Automated machining method and computer program product for extracted microscopic samples

Abstract

The invention relates to a method for automated machining of an extracted microsample. The machining is performed by irradiating the microsample with a focused ion beam. To this end, a milling box is placed so as to delimit the region of the microsample to be machined. The method comprises the following steps: - attaching the extracted microsample to a holding device, wherein the holding device is configured to hold the microsample during the machining procedure; - performing a position determination procedure for the microsample while the microsample is attached to the holding device; - based on the results of the position determination procedure, aligning and / or compensating for deviations from an ideal position with respect to the microsample of the milling box; - machining the microsample by irradiating the region that has been determined by the milling box. The invention also relates to a computer program product.

Description

Technical Field

[0001] This invention relates to the localization and processing of separated microscopic samples for subsequent machining procedures, such as automated thinning processes. For example, the separated microscopic sample may be a extracted TEM sheet. Background Technology

[0002] A TEM slide is a microscopic sample required for analysis in transmission electron microscopy (TEM). Typically, a TEM slide has a very flat cuboid shape. At least one region of this flat cuboid must be thin enough to be transparent to electrons. This usually requires a special thinning procedure. During analysis with TEM, electrons are transmitted through the sample and can be used to generate transmission images of the transparent regions.

[0003] Typically, TEM slides (or any other type of microscopic sample) are prepared from a sample block. After the TEM slide has been exposed from the sample block, it must be extracted. This means that the TEM slide is cut from the sample block and simultaneously transferred to a suitable holding device, such as the fingers of a TEM grid, using a micromanipulator needle or microclamp. Therefore, extraction methods involve both sample extraction itself (removing the slide from the sample block) and transferring and attaching the extracted sample to or to a target location on the grid fingers. In recent years, these extraction methods have been automated in many ways. However, in particular, inaccuracies, such as misalignment and / or distortion, can occur in the transfer and attachment steps.

[0004] Furthermore, the overall workflow typically does not end with the attachment step and the release of the micromanipulator needle or microclamp. In many cases, additional preparation of the TEM sheet is required, such as further thinning of the TEM sheet. Advantageously, this step should also be included in the automated workflow.

[0005] It turns out that using a combined FIB / SEM system to prepare TEM slides or other microscopic samples is beneficial. The FIB / SEM system allows for the generation of electrons and focused ions for imaging and / or machining of the sample.

[0006] SEM (Scanning Electron Microscopy) capabilities allow for the imaging of samples. By using a focused ion beam (FIB), the material of the sample can be ablated and / or imaged. Furthermore, it is conceivable that material can be deposited onto the sample surface using an ion beam in conjunction with a suitable process gas.

[0007] Typically, one or more so-called milling boxes are used to define sample areas where material should be ablated (or deposited) using a focused ion beam. The milling box (also called the thinning shape when used for sample thinning) can be displayed in the user interface of a FIB / SEM combined system. However, the milling box can also be invisible to the user, as it can be a virtual tool.

[0008] Typically, a milling box not only defines the location, shape, and size of the sample area to be machined, but also defines other machining parameters (namely, milling parameters), such as (especially) milling current and milling time.

[0009] One of the biggest challenges during sample processing is the positioning and alignment of the desired geometry relative to the TEM sheet held in place by the carrier fingers. Positioning and alignment of the desired geometry must be achieved by milling the location and shape of the box. This process is further complicated by the fact that the TEM sheet does not yet exist when the user sets up the automated workflow, let alone be attached to the carrier fingers.

[0010] Therefore, it is desirable to propose a method that allows for the removal of the exact bounding portion of the sample volume based on the actual misalignment and / or distortion of the microscopic sample after extraction.

[0011] Different methods of preparing TEM samples are known in the art.

[0012] For example, US 9,837,246 B1 (Geurts et al.) discloses a method for thinning samples (such as TEM samples). In this method, reference points are formed on the workpiece so that the ion beam can be aligned using these reference points.

[0013] EP 3 018 693 A1 (Brogden et al.) describes another method for automating sample preparation, in which reference points are milled and used to determine the position and / or alignment of the sample.

[0014] Salzer et al. (2009, Standard Free Thickness Determination of Thin TEM Samples via Backscatter Electron Image Correlation, Microscopy and Microanalysis, Vol. 15 (Supplement 2), p. 340) describe a method for monitoring the thickness of TEM samples during FIB milling.

[0015] DE 10 2012 110 651 A1 (Nakatani et al.) discloses a method for producing thin sections using a FIB / SEM combined system.

[0016] US 2011 / 0226948 A1 (Tanaka et al.) relates to a method for processing and observing cross-sections of samples. This involves forming thin sections using a fibril (FIB).

[0017] US 2012 / 0187292 A1 (Tomimatsu et al.) describes a method for monitoring film thickness using an electron beam.

[0018] DE 102020 122 535 B4 (Orchowski) discloses a method for operating a particle beam system in which the particle beam can be easily adjusted with reference to the object after the object holder has been displaced.

[0019] DE 10 2021 205 001 B4 (Biberger et al.) discloses a method for positioning movable objects in the sample chamber of a particle beam microscope by means of a so-called barrier method.

[0020] EP 3 163 283 B1 (Uemoto et al.) discloses an automated sample preparation apparatus. The document also describes the preparation of thin sheets using a prong and the fixing of the thin sheets to the support of a holding device. Summary of the Invention

[0021] The object of this invention is to provide a method for automated machining of extracted microscopic samples, according to a procedure described below. Advantageously, the overall procedure is automated, eliminating the need for user intervention.

[0022] This invention provides an automated machining method for an extracted microscopic sample, wherein the machining is performed by irradiating the microscopic sample with a focused ion beam generated within a particle beam apparatus, wherein the particle beam apparatus is configured to be controlled by an evaluation and control unit, wherein a milling box is positioned to define the area of ​​the microscopic sample to be machined; wherein the method includes the following steps:

[0023] - The extracted microscopic sample is attached to a holding device, wherein the holding device is configured to hold the microscopic sample during the machining process;

[0024] - While the microscopic sample is attached to the holding device, a position determination procedure is performed on the microscopic sample;

[0025] - Based on the results of the position determination procedure, align the milling box and / or compensate for deviations relative to the ideal position of the microscopic sample;

[0026] - The microscopic sample is machined by irradiating the area already defined by the milling box.

[0027] Beneficial embodiments are presented below.

[0028] In some embodiments, the location determination procedure determines at least one of the following characteristics of the microscopic sample:

[0029] - The dimensions of the microscopic sample;

[0030] - The geometry of the microscopic sample;

[0031] - The location of the microscopic sample at the holding device;

[0032] - The orientation of the microscopic sample relative to the optical axis of the particle beam apparatus.

[0033] In some embodiments, the location determination procedure is selected from the following procedures: shape recognition, direct segmentation, indirect segmentation, edge detection, line and / or strip scanning methods, and barrier methods.

[0034] In some embodiments, the method includes the following additional step: - checking the reasonableness of the result of the previous position determination procedure.

[0035] In some embodiments, the reasonableness is checked by comparing the result of the location determination procedure with the confidence level provided in the algorithm used.

[0036] In some embodiments, the location of the microsample is determined based on the geometric properties of the microsample during the location determination procedure.

[0037] In some embodiments, the location determination procedure is performed without using a reference point.

[0038] In some embodiments, multiple milling boxes are aligned.

[0039] In some embodiments, at least two different types of position determination procedures are performed using the same arrangement of the microscopic sample and the holding device.

[0040] In some embodiments, irradiation of the microscopic sample causes the removal of material from the microscopic sample.

[0041] In some embodiments, the microscopic sample is a TEM sheet, and the automated machining is an automated thinning of the TEM sheet.

[0042] Another object of the present invention is a computer program product according to the description below. The present invention also provides a computer program product comprising a series of control commands, wherein, when executed, these control commands cause a particle beam device to perform the methods described in any of the foregoing descriptions.

[0043] The core of this invention is the concept that the post-machining process can be adapted to the actual misalignment of the microscopic sample. This misalignment occurs during the transfer of the microscopic sample and is established when the microscopic sample is attached to the holding device.

[0044] The term “holding device” is understood to be a device configured to hold a microscopic sample while it is being machined (e.g., milled or thinned or subjected to deposition).

[0045] Conversely, during the overall workflow, the sample can also be transitionally secured to a transfer device, such as a micromanipulator needle or microclamp (sometimes also referred to as a nanomanipulator needle or nanoclamp, respectively), which is a device configured to manage the extraction and transfer of the microscopic sample.

[0046] Of course, attaching the microscopic sample to the transfer device is not as important as attaching it to the holding device: the precise attachment of the microscopic sample to the holding device is crucial because it determines the quality of subsequent machining steps (e.g., thinning procedures).

[0047] To perform the post-extraction machining process as accurately as possible, it is useful to understand the existing misalignments. Therefore, after attaching the extracted microscopic sample to the holding device, the microscopic sample undergoes a position determination procedure.

[0048] The location determination procedure can be an image recognition procedure, in which the current image of the sample is analyzed. For this purpose, a microscopic sample (e.g., a TEM slide) is extracted and positioned at a holding device (e.g., a TEM grid). Then, while attaching the microscopic sample to the holding device, an image of the microscopic sample is recorded. The image can be a SEM image or a FIB image. It can be advantageous to image the microscopic sample together with a portion of the holding device. In any case, the image is subsequently subjected to at least one image recognition procedure, which is performed without the use of reference markers.

[0049] On the other hand, the location determination procedure can also be an image-free method, where information is collected by detecting signals using a suitable detector as a charged particle beam scans on or near the sample. In this paper, extensions in the x and y directions are ignored, making the detected signals unrecognizable as an image.

[0050] Generally, location determination procedures allow for the determination of several characteristics of a microscopic sample, such as its geometry and size, as well as the location and spatial orientation of the holding device.

[0051] In this respect, it was advantageous when the location determination procedure was based on the external shape of the microscopic sample, so that no manual or natural alignment marks (e.g., reference points) were required.

[0052] The method of the present invention can be executed by one or more location determination procedures, which are different in type, as explained in more detail below.

[0053] As already mentioned, methods can be differentiated based on images, such as shape recognition, image segmentation (direct or indirect), and edge detection. On the other hand, there are procedures that are not necessarily image-based, such as line scan methods, bar scan methods, or barrier methods.

[0054] Shape recognition methods are examples of image-based procedures. This approach allows for the identification of characteristic shapes, especially when the geometry and dimensions of the microscopic sample are known.

[0055] Furthermore, the sample images can be used to perform image segmentation methods. For this purpose, the image is divided into multiple image segments. Using indirect image segmentation methods, known objects or objects depicted in the image that are simpler than the TEM slice itself are detected. The TEM slice can then be obtained from the segmentation results. Alternatively, the TEM slice can be directly segmented (so-called direct image segmentation).

[0056] Edge detection methods are based on the fact that scanning signals from a dedicated region allow for the determination of characteristic contrast patterns caused by the edges of the imaged object (e.g., a TEM slice). In other words, when a dedicated region is scanned with a particle beam, characteristic contrast patterns caused by the edges of the TEM slice located within the scanned region can be identified. Specifically, if the signature of the microscopic sample is known, changes in the contrast pattern can be easily detected, allowing for the monitoring of the slice's location and orientation.

[0057] Alternatively, the location determination procedure can be an image-free method, such as the so-called line scan or bar scan method. Both methods use variations in grayscale values ​​along a spatial dimension to generate a grayscale histogram.

[0058] Another image-free method is the so-called barrier method. In this method, a barrier region is defined in three-dimensional space. The barrier region is scanned with a particle beam, and the resulting signal is continuously recorded. When an object is located within the scanned barrier region (more precisely, within the scanned barrier volume), the detected signal changes. These changes can be monitored to obtain information about the object's location and orientation. The barrier region can be shifted in three-dimensional space, allowing for complete monitoring of the space surrounding the object.

[0059] It is advantageous to perform the method according to the invention using a combined FIB / SEM system. This system should be understood as a particle beam system that, in addition to a scanning electron microscope (SEM), includes an ion beam column for generating a focused ion beam (FIB). Using the ion beam, the material of the microscopic sample can be ablated via a so-called milling process. Simultaneously, the microscopic sample can be imaged using the SEM function of the FIB / SEM combined system. Furthermore, it is conceivable to deposit material onto the sample using the focused ion beam. For this purpose, a suitable process gas can be fed into the sample chamber via a gas injection system (GIS).

[0060] A common application of FIB milling is the thinning of microscopic samples. For example, it is necessary to thin the sample to prepare a TEM sheet in a way that makes the TEM sheet transparent to electrons.

[0061] A TEM slice can be considered a typical example of a microscopic sample that must be cut out and extracted from a sample block. Typically, a TEM slice has a roughly cuboid design. The cuboid consists of two opposite, flat sides and four short sides. The flat sides are usually thinned by milling with a grazing ion beam until an electron-transparent region is obtained.

[0062] To determine the area to be thinned, a computer-controlled FIB / SEM combination system typically offers the possibility of defining and performing what is known as milling boxes.

[0063] The milling box (e.g., a so-called thinning shape used to thin a microscopic sample) marks the area to be milled (i.e., irradiated) with a focused ion beam and defines the milling parameters to be performed, such as milling current and milling time.

[0064] Therefore, the milling box can be displayed in the user interface of the FIB / SEM combined system. For example, the milling box can be superimposed on the currently recorded image of the sample. However, it is also conceivable that the milling box is not shown to the user at all, as it can be designed as a virtual tool for operating the focused ion beam by means of control software.

[0065] In any case, it is advantageous to use a milling box to define the milling parameters and the sample area that should be processed by a focused ion beam machine.

[0066] According to the present invention, the exact alignment of the milling box can be predicted based on the results of one or more previously executed position identification procedures. In doing so, the exact location and shape of the milling box can be correctly aligned. This will significantly improve the accuracy of procedures such as automated thinning, because this step in the overall workflow no longer depends on the cumulative displacement of previous steps.

[0067] Furthermore, it is conceivable that, according to the invention, two or more milling boxes can be aligned to define a machining step, such as the thinning of a TEM sheet.

[0068] However, the present invention is not limited to the thinning of microscopic samples as described above. In fact, the proposed method can be used for any other post-transfer process of microscopic samples, such as atomic probe tomography, where specially formed sample supports are required.

[0069] Furthermore, if the exact location (i.e., the part in three-dimensional space) and spatial orientation of the microscopic sample have been determined through the procedures described herein, this information can be used not only to align the milling box for further processing, but also to include a check on the plausibility of the obtained results.

[0070] It can also compensate for deviations from the ideal position of the microscopic sample. Compensation can be performed by mechanically altering the position of the microscopic sample, i.e., by moving the microscopic sample (e.g., by shifting and / or rotating the sample stage). In many cases, tilting the microscopic sample (i.e., rotating it about a tilted axis perpendicular to the optical axis of the particle beam apparatus) can be helpful. Alternatively or additionally, compensation can be performed electronically, for example, by electronically shifting the beam scan. Attached Figure Description

[0071] Figure 1 An image of a particle beam attached to a TEM sheet with a grid-like structure is shown, illustrating the potential difficulties.

[0072] Figure 2a and 2b This demonstrates the problems that may occur when imaging the attached TEM film.

[0073] Figure 3 The workflow of a first embodiment of the invention is illustrated. Exemplarily, the microscopic sample to be machined is a TEM slide, which should be thinned after the extraction process. For this purpose, different types of position recognition procedures can be applied.

[0074] Figure 4 The placement of a milling box (e.g., a thinned shape) is shown. Milling boxes can be stacked in a particle optics image to precisely define the expected ablation of the sample material.

[0075] Figures 5a to 5c The principle of the line scan / strip scan method, as an example of a location determination method, is demonstrated.

[0076] Figures 6a to 6c The principle of the barrier method, as an alternative example of a location determination method, is shown.

[0077] Figures 7a to 7cAn example of a segmentation method is shown as an alternative to the location determination method.

[0078] Figure 8 A flowchart illustrating an alternative embodiment of the invention is provided. The procedure of this embodiment includes at least one query step for checking the reasonableness of the results of the location identification procedure. Furthermore, this embodiment may include a step of compensating for detected biases.

[0079] Figure 9 A FIB / SEM combination system is schematically shown as an example of a particle beam apparatus that can be used to perform the methods of the present invention. Detailed Implementation

[0080] Figure 1 A particle beam image of a TEM slide 1 attached to a grid finger 3 is shown. The imaged scene represents a stage in the overall workflow that can be achieved using a FIB / SEM combined system. In the overall workflow, the TEM slide 1 (or any other microscopic sample) is produced and machined. Typically, the microscopic sample is obtained from a sample block. After the microscopic sample is cut from the sample block, it is extracted by transferring the microscopic sample in a so-called extraction procedure.

[0081] Figure 1 The image shows the sample after extraction. The TEM sheet 1 has been transferred from the sample block (not shown) to the mesh finger 3 of the TEM mesh 4 using a micromanipulator needle 2. The TEM mesh 4 acts as a holding device, configured to hold the microscopic sample during subsequent machining steps. A TEM mesh (sometimes also called a FIB extraction mesh) does not necessarily need to include a mesh or grid. In fact, the term is used broadly to refer to a holding device for a TEM sample or any other carefully prepared microscopic sample.

[0082] like Figure 1 As indicated by the horizontal auxiliary line 9, the fixed TEM sheet 1 is not perfectly aligned with the geometry of the mesh fingers 3. The edges of the TEM sheet 1 do not appear to be parallel to the edges of the mesh fingers 3 and are not oriented accordingly.

[0083] Generally speaking, after the transfer step, the position of the TEM sheet 1 at the grid finger 3 may deviate from the ideal expected position.

[0084] The sample location (also abbreviated as "location") includes both the physical location and spatial orientation of the microscopic sample. Here, the term "location" refers to the position of the microscopic sample in three-dimensional space, which can be described by specifying x, y, and z coordinates. Therefore, the exact location of the transferred microscopic sample at the holding device can be described. "Spatial orientation" should be understood as the arrangement of the microscopic sample in three-dimensional space. Spatial orientation is typically specified relative to at least one of the optical axes of the particle beam system employed.

[0085] During presentation and transfer, the microscopic sample may move and / or twist unintentionally in many different ways, resulting in a deviation from its ideal position. For example, the microscopic sample may deform due to impact with an object in the sample chamber or due to internal tension in the sample material.

[0086] This misalignment of the microscopic sample will have a strong impact on further machining processes (such as the thinning of the microscopic sample), especially when the degree of misalignment is not known precisely.

[0087] Figure 2a and 2b The problems that may occur when a TEM slice is transferred to the grid fingers of a TEM grid are shown in more detail. Typically, the TEM slice is fixed in place, resulting in some displacement and / or distortion. However, this misalignment is not always obvious and is therefore difficult to detect.

[0088] exist Figure 2a and 2b In this example, TEM slice 1 should have a rectangular shape. TEM slice 1 can be imaged in two different views using a FIB / SEM combined system: it can be imaged in both the SEM and FIB views. The SEM and FIB columns of the particle beam apparatus are arranged at an angle to each other, in this example, at 54°. Therefore, the two views produce images of the same sample (i.e., TEM slice 1) from different perspectives.

[0089] Figure 2a Different views are illustrated using arrows V1 to V4. Arrow V1 represents a SEM view on a TEM sheet 1 attached to a grid finger 3. In the example shown, the electron beam is grazing relative to the first flat side 7, the second flat side 7a, and the first short side 5 of the TEM sheet 1. The image obtained from view V1 is... Figure 2b The left side is shown.

[0090] Arrow V2 indicates the FIB view on TEM section 1. SEM view V1 and FIB view V2 are arranged at an angle to each other, for example, 54°. Therefore, FIB view V2 produces different perspectives of TEM section 1, such as... Figure 2b As shown in the middle section.

[0091] In this example, in FIB view V2, the second flat side 7a (which is planar parallel and opposite to the first flat side 7) and the second short side 6 of TEM sheet 1 are imaged.

[0092] Due to the perspective of SEM view V1 and FIB view V2, the second flat side 7a and the second short side 6 are depicted in an oblique manner. Therefore, it is helpful to introduce the concept of virtual views V3 and V4, in which each view (V3 or V4) provides a view perpendicular to the depicted side of the TEM sheet.

[0093] V3 is called the Virtual Front View because it provides a view directly facing the flat side 7 (such as...). Figure 2b As shown (on the right), while the virtual side view V4 provides an image perpendicular to the first short side 5.

[0094] Figure 2b Two examples of TEM sheets 1 and 8 are shown in different views. TEM sheets 1 and 8 are fixed to the carrier finger 3 to illustrate the variability of the transfer process.

[0095] Figure 2b The upper portion shows TEM slice 1, which is ideally transferred to the grid. In the SEM view (left), the short side of TEM slice 1 is imaged.

[0096] In the FIB view (center), the short side and flat side of TEM slice 1 are imaged. Here, the flat side appears stereoscopically shortened due to the different viewing angle, while in the virtual front view (right side), TEM slice 1 appears rectangular.

[0097] In comparison, Figure 2b The lower portion illustrates the situation when the second TEM sheet 8 has not yet been fully transferred. In the depicted example, the second TEM sheet 8 is unintentionally tilted: the second TEM sheet 8 is fixed at an angle α relative to the carrier fingers. Therefore, the second TEM sheet 8 appears to be three-dimensionally shortened, as... Figure 2b The middle (FIB view) and right (virtual front view) are shown.

[0098] Therefore, from only one perspective, it is impossible to reliably distinguish between the different types of displacements that may occur.

[0099] Therefore, it is difficult to align the milling boxes, which define the areas subsequently milled using a focused ion beam. This problem is exacerbated when milling boxes are defined in an automated process. If so, the milling boxes should be defined in advance, which may mean defining the milling boxes before the sample is even present, since the sample has not yet been cut from the sample block.

[0100] Therefore, it is desirable to propose a method for accurately determining the position (i.e., location and spatial orientation) of a microscopic sample at the holding device, the external shape of the sample, and the dimensions of the microscopic sample. Dimensions can be characterized by specifying the length, width, and depth of the microscopic sample.

[0101] Knowing these parameters allows subsequent thinning procedures or any other subsequent machining processes to be adapted to the actual conditions.

[0102] Figure 3 A first exemplary embodiment of the workflow of the present invention is shown, wherein the thinning of the TEM sheet that occurs after the extraction procedure is adapted according to the identified actual misalignment. Figure 3 The method includes at least steps S1 to S5. The workflow diagram shows that different types of location identification procedures (S2, S3) can be performed alternatively or additionally.

[0103] In the first step S1, after being withdrawn, the TEM sheet is attached to the carrier fingers. The carrier fingers act as a holding device, so that the TEM sheet is positioned for the workflow according to the invention.

[0104] Then (step S2), while attaching the TEM sheet to the carrier finger, a position recognition procedure is performed on the TEM sheet.

[0105] There are several alternative or common methods for location identification.

[0106] For example, a line and / or strip scanning method can be performed (step S2a) or a barrier method can be applied (step S2b). Both methods are described in more detail in... Figures 5a to 5c and Figures 6a to 6c middle.

[0107] It is also conceivable to record an image of the TEM slice at the grid-supported finger (step S3). The recorded image can then be used to perform image-based methods, such as shape recognition (step S3a), direct segmentation (step S3b), indirect segmentation (step S3c), or edge detection (step S3d). For example, the indirect segmentation method is described in more detail in... Figures 7a to 7c middle.

[0108] When using image-based methods, it is advantageous to identify the geometric or material properties of TEM slices. For example, geometric properties could be the external shape (i.e., profile) of the TEM slice. Material properties could be, for example, roughness or other typical surface properties observable in the image. Furthermore, it is conceivable that characteristic edges within the sample are observable.

[0109] Discovery identification of geometric or material properties is advantageous because it eliminates the need for reference points or other markings at TEM sections or mesh-supported fingers. This means that the additional step of creating reference points is unnecessary.

[0110] Generally, due to the location recognition process, several features of a microscopic sample can be identified. For example, it is conceivable to identify the geometry of the microscopic sample. Furthermore, the size of the sample can be identified. This is possible because the pixel size is known, allowing the absolute size of the imaged object to be determined.

[0111] It is also conceivable to determine the exact location of the microscopic sample at the holding device and / or the orientation of the microscopic sample relative to the optical axis of the particle beam device.

[0112] Based on the results of the position recognition step, the milling box is placed and aligned (step S4), as described below. Figure 4 As shown in the figure, this is to approximate the optimal conditions for thinning.

[0113] This means that the milling box is aligned in a way that allows the required sample volume to be milled in the next step of the procedure.

[0114] Finally, the TEM section is thinned by performing a milling operation using the aligned milling box (step S5). This means that the area previously defined by the alignment using the milling box is irradiated with a focused ion beam. Irradiation causes material removal from the microscopic sample, resulting in sample thinning.

[0115] Furthermore, knowing the exact location of the TEM slices obtained in steps S2 and S3 can be used not only for placing the milling box (i.e., the thinned shape) for further processing, i.e., for aligning the milling box, but also for compensating for deviations relative to the ideal arrangement by moving the TEM slices, thereby changing the location and / or spatial orientation of the TEM slices. In particular, the rotation of the sample stage used to move the TEM slices about the tilt axis is very helpful for thinning in cases of excessive or insufficient tilting. On the other hand, it is conceivable to shift the particle beam in a way that compensates for the deviations.

[0116] However, it turns out that it is advantageous to perform another position determination step after the deviation has already been compensated.

[0117] Figure 4 As shown in Figure 3 An example of the placement of typical milling boxes 10, 11 performed in step S4 of the embodiment.

[0118] To define the area of ​​material to be ablated in the TEM sheet 1, two milling boxes 10 and 11 (sometimes referred to as thinning shapes) were placed. For this purpose, the milling boxes can be overlaid in the image (right side) showing the arrangement of the microscopic sample and the holding device. This is done using control software that allows the milling boxes to be stacked to define the areas and intensity of machining.

[0119] However, the milling box can also be invisible to the user, as it can be a virtual tool used to manipulate the focused ion beam generated in the FIB / SEM combined system.

[0120] The locations of milling boxes 10 and 11 determine the areas to be milled. Furthermore, defining the milling boxes defines the shape of the area to be milled and all necessary milling parameters, such as milling speed and milling time.

[0121] exist Figure 4 In the example, the first milling box 10 and the second milling box 11 are precisely aligned on both sides of the TEM sheet 1 to be thinned. The material of the TEM sheet is then ablated by milling the milling boxes 10 and 11 with a focused ion beam (FIB).

[0122] However, the method according to the invention is not limited to the automated thinning of TEM sections. Rather, the method can be applied to any other microscopic sample that should be machined after the extraction procedure.

[0123] Figures 5a to 5c and Figures 6a to 6c An example of a location recognition method performed without an image is shown.

[0124] Figures 5a to 5c The principle of line scanning and / or bar scanning methods is illustrated. These methods utilize grayscale value variations along a single pixel width (so-called line scanning) or by averaging the values ​​of a rectangular shape along the lateral direction (so-called bar scanning) to obtain a spatial dimension.

[0125] Specifically, the embodiment shown is an example of a bar scanning method. Figure 5a The TEM sheet 1 is shown fixed to the carrier finger 3, with the TEM sheet 1 intentionally over-tilted. Within the scanning area 15, a charged particle beam scans in a rectangular pattern. Arrow 13 indicates the scanning direction. The charged particle beam scans back and forth within the boundaries defined by the scanning area 15. Simultaneously, a permanent detector signal (grayscale value) is detected.

[0126] Figure 5b The TEM sheet 1 is shown to include several edges that constrain the body of the sheet, such as the first sheet edge A, the second sheet edge B, and the third sheet edge C.

[0127] During the scan, the charged particle beam also touches these edges A, B, and C of the TEM slice 1. This means that the charged particle beam transitions from the background (where the charged particle beam does not collide with any material) to the TEM slice (where the charged particle beam collide with the material of the TEM slice). This transition causes a change in contrast (i.e., grayscale value).

[0128] Figure 5c The resulting grayscale image is shown, where the grayscale values ​​are plotted relative to distance (length). Plotting direction 14 corresponds to scanning direction 13. In graph 12, the changes indicated by the grayscale values ​​named A', B', and C' correspond to... Figure 5b The edges A, B, and C of the thin sheet are shown.

[0129] In summary, both bar scanning and line scanning methods generate a histogram of grayscale values ​​along a length dimension, where the x-position of a feature refers to its relative position along a line or bar. Together with the known placement of the feature within the field of view (FOV) in a line or bar scan, it allows determining the absolute position of the feature within the FOV.

[0130] On the one hand, the location of well-known features of a TEM slice can be determined by knowing the line / strip scan signatures at different locations and detected by different detectors. On the other hand, the signatures can be used for plausibility checks because the geometry of the TEM slice results in a predictable spatial sequence of features in the histogram.

[0131] Placing line / strip scans at different locations or assessing the details of a signature can even allow for determining the exact orientation of a sheet.

[0132] The scan signal of the dedicated area (reduced raster image and barrier area) allows for the determination of the characteristic contrast pattern caused by the edge of the sheet.

[0133] Therefore, the position of the TEM slice (i.e., its location and spatial orientation) can be learned. Furthermore, the exact thickness of the TEM slice can be determined, and its edges can be located.

[0134] Figures 6a to 6c The principle of the barrier method is illustrated schematically. The barrier method is related to the line scan or strip scan method in some respects because it is based on a series of average gray values ​​(signals from the detector) within a so-called barrier region.

[0135] However, the barrier method additionally requires a certain "dynamic," that is, either the object (e.g., the TEM sheet 63) must be moved, or the position of the barrier region 62 needs to be gradually changed (i.e., in a series of steps). The latter is called the "moving barrier method" and can be used in particular to determine the position of the TEM sheet 63 when it is attached to the carrier finger 61.

[0136] For the barrier method, the TEM sheet 63 is positioned near the mesh finger 61 at the micromanipulator 64, but at a certain distance from the mesh finger 61. This ensures that the TEM sheet 63 will not accidentally collide with the mesh finger 61.

[0137] Then, image 60 is recorded, showing the mesh fingers 61 and the barrier region 62, as follows. Figure 6a As shown. The barrier region 62 is located near the carrier finger 61 and defines the area where the integrator detector signal will be continuously detected.

[0138] Next ( Figure 6b The TEM sheet 63' is moved toward the grid finger 61, as indicated by arrow 65. Simultaneously, the integrator detector signal is continuously detected.

[0139] Figure 6c An example is shown of the integral detector signal 66 plotted relative to time. At a certain point (indicated by arrow 67), the detector signal 66 shows a striking change. Here, the TEM slice 63 enters the barrier region 62. This change indicates the position of the edge of the TEM slice 63 along the direction of movement.

[0140] In the example shown, the TEM sheet moves toward the mesh fingers.

[0141] However, it is also conceivable that the barrier region itself is displaced, i.e., its location changes. This could help determine the exact position of the TEM sheet during its attachment to the carrier finger.

[0142] It is also conceivable to use more than one barrier region, such as a first barrier shifted along a first axis and a second barrier shifted along a second axis.

[0143] Alternatively or additionally, images of the arrangement of the microscopic sample with the holding device can be recorded, and these images can be subjected to several image-based position determination methods, such as shape recognition methods. This is particularly advantageous if the geometry and dimensions (i.e., length, width, and depth) of the microscopic sample are known.

[0144] During shape recognition methods, a shape (e.g., a rectangle) of a certain size can be found and identified in an image. This is possible because the pixel size of the corresponding image is known. The pixel size and magnification can be read from the metadata of the image data. Therefore, the location of the TEM slice can be determined. Even if the dimensions are not given, it is conceivable that the size of the TEM slice can be determined if the pixel size and magnification are known.

[0145] Furthermore, image segmentation (directly or indirectly) can be useful in obtaining desired information about the misalignment of samples, such as... Figures 7a to 7c As shown.

[0146] By direct image segmentation ( Figure 7b This involves searching for objects in an image, identifying them, and isolating them from the rest of the image. This is achieved through indirect image segmentation (…). Figure 7c It identifies known backgrounds in an image and identifies areas that do not belong to the known background as objects.

[0147] Figures 7a to 7c SEM images representing TEM sections. Figure 7a On the left, the original SEM image 70 is reproduced. Figure 7a The right side shows the same image as line graph 70'. Select the representation of line graph 70' (see...). Figures 7a to 7c In order to clearly explain the differences in segmentation methods, the TEM sheet 72 is attached to the carrier fingers 71 of the holding device (TEM carrier).

[0148] Figure 7b An example of direct image segmentation is shown. A line graph 70' (left) of SEM image 70 is shown, in which image region 72 can be directly identified as a TEM slice (right).

[0149] On the contrary, Figure 7c An indirect segmentation is shown. Here, the background is well-known, such as... Figure 1 c (left side) shows the area. It is assumed that areas not belonging to the known background are objects, such as... Figure 7c (As indicated on the right). Finally, TEM sheet 72 can be identified and shown in an isolated manner. Figure 7c lower part).

[0150] Figure 8 Alternative embodiments of the invention are shown, wherein at least one step (S12, S19) is involved to check the reasonableness of the results of the position identification procedure. If there is doubt about reasonableness, this embodiment specifies performing two or more position identification methods for the same arrangement of the microscopic sample and the holding device. It is also conceivable that the reasonableness check be configured as a termination criterion.

[0151] In the first step S10, the microscopic sample (e.g., a TEM slide) is placed at the holding device after being extracted.

[0152] Then, while holding the microscopic sample in the holding device, the first position identification procedure (step S11) is performed as described above.

[0153] Similarly, the location recognition procedure may be selected from the following procedures: line and / or bar scanning methods, barrier methods, shape recognition, direct segmentation, indirect segmentation, edge detection, or any other suitable method.

[0154] If an image-based location recognition method is chosen, it is advantageous to perform image recognition using the geometric properties of the microscopic sample. This means that image recognition focuses on the external shape (i.e., contour) of the microscopic sample, eliminating the need for additional markers or reference points.

[0155] In the next step (S12), the reasonableness of the result is checked. This is done through the first query: Is the result of step S11 reasonable? To evaluate the result of the location recognition procedure, the result is compared with a confidence level already integrated in the algorithm for this purpose. The comparison with the confidence level allows for an internal evaluation of the quality of the result. Furthermore, a statement about the probability of the result's correctness can be conceived.

[0156] If the answer to the first query step (step S12) is "yes", the procedure can continue to steps S13 to S15, which may include three alternative paths I, II and III.

[0157] According to the first path I, the milling box is aligned based on the result of the position recognition procedure (S13). The alignment of the milling box includes determining the exact location of the milling box and defining the milling parameters used for machining the microscopic sample, such as... Figure 3 As described. Then (step S15), the microscopic sample is machined, for example, by milling, i.e., the sample material is removed.

[0158] According to the second path II, after a positive rationality check, a compensation step (S14) is performed. This means that after obtaining detailed information about the deviations from the ideal position of the microscopic sample, these misalignments are compensated. For this purpose, the position of the microscopic sample can be changed by moving the sample stage (i.e., by shifting and / or rotating the microscopic sample). Alternatively or additionally, the beam scan can be shifted and / or rotated. After the misalignment has been compensated, the milling box can be aligned in step S13, and the microscopic sample can be milled (step S15).

[0159] However (as shown in path III), it is also conceivable that step S14 is not followed by alignment with step S13. Instead, after compensating for the deviation (S14), the microscopic sample can be directly subjected to the machining (i.e., milling) step S15.

[0160] If the answer to the first query (step S12) is below the confidence level, i.e., the answer to the query is "no", then a second image recognition procedure (step S18) is performed using a microscopic sample, i.e., using the same arrangement of the microscopic sample and the holding device. Advantageously, the second procedure is different in type from the first position recognition procedure.

[0161] Then, the second query is performed (step S19). Again, the reasonableness of the result of the location recognition procedure is checked, that is, the result is compared with the confidence level provided in the algorithm.

[0162] If the answer to the query is "yes", the procedure continues to steps S13 to S16.

[0163] If the answer to the query is "no", another position identification step S20 can be performed for the same arrangement of the microscopic sample and the holding device, wherein a position identification method of a different type from the previously performed position identification method is used.

[0164] However, it is also conceivable that when the answer to the query is "no", this is used as a criterion to stop the machining process (step S17).

[0165] Steps S19 to S20 can be iterated to improve the alignment step (S13) and / or the compensation step (S14). This can also be done by using artificial intelligence (AI) to evaluate the results of the position recognition procedure and improve the overall machining procedure.

[0166] Figure 9 The FIB / SEM combined system 90 is schematically shown, which is a particle beam device that can be used to perform the methods of the present invention.

[0167] The microscopic sample 105 may be disposed on a sample holder 104, which is mounted on a sample stage 103. Alternatively, the microscopic sample 105 may be disposed on a holding device 114 (e.g., a TEM grid), which is in turn mounted on the sample holder 104.

[0168] Advantageously, the FIB / SEM combined system 90 includes a transfer device 112 for transferring the microscopic sample 105, such as a micromanipulator or microclamp.

[0169] The sample stage 103 is located within the sample chamber 99 of the FIB / SEM combined system 90. During operation, a vacuum condition is maintained inside the sample chamber 99.

[0170] Advantageously, the sample stage 103 is designed as a multi-axis stage, providing several translational and rotational axes. An example of such a sample stage is a five-axis stage, including translational axes x, y, and z, and rotational axes R and T (tilt). Thus, the translational axes are arranged perpendicular to each other. Typically, the rotational axis R is parallel to the z-axis, while the tilting axis T is arranged perpendicular to the rotational axis R.

[0171] The FIB / SEM combined system 90 includes two particle beam columns: an electron beam column 91 for generating an electron beam and an ion beam column 111 for generating an ion beam. Both particle beams are guided to a microscopic sample 105, which is typically located at the point of overlap of the two particle beams.

[0172] During the operation of the FIB / SEM combined system 90, electrons are generated in the electron source 92. The primary electrons propagate along the optical axis 94 of the electron column 91, are focused by the concentrator systems 93 and 95, parallelized, and cut by at least one aperture 96.

[0173] Furthermore, the electron beam column 91 includes a first deflection system 97 and an objective lens system 98, the first deflection system allowing the primary electron beam to scan the surface of the sample. The objective lens system 98 enables the primary beam to be focused onto the microscopic sample 105.

[0174] Furthermore, the FIB / SEM combined system 90 includes an ion beam column 111, which includes an ion source 109, a second deflection system 107, a second aperture 113, a lens system (FIB) 106, and an objective system (FIB) 110. The ion source 109 can be, for example, a liquid metal ion source (LMIS), such as a gallium ion source.

[0175] Ions generated inside the ion source 109 are accelerated along the optical axis 108 of the ion beam column 111. Then, the ion beam is focused onto the sample 105 using the objective system 110.

[0176] Ions incident on the microscopic sample 105 can be used to ablate the material of the sample 105 and / or to image the sample 105.

[0177] Optionally, the FIB / SEM combined system 90 may include a gas injection system 102 for feeding process gases into the sample chamber 99. The process gases may allow for accelerated removal of sample material or deposition of material onto the sample surface.

[0178] Furthermore, the FIB-SEM combined system 90 includes at least one detector 100 for detecting interaction products of electrons and / or ions interacting with the material of the microscopic sample 105. For example, the detector 100 may be an SE detector or a BSE detector for detecting secondary electrons (SE) and backscattered electrons (BSE), respectively.

[0179] Additionally, the FIB-SEM combined system 90 includes an evaluation and control unit 101. The evaluation and control unit 101 can receive and read computer program products.

[0180] The computer program product includes a series of control commands that, when executed, cause a particle beam device (e.g., FIB / SEM combined system 90) to perform the method of the present invention.

[0181] Figure Labels

[0182] 1TEM thin film

[0183] 2 Micromanipulator needles

[0184] 3-load net finger-shaped object

[0185] 4TEM carrier network (holding device)

[0186] The first short side of the 5TEM thin film

[0187] Second short test of 6TEM thin film

[0188] The first flat side of the 7TEM thin film

[0189] The second flat side of the 7aTEM sheet (opposite to the first flat side 7 and parallel to the plane).

[0190] 8 Second TEM thin film

[0191] 9 auxiliary lines

[0192] V1SEM view

[0193] V2FIB view

[0194] V3 Virtual Front View

[0195] V4 Virtual Side View

[0196] 10 First Milling Box

[0197] 11 Second Milling Box

[0198] 12-line graph

[0199] 13 scanning directions

[0200] 14 Drawing Direction

[0201] 15 scanning areas

[0202] A First Thin Edge

[0203] B, the edge of the second thin sheet

[0204] C Third Thin Edge

[0205] The first change in grayscale value A' (corresponding to A)

[0206] The second change in grayscale value (corresponding to B) is B'.

[0207] The third change in the grayscale value of C' (corresponding to C)

[0208] 60 images

[0209] 61-loaded net finger-shaped object

[0210] 62 Barrier Area

[0211] 63TEM thin film

[0212] 63' Moving TEM sheet

[0213] 64 micromanipulator

[0214] 65 Arrows indicating the direction of movement of the TEM slice

[0215] Graphical representation of the 66-integral detector signal

[0216] 67 Arrow indicating signal change

[0217] 70 SEM images

[0218] 70' SEM image 70 line graph representation

[0219] 71-carrying net-like objects

[0220] 72TEM thin film

[0221] 90FIB / SEM combined system

[0222] 91 electron beam column

[0223] 92 electronic sources

[0224] 93 First Concentrator System

[0225] The optical axis of the 94 electron beam column

[0226] 95 Second Concentrator System

[0227] 96 aperture

[0228] 97 First Deflection System

[0229] 98 Objective System (SEM)

[0230] 99 Sample Room

[0231] 100 detectors

[0232] 101 Assessment and Control Unit

[0233] 102 Gas Injection System

[0234] 103 Sample Platform

[0235] 104 Sample Holders

[0236] 105 microscopic samples

[0237] 106 Lens System (FIB)

[0238] 107 Second Deflection System

[0239] The optical axis of the 108 ion beam column

[0240] 109 ion source

[0241] 110 Objective System (FIB)

[0242] 111 Ion Beam Column

[0243] 112 Transfer device (e.g., micromanipulator)

[0244] 113 Second Aperture

[0245] 114 Holding devices (e.g., TEM carriers)

[0246] X x axis

[0247] Yy axis

[0248] Zz axis

[0249] R rotation axis

[0250] T-axis

[0251] S1 places the TEM sheet at the grid finger.

[0252] S2 performs a position recognition procedure on the TEM film at the finger-shaped object of the carrier.

[0253] S2a line and / or bar scanning method (grayscale histogram)

[0254] S2b barrier method

[0255] S3 records images of the TEM film at the mesh finger.

[0256] S3a Shape Recognition

[0257] S3b direct segmentation

[0258] S3c indirect segmentation

[0259] S3d edge detection

[0260] S4 aligns the milling box based on the position recognition result.

[0261] S5 thins the TEM sheet.

[0262] S10 places the microscopic sample at the holding device.

[0263] S11 executes the first location identification procedure

[0264] S12 First Query: Is the result of the first location identification reasonable?

[0265] S13 aligns the milling box

[0266] S14 Compensation for Deviation

[0267] S15 performs machining on the microscopic sample.

[0268] S17 Termination Procedure

[0269] S18 executes a second location identification procedure (for the same layout).

[0270] S19 Second Query: Is the result of the second identification reasonable?

[0271] S20 performs another location identification procedure

[0272] First Path

[0273] II. Second Path

[0274] III. Third Path

Claims

1. An automated machining method for extracted microscopic samples (1, 20, 10⁵), in, The machining was performed by irradiating the microscopic sample with a focused ion beam. The focused ion beam is generated within the particle beam apparatus (90). Furthermore, the particle beam device is configured to be controlled by an evaluation and control unit (101). Among them, milling boxes (10, 11) are placed to define the area of ​​the microscopic sample (1, 20, 105) to be machined; The method includes the following steps: - The extracted microscopic samples (1, 20, 105) were attached to the holding device (21, 114). The holding device (21, 114) is configured to hold the microscopic sample during the machining process (S1, S10); - While the microscopic sample is attached to the holding device (S3, S11), a position determination procedure is performed on the microscopic sample (1, 20, 105); - Based on the results of the position determination procedure, align the milling box and / or compensate for deviations relative to the ideal position of the microscopic sample; - The microscopic sample (1, 20, 105) is machined by irradiating the area already defined by the milling box (S5, S15).

2. The method as described in claim 1, wherein, Due to this location determination procedure, at least one of the following characteristics of the microscopic sample is determined: - The dimensions of the microscopic sample; - The geometry of the microscopic sample; - The location of the microscopic sample at the holding device; - The orientation of the microscopic sample relative to the optical axis of the particle beam apparatus.

3. The method as described in claim 1 or 2, wherein, The location determination procedure is selected from the following procedures: Shape recognition (S3b), direct segmentation (S3d), indirect segmentation (S3c), edge detection (S3a), line and / or strip scanning method (S2a), barrier method (S2b).

4. The method according to any one of the preceding claims, wherein, The method includes the following additional steps: - Check the validity of the result of the procedure in the previous position.

5. The method according to claim 4, wherein, The reasonableness of this determination procedure is checked by comparing the result with the confidence level provided in the algorithm used (S12, S19).

6. The method according to any one of the preceding claims, wherein, During the location determination procedure, the location of the microscopic sample is determined based on its geometric properties.

7. The method according to any one of the preceding claims, wherein, Perform this location determination procedure without using a reference point.

8. The method according to any one of the preceding claims, wherein, Align the multiple milling boxes (10, 11).

9. The method according to any one of the preceding claims, in, At least two different types of position determination procedures are performed using the same arrangement of the microscopic sample and the holding device.

10. The method according to any one of the preceding claims, in, Irradiation of the microscopic sample causes the removal of material from the microscopic sample.

11. The method according to claim 10, wherein, The microscopic sample is a TEM sheet (1, 20, 72), and the automated machining is an automated thinning of the TEM sheet.

12. A computer program product comprising a series of control commands, wherein, When executed, these control commands cause the particle beam device to perform the method according to any one of claims 1 to 11.

Images