Method for three-dimensional tomography of elongated samples
By accessing the sample multiple times within the tilt range of interest in electronic tomography and translating the sample, combining strobe imaging and real-time correction technology, the problem of low image quality and reconstruction efficiency of needle-shaped samples in the prior art is solved. , achieving more efficient three-dimensional tomography.
Patent Information
- Application Number
- CN202410574558.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-06-01
- Filing Date
- 2024-05-10
- Publication Date
- 2024-12-03
AI Technical Summary
Existing electronic tomography methods have problems with poor image quality and low reconstruction efficiency when acquiring three-dimensional images of needle-like samples, especially in radiation-sensitive samples, which are difficult to ensure equal amounts of radiation and splicing individual 3D reconstruction volumes or tilt series. The process is complicated.
Using accessing samples only once per angle within the tilt of interest, obtaining multiple image frames and translating the samples to reduce motion blur, using strobe imaging or coding exposure technology, the sample tilt and curvature are tracked in real time to correct beam and stage movement, The samples are moved alternately to minimize the movement required for image acquisition.
Improve image quality and reconstruction efficiency, ensure equal radiation in sample areas, simplify the splicing process, and provide real-time information about sample location to optimize 3D reconstruction.
Smart Images

Figure CN119064389A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to three-dimensional tomographic analysis of samples. Background Art
[0002] Tomography is a group of imaging techniques that involves taking multiple images of an object using energy that penetrates the interior of the object, and then using the images of the object to produce a three-dimensional reconstruction of the object using one or more reconstruction algorithms (such as filtered back projection or iterative reconstruction). Tomography examples include magnetic resonance imaging, CT scanning, and ultrasound tomography, but there are many other techniques and radiation energy sources. Electron-based tomography such as electron tomography or cryo-electron tomography is also a tomographic method and is generally applicable to the tomography of small samples, for example, from micrometers to atomic scales, given the shorter wavelengths of the imaging particles.
[0003] Transmission electron microscopy (TEM) tomography of needle-shaped samples (rather than the more common plate-shaped samples) is becoming increasingly popular, mainly due to new tools for rapidly manufacturing such samples (e.g., by cryogenic focused ion beam milling). Using currently available technology, needle-shaped samples can be tilted around their elongated axis at any angle between -90 degrees and +90 degrees, thereby advantageously eliminating the so-called missing wedge problem. It is often desirable to obtain 3D volume information of the elongated portion of the needle. More generally, for samples of any shape, it may be desirable to expand the sample volume from which TEM image data is collected. Currently, this is achieved by acquiring several tilted series of image acquisitions, or by spiral tomography, which includes tilting while also shifting along the needle (if the sample is so shaped) to obtain a spiral trajectory. The resulting images are then reconstructed and spliced (or vice versa).
[0004] In an electron microscope, such as a transmission electron microscope (TEM), an electron source generates a high-energy electron beam, with a selectable energy, for example, between 80 keV and 300 keV, although higher and lower energies are known to be used. The beam is manipulated (focused, deflected) by optical elements (electron lenses, deflectors, multipoles) and illuminates a sample placed in a sample holder. The sample holder holds the sample, which typically exhibits three translational degrees of freedom (x, y, z) and one rotational degree of freedom (α) or more. The sample is typically thin enough to be transparent to electrons. Although some electrons are absorbed, many electrons enter and pass through the sample, either scattered or unscattered. Typically, for biological samples, the thickness of the sample is between 50 nm and 1 μm (for cryogenic samples, the thickness of the sample is between 50 nm and 200 nm), and for semiconductor samples, the thickness is less than 100 nm. Another set of optical elements forms a magnified image of the sample on a detector, such as a CMOS camera, a CCD camera, a fluorescent screen, or a combination of a fluorescent screen and a camera connected by an optical fiber. The image formed on the detector is a two-dimensional (2D) projection of the sample.
[0005] In order to achieve a three-dimensional (3D) image, a so-called tilt series of projection images is first acquired. Tomographic imaging electron microscopy is based on acquiring sample images at multiple exposure angles. These angles are usually set by stepping the angular orientation of the sample through a series of ascending angles, where the axis of rotation is substantially consistent with the long direction of the needle. In other examples, the angle is set by tilting the sample back and forth. One difficulty of these conventional methods is the limited range of tilt angles. In order to achieve a 3D image, a series of images of the sample at different rotational orientations are acquired with a tilt series. Typically, the tilt series covers a tilt range from -70 to +70 degrees, with regular intervals of 1 or 2 degrees. The images of the tilt series are then aligned to each other and 3D reconstruction is performed using appropriate software and then visualized.
[0006] As an example of conventional tilt-series image acquisition, Figure 1A and Figure 1B Schematically illustrates the angular orientation and positioning of three different tilted series of electron microscope images (not shown) of an elongated sample assumed to be aligned substantially parallel to and coincident with the axis 117 of an imaginary cylinder 152 having a first end 153a and a second end 153b. The first, second and third series of tilted images are indicated by the reference numerals p1 to p2, respectively. m 、p m+1 to p 2m and p 2m+1 to p 3m The three sets of image frames represented by the collection of shaded boxes are schematically depicted. Figure 1AEach such shaded box in represents the intersection of an incident charged particle beam (such as an electron beam of a transmission electron microscope) with an imaginary cylinder 152 that surrounds the sample at axis 117 during acquisition of a single image of the sample. Figure 1B , the shaded box is schematically depicted as if the surface of the cylinder 152 is flattened onto a plane, where Figure 1B The horizontal axis represents the angle of rotation and the vertical axis represents the position measured along the length of the cylinder 152 (ie, measured parallel to the axis of rotation).
[0007] exist Figure 1A and Figure 1B During the acquisition of the various image frames depicted in FIG. 1 , the sample is assumed to rotate unidirectionally in a clockwise direction (as viewed from above toward the first end 153a) about the axis 117, and it can be assumed that the closed cylinder 152 rotates with the sample. Under these conditions, the rotation causes the image frames to be acquired in a sequence represented by the arrows surrounding the cylinder. During the acquisition of the image frames represented by the shaded boxes p1 to p2 m After the first tilt series of images represented by the shaded box p , the sample and / or the charged particle beam is translated in a direction parallel to the axis 117 in preparation for acquiring the image represented by the shaded box p m+1 to p 2m The second tilt series of images is shown in Figure 2A and Figure 2B As schematically depicted in , it is assumed (according to this example) that the second series of images are acquired at the same sample rotation angle and in the same order as the first series of images.
[0008] Helical tomography generally involves taking multiple images of a sample in a helical fashion (ie, at multiple angles or projection positions around the sample), and using these images to form a 3D reconstruction of the sample. FIG. 2A to FIG. 2B An example of a set of image acquisitions that may be obtained during a helical tomography procedure is illustrated. Figure 2A The shaded box depicted in Figure 1A , where each such shaded box represents the intersection of an incident charged particle beam (such as an electron beam of a transmission electron microscope) with an imaginary cylinder 152 that surrounds the sample at the axis 117 of the cylinder during acquisition of a single image of a portion of the sample. Figure 1B The depiction of Figure 2B152 is schematically depicted as if the surface of a cylinder 152 is flattened onto a plane. The image acquisition of the helical tomography process includes a plurality of series of image acquisitions (e.g., series p, series Q, and series r), wherein during each such series of image acquisitions, the sample is gradually rotated about the elongated axis 117 (which in this example is assumed to be substantially coincident with the device rotation axis 115) such that each image of the series is obtained at a different corresponding rotation angle. For example, series p includes a set {P}p1,p2,…,p2 corresponding to the shaded box image acquisitions i ,…,p m-1 ,p m Similarly, the series Q includes the set {Q}Q1,Q2,…,Q corresponding to the image acquisition of the shadow box i ,…,Q n-1 ,Q n Similarly, the sequence r includes the set {r}r1,r2,…,r corresponding to the image acquisition of the shadow box i ,…,r s-1 ,r s Although Figure 2A Three such series are depicted in Figure 2B A total of six such series (three of which are not labeled) are depicted in , but any number of similar series of images may be acquired during a single imaging session. The values of m, n, and s may or may not be equal to one another.
[0009] By comparison FIG. 2A to FIG. 2B and Figure 1A to Figure 1B , it can be observed that the image acquisition ( FIG. 2A to FIG. 2B ) sequences p, Q and r are similar to the image acquisition of the tilt sequence T1, T2, T3, ... ( Figure 1B ). However, while there is substantially no translational movement of the sample relative to the incident beam parallel to the elongated axis 117 between consecutive image acquisitions during the acquisition of a tilt series, there is an intentional translation of the sample relative to the incident beam between consecutive image acquisitions of series p, series Q, series r, etc. during the helical tomography process.
[0010] Although such conventional tomography methods (e.g., tilt series tomography and spiral tomography) have been successfully adopted, there are still several challenges that may affect image quality and image reconstruction efficiency. For example, when using normal stitching of images or spiral tomography, it is difficult to ensure that all areas of the sample receive an equal amount of electrons, which is important for preventing image artifacts or resolution loss in the case of radiation-sensitive samples. Second, the process of stitching separate 3D reconstruction volumes or separate tilt series is known to be a well-known difficult mathematical problem. Third, the imaging process of discretely performing sample tilt or rotation steps (such as in normal tilt series or spiral image acquisition) requires a considerable position stabilization time interval between each tilt or rotation step. In addition, the quality of spiral tomography three-dimensional structural reconstruction is generally poor or variable, and given the required accuracy and technical requirements, the spiral tomography process may be associated with many difficulties. For example, the position shift of the sample after the movement of the sample during the image acquisition process may result in a blurred image. Therefore, there is still a need to improve the electron-based tomography method.
[0011] U.S. Patent No. 8,912,491 in the name of inventors Schoenmakers et al. discloses a method of performing tomographic imaging of a sample, the method comprising: providing a charged particle beam; providing the sample on a tiltable sample holder; in an imaging step, directing the beam through the sample to image the sample; repeating the process at each of a series of sample tilts to acquire a set of images; in a reconstruction step, mathematically processing the images from the set to construct a composite image, whereby in the imaging step, for a given sample tilt, a sequence of component images is captured with a corresponding sequence of focus settings; and in the reconstruction step, for at least one member of the series of sample tilts, using multiple members of the sequence of component images in the mathematical image processing. Such methods present a 3D imaging cube at a given sample tilt rather than a 2D imaging slice.
[0012] U.S. Patent No. 7,884,326 in the name of inventors van de Water et al. discloses a manipulator for use, for example, in a transmission electron microscope (TEM), which is capable of rotating and translating a sample holder. The manipulator clamps a circular sample holder between two members mounted on an actuator. Moving the actuator in the same direction results in translation of the sample holder, while moving the actuator in the opposite direction results in rotation of the sample holder.
[0013] U.S. Patent No. 9,147,551 in the name of inventor Luecken et al. discloses a sample holder capable of holding a sample and capable of continuously rotating the sample in a single direction while the sample is exposed to a charged particle beam or other radiation source. The sample holder can rotate over one complete rotation of the sample. The charged particle beam image is used for tomographic reconstruction, and in some cases, relative rotation angles are used for the reconstruction without inputting absolute rotation angles. Methods of tomography that rely on such continuous rotation of the sample in a single direction during exposure of the sample to a charged particle beam or other radiation source are referred to as "continuous tilt tomography".
[0014] U.S. Patent No. 8,854,430 in the name of inventors Varslot et al. discloses a method for processing image data of a sample, the method comprising: aligning a first image and a second image of at least partially overlapping spatial regions of the sample, and processing data from the aligned images to obtain integrated image data including information about the sample, the information being additional to information obtainable from the first image and the second image.
[0015] U.S. Patent No. 10,254,419 in the name of inventors Kingston et al. discloses a method of studying a specimen using a tomographic imaging apparatus using a stage for producing relative motion of a source with respect to the specimen (so as to allow the source and detector to image the specimen along a series of different observation axes) and processing means for assembling a tomographic image of at least a portion of the specimen. The study is performed by considering a virtual reference surface surrounding the specimen and substantially centered thereon, considering the incoming intersection of each of the observation axes with the reference surface, thereby generating a set of such intersection points corresponding to a series of observation axes, and selecting discrete observation axes in the series so that the set comprises a two-dimensional lattice of points located on the reference surface with a substantially uniform distribution area.
[0016] U.S. Patent No. 9,618,460 in the name of inventors Schoenmakers et al. discloses a method for performing tomographic imaging, the method involving repeatedly directing a charged particle beam through a sample for a series of sample tilts to acquire a corresponding set of images and mathematically combining the images to construct a composite image, wherein the mathematical combination includes using a spectral detector to produce a spectral map of the sample at each of a second series of sample tilts to acquire a set of spectral maps; analyzing the spectral maps to obtain compositional data of the sample; and employing the compositional data in constructing the composite image.
[0017] U.S. Patent No. 10,937,625 in the name of inventor Franken et al. discloses a method for imaging a sample, wherein the sample is mounted on a sample holder in an electron microscope, the electron microscope comprising an electron source for generating a high energy electron beam along an optical axis and optical elements for focusing and deflecting the beam so as to irradiate the sample with the electron beam. The sample holder is capable of positioning and tilting the sample relative to the electron beam. The method comprises the steps of acquiring a series of tilted images by irradiating the sample with the electron beam, and changing the position of the sample in parallel during the acquisition of the images so that each image is acquired with an associated unique tilt angle and an associated unique position. In one embodiment, "stroboscopic illumination" of the sample is used. The sample is illuminated in a stroboscopic manner, and the frequency of the illumination pulses can be synchronized with the camera frame readout.
[0018] U.S. Patent No. 11,257,656 in the name of inventor Janssen et al. discloses a rotating sample holder for random angle sampling in tomography. The sample holder holds the sample and can rotate the sample continuously in a single direction when the sample is exposed to a charged particle beam or other radiation source. Typically, the charged particle beam is gated to produce a series of charged particle beam images at random or arbitrary rotation angles. The sample holder can rotate over a complete rotation of the sample. The charged particle beam image is used for tomography reconstruction, and in some cases, the relative rotation angle is used for the reconstruction without the need to enter the absolute rotation angle.
[0019] Pre-grant publication No. 2022 / 0310354-A1 method and related systems and devices, including or providing: holding a sample using a moving stage configured to rotate the sample around a rotation axis, guiding an imaging beam to a first sample position and detecting a first US transmission imaging beam image when the sample is at a first rotational position around the rotation axis, rotating the sample to a second rotational position around the rotation axis using the moving stage, and guiding the imaging beam to a second sample position by deflecting the imaging beam relative to an optical axis of the imaging beam and detecting a second transmission imaging beam image, wherein the second sample position is spaced apart from the first sample position at least relative to the optical axis. Some examples also include guiding the imaging beam to one or more additional sample positions at one or more additional rotational positions and detecting corresponding transmission imaging beam images, wherein the first sample position at the first rotational position, the second sample position at the second rotational position, and the one or more additional sample positions at the one or more additional rotational positions include an imaging beam spiral imaging trajectory for the sample. Summary of the invention
[0020] Disclosed herein are novel alternative methods for three-dimensional tomographic imaging of elongated sections of needle-like or plate-like samples. These novel methods are characterized by the following features:
[0021] Each unique sample tilt within the tilt range of interest is visited only once, and at each tilt angle, multiple image frames or a single camera acquisition in "movie mode" are obtained across the length of the feature of interest while translating the sample relative to an incident beam parallel to the elongated axis of the sample or sample portion. The translation can be performed continuously and smoothly, or alternatively in well-defined discrete steps. The translational movement can be produced using optical adjustments (e.g., using a beam deflector to adjust the imaged portion of the sample), sample stage motion (using x, y actuators), or a combination of both.
[0022] Optionally, stroboscopic imaging or coded exposure imaging (techniques known from optical photography) are used to enhance high-frequency spatial information, thereby reducing or eliminating motion blur. According to one method of stroboscopic imaging, the incident beam may be periodically or otherwise intermittently introduced onto the sample as a result of repeated beam blanking. According to a second method of stroboscopic imaging, the image beam may be periodically or otherwise intermittently interrupted by blanking the beam between the sample and the detector (e.g., a camera). Such forms of stroboscopic imaging can effectively reduce motion blur and may be employed in certain embodiments. However, stroboscopic imaging can severely limit the achievable dose rate of charged particles at the camera, resulting in poor image quality because the beam will be blanked most of the time. Coded exposure imaging using any of the above-mentioned stroboscopic imaging methods provides an alternative that allows more efficient use of the available "dose budget" by using a (pseudo) random beam blanking timing pattern. The resulting data can be reliably deconvolved to recover spatial information from blurred images.
[0023] Optionally, the tilt and / or curvature of the sample can be tracked in near real time by analyzing the needle center position in a movie frame or in multiple still images. This information can be used to provide near real time corrections of the direction of beam translation or stage movement, as well as corrections of beam focus. Focus correction can be achieved by stage movement parallel to the z-axis (i.e., the direction of the incident beam) or by objective lens adjustment. This optional tracking feature is important in cases where the needle is mounted at a non-zero angle relative to the tilt axis, or when the sample is significantly damaged or bent compared to the size of the field of view or the maximum acceptable change in focus.
[0024] - Optionally, for each tilt angle, the direction of movement of the sample relative to the incident beam may alternate between forward and backward movement along the length of the sample to further minimize necessary movement during image acquisition.
[0025] After (or during) acquiring a plurality of images at each tilt angle, a two-dimensional elongated projection image of the sample is constructed using data from the images acquired at the corresponding tilt angles. The resulting set of two-dimensional images (each corresponding to a different corresponding tilt angle) is used as input to one or more three-dimensional tomographic reconstructions of the structure of the sample, which can be performed by known techniques. In some cases, for example, a full or partial 3D reconstruction can be obtained while acquiring additional image frames. In some cases, construction of a two-dimensional composite image can alternate with a 3D tomographic reconstruction.
[0026] According to a first aspect of the present teachings, a method of three-dimensional (3D) tomography is provided, the method comprising: (a) providing a sample comprising a slender portion, the elongation of the slender portion defining an elongated axis having an axial length; (b) acquiring a plurality of two-dimensional (2D) composite images of the elongated portion along its axial length, each image corresponding to a projection of the elongated portion parallel to the elongated axis, wherein each image is acquired at a different respective rotation angle of the sample relative to an initial orientation of the sample about a rotation axis substantially coincident with the elongated axis; and (c) combining the plurality of projection images to obtain a 3D tomographic representation of the elongated portion of the sample.
[0027] In various embodiments of the method, each acquisition of a 2D image of a slender portion across its length includes acquiring a series of image frames, all images of the series corresponding to the same rotation angle of the sample around the rotation axis relative to the initial orientation, and each image frame corresponding to a 2D projection image of a portion of the length of the slender portion along the 2D projection image. Each such image frame of each series may overlap with at least one other adjacent image frame in the same series. In this case, the overlapping portion of the image frame can be used to align the frames to prepare for generating corresponding 2D projection images from the image frames. Alternatively, there may be no overlap of image frames within each series, and in this case, each 2D image of the slender portion across its length can be generated from the non-overlapping image frames of the image frames of the corresponding series by digital reconstruction of sample features between adjacent frames. As another alternative, in the case of very precise image positioning, there may be no image overlap, but there is also no lack of coverage between consecutive images. In this case, splicing to create a 2D composite image is unimportant.
[0028] In various embodiments of the method, acquiring each 2D image includes acquiring at least one transmission electron microscope (TEM) image. The thickness of the elongated portion taken perpendicular to the elongated axis may be less than or equal to 1 micron (μm). In various embodiments, acquiring each 2D image may include acquiring at least one image of the emission of photons from the sample in response to charged particles impinging on the sample. In various embodiments, providing the elongated portion of the sample may include generating the elongated portion by ion beam milling of the sample. In various embodiments, a plurality of rotation angles of the sample relative to an initial orientation of the sample may be in a range between ninety degrees and negative ninety degrees.
[0029] In various embodiments of the method, acquiring each series of image frames may include changing the position of the sample relative to an incident flux of photons, electrons, or ions directed onto the sample between the acquisition of consecutive image frames. Changing the position of the sample relative to the incident flux may include changing the position of the sample, changing the position of the incident flux, or changing both the position of the sample and the incident flux. After acquiring each 2D image, the direction of relative movement may be reversed. In various embodiments, the relative position of the sample and the incident flux may be continuously changed during the acquisition of each series of image frames. During this continuous position change, the incident flux may be pulsed, whereby each image frame in each series of image frames corresponds to a corresponding pulse. In various embodiments, the timing of the pulses of the incident flux may be consistent with a pre-generated pseudo-random sequence.
[0030] In a variation, the method may also include: (d) detecting a second elongated portion of the sample based on at least one of the 2D images, the elongation of the second elongated portion defining a second elongated axis having a second axial length; (e) acquiring a plurality of 2D images of the second elongated portion along the second axial length, each image corresponding to a projection of the second elongated portion parallel to the second elongated axis, wherein each image is acquired at a different corresponding rotation angle of the sample relative to an initial orientation of the sample around a second rotation axis substantially coinciding with the second elongated axis; and (f) combining the plurality of projection images of the second elongated portion to obtain a 3D tomographic representation of the second elongated portion of the sample.
[0031] According to a second aspect of the present teachings, a charged particle beam device is disclosed, wherein the device includes: (a) a charged particle source; (b) a charged particle beam column, which is configured to direct a charged particle beam from the charged particle source onto a slender portion of a sample mounted on a rotatable and translatable sample stage; (c) a charged particle beam optical device or a photon optical device, which is configured to transfer charged particles or photons transmitted through the sample or generated at the sample to a detector in response to the charged particle beam impinging on the sample, wherein transferring the charged particles or photons includes generating an image frame, which includes a projection of the slender portion along a portion of its length; and (b) a controller environment, which is electrically coupled to one or more of the charged particle source, the rotatable and translatable sample stage, the charged particle beam optical device or the photon optical device and the detector, the controller environment including program instructions that are operable and configured to execute any one of the method embodiments disclosed herein.
[0032] According to a third aspect of the present teachings, a computer program product is disclosed comprising program instructions operable to perform the method embodiments disclosed herein when executed on a controller environment of a three-dimensional tomographic imaging system. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to best understand the features and advantages of the teachings of the present disclosure, the reader is referred to the accompanying drawings, which are to be viewed in conjunction with the detailed description of certain examples provided below. It should be understood that these drawings depict only exemplary embodiments of the invention and therefore should not be considered limiting of its scope, and these examples will be described and explained with reference to the accompanying drawings, in which:
[0034] Figure 1A and Figure 1B is a schematic diagram of an image acquisition sequence that may be employed during the practice of routine tilt-series tomographic analysis;
[0035] Figure 2A and Figure 2B is a schematic diagram of an image acquisition sequence that may be employed during the practice of conventional helical tomographic analysis;
[0036] Figure 3 is a schematic diagram of an exemplary charged particle microscope imaging system that can be used to perform methods according to the present teachings;
[0037] Figure 4A and Figure 4B is a schematic diagram of an image acquisition sequence that may be employed during the practice of three-dimensional tomographic analysis in accordance with the present teachings; and
[0038] Figure 5is a schematic diagram of acquiring a sequence of overlapping image frames across the length of an elongated sample in preparation for constructing an elongated composite image from the image frames in accordance with the present teachings;
[0039] Fig. 6A is another schematic diagram of acquiring a sequence of overlapping image frames across the length of an elongated sample in preparation for constructing an elongated composite image from the image frames in accordance with the present teachings;
[0040] Figure 6B is a schematic diagram of a first method of acquiring multiple images along the length of an elongated sample, whereby a series of short pulses of a charged particle beam are incident on the sample as the sample is continuously moved through the field of view of a charged particle beam imaging system;
[0041] Figure 6C is a schematic diagram of a second method of acquiring multiple images along the length of an elongated sample, wherein during each image acquisition, the sample remains stationary relative to the incident charged particle beam, and the sample and / or the charged particle beam are moved relative to each other between acquisitions;
[0042] Fig.6D is a schematic comparison between imaging using a conventional single exposure of an object and imaging using coded exposures of a moving object;
[0043] Figure 7 is a diagram of a hypothetical, generally elongated needle-shaped sample that does not conform to a single elongated axis.
[0044] Figure 8 is a flow chart of a first exemplary method of three-dimensional tomographic analysis according to the present teachings; and
[0045] Fig. 9 is a flow chart of a second exemplary method of three-dimensional tomographic analysis according to the present teachings. DETAILED DESCRIPTION
[0046] Typically, during inspection and / or analysis of a sample by charged particle tomography, the sample may be positioned on a rotatable sample stage of a charged particle microscope for repeated exposure to a charged particle beam (CPB). As used herein, the term "charged particle" encompasses electrons or ions (typically positive ions, such as gallium ions or helium ions, for example, although negative ions may also be possible). For example, it may also be a proton. The term "charged particle microscope" refers to a device that uses a charged particle beam to create a magnified image of an object, feature, or component that is typically too small to be seen with satisfactory detail by the naked human eye. In addition to having an imaging function, such a device may also have a processing function; for example, it may be used to locally modify a sample by removing material from the sample ("grinding" or "ablation") or adding material to the sample ("deposition"). The imaging and processing functions may be provided by the same type of charged particles, or may be provided by different types of charged particles; for example, a focused ion beam (FIB) microscope may employ a (focused) ion beam for processing purposes and an electron beam for imaging purposes (a so-called "dual-beam" microscope), or it may perform processing using a relatively high-energy ion beam and imaging using a relatively low-energy ion beam.
[0047] The tomography examples disclosed herein are generally described with reference to a transmission electron microscope, but other CPBs may also be used. As used herein, a "column" generally refers to one or more CPB optical elements, or a combination of elements such as a CPB source, a CPB lens, a CPB deflector, a CPB aperture, an stigmator, or other CPB optical elements. Such optical elements can be used to generate pulsed CPBs, which can be directed to the sample to provide pulsed exposures. Such pulsed exposures are generally referred to as "stroboscopic" exposures to indicate that the effective exposure time is short enough compared to the time required to rotate or translate the sample by a detectable amount, so that a suitable image can be acquired, i.e., an image without excessive motion-induced blur. In some disclosed examples, the CPB or CPB column may be energized to generate a stroboscopic CPB exposure. The term "sample holder" refers to any type of workbench, platform, arm, etc., on which a sample can be mounted and held in place. Generally speaking, such a sample holder will be included in a sample holder stage assembly, with which the sample holder can be precisely positioned with several degrees of freedom, for example, by means of an electric actuator.
[0048] As used in the following description and in the claims, the singular forms "a", "an" and "the" include plural forms unless the context clearly indicates otherwise. In addition, the term "comprising" means "including". As used herein, the term "substantially" when used to indicate that an object, system, variable or parameter "substantially" has, is, includes or is equal to a specific shape, size, orientation, value or other property, means that the referenced object, system, variable or parameter has, is, includes or is equal to the specific shape, size, orientation, value or other property within the normal range of experimental error. In this sense, as known to those of ordinary skill in the art, normal experimental error is considered to be a conventional and generally accepted error value, which depends on the identity and technical complexity of the referenced object, system, variable or parameter. As used herein, the term "substantially" when used to indicate that an object, system, variable, or parameter "substantially" has, is, includes, or is equal to a specific shape, size, orientation, value, or other property, means that the referenced object, system, variable, or parameter merely intentionally approximates the specific shape, size, orientation, value, or other property, or is an approximation of the specific shape, size, orientation value, or other property.
[0049] The systems, devices, and methods described herein should not be construed as limiting in any way. Rather, the present disclosure is directed to all novel and non-obvious features and aspects of the various disclosed embodiments, either individually or in various combinations and sub-combinations with one another.
[0050] Although the operations of some of the methods disclosed herein are described in a particular sequential order for ease of presentation, it should be understood that this manner of description encompasses rearrangement unless a particular order is required by the specific language described below. For example, the operations described in sequence may be rearranged or performed simultaneously in some cases. In addition, for simplicity, the accompanying drawings may not show the various ways in which the disclosed systems, methods, and devices can be used in conjunction with other systems, methods, and devices. In addition, the specification sometimes uses terms like "generate" and "provide" to describe the disclosed methods. These terms are high-level abstractions of the actual operations performed. The actual operations corresponding to these terms will vary depending on the specific implementation and can be easily discerned by a technician in the field.
[0051] The term "image" is used herein to refer to a display image such as on a computer monitor, or a digital or analog representation that can be used to generate a display image. The digital representation can be stored in various formats, such as JPEG, TIFF, Medical Research Council (MRC), Electronic Event Representation (EER) or other formats. An array detector or a single element detector can be used to generate an image signal together with a suitable scan of the sample. In most practical examples, the image is two-dimensional. As used herein, the term "image" also encompasses the terms "image frame" and "composite image", and their plural forms. In other words, an "image frame" is also an "image" and can be referred to as an image. Similarly, a "composite image" is also an "image" and can be referred to as an image. In the specific use of this article, an "image frame" is an image obtained from a portion of a sample that is smaller than an area of interest because the sample portion is exposed to photons or charged particles (e.g., electrons or ions). In contrast, the term "composite image" as used herein is an image obtained by combining the information of two or more image frames, and the image belongs to a sample volume or sample area that is larger than the sample area or sample volume imaged by any one of the separate image frames from which the image is obtained. Thus, as the term is used herein, a composite image includes data from more than one image frame.The area extent of a sample imaged by any individual such image frame may or may not overlap with the area extent of a sample imaged by any other such image frame.
[0052] The sample stage rotation and translation as discussed herein may be provided with various types of motors and actuators, including DC motors, stepper motors, rotary piezoelectric motors, AC motors or other devices. The rotation angle and translational movement distance may be detected with an optical encoder, a magnetic encoder or other device. The terms "continuous rotation" and "continuous translation" refer to rotation and translation that are allowed to continue without stopping during image acquisition. For example, a sequence of drive signals applied to a stepper motor to produce rotational or translational movement may continue during image acquisition, and no waiting time is required at a selected imaging angle or position. Although a stepper motor is used, the rotation is referred to as continuous. In other examples, a continuous (i.e., non-stepping) motor may be allowed to freely rotate or translate the sample. Images may be acquired at random times.
[0053] Figure 3An exemplary charged particle microscope imaging system 100, specifically, a transmission electron microscope, that can be used to perform the methods of the present teachings is shown. The sample 102 can have various shapes and cross-sections. However, for the purpose of discussion, it is assumed that the sample includes at least one portion having an elongated shape (such as a needle-like or elongated plate-like shape). The imaging system 100 includes an electron beam source 104, which is configured to generate a beam 106 along an optical axis 107 (an axis that is generally aligned with the z direction of the device as shown in the figure). The imaging system 100 also includes an electron optical device of a charged particle beam column 180, such as one or more objective lens elements 108, 110 configured to guide the light beam 106 onto the sample 102 and through the sample. The light beam 106 is typically received by an electronic imaging detector 112 after being transmitted through the sample 102, so as to provide an image of the sample 102 at a specific orientation. Suitable electronic detectors may include direct and indirect detectors or sensors, including CCD, CMOS, pixel arrays, photomultiplier tubes, etc. The light beam 106 may be focused and / or collimated one or more times before being received by the detector 112, including at or near the sample.
[0054] Still reference Figure 3 , a particular orientation of the sample 102 is provided by a moving stage 114, which is configured to maintain and change the orientation of the sample 102 by rotating about an axis 115 (shown in the figure as aligned with the x-direction) and translating along one or more axes (e.g., x, y, and z). Thus, the moving stage 114 includes both a translation stage component and a rotation stage component. The moving stage 114 can be configured to move and position the sample (such as at a center or offset position) or compensate for drift. In further examples, the beam source 104 can include other energy sources such as x-rays, ultrasound, etc., the optical device can be configured to direct an imaging beam other than electrons, and the detector 112 can be configured to detect an imaging beam other than electrons (e.g., x-ray photons).
[0055] The imaging system 100 also includes an image deflector 120 configured to direct the image beam 106 onto the imaging detector 112 after having passed through a portion of the sample 102 that is offset from the optical axis 107. An optional beam deflector 116 may be configured to direct the beam 106 away from the axis 107 before the beam 106 impinges on the sample 102, such that the beam 106 impinges on the sample at a translational position relative to the intersection of the axis 107 and the sample 102, e.g., in the x-direction or in both the x- and y-directions. Figure 3In the illustrative depiction of FIG. 1 , the beam 106 is illustrated as being translated in the x-direction by an optional beam deflector 116. In some examples, the beam deflector 116 includes a beam deflection coil 118, such as a magnetic or electrostatic coil, a DC coil, an AC coil, a fast scanning coil, etc. An image deflector 120 is provided with a coil 122 similar to the coil 118. The coil 122 can be used to deflect the beam 106 back toward the optical axis 107. By operating the image deflector 120 in conjunction with the motion of the moving stage 114 and / or the operation of the beam deflector 116 (if present), various imaging sequences can be efficiently collected in tomography and other applications, for example, forming spiral trajectories as well as additional or other trajectories or shapes, including double spirals, circles, serpentines, saddles, 2D lattices of dots, 3D lattices, etc. Examples of the imaging system 100 may also include a beam blanking unit 123 that is configured to block the beam for a selected duration, such as between imaging times or during beam deflection or stage movement.
[0056] The various components of the imaging system 100 may be coupled to a controller environment 128 configured with one or more controllers, computing devices, processors, memory, and / or other components to coordinate the directing of the beam 106 to the sample 102, the orientation and movement of the sample 102 by the motion stage 114, and the deflection of the beam 106 using the beam deflector 116. The controller environment 128 may also be configured to control the lens elements 108, 110, the emission of the beam 104, the blanking using the beam blanker 123, the temperature and / or pressure of the cryogenic volume 124, and the timing and / or synchronization of signals of the detector 112 with the movement (i.e., rotation and translation), beam blanking or pulsing, and / or image data acquisition of the sample. The control environment 128 may be configured with computer-readable instructions that automatically cause the various components of the imaging system 100 to operate in a coordinated manner so as to perform the novel tomographic methods disclosed in this document.
[0057] The control components of the controller environment 128 may include a desktop or portable computer, a mobile device, a tablet computer, a logic controller, etc. The processor may include a CPU, a GPU, an ASIC, a PLC, an FPGA, a PLD, a CPLD, etc., which may perform various data processing or I / O functions associated with the controller environment 128, such as control and / or data processing. The memory may be volatile or non-volatile (e.g., RAM, ROM, flash memory, hard drive, optical disk, etc.), fixed or removable, and connected to the processor. The memory may provide storage capacity for one or more computer-readable media. The controller environment 128 may also be located in a distributed form so that applications and tasks are executed by remote processing devices linked through a communication network. Program modules and logic may be located in local and remote memory storage devices. In some examples, components in the controller environment 128 do not need to be linked to other components.
[0058] FIG. 4A to FIG. 4B A sequence of image acquisitions that may be performed according to the method of 3D tomography described herein is schematically illustrated. Figure 4A The shaded box depicted in Figure 1A and Figure 2A The shaded boxes shown in the drawings are because each such shaded box relates to the acquisition of a single image of a portion of the sample and represents the intersection of the incident charged particle beam with an imaginary cylinder 152 (not shown) surrounding the sample, which is at least approximately disposed on a rotation axis 115 of the sample holder, wherein the rotation axis 115 coincides with the cylinder axis 117. Although the rotation axis is instrument dependent, the cylinder axis 117 may be defined or selected based on the particular geometry of the sample. Figure 4B The shaded box depicted in Figure 1B and Figure 2B The shaded boxes depicted in the figure are similar in that, for ease of viewing, they are schematically represented with Figure 4A The same shaded box as in FIG. 1 , but with the surface of cylinder 152 laid flat on the plane.
[0059] The method according to the present teachings involves a tomographic reconstruction of the structure of an elongated portion of a sample, the elongation direction or elongation axis of the sample being approximately parallel to the axes 117, 115. Preferably, but not necessarily, the elongated portion of the sample comprises an elongated axis 117 which is substantially parallel to the axis of rotation 115 and substantially coincides with the axis of rotation. Such conditions are substantially satisfied if the elongated portion itself is cylindrical or approximately cylindrical (i.e., an elongated needle or column) or comprises a symmetry plane whose intersection defines an axis of symmetry (e.g., a plate-like sample). Such sample shapes may be manufactured during sample preparation prior to analysis by, for example, ion beam milling. However, actual elongated samples may include a thickness that varies along their length, or may include an irregularly shaped surface, such as a surface with various protrusions or pits, which does not completely conform to the axis of symmetry. In this case, the method according to the present teachings may include the following steps: defining or selecting the elongated axis of the sample or the elongated portion of the sample as a line passing through the maximum length of the sample or the elongated portion, which may coincide, align or be parallel to the axis of rotation 115. Furthermore, actual needle- or plate-shaped samples may not be perfectly straight, but may include bends or bends. Methods for imaging such samples are described later in this document.
[0060] Return to FIG. 4A to FIG. 4B 1, it should be noted that each shaded box represents the acquisition of a single corresponding image frame. The first five acquired image frames are represented by shaded boxes f1, f2, f3, f4, and f5 in their acquisition sequence. For the purpose of example, it is assumed that only four such image frames (e.g., image frames f1, f2, f3, and f4) are needed to image the full length W of the region of interest of the sample as measured parallel to the axis 115. Figure 4A 1, the length W is indicated in a vertical orientation for illustrative purposes only. The lengths of image frames taken parallel to axis 115 are indicated as w1, w2, w3, and w4. In many cases, these image frame lengths may be equal to one another. Thus, Figure 4A and Figure 4B As shown, w1=w2=w3=w4. However, the image frame lengths are not necessarily all equal to each other. For ease of illustration, the image frames are illustrated in the drawings as being elongated parallel to the axes 115, 117. In practice, the image frames may not be elongated. For example, the camera frames captured by the TEM camera may be square or approximately square (e.g., see Figure 5 ).
[0061] like FIG. 4A to FIG. 4B As further shown, consecutive frames may overlap each other. Such regions are Figure 4B In FIG. 1 , the overlapping regions o1, o2, o3, and o4 are indicated by dark shadows. For example, the overlapping region o1 is formed by the overlap of the image frame f1 and the image frame f2. Similarly, Figure 4BThe residual overlap depicted in is the area where the area imaged by the later image frame overlaps a portion of the area imaged by the earlier image frame. The overlapping area allows identification of common image features that can be used to properly align the image frames with each other and to equalize the exposure of a set of images in preparation for generating a composite image from multiple image frames, as discussed further below.
[0062] like Figure 4A and Figure 4B As shown, the image frames f1, f2, f3 and f4 are all acquired without the intervention of the rotating stage; that is, the image frames f1, f2, f3 and f4 are all acquired at the initial angle setting of the rotating stage, and the initial angle setting can be defined as zero degrees, such as Figure 4B This set of image frames was acquired with the rotating stage set to zero degrees. Figure 4B Each of these image frame acquisitions generates a different corresponding portion of the elongated sample or a projection image of the elongated sample portion, with some overlap between the different portions. The charged particle beam is directed to a first side (e.g., the "top" side; see Figure 3 ) and acquires each image frame by detecting charged particles and / or photons that emerge from an opposite side (e.g., the “bottom” side) of a portion of the sample in response to the light beam impinging on the sample.
[0063] To reposition the imaging position from one portion of the sample to another portion (e.g., from image frame f1 to image frame f2) during acquisition of image frame set C1, any of the following is performed: (a) in a first direction (e.g., toward Figure 4A (b) moving the incident light beam in a second direction opposite to the first direction (e.g., toward Figure 4A (a) translating the sample stage and the sample mounted thereon parallel to the axis of rotation; or (c) a combination of a and b, i.e., moving the incident beam in a first direction and moving the sample stage in a second direction. Figure 3 As shown, the movement of the incident light beam parallel to the rotation axis 115 can be controlled by adjusting the voltage applied to the electrodes of the beam deflection unit 118 and / or other electrodes of the charged particle beam column 180, which adjustment may be supplemented by supplementary adjustment of other electrodes of the descanning unit 122 and / or the sub-stage image transfer assembly 182. The translational movement of the sample stage can be controlled by an actuator (such as a piezoelectric actuator), as is known in the art.
[0064] After the acquisition of all image frames of the image frame set C1 is completed, the rotating stage mechanism is operated to rotate the elongated sample or sample portion by a predetermined angular increment, such as thirty degrees, as shown in FIG. Figure 4B Then, the second set of image frames is acquired while the rotating stage is kept at a fixed angle of 30 degrees. Figure 4B As shown, the acquired second set of image frames is labeled as image frame set C2, which, according to the example system, consists of image frames f5 to f8. In general, various tilt angles may be accessed in any order, depending on the needs of the experiment. For example, for the purpose of minimizing the dose of electrons received by certain parts of the sample, some sequences of tilt angles may be preferred. In some sequences, positive tilt angles may alternate with negative tilt angles, where the magnitude of the tilt angle increases over time. An example of this type of sequence would be tilt angles in the order of 0, +3, -3, -6, +6, +9, -9, -12, +12, etc. (all in degrees). Alternatively, a sequence of coarse tilt angle increments (e.g., in 30 degree steps) may be followed by a refinement of a sequence using smaller tilt angle increments (e.g., in 10 degree increments). With the latter technique, data can be obtained relatively early in the imaging sequence using tilt angles that provide the most informative structural information when the total received electron dose is still low.
[0065] To reposition the imaging position from one part of the sample to another (e.g., from the position corresponding to image frame f5 toward the position corresponding to image frame f8), the incident beam and / or the translation stage are moved parallel to the rotation axis 115 as previously described. However, for efficiency of data collection, it may be the case that the direction of relative motion between the sample and the incident beam during acquisition of image frame set C2 is opposite to the direction of relative motion utilized during acquisition of image frame set C1. This change of direction between image frame sets is determined by Figure 4A . According to this mode of operation, image frames of C2 are acquired in sequence from image frame f5 to image frame f8, all without any rotation of the rotation stage. After acquiring image frame f8 of image frame set C2, the rotation stage mechanism is operated so as to rotate the elongated sample or sample portion to a new tilt angle. After the rotation, image frames f9 to f10 of the next image frame set are acquired in sequence during translation of the incident beam and / or the translation stage. 12. As described above, the tilt angles may be accessed in any suitable order. For example, the sample may be rotated at another predetermined angular increment that may be equal to the angular increment employed in the previous rotation. Alternatively, the sample may be rotated according to a particular tilt angle sequence that, as noted in the preceding paragraph, is selected to minimize total electron dose or to utilize tilt angles that provide useful structural information early in the imaging sequence. In addition, for each next acquired set of image frames, the direction of movement of the sample relative to the incident charged particle beam may be reversed relative to the direction of movement during acquisition of the immediately preceding set of image frames. For example, such a change in relative movement is determined by Figure 4A Indicated by the dotted arrow in .
[0066] Figure 5 According to the method taught in this invention, each image frame set (C1, C2, ..., C j ,……,C z ) are combined into a single corresponding composite image. Each such composite image represents a projection of a portion of an elongated region of the sample corresponding to a single rotation angle. For example, referring to Figure 4B , the image data of the image frames f1, f2, f3 and f4 are combined into a first elongated composite image (at FIG. 4A to FIG. 4B ), the first elongated composite image is a projection of features and / or structures of the sample obtained at an angular rotation of zero degrees of the sample (i.e., obtained at a reference orientation). Similarly, the image data of image frames f5 to f8 are combined into a second elongated composite image, which is a projection of features and / or structures of the sample obtained at an angular rotation of thirty degrees relative to the reference angle.
[0067] Preferably, but not necessarily, during the acquisition of the separate sets of image frames, the movement of the sample relative to the incident charged particle beam is such that at least a portion of the sample is imaged in more than one image frame. In other words, the relative movement is such that there is overlap between a first region of the sample imaged in a first image frame and a different region of the sample imaged in a second image frame. For example, such an overlapping region is FIG. 4A to FIG. 4B Some of these are denoted as overlapping regions o1, o2, and o3. Overlapping regions are also indicated by the boundaries of the image frames, as indicated by the frame length dimension markers showing the image frame lengths w1, w2, w3, and w4. Such overlapping regions can be useful in aligning the image frames during subsequent reconstruction of the composite image. Alignment between images may often be required in order to provide a reconstruction of adequate image quality. Generally speaking, alignment involves determining from which direction and from which position the image is acquired relative to the sample. Even with the use of high-quality translation stages, beam precision, and other high-quality instrumentation, angstroms are required. Positioning resolution, so as to obtain angstrom resolution in tomographic reconstruction; this level of motion accuracy is not realistic. Therefore, in some cases, alignment may be required to determine where the acquired image projection belongs and is aligned with other image projections. In case of very precise image positioning, there may be no image overlap, but also no gaps between consecutive images. In this case, alignment may not be necessary.
[0068] Figure 5 A more complex example of overlapping image frames is schematically depicted. Figure 5 In the embodiment, the elongated sample 102 is aligned parallel to the axis 117 and is held by the sample holder 114h, which is a movable stage assembly 114 ( Figure 3 ), the moving stage assembly having at least the ability to rotate the sample 102 about a rotation axis 115 that is substantially parallel to and substantially coincident with the axis 117. The moving stage assembly may also have the ability to translate the sample at least in a direction of movement that is substantially parallel to the axis 117. Figure 5 Also depicted are representations of thirty-three different image frames, denoted as image frames f1 to f 33 For purely visual purposes, the image frames are depicted with different shading; the different shading should not have other meanings. For the purpose of illustration, all depicted image frames have a common image length w0 taken parallel to the axis 117. However, in general, the image lengths do not need to be the same. In addition to the image frame f 33 Furthermore, only a portion (ie, one eighth) of each image frame is depicted in the illustrated example, as each subsequent image frame is depicted overlying a corresponding previous image frame.
[0069] exist Figure 5 Within the "center" portion of the sample 102, shown in the figure as having a length W, each sample point is imaged by eight separate image frames. The number of image frames in which a particular sampling point appears gradually decreases toward the "left" and "right" of the center sample portion (as depicted in the figure). For the purposes of aligning images, discarding defective images, or providing an adequate dose of electrons to the sample, the creation of a composite image of the center portion of the sample benefits from maximum overlap. Therefore, in some cases, the length of a subsequently constructed composite image may be limited to the center portion of the sample 102, which is shown as having a width W. However, with the exception of a portion of the sample closest to the sample holder 114, each point on the sample is imaged in at least two image frames.
[0070] Fig. 6Ais a schematic diagram of a collection 600 of eight overlapping image frames f1 to f8, each of length w0 as measured parallel or substantially parallel to the axis about which the sample can rotate, wherein half of the length of each previous image frame is overlapped by the corresponding subsequent image frame. The different shading patterns of the image frames have no meaning other than as a means of distinguishing the frames. Figure 6B , Figure 6C and Fig.6D Schematically depicted is a time sequence of events for three separate modes acquired from image data in accordance with the present teachings.
[0071] According to the first operation mode ( Figure 6B ), causing the sample to be continuously moved according to a velocity vector 605 relative to the position of the device causing the pulse 602 of charged particles to impact the sample, wherein each pulse corresponds to an exposure of a respective image frame that generates a set of image frames. Such repeated exposures are referred to herein as "stroboscopic imaging", and the impact of each such pulse on the sample is referred to herein as an "exposure", similar to optical photography. According to the present teachings, vector 605 ( Fig. 6A ) is parallel or substantially parallel to the axis about which the sample can rotate. Typically, the exposure will be at a constant time interval t f Repeatedly, this constant time interval is referred to herein as the image frame time. Figure 6B The operating mode depicted in the image frame time t f The larger of the following times: (a) the time required for the sample to be continuously moved so that the sample is translated a distance of (w0 / 2) to prepare for the next exposure; or (b) the time required to read the image data of the previous exposure from the detector to the computer memory. Generally speaking, in order to avoid image blur, the duration of each exposure can be made t e T f Much shorter.
[0072] According to the second operation mode (eg, Figure 6C ), the sample remains stationary during each exposure and moves only between exposures to the incident charged particle beam. Such a process may be employed, for example, where relatively long exposures are required to improve image contrast or when the image magnification is so great that any movement of the sample during exposure would result in unacceptable blurring.
[0073] Figure 6C The position of the elongated sample as measured parallel to the axis of rotation and / or the sample elongated axis is depicted as a function of time during such a process. In this figure, the time when the sample is stationary is indicated as a plateau 603, and the time when the sample is translated in preparation for a subsequent exposure is indicated as a ramp 604. Figure 6C is represented as t eThe duration of the exposure (ie, the “exposure time”) corresponds to the plateau 603 where the sample does not move. e All exposures corresponding to a set of image frames or to a given sample will typically be the same, but it is contemplated that in some cases the various exposure times may vary within or across a set of image frames. Figure 6C The duration between successive exposures of a set of image frames of the ramp 604 in FIG. 6 (ie, the “inter-exposure time”) is denoted as t rm During the duration t rm During each inter-exposure period of t, the sample is translated substantially parallel to the instrument rotation axis or the sample rotation axis in preparation for the subsequent exposure. During this same period of time, image data corresponding to the previous exposure can be read out from the sensor into a computer memory, and the charged particle beam can be blanked. The inter-exposure time may also include some dead time during which any residual position drift associated with the previous movement is allowed to dissipate. Thus, the image frame time t f is the exposure time t e and the exposure time t rm For overall efficiency, it is preferred that the exposure time t rm As short as possible. Therefore, a charged particle microscope stage of an apparatus on which a sample is imaged should include a stage translation actuator having a fast response time and a short duration stabilization time.
[0074] As described in the foregoing discussion, a three-dimensional (3D) tomography method according to the present teachings employs constructing a 3D structural representation of an elongated sample or sample portion from a plurality of two-dimensional image strips, each of which is constructed by first acquiring a plurality of image frames without rotating the sample while moving the sample substantially parallel or substantially parallel to the elongation axis or rotation axis. According to the first and second methods described above, any potential motion blur in the image frames that may potentially be caused by sample movement is minimized (i.e., by using short duration pulse exposures as in the described first mode of operation) or substantially eliminated entirely (i.e., by moving the sample only between exposures as in the described second mode of operation).
[0075] According to a third mode of operation according to the present teachings, motion blur is allowed to occur in the raw image frames as the sample is continuously translated without stopping between image frame acquisitions. In addition, according to this third mode of operation, the motion blur is subsequently removed by digital deblurring techniques, such as described by Raskar et al. for conventional optical photography (Raskar, Ramesh, Amit Agrawal and Jack Tumblin, "Coded exposure photography: motion deblurring using fluttered shutter", Acm Siggraph 2006 Papers, pp. 795-804, 2006). Digital deblurring techniques require encoding additional information within each set of raw image frames, wherein the encoding relates to the motion of the imaged object. As described by Raskar et al., when an object moves within the field of view of the camera, during the acquisition of an image of the object, the image is deblurred by a period of time t that would normally correspond to a single conventional exposure 614 (e.g., a single opening and closing of the shutter). e The encoding is performed by acquiring a series of shorter duration images 624 within a period of time, the timing and duration of which are determined according to a binary pseudo-random sequence. Raskar et al. refer to this process as "shaking" the shutter of the camera. However, with respect to the present disclosure, these shorter duration exposures 624 are referred to herein as "partial image frame exposures."
[0076] Using the shaking shutter image acquisition process described by Raskar et al., blur will be observed within a single image frame obtained as a result of accumulating signal (pixel intensity) at the detector during the exposure of such a set of partial image frames. However, within such a blurred image, detail will be encoded within the blur pattern as recognizable changes in pixel intensity in the direction of motion. According to Raskar et al., the process provides a wideband filter that preserves high-frequency spatial detail in the blurred image, making subsequent deconvolution reconstruction of the detail a well-posed problem.
[0077] Charged particle beam microscopy does not necessarily include a conventional mechanical shutter as is commonly employed in optical photography. However, the charged particle beam incident on the sample can be periodically blanked according to a binary pseudo-random sequence (e.g., Figure 3The same effect of shutter "jitter" as described by Raskar et al. can be achieved by applying appropriate voltages to a sub-stage component (e.g., an element of the objective system 110 or the descanning unit 122) disposed between the sample and the detector according to a pseudo-random sequence to achieve a jitter effect by periodically diverting charged particles that have passed through the sample (e.g., sample 102) away from the detector (e.g., detector 112). Alternatively, if the data readout is fast enough, the same effect can be achieved by reading the data from the detector according to a pseudo-random sequence.
[0078] According to the method of the present teachings, after a single set of image frames (e.g., Figure 4B After all individual image frames of one of the image frame sets C1, C2, etc. shown in the figure are acquired, a single composite image including information from all image frames of the set is constructed. Each composite image is an elongated projection of features and / or structures of a portion of the sample, as captured at a particular angular orientation of the sample relative to the rotation axis and / or the elongated axis of the sample (i.e., the common angular orientation at which all image frames constituting the image frame set are acquired). The construction of the composite image may be performed before or during the subsequent acquisition of image frames corresponding to the next image frame set (if any). Alternatively, the construction of all composite images (each composite image corresponding to a respective image frame set) may be performed only after all image frames of all image frame sets of a region of interest of the sample have been acquired.
[0079] The process of constructing a composite image from a set of image frames may include correcting one or more individual image frames for astigmatism and / or other image distortions. If necessary, the process of constructing a composite image may include deblurring of one or more individual image frames. The process of constructing a composite image may include aligning the individual image frames to each other by aligning the portion of each of the one or more overlapping images corresponding to the same area of the sample, preferably after any image distortion and / or deblurring steps. If there is no overlapping area between a pair of adjacent image frames or between several pairs of adjacent image frames, the composite image including those frames can still be constructed by estimating the displacement between non-overlapping images based on the approximate mechanical displacement vector recorded by the sensor during the movement of the sample stage during image acquisition. In this case, there may be gaps corresponding to the unimaged portion of the sample in the composite image. However, such incomplete composite images can convey useful information about the structure of the sample. In some cases, missing structural features can be reconstructed in a two-dimensional composite image by extrapolating features from the boundary image frames into the gaps.
[0080] In some cases, a generally elongated sample may still not conform to a single elongated axis. According to some methods according to the present teachings, a process of acquiring multiple images along the length of the sample as previously described herein may be employed to detect such nonlinearities and other irregularities in the shape of the sample. For example, Figure 7 A hypothetical, generally elongated needle-shaped sample 102 is depicted mounted on a moving stage 114 having the ability to rotate the sample about an x-axis and to translate the sample parallel to the x-axis. As shown, the needle-shaped sample is bent in region 702, which separates straight elongated regions 701 and 703 having non-uniform elongated axes 117 and 117b, respectively, where axis 117 is also the axis of rotation (e.g., the x-axis of the stage device) about which the sample can be rotated.
[0081] According to the various operating modes described herein (see Figure 5 and FIG. 6A to FIG. 6C According to some variations of these methods, automated software may perform preliminary analysis of certain aspects of the images as they are acquired, including edge detection analysis to determine potential irregularities in the shape of the sample, for quality control purposes. The values of the monitored parameters may be recorded relative to time so that later acquired images can be compared to earlier images. Preferably, such comparisons are performed automatically in real time.
[0082] Therefore, about Figure 7 For example, the software may determine that from the image frame f 14 Initially, the apparent center of the needle image (i.e., the midpoint between the "top-most" boundary of needle image 708a and the "bottom-most" boundary of needle image 708b as determined by edge detection) has already deviated in the negative y-axis direction from axis 117. The software can then detect that the rate of further deviation from axis 117 is increasing at image frame f 15 and image frame f 21 , thereby indicating the curvature of the needle. In some cases, the software may alert the user to the deviation from linearity and send a warning to the user or abort the imaging process. Alternatively, if the mobile stage 114 is configured with an actuator that enables it to translate parallel to the y-axis, the software may initiate translation of the mobile stage 114 in the direction of the positive y-axis to bring the apparent center of the image of the sample into the center of the field of view. This process may continue to continue acquiring image frames f 22 to f 31 .
[0083] Given relative to Figure 7With extreme deviations from linearity as shown in , it may be impractical to rotate the sample about axis 117 to obtain additional composite images spanning the full length of the sample, as such necessary rotation about axis 117 would cause portions of sample 102 to be out of focus. Figure 7 For the sample shape shown in , it will only be possible to perform a full 3D tomographic analysis of the sample portion closest to the moving stage. However, by utilizing automatic edge detection in combination with automatic translation adjustment to compensate for nonlinearity, at least one complete two-dimensional composite image can still be generated. In extreme cases where the deviation from linearity is small, the full 3D tomographic method described herein can be continued by continuing to obtain additional two-dimensional composite images by rotating the sample about axis 117.
[0084] Alternatively, if the translation stage includes two or more independent rotation axes, each of the sample elongated axes 117 and 117b can be individually aligned with the device rotation axis 115. For example, if the device rotation axis 115 is assumed to be fixed relative to the device reference frame, and the translation stage 114 also has the ability to rotate the sample about a second rotation axis (not shown) that is parallel to the z-axis and is also assumed to be fixed relative to the device reference frame, then rotation of the sample about the z-axis can cause the second elongated axis 117b to coincide with the fixed rotation axis 115. Subsequently, as described above, a tomogram of the sample region 703 can also be constructed using methods according to the present teachings. Preferably, the translation stage (and / or beam deflector) also has the ability to translate the sample parallel to the y-axis relative to the charged particle beam.
[0085] Figure 8 800 is a flow chart of a first exemplary method of three-dimensional tomographic analysis of an elongated sample or elongated sample portion according to the present teachings. Method 800 involves the collection of image frames as the position of the sample relative to the incident charged particle beam is continuously moved. The movement of the sample relative to the incident charged particle beam may include one or both of the following: (a) movement of a moving stage on which the sample is mounted; and (b) movement of the incident charged particle beam. The latter type of movement may include deflecting the beam by beam deflector 116 and / or repositioning the beam by sub-stage descanning unit 120 ( Figure 3 ).
[0086] In the first step (step 802) of method 800, a region of interest along the length of the sample is determined or selected. In addition, a range of tilt angles to be employed and an order in which the tilt angles are to be accessed during the acquisition of a set of image frames are selected. This step may include selecting an initial tilt angle relative to a reference angular position, and may also include selecting an angular increment by which the sample is to be rotated between acquisitions of a set of image frames. These selections may be made based on previous examination of the sample by optical microscopy or charged particle beam microscopy.
[0087] The determination or selection of the sample region of interest in step 802 means the determination of the sample region to be imaged. The region to be imaged is not necessarily the same as the region to be imaged. The determination of the latter region depends on whether the desired region of interest is to be imaged with sufficient redundancy. For example, only the region of interest captured by the image frame in each of the eight separate image frames is captured. Figure 5 The dimension W in W0 specifies sample points within a portion of the full length of the sample 102. In practice, the actual number of image frames in which a given portion of the sample is imaged is not limited or restricted to any particular number. In many cases, the dimension W0 of the image frame is determined by the field of view of the imaging system. In this case, if it should happen that the portion of the sample specified by the dimension W is actually a region of interest and it is desired to capture every point within the region of interest in at least eight image frames, additional image frames f1 to f7 must also be imaged.
[0088] In method 800 ( Figure 8 ), the sample is rotated to a first tilt angle in step 806. The following five steps, namely steps 808, 809, 810, 812 and 814, are shown as being performed in parallel because according to method 800, image frames are repeatedly acquired with the tilt angle held fixed (step 810), and while the acquisitions are being repeated, the position of the sample is continuously translated relative to the charged particle beam in a direction substantially parallel to the axis of rotation (step 808). Step 808 also encompasses initial translation of the sample to a starting position along the length of the elongated sample after the tilt angle is set (in a previous execution of step 806) before acquiring image frames at the fixed tilt angle. Each image frame acquisition begins at the start of an exposure (step 809) and ends at the end of the same or different exposure (step 809), and may include additional exposures (step 809), such as Fig.6D 624. During the acquisition of multiple exposures, the position of the sample relative to the center of the field of view can be tracked (step 812), and the deviation can be appropriately corrected by adjusting the sample position relative to the incident charged particle beam or by adjusting the beam focus. After each exposure, the position of the last acquired image frame is compared with the boundary of the region of interest (ROI), as selected in step 802; thus continuing to perform steps 808, 809, 810, 812 and 814 until the comparison indicates that the full length of the ROI has been imaged. Control of the exposure (step 809) may include one or more of the following: (a) blocking or unblocking the transmission of the charged particle beam on the sample; (b) blocking or unblocking the transmission of charged particles or photons from the sample to the detector. Any of blanking, unblanking, blocking and unblocking may include operating a shutter, which may include an electronic shutter or a mechanical shutter.
[0089] Still referring to method 800 ( Figure 8), after the full length of the region of interest has been imaged, as determined in step 814, then at least step 819 is performed. Additionally, steps 815 and 817 may be performed prior to performing step 819. However, since steps 815 and 817 are image processing steps, the order of execution may be changed such that steps 815 and 817 are performed after all image frames have been acquired at all tilt angles. Accordingly, steps 815 and 817 are depicted with dashed lines to indicate that they may alternatively be performed after exiting decision step 819 along their "yes" branches and typically, but not necessarily, prior to performing step 821. Furthermore, steps 815, 817, and 821 may be swapped, interchanged, or iterated in order to obtain an optimal two-dimensional image reconstruction or three-dimensional tomographic reconstruction. For purposes of description, steps 815 and 817 are discussed below as if described above. Figure 8 The order shown is the same.
[0090] Optional step 815 may be required to remove image blur caused by continuous motion of the sample during an exposure sequence used to acquire image frames at a particular tilt angle (e.g., the most recently taken tilt angle as set in step 806). The deblurring process may include decoding spatial details that are encoded in changes in pixel intensities taken parallel to the direction of motion. Once all image frames have been acquired at the most recently set tilt angle and after any required deblurring operations have been performed, a composite two-dimensional image is constructed from the complete set of image frames acquired at that tilt angle (step 817). Construction of the composite image may include one or more of the following steps: digitally correcting the image frames for astigmatism and other distortions; aligning image frames of a set of image frames by identifying imaging features common to more than one image frame; and, after alignment, accumulating or averaging the image intensities of the set of image frames on a pixel-by-pixel basis. The accumulating or averaging process may include appropriate corrections to account for varying degrees of overlap in images across the sample (see Figure 5 and its discussion).
[0091] Execution of method 800 repeatedly loops through a set of steps from 800 to 819, with each iteration of the loop including: (a) setting a new tilt angle; and (b) subsequently acquiring a plurality of image frames along the length of the sample, wherein all of the image frames are acquired at the same tilt angle. Although the number of tilt angles employed and the order in which the tilt angles are accessed may be predetermined prior to execution of method 800, some of the tilt angles may be determined during execution of the method based on results observed during previous image acquisitions. Optionally, each iteration of the loop may include image data processing steps 815 and 817, wherein data from a set of recently acquired image frames is optionally deblurred and used to construct a two-dimensional (2D) image corresponding to the recently set tilt angle. Otherwise, steps 815 and 817 may be moved outside the loop and prior to step 821, in which case they include optional deblurring of all image frames acquired at all tilt angles and construction of a plurality of 2D images, each of which corresponds to a corresponding tilt angle. The resulting plurality of elongated 2D composite images comprises a simple tilt series which can then be employed in step 821 to generate a three-dimensional model of the structure of the sample within the region of interest using conventional tomographic techniques.
[0092] Fig. 9 is a flow chart of a second exemplary method of three-dimensional tomographic analysis of an elongated sample or a portion of an elongated sample according to the present teachings. Fig. 9 Method 900 of the present invention relates to image frame collection during discontinuous movement of a sample relative to an incident charged particle beam, whereby image frame acquisition occurs only when there is no relative motion between the sample and the incident light beam (e.g., when the sample is stationary), and such motion occurs only between image frame acquisitions. Steps 902, 906, 914, 917, 919, and 921 of method 900 are the same or similar to steps 802, 806, 814, 817, 819, and 821 of method 800, respectively, and therefore these steps are not further described herein. Steps 908-912 of method 900 are similar to steps 808-812 of method 800. However, although steps 800-812 are repeated in parallel, steps 908-912 are performed in an appropriate sequence. Therefore, movement of the sample and / or light beam to the next field of view (step 908) must occur before opening the shutter or introducing the charged particle beam onto the sample (step 909). Similarly, the step of shutter opening or beam introduction (step 909) must occur before the acquisition of the image frame (step 910), which is naturally followed by shutter closing or beam blanking (step 911) and sample position tracking (step 912). Similarly, for step 817, step 917 only involves the processing of image data and can therefore be performed after exiting decision step 919 along its "yes" branch and before executing step 921.
[0093] As described herein, certain steps of the methods disclosed herein involve image data processing, including tomographic reconstruction. In certain examples, tomographic reconstruction and other data processing steps may be performed remotely, for example in a distributed computing environment, to increase processing speed or to address heavy computational loads. Various tomographic reconstruction techniques may be used, including iterative methods, searching, expectation maximization, weighted back projection, etc. Images may be stored and reused as necessary so that acquired image sequences may be customized or adjusted according to standard reconstruction techniques.
[0094] Image acquisition algorithms utilizing beam deflection, beam blocking, beam blanking, detector shutters, etc. may be embodied as software or firmware instructions executed by a digital computer local to a charged particle beam (CPB) microscope system. For example, any of the disclosed techniques may be performed by a computer or other computing hardware that is part of or coupled to a CPB microscope system, such as a transmission electron microscope (TEM) system or a scanning electron microscope (STEM) system. For example, software for performing any disclosed embodiment may be stored as computer executable instructions on one or more volatile, non-transitory computer readable media that, when executed by one or more processors, cause the one or more processors to perform any disclosed technique. The results of image acquisition, calculation, or reconstruction may be stored (e.g., using a suitable data structure or lookup table) in one or more tangible, non-transitory computer readable storage media and / or may also be output to a user, for example, by displaying data, images, tomographic image slices, 3D reconstructions, etc. on a display.
[0095] The three-dimensional tomography method described herein provides several advantages over conventional tomography methods that rely on tilt series or spiral image acquisition. First, with respect to radiation-sensitive samples, the method disclosed herein can ensure an equal dose of charged particles over the entire region of interest or over the entire sample during image acquisition. This is in contrast to image acquisition using normal tilt series or spiral tomography, during which it is difficult to ensure that all areas of the sample receive an equal amount of charged particles, a condition that is critical to preventing image artifacts. Second, the disclosed method can simplify the stitching problem. The act of stitching separate three-dimensional reconstructed volumes or separate tilt series is notoriously difficult. By adopting the novel method disclosed herein, the only image stitching that may be required is the less difficult process of stitching adjacent two-dimensional images with large overlap. In some cases (e.g., when using a well-calibrated image beam shifting optical device that maintains an accurate record of the image frame position), stitching may not be required at all. Third, compared with spiral or ordinary tomography acquisition that requires more tilt steps, the method disclosed herein facilitates faster image acquisition. Finally, the methods disclosed herein enable tracking of the position of an elongated sample or samples while imaging at each tilt angle, which provides near real-time information about the tilt position and curvature of the sample or sample portion, and thereby allows stage motion corrections to be applied in near real-time. Such information may also provide useful additional metadata that can be used to constrain the 3D reconstruction algorithm.
[0096] Additional Embodiments
[0097] In addition to the foregoing description, the method according to the present disclosure may comply with any one or more of the following items:
[0098] Item 1. A three-dimensional (3D) tomography method, comprising:
[0099] providing a sample including an elongated portion, the elongation of the elongated portion defining an elongated axis having an axial length;
[0100] acquiring, using a charged particle beam microscope, a plurality of two-dimensional (2D) composite images of the elongated portion along its axial length, each 2D composite image being generated by a respective series of image frames and corresponding to a projection of the elongated portion parallel to the elongated axis, wherein each 2D composite image is acquired at a different respective rotation angle of the sample relative to an initial orientation of the sample about a rotation axis substantially coincident with the elongated axis; and
[0101] The plurality of 2D composite images are combined to obtain a 3D tomographic representation of the elongated portion of the sample.
[0102] Item 2. A method of 3D tomography according to Item 1, wherein all images in each series of image frames correspond to the same rotation angle of the sample around the rotation axis relative to the initial orientation, and each image frame in each series of image frames corresponds to a 2D projection image of a portion of the elongated portion along its length.
[0103] Item 3. A method of 3D tomography according to Item 2, wherein each image frame in each series overlaps with at least one other adjacent image frame in the same series.
[0104] Item 4. A method of 3D tomography according to Item 2, wherein there is no overlap of image frames within each series, and wherein each 2D composite image of the elongated portion across its length is generated from non-overlapping image frames in a corresponding series of image frames by digitally reconstructing sample features between consecutive frames.
[0105] Item 5. A method of 3D tomography according to any one of Items 2 to 4, wherein each image frame in each series of image frames is generated by detecting the response of the sample to a pulse of an incident electron or ion beam incident on a portion of the sample, wherein the relative position of the sample and the beam changes between consecutive pulses.
[0106] Item 6. A method of 3D tomography according to any one of Item 5, wherein acquiring each 2D composite image comprises acquiring at least one image frame using a transmission electron microscope (TEM) image.
[0107] Item 7. A method of 3D tomography according to any preceding item, wherein the thickness of the elongated portion taken perpendicularly to the elongated axis is no greater than 1 micrometer (μm).
[0108] Item 8. A method of 3D tomography according to any of the preceding items, wherein acquiring each 2D image comprises acquiring at least one image of photons emitted from the sample in response to electrons or ions impinging on the sample.
[0109] Item 9. A method of 3D tomography according to any one of Items 1 to 8, wherein providing the sample comprising the elongated portion comprises producing the elongated portion by ion beam milling the sample.
[0110] Item 10. A method of 3D tomography according to any one of Items 1 to 9, wherein the multiple rotation angles of the sample relative to the initial orientation of the sample are in a range between ninety degrees and negative ninety degrees.
[0111] Item 11. A method of 3D tomography according to any one of Items 2 to 4, wherein acquiring each series of image frames includes changing the position of the sample relative to an incident flux of photons, electrons or ions directed onto the sample between acquisition of consecutive image frames.
[0112] Item 12. A method of 3D tomography according to Item 11, wherein the relative positions of the sample and the incident flux are continuously changed during acquisition of each series of image frames.
[0113] Item 13. A method of 3D tomography according to Item 12, wherein each image frame in each series of image frames corresponds to a corresponding time increment, during which the sample is exposed to the incident flux, wherein the timing of the pulses of the incident flux is consistent with a pseudo-random sequence.
[0114] Item 14. A method of 3D tomography according to Item 11, wherein the movement of the sample relative to the incident flux is reversed after acquiring each series of image frames.
[0115] Item 15. The 3D tomography method according to any of the preceding items, further comprising:
[0116] detecting a second elongated portion of the sample based on at least one of the 2D images, an elongation of the second elongated portion defining a second elongated axis having a second axial length;
[0117] A second plurality of 2D composite images of the second elongated portion along the second axial length are acquired using the charged particle beam microscope, each of the second plurality of 2D composite images being generated from a second image corresponding to a projection of the second elongated portion.
[0118] Item 16. A method of 3D tomography according to any one of claims 2 to 4, wherein acquiring each image frame comprises accumulating image data from multiple partial image frame exposures, wherein the timing of each partial image frame exposure is controlled by a predetermined binary pseudo-random sequence.
[0119] Item 17. The method of 3D tomography according to any one of claims 2, 3, 4 and 16, further comprising deblurring the image frames.
[0120] Item 18. A charged particle beam device, comprising:
[0121] Charged particle sources;
[0122] a charged particle beam column configured to direct a charged particle beam from the charged particle source onto an elongated portion of a sample mounted on a rotatable and translatable sample stage;
[0123] charged particle beam optics or photon optics configured to transfer charged particles or photons transmitted through or generated at the sample to a detector in response to the charged particle beam impinging on the sample, wherein transferring the charged particles or photons comprises generating an image frame that is a projection of features and / or structures of the elongated portion along a portion of its length; and
[0124] A controller environment electrically coupled to the charged particle source, the rotatable and translatable sample stage, the charged particle beam optics or photon optics and one or more of the detectors, the controller environment comprising program instructions operable to perform a method according to any one of items 1 to 17.
[0125] Item 19. A computer program product comprising program instructions, which, when executed on a controller environment of a 3D tomographic imaging system, are operable to perform a method according to any one of items 1 to 17.
Claims
1. A method of three-dimensional (3D) tomography, comprising: providing a sample including an elongated portion, the elongation of the elongated portion defining an elongated axis having an axial length; acquiring, using a charged particle beam microscope, a plurality of two-dimensional (2D) composite images of the elongated portion along its axial length, each 2D composite image being generated by a respective series of image frames and comprising a projection of the elongated portion parallel to the elongated axis, wherein each 2D composite image is acquired at a different respective rotation angle of the sample relative to an initial orientation of the sample about a rotation axis substantially coincident with the elongated axis; as well as The plurality of 2D composite images are combined to obtain a 3D tomographic representation of the elongated portion of the sample.
2. A method for 3D tomography according to claim 1, wherein all image frames in each series of image frames correspond to the same rotation angle of the sample around the rotation axis relative to the initial orientation, and each image frame in each series of image frames corresponds to a 2D projection image of a portion of the slender portion along its length.
3. The method of 3D tomography of claim 2, wherein acquiring each series of image frames comprises changing the position of the sample relative to an incident flux of charged particles directed onto the sample between acquisition of consecutive image frames.
4. The method of 3D tomography according to claim 3, wherein each image frame in each series overlaps with at least one other adjacent image frame in the same series.
5. A method of 3D tomography according to claim 3, wherein there is no overlap of image frames within each series, and wherein each 2D image of the elongated portion across its length is generated from non-overlapping image frames in a corresponding series of image frames by digitally reconstructing sample features between adjacent image frames.
6. The method of 3D tomography according to any one of claims 1 to 5, wherein the plurality of rotation angles of the sample relative to the initial orientation of the sample are in a range between ninety degrees and negative ninety degrees.
7. A method of 3D tomography according to any one of claims 3 to 6, wherein the direction of the relative movement of the sample with respect to the incident flux is reversed after acquisition of each series of image frames.
8. A method of 3D tomography according to any one of claims 3 to 7, wherein each image frame in each series of image frames corresponds to a respective time increment during which the sample is exposed to a plurality of pulses of the incident flux, wherein the timing of the pulses of the incident flux is consistent with a pseudo-random sequence.
9. A method of 3D tomography according to any one of claims 3 to 7, wherein each image frame in each series of image frames corresponds to a respective time increment during which a detector is exposed to a plurality of pulses of a flux of electrons from the sample to the detector, wherein the timing of the pulses of the flux is consistent with a pseudo-random sequence.
10. The method of 3D tomography according to any preceding claim, further comprising: detecting a second elongated portion of the sample based on at least one of the 2D images, an elongation of the second elongated portion defining a second elongated axis having a second axial length; acquiring, using the charged particle beam microscope, a second plurality of 2D composite images of the second elongated portion along the length of the second axis, each of the second plurality of 2D composite images being generated by a respective series of image frames and comprising a projection of the second elongated portion parallel to the second elongated axis, wherein each of the second plurality of 2D composite images is acquired at a different respective rotation angle of the sample relative to an initial orientation of the sample about a second rotation axis substantially coincident with the second elongated axis; as well as The second plurality of 2D composite images of the second elongated portion are combined to obtain a 3D tomographic representation of the second elongated portion of the sample.
11. A method of 3D tomography according to any preceding claim, wherein the thickness of the elongate portion taken perpendicularly to the elongate axis is no greater than 1 micrometer (μm).
12. A method of 3D tomography according to any preceding claim, wherein acquiring each 2D composite image comprises acquiring at least one image of photons emitted from the sample in response to electrons or ions impinging on the sample.
13. A method of 3D tomography according to any one of claims 3 to 7, wherein the relative position of the sample and the incident flux is continuously changed during the acquisition of each series of image frames.
14. A charged particle beam device comprising: Charged particle sources; a charged particle beam column configured to direct a charged particle beam from the charged particle source onto an elongated portion of a sample mounted on a rotatable and translatable sample stage; charged particle beam optics or photon optics configured to transfer charged particles or photons transmitted through or generated at the sample to a detector in response to the charged particle beam impinging on the sample, wherein transferring the charged particles or photons comprises generating an image frame that is a projection of the elongated portion along a portion of its length; and A controller environment electrically coupled to the charged particle source, the rotatable and translatable sample stage, the charged particle beam optics or photon optics and one or more of the detectors, the controller environment comprising program instructions operable to perform the method according to any one of claims 1 to 13.
15. A computer program product comprising program instructions operable to perform the method according to any one of claims 1 to 13 when executed on a controller environment of a 3D tomographic imaging system.
Citation Information
Patent Citations
Acquisition and processing of data in a tomographic imaging apparatus
US10254419B2
Method of imaging a sample using an electron microscope
US10937625B2
Rotating sample holder for random angle sampling in tomography
US11257656B2
Manipulator for rotating and translating a sample holder
US7884326B2
Image data processing
US8854430B2
