Method of operating a multi-beam scanning charged particle microscope and circuitry for controlling operation of a multi-beam scanning charged particle microscope

Abstract

This invention discloses a technique for determining at least one aberration of multiple primary charged particle beams (3.1, 3.2, 3.3) in a multi-beam scanning charged particle microscope, such as deflection or spot distortion of the multiple primary charged particle beams. This at least one aberration is determined based on an estimate of the charge accumulation (1050) of the sample (7), i.e., the local surface charge on the sample surface. Thus, a prediction of future point charge accumulation can be determined.

Description

[Technical Field]

[0001] Various embodiments of the present invention are generally related to techniques for mitigating the charge accumulation effect when a sample is irradiated by a charged particle beam in a multi-beam scanning charged particle microscope. Various embodiments of the present invention are specifically related to determining aberration estimates of multiple primary charged particle beams based on said charge accumulation, and subsequently compensating for such aberrations. [Previous Technology]

[0002] In scanning electron microscopy (SEMs) or other types of scanning charged particle microscopy (e.g., scanning helium ion microscopy), among other effects, the accumulation of charge on the sample can cause so-called "drift," which is a gradual shift of the field of view (FOV) relative to the imaged object / sample. Electrostatic discharge fields above the sample surface can cause localized distortion.

[0003] Generally, charge accumulation primarily occurs when a primary electron beam interacts with insulating or poorly conductive areas of the sample after passing through the beam-generating device (the "irradiation side" of the SEM, not the "detection side"). When a primary electron strikes the sample, the sample emits secondary electrons, resulting in a net accumulation of, for example, positive charge in the irradiated area. If the sample is non-conductive or has low conductivity, this charge will not dissipate or will dissipate slowly. The resulting charge accumulation or buildup distorts the electric field in that area. This introduces artifacts such as brightness variations and image distortion, thereby degrading image quality.

[0004] Prior art mitigation techniques include depositing a conductive layer on the sample or using a low vacuum within the SEM chamber. These prior techniques face certain limitations. For example, coating the sample with a conductive layer affects the sample integrity, which is equivalent to invasive imaging. This is not always feasible, for example, in in-line testing of true semiconductor structures. Using a low vacuum environment will degrade image quality.

[0005] In recent years, multi-beam scanning electron microscopy (MSEMs) has been used to provide large-scale composite images of samples. Here, multiple images of multiple fields of view (FOVs) are simultaneously acquired, each FOV defined by a pair of patterned primary and secondary electron beams, thereby defining the composite FOV. The resulting relative images are then stitched together to form a composite image.

[0006] It has been observed that charge accumulation can cause severe distortion in composite images. For example, different fields of view (FOVs) may be affected differently by charge accumulation, resulting in FOV shifts. Consequently, artifacts appear in composite images. [Summary of the Invention]

[0007] Therefore, it is necessary to reduce the effects of charge accumulation on samples imaged using multi-beam scanning charged particle microscopy (such as MSEM).

[0008] The features of the independent request item satisfy this requirement. The features of the subsidiary request item define several embodiments.

[0009] Hereinafter, several techniques are used to determine the charge accumulation estimate of a sample irradiated by multiple charged particle beams in a multi-beam scanning charged particle microscope (such as MSEM).

[0010] According to the present invention, one or more measurements can be performed at the detector side of a multi-beam scanning charged particle microscope, and an estimate of charge accumulation can be inferred / determined based on such detector-side measurements. For example, by considering the variation of the spot characteristics of the secondary charged particle beams on an image of the relative pattern of the secondary charged particle beams, a spatial dependence estimate of charge accumulation can be determined, that is, the charge accumulation at multiple locations on the sample can be estimated. In other cases, a "global" estimate of charge accumulation can be determined, for example, estimating the average charge accumulation over the entire sample.

[0011] According to the example, once the estimate of charge accumulation is determined, the sample can be discharged. This discharge is also called charge compensation. For example, for positive charge accumulation, a mirror-mode discharge can be performed; the duration of this mirror-mode discharge can be estimated to optimize the yield. Depending on the surface charge (i.e., the amount of charge accumulation), the irradiation time of a single electron with an impact energy of "0 eV" (so-called mirror mode) may vary. This discharge time is preferably set as short as possible to avoid a significant decrease in sample yield. This mirror-mode irradiation time can be determined based on the amount of charge accumulation.

[0012] In the alternative or addition of such discharge, the aberrations of the primary charged particle beam can be determined based on an estimate of charge accumulation. Charge accumulation generates an electric field; charged particles (e.g., electrons) are deflected by this electric field, resulting in aberrations of the primary charged particle beam.

[0013] This could be different types of aberrations. For example, aberrations can be beam-specific. This means that one of the multiple primary charged particle beams exhibits another beam-specific aberration; another of the multiple primary charged particle beams exhibits another beam-specific aberration. Other types of aberrations are those related to the pattern formed by the multiple primary charged particle beams. The size / magnification of the pattern can be changed. The pattern can be rotated. The pattern can be tilted.

[0014] Next, some examples of beam-specific aberrations are provided. For example, it is possible to estimate the deflection of each of a plurality of primary charged particle beams. The deflection of the plurality of primary charged particle beams corresponds to one or more of these primary charged particle beams that deviate from their target spot position in the sample plane.

[0015] Another example of beam-specific aberration is spot distortion of the spot shape of a particular one of a plurality of primary charged particle beams. The spot shape of each of these plurality of primary charged particle beams may be affected. For example, an ideal spot shape may be circular. However, the spot shape may also be distorted, for example, becoming elliptical.

[0016] Another example of a specific aberration is defocus. A focal position is defined along each of a plurality of primary charged particle beams, where the focal point size (i.e., the cross-sectional area of ​​the beam perpendicular to the direction of charged particle propagation) is minimized. Sometimes, this focal position may be offset relative to a nominal position, such as an offset relative to the sample plane, resulting in defocus.

[0017] As mentioned above, in some cases, estimates of beam-specific aberrations can be determined for each of a plurality of primary charged particle beams. This is based on the finding that the intensity of beam-specific aberrations (e.g., defocus or deflection) may vary with the position of the primary charged particle beams in the relative pattern, or in other words, with the spatial relationship to the accumulation of charge on the sample.

[0018] Based on the knowledge of at least one aberration of multiple primary charged particle beams, countermeasures can be taken. For example, physical compensation for at least one aberration can be implemented. This may include applying a magnetic field and / or an electric field to the pattern of the primary charged particle beams; thereby, a common baseline of at least one aberration of the primary charged particle beams can be compensated. The electric or magnetic field may also present a gradient on the pattern of the primary charged particle beams. thereby, for example, compensation for stronger deflection or defocus at one side of the pattern can be made. Sometimes, a multi-beam deflection unit capable of deflecting individual beams among multiple primary charged particle beams can be used, for example, in a porous plate near the charged particle source on the irradiation side of a multi-beam scanning charged particle microscope; then, by controlling the multi-beam deflection unit, per-beam compensation can be performed for any beam-specific aberration estimated for each of the multiple primary charged particle beams. In some cases, as an alternative to or addition to this physical compensation, retrospective compensation is applied, i.e., after image acquisition of the sample image is completed. Here, a transformation can be applied to the sample image. This transformation can be determined based on at least one aberration estimated for multiple single-stage charged particle beams.

[0019] A method for operating a multi-beam scanning charged particle microscope is disclosed. The method may include loading a sample into an imaging position of the multi-beam scanning charged particle microscope. The method also includes irradiating the sample with multiple primary charged particle beams, for example, when the sample is in the imaging position. The method may also include measuring at least one characteristic of a secondary charged particle beam. The secondary charged particle beam is associated with multiple primary charged particle beams. The method also includes determining an estimate of charge accumulation in the sample based on at least one characteristic of the secondary charged particle beam. The method further includes determining an estimate of at least one aberration of the multiple primary charged particle beams. The at least one aberration is caused by charge accumulation in the sample. The method further includes performing compensation for the at least one aberration based on the estimate of the at least one aberration of the multiple primary charged particle beams.

[0020] A program code executed by at least one processor includes a plurality of control instructions for operating a multi-beam scanning charged particle microscope. When the program code is loaded and executed by the at least one processor, the at least one processor performs a method. The method may include loading a sample into an imaging position of the multi-beam scanning charged particle microscope. The method also includes irradiating the sample with a plurality of primary charged particle beams, for example, when the sample is in the imaging position. The method may also include measuring at least one characteristic of a secondary charged particle beam. The secondary charged particle beam is associated with a plurality of primary charged particle beams. The method also includes determining an estimate of charge accumulation in the sample based on at least one characteristic of the secondary charged particle beam. The method further includes determining an estimate of at least one aberration of the plurality of primary charged particle beams. The at least one aberration is caused by charge accumulation in the sample. The method further includes performing compensation for at least one aberration based on the estimate of at least one aberration of the plurality of primary charged particle beams.

[0021] A circuit for controlling the operation of a multi-beam scanning charged particle microscope is configured to control the multi-beam scanning charged particle microscope to load a sample into an imaging position of the microscope and to control the microscope to irradiate the sample when the sample is in the imaging position. The irradiation is performed using a plurality of primary charged particle beams. The circuit is further configured to measure at least one characteristic of a secondary charged particle beam associated with the plurality of primary charged particle beams and to determine an estimate of charge accumulation in the sample based on the at least one characteristic of the secondary charged particle beams. The circuit is further configured to determine an estimate of at least one aberration of the plurality of primary charged particle beams caused by charge accumulation in the sample and to compensate for the at least one aberration based on the estimate of the at least one aberration.

[0022] It should be understood that the foregoing features and the features to be explained below can be used not only in the indicated relative combinations, but also in other combinations or individually, without departing from the scope of the invention.

Implementation Method

[0037] Some embodiments of the present invention are generally provided for use in a plurality of circuits or other electronic devices. All references to circuits and other electronic devices and the functions they provide are not intended to be limited to what is illustrated and described herein. While specific labels may be assigned to the various circuits or other electronic devices disclosed, such labels are not intended to limit the scope of operation of the circuits and other electronic devices. Such circuits and other electronic devices may be combined and / or separated in any way based on a particular type of electronic implementation required. It should be understood that any circuit or other electronic device disclosed herein may include any number of microcontrollers, a graphics processing unit (GPU), integrated circuits, memory devices (e.g., flash memory, random access memory (RAM), read-only memory (ROM), electronically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or other software that cooperates with each other to perform the operations disclosed herein. Furthermore, any one or more of the electronic devices may be configured to execute program code, which may be embodied in a non-transitory computer-readable medium, programmable to perform any number of functions as disclosed.

[0038] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be understood that the following description of the embodiments is not intended to limit the present invention. The scope of the present invention is not intended to be limited to the embodiments or drawings described below, which are merely illustrative.

[0039] The drawings should be considered schematic diagrams, and the elements shown in the drawings are not necessarily drawn to scale. Rather, the various elements are presented so that those skilled in the art can clearly understand their function and general purpose. Any connection or coupling between functional blocks, devices, components, or other entities or functional units shown in the drawings or described herein may also be implemented through indirect connection or coupling. Coupling between components may also be established through wireless connection. Functional blocks may be implemented using hardware, firmware, software, or a combination thereof.

[0040] The various techniques disclosed herein relate to imaging samples using a multi-beam charged particle scanning microscope. The following will disclose techniques for operating a multi-beam scanning charged particle microscope. Specifically, we will explain the techniques for MSEM, but other types of multi-beam scanning charged particle microscopes may also be used.

[0041] MSEM combines scanning of the sample with multiple primary electron beams. Corresponding secondary electron beams are detected. Multiple images are formed, one image for each pair of primary-secondary electron beams. Each image has a corresponding field of view (FOV). A stitching operation is then used to determine a composite image. The composite FOV corresponds to the aggregation of multiple FOVs. The composite FOV of the composite image captures the entire irradiated area of ​​the sample. Irradiating the sample with electron beams can lead to charge accumulation.

[0042] Based on various examples, the charge accumulation of a sample can be estimated. For example, the charge accumulation in the irradiated area of ​​the sample can be estimated by scanning multiple primary charged particle beams on the sample. Based on the relative estimate of the sample charge accumulation, an estimate of at least one aberration of the multiple primary charged particle beams generated by the sample charge accumulation can then be determined. In other words, the aberrations of the secondary charged particle beams are taken into account, as well as the effect of the sample charge accumulation on the primary charged particle beams.

[0043] An estimate of at least one specific aberration of each of a plurality of primary charged particle beams may be determined; alternatively or additionally, an estimate of pattern-specific aberrations relating to the pattern formed by the plurality of primary charged particle beams may be determined, for example in a reference plane (such as a sample plane).

[0044] In other words, at least one aberration can be estimated for each of the plurality of primary charged particle beams. For example, a charged particle beam located near the edge of a pattern of the plurality of primary charged particle beams may be located closer to the edge of the irradiated area where charge accumulation occurs (outside the irradiated area, charge accumulation can be estimated to be zero). Therefore, due to the edge effect, different distortions or deflections may occur compared to primary charged particle beams located near the center of the pattern or at the center of the pattern.

[0045] Besides these edge effects, there may be other reasons for the different values ​​of a particular aberration observed in different individuals within a primary charged particle beam: in some cases, the estimate of charge accumulation may depend on the spatial location of the sample / on the primary and secondary electron beam patterns. Different regions of the sample may exhibit different degrees of charge accumulation. The reason may lie in the structure of the sample. For example, semiconductor samples may employ local metallization and local isolation, resulting in different intensities of charge accumulation.

[0046] Based on the estimation of at least one aberration, at least one aberration can be compensated. Typically, such compensation for at least one aberration can be physical compensation, for example, performed during sample imaging. One or more beam control elements of a multi-beam charged particle microscope can be controlled to deflect multiple primary charged particle beams, thereby canceling at least one aberration of the multiple primary charged particle beams.

[0047] In addition to or as an alternative to this physical compensation, digital compensation can also be employed. Performing compensation may include digital post-processing of images of the sample obtained using a multi-beam scanning charged particle microscope. Based on an estimate of at least one aberration of multiple primary charged particle beams, the misalignment of each pixel or group of pixels in the acquired image can be estimated. The misalignment of each pixel or group of pixels can then be corrected. The image remains undistorted.

[0048] Generally, the level of detail in determining the charge accumulation estimate can vary according to various embodiments of the invention. For example, in some embodiments, the charge accumulation estimate can be determined without any spatial or temporal dependence. That is, it can be assumed that the charge accumulation remains constant in a certain region of the sample, for example, in the irradiation region of the sample struck by multiple primary charged particle beams. On the other hand, in other cases, the charge accumulation estimate can be determined as a function of position on the sample and / or time. For example, in order to determine the charge accumulation estimate as a function of position (e.g., in the irradiation region of the sample scanned by multiple charged particle beams, determining spatial dependence), the individual characteristics of each of the multiple secondary charged particle beams can be considered. For example, the spot size, spot offset, spot shape, or spot brightness of the individual beam spots of the individual secondary charged particle beams can be considered to determine the estimate of local charge accumulation. To obtain temporal dependence, the temporal evolution of one or more characteristics of the secondary charged particle beams can be considered. For example, the magnification of the pattern of the multiple secondary charged particle beams can vary as a function of time, for example, until a steady state is reached. This time dependence indicates the time dependence of charge accumulation, which gradually increases until it plateaus.

[0049] The level of detail in determining the estimate of at least one aberration corresponds to the level of detail in determining the estimate of charge accumulation. In other words, if charge accumulation is accompanied by a time-resolution estimate, then similarly, at least one aberration can also be accompanied by a time-resolution estimate.

[0050] Figure 1 is a schematic diagram of MSEM 1. Further information on this type of MSEM and the components used therein, such as, for example, particle sources, porous plates and lenses, can be obtained from international patent applications WO 2005 / 024881, WO 2007 / 028595, WO 2007 / 028596, WO 2011 / 124352 and WO 2007 / 060017 and German patent applications DE 10 2013 016 113 A1 and DE 10 2013 014 976 A1, the scope of which is incorporated herein by reference in its entirety.

[0051] MESM1 uses a plurality of charged particle beams (also known as small beams or simply beams) to image sample 7. MSEM 1 generates a plurality of J primary electron beams 3.1, 3.2, 3.3, which bombard sample 7 to produce interaction products, such as secondary electrons, which are emitted from sample 7 to form secondary electron beams 9.1, 9.2, 9.3, and are subsequently detected.

[0052] The primary and secondary electron beams 3.1, 3.2, 3.3, 9.1, 9.2, and 9.3 are each formed and guided by a relative imaging subsystem of MSEM 1. Each imaging subsystem incorporates a relative field of view (FOV). The images captured by a relative imaging subsystem depict the relative field of view. Multiple FOVs are configured into a spatial pattern to define a composite FOV.

[0053] The primary electron beams 3.1, 3.2, and 3.3 are formed by electrons incident on the surface of the sample 7 at multiple locations, and produce multiple primary electron beam spots 5.1, 5.2, and 5.3 that are spatially separated from each other.

[0054] The sample 7 to be inspected can be any desired type, such as a semiconductor wafer or a semiconductor photomask, and can contain a configuration of miniaturized components.

[0055] The surface of sample 7 is disposed in a sample plane 101 of the objective lens system 102 of a first particle optical unit (also known as an illumination system) 100.

[0056] The diameters of the smallest beam spots or focal spots 5.1, 5.2, 5.3 formed in the sample plane 101 can be very small. An exemplary value for this diameter is less than four nanometers, for example, three nanometers or less. The objective system 102 is used to focus the primary electron beams 3.1, 3.2, 3.3 to form the focal spots 5.1, 5.2, 5.3. In this case, the objective system 102 can include a magnetically immersed lens. Further examples of focusing components have been described in German Patent DE 102020125534 B3, the entire contents of which are incorporated herein by reference.

[0057] The number J of primary electron beams 3.1, 3.2 and 3.3 (i.e. the number of FOVs) can be 5, 25, 90 to 100 or more (for simplicity, only three primary electron beams 3.1, 3.2 and 3.3 and their corresponding focal points 5.1, 5.2 and 5.3 are shown in Figure 1).

[0058] In practice, the number of electron beams J (and thus the number of FOVs) can be chosen to be significantly large, such as J = 10 x 10, J = 20 x 30, or J = 100 x 100. Example values ​​for the spacing between the incident position and the FOV are 1 micrometer, 10 micrometers, or larger, such as 40 micrometers.

[0059] The number of primary and secondary electron beams J defines the number of FOVs. Each imaging subsystem has a relative FOV. The relative FOV is defined by the relatively paired primary and secondary electron beams (e.g., beams 3.1 and 9.1) scanning the sample 7 in the relative FOV.

[0060] The primary electron beams 3.1, 3.2, and 3.3 impacting sample 7 produce interaction products, such as secondary electrons emitted from the surface of sample 7, backscattered electrons, or primary particles that have undergone reverse motion for other reasons. Depending on the charge dissipation characteristics of sample 7, charge accumulation may occur. The interaction products emitted from the surface of sample 7 are shaped by objective lens system 102 to form secondary electron beams 9.1, 9.2, and 9.3. The secondary electrons included in secondary electron beams 9.1, 9.2, and 9.3 are used for imaging.

[0061] MSEM 1 provides a detection beam path on the detection side for guiding a plurality of secondary electron beams 9.1, 9.2, 9.3 to a secondary electron imaging system 200. The secondary electron imaging system 200 includes a plurality of electron optical lenses 205.1 to 205.5 for guiding the secondary electron beams 9.1, 9.2, 9.3 to a spatial resolution detector system 600.

[0062] Imaging using the secondary electron imaging system 200 achieves extremely high magnification, resulting in particularly magnified imaging of the pattern of the primary electron beam on the wafer surface, as well as the size and shape of the primary electron beam focal point. For example, a scaling factor / magnification is between 100x and 300x, resulting in nanometer (nm) structures on the wafer surface being magnified to between 100 nm and 300 nm. In one example, the image field of a multi-beam device with a diameter of, for example, 100 μm is magnified to approximately 30 mm.

[0063] Primary electron beams 3.1, 3.2, and 3.3 are generated in a beam generating apparatus 300 at the irradiation side. This beam generating apparatus includes at least one particle source 301 (e.g., an electron source), at least one collimating lens 303, a porous configuration 305, a first field lens 331, and a second field lens 333. The particle source 301 generates at least one diverging particle beam 309, which is at least substantially collimated by the at least one collimating lens 303 and irradiates the porous configuration 305. The porous configuration 305 includes a porous plate (MAP) 304 (also called a filter plate or multi-hole plate) with a plurality of J openings formed therein in a first grating configuration. Particles irradiating the particle beam 309 pass through the J holes or openings of the MAP 304, forming a plurality of J primary electron beams 3.1, 3.2, and 3.3. The particles of the irradiating particle beam 309 impacting the first aperture plate 304 are absorbed by the first aperture plate and do not form primary electron beams 3.1, 3.2, 3.3. The multi-aperture configuration 305 sometimes has at least one further MAP 306, which includes beam deflection components, such as a lens array, an astigmatism array, or an array of deflecting elements. Such beam deflection components can individually deflect each of the multiple primary electron beams 3.1, 3.2, 3.3.

[0064] Together with the field lens 331 and a second field lens 333, the aperture configuration 305 focuses each of a plurality of primary electron beams 3.1, 3.2, 3.3, wherein a plurality of focal points are formed in an intermediate image surface 321. Alternatively, the beam focal points and the intermediate image surface 321 may be virtual. The intermediate image surface 321 can be bent to pre-compensate the image field curvature of the imaging system configured downstream of the intermediate image surface 321.

[0065] At least one field lens 103 and objective lens system 102 provide a first imaging particle optics unit for imaging the surface 321 forming the beam focal point onto the sample plane 101, such that a second pattern of the focal spots 5.1, 5.2, 5.3 of the primary electron beams 3.1, 3.2, 3.3 is formed therein. Typically, the surface 25 of the sample 7 is disposed in the sample plane 101, and the focal spots 5.1, 5.2, 5.3 are correspondingly formed on the surface 25. The plurality of primary electron beams 3.1, 3.2, 3.3 form an intersection point 108, and a first deflection scanner 110 is disposed near this intersection point. The first deflection scanner 110 is used to simultaneously and completely deflect the plurality of primary electron beams 3.1, 3.2, 3.3, such that the plurality of focal spots 5.1, 5.2, 5.3 simultaneously and jointly scan the surface 25 of the sample 7. A grating scanning system is implemented to image the sample 7 thereby. The first deflection scanner 110 is driven by a scanning control unit 860 to acquire a plurality of two-dimensional image data of the surface in a detection operation mode. Additionally, the MSEM 1 may include a plurality of further electrostatic deflectors configured to adjust the positions of the plurality of primary electron beams 3.1, 3.2, 3.3.

[0066] The objective lens system 102 and the projection lens 205 provide a secondary electronic imaging system 200 for imaging the sample plane 101 onto the imaging plane 225. The objective lens system 102 is therefore a lens or lens system that is part of both the first and second particle optical units, while the field lenses 103, 331, and 333 belong only to the first particle optical unit 100, and the projection lens 205 belongs only to the secondary electronic imaging system 200.

[0067] A beam splitter 400 is disposed in the beam path of the first particle optical unit 100 between the field lens 103 and the objective lens system 102. The beam splitter 400 is also part of a second optical unit in the beam path between the objective lens system 102 and the projection lens 205.

[0068] The first deflector scanner 110 is configured in either the primary electron beam path or a combined electron beam path. In the example shown in FIG1, the secondary electron sub-beams 9.1, 9.2, and 9.3 output the first deflector scanner 110 in the opposite direction during use, partially compensating for the scanning motion of the secondary electron beams 9.1, 9.2, and 9.3. Secondary electrons typically have different kinetic energies than primary electrons. Therefore, only partial compensation is made for the scanning motion of the moving irradiation position. To fully compensate for the scanning motion of the secondary electron beams 9.1, 9.2, and 9.3, a combined beam deflector 222 is configured in the secondary electron beam path.

[0069] The secondary electron imaging system 200 includes a second collective beam deflector 222 disposed near an intersection of the secondary electron beams 9.1, 9.2, and 9.3. The second collective beam deflector 222 is operationally synchronized with the first beam deflector 110 and compensates for beam deflection of the secondary electron beams 9.1, 9.2, and 9.3 during use, such that the beam spot 15 of the secondary electron beam 9 remains at a constant position on the imaging plane 225. In this way, each secondary electron beam 9 is held within a region of a set of detection elements, which is allocated to the individual secondary electron beam 9.

[0070] The secondary electron imaging system 200 includes electron optical lenses 205.1 to 205.5 to adjust the focal plane of the secondary electron beams 9.1, 9.2, and 9.3. Defocusing can be applied. Electron optical lenses 205.1 to 205.5 can therefore implement correction elements for correcting the focal plane. Electron optical lenses 205.1 to 205.5 are shown as magneto-optical elements, but are not limited to magneto-optical elements, and can also include electrostatic lens elements or astigmatism correctors. Using electron optical lenses 205.1 to 205.5, the secondary electron beams 9.1, 9.2, and 9.3 can be focused into the imaging plane 225 of the secondary electron imaging system 200.

[0071] The secondary electron imaging system 200 may include a plurality of additional correction elements, such as at least one of a multi-aperture array element, a deflector, or a replaceable aperture stop. The lens-objective system 102 can be used to focus the secondary electron beams 9.1, 9.2, 9.3 onto the spatial resolution detector system 600, and in the process, can correct or compensate for magnification and rotation of the pattern of the secondary electron beams 9.1, 9.2, 9.3 on the imaging plane 225. This allows for the stabilization of the pattern of the plurality of secondary electron beams 9.1, 9.2, 9.3. For example, a first and second magnetic lenses 205.4 and 205.5 (as a further example of correction elements) are designed in opposite order and have magnetic fields in opposite directions. By appropriately applying control signals to (driving) the magnetic lenses 205.4 and 205.5, Larmor rotation of the secondary electron beams 9.1, 9.2, 9.3 can be compensated. In the example shown, the secondary electron imaging system 200 includes additional correction elements, specifically in the porous plate 216.

[0072] Furthermore, MSEM 1 incorporates a processing device 800 configured to control individual particle optics components of multiple particle beam systems and to evaluate and analyze signals acquired by detector system 600. Processing device 800 can be separate from MSEM 1 or integrated into MSEM 1. For example, processing device 800 can be configured to acquire multiple pairs of test images and then estimate these test images to determine values ​​of one or more imaging parameters. In this case, control or processing device 800 can be composed of multiple individual computers or electrical components. For example, processing device 800 includes a control processor 880; a control module 840 for controlling the electro-optical elements of secondary electron imaging system 200; and a control module 830 for controlling the electro-optical elements of primary electron beam generating unit. Processing device 800 is further connected to a control module 503 for supplying voltage to sample 7, also referred to as lead-out voltage. This creates a lead-out field between objective system 102 and surface 25 of sample 7 during use. During use, the extraction field decelerates the primary electrons of the primary electron beams 3.1, 3.2, and 3.3 before they reach the objective lens surface 25, and produces an additional focusing effect on the plurality of primary electron beams 3.1, 3.2, and 3.3. Simultaneously, the extraction field accelerates secondary particles away from the surface 25 of the sample 7 during use. Furthermore, the processing apparatus 800 includes a scanning control unit 860 for raster scanning.

[0073] The detector system 600 includes a plurality of sets of detection elements, wherein each secondary electron beam 9 has a set of detection elements for providing a highly magnified image for each FOV. During use, each set of detection elements is configured to record the intensity signal of the assigned secondary electron beam 9. The plurality of intensity signals of the plurality of secondary electron beams 9.1, 9.2, 9.3 are transmitted to the image data acquisition unit 810, where the image data is processed and stored in memory 890. Thus, multiple images are acquired, one image corresponding to each imaging subsystem. These multiple images (or aggregated images determined based on a relative sequence of images) can be combined into a composite image with a composite FOV.

[0074] As mentioned above, MSEM 1 includes components for generating a plurality of primary electron beams 3.1, 3.2, 3.3, which are configured into a first pattern. This first pattern is defined by a plurality of holes in a perforated plate 304 (also referred to as a porous plate). Figure 2 illustrates an example of the first pattern 41.1.

[0075] Figure 2 shows a first aperture plate 304, in which a plurality of apertures 85 form a first pattern 41.1. In this example, the first pattern 41.1 is a hexagonal grating with a grating spacing p1 of, for example, 100 µm. The first pattern 41.1 defines a reference for the pattern of a plurality of primary electron beams 3.1, 3.2, 3.3 in the sample plane.

[0076] Figure 3 shows the origin of the secondary electron beams 9.1, 9.2, 9.3 formed by the focal spots 5.1, 5.2, 5.3 of the primary electron beams 3.1, 3.2, 3.3 (see Figure 1). When the primary electron beams 3.1, 3.2, 3.3 are used to irradiate each location on the surface 25 of a sample 7, secondary electrons are generated, forming a plurality of secondary electron beams 9.1, 9.2, 9.3. Therefore, the endpoints of the primary electron beams 3.1, 3.2, 3.3 and the origin of the plurality of secondary electron beams 9.1, 9.2, 9.3 form a second pattern 41.2 influenced by the first pattern 41.1 of the primary electron beams 3.1, 3.2, 3.3. The second pattern 41.2 may be affected by the aberrations of the primary electron beams 3.1, 3.2, 3.3.

[0077] Figure 4 shows spots 15 of the secondary electron beams 9.1, 9.2, and 9.3 on the imaging plane 225. During use, spots 15 of the secondary electron beams 9.1, 9.2, and 9.3 form a third pattern 41.3. The third spacing 1061 is illustrated in the figure, where p3 = 1000 μm.

[0078] In the absence of stability, the third pattern 41.3 may differ from the first pattern 41.1 and the second pattern 41.2. For example, the third pattern may be distorted, translated, rotated, tilted, and / or magnified. Furthermore, defocusing may occur. This will change the size (i.e., width) of the secondary electron beams 9.1, 9.2, and 9.3 in the imaging plane. Such deviations can degrade image quality or even lead to complete signal loss. Furthermore, the second pattern may also differ from the first pattern 41.1. For example, the second pattern may be distorted, translated, rotated, tilted, and / or magnified. In other words, the primary electron beams 3.1, 3.2, and 3.3 and the secondary electron beams 9.1, 9.2, and 9.3 may be distorted. This distortion typically originates from charge accumulation on the sample 7. Additionally, individual spots 15 may be brighter or darker than others. Some spots 15 may shift from the reference position defined by the first pattern 41.1. Some spots 15 may be larger than others. Furthermore, these characteristics of individual spots 15 may depend on charge accumulation on sample 7. Since charge accumulation may be spatially dependent, individual spots 15 may be affected differently.

[0079] Figure 5 schematically illustrates the effect of charge accumulation 1050 at sample 7. Figure 5 shows a cross-section of sample 7. Due to the extraction of secondary electrons, charge accumulation 1050 occurs within the irradiated region 1060. Charge accumulation 1050 results in at least one aberration of the secondary electron beams 9.1, 9.2, and 9.3. Charge accumulation 1050 also results in at least one aberration of the primary electron beams 3.1, 3.2, and 3.3.

[0080] As shown in Figure 5, at least one aberration of the primary electron beams 3.1, 3.2, and 3.3 may specifically include a misalignment / deflection of the scan position of each of the primary electron beams 3.1, 3.2, and 3.3 on the sample 7. That is, the focal points 5.1, 5.2, and 5.3 are shifted. This results in distortion of the captured image. For example, Figure 6 shows an exemplary image 2020, such as in the FOV of the primary electron beam 3.1. Due to the temporal dependence of charge accumulation, the distortion gradually increases from bottom to top. Referring upwards, Figure 7 illustrates the corresponding image 2021 without distortion. Charge accumulation generally depends on the spatial position on the sample 7 (as previously shown in Figure 2); charge accumulation also depends on time. As irradiation increases, charge accumulation also increases and generally stabilizes to an equilibrium value after a certain period of time.

[0081] Various techniques are based on the discovery that the estimate of charge accumulation 1050 can be determined based on one or more characteristics of the secondary electron beams 9.1, 9.2, and 9.3 measured at the MSEM 1 detector side. Specifically, multi-pixel images of the spots of the secondary electron beams 9.1, 9.2, and 9.3 on the imaging plane can be acquired. For this purpose, an auxiliary monitoring system for monitoring one or more characteristics of the secondary electron beams 9.1, 9.2, and 9.3 can be used in the MSEM 1 detector system 600. This is shown in Figure 8.

[0082] Figure 8 illustrates an example of the detector system 600 of MSEM 1. Detector 600 includes an electron-to-light conversion element 602 disposed in an imaging plane 225. The electron-to-light conversion element 602 is configured to convert secondary electrons from secondary electron beams 9.1, 9.2, and 9.3 into light. The detector further includes an optical relay system having multiple optical elements 605 and 611 for imaging and guiding the excitation light from the electron-to-light conversion element 602 to a detector element 623. The signal from the detector element 623 is used to determine the pixel value of the image pixel acquired by each FOV, for ultimately determining the microscopic, high-magnification composite image of sample 7. For this purpose, the optical relay system may include a zoom lens system 611, multiple mirrors 607, multiple rotating prisms (not shown), and multiple optical fibers 615. In the example of Figure 8, detector 600 is configured to image excitation light from electron-to-light conversion element 602 onto image plane 613 of primary electron beam detector 612, wherein a plurality of entrances 613 of optical fiber 615 are configured. Each entrance 613 incorporates a secondary electron beam.

[0083] Figure 9 shows a fourth pattern 41.4 of these inlets 613. The fourth pattern 41.4 is defined by the configuration of the inlets 613 of the optical fiber 615 and by the magnification of the optical system containing the lens 605 and the zoom lens 611. If the third pattern 41.3 changes over time (e.g., due to charge accumulation), the secondary electron beam may drift between the inlets 613; this degrades image quality. This is similar to inter-beam crosstalk.

[0084] To mitigate the effects of charge accumulation, the detector system 600 further includes a monitoring system 230 having a multi-pixel detector 232 containing multiple pixels 626; the monitoring system 230 also includes an optical relay lens 235. The monitoring system 230 is coupled to a beam splitter 237. The multi-pixel detector 232 typically operates at a low frame rate, for example, 0.1 to 1 kHz, and therefore cannot collect intensity signals at raster scan rates of approximately 20 MHz to 80 MHz. The multi-pixel detector 232 acquires multi-pixel images of secondary electron beams 9.1, 9.2, and 9.3. The third pattern 41.3 can be measured relatively quickly. The measurement bandwidth of the multi-pixel detector 232 is higher than that of the detector element 623; however, the sensitivity is generally lower. The speckle images of secondary electron beams 9.1, 9.2, and 9.3 acquired using the multi-pixel detector 232 are still sufficient to accurately determine the estimate of charge accumulation 1050.

[0085] Figure 10A is a flowchart of a method according to various examples. Figure 10A is a flowchart of a method for operating a multibeam scanning electron microscope. The method of Figure 10A can be executed by a processor, loading program code from memory and executing the program code. For example, the method of Figure 10A can be executed by the processing device 800 of MSEM 1 shown in Figure 1. The method of Figure 10A will be explained with reference to MSEM 1; however, it can also be executed by other types of multibeam scanning electron microscopes.

[0086] In step 3005, the sample (see Figure 1: Sample 7) is loaded into the MSEM imaging position. The imaging position is similar to the positioning of the sample in the vacuum chamber of the MSEM, where an imaging mode can be performed, i.e., an image of the sample can be obtained at this position. Sometimes, a sample stage can move between multiple positions in the vacuum chamber of the MSEM, such as the imaging position and the calibration position. At the calibration position, imaging may not be possible. On the other hand, the calibration sample can then be positioned so that multiple primary charged electron beams can strike the calibration sample, while the sample / sample stage under review is deviated from the imaging position.

[0087] Step 3005 may include placing the sample in the load lock of the vacuum chamber. Then, the load lock is evacuated, and the sample can be moved by a motorized platform (see Figure 1: sample carrier 500) to the imaging position for primary electron beam irradiation.

[0088] For example, the sample may be a semiconductor sample, such as a wafer, die, or piezoelectric chip carrying a semiconductor structure. For example, the semiconductor structure may be a logic gate, a power line, a memory cell, etc., to name just a few examples.

[0089] In step 3010, the sample is then irradiated at the imaging location using multiple primary electron beams. Step 3010 may include grating scanning of the sample with multiple primary electron beams using a collective deflection scanner (see Figure 1, deflection scanner 110). Due to the irradiation of the sample, charge accumulation may occur.

[0090] Step 3010 may include imaging the sample. This includes performing a single electron beam grating scan of the sample using a single detection mode with high sensitivity but relatively small measurement bandwidth (e.g., detection element 623 in the example detector system 600 illustrated in FIG. 10A), and obtaining relative intensity values ​​for each scan location. Based on one or more images in each field of view, a composite image is stitched together. After obtaining the composite image, the sample stage may be moved to obtain further composite images of other locations on the sample.

[0091] Step 3010 may include acquiring a sequence of multiple images for each field of view and combining the multiple images for each field of view into an aggregated image. Frame aggregation can be performed.

[0092] In all cases, it is not necessary to image the sample to irradiate it. For example, a calibration procedure may be performed before imaging the sample to determine the charge accumulation on the sample. In this case, a smaller exposure dose can be used in step 3010 compared to the previous condition of the imaged sample.

[0093] Next, in step 3015, one or more characteristics of the secondary electron beam relating to the primary electron beam in step 3010 are measured. This measurement is performed on the detector side of the MSEM.

[0094] Generally speaking, one or more characteristics of the secondary electron beam may include one or more characteristics of individual spots of the secondary electron beam and / or may include one or more characteristics of the secondary electron beam pattern, see Figure 4: Pattern 41.3.

[0095] The measurements at step 3015 may be based on images or image sequences obtained by an auxiliary multi-pixel detector (such as multi-pixel detector 232 discussed in conjunction with FIG10A).

[0096] Example characteristics include the magnification of pattern 41.3 of the secondary electron beam. That is, the scaling factor of pattern 41.3 can be determined relative to a reference pattern. It has been found that this magnification of pattern 41.3 indicates global charge accumulation over the irradiated area. Local variations in charge accumulation, i.e., the spatial dependence of charge accumulation within the irradiated area, may be more difficult to determine based on this magnification of pattern 41.3. The magnification / scaling factor of pattern 41.3 may be related to the average beam spacing. The greater the magnification, the larger the observed average beam spacing.

[0097] Alternatively or additionally, the beam spacing variation of each of the plurality of secondary electron beams can be determined (see Figure 4: Spacing 1061). This beam spacing variation means that the distance between adjacent spots 15 deviates from a reference value. For this purpose, an image acquired using a multi-pixel detector 232 can be analyzed. For example, this image analysis can be based on peak detection, which is used to identify bright spots in the multi-pixel image. A predefined reference pattern can be fitted, and the deviation of the spot position can be determined based on the fitting result. Alternatively or additionally, the line spacing can be determined (see Figure 4: Line Spacing 1062). As mentioned above, such spacing variation indicates local variations in charge accumulation in these regions.

[0098] As mentioned above, image domain analysis using the images acquired by the multi-pixel detector 232 can reveal these characteristics of individual secondary electron beams. For example, a machine learning model operating on individual images acquired for the secondary electron beam pattern can be employed. This machine learning model can provide a scalar output to quantify the magnification of the pattern 41.3. For example, such a machine learning model can be implemented as a deep neural network containing multiple convolutional layers.

[0099] Generally, in addition to or as an alternative to image domain analysis using the image acquired by the multi-pixel detector 232, Fourier domain analysis can also be performed on the image. That is, a Fourier transform can yield a Fourier domain representation of the image. The Fourier domain is sometimes also called the k-space or spatial frequency domain. Due to the geometry of pattern 41.3, this Fourier domain representation of the image is particularly useful for determining one or more features of pattern 41.3. For example, PCT / EP2024 / 051248 filed January 19, 2024, and German Patent Application No. 10 2023 101 358.0 filed January 19, 2023, disclose a fast technique for determining one or more features of pattern 41.3 based on Fourier domain sparse matrix multiplication.

[0100] In addition to one or more features of pattern 41.3, one or more characteristics of individual secondary electron beams may also be considered. For this purpose, each individual spot 15 may be analyzed. This includes, for example, the spot displacement of one or more secondary electron beams 9.1, 9.2, 9.3 (e.g., displacement relative to an absolute reference position, but not relative to its adjacent points), the spot size of one or more secondary electron beams 9.1, 9.2, 9.3, the spot shape of one or more secondary electron beams 9.1, 9.2, 9.3, and / or the beam brightness of one or more secondary electron beams 9.1, 9.2, 9.3. These are electron beam-specific features that can be used to identify local charge accumulation at the origin locations of secondary electron beams 9.1, 9.2, 9.3 in the sample.

[0101] Generally, this measurement can be performed using a time-resolved method. That is, when irradiating the sample, for example, when a primary electron beam scans the sample to obtain one or more images in each field of view, multiple multi-pixel images of the secondary electron beam can be obtained. For example, Figure 10B illustrates the temporal dependence of the magnification 7001 of pattern 41.3. The initial increase in magnification (and the resulting charge accumulation) and subsequent saturation can be clearly seen. However, the change in magnification depends on the charged sample. That is, the magnification may increase or decrease when different charged samples are used. Figure 10B further illustrates the offset 7002 of a specific beam spot (in the x and y directions) of one of the multiple secondary electron beams; and the beam spot size 7003 of the specific beam spot. In addition, Figure 10C illustrates the temporal dependence of the beam spacing 7004.

[0102] Note that although the increase in beam spacing 7004 is illustrated in Figure 10C (similar to the increase in magnification 7001 shown in Figure 10B), in other cases, a decrease in beam spacing 7004 and magnification 7001 may be observed depending on the polarity of the sample and / or the charged particles in the charged particle microscope.

[0103] These characteristic curves 7001, 7002, 7003, and 7004 can correspond to the charge accumulation of the sample. Since these characteristic curves 7001, 7002, 7003, and 7004 exhibit time dependence, the charge accumulation of the sample also exhibits time dependence.

[0104] Please refer back to Figure 10A: In step 3016, a prediction of at least one characteristic may be selectively determined; this is based on the measurement results of step 3015. This prediction may be related to one or more future time points. As described in steps 3015, Figures 10B and 10C, various characteristics of the secondary charged particle beam (e.g., the magnification of the secondary electron beam pattern related to the average beam spacing) may exhibit characteristic time dependencies. These time dependencies are reproducible because they depend on well-defined charging effects of the sample. Depending on the example, a machine learning model (predictive model) may be used to provide such predictions. The predictive model may take an input sequence and provide one or more outputs, for example, each output being related to a prediction of a relative characteristic relative to a future time point.

[0105] For example, a sequence-to-sequence prediction model can be used to predict the characteristics of secondary electron beams at multiple future time points.

[0106] The input to this prediction model may include a sequence of images depicting a secondary electron beam pattern. The output of this prediction model may include a sequence of beam spacing or magnification / scaling factors of the secondary electron beam pattern, at least some of which are related to future time points. Figure 10D illustrates the prediction of a scalar from images. Here, during a measurement interval 910, multiple images 905, 906, 907, and 908 are acquired, depicting the patterns of the secondary electron beam at relative measurement time points 911-914, respectively. These images 905, 906, 907, and 908 are jointly processed in the prediction model to provide a scalar value as output indicating a magnification factor 7001 for future time points 921-924, distributed across the prediction range 920. The output is provided at the current time point 930; a processing delay 931 is also illustrated.

[0107] In addition to this image-to-scalar implementation of the prediction model, a scalar-to-scalar implementation of the prediction model is also an option. For example, a sequence of beam spacings determined based on images acquired by a secondary electron beam can be input into the prediction model, and a machine learning model can provide a sequence of beam spacings determined for future time points. This is illustrated in Figure 10E, where images 905-908 are first processed to obtain a relative value of magnification 7001. Then, the time series of this equivalent value of magnification 7001 is processed in the prediction model (e.g., a regression model) to obtain predicted values ​​of magnification 7001 for future time points 921-924 within a prediction range 920. Compared to the scenario illustrated in Figure 10D, it is found that the processing delay 931 illustrated in Figure 10E may be smaller because the prediction model may be smaller, i.e., using fewer weights and operating faster.

[0108] Various architectures or prediction methods are possible, depending on the specific implementation of the prediction model. For example, several examples include linear regression, random forest, recurrent deep neural networks, etc., and only a few examples are given.

[0109] Regression can refer to a statistical method used to establish relationships between variables, such as the time-dependent changes in secondary electron beam characteristics and the relative charge accumulation on a sample. For example, linear or nonlinear regression techniques can be used to simulate how characteristics such as magnification, electron beam spacing, or spot displacement change over time due to the charging effect.

[0110] Random forest can refer to an ensemble learning method that combines multiple decision trees to improve prediction accuracy and robustness. In charge accumulation estimation, random forest models can analyze patterns in multi-pixel images obtained by a monitoring system, such as spatial variations in the brightness or displacement of secondary electron beam spots.

[0111] Recurrent deep neural networks can refer to a category of artificial neural networks that process sequential data, such as the temporal dependence of secondary electron beam characteristics. These networks can capture temporal patterns in the evolutionary process and predict future states based on historical data. For example, a recurrent deep neural network can simulate the magnification or beam spacing of a secondary electron beam pattern over time to predict aberrations caused by charge accumulation at future time points.

[0112] Depending on various examples, multiple frames / images can be obtained to build a training dataset. The data can include the current frame, the last frame, the first two frames, etc., up to the first m frames. This data acquisition can capture the temporal and spatial variations in the characteristics of the secondary electron beam, which can indicate charge accumulation. The resulting dataset can be divided into training samples and validation samples. Then, a prediction model containing a regression architecture can be fitted to the training samples, where the prediction model is based on historical frames to predict one or more characteristics of the secondary electron beam. Therefore, unsupervised learning is possible.

[0113] The performance of the prediction model can be evaluated on a validation sample to assess its prediction accuracy. The evaluation process may involve comparing predicted and actual values ​​of target characteristics (such as beam spot displacement, beam spacing, or magnification). Based on this evaluation, the performance of the prediction model can be optimized by adjusting parameters or selecting the most suitable prediction model.

[0114] This method may further include comparing the performance of multiple prediction models to determine the best performing model. This comparison may be based on metrics such as mean squared error, R-squared value, or other relevant performance indicators.

[0115] Please refer back to Figure 10A: Based on one or more of these (measured and / or predicted) characteristics of the secondary electron beam, an estimate of the charge accumulation of the sample can be determined in step 3020. The charge accumulation can be estimated for a specific time point for estimating one or more characteristics. If a prediction was performed in step 3016, these may be future time points.

[0116] Step 3020 is based on the fact that such characteristics of secondary electron beams, as described above, are usually primarily or solely derived from charge accumulation.

[0117] For example, an estimate of charge accumulation is determined based on a predefined lookup table that links one or more characteristic values ​​of the secondary electron beam to the charge accumulation. For example, a lookup table could link the magnification of the pattern 41.3 of the secondary electron beam to the charge accumulation within the irradiation area. In this case, the charge accumulation may not be considered to vary within the irradiation area. Such a lookup table is shown in Table 1. Scaling factor Charge accumulation (V) -1.2 -20V -1.1 -10V 1 0V 1.1 +10V 1.2 +20V Table 1: An exemplary lookup table that maps the magnification of the secondary electron beam pattern 41.3 in the imaging plane to the charge accumulation on the sample. The charge accumulation is assumed to be constant within the irradiated area.

[0118] Even if the charge accumulation is considered constant within the irradiation area, i.e., the estimate of charge accumulation is not spatially dependent, different primary and secondary electron beams will still be affected differently by this charge accumulation, depending on their positions within the irradiation area. For example, due to edge effects, primary and secondary electron beams positioned near the edges of the pattern may be affected differently than those positioned near the center of the pattern.

[0119] Such a lookup table as in Exemplary Table 1 can be determined during calibration or based on ray tracing simulations of a secondary electron beam with varying charge accumulation. If calibration is used, a sample with perfectly defined charge dissipation characteristics can be used as a calibration tool; for example, a metal print test structure is a phase-selective method. Alternatively, microelectromechanical structures (e.g., stand-alone structures) can be used because these MEMS structures have perfectly defined charge dissipation characteristics. MEMS structures can also be fabricated using capacitors for in-situ sensing of charge accumulation. For example, a known charge accumulation (e.g., using a secondary measurement technique) can be linked to different values ​​of the pattern's magnification.

[0120] It should be understood that the lookup table linking the magnification of pattern 41.3 with the charge accumulation estimate is a relatively simple technique that can quickly determine the charge accumulation estimate. Therefore, this method is particularly useful if rapid compensation is required (e.g., in-situ entity compensation performed in parallel with sample imaging). However, in some conceivable cases, the accuracy of the charge accumulation estimate is more important than speed. This situation may relate, for example, to retrospective offline compensation, i.e., performed after imaging is complete. In this case, while acquiring the data used for imaging, a multi-pixel image of the secondary electron beam can be acquired as auxiliary information, upon which retrospective offline compensation will later be performed. In such retrospective offline compensation, more sophisticated techniques can be employed to determine the charge accumulation estimate.

[0121] For example, charge accumulation can be estimated based on a predefined model. Such a predefined model can link the value of at least one feature of the secondary electron beam to charge accumulation. The model may include functional dependencies. These functional dependencies may be based on heuristic methods, such as ray tracing simulations. Alternatively, machine learning models, such as deep neural networks, can be used. Such machine learning models can accept images obtained by imaging a pattern of the secondary electron beam using a multi-pixel detector as input. Such machine learning models can output a charge accumulation map that quantifies the estimate of local charge accumulation in the composite field of view. Training data for training such machine learning models can be obtained by performing ray tracing simulations of multiple secondary electron beams passing through a multi-beam scanning electron microscope detector system under different charge accumulation conditions. Regardless of the specific implementation of the model, using such a model, regression inference can be performed to determine an estimate of charge accumulation in a continuous result space, rather than simply inferring discrete values ​​such as in a lookup table.

[0122] In some cases, the estimate of charge accumulation in a sample can even be determined using a multi-beam scanning electron microscope detector system based on ray tracing simulations of multiple secondary electron beams. The parameters of the ray tracing simulation can then be repeatedly changed until the pattern observed in the ray tracing simulation matches the measured image symbol of the pattern of multiple secondary electron beams. This approach tends to provide relatively accurate results; however, it also requires significant computational resources. This may limit its applicability in offline scenarios, where the estimate of charge accumulation is determined after sample imaging is complete.

[0123] As described above with respect to Figure 10B, one or more characteristics of the secondary electron beam can be analytically measured in the time domain. Therefore, in step 3020, time resolution can be selectively used to determine an estimate of the charge accumulation of the sample. For example, a model can be inferred once for each time sample of one or more characteristics of the secondary electron beam via multiple inferences.

[0124] In selective step 3025, the charge accumulation of the sample may conform to the structural types of multiple structures of the sample. For example, the sample may include an array of structures, where certain structures of a particular type reappear at different locations on the sample. A typical example is a semiconductor sample in which memory elements (e.g., filled trenches) are configured in a specific pattern. Then, based on this spatial dependence estimate of charge accumulation conforming to the structural types of multiple structures, the spatial dependence estimate of charge accumulation of the sample may be enhanced based on the repeated configuration of the same type of structures on the sample surface. In short, if a certain type of structure (e.g., a transistor gate oxide) exhibits a relatively high charge accumulation due to its insulating properties, this finding based on measurements of a particular transistor gate oxide can be interpolated into multiple other transistor gate oxides configured at other sample locations. This repeated configuration of the same type of structure can be derived from prior knowledge of the sample. For example, for semiconductor structures, repeated configurations may be used in the photomask design data of a lithography mask. The advantage of this enhancement is that it is not necessary to determine the charge accumulation at each location on the sample based on relative measurement step 3015. For example, by moving the sample stage, multiple composite images of the sample can be obtained for different areas on the sample. Then, steps 3015 and 3020 do not need to be performed for each sample stage position.

[0125] Next, in step 3030, at least one aberration of the plurality of primary electron beams generated by the accumulation of sample charge is estimated. This is for estimating one or more time points of charge accumulation. For example, if step 3016 is performed, a predictive estimate can be made.

[0126] This estimation can be determined using a predefined model that provides the charge-induced stray field of the sample surface (i.e., at each location on the sample surface) for each of a plurality of primary electron beams. The farther the spot location of a particular one of the plurality of primary electron beams is from the local charge accumulation, the smaller the amplitude of the electric stray field.

[0127] In other words, based on the estimate of charge accumulation determined in step 3020, the local charges or potentials of discrete finite elements on the sample surface can be determined. Then, using electrostatic equations, the stray electric fields caused by these local charges or potentials and their effects on each primary electron beam can be derived. This is illustrated in Figure 11, specifically for deflection aberration. Figure 11 illustrates a model for calculating the deflection dρ of the primary electron beam 3.1. The sample system is discretized into such charged disks (finite elements) of size A and potential U. Then, based on the electrostatic solid law, the deflection outside and inside a particular charged disk is calculated, where a is a function of A, U, the impact energy LE, and the distance ρ to the charged disk. Figure 12 shows an exemplary dependency of the deflection as a function of the distance to the charged disk charge accumulation 1050 (inside and outside the disk-shaped charge accumulation 1050). This calculation can be repeated for each finite element (e.g., each charged disk) (inner loop). This calculation can be repeated for each primary electron beam (outer loop).

[0128] The model presented above is a relatively simple example. More complex implementations are also possible. For example, the model could include a charge diffusion effect. The charge diffusion term in the model could simulate the spatial diffusion of local charge on a surface. This could provide the time dependence of charge accumulation. The model could also include a time dependence that simulates the change in local charge accumulation on the sample surface over time, which is a function of the sample being exposed to multiple primary electron beams.

[0129] Furthermore, the model is not limited to determining the deflection of each of the multiple primary electron beams. For example, a distortion of the beam spot 15 on the sample surface can also be modeled. Aberrations can be modeled. Other examples include defocus and astigmatism. The attenuation of electron flux can also be modeled. Furthermore, not only beam-specific aberrations in charge accumulation can be modeled, but alternatively or additionally, at least one pattern-specific aberration can also be modeled. Here, the pattern of the primary electron beam on the sample surface may undergo, for example, magnification changes, rotation or translation, or may be tilted.

[0130] Next, some specific examples of the dependencies between local charge accumulation and other modelable aberrations are provided. For example, studies have found that for uniformly accumulated sample charge, the primary electron beam defocusing introduced by the local electric field at the sample is most significant at lower impact energies. There is a clear dependency between sample surface voltage and the amount of defocusing. The amount of defocusing per unit charging voltage also depends on the size of the charged area of ​​the sample surface, i.e., how the charge is distributed due to diffusion or previous scans: the larger the charged area, the greater the amount of defocusing per unit charging voltage. Similar relationships exist between magnification changes or beam astigmatism and the charge distribution on the sample surface. These and other dependencies can be captured by models or lookup tables.

[0131] Based on the knowledge of at least one aberration of the primary electron beam, in step 3035, the sample may be selectively discharged. This may be conditional, for example, by performing step 3034. Exemplary triggering criteria to be checked in step 3034 may include whether the estimated charge accumulation exceeds a certain threshold. The sample discharge in step 3035 may be based on irradiating the sample with neutral charged particles of opposite charge to those used in the primary and secondary charged particle beams in step 3010. The sample discharge in step 3035 may also be based on providing the sample with a neutral gas (e.g., including free radicals). The discharge may be based on mirror mode operation of the MSEM. In mirror mode operation, the accelerating voltage applied to electrons near the source is reduced. Therefore, the primary electrons do not have enough energy to strike the sample. However, the sample may still influence the secondary electron beam based on electrostatic forces.

[0132] Next, we will briefly explain mirror-mode operation discharge. For a neutral sample (no surface charge, zero charge accumulation), the primary electrons in the so-called mirror-mode operation do not interact with the sample but are reflected close above the sample. If the sample surface is positively charged, the surface electric field will attract these electrons and neutralize the surface positive charge. When the surface charge is zero, the process stops because the electrons will no longer interact with the sample.

[0133] Sample discharge can be configured based on the estimate of charge accumulation in step 3020. Sample discharge may not fully compensate for charge accumulation, but it may significantly reduce it. The remaining charge accumulation can be compensated in a subsequent step 3040.

[0134] Note that steps 3034 and 3035 are optional.

[0135] In step 3040, the accumulation of (residual) charge can be compensated.

[0136] For example, if the charge accumulation and at least one aberration of the primary electron beam are predicted based on predictions of one or more characteristics of the secondary electron beam obtained from step 3016, this compensation can be performed in advance. The advantage of advance compensation is that the processing delays for determining one or more characteristics of the secondary electron beam in step 3015, determining the estimate of charge accumulation in step 3020, determining the aberrations in step 3030, and determining the appropriate compensation operation in step 3040 can be taken into account. Specifically, a prediction range can be on the same order of magnitude as or larger than the processing delay. Therefore, even if the measurement in step 3015 has been performed in advance, the specific charge accumulation and aberration that need to be compensated can be estimated at the precise point in time when step 3040 is performed. For example, for an MSEM device equipped with standard hardware, a typical processing delay is approximately 400 milliseconds to 1 second. Due to the large image size, this time is required to perform image analysis to determine the magnification of the secondary electron beam pattern. Therefore, the processing delay can be calculated by predicting the magnification within a prediction range of 0.5 seconds to 2 seconds.

[0137] Step 3040 may include a physical compensation (step 3045), wherein one or more beam control elements of the beam generation apparatus of the multi-beam scanning electron microscope (see Figure 1: beam generation apparatus 300) are controlled based on an estimate of at least one aberration of the multiple primary electron beams. For example, a field lens (such as field lens 333) may not individually affect the primary electron beams 3.1, 3.2, 3.3. Instead, it will have a coherent effect on all primary electron beams 3.1, 3.2, 3.3. For example, pattern-specific aberrations may be compensated by the field lens. For example, magnification variations in the patterns of the primary electron beams 3.1, 3.2, 3.3 may be compensated using lens excitation or multipole excitation. Since magnification variations manifest as a distortion of the image, numerical correction of the image may also be performed (as described below). For a specific aberration, if there are differences between primary electron beams, this global compensation may only compensate for the baseline or linear gradient of the specific aberration observed in all primary electron beams; after this global compensation, the residual effect of the magnitude of the variation in each individual primary electron beam can be observed. Sometimes, a MAP such as MAP 306 can individually affect each of the multiple primary electron beams. That is, each aperture of the porous plate can be combined with a corresponding beam-shaping unit that locally acts on that particular primary electron beam. In this case, one or more beam characteristics of each primary electron beam are locally compensated on the porous plate. For example, micro-optical elements used as multipoles on individual beams can be used to correct the astigmatism of individual beams. This physical compensation for at least one aberration of multiple primary electron beams can also be called pre-compensation. That is, the multiple primary electron beams are affected before they are affected by the accumulation of local charge, such as deflection, distortion, spot size changes, etc. Therefore, each of the multiple primary electron beams can be pre-treated before reaching the sample: for example, a pre-deflection can be applied that is approximately opposite to the deflection caused by charge accumulation downstream of the particle beam at the sample. The net deflection is then zero (or at least close to zero).

[0138] This pre-compensation for at least one aberration can be performed in a time-dependent manner. That is, in some cases, the measurement system at step 3015 can be performed using a time-resolved method. Then, the charge accumulation estimate determined at step 3020 can be time-dependent. This time dependence can be used to estimate the time dependence of at least one aberration of a primary electron beam at step 3030. For longer irradiation durations, there is typically a higher charge accumulation. However, the charge accumulation may saturate to a certain value.

[0139] Optionally or additionally, at least one aberration may be compensated in the digital post-processing of the sample image acquired using a multi-beam scanning electron microscope (see step 3050). For example, pixel-level conversion may be used for such images, wherein the conversion parameters are determined based on the at least one aberration, which may include one or more beam-specific aberrations.

[0140] For each field of view, acquire each image in a sequence of images to perform digital post-processing. Typically, a sequence of images is acquired for each field of view. Each of these images can be acquired using a relatively short dwell time of an electron beam at each scan position. This is done to limit charge accumulation. Nevertheless, for multiple subsequent images, certain positions on the sample are scanned multiple times, so charge accumulation may increase with different images in the sequence of images. As previously described with respect to Figure 10B, the estimate of charge accumulation can be temporally dependent (based on the temporally dependent measurement in step 3015). Then, before combining each image in the sequence of images as part of frame averaging processing, each of these images in the sequence of images can be related to relative charge accumulation, and furthermore, may also be related to a relative estimate of at least one aberration of multiple primary electron beams. Therefore, in the compensation in step 3050, the different intensities / values ​​of at least one aberration can be considered for each image in the sequence of images acquired for each field of view.

[0141] Generally, a combination of steps 3045 and 3050 can be considered. For example, step 3045 can be performed while imaging the sample. Then, when step 3050 is performed, residual effects of charge accumulation can be compensated after imaging is completed.

[0142] Steps 3010 to 3035 can be repeated, for example, for multiple different positions on the sample stage or different (parts) of the composite field of view. This is shown by the dashed arrow and the repeated 3009.

[0143] In summary, the disclosed technique can infer at least one aberration, specifically deflection, of multiple primary charged particle beams based on measurements of multiple secondary charged particle beams in a multi-beam scanning charged particle microscope. However, the type of aberration is not limited to deflection of multiple primary charged particle beams. Other types of aberrations, such as astigmatism and defocusing, can also be considered. This calculation is performed by an intermediate step that determines the charge accumulation of the sample. In online processing involving in-situ measurements, each of the at least one aberration can be pre-compensated physically. Alternatively or additionally, post-compensation can be performed for each of the at least one aberration in post-processing of images acquired using a multi-beam scanning charged particle microscope.

[0144] To further summarize, at least the following examples have been revealed:

[0145] Although the present invention has been shown and described in conjunction with certain preferred embodiments, various equivalents and modifications will arise for those skilled in the art upon reading and understanding this specification. The present invention includes all such equivalents and modifications, and is limited only to the scope of the claims made below.

[0146] For ease of explanation, various examples have been specifically disclosed in conjunction with MSEM. However, similar techniques can also be applied to other multi-beam scanning charged particle microscopes, such as those using helium ions.

[0147] To further illustrate, various examples of deflection aberrations are discussed in this paper. However, the techniques disclosed are not limited to determining the deflection estimate of each of multiple primary charged particle beams. Other types of aberrations can also be estimated.

[0148] To further illustrate, various examples have been previously disclosed relating to determining at least one aberration of a plurality of primary charged particle beams in a multi-beam scanning charged particle microscope. However, similar techniques can also be used to determine at least one aberration of a single primary charged particle beam in a single-beam scanning charged particle microscope (such as a conventional SEM).

[0149] To further illustrate, various examples have been disclosed in which, in a first step, an estimate of the sample charge accumulation is determined, and then in a second step, an estimate of at least one aberration of the plurality of primary charged particle beams is determined. Generally, the second step of determining the estimate of at least one aberration of the plurality of primary charged particle beams is selective. The sample charge accumulation is estimated, and compensation for related aberrations is performed based on the sample charge accumulation. For example, some techniques have been previously disclosed in which characteristics of secondary charged particle beam patterns are predicted based on relative measurements. These techniques for predicting secondary charged particle beam pattern characteristics can be used to predict charge accumulation. In this case, compensation can be applied based on the predicted charge accumulation, for example, pre-compensation, and it is not even necessary to explicitly or implicitly determine the estimate of the aberrations of the plurality of primary charged particle beams caused by charge accumulation. [Simplified Explanation of the Diagram]

[0023] Figure 1 schematically illustrates MSEM based on various examples.

[0024] Figure 2 schematically illustrates a porous plate of an MSEM, which defines a pattern of a primary electron beam according to various examples.

[0025] Figure 3 schematically illustrates the focal position of a primary electron beam on the sample surface according to various examples.

[0026] Figure 4 schematically illustrates the pattern of the secondary electron beam in the imaging plane according to various examples.

[0027] Figure 5 schematically illustrates the deflection of the primary and secondary electron beams caused by charge accumulation in the sample according to various examples.

[0028] Figure 6 schematically illustrates a distorted image of the field of view defined by a specific primary electron beam on the scanned sample.

[0029] Figure 7 schematically illustrates the undistorted image of the field of view shown in Figure 6.

[0030] Figure 8 schematically illustrates a detector system of MSEM according to various examples.

[0031] Figure 9 schematically illustrates the patterns of optical detection fibers according to various examples.

[0032] Figure 10A is a flowchart of the methods based on various examples.

[0033] Figures 10B and 10C schematically illustrate the time dependence of multiple characteristics of secondary electron beams according to various examples.

[0034] Figures 10D and 10E illustrate the characteristics of the predicted secondary electron beam based on various examples.

[0035] Figure 11 schematically illustrates a model used to determine the deflection estimate of a primary electron beam in the case of charge accumulation on a sample.

[0036] Figure 12 illustrates the calculated estimate of the electron beam deflection obtained from the model in Figure 11.

Claims

1. A method of operating a multi-beam scanning charged particle microscope (1), comprising: - loading (3005) a sample (7) into an imaging position of the multi-beam scanning charged particle microscope (1); - irradiating (3010) the sample (7) with a plurality of primary charged particle beams (3.1, 3.2, 3.3) while the sample (7) is in the imaging position; - measuring at least one characteristic of secondary charged particle beams (9.1, 9.2, 9.3) relating to the primary charged particle beams (3.1, 3.2, 3.3); - determining (3020) an estimate of the charge accumulation (1050) of the sample (7) based on the at least one characteristic (7001, 7002, 7003, 7004) of the secondary charged particle beams; Determine (3030) an estimate of at least one aberration of the primary charged particle beam generated by the charge accumulation (1050) of the sample (7); and - perform (3040, 3045, 3050) compensation of the at least one aberration based on the estimate of the primary charged particle beam.

2. The method as described in claim 1, wherein the estimate of the at least one aberration is determined using a predetermined model that provides a charge-induced stray field at the sample surface of the sample (7) at multiple locations in each of the primary charged particle beams.

3. The method as described in claim 2, wherein the model includes a charge diffusion term to simulate the spatial diffusion of local charges on the surface.

4. The method as described in claim 2 or 3, wherein the model includes a time dependency that simulates the time dependency of the charge accumulation, the time dependency being a function of the sample (7) being exposed to the primary beam of charged particles during the irradiation period.

5. The method as described in claim 1, wherein the estimate of the charge accumulation (1050) is time-dependent, and wherein the estimate of the at least one aberration is time-dependent.

6. The method as described in claim 5, wherein the compensation is time-dependent.

7. The method as described in claim 1, further comprising: - predicting at least one characteristic of the secondary charged particle beams based on the measurement, wherein, The prediction based on at least one characteristic is used to predict an estimate of the charge accumulation at a future point in time.

8. The method as described in claim 7, wherein the prediction is based on a machine learning model operated on an input sequence.

9. The method as claimed in claim 1, wherein the at least one characteristic of the secondary charged particle beams comprises one or more characteristics of the pattern (41.3) of the secondary charged particle beams (9.1, 9.2, 9.3).

10. The method as claimed in claim 9, wherein one or more characteristics of the pattern (41.3) of the secondary charged particle beams include a magnification of the pattern (41.3).

11. The method as claimed in claim 9, wherein one or more characteristics of the pattern (41.3) of the secondary charged particle beams (9.1, 9.2, 9.3) include at least one of a beam spacing (1061) or a line spacing (1062).

12. The method as claimed in claim 1, wherein the at least one characteristic of the secondary charged particle beams includes at least one of the following: a spot (15) displacement of one or more secondary charged particle beams, a spot (15) size of one or more secondary charged particle beams, a spot (15) shape of one or more secondary charged particle beams, or a spot (15) brightness of one or more secondary charged particle beams.

13. The method as described in claim 1, wherein the estimate of the charge accumulation (1050) of the sample (7) is based on a predefined lookup table that links the value of the at least one characteristic to the charge accumulation (1050).

14. The method as claimed in claim 1, wherein the estimate of the charge accumulation (1050) of the sample (7) is determined based on a predetermined model that links the value of the at least one characteristic to the charge accumulation (1050).

15. The method as described in claim 1, wherein the estimate of the charge accumulation (1050) of the sample (7) is determined based on ray tracing simulations of the secondary charged particle beams.

16. The method as described in claim 1, wherein the estimate (1050) of the charge accumulation is spatially dependent.

17. The method as described in claim 16 further comprises: - matching the spatial dependence estimate of the charge accumulation (1050) of the sample (7) with the structure type of a plurality of structures of the sample (7).

18. The method as described in claim 17 further comprises: - enhancing (3025) the spatial dependence estimate of charge accumulation (1050) of the sample (7) based on the repeated configuration of the same type of structure on the sample surface of the sample (7).

19. The method as described in claim 1, wherein the at least one aberration comprises at least one bundle of specific aberrations.

20. The method as claimed in claim 1, wherein the at least one aberration comprises at least one pattern aberration of the pattern (41.2) of the primary charged particle beams (3.1, 3.2, 3.3).

21. The method as described in claim 1, wherein the at least one aberration includes the deflection of each of the primary charged particle beams.

22. The method as described in claim 1, wherein the at least one aberration includes spot distortion of each of the primary charged particle beams.

23. The method as described in claim 1, wherein the at least one aberration includes defocusing of each of the primary charged particle beams.

24. The method as described in claim 1, wherein the at least one aberration comprises astigmatism of each of the primary charged particle beams.

25. The method as claimed in claim 1, wherein the at least one aberration comprises a magnification change of the pattern of the primary charged particle beams.

26. The method as described in claim 1, wherein the at least one aberration includes rotation of each of the first-order charged particle beams.

27. The method as described in claim 1, wherein the execution of the compensation comprises: - controlling (3045) one or more beam control elements of the multi-beam charged particle microscope to deflect the primary charged particle beams thereby canceling the at least one aberration.

28. The method as described in claim 27, wherein at least one of the one or more beam control elements acts coherently on the primary charged particle beams.

29. The method as described in claim 27, wherein at least one of the one or more beam control elements individually acts on each of the primary charged particle beams.

30. The method as described in claim 1, wherein the execution of the compensation comprises: - digital post-processing (3050) of an image of the sample (7) obtained using a multi-beam scanning charged particle microscope.

31. The method as described in claim 1 further comprises: - discharging the sample (7) (3035) based on the estimate of the charge accumulation (1050) of the sample (7).

32. A circuit for controlling the operation of a multi-beam scanning charged particle microscope (1), the circuit being configured to perform: - controlling the multi-beam scanning charged particle microscope to load (3005) a sample (7) into an imaging position of the multi-beam scanning charged particle microscope (1); - controlling the multi-beam scanning charged particle microscope to irradiate the sample (7) with a plurality of primary charged particle beams (3.1, 3.2, 3.3) when the sample (7) is in the imaging position; - measuring at least one characteristic of secondary charged particle beams (9.1, 9.2, 9.3) relating to the primary charged particle beams (3.1, 3.2, 3.3); - determining (3020) an estimate of the charge accumulation (1050) of the sample (7) based on the at least one characteristic (7001, 7002, 7003, 7004) of the secondary charged particle beams; Determine (3030) an estimate of at least one aberration of the primary charged particle beam generated by the charge accumulation (1050) of the sample (7); and - based on the estimate of the at least one aberration of the primary charged particle beam, perform (3040, 3045, 3050) compensation for the at least one aberration.

33. The circuit as described in claim 32, wherein the circuit is configured to perform the method as described in any one of claims 1 to 31.

Images