure 8 schematically shows the arrangement of a sample 8, 28. In Figure 8a, a wafer 28 is arranged on a sample stage 15. It lies flat on the sample stage 15. A mini stage 18 is additionally provided for relatively small samples 8 and in addition to a height adjustment also includes a tilt adjustment for arranging a sample 8 in a sample space. This is illustrated in Figures 8b and 8c. In Figure 8b, a sample 8 is depicted on a mini stage 18 which is non-tilted. The sample surface consequently extends in the y-direction; the z-direction corresponds to the sample normal. In Figure 8c, the sample 8 on the mini stage 18 has been tilted through the angle 0. Thus, the height z of the sample surface changes in the y-direction. These different setting options can be made use of in a calibration method according to the invention:

Figure 9 schematically shows raster-scanning of a sample 8 for different arrangements of the sample 8. An accurately known test structure, whose dimensions are known accurately, is present on the sample surface 8a. Two-dimensional test samples of this type already exist. Thus, it is possible to resort to this type of test samples for calibration purposes.

In Figure 9a, the sample 8 is not tilted, i.e. the normal to the sample surface 8a) corresponds to the z-direction. Moreover, the height of the sample surface does not change in the y- direction. The individual particle beams 121.1 , 121.2 and 121.3 depicted by way of example are each incident in focus on the sample surface 8a at the same height z. By contrast, in Figure 9b, the sample 8 is tilted through the angle 0 with its sample surface 8a. Now, in the event of a scan in the y-direction, the particle beam is only still incident in focus on the sample surface 8a at the y-position y3. Structures of the sample recorded at the positions y1 and y2 are rendered unsharp. Nevertheless, a position of the test structure may be ascertained in principle if the tilt angle 0 is known and if the structure of the surface 8a is known. If, in the example of Figure 9b, the depth focus setting is now modified in the z-direction but the angle 0 is not modified, then shifts in the test structure observed after the raster-scanning can be traced back to a lateral focal shift fs. This may optionally be repeated for different depth focusing settings z, e.g. for z1 , z2 and z3.

Figure 10 schematically shows a flowchart for a calibration method according to the invention, according to a first aspect of the invention.

A particle imaging apparatus 100 that operates with at least one charged particle beam 122 is provided in a first method step S1. For example, the charged particles can be electrons, positrons, muons or ions or other charged particles. Moreover, the particle imaging apparatus 100 may take different forms, for example a particle microscope, an SEM, a TEM, an STEM, an SEM-STEM, a dual beam system, a multi-beam particle microscope or a mask repair system, or another form.

A test object 8 is provided in a method step S2, the test object having a test structure on its surface 8a. This test structure has geometric dimensions that are accurately known, for example a special pattern made of mutually orthogonal line structures. The test structures of the test object 8 may be produced with great accuracy and therefore serve as a reference object.

The test object 8 is arranged in an object plane of the particle imaging apparatus 100 in the particle-optical beam path of the particle imaging apparatus 100 in method step S3, in such a way that a surface normal of the test object 8 makes an angle 01 with the particle-optical axis Z of the particle imaging apparatus 100. This angle 01 may be 0° but need not be 0°. Given an angle 01 of 0°, the at least one charged particle beam 122 would be incident on the test structure of the test object 8 at right angles or in telecentric fashion.

A first focusing setting z1 of the particle imaging apparatus 100 is set in a method step S4, in such a way that the at least one particle beam 122 of the particle imaging apparatus 100 is focused onto the surface of the test object 8, at least at points. If the angle 01=0°, then the first focusing setting z1 is set such that the surface 8a of the test object 8 is in focus not only at a few points but at multiple points and in particular at all points on the test object 8 that should be scanned at a later stage.

Raster-scanning or scanning over the test structure with the particle beam 122 with the first (fixed) focusing setting z1 is implemented in a method step S5, and a first particle-optical image of the test structure is created. In the context of this application, a particle-optical image is understood to mean an image which is created by means of the detection of particle radiation. A particle detector is used to detect the particle radiation, and it is possible to combine a particle detector with an optical detector. The detected particles can be for example electrons or other charged particles. It is possible for example to detect secondary electrons or backscattered electrons or mirror electrons or mirror ions for the creation of the particle-optical image. The particle-optical images are images in which the intensity of the detected particle radiation is represented as a function of an image position for a pixel.

A first position P1 of the test structure in the first particle-optical image is determined in a method step S6. This first position P1 of the test structure may denote the overall position of the test structure. However, it is also possible that the first position P1 of the test structure denotes only one position of multiple positions of a part of the test structure. Moreover, it is of course possible that multiple first positions P1 of the test structure are determined, from which it is then for example also possible to derive the overall position P1 of the test structure.

The test object 8 is arranged in the particle-optical beam path of the particle imaging apparatus 100 in a method step S7, in such a way that a surface normal of the test object 8 makes a second angle 02 with the particle-optical axis Z of the particle imaging apparatus 100, wherein this second angle 02 differs from the first angle 01 , i.e. 02 01. By preference, the second angle 02 is an angle that differs from 0°. Thus, this preferably is a true oblique position of the surface of the test object 8 with respect to the particle-optical axis Z of the particle imaging apparatus 100.

The focusing setting is changed and a second focusing setting z2 of the particle imaging apparatus 100 is set in a method step S8, in such a way that a particle beam 122 of the particle imaging apparatus 100 is focused onto the surface of the test object, at least at points, wherein this second focusing setting z2 differs from the first focusing setting z1 , i.e. z2 z1 applies.

Raster-scanning over the test structure with the particle beam 122 with the second fixed focusing setting z2 is implemented in a method step S9, and a second particle-optical image of the test structure is created. A second position P2 of the test structure in the second particle-optical image is determined in a method step S10. Thus, the test structure is imaged multiple times overall, to be precise with different focusing settings and angles. In so doing, the statements already made as regards the determination of the first position P1 of the test structure in the first particle-optical image in principle apply to the determination of the second position P2 of the test structure in the second particle-optical image.

A positional shift d21 between the second position P2 of the test structure in the second particle-optical image and the position P1 of the test structure in the first particle-optical image is determined in a method step S11. Naturally, it is once again possible that multiple positional shifts d21 are determined overall for different partial structures or characteristic points of the test structure. However, there is at least one determination of a positional shift d21.

A focal shift fs of the particle beam 122 in a direction orthogonal to the particle-optical axis Z is determined in a method step S12 depending on the focusing setting change Z = z2 - z1 , with the determination of the focal shift fs being implemented on the basis of the determined lateral positional shift d21 of the test structure and on the basis of the angular difference A0 = 02 - 01 . Thus, the focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z is a lateral focal shift. This lateral focal shift is the cause of distortions that occur in the event of a focusing setting change. Since the test structure or test structures of the test sample 8, which represents a reference sample, are accurately known, the lateral focal shift fs can be inferred from the measured positional shifts d21. According to a preferred embodiment of the invention, the focal shift fs is determined on the basis of a discrepancy between an expected positional shift d21' of the test structure and the actual positional shift d21 of the test structure that was ascertained in step (1 k). The angular difference A0 = 02 - 01 is known, allowing inference about the expected positional shift d21' of the test structure. Deviations from this can be ascribed to the lateral focal shift fs.

The calibration method ends in step S13.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying. The two directions x1 and x2 are preferably orthogonal to one another, and the directions may also be referred to as x-direction and y-direction in a Cartesian coordinate system. In the context of this patent application, a y-direction denotes the direction in which the sample 8 is tilted, especially for a tilted test sample 8 or sample 8. In this respect, reference is initially made to two directions x1 and x2, in which the focal shift fs is ascertained in each case.

According to a preferred embodiment of the invention, the first angle is 01 = 0°. This allows particularly simple determination of the focal shift fs.

According to a preferred embodiment of the invention, method steps S8 to S12 are carried out repeatedly for a further focusing setting or for multiple further focusing settings. Thus, the corresponding method steps are repeated for further focusing settings z3, z4, z5 and so on, and corresponding positional shifts d31 , d41 , d51 and so on are ascertained. This allows a further determination of the lateral focal shift fs depending on different focusing setting changes or focusing settings. In this way, the particle imaging apparatus 100 may be calibrated for large changes in the focusing setting Az, which are required during an inspection of 3D features, for example channel structures 60, 61 in semiconductor samples.

According to a preferred embodiment of the invention, the following relation applies to the constant value dz: 1 pm < dz < 50 pm. Depending on the focal ranges or focal variations over which a particle imaging apparatus 100 should be settable overall, the incremental value dz may be chosen in matched fashion.

According to a preferred embodiment of the invention, the test object 8 is arranged on a mini stage 18 with a settable tilt angle 0, and the method moreover comprises tilting of the mini stage 18, in particular for setting the angle 02. The angle 01 may for example be 0° for a nontilted mini stage.

According to a further preferred embodiment of the invention, the calibration method according to the invention is carried out multiple times for different angles 0. This may be achieved particularly easily using the described mini stage 18. For example, the method may be carried out for 3, 4, 5, 10 or even more angles 0i. As a result, the accuracy of the calibration method may be increased overall.

Figure 11 schematically shows raster-scanning of a sample 8 for different arrangements of the sample 8 with inclination-matched focusing settings z or adaptive focusing settings. The focusing setting z is changeable in the y-direction in Figure 11b) - unlike in Figure 9 - and matched to the inclination 0. If the inclination 0 is known, the required focusing setting z as a function of a raster-scan position Rij is naturally also known in principle. In this respect, a sample surface 8a) with a known test structure can in each case be raster-scanned in focus. However, distortions that may be observed in the associated particle-optical image possibly occur on account of a lateral focal shift fs in the case of this raster-scanning that is regular per se.

Figure 12 schematically shows a flowchart for a calibration method according to the invention, according to a second aspect of the invention, wherein work is carried out with an inclination- matched focusing setting:

A particle imaging apparatus 100 that operates with at least one charged particle beam 122 is provided in a first method step S20. Incidentally, statements already made in conjunction with the first aspect of the invention apply to the particle imaging apparatus 100.

A test object 8 is provided in a method step S21 , the test object 8 having a test structure on its surface 8a. As regards the details of a test object 8 or a test structure, reference is made to the statements already made in conjunction with the first aspect of the invention.

The test object 8 is arranged in an object plane of the particle imaging apparatus 100 in the particle-optical beam path of the particle imaging apparatus 100 in a method step S22, in such a way that a surface normal of the test object 8 makes a first angle 01= 0° with the particle- optical axis Z of the particle imaging apparatus 100.

A first focusing setting z1 of the particle imaging apparatus 100 is set in a method step S23, in such a way that the particle beam 121 of the particle imaging apparatus 100 is focused onto the surface of the test object 8.

Raster-scanning over the test structure with the at least one particle beam 122 with the fixed first focusing setting z1 is implemented in a method step S24, and a first particle-optical image of the test structure is recorded. Up until this point, the calibration method according to the second aspect of the invention may correspond to method steps S1 to S5 according to the first aspect of the invention, provided the angle 01 is 0° there.

The test object 8 is arranged in the particle-optical beam path of the particle imaging apparatus 100 in a method step S25, in such a way that a surface normal of the test object 8 makes a second angle 02, which differs from the first angle 01 , with the particle-optical axis Z of the particle imaging apparatus 100, i.e. wherein 02 01 applies.

Raster-scanning over the test structure with the particle beam 122 with a variable focusing setting AZ that is adapted to the respective raster-scan position Rij of the particle beam 122 is implemented in a method step S26, and a second particle-optical image of the test structure is recorded. In other words, the variable focusing setting Z is matched to the tilt of the test object 8 such that (at least without lateral focal shift fs) the raster-scanning particle beam would in theory be incident on the surface 8a of the test object 8 in focus at each raster-scan position Rij.

A distortion of the test structure in the second particle-optical image relative to the first particle- optical image is determined in a method step S27. Once again, the distortion of the test structure may be determined on the basis of determining positional changes of the test structure or of parts thereof, or else on the basis of deviations from the regularity of the test structure. In principle, there are a number of options available to this end, and these are known to a person skilled in the art.

A focal shift fs of the particle beam 122 in a direction orthogonal to the particle-optical axis Z, or in the depth direction T, is determined in a method step S28 depending on a focusing setting change AZ in the direction of the particle-optical axis Z with the determination of the focal shift fs being implemented on the basis of the determined distortion of the test structure and on the basis of the angle 02. In this case the focusing setting change AZ corresponds to the variable focusing setting AZ that in each case is known or has been adaptively set for a respective raster-scan position Rij of the particle beam.

The method ends in step S29.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying. What may apply in particular in this case is that the two directions x1 and x2 are in turn orthogonal to one another. One of the directions may correspond to a y-direction, into which the test object 8 has been tilted. According to a preferred embodiment of the invention, the test object 8 is arranged on a mini stage 18 with a settable tilt angle, and the method moreover comprises tilting of the mini stage 18 for setting the angle 02.

According to a preferred embodiment of the invention, the following relation applies to the edge length L of a field of view FOV raster-scanned with the particle beam 122: L > 10 pm. In this way, a sufficiently large focusing setting change AZ can be represented during the rasterscanning process. Moreover, a sufficiently large region of the test object 8 is represented, allowing a sufficiently accurate determination of the arising distortion.

Figure 13 schematically shows a focal shift during a refocusing. This effect may likewise be used for calibration purposes: In Figure 13a, the particle beam 122 is focused onto the sample surface 8a, to be precise at an exactly known position. For example, a characteristic feature or a position mark may be used to this end. Now, the focusing depth is modified in Figure 13b, to be precise by the value Z, without a lateral focal shift fs occurring in the process. For example, using a particle imaging apparatus 100 of the type described at the outset, this may be achieved by a modified acceleration voltage EHT at the particle source. To this end, it is not necessary to modify the magnetic lens 11 or the optionally present electrostatic lens component 10. The adjustment by way of the assembled objective lens system 11 , 10 is maintained when the acceleration voltage EHT is modified. The change in the acceleration voltage EHT is indicated in Figure 13 by the specification AV 0V.

Refocusing is now implemented in Figure 13c, to be precise exclusively by modifying the control of the magnetic objective lens 11 : This results in a lateral focal shift fs. The absolute value of this lateral focal shift fs depends on the refocusing Az.

Figure 14 schematically shows a flowchart for a calibration method according to the invention, according to the fourth aspect of the invention as has already been sketched out in principle in Figure 13:

A particle imaging apparatus 100 is provided in a method step S30. In this context, the particle imaging apparatus 100 comprises a particle source 1 for creating a particle beam 122 with charged particles, with an acceleration voltage EHT being applied to the particle source 1 during operation. Moreover, the particle imaging apparatus 100 comprises at least two focusing lenses that are traversed by the particle beam 122 and comprise a magnetic objective lens 11 , 1T on the one hand and an electrostatic lens 10, 10' on the other hand, with the electrostatic lens 10' being arranged downstream of the magnetic objective lens 1T in the direction of the particle-optical beam path, the magnetic objective lens 11 , 1T and the electrostatic lens 10, 10' focusing the particle beam 122 onto an object 8 at a distance z from the magnetic objective lens. The particle imaging apparatus 100 furthermore comprises a deflector unit 24 that is configured to deflect the particle beam 122 towards the centre of the electrostatic lens 10, 10'. Moreover, the particle imaging apparatus 100 comprises an object stage 15 or object holder that is configured to hold an object 8 at a working distance WD from the magnetic objective lens 11. Moreover, the particle imaging apparatus 100 comprises a detection unit 17 for detecting interaction particles that emanate from the object 8, for example for detecting secondary electrons. Moreover, the particle imaging apparatus 100 comprises a controller 20 for controlling the particle imaging apparatus 100. In this case, the controller 20 is configured to control the particle source 1 for the provision of the acceleration voltage EHT. The controller 20 is moreover configured to control the magnetic objective lens 11 or for example to vary the current in a coil 31 of the magnetic objective lens 11. Moreover, the controller 20 is configured to control the deflector unit 24, whereby the particle beam 122 is deflectable exactly onto the centre of the electrostatic lens 10, 10', and this serves the adjustment of the particle imaging apparatus. Moreover, the particle imaging apparatus 100 is adjusted into such an energyindependent state that a change in the acceleration voltage EHT has no influence on the beam position upon incidence on the object 8. In the case of such an adjustment, a change in the acceleration voltage EHT only has the change of a beam diameter on the object 8 as a consequence, i.e. the image sharpness of the recorded particle-optical image is modified but not the beam position.

Thereupon, an object 8 is provided at the distance z in a method step S31. Once again, this object 8 may be a test object 8 or a test sample 8 with regular structures; however, it may also be a different object with a characteristic feature, for example with a position mark on a sample surface.

A reference feature of the object 8 is located in a step S32.

A first particle-optical image is recorded by means of the focusing setting z in a step S33. Thus, the focusing setting z in this case corresponds exactly to the working distance WD, which is why the recorded particle-optical image is in principle recorded in focus. A position P1 of the reference feature in the first particle-optical image is determined in a method step S34.

The acceleration voltage EHT is modified by an offset AV in a method step S35.

The particle imaging apparatus 100 is refocused by Az in a method step S36 by modifying the control of the magnetic objective lens 11. As described, the consequence of modifying the acceleration voltage EHT by the offset AV is that the particle beam 122 is no longer imaged on the object 8 in focus but in blurred fashion. However, the position of the particle beam 122 upon incidence on the object 8 was left unchanged in the process. By contrast, there may be a change in the position of the particle beam 122 on the object 8 as a result of refocusing the particle imaging apparatus or refocusing the particle beam 122 by Az by modifying the control of the magnetic objective lens 11 if the magnetic objective lens 11 , 1T and the electrostatic lens 10, 10' are not aligned exactly to one another; the latter is practically never the case with greatest accuracy. By contrast, the control of the electrostatic lens 10 is preferably not modified when refocusing the particle imaging apparatus 100 by Az by modifying the control of the magnetic objective lens 11. Possible lateral shifts or focal shifts fs can then be traced back exclusively to the changed control of the magnetic objective lens 11 , enabling a calibration.

A second particle-optical image is recorded in method step S37 in the setting of the particle imaging apparatus 100 that has been refocused by Az.

A position P2 of the reference feature in the second particle-optical image is determined in method step S38.

A positional shift d21 between the second position P2 of the reference feature and the first position P1 of the reference feature is determined in method step S39. Naturally, it is also possible to determine multiple positional shifts of multiple reference features.

A focal shift fs of the particle beam 122 in a direction orthogonal to the particle-optical axis Z is determined in method step S40 depending on the refocusing Az by means of the magnetic objective lens 11 , with the determination of the focal shift fs being implemented on the basis of the determined positional shift d21 of the reference feature. In particular, the positional shift d21 may correspond exactly to the focal shift fs in the process.

According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying. In particular, the two directions x1 and x2 may once again be orthogonal to one another.

According to a preferred embodiment of the invention, method steps S35 to S40 are performed repeatedly for different offsets AV in the acceleration voltage EHT. This enables a more accurate calibration and a calibration over a range that is as large as possible.

According to a preferred embodiment of the invention, the offset AV is in each case modified incrementally by a constant value dV. As a result, the calibration method can be executed particularly easily in automated fashion.

According to a preferred embodiment of the invention, the electrostatic lens is an electrostatic objective lens 10'. What is possible here in particular is that this electrostatic lens is provided by the provision of a potential at an end region 10 of a beam tube within the magnetic objective lens 11 (a so-called "Gemini lens"). In this case, each pole shoe 29 of the magnetic objective lens 11 is at earth potential; the object 8 to be examined is preferably likewise at earth potential. Thus, work in this case is not carried out with an electrostatic immersion field.

According to an alternative embodiment of the invention, the electrostatic lens 10' develops its lens effect by means of a potential difference between the object 8 on the one hand and the magnetic objective lens 11 on the other hand. For example, this may be achieved by virtue of the object 8 being at a potential that differs from 0 while the magnetic objective lens is at earth potential. For example, a sample voltage of up to -4 kV may be applied to a wafer 8 should the charged particles of the particle beam be electrons.

However, it is also possible that the electrostatic lens 10' according to the third aspect of the invention has a different embodiment. It is also possible to combine the two above embodiment variants of electrostatic lenses 10'.

Figure 15 shows a schematic illustration of the usual cross-sectional image method for obtaining a 3D volume image of an integrated semiconductor sample 8. Using the cross- sectional method, also referred to as slice and image method, a three-dimensional (3D) volume image acquisition is obtained according to the "step and repeat" principle. Initially, the integrated semiconductor sample 8 is prepared for the conventional cross-sectional image method using known methods. A small block or a piece is taken from a wafer and then subjected to further analysis. Hereinafter, the terms "cross-sectional image" and "slice" are used as synonyms. A thin surface layer of the material is removed in one step. This material may be removed using various ways known from the art, including the use of a focused ion beam for milling or polishing at an angle by way of a focused ion beam (FIB) column 50. For example, the focused ion beam 51 propagates virtually parallel to the z-axis and is scanned in the y-direction in order to mill the upper surface 55 of the sample 8 (which may be part of a wafer, for example) and expose a new cross-sectional surface 52 in a yz-plane. As a result, the newly exposed cross-sectional surface 52 is accessible to imaging. In a further step, the cross-sectional surface 52 is scanned using an imaging system with a charged particle beam (CPB) 40, e.g. a scanning electron microscope (SEM) or a second FIB, in order to obtain a cross-sectional image 1000.1. The optical axis 42 of the imaging system for charged particles 40 may be arranged parallel to the x-direction or inclined at an angle to the x-direction. Both secondary electrons and backscattered electrons are captured by a detector (not depicted here) in order to indicate a material contrast within the integrated semiconductor sample, and these are visible in the cross-sectional image 1000.1 as different greyscales. Metal structures create lighter measurement results. The surface layer removal by milling and the creation of a cross-sectional image are repeated for the cross-sectional surfaces 53 and 54 and further cross-sectional surfaces at the same spacing d, and a sequence of 2D cross-sectional images 1000, which for example consists of N cross-sectional image slices 1000.2, 1000.3, ... 1000.N at different depths, is produced in order to create a three-dimensional 3D data record. The representative cross-sectional image detail 1000.1 is obtained by measurement of a commercially available processor-integrated semiconductor chip from Intel with 14 nm technology.

In the method, at least first and second cross-sectional images are created by virtue of the cross-sectional surfaces being successively milled into the integrated semiconductor sample using a focused ion beam in order to expose or make accessible a sequence of cross-sectional surfaces to imaging, and each cross-sectional surface of the integrated semiconductor sample is imaged using an imaging system with charged particle beam 40. A 3D image of the integrated semiconductor structure is reconstructed from the sequence of N 2D cross-sectional image slices 1000. The spacing d between the cross-sectional images 1000.1 , 1000.2, 1000.3 can be controlled by the FIB milling or polishing process and may be between e.g. 1 nm and 30 nm.

In the example above, the cross-sectional image planes are aligned perpendicular to the upper surface 55 of the integrated semiconductor wafer 8, wherein the normal to the upper surface 55 of the wafer is aligned parallel to the z-direction, as shown in Figure 15. This yields 2D cross-sectional images that are aligned parallel to the yz-plane or, expressed differently, the cross-sectional image planes include the z-axis or the wafer normal axis, and the imaging direction x is parallel to the wafer surface. The conventional slice image method in this conventional geometry is therefore only applicable to samples that have been taken from a wafer.

Naturally, the positions of the FIB column 50 and the CPB imaging column 40 may also be reversed. In that case, the optical axis of the CPB may be aligned perpendicular to the upper surface 55 of the sample 8. This arrangement is advantageous for imaging HAR structures that reach into the depth of the sample 8.

In both of the above-described arrangements, a working distance of the FIB columns and of the CPB imaging columns may be set individually. It is therefore possible to optimize both working distances. As a general rule: the shorter the working distance, the better the image resolution. Thus, high-resolution imaging with the CPB image columns 40 is possible since it is possible to set a short working distance without geometric restrictions.

However, a general problem remains the same: A measurement site on a wafer must be located in the vicinity of an edge of the wafer in order to enable a measurement using the depicted geometric arrangement of the columns 40, 50. Otherwise, the wafer must be destroyed and a sample 8 or a piece must be taken from the wafer in order to artificially create an edge and hence a situation suitable for further analysis.

Figure 16 shows a method for creating a 3D volume image using the slice and image method that is applied to a test volume within a wafer in what is known as the "wedge cut" procedure or wedge cut geometry, without a sample 8 needing to be removed from the wafer. The slice and image method is applied to a test volume with dimensions of a few pm, for example 5 pm to 10 pm, of a lateral extent in 200 mm or 300 mm wafers, without samples needing to be removed from the wafer 8. A groove or edge is milled into the top side 55 of an integrated semiconductor wafer 8 in order to make a cross-sectional surface at an angle to the top side 55 accessible. 3D volume images of test volumes are recorded at a limited number of measurement sites, e.g. at representative sites of dies 6.1 , 6.2, e.g. at process control monitors (PCM) or at sites identified by other testing tools. In the slice and image method, the wafer 8 is destroyed only locally and other dies may continue to be used, or the wafer 8 can continue to be used for further processing.

Figure 16 shows the wafer inspection system 500 that is configured for a slice and image method under wedge cut geometry using a two-beam apparatus 1. Multiple measurement sites, including the measurement sites 6.1 and 6.2, are defined for a wafer 8 in a position plan or in an inspection list, which are generated by an inspection tool or from structural information. The wafer 8 is placed on a wafer support stage 15. The wafer support stage 15 is assembled on a stage 155 with actuators and position control. Actuators and means for precisely controlling a wafer stage, e.g. laser interferometers, are known in the art. A control unit 16 is configured such that it controls the wafer stage 155 and sets a measurement site 6.1 of the wafer 8 at the intersection 43 of the two-beam apparatus 100. The two-beam apparatus 100 comprises an FIB column 50 with an optical FIB axis 48 and an imaging system 40 for charged particle beams (CPB) with an optical axis 42. The wafer surface 55 is arranged at an oblique angle GF to the FIB axis 48 at the intersection 43 of the two optical axes of the FIB and CPB imaging system. The FIB axis 48 and the axis of the CPB imaging system 42 make an angle GFE, and the optical axis of the CPB imaging system 42 makes an angle GE with the perpendicular to the wafer surface 55. In the coordinate system of Figure 16, the normal to the wafer surface 55 is given by the z-axis. The focused ion beam (FIB) 51 is created by the FIB column 50 and is incident on the surface 55 of the wafer 8 at the angle GF. Oblique cross- sectional surfaces at approximately the oblique angle GF are milled into the wafer at the test site 6.1 by ion beam milling. In the example of Figure 16, the oblique angle GF is approximately 30°. The actual inclination angle of the oblique cross-sectional surface may deviate from the inclination angle GF by up to 1° to 4° on account of the beam divergence of the focused ion beam, e.g. a gallium ion beam. Images of the milled surfaces are recorded by a charged particle beam using the imaging system 40 that is inclined to the wafer normal by the angle GE. In the example of Figure 16, the angle GE is approximately 15°. However, other arrangements are also possible, e.g. with GE = GF such that the axis of the CPB imaging system 42 is perpendicular to the Fl B axis 48 or GE = 0° such that the axis of the CPB imaging system 42 is perpendicular to the wafer surface 55.

During the imaging, a beam of charged particles 44 is scanned over a cross-sectional area of the wafer 8 along a scanning path at the measurement site 6.1 by a scanning unit of the imaging system for charged particle beams 40, with secondary particles and scattered particles being created. The particle detector 17 collects at least some of the secondary particles and scattered particles and transmits the number of particles to a control unit 19. Other detectors for other types of interaction products may likewise be present. The control unit 19 controls the imaging column 40 and the FIB column 50 and is connected to a control unit 16 in order to control, by means of the wafer stage 155, the position of the wafer 8 assembled on the wafer support stage 15. The control unit 19 communicates with the operational control unit 20, which triggers the placement and alignment for example of the measurement site 6.1 of the wafer 8 at the intersection 43 by way of the movement of the wafer stage and repeatedly triggers processes of FIB milling, image recording and stage movements. Each new cut surface is milled by the FIB beam 51 and imaged by the charged particle imaging beam 44, which for example is a scanning electron beam or a helium ion beam of a helium ion microscope (HIM).

However, since both the FIB column 50 and the CPB imaging column 40 are positioned above the wafer surface 55, there are geometric restrictions in respect of their arrangement. It is no longer possible to individually optimize both working distances since the columns 40, 50 block one another at already short working distances. What applies as a general practical rule is that the position of the FIB column 50 limits and hence "defines" the working distance of the CPB imaging column 40. Normally, the working distance of the CPB imaging column 40 cannot be chosen to be shorter than 4 to 5 mm. In contrast, the working distance of the CPB image column 40 as depicted in Fig. 16 may be reduced to approximately 2 mm. Therefore, the image resolution for the wedge-shaped arrangement depicted in Fig. 16 is typically one order of magnitude smaller than in the case of the edge arrangement depicted in Fig. 15.

It is also possible to perform a calibration method based on different depths of field for a particle imaging apparatus 100 and a lateral focal shift fs that occurs in the case of a modification to the focus. Figure 17 schematically illustrates different depths of field for an imaging particle beam 122. Figure 17, left, depicts a first particle beam 122.1 , the beam cone of which is focused onto the focal plane F. The beam cone of the particle beam 122.1 has an opening angle cp1 with respect to the particle-optical axis Z. This opening angle cp1 is comparatively small. Hence, the size of the beam spot in the focal plane F changes only slowly in the z- direction. Thus, the particle beam 122.1 has a comparatively large depth of field.

Figure 17, right, depicts a further charged particle beam 122.2 that is also focused onto the focal plane F. Its beam cone has a comparatively large opening angle <p2 with respect to the particle-optical axis Z. Using the focal plane as a starting point, the diameter of a beam spot thus changes more quickly in the z-direction than in the case of the example depicted to the left. Thus, the depth of field of the particle beam 122.2 is comparatively small. The depth of field TS is mathematically defined as the defocus, i.e. the distance in the z-direction from the focal plane F, at which the diameter of the beam spot increases by 10%. The diameter of a beam spot in the focal plane is typically approximately 1 nm to 2 nm. Depending on the opening angle cp, the depth of field is then typically between 200 nm and 500 nm.

For recording with a large depth of field (left in Figure 17), only a lower resolution can be obtained on account of a smaller numerical aperture linked with this case. The situation is different in the example depicted on the right in Figure 17, where the numerical aperture is comparatively large, and so a high resolution is possible.

Figure 18 schematically shows a flowchart for a calibration method according to the invention according to the fourth aspect of the invention, wherein the above-described relationships with the depth of field TS and optionally also the described slice and image method are resorted to:

A particle imaging apparatus 100 that operates with at least one charged particle beam 122 and that may operate in a first mode of operation and in a second mode of operation is provided in a first method step S50. In this case, the particle imaging apparatus 100 works with a first depth of field TS1 and a first resolution A1 in the first mode of operation, and the particle imaging apparatus works with a second depth of field TS2 and a second resolution A2 in the second mode of operation. In this case, the first depth of field TS1 is greater than the second depth of field TS2. Moreover, the first resolution A1 is lower than the second resolution A2. In this case, the depth of field TS is in each case defined as the defocus df at which the diameter of a beam waist increases by 10% (for the smallest, best-possible beam diameter). In general, there is an optimal opening angle, at which a beam diameter is minimal, and the resolution is optimal. The beam diameter increases both for larger opening angles (deterioration as a result of lens aberrations) and for smaller opening angles (deterioration on account of diffraction effects). Within the scope of the invention, a smaller opening angle that benefits a better depth of field but has an attendant increase in the beam waist and hence a non-optimal resolution for image recording is chosen at least intermittently.

A test object 8 with a wedge-like cutout and 3D structures is provided in method step S51 , wherein the cutout is defined by means of an angle GF with respect to the surface of the test object 8. For example, the 3D structures may be channel structures 60, 61 that extend into the depth of the test object over a large distance, for example over more than 10 pm, more than 50 pm or more than 100 pm. The test object with a wedge-like cutout might have already been created before method step (S51) but might likewise only be created during method step (S51). According to a preferred embodiment of the invention, the provision of the test object 8 with a wedge-like cutout and the 3D structures comprises a wedge-like ablation of the test object with the 3D structures by means of a focused ion beam (FIB).

The test object 8 is arranged relative to the particle imaging apparatus 100 in method step S52 such that the particle-optical axis Z of the particle imaging apparatus 100 and the uncut surface of the test object 8 are orthogonal to one another, with the 3D structures being exposed at different depths z of the test object in the direction of the particle-optical axis Z of the particle imaging apparatus 100. Thus, ideally, the direction of the 3D structures corresponds or substantially corresponds to the particle-optical axis Z.

Raster-scanning the test object 8 with a wedge-like cutout in the first mode of operation and with a fixed setting z1=z1 ij for the focusing setting is implemented in method step S53, and a first particle-optical image of the test object with the 3D structures is created. Imaging is thus implemented with a greater depth of field TS1 , wherein the different arrangement of the 3D structures as regards the height or depth z is rather irrelevant to the imaging. The image representation of the 3D structures in the particle-optical image has no distortions; it only varies slightly as regards the sharpness of the depicted 3D structures.

Positions P1 of the 3D structures in the first particle-optical image are determined in method step S54.

Raster-scanning the test object 8 with a wedge-like cutout in the second mode of operation and with an adaptive focusing setting z2ij that is adapted to the respective raster-scan position Rij of the particle beam is implemented in method step S55, and a second particle-optical image of the test object with the 3D structures is created. In this case, the adaptive focusing setting z2ij describes a focusing setting in which raster-scan positions of the test object 8 are focused on, in each case in a manner that is sharp or focused at points, i.e. with a minimal beam waist. Hence, the raster-scanning and creation of the second particle-optical image is implemented with a smaller depth of field TS2 but in return with a better resolution A2 or, expressed differently, with a greater numerical aperture NA2.

Positions P2 of the 3D structures in the second particle-optical image are determined in method step S56.

The positions P1 and P2 are compared to one another in method step S57, and respective shifts d21=P2-P1 of the 3D structures are ascertained. These shifts d21 are the result of a lateral focal shift fs of the particle beam 122 during the respective focusing setting change AZ = z2ij - z1.

A lateral focal shift fs of the particle beam 122 in a direction orthogonal to the particle-optical axis Z is determined in method step S58 depending on a focusing setting change Z = z2ij - z1 , with the determination of the focal shift fs being implemented on the basis of the determined shifts d21 of the 3D structures. According to a preferred embodiment of the invention, the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

According to a preferred embodiment of the invention, method steps S51 to S58 are repeated, in particular repeated multiple times, to be precise for different angles GF. Thus, the method is repeated for a plurality of test objects 8 with different wedge-like cutouts and 3D structures. It is also possible to yet again ablate a first wedge-like cutout such that this gives rise to a different test object with a wedge-like cutout that has a different angle GF. For example, a focused ion beam can be used to this end, in particular according to the "slice and image method".

The method of the shown example ends in a step S59.

Figure 19 schematically illustrates an inclined sample surface 8a with cut 3D structures 60 and an associated particle-optical image 1000.1.

Figure 19 schematically illustrates positional shifts of 3D structures 60 with oblique cutouts in particle-optical images that were generated using two different recording modes. Figure 19a shows a perspective illustration, and Figure 19b shows the corresponding plan view. Shown in each case is an overlay of two particle-optical images in an overlay image 1000.1. The first particle-optical image was recorded in the first mode according to the method of Figure 18, and the second particle-optical image was recorded in the second mode according to the method of Figure 18. In the first mode with a large depth of field, the position of the 3D structures 60.1 with cutouts is very regular or provided in the illustrated example at positions that are represented by dark hatched circles in Figure 19. There was no lateral focal shift fs in this recording of the 3D structure. The situation is different in the recording of the second mode of operation with the smaller depth of field but adapted focusing setting: In this case, the position of the 3D structures 60 is represented by the brightly dotted circles. From the upper row of Figure 19, it is evident that the position of the structures 60.1 is congruent in both recordings (and hence no dark hatching can be seen). The positions of the 3D structures in the two recordings increasingly differ with a focusing setting increasingly modified into the depth or into the z-direction. A lateral positional shift fs is evident most clearly in the lowermost row with the greatest change in focusing setting in the z-direction. This is plotted in Figure 19b in exemplary fashion at the 3D structure 60.2: The lateral focal shift fs is composed of a shift dx in the x-direction and a shift dy in the y-direction, wherein the y-direction coincides with the direction in which the sample 8 is tilted or has an oblique cutout and in which the z-focus setting has been simultaneously adapted for the image recording in the second mode. In the example shown, the lateral focal shift fs that depends on the focusing setting change Az can be ascertained by a line-by-line ascertainment of the lateral shifts dx and dy.

Figure 20 schematically shows a flowchart for a further calibration method according to the invention, to be precise for a calibration method according to the fifth aspect of the invention. This calibration method serves to determine a lateral focal shift fs in a direction orthogonal to the particle-optical axis Z depending on a strength of an electrostatic immersion field in the case of test objects 8 with a wedge-like cutout. Moreover, a calibration may also be undertaken depending on the geometry of the wedge-like cutout.

A particle imaging apparatus that operates with at least one charged particle beam and that may operate in a first mode of operation and in a second mode of operation is provided in a first method step S70. An object to be imaged is not arranged in an electrostatic immersion field in the first mode of operation, while the object to be imaged is arranged in an electrostatic immersion field in the second mode of operation. For example, this may be realized by virtue of a controller of the particle imaging apparatus being configured to apply an electric potential to the object to be imaged or to a sample stage holding the object, or to switch on or else switch off said electric potential.

A first test object with a wedge-like cutout and 3D structures is provided in method step S71 , wherein the cut is defined by means of a first angle GF1 with respect to the surface of the test object.

The test object is arranged relative to the particle imaging apparatus in method step S72 such that the particle-optical axis Z of the particle imaging apparatus and the uncut surface of the test object are in particular orthogonal to one another, with the 3D structures being exposed at different depths Z of the test object in the direction of the particle-optical axis of the particle imaging apparatus.

Raster-scanning the first test object 8 with a wedge-like cutout with a calibrated particle beam 122 in the first mode of operation without electrostatic immersion field is implemented in method step S73, and a first particle-optical image of the test object 8 with the 3D structures is created.

Positions P1 of the 3D structures in the first particle-optical image are determined in method step S74.

Raster-scanning the first test object with a wedge-like cutout with an adjusted particle beam 122 in the second mode of operation with a first electrostatic calibrated immersion field is implemented in method step S75, and a second particle-optical image of the test object 8 with the 3D structures is created.

Positions P2 of the 3D structures in the second particle-optical image are determined in method step S76.

The positions P1 and P2 are compared to one another in method step S77, and respective shifts d21 = P2 - P1 of the 3D structures are ascertained.

A lateral focal shift fs in a direction orthogonal to the particle-optical axis Z is determined in method step S78 depending on a strength of the electrostatic immersion field, with the determination of the focal shift fs being implemented on the basis of the determined shifts d21 of the 3D structures.

In this embodiment of the invention, the lateral focal shift fs is based on small bends in equipotential lines of the electrostatic field at the surface with a wedge-like cutout of the test object or results from the geometry of the test object with a wedge-like cutout. In this case, the lateral focal shift can be traced back exclusively to this effect. A lateral focal shift on account of a non-exact alignment of successive particle-optical lenses 10', 1T has already been corrected or calibrated for in advance. In this respect, raster-scanning in steps S73 and S75 is implemented in each case with an already calibrated particle beam 122 or with a particle imaging apparatus 100 calibrated in this respect.

According to a preferred embodiment of the invention, the lateral focal shift fs is in each case determined for a plurality of raster-scan positions Rij. This is meaningful since the shape of the test object with a wedge-like cutout is defined by way of the respective raster-scan positions. In this case, the raster-scan may be chosen with different degrees of fineness. It must be chosen to be fine enough such that the geometric bends of the potential lines at the surface of the test sample with a wedge-like cutout can be reconstructed accurately enough during the raster-scanning process.

According to a preferred embodiment of the invention, method steps S75 to S78 are carried out repeatedly, in particular in a manner repeated multiple times, to be precise with a different strength of the electrostatic immersion field in each case. In this case it is possible that the strength of the electrostatic immersion field is modified incrementally, for example over a range from approximately 500 V/mm to approximately 4000 V/mm.

According to a preferred embodiment of the invention, method steps S75 to S78 are carried out repeatedly, in particular in a manner repeated multiple times, to be precise for test objects with a wedge-like cutout that have different geometries in each case. In that case, determining the focal shift fs in a direction orthogonal to the particle-optical axis Z is implemented depending on the geometry of the test object with a wedge-like cutout.

According to a preferred embodiment of the invention, the angle GF is in each case varied for the test objects 8 with a wedge-like cutout that have different geometries. In addition to that or in an alternative, the maximum cut depth T is varied for the test objects 8 with a wedge-like cutout that have different geometries. In addition to that or in an alternative, the maximum cut width B is varied for the different test objects 8 with a wedge-like cutout that have different geometries. In this way it is possible to separately ascertain the separate influence of different parameters of the geometry on the focal shift fs. The calibration can thus be implemented very finely. According to an advantageous embodiment of the invention, the various parameters are varied separately and successively, and their influence on the focal shift fs is ascertained separately.

In the example shown, the method ends in a step S79.

According to a preferred embodiment of the invention, the particle imaging apparatus 100 is an apparatus from the following list of particle imaging apparatuses 100: a particle microscope, an SEM, a TEM, an STEM, an SEM-STEM, a dual particle beam system, a multi-beam particle microscope, a mask repair system. In this context, this list should not be construed as exhaustive. The described lateral focal shift may occur in all of the aforementioned particle imaging apparatuses 100, be it on account of an alignment of successive particle-optical lenses that is not 100% exact or else on account of electrostatic or magnetic immersion fields at an object or sample surface.

In this context, in particular an end region of a beam tube that projects into the magnetic objective lens may be at an electric potential, for example at a potential of several kV or keV.

Figure 21 schematically shows various test objects with a wedge-shaped cutout: Figure 21a shows a first test object with a wedge-shaped cutout, with a width B1 and a maximum depth T1. Furthermore, the cutout angle GF1 is plotted.

Figure 21 b depicts a second test object with a different geometry: The cutout angle GF2 is greater than the cutout angle GF1 in Figure 21a. The width B2 corresponds to the width B1 in Figure 21a. The overall depth of the wedge-like cutout T2 is greater than the depth T1 in Figure 21a.

A third test sample 8 with an oblique cutout is depicted in Figure 21c. Its overall width B3 is larger than in the preceding examples; the depth T3 corresponds to the depth T2 from Figure 21b. The cutout angle GF3 has been chosen to be smaller than in the other examples. These variations of geometric parameters can be continued. Moreover, it is naturally possible to also use differently designed geometries instead of the wedge-shaped cutout. However, the wedge- shaped cutouts are naturally particularly practical since these can be produced very well by means of a focused ion beam FIB.

Figure 22 shows, by way of example, a flowchart for a method for operating a particle imaging apparatus 100. According to a sixth aspect of the invention, the latter relates to a method for operating a particle imaging apparatus 100. In this context, the particle imaging apparatus 100 may be a particle imaging apparatus as described above in multiple embodiment variants.

A particle imaging apparatus 100 that operates with at least one charged particle beam 122 is provided in a first method step S80.

The particle imaging apparatus 100 is calibrated in a method step S81 in respect of a focal shift fs of the particle beam 122 in a direction orthogonal to the particle-optical axis Z depending on a focusing setting change Az or a refocusing Az of the particle imaging apparatus 100; in particular, the calibration is implemented according to one of the calibration methods as described above according to the first to fourth aspect of the invention. In addition to that or in an alternative, the particle imaging apparatus 100 is calibrated in respect of a focal shift fs of the particle beam 122 on the basis of the strength of an electrostatic immersion field at an object 8 to be imaged and/or on the basis of a geometry of an object 8 to be imaged in an electrostatic immersion field, wherein this calibration may be implemented in particular according to a calibration method that was described in conjunction with the fifth aspect of the invention, although this need not be the case.

At least one particle-optical image of an object is created by means of the particle imaging apparatus in a method step S82. As a result, a distortion in the particle-optical image on account of a lateral focal shift does not occur in the first place, offering advantages in metrological applications in particular.

According to a preferred embodiment of the invention, the calibrated particle imaging apparatus is used to create a plurality of slice images, aligned parallel to one another, through a 3D structure, and the following further method steps are performed:

A volume image is created from the slice images in a step S83, and a tilt of the 3D structure is determined in a step S84 on the basis of the slice images.

By preference, the calibrated particle imaging apparatus 100 may be a dual particle beam system which for example operates by means of a system made of a particle microscope and a focused ion beam, for example using a slice and image method. According to a preferred embodiment of the invention, the 3D structure comprises a NAND structure.

The method described by way of example ends in step S85.

The various embodiments and aspects of the invention may be combined wholly or partly with one another, provided that no technical contradictions arise as a result. In all other respects, explicit reference is made to the fact that the described exemplary embodiments of the invention should not be construed as limiting the invention.

Multiple calibration methods for particle imaging apparatuses 100 are proposed, which allow a lateral focal shift fs to be determined depending on a focusing setting change Az or a refocusing Az of a particle imaging apparatus 100. Distortions in particle-optical images created by correspondingly calibrated particle imaging apparatuses 100 can be reduced in this way, and metrological examinations in the particle-optical images and in particular in 3D tomographic images may be implemented with greater accuracy. Using correspondingly calibrated particle imaging apparatuses 100, it is possible in particular to measure 3D samples such as deep channels in semiconductor samples more accurately. It is also possible to calibrate and hence correct distortion effects that arise on account of sample geometries in electrostatic immersion fields.

List of reference signs

1 Particle source

4 Condenser lens

6 Measurement site, measurement region, examination region

8 Sample, object, wafer

8a Object surface with structure, test sample

10 End region of the beam tube at a potential, electrostatic objective lens

10' Optical analogue of the electrostatic objective lens

11 Magnetic objective lens

1T Optical analogue of the magnetic objective lens

15 Sample stage

16 Sample stage control unit

17 Detection system

18 Mini stage

19 Control unit

20 Controller

21 Pole shoe

23 Coil

24 xy-stigmator

25 Beam deflector, scanning device

29 Pole shoe

31 Coil

40 Imaging particle beam column

42 Optical axis of the imaging system

43 Intersection

44 Imaging particle beam

46 Scanning path, raster-scanning path

48 Optical axis of the FIB

50 FIB column

51 Focused ion beam

52 Cut surface, surface

53 Cut surface, surface

54 Cut surface, surface

55 Surface of the object, surface of the sample, surface of the wafer

60 3D structure, channel

61 3D structure, channel 100 Particle beam system, particle imaging apparatus, scanning electron microscope, SEM

155 Wafer stage

500 Wafer inspection system

1000 3D image, volume image

1000.1 Cross-sectional image

1000.2 Cross-sectional image

1000.3 Cross-sectional image

WD Working distance

GF Angle, cut angle

GE Angle, angle in the imaging procedure

GFE Angle, arrangement angle between the columns

Us Potential at the object, voltage at the object

U Potential at the end region of the beam tube/in the interior of the objective, voltage at the end region of the beam tube/in the interior of the objective x direction y direction (inclination direction) z direction

Z Particle-optical axis

Claims

1. Calibration method for a particle imaging apparatus, including the following steps:

(3a) providing a particle imaging apparatus comprising the following: a particle source for creating a particle beam with charged particles, with an acceleration voltage EHT being applied to the particle source during operation, at least two focusing lenses that are traversed by the particle beam and comprise a magnetic objective lens on the one hand and an electrostatic lens on the other hand, with the electrostatic lens being arranged downstream of the magnetic objective lens in the direction of the particle-optical beam path, the magnetic objective lens and the electrostatic lens focusing the particle beam onto an object at a distance z from the magnetic objective lens, a deflector unit that is configured to deflect the particle beam towards the centre of the electrostatic lens, an object stage or object holder that is configured to hold an object at a working distance WD from the magnetic objective lens; a detection unit for detecting interaction particles that emanate from the object; and a controller for controlling the particle imaging apparatus, wherein the controller is configured to control the particle source for the provision of the acceleration voltage EHT, wherein the controller is configured to control the magnetic objective lens, and wherein the controller is configured to control the deflector unit, and wherein the particle imaging apparatus is adjusted into such an energyindependent state that a change in the acceleration voltage EHT has no influence on the beam position upon incidence on the object;

(3b) providing an object at the distance z;

(3c) locating a reference feature of the object;

(3d) recording a first particle-optical image by means of the focusing setting z;

(3e) determining a position P1 of the reference feature in the first particle-optical image;

(3f) modifying the acceleration voltage EHT by an offset AV;

(3g) refocusing the particle imaging apparatus by Az by modifying the control of the magnetic objective lens;

(3h) recording a second particle-optical image in the setting of the particle imaging apparatus that has been refocused by Az;

(3i) determining a position P2 of the reference feature in the second particle-optical image;

(3j) determining a positional shift d21 between the second position P2 of the reference feature and the first position of the reference feature; and

(3k) determining a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on the refocusing Az by means of the magnetic objective lens, with the determination of the focal shift fs being implemented on the basis of the determined positional shift d21 of the reference feature.

2. Calibration method according to the preceding claim, wherein the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

3. Calibration method according to either of Claims 1 and 2, wherein method steps (3f) to (3k) are performed repeatedly for different offsets AV of the acceleration voltage EHT.

4. Calibration method according to the preceding claim, wherein the offset AV is in each case modified incrementally by a constant value dV.

5. Calibration method according to any of the preceding Claims 1 to 4, wherein the electrostatic lens is an electrostatic objective lens.

6. Calibration method according to any of the preceding Claims 1 to 4, wherein the electrostatic lens develops its lens effect by means of a potential difference between the object and pole shoes of the magnetic objective lens, in particular wherein the object is at a potential that differs from zero.

7. Calibration method for a particle imaging apparatus, including the following steps:

(4a) providing a particle imaging apparatus that operates with at least one charged particle beam and that may operate in a first mode of operation and in a second mode of operation, wherein the particle imaging apparatus works with a first depth of field TS1 and a first resolution A1 in the first mode of operation, and wherein the particle imaging apparatus works with a second depth of field TS2 and a second resolution A2 in the second mode of operation, wherein the first depth of field TS1 is greater than the second depth of field TS2, and wherein the first resolution A1 is lower than the second resolution A2; (4b) providing a test object with a wedge-like cutout and 3D structures, wherein the cutout is defined by means of an angle GF with respect to the surface of the test object;

(4c) arranging the test object relative to the particle imaging apparatus such that the particle-optical axis Z of the particle imaging apparatus and the uncut surface of the test object are orthogonal to one another, with the 3D structures being exposed at different depths z of the test object in the direction of the particle-optical axis of the particle imaging apparatus;

(4d) raster-scanning the test object with a wedge-like cutout in the first mode of operation and with a fixed setting z1=z1 ij for the focusing setting, and creating a first particle- optical image of the test object with the 3D structures;

(4e) determining positions P1 of the 3D structures in the first particle-optical image;

(4f) raster-scanning the test object with a wedge-like cutout in the second mode of operation and with an adaptive focusing setting z2ij that is adapted to the respective rasterscan position Rij of the particle beam, and creating a second particle-optical image of the test object with the 3D structures;

(4g) determining positions P2 of the 3D structures in the second particle-optical image;

(4h) comparing the positions P1 and P2 to one another, and ascertaining respective shifts d21=P2-P1 of the 3D structures; and

(4i) determining a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on a focusing setting change Az=z2ij-z1 , with the determination of the focal shift fs being implemented on the basis of the determined shifts d21 of the 3D structures.

8. Calibration method according to Claim 7, wherein the provision of the test object with a wedge-like cutout and the 3D structures comprises a wedge-like ablation of the test object with the 3D structures by means of a focused ion beam.

9. Calibration method according to the preceding claim, wherein the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

10. Calibration method according to any of Claims 7 to 9, wherein method steps (4b) to (4i) are carried out repeatedly, in particular in a manner repeated multiple times, to be precise for different angles GF.

11. Calibration method according to any of Claims 7 to 10, wherein the following relation applies to a ratio of first depth of field TS1 to second depth of field TS2: 5 < TS1/TS2 < 30; and/or wherein the following relation applies to a ratio of first resolution A1 to second resolution A2: 2 < A1/A2 < 6.

12. Calibration method for a particle imaging apparatus, including the following steps:

(5a) providing a particle imaging apparatus that operates with at least one charged particle beam and that may operate in a first mode of operation and in a second mode of operation, wherein an object to be imaged is not arranged in an electrostatic immersion field in the first mode of operation, and wherein the object to be imaged is arranged in an electrostatic immersion field in the second mode of operation;

(5b) providing a first test object with a wedge-like cutout and 3D structures, wherein the cutout is defined by means of a first angle GF1 with respect to the surface of the test object;

(5c) arranging the test object relative to the particle imaging apparatus such that the particle-optical axis Z of the particle imaging apparatus and the uncut surface of the test object are in particular orthogonal to one another, with the 3D structures being exposed at different depths z of the test object in the direction of the particle-optical axis of the particle imaging apparatus;

(5d) raster-scanning the first test object with a wedge-like cutout with a calibrated particle beam in the first mode of operation without electrostatic immersion field, and creating a first particle-optical image of the test object with the 3D structures;

(5e) determining positions P1 of the 3D structures in the first particle-optical image;

(5f) raster-scanning the first test object with a wedge-like cutout with an adjusted particle beam in the second mode of operation with a first electrostatic immersion field, and creating a second particle-optical image of the test object with the 3D structures;

(5g) determining positions P2 of the 3D structures in the second particle-optical image;

(5h) comparing the positions P1 and P2 to one another, and ascertaining respective shifts d21=P2-P1 of the 3D structures; and

(5i) determining a focal shift fs in a direction orthogonal to the particle-optical axis Z depending on a strength of the electrostatic immersion field, with the determination of the distortion D being implemented on the basis of the determined shifts d21 of the 3D structures.

13. Calibration method according to the preceding claim, wherein the focal shift fs is in each case determined for a plurality of raster-scan positions Rij.

14. Calibration method according to either of Claims 12 and 13, wherein the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

15. Calibration method according to any of Claims 12 to 14, wherein method steps (5f) to (5i) are carried out repeatedly, in particular in a manner repeated multiple times, to be precise with a different strength of the electrostatic immersion field in each case.

16. Calibration method according to any of Claims 12 to 15, wherein method steps (5b) to (5i) are carried out repeatedly, in particular in a manner repeated multiple times, to be precise for test objects with a wedge-like cutout that have different geometries in each case; and wherein determining the focal shift fs in a direction orthogonal to the particle-optical axis Z is implemented depending on the geometry of the test object with a wedge-like cutout.

17. Calibration method according to the preceding claim, wherein the angle GF in each case varies for the test objects with a wedge-like cutout that have different geometries; and/or wherein the maximum cut depth varies for the test objects with a wedge-like cutout that have different geometries; and/or wherein the maximum cut width varies for the test objects with a wedge-like cutout that have different geometries.

18. Calibration method for a particle imaging apparatus, including the following steps:

(la) providing a particle imaging apparatus that operates with at least one charged particle beam;

(l b) providing a test object, the test object having a test structure on its surface;

(lc) arranging the test object in an object plane of the particle imaging apparatus in the particle-optical beam path of the particle imaging apparatus, in such a way that a surface normal of the test object makes an angle 01 with the particle-optical axis of the particle imaging apparatus;

(1 d) setting a first focusing setting z1 of the particle imaging apparatus, in such a way that the particle beam of the particle imaging apparatus is focused onto the surface of the test object, at least at points; (le) raster-scanning over the test structure with the particle beam with the first focusing setting z1 , and creating a first particle-optical image of the test structure;

(lf) determining a first position P1 of the test structure in the first particle-optical image;

(lg) arranging the test object in the particle-optical beam path of the particle imaging apparatus, in such a way that a surface normal of the test object makes an angle 02 with the particle-optical axis of the particle imaging apparatus, with 02 01 applying;

(1 h) changing the focusing setting, and setting a second focusing setting z2 of the particle imaging apparatus, in such a way that a particle beam of the particle imaging apparatus is focused onto the surface of the test object, at least at points, with z2 z1 applying;

(1 i) raster-scanning over the test structure with the particle beam with the second focusing setting z2, and creating a second particle-optical image of the test structure;

(1 j) determining a second position P2 of the test structure in the second particle- optical image;

(lk) determining a positional shift d21 between the second position P2 of the test structure in the second particle-optical image and the position P1 of the test structure in the first particle-optical image;

(ll) determining a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on the focusing setting change Az =z2 - z1, with the determination of the focal shift fs being implemented on the basis of the determined lateral positional shift d21 of the test structure and on the basis of the angular difference A0 = 02- 01.

19. Calibration method according to Claim 18, wherein the focal shift fs is determined on the basis of a discrepancy between an expected positional shift d21' of the test structure and the actual positional shift d21 of the test structure that was ascertained in step (1k).

20. Calibration method according to claim 18 or 19, wherein the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

21. Calibration method according to any of claims 18 to 20, wherein the first angle is 01 = 0°.

22. Calibration method according to any of claims 18 to 21 , wherein method steps (1h) to (11) are carried out repeatedly for a further focusing setting or for multiple further focusing settings.

23. Calibration method according to the preceding claim, wherein the focusing setting is modified incrementally by a constant value dz.

24. Calibration method according to the preceding claim, wherein the following relation applies to the constant value dz: 1 pm < dz < 50 pm.

25. Calibration method according to any of claims 18 to 24, wherein the test object is arranged on a mini stage with a settable tilt angle, and wherein the method moreover comprises tilting of the mini stage, in particular for setting the angle 02.

26. Calibration method according to any of claims 18 to 25, wherein the method is carried out multiple times for different angles 0.

27. Calibration method for a particle imaging apparatus, including the following steps:

(2a) providing a particle imaging apparatus that operates with at least one charged particle beam;

(2b) providing a test object, the test object having a test structure on its surface;

(2c) arranging the test object in an object plane of the particle imaging apparatus in the particle-optical beam path of the particle imaging apparatus, in such a way that a surface normal of the test object makes an angle 01 = 0° with the particle-optical axis of the particle imaging apparatus;

(2d) setting a first focusing setting z1 of the particle imaging apparatus, in such a way that the particle beam of the particle imaging apparatus is focused onto the surface of the test object;

(2e) raster-scanning over the test structure with the particle beam with the fixed first focusing setting z1 , and recording a first particle-optical image of the test structure;

(2f) arranging the test object in the particle-optical beam path of the particle imaging apparatus, in such a way that a surface normal of the test object makes an angle 02 with the particle-optical axis of the particle imaging apparatus, with 02 01 applying;

(2g) raster-scanning over the test structure with the particle beam with a variable focusing setting Az that is adapted to the respective raster-scan position Rij of the particle beam, and recording a second particle-optical image of the test structure; (2h) determining a distortion of the test structure in the second particle-optical image relative to the first particle-optical image; and

(2i) determining a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on a focusing setting change Az in the direction of the particle-optical axis Z, with the determination of the focal shift fs being implemented on the basis of the determined distortion of the test structure and on the basis of the angle 02.

28. Calibration method according to the preceding claim, wherein the focal shift fs is in each case determined in a first direction x1 orthogonal to the particle-optical axis Z and in a second direction x2 orthogonal to the particle-optical axis Z, with x1 x2 applying.

29. Calibration method according to either of Claims 27 and 28, wherein the test object is arranged on a mini stage with a settable tilt angle, and wherein the method moreover comprises tilting of the mini stage for setting the angle 02.

30. Calibration method according to any of Claims 27 to 29, wherein the method is carried out multiple times for different angles 0.

31. Calibration method according to any of Claims 27 to 30, wherein the following relation applies to the edge length L of a field of view FOV raster-scanned with the particle beam: L > 10 pm, in particular L > 20 pm or L > 50 pm.

32. Calibration method according to any of Claims 1 to 31 , wherein the particle imaging apparatus is an apparatus from the following list of particle imaging apparatuses: a particle microscope, an SEM, a TEM, an STEM, an SEM-STEM, a dual particle beam system, a multi-beam particle microscope, a mask repair system.

33. Calibration method according to any of Claims 1 to 32, wherein the particle imaging apparatus comprises a magnetic objective lens and an electrostatic objective lens.

34. Calibration method according to the preceding claim, wherein the electrostatic objective lens is arranged or formed within the pole shoes of the magnetic objective lens.

35. Method for operating a particle imaging apparatus, including the following steps:

(6a) providing a particle imaging apparatus that operates with at least one charged particle beam;

(6b) calibrating the particle imaging apparatus in respect of a focal shift fs of the particle beam in a direction orthogonal to the particle-optical axis Z depending on a focusing setting change Az or a refocusing Az of the particle imaging apparatus, in particular a calibration according to any of Claims 1 to 11 , 18 to 31 , and/or calibrating the particle imaging apparatus in respect of a focal shift fs of the particle beam on the basis of the strength of an electrostatic immersion field at an object to be imaged and/or on the basis of a geometry of an object to be imaged in an electrostatic immersion field, in particular a calibration according to any of Claims 12 to 17;

(6c) creating at least one particle-optical image of an object by means of the particle imaging apparatus.

36. Method for operating a particle imaging apparatus according to the preceding claim, wherein the calibrated particle imaging apparatus is used to create a plurality of slice images, aligned parallel to one another, through a 3D structure, and wherein the method moreover includes the following steps:

(6d) creating a volume image from the slice images;

(6e) determining a tilt of the 3D structure on the basis of the volume image.

37. Method for operating a particle imaging apparatus according to the preceding claim, wherein the 3D structure comprises a NAND structure.