Multi-beam digital scanning and image acquisition

Abstract

A multi-beam charged particle microscope and a method for operating the multi-beam charged particle microscope for wafer inspection with high throughput, high resolution, and high reliability. The operating method and the multi-beam charged particle microscope include means for performing a synchronized scanning operation and image acquisition using a plurality of charged particle beamlets according to a selected scanning program, wherein the selected scanning program can be selected based on inspection tasks from different scanning programs.

Description

Technical Field

[0001] The present invention relates to a multi-beam charged particle beam scanning inspection system, comprising a highly flexible and high-throughput multi-beam scanning and image acquisition system, a related method, and a computer program product. The multi-beam charged particle beam inspection system includes an improved scanner architecture and deflection scanner operation controller, and an improved image acquisition module for synchronized image acquisition. Background Art

[0002] The continued development of increasingly smaller and more complex microstructures, such as semiconductor devices, necessitates further development and optimization of planar manufacturing technologies and inspection systems for the fabrication and inspection of small-scale microstructures. Semiconductor device development and manufacturing require design verification, such as on test wafers, and planar manufacturing technologies involve process optimization for reliable, high-throughput manufacturing. Furthermore, there is a growing need for analysis of semiconductor wafers for reverse engineering and customized, individual configurations of semiconductor devices. Consequently, high-throughput inspection tools are needed to accurately test microstructures on wafers.

[0003] Typical silicon wafers used to manufacture semiconductor devices have a diameter of up to 12 inches (300 mm). Each wafer is divided into 30 to 60 repeating regions ("bare dies"), with a maximum area of approximately 800 square millimeters. Semiconductors consist of multiple semiconductor structures fabricated layered on the wafer surface using planar integration techniques. Due to the manufacturing processes involved, semiconductor wafers typically have a flat surface. The feature sizes of integrated semiconductor structures range from a few μm down to a critical dimension (CD) of 5 nm, and in the near future, feature sizes will continue to decrease, for example, to feature sizes or CDs below 3 nm (nanometers) (e.g., 2 nm), or even below 1 nm. With these small feature sizes, defects of critical dimensions must be identified over a large area within a short period of time. For some applications, the measurement accuracy specifications imposed by inspection equipment are even higher, for example, by a factor of two or an order of magnitude. For example, the width of semiconductor features must be measured with an accuracy of less than 1 nm, for example, 0.3 nm or even less, and the relative positions of semiconductor structures must be determined with an overlay accuracy of less than 1 nm, for example, 0.3 nm or even less.

[0004] Therefore, it is an object of the present invention to provide a charged particle system and a method for operating a charged particle system that allow high-throughput inspection of integrated semiconductor features with a resolution of at least critical dimensions during development or manufacturing, or for reverse engineering of semiconductor devices. Another object of the present invention is to provide a charged particle system and a method for operating a charged particle system that allow high-precision measurement of semiconductor features with an accuracy of less than 1 nm, less than 0.3 nm, or even 0.1 nm.

[0005] A recent development in the field of charged particle microscopy (CPM) is the MSEM, a multi-beam scanning electron microscope (MSEM). For example, US Pat. No. 7,244,949 and US Pat. No. 2019,0355,544 disclose a multi-beam charged particle microscope. In a multi-beam charged particle microscope, such as a multi-beam electron microscope or MSEM, a sample is irradiated by an array of electron beamlets comprising, for example, J = 4 to up to J = 10,000 electron beams (as primary radiation), whereby each electron beam is separated from its next adjacent electron beam by a distance of 1–200 micrometers. For example, an MSEM has approximately J = 100 separated electron beams or beamlets arranged in a hexagonal array, where the electron beamlets are separated by a distance of approximately 10 μm. Multiple primary charged particle beamlets are focused by a common objective onto the surface of the sample under investigation, such as a semiconductor wafer fixed to a wafer chuck mounted on a movable platform. During the irradiation of the wafer surface with a primary charged particle beamlet, interaction products, such as secondary electrons, originate from a plurality of intersection points formed by the focal point of the primary charged particle beamlet, and the number and energy of the interaction products depend on the material composition and morphology of the wafer surface. The interaction products form a plurality of secondary charged particle beamlets, which are collected by a common objective lens and directed to a detector arranged on a detector plane by a projection imaging system of a multi-beam inspection system. The detector includes a plurality of detection areas, each of which includes, for example, a plurality of detection pixels. The detector detects the intensity distribution of each of the plurality of secondary charged particle beamlets and obtains a digital image patch, for example, of 100 μm×100 μm.

[0006] During image acquisition by an MSEM, a plurality of J primary beamlets are scanned over a sample (e.g., a wafer surface) using a collective charged particle deflection system. At J scanning positions on the sample surface, secondary electrons are generated, thereby forming a plurality of J secondary electron beamlets. The J secondary beamlets are imaged onto an array of at least J detectors, each detector receiving a signal corresponding to the number of secondary electrons or the intensity of an individual secondary electron beamlet. Thus, a plurality of J image subfields are scanned during an image scan. The J digital image segments corresponding to the J image subfields are stitched together to form one or more digital image tiles. The deflection scanners of prior art MSEM systems have been linear image scanners having linear image processing configured for a linear data stream of intensity values for a plurality of J image subfields. It has been shown that a more flexible image scanner and a more flexible image data processing architecture are needed.

[0007] For single-beam electron microscopy, different scanning strategies are known. For example, in US 10720306 B2, an interlaced scanning strategy is applied, which is well known in cathode ray tube (CRT) television systems. As with CRT systems, this interlaced scanning strategy is known to reduce the charging effect of the sample. A similar method for single-beam microscopy is described in SCANNING VOL. 23, 395–402 (2001) by JTLTHONG, KWLEE, WKWONG et al. under the title “Reduction of Charging Effects Using Vector Scanning in the Scanning Electron Microscope”.

[0008] However, multi-beam charged-particle microscopy presents additional challenges. First, the data throughput is much higher, and image data must be generated for multiple beamlets in parallel. Second, recent requirements necessitate much higher precision in scanning and data acquisition control than was required just a few years ago. Third, multiple subfields must be scanned and imaged in parallel. This brings with it additional requirements and limitations, such as the distance or spacing between the multiple primary charged-particle beamlets, which must be considered, as well as distortion caused by scanning or image post-processing requirements.

[0009] In a single-beam scanning electron microscope (SEM), pre-compensation of the distortion caused by scanning is well known in the art and is intuitive. In addition, the choice of scan path configuration for a single-beam scanning microscope is almost unlimited. For example, a single-beam raster scanner can modify the magnification or resolution almost arbitrarily. However, this is different in a multi-beam charged particle scanning microscope, in which a plurality of J primary charged particle beamlets are arranged in a raster configuration at a given spacing. A plurality of J primary charged particle beamlets are collectively scanned in parallel, and a plurality of J digital image data from a plurality of J image subfields are collected in parallel. For example, the magnification or orientation of the scanning operation cannot be changed arbitrarily.

[0010] Therefore, one problem of the present invention is to provide a multi-beam charged particle scanning inspection system having a plurality of J primary charged particle beamlets for performing wafer inspection with high imaging accuracy, high speed or high throughput. Another problem of the present invention is to provide a multi-beam charged particle inspection system having improved mitigation of charging effects during wafer inspection. Another problem of the present invention is to provide a multi-beam charged particle inspection system that reduces the need for post-processing of acquired image subfields and for forming image blocks from image subfields. Another problem of the present invention is to provide a multi-beam charged particle scanning inspection system having a plurality of J primary charged particle beamlets, the system being configured to perform collective scanning imaging of a plurality of J image subfields in different scanning procedures or scanning modes. Summary of the Invention

[0011] According to the present invention, an improved multi-beam scanning and image acquisition unit for a multi-beam charged particle scanning microscope having a plurality of J primary charged particle beamlets is provided. With the improved multi-beam scanning and image acquisition unit, a plurality of primary charged particle beamlets can be collectively scanned with high precision and synchronized, rapid image acquisition. According to one aspect of the present invention, the improved multi-beam scanning and image acquisition unit enables different scanning procedures or scanning modes with high precision, for example, with a distortion of less than 1 nm. The image acquisition of a plurality of image subfields of an image tile is synchronized with the selected scanning procedure or scanning mode. In one example, image acquisition is enabled without extensive image post-processing. According to one aspect of the present invention, the improved multi-beam scanning and image acquisition unit enables different scanning procedures or scanning modes of a plurality of primary charged particle beamlets and synchronized, rapid image acquisition of a plurality of image subfields of an image tile. According to a further aspect of the present invention, the improved multi-beam scanning and image acquisition unit enables interruption of the scanning procedure or scanning mode of a plurality of primary charged particle beamlets and image acquisition of a plurality of image subfields of an image tile. During the interruption of the first scanning procedure, further tasks or operations are performed, including collectively scanning the plurality of primary charged particle beamlets according to a second scanning procedure.

[0012] Using a method according to a first embodiment, an improved method for operating a multi-beam scanning and image acquisition unit is provided, which minimizes distortion caused by scanning in a multi-beam charged particle scanning microscope. Using a second embodiment, an improved method for operating a multi-beam scanning and image acquisition unit is provided, which is capable of performing collective raster scanning and image acquisition according to a selected scanning program or scanning mode. In a third embodiment, an improved multi-beam scanning and image acquisition unit is provided. Using embodiments of the present invention, imaging performance is improved, greater flexibility is provided, and the throughput of inspection tasks is increased.

[0013] In the multi-beam scanning and image acquisition method according to the second embodiment, a scanning program is selected and provided in the configuration step. The multi-beam image scanning and acquisition method is separated in the first general scan processing step, the second specific scan deflection control step, and the third image data acquisition step.

[0014] The first general scan processing step includes at least one of a scan command receiving step and a scan command processing step. The improved multi-beam scanning and image acquisition unit receives a selected scan program and converts it into a series of unit scan coordinates in unit subfield coordinates, such as pixel coordinates with integer precision in an image subfield. The first general scan processing step also includes a vertex post-processing step. Based on the selected scan program, a series of pre-compensated digital scan commands are generated in the first general scan processing step and provided to the second specific scan deflection control step. Each scan deflection control step includes at least one of a specific conversion step, a vertex post-processing step, a digital-to-analog conversion step, and an amplification step. The third image data acquisition step includes an A / D conversion step, a digital image data selection step, and a digital image data addressing and writing step. In the digital image data addressing and writing step, a plurality of image data are written to a parallel access temporary storage memory corresponding to a plurality of pixel coordinates of a plurality of image data corresponding to a plurality of image subfields. In a further parallel reading and image processing step, the plurality of image data are read from the parallel access memory and post-processed.

[0015] The modular separation between the first general scan processing step and the second scan deflection control step has the advantage that each specific scan deflection control step can be adjusted or calibrated for a specific first collective multi-beam raster scanning system or second collective multi-beam raster scanning system without having to adapt or modify the general scan processing step. Thus, the specific scan deflection control step can be modified, for example, based on the needs of the specific first collective multi-beam raster scanning system, such as when using an octopole scanner, a quadrupole scanner sequence, a dipole scanner sequence, or another collective multi-beam raster scanning system. By separating the modules between the first general scan processing step and the second scan deflection control step, the systematic aberrations of the multi-beam charged particle system are separated from the specific aberrations or nonlinear effects of the actual collective scanning system. The systematic aberrations of the multi-beam charged particle system are pre-compensated in the first general scan processing step. The specific aberrations or nonlinear effects of the actual collective scanning system (e.g., the first collective multi-beam raster scanning system or the second collective multi-beam raster scanning system, including the nonlinearity of the voltage amplification during the amplification step, respectively) are pre-compensated in the specific second scan deflection control step. In one example, a multi-beam scanning and image acquisition method includes at least a first scanning deflection control step for operating control of a first collective multi-beam raster scanner and a second scanning deflection control step for operating control of a second collective multi-beam scanner. The multi-beam scanning and image acquisition method may include a further scanning deflection control step.

[0016] An advantage of module separation in the first general scan processing step and the third image data acquisition step is that different and complex scan programs can be selected and configured for image acquisition of multiple J subfields, synchronized with the collective scanning operation of multiple J charged particle beamlets. For example, as a unit scan command is generated and provided to the image data acquisition step, the J fluctuation voltage stream from the image sensor is converted, selected, sorted, and written to a parallel access memory at multiple J addresses, each based on a unit scan coordinate. For different scan programs, multiple J memory address pointer sequences are generated for multiple J image data corresponding to the multiple J image subfields, and the multiple J image data are written to a parallel access memory at the multiple J memory address pointer sequences. Using this approach, different scan programs can be enabled, including those with arbitrary or random scan patterns, separate scan paths, or scan patterns with cascaded resolution enhancement.

[0017] In another aspect of the present invention, a selected first scanning procedure can be interrupted, and a second scanning procedure or scanning mode can be selected during the interruption of the first scanning procedure. For example, during the interruption of the first scanning procedure, calibration measurements or repetitive measurements of multiple J image segments in multiple J image subfields can be performed according to the second scanning mode, and actual performance characteristics such as drift of the multi-beam charged particle microscope can be monitored. Digital image data corresponding to the second scanning procedure or scanning mode is written to a different memory address and treated as digital image data from the first scanning procedure.

[0018] A multi-beam scanning and image acquisition method utilizes a common clock signal to synchronize operations and to time and stream data streams or sequences generated during the multi-beam scanning and image acquisition method. The common clock signal is generated, for example, in a common scan processing step and provided to a second scan deflection control step and a third image data acquisition step. The parallel readout and image processing steps can operate at different or the same clock frequencies. The common control unit of the multi-beam charged particle scanning microscope can operate at different or the same clock frequencies. In the multi-beam scanning and image acquisition method, for example, the DA conversion step and the AD conversion step are synchronized to provide synchronization between raster scan deflection and image pixel data collection. Delays in data processing after the AD conversion step, as well as delays in the amplification step and the collective deflection step after the DA conversion step, can be calibrated and accounted for, for example, in the digital image data selection step. Thus, a plurality of digital image pixel data corresponding to actual scan positions in local subfield coordinates (p, q) are written to a plurality of memory addresses, and an address corresponding to a unit scan command corresponding to the actual scan position in the subfield coordinates (p, q) is written to each of the plurality of memory addresses.

[0019] Using a method according to a first embodiment, scan-induced distortion is minimized overall, regardless of its origin. According to the first embodiment, a method for calibrating a multi-beam charged-particle scanning electron microscope is provided, thereby reducing maximum scan-induced distortion. The method comprises a first step of performing calibration measurements by raster scanning a plurality of J primary charged-particle beamlets over a calibration sample surface using a collective multi-beam raster scanner with a first drive signal V1(p,q). In a second step of deriving a plurality of scan-induced distortion patterns from the calibration measurements, a scan-induced distortion pattern is derived for each of the plurality of primary charged-particle beamlets. In a third step, the plurality of scan-induced subfield distortion patterns are analyzed, for example, by a statistical method, and a correction signal C(p,q) is derived. In one example, the third step of analyzing comprises deriving a reference distortion pattern for the plurality of scan-induced distortion patterns by a statistical method, the statistical method comprising calculating an average value, a weighted average value, or a median value. The correction signal C(p,q) is derived from the reference distortion pattern. In the fourth step of the method, the first drive signal V1(p,q) is modified using the correction signal C(p,q), and a modified drive signal V2(p,q) is derived for driving the first collective multi-beam raster scanner. By applying the modified drive signal V2(p,q) to the first collective multi-beam raster scanner, the reference distortion pattern is pre-compensated for each of the multiple scan-induced distortion patterns, and the distortion caused by the maximum scan is minimized. In one example, steps 1 through 4 are repeated until the distortion caused by the maximum scan is minimized below a predetermined threshold.

[0020] According to a first embodiment, a calibration method for a multi-beam charged particle scanning electron microscope 1 includes:

[0021] a first step of performing a calibration measurement by raster scanning the plurality of primary charged particle beamlets 3 with a first drive signal V1 (p, q) over the surface 25 of the calibration sample with a collective multi-beam raster scanner 110,

[0022] - a second step of deriving a plurality of scanning-induced sub-field distortion patterns from the calibration measurements, comprising deriving a scanning-induced distortion pattern for each of the plurality of primary charged particle beamlets 3,

[0023] - The third step is to analyze the subfield distortion patterns caused by multiple scans and derive a correction signal C(p,q),

[0024] - a fourth step of modifying the first drive signal V1(p,q) with the correction signal C(p,q) and deriving a modified drive signal V2(p,q) for driving the collective multi-beam raster scanner 110,

[0025] - This reduces the maximum scanning-induced distortion.

[0026] In one example, the third step of analysis includes the following further steps:

[0027] - deriving a reference distortion pattern of the distortion patterns caused by the plurality of scans by a statistical method, the statistical method comprising calculation of an average, a weighted average, or a median; and

[0028] - deriving a correction signal C(p,q) from the reference distortion pattern.

[0029] In one example, steps 1 to 4 are repeated until the maximum scan-induced subfield distortion of each of the plurality of primary charged particle beamlets is minimized below a predetermined threshold. In one example, the maximum scan-induced subfield distortion of most of the plurality of primary charged particle beamlets 3 is reduced below the predetermined threshold, and the maximum scan-induced subfield distortion of a few individual primary charged particle beamlets exceeds the predetermined threshold. The few individual primary charged particle beamlets exceeding the predetermined threshold are marked, and the scan-induced subfield distortion of the image subfield corresponding to the marked primary charged particle beamlets is compensated, for example, by digital image post-processing. In one example, a correction signal C(p,q) is stored in a memory of a control unit of the collective multi-beam raster scanner to pre-compensate for scan-induced distortion according to a selected scan program. During a scanning operation, the control unit applies the correction signal C(p,q) during the collective raster scanning of the plurality of primary charged particle beamlets according to the selected scan program, and pre-compensates the average scan-induced distortion according to a reference distortion pattern.

[0030] Using the method of the first embodiment, a correction function C(p,q) is generated, by which the subfield distortion caused by scanning of the plurality of primary charged particle beamlets is minimized, for example reduced by a factor of two, or, for example, achieved as a residual scanning-induced distortion below a threshold, for example below 1 nm or below 0.5 nm. The correction function C(p,q) is applied to the scanning voltage signal during use.

[0031] According to a second embodiment, a multi-beam scanning and image acquisition method is provided for controlling collective scanning of a plurality of J primary charged particle beamlets over a plurality of J image subfields and providing for acquiring a plurality of J digital image data corresponding to the plurality of J image subfields using a multi-beam charged particle microscope. The method comprises:

[0032] - a configuration step for providing a plurality of scanning programs and for selecting a selected scanning program;

[0033] - a general scan processing step in which a selected scan program is received and at least a first sequence of pre-compensated digital scan commands and selection control signals is generated from the selected scan program;

[0034] a specific scan deflection control step, in which drive voltages of at least a first amplification sequence are generated from at least a first sequence of pre-compensated digital scan commands, and in which the drive voltages of the at least first amplification sequence are supplied to a collective deflection step for collectively deflecting a plurality of J primary charged particle beamlets over a plurality of J image subfields on a sample surface;

[0035] - an image data acquisition step, in which the J fluctuating voltage stream collected from an image sensor unit in the analog data collection step is converted and selected to form a J digital image data value stream, which is written into a common access memory at a plurality of J memory locations to form a plurality of J digital image data corresponding to a plurality of J image subfields, such that the selection and writing are controlled by a selection control signal generated and provided by the common scanning processing step.

[0036] The method further includes a parallel reading and image processing step, in which a plurality of J digital image data corresponding to the plurality of J image subfields is read from the shared access memory and image processing is performed. The image processing may include one of image filtering, image registration, thresholding, object detection, image object size measurement, distortion compensation, contrast enhancement, deconvolution, or image correlation. The image processing may also include a stitching operation to form a single digital image file from the plurality of J digital image data.

[0037] In one example, during a general scan processing step, at least a first sequence of unit scan commands is generated in normalized subfield coordinates (u, v). The first sequence of unit scan commands is converted into a sequence of pre-compensated digital scan commands in image subfield coordinates (p, q) by applying one of an operation including rotation, rescaling, or taking into account a predetermined correction function C(p, q) received from a memory.

[0038] In one example, the selection control signal generated during the common scan processing step includes at least a first sequence of unit scan commands, and during the image data acquisition step, the J digital image data value stream is written to the common access memory at a plurality of J memory locations corresponding to a plurality of J image subfield coordinates and the first sequence of unit scan commands.

[0039] In one example, during the general scan processing step, a synchronization control command is exchanged with a scan synchronization control step.

[0040] In one example, a plurality of J primary charged particle beamlets are configured in a raster configuration that is rotated by a rotation angle relative to the orientation of a coordinate system of a gantry or a sample mounted on the gantry during use, and wherein in a common scan processing step, the unit scan command sequence is adjusted to compensate for the rotation of the raster configuration.

[0041] The operating method and multi-beam charged particle beam microscope include a device for performing synchronous scanning operations and image acquisition through multiple charged particle beamlets according to a selected scanning program, wherein the selected scanning program can be selected based on the inspection tasks from different scanning programs stored in the memory of the multi-beam charged particle scanning microscope control unit.

[0042] According to an embodiment, there is provided a multi-beam charged-particle microscope configured to perform any method or method step of the first or second embodiment.

[0043] According to a third embodiment, an improved multi-beam scanning and image acquisition unit includes a scan control unit connected to an image acquisition unit. The scan control unit is configured to generate at least one digital signal stream representing scan coordinates in each image subfield according to a selected scan program during use, and convert the digital signal stream into at least one drive voltage sequence or stream for driving a collective multi-beam raster scanner. During use, the collective multi-beam raster scanner is configured to scan a plurality of J primary charged particle beamlets deflected over a plurality of J image subfields according to the drive voltage sequence. A digital image acquisition unit is configured to acquire a plurality of J sequences or streams of digital image data corresponding to the plurality of J image signals from the plurality of J image subfields. The digital image acquisition unit is connected to the scan control unit and configured to synchronize image acquisition with the selected scan program and write the plurality of J digital image data streams to a plurality of J memory locations according to the scan coordinate sequence of the selected scan program. The selected scan program can be one of several different scan programs and can be described by a series of scan commands, e.g., including line and point coordinates, or by a functional description, e.g., including a repeating functional cycle. Different scanning programs can be stored in the memory of the control unit and selected by the multi-beam charged-particle scanning microscope 1 for a specific inspection task.

[0044] According to one embodiment of the present invention, a multi-beam charged particle microscope 1 comprises:

[0045] at least one first collective raster scanner 110 for collectively scanning the plurality J of primary charged-particle beamlets 3 over the plurality J of image subfields 31 . 11 to 31 . MN; and

[0046] a detection system 200 comprising a detector 207 for detecting a plurality of J secondary electron beamlets 9, each beamlet corresponding to one of the J image subfields 31.11 to 31.MN; and

[0047] - Imaging control module 820, the imaging control module 820 includes:

[0048] a scanning control unit 930 connected to the first collective raster scanner 110 and configured to control, during use, the raster scanning operation of the plurality J of primary charged particle beamlets 3 using the first collective raster scanner 110 according to a first selected scanning program 762,

[0049] - an image data acquisition unit 810, which is connected to the scan control unit 930 and the detector 207, and is configured to acquire and select a plurality of J image data from the detector 207 during use, in synchronization with a clock signal provided by the scan control unit 930, and is configured to write the plurality of J image data into the parallel access memory 1816 at a memory location according to the first selected scan program 762.

[0050] In one example, the scan control unit 930 includes a clock signal generator 938 configured to provide a clock signal to the scan control unit 930 and the image acquisition unit 810 during use. The scan control unit 930 may be further connected to at least one further system 960 configured to operate synchronously with the raster scanning operation. A further system 960 may be a collective deflector 350 configured to collectively deflect a plurality of J primary charged particles into a beam-dump 130 during use.

[0051] The multi-beam charged particle microscope 1 may further include a second collective raster scanner 222 disposed in the detection system 200 . The scanning control unit 930 is further connected to the second collective raster scanner 222 .

[0052] The imaging control module 820 of the third embodiment may include a voltage supplier 925 configured to provide voltage to the scanning control unit 930 and the image acquisition unit 810, and configured to provide a driving voltage to the first collective raster scanner 110 or the second collective raster scanner 222 during use.

[0053] According to one embodiment, the scanning control unit 930 of the multi-beam charged particle scanning microscope 1 includes:

[0054] - a scan generator module 932 connected to the clock unit 938; and

[0055] a first amplifier module 936 . 1 connected to the first collective raster scanner 110 ; and

[0056] a second amplifier module 936 . 2 connected to the second collective raster scanner 222 ;

[0057] - the scan generator module 932 is configured to generate and provide a series of pre-compensated digital scan commands to the first amplifier module 936.1 and the second amplifier module 936.2 during use; and

[0058] - the first amplifier module 936.1 is configured to generate, during use, at least a first amplified sequence of drive voltages to the electrodes of the first collective raster scanner 110;

[0059] - The second amplifier module 936.2 is configured to generate at least one second amplified sequence of drive voltages to the electrodes of the second collective raster scanner 222 during use.

[0060] The scan control unit 930 may further comprise at least one further amplifier module 936 . 3 connected to a system 960 . 3 configured to operate synchronously with the raster scan deflection during use. The scan generator module 932 may further be connected to the image acquisition module 810 .

[0061] According to one embodiment, the image data acquisition unit 810 of the multi-beam charged particle scanning microscope 1 includes:

[0062] an ADC module 1808 comprising a plurality of AD converters connected to the image sensor 207 and configured to convert the plurality of S undulating voltages 786 into a plurality of S digital sensor data streams 788 during use; and

[0063] an acquisition control unit 1812 connected to the ADC module 1808 and the scan control unit 930 and configured to select, during use, the plurality of S digital sensor data streams 788 and, during use, the plurality of S digital image data value streams 790 from a selection control signal 744 provided by the scan control unit 930 in accordance with a selected scan program 762; and

[0064] - an image data classifier 1820 connected to the acquisition control unit 1812, the scanning control unit 930 and the parallel access memory 1816;

[0065] The image data classifier 1820 is configured to write, during use, a plurality of S digital image data value streams 790 to the parallel access memory 1816 at a plurality of memory addresses corresponding to scan positions of the plurality of S primary charged particle beamlets 3 according to a selected scanning program 762. In one example, the imaging control module 820 includes a plurality of L image data acquisition units 810, where S x L = J, where J is the number of the plurality of primary beamlets 3 generated and utilized during use of the multi-beam charged particle microscope 1. In one example, the imaging control module 820 includes one image data acquisition unit 810, where S = J, where J is the number of the plurality of primary beamlets 3 generated and utilized during use of the multi-beam charged particle microscope 1. In one example, the number L of image data acquisition units 810 is selected based on the number S of fluctuation voltages 786, each fluctuation voltage corresponding to a primary beamlet 3 that can be processed by a single image data acquisition unit 810. The number S may be, for example, S=6, 8, 10, 12 or more undulating voltages 786. L may be L=1, 8, 10, 12, 20 or more, for example, L may be 100.

[0066] The ADC module may be connected to a clock unit and may be configured to receive a clock signal from the clock unit 938 during use and synchronize the operation of the plurality of AD converters to convert the plurality of J fluctuating voltages 786 into a plurality of J digital sensor data streams 788 during use. The clock unit 938 may be connected to the control unit 800 and may be configured to receive a control signal from the control unit 800 and may be configured to change the clock frequency of the clock unit 938 during use.

[0067] According to one embodiment, a multi-beam charged particle microscope 1 comprises:

[0068] - a multi-beam generator 300 for generating a plurality of primary charged particle beamlets,

[0069] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0070] - a detection unit 200 comprising a detector 207; and

[0071] - an imaging control module 820 , comprising a scanning control unit 930 and an image acquisition unit 810 , configured to scan and image the inspection site of the sample using a selected scanning program 762 ;

[0072] The scanning control unit 930 comprises: a universal scanning generator module 932 and at least one first amplifier module 936.1 for providing at least one sequence of high voltages to the electrodes of the first collective raster scanner 110; and a second amplifier module 936.2 for providing at least one sequence of high voltages to the electrodes of the second collective raster scanner 220.

[0073] The scanning control unit 930 is adapted to selectively include a third or further amplifier module 936.3 or 936.n for controlling a third or further operating unit 960.3 or 960.n to operate synchronously with the selected scanning program 762 during use.

[0074] In one example, the universal scan generator module 932 includes a vertex post-processing unit that, during use, is configured to pre-compensate for aberrations caused by systematic scanning of the multi-beam charged particle microscope 1. Each amplifier module includes a vertex post-processing unit, a digital-to-analog converter, and an amplifier, whereby each operating unit 960.i in each amplifier module is individually pre-compensated for operating nonlinearities of the amplifier modules combined with the operating units 960.i for synchronous operation. The operating units 960.i can be the first collective multi-beam raster scanner 110 or the second collective multi-beam raster scanner 222.

[0075] In one example, the multi-beam charged particle microscope includes a multi-beam generator configured to generate a plurality of primary charged particle beamlets in a hexagonal grating configuration, and the imaging control module is configured to scan and image an inspection portion of a sample (e.g., a wafer) using a plurality of image subfields (each of the plurality of image subfields having a hexagonal shape). Therefore, a multi-beam charged particle microscope 1 includes:

[0076] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3 in a hexagonal grating configuration,

[0077] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0078] - a detection unit 200 comprising a detector 207; and

[0079] An imaging control module 820 , comprising a scanning control unit 930 and an image acquisition unit 810 , is configured to scan and image the inspection site of the sample through a plurality of J image subfields 31 , each of the J image subfields 31 having a hexagonal shape.

[0080] In one example, a multi-beam charged particle microscope includes:

[0081] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3,

[0082] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0083] - a detection unit 200 comprising a detector 207; and

[0084] - objective lens 102;

[0085] - an imaging control module 820 comprising a scanning control unit 930 and an image acquisition unit 810, for controlling the scanning of the plurality of J primary charged particle beamlets 3, and for controlling the acquisition of image patches of the sample surface via a plurality of scanning lines arranged in a plurality of J image subfields 31,

[0086] The imaging control module 820 is configured to change the orientation of the plurality of scan lines relative to the sample orientation. In one example, the imaging control module 820 is further configured to change at least the length of the scan lines or the number of scan lines to change the size of the plurality of J image subfields 31 used to cover the image patch. In one example, the multi-beam charged particle microscope 1 further includes a control unit 800 for controlling the operating conditions of the objective lens 102, and wherein the orientation of the plurality of scan lines changes based on the rotation of the raster configuration of the plurality of J primary charged particle beamlets 3 caused by the change in the operating conditions of the objective lens 102. The change in the operating conditions can, for example, be a change in the working distance between the sample surface and the reference surface of the multi-beam charged particle microscope 1, a change in the focal plane of the multi-beam charged particle microscope 1, or a change in the magnification of the multi-beam charged particle microscope 1.

[0087] According to one embodiment, the multi-beam scanning and image acquisition method includes selecting a selected scanning procedure and, in accordance with the selected scanning procedure, collectively deflecting a plurality of J primary charged particle beamlets over a plurality of J image subfields on a sample surface. The method also includes acquiring a J fluctuating voltage stream and converting the J fluctuating voltage stream at a clock rate to form a J digital image data value stream. The method also includes processing at least two digital image data values from each of the J digital image data value streams to form a sum, an average digital image data value, or a difference in digital image data values; and writing the J digital image data value stream, including the sum, the average digital image data value, or the difference in digital image data values, to a shared access memory at a predetermined memory location. In this example, each pixel location within each of the plurality of image subfields is illuminated by a primary charged particle beamlet having an extended dwell time. The extended dwell time may correspond to G multiplied by the inverse of the clock rate, where G is an integer of G=2, 3, 4, or greater. In another example, the scanning process includes a single scan pattern and further includes repeated scanning illumination of each pixel position 55 within the single scan. By this method, for example, including pixel or line averaging, the signal-to-noise ratio (SNR) of the digital image data collected from the object surface area can be increased, and, for example, an overview image of the object surface area can be combined with an image of a selected area with a higher SNR.

[0088] Utilizing embodiments of the present invention, more flexible image scanning and acquisition methods are achieved, and a more flexible image scanning and data processing architecture is provided. Utilizing examples and embodiments of the present invention, the imaging accuracy and throughput of multi-beam charged particles are improved, for example, by pre-compensating for scanning nonlinearities to minimize scanning-induced aberrations. For example, by utilizing hexagonal image subfields, the maximum scan deflection is reduced compared to rectangular image subfields, and a large surface area of an image block can be scanned and imaged with lower induced aberrations and enhanced imaging performance at the same throughput or time per image acquisition. For example, each image subfield is divided into different, selectively separated scanning patterns, and the charging of the sample is controlled. For example, the need for image post-processing is reduced, and image data is generated with higher accuracy and higher throughput. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Further details are provided in the examples of embodiments. Further embodiments include combinations or variations of the examples and embodiments described below. Further details are disclosed below with reference to the accompanying drawings. As shown:

[0090] Figure 1 is a diagram of a multi-beam charged particle microscope system according to one embodiment.

[0091] Figure 2Schematic diagram showing further details of a multi-beam charged particle microscope system including a deflector and a beam dump.

[0092] Figure 3 Graph showing coordinates of a first test location and a second test location comprising first and second image patches.

[0093] Figure 4 FIG. 1 is a diagram illustrating the operation of a first collective multi-beam raster scanner for an example of a selected primary charged particle beamlet.

[0094] Figure 5 Graphs of maximum scan-induced distortion for each of a plurality of primary charged particle beamlets (a) before pre-compensation and (b) after pre-compensation of the scan-induced distortion.

[0095] Figure 6 Graph showing compensation of the nonlinear behavior of the first collective multi-beam raster scanner by a nonlinear driving voltage.

[0096] Figure 7 Schematic diagram of the multi-beam scanning and image acquisition method.

[0097] Figure 8 This is an architectural diagram of the imaging control module and related modules.

[0098] Figure 9 A diagram of a scanning control unit and related units or modules.

[0099] Figure 10 A diagram of the amplifier module.

[0100] Figure 11 A diagram illustrating an image data acquisition unit and related units or modules.

[0101] Figure 12 (a) is a diagram of a scan path of an exemplary primary beamlet according to a first scanning procedure, Figure 12 (b) is a diagram illustrating some of the signals exchanged within the imaging control module and related modules.

[0102] Figure 13 (a) is a diagram of a scan path of an exemplary primary beamlet according to a scanning procedure, Figure 13 (b) is a diagram illustrating some of the signals exchanged within the imaging control module and related modules.

[0103] Figure 14 A diagram illustrating a scan path of a further scanning procedure.

[0104] Figure 15 Schematic diagram of the scanning procedure for a hexagonal image subfield according to a hexagonal raster configuration with primary charged particle beamlets.

[0105] Figure 16 Illustration of compensation for rotation of a hexagonal grating configuration using an adapted scanning procedure.

[0106] Figure 17 is a diagram illustrating a method according to a first embodiment.

[0107] Figure 18 A diagram illustrating a memory configuration according to a multi-beam scanning and image acquisition method.

[0108] Figure 19 A diagram illustrating a memory configuration according to a multi-beam scanning and image acquisition method using pixel averaging.

[0109] Figure 20 A diagram illustrating a memory configuration according to a multi-beam scanning and image acquisition method using line averaging.

[0110] Figure 21 A diagram showing a memory configuration according to a multi-beam scanning and image acquisition method using a temporary memory. DETAILED DESCRIPTION

[0111] In the exemplary embodiments of the present invention described below, components that are similar in function and structure are denoted by similar or identical reference numerals as much as possible.

[0112] Figure 1 The schematic diagram shows the basic features and functions of a multi-beam charged particle microscope system 1 according to an embodiment of the present invention. It should be noted that the symbols used in the figure have been selected to symbolize their respective functions. The type of system shown is a multi-beam scanning electron microscope (MSEM or Multi-SEM), which uses a plurality of primary electron beamlets 3 to generate a plurality of primary charged particle beam spots 5 on the surface 25 of an object 7 (e.g., a wafer having a top surface 25 located in the object plane 101 of the objective lens 102). For simplicity, only five primary charged particle beamlets 3 and five primary charged particle beam spots 5 are shown. Electrons or other types of primary charged particles (e.g., ions, in particular helium ions) can be used to achieve the characteristics and functions of the multi-beamlet charged particle microscope system 1. More details of the microscope system 1 are provided in German patent application 102020209833.6 filed on August 5, 2020, which is incorporated herein by reference in its entirety.

[0113] The microscope system 1 comprises an object illumination unit 100 and a detection unit 200, as well as a beam splitter unit 400 for separating the secondary charged particle beam path 11 from the primary charged particle beam path 13. The object illumination unit 100 comprises a charged particle multi-beam generator 300 for generating a plurality of primary charged particle beamlets 3 and adapted to focus the plurality of primary charged particle beamlets 3 in an object plane 101, wherein the surface 25 of the wafer 7 is positioned by a sample stage 500.

[0114] The primary beam generator 300 generates a plurality of primary charged particle beamlet spots 311 within an intermediate image surface 321, which is typically a spherically curved surface to compensate for the field curvature of the object irradiation unit 100. The primary beamlet generator 300 includes a source 301 of primary charged particles (e.g., electrons). The primary charged particle source 301 emits a diverging primary charged particle beam 309, which is collimated by at least one collimating lens 303 to form a collimated beam. The collimating lens 303 typically comprises one or more electrostatic or magnetic lenses, or a combination of electrostatic and magnetic lenses. The collimated primary charged particle beam is incident on a primary multi-beam forming unit 305. The multi-beam forming unit 305 essentially comprises a first porous plate 306.1 illuminated by the primary charged particle beam 309. The first porous plate 306.1 includes a plurality of apertures in a grating configuration for generating a plurality of primary charged particle beamlets 3, which are generated by transmitting the collimated primary charged particle beam 309 through the plurality of apertures. The multi-beamlet forming unit 305 comprises at least one further porous plate 306.2, which is located downstream of the first porous plate 306.1 relative to the direction of motion of the electrons in the electron beam 309. For example, the second porous plate 306.2 has the function of a microlens array and is preferably set to a defined potential so as to adjust the focus position of the plurality of primary beamlets 3 within the intermediate image surface 321. A third active porous plate arrangement (not shown) comprises a separate electrostatic element for each of the plurality of apertures to influence each of the plurality of beamlets separately. The active porous plate arrangement consists of one or more porous plates with electrostatic elements, such as circular electrodes for microlenses, multipole electrodes, or a sequence of multipole electrodes, to form a static deflector array, a microlens array, or an stigmator array. The multi-beamlet forming unit 305 consists of an adjacent first electrostatic field lens 307, which, together with the second field lens 308 and the first or second porous plate, focuses the plurality of primary charged particle beamlets 3 within or near the intermediate image surface 321. Downstream of the multi-beamlet forming unit 305 , a scanning distortion compensator array 601 may be configured.

[0115] In or near the intermediate image plane 321, a static beam steering porous plate 390 is configured with multiple holes having electrostatic elements (e.g., deflectors) to individually steer each of the multiple charged particle beamlets 3. The holes of the beam steering porous plate 390 are configured to have a larger diameter to allow the multiple primary charged particle beamlets 3 to pass through, even when the focus of the primary charged particle beamlets 3 deviates from its designed position. Near the intermediate image plane 321, a scanning compensator array 602 for compensating for telecentricity errors caused by scanning can be configured. The primary charged particle source 301 and the active porous plate configuration 306.1...306.2, the scanning distortion compensator array 601, the beam steering porous plate 390, and the scanning compensator array 602 for compensating for telecentricity errors caused by scanning are controlled by a primary beamlet control module 830, which is connected to the control unit 800.

[0116] The multiple focal points of the primary charged particle beamlets 3 passing through the intermediate image plane 321 are imaged by the field lens assembly 103 and the objective lens 102 into the image plane 101, where the surface 25 of the wafer 7 is located. The object illumination system 100 further includes a collective multi-beam raster scanner 110 near the first beam intersection 108, so that the multiple charged particle beamlets 3 can be deflected by the scanner in a direction perpendicular to the beam propagation direction. The collective multi-beam raster scanner 110 is configured to scan the deflection of the multiple primary charged particle beamlets 3 passing through the collective multi-beam raster scanner 110 at different propagation angles β. The objective lens 102 and the collective multi-beam raster scanner 110 are centered on the optical axis 105 of the multi-beamlet charged particle microscope system 1, which is perpendicular to the wafer surface 25. The multiple primary charged particle beamlets 3 are synchronously scanned on the wafer surface 101 to form multiple raster-arranged beam spots 5. In one example, the grating configuration of the focal points 5 of the plurality of primary charged particle beams 3 is a hexagonal grating of about one hundred or more primary charged particle beamlets 3. The primary beam spots 5 have a distance of about 6 μm to 15 μm and a diameter of less than 5 nm, such as 3 nm, 2 nm or even less. In one example, the beam spot size is about 1.5 nm, and the distance between two adjacent beam spots is 8 μm. At each scanning position of each of the plurality of primary beam spots 5, a plurality of secondary electrons are generated respectively, forming a plurality of secondary electron beamlets 9 with the same grating configuration as the primary beam spots 5. The intensity of the secondary charged particle beamlet 9 generated at each beam spot 5 depends on the intensity of the impacted primary charged particle beamlet 3, the corresponding point 5 illuminated, the material composition and morphology of the object 7 under the beam spot 5, and the charging condition of the sample at the beam spot 5. The secondary charged particle beamlets 9 are accelerated by the electrostatic field generated by the sample charging unit 503, collected by the objective lens 102, and directed to the detection unit 200 by the beam splitter 400. The detection unit 200 images the secondary electron beamlets 9 onto the image sensor 207, forming a plurality of secondary charged particle image spots 15 therein. The detector or image sensor 207 includes a plurality of detector pixels or individual detectors. The intensity of each of the plurality of secondary charged particle beam spots 15 is detected separately, and the material composition of the wafer surface 25 is detected at high resolution over a large image patch of the wafer with high throughput. For example, for a 10×10 beamlet raster with an 8 μm pitch, a single image scan using the collective multi-beam raster scanner 110 produces an image patch of approximately 88 μm×88 μm at an image resolution of, for example, 2 nm or less. The image tiles are sampled at half the beam spot size, so for each beamlet the number of pixels per image line is 8000 pixels, so that an image tile produced by 100 beamlets comprises 64 gigapixels.The control unit 800 collects digital image data.Details of digital image data collection and processing using, for example, parallel processing are described in German patent application 102019000470.1 and US Pat. No. 9,536,702, which are hereby incorporated by reference.

[0117] The plurality of secondary electron beamlets 9 pass through the first collective multi-beam raster scanner 110 and are deflected and scanned by the first collective multi-beam raster scanner 110 and guided by the beam splitter unit 400 to follow the secondary particle beam path 11 of the detection unit 200. The plurality of secondary electron beamlets 9 travel in opposite directions to the primary charged particle beamlets 3, and the beam splitter unit 400 is configured to separate the secondary particle beam path 11 from the primary particle beam path 13, typically by a magnetic field or a combination of a magnetic field and an electrostatic field. Optionally, an additional magnetic correction element 420 is present in the primary particle beam path and the secondary particle beam path. The projection system 205 also includes at least one second collective raster scanner 222, which is connected to the scanning and imaging control unit 820. The control unit 800 and the imaging control unit 820 are configured to compensate for residual differences in the positions of the plurality of focal spots 15 of the plurality of secondary electron beamlets 9 so that the positions of the plurality of electron focal spots 15 remain constant on the image sensor 207.

[0118] The projection system 205 of the detection unit 200 includes further electrostatic or magnetic lenses 208, 209, 210 and a second intersection 212 of the plurality of secondary electron beamlets 9, at which an aperture 214 is located. In one example, the aperture 214 further includes a detector (not shown) connected to the imaging control unit 820. The imaging control unit 820 is further connected to at least one electrostatic lens 206 and a third deflection unit 218. The projection system 205 also includes at least one first multi-aperture corrector 220 having apertures and electrodes for separately influencing each of the plurality of secondary electron beamlets 9, and a further optional active element 216 connected to the control unit 800 or the imaging control unit 820.

[0119] The image sensor 207 consists of an array of sensing areas in a pattern compatible with the raster configuration of the secondary electron beamlets 9 focused onto the image sensor 207 by the projection lens 205. This enables detection of each individual secondary electron beamlet independently of the other secondary electron beamlets incident on the image sensor 207. Figure 1The image sensor 207 shown can be an electron-sensitive detector array, such as a CMOS or CCD sensor. This electron-sensitive detector array can include electron-to-photon conversion units, such as a scintillator element or an array of scintillator elements. In another embodiment, the image sensor 207 can be configured as an electron-to-photon conversion unit or a scintillator plate arranged in the focal plane of multiple secondary electron particle image spots 15. In this embodiment, the image sensor 207 can also include a relay optical system for imaging and directing the photons generated by the electron-to-photon conversion unit at the secondary charged particle image spots 15 on dedicated photon detection elements such as multiple photomultiplier tubes or avalanche photodiodes (not shown). Such an image sensor is disclosed in US 9,536,702, which is cited above and incorporated by reference. In one example, the relay optical system also includes a beam splitter for splitting and directing the light to a first slow light detector and a second fast light detector. The second fast light detector is constituted, for example, by a photodiode array such as avalanche photodiodes, which are fast enough to resolve image signals of the plurality of secondary electron beamlets 9 according to the scanning speed of the plurality of primary charged particle beamlets 3. The first slow light detector is preferably a CMOS or CCD sensor, which provides high-resolution sensor data signals for monitoring the focus 15 or the plurality of secondary electron beamlets 9 and for controlling the operation of the multi-beam charged particle microscope.

[0120] The platform 500 is preferably not moved during the acquisition of an image tile by scanning the plurality of primary charged particle beamlets 3, and after an image tile is acquired, the platform 500 is moved to the next image tile to be acquired. In an alternative embodiment, the platform 500 is continuously moved in the second direction while an image is acquired by scanning the plurality of primary charged particle beamlets 3 in the first direction using a collective multi-beam raster scanner 110. The platform movement and platform position are monitored and controlled by sensors known in the art, such as a laser interferometer, a grating interferometer, a confocal microlens array, or the like.

[0121] According to an embodiment of the present invention, a plurality of electrical signals are created and converted into digital image data and processed by the control unit 800. During an image scan, the control unit 800 is configured to trigger the image sensor 207 to detect a plurality of timely resolved intensity signals from the plurality of secondary electron beamlets 9 at predetermined time intervals, and digital images of image tiles are accumulated and stitched together from all scan positions of the plurality of primary charged particle beamlets 3. More details will be described below.

[0122] Figure 2 illustrating further aspects of the object illumination unit 100. Figure 1In addition to the elements described above, the object illumination unit 100 includes a set of deflectors, including at least a first deflector 350 and a second deflector 351, for deflecting the plurality of primary beamlets 3 from a first beam path along the primary beam path 3a to a second beam path along the primary beam path 3b, and includes a beam dump 130 arranged in the intersection plane 109. The beam dump 130 is arranged off-axis with respect to the central z-axis 105 and is configured to absorb the plurality of primary charged particle beamlets 3. At least the first deflector 350 is connected to the imaging control unit 820. When at least the first deflector 350 is in the off state, the plurality of primary beamlets 3 follow the beam path 3a through the beam dump 130 and are focused onto the sample surface by the objective lens 102. With at least the first deflector 350 in an on state, the plurality of primary beamlets 3 are selectively deflected and directed into a beam dump 130 near the beam crossing point 108 in the beam crossing plane 109 and offset from the z-axis by a distance d.

[0123] Figure 3 A method for inspecting a wafer by acquiring image tiles is described in more detail. A wafer, along with its wafer surface 25, is placed in the focal plane of a plurality of primary charged particle beamlets 3, with a center 21.1 of a first image tile 17.1. The predetermined positions of the image tiles 17.1...k correspond to inspection locations on the wafer for semiconductor feature inspection. The predetermined positions of the first inspection location 33 and the second inspection location 35 are loaded from an inspection file in a standard file format. The predetermined first inspection location 33 is divided into a plurality of image tiles, such as a first image tile 17.1 and a second image tile 17.2, and the first center position 21.1 of the first image tile 17.1 is aligned below the optical axis 105 of the multi-beam charged particle microscope 1 for the first image acquisition step of the inspection task. The first center of the first image tile 21.1 is selected as the origin of a first local wafer coordinate system for acquiring the first image tile 17.1. Methods for aligning the wafer 7 to register the wafer surface 25 and generate a local coordinate system of wafer coordinates are well known in the art.

[0124] A plurality of primary beamlets 3 are distributed in each image tile 17.1 ... k in a regular raster configuration and are scanned by a raster scanning mechanism to produce a digital image of the image tile. In this example, the plurality of primary charged particle beamlets 3 are arranged in a rectangular raster configuration, with N primary beam spots 5.11, 5.12 to 5.1N in a first row of N beam spots, and beam spots 5.11 to 5.MN in the Mth row. For simplicity, only M = five by N = five beam spots are shown, but the number of beam spots J = M by N can be larger, for example J = 61 beamlets, or approximately J = 100 beamlets or more, and the plurality of beam spots 5.11 to 5.MN can have different raster configurations, for example, a hexagonal or circular raster.

[0125] Each primary charged particle beamlet is scanned across the wafer surface 25, as shown in the example of a primary charged particle beamlet having beam spots 5.11 and 5.MN and scan paths 27.11 and 27.MN. For example, scanning of each of the plurality of primary charged particle beamlets is performed by moving back and forth along scan paths 27.11 ... 27.MN, and the collective multi-beam scanning deflector system 110 collectively moves each focal point 5.11 ... 5.MN of each primary charged particle beamlet in the x-direction starting from the start position of an image subfield line, which in this example is, for example, the leftmost image point of image subfield 31.MN. Each focal point 5.11 ... 5.MN is then centrally scanned by centrally scanning the primary charged particle beamlet 3 to the correct position, and the collective multi-beam raster scanner 110 then moves each of the plurality of charged particle beamlets 3 in parallel to the line start position of the next line in each corresponding subfield 31.11 ... 31.MN. The movement back to the line starting position of the next scan line is called fly-back. In one example, during fly-back, the plurality of primary beamlets 3 are selectively deflected by the first deflector 350 and directed to the beam dump 130. It will be understood that in this example, the primary charged particle beamlets 3 do not reach the sample surface 25.

[0126] Multiple primary charged particle beamlets 3 follow parallel scanning paths 27.11 to 27.MN, thereby acquiring multiple scanned images of each subfield 31.11 to 31.MN in parallel. For image acquisition, as previously described, multiple secondary electrons are emitted at focal points 5.11 to 5.MN, generating multiple secondary electron beamlets 9. These beamlets 9 are collected by objective lens 102, passed through a first collective multi-beam raster scanner 110, and directed to detection unit 200 for detection by image sensor 207. Sequential data streams for each of the multiple secondary electron beamlets 9 are synchronously transformed with the multiple 2D data sets within the scanning paths 27.11…27.MN, thereby forming digital image data for each image subfield 31.11 to 31.MN. Finally, the multiple digital images of the multiple image subfields 31.11 to 31.MN are stitched together by an image stitching unit to form a digital image of the first image tile 17.1. Each image subfield 31 . 11 to 31 . MN is constructed to have a small overlapping region with an adjacent image subfield, as shown by the overlapping region 39 of subfield 31 . mn and subfield 31 . m(n+1).

[0127] Obviously, the raster configuration of multiple primary charged particle beamlets is not limited to Figure 3 However, other grating configurations are possible, such as a one-dimensional grating configuration with a plurality of primary charged particle beamlets on a straight line, a circular grating configuration with a plurality of primary charged particle beamlets arranged on at least one circle, or as follows Figure 15 and Figure 16 Hexagonal grating configuration shown.

[0128] Next, we will describe various requirements or specifications for wafer inspection tasks. For high-throughput wafer inspection, the image acquisition time for each image tile 17.1 ... k (including the time required for image post-processing) must be fast. On the other hand, stringent image quality specifications, such as image resolution, image accuracy, and repeatability, must be maintained. For example, image resolution requirements are typically 2 nm or less, with high repeatability. Image accuracy is also referred to as image accuracy. For example, the edge position of a feature, or the absolute positional accuracy of the feature, is typically determined with high absolute accuracy. Typically, the positional accuracy requirement is approximately 50% of the resolution requirement or even lower. For example, measurement tasks require absolute accuracy of semiconductor feature dimensions below 1 nm, below 0.5 nm, or even 0.3 nm. Therefore, the lateral positional accuracy of each focus 5 of the multiple primary charged particle beamlets 3 must be less than 1 nm, for example, less than 0.5 nm or even less than 0.3 nm. High image repeatability requires that repeated image acquisitions of the same area produce first and second duplicate digital images, with the difference between the first and second duplicate digital images being below a predetermined threshold. For example, the difference in image distortion between the first and second repeated digital images must be less than 1 nm, such as 0.5 nm, or even preferably less than 0.3 nm, and the difference in image contrast must be less than 10%. This allows similar image results to be obtained even with repeated imaging operations. This is important, for example, for image acquisition and comparison of similar semiconductor structures in different wafer dies, or for comparing acquired images with representative images obtained from CAD data, databases, or image simulations of reference images.

[0129] One of the many requirements or specifications for wafer inspection tasks is throughput. The measurement area per acquisition time is determined by the dwell time, resolution, and number of beamlets. Typical dwell times range from 20 ns to 80 ns. Consequently, the pixel rate at the fast image sensor 207 is in the range of 12 MHz to 50 MHz, and approximately 15 to 20 image tiles or frames can be acquired per minute. For 100 beamlets in high-resolution mode with a pixel size of 0.5 nm, a typical throughput is approximately 0.045 sqmm / min (square millimeters per minute). For a larger number of beamlets, such as 10,000 beamlets and a dwell time of 25 ns, the throughput can exceed 7 sqmm / min. However, in prior art systems, the requirements for digital image processing significantly limit throughput. For example, digital compensation of scan distortion in prior art systems is very time-consuming and therefore undesirable. In embodiments of the present invention, the requirements for image post-processing are reduced, and the throughput of high-precision measurement tasks is increased. Embodiments of the present invention enable high throughput wafer inspection tasks with pixel scan speeds exceeding 2Tpixels / hour, for example, approximately 8 to 10Tpixels / hour, while maintaining image performance specifications well within the aforementioned requirements. For example, using embodiments of the present invention, the positional accuracy of each image pixel is maintained below 1nm.

[0130] The collective multi-beam raster scanner 110 typically exhibits nonlinear behavior, and the deflection angle α generated by the collective multi-beam raster scanner 110 is not linearly related to the voltage applied to the electrodes of the collective multi-beam raster scanner 110. Furthermore, the deflection angle α is different for each of the plurality of primary charged particle beamlets. The deviation in the deflection angle α increases as the angle of incidence β of the primary beamlets increases, and the scanning-induced distortion generated by the collective multi-beam raster scanner 110 also increases. In a first embodiment of the present invention, the scanning-induced distortion is reduced through modified voltage signals provided to the deflection scanners, a modified design of the deflection scanner control circuitry, and improved operation of the collective multi-beam raster scanner 110.

[0131] In the multi-beam system 1, multiple charged particle beamlets 3 are scanned in parallel with the same, collective multi-beam raster scanner 110, and the same voltage difference VSp(t) is applied to the deflection electrodes of each primary beamlet 3 according to the functional relationship between the deflection angle sin(α) and the scanning voltage difference VSp(t). Figure 4The distortion caused by scanning in the example of a selected primary beamlet 3 is illustrated, which enters the collective multi-beam raster scanner 110 at a propagation angle β. The collective multi-beam raster scanner 110 is represented by deflection electrodes 153.1 and 153.2 and a voltage source according to an embodiment, which provides a scanning voltage difference VSp(t). For simplicity, only the deflection scanner electrodes for raster scanning deflection in a first direction are illustrated. During use, a scanning deflection voltage difference VSp(t) is applied, and an electrostatic field is formed in the intersection volume 189 between the electrodes 153.1 and 153.2. In the case of a first voltage VSp(t1)=0V, the primary beamlet traverses the intersection volume 189 along the path 157a and is not deflected. The objective lens 102 forms a focus or beam spot at the center position 29.o of the image subfield 31.o. In the case of a second voltage VSp(t2) = V1, the primary beamlet passes through the intersection volume 189 along a path 157z and is deflected by an angle α1, thereby forming a beam spot at a position slightly offset from the ideal position 29.1 by the distortion vector dpz. In the case of a third voltage VSp(t3) = V2, the primary beamlet passes through the intersection volume 189 along a path 157f and is deflected by an angle α2, thereby forming a beam spot at a position slightly offset from the ideal position 29.2 by the distortion vector dpf. The subfield coordinates are given in relative coordinates (p, q) relative to the center point 29.o of the subfield 31.o. The subfield center coordinates X or 29.o are proportional to the angle of incidence β of the primary charged particle beamlet 157a.

[0132] The distortion vectors dpz and dpf depend on the angle of incidence β. Thus, a specific scan-induced subfield distortion pattern is obtained for each primary charged particle beamlet 3, and typically, the scan-induced subfield distortion pattern differs slightly for each primary beamlet and each image subfield. The difference in scan-induced subfield distortion can be of the order of a few nm, for example up to 2 nm or 3 nm.

[0133] Figure 17 A first embodiment is illustrated. According to the first embodiment, the nonlinearity of the deflector is minimized for a plurality of primary charged particle beamlets. In a first step S1, the subfield distortion pattern caused by scanning is determined, for example, by measuring a calibration pattern of a predetermined shape. Thus, linear voltage ramps VLp(t) and VLq(t) are applied to a collective deflection scanner to collectively scan the plurality of beamlets in a first direction, or p-direction, and a second direction, or q-direction, and the subfield distortion pattern caused by the plurality of scans is measured.

[0134] In the second step S2, the subfield distortion patterns caused by the multiple scans are statistically analyzed to generate a reference distortion pattern. In the first example, the reference distortion pattern is generated by averaging the subfield distortion patterns caused by the multiple scans. In the second example, a maximum threshold is used to optimize the reference distortion pattern, which will be described in more detail below.

[0135] In a third step S3, the reference distortion pattern is converted into a correction function C(p,q).

[0136] In an optional iteration step, the correction function C(p,q) is applied to the linear voltage ramps VLp(t) and VLq(t), and first correction voltage ramps VCp(t) and VCq(t) are obtained. Steps S1 and S2 are repeated using the first correction voltage ramps VCp(t) and VCq(t) to obtain a residual average distortion pattern. The iteration steps are repeated until the residual average distortion pattern does not exhibit any deviations greater than a threshold value (e.g., less than 0.3 nm).

[0137] In step S4, the correction function C(p,q) or the residual correction voltage ramps VCp(t) and VCq(t) are stored in a memory and provided to the scan control module, as described in the second embodiment of the present invention. Using the optimized correction function C(p,q) or the residual correction voltage ramps VCp(t) and VCq(t), a scan-induced distortion pattern with reduced maximum scan-induced distortion is generated for the plurality of primary charged particle beamlets 3. It will be appreciated that the scan-induced distortion pattern typically exhibits a two-dimensional distortion vector D = [dp, dq] (p, q) as a function of the local subfield coordinates p and q for each subfield, and that the voltage signal that generates the deflection to a particular (p, q) coordinate may depend on p and q.

[0138] In a second example of step S2, a reference distortion pattern is optimized using a maximum threshold. In this example, the reference distortion pattern is optimized from the subfield distortion patterns induced by multiple scans to minimize the distortion induced by the scans. The reference distortion pattern is subtracted from the subfield distortion patterns induced by the multiple scans, and the reference distortion pattern is changed whenever the residual distortion exceeds a predetermined threshold, such as 1 nm or 0.5 nm. In this example, a residual distortion below the threshold may not be achieved for all primary beamlets, and certain image subfields where the residual distortion exceeds the threshold are marked as exceeding the accuracy requirements of the metrology task.

[0139] Due to the above-mentioned processing, the distortion caused by the multi-beam scanning is minimized overall and reduced by a factor of 2 or more, for example. It should be noted that by minimizing the distortion caused by the multi-beam scanning, the sub-field distortion caused by the residual scanning of the surrounding primary beamlets with β>0 can be increased, for example, by the residual scanning of the axial primary beamlets with β=0. An example is shown in Figure 5 middle. Figure 5 a shows the maximum value of the subfield distortion caused by each scan for J=61 primary charged particle beamlets. Figure 5 b shows the residual maximum value of the subfield distortion caused by the scanning after correction using the correction function C(p,q). The maximum distortion value of subfield 31.15 has been significantly reduced, for example by a factor of 2, while the distortion caused by the scanning of the central beamlet 31.55 has been increased. Figure 6 An example of corresponding voltage differences applied to the collective deflection scanner 110 in dependence on the subfield coordinate p is illustrated. The linear dependency VL is corrected to the non-linear deflection voltage VC by C(p,q), thereby obtaining a minimal residual scanning-induced subfield distortion.

[0140] The subfield distortion induced by scanning the plurality of primary charged particle beamlets 3 arises not only from the nonlinear behavior of the collective deflection scanner 110, but also from distortions or other aberrations of other charged particle optical elements, such as the objective lens 102. The objective lens 102 can be a motorized or electrostatic optical immersion lens and can form the plurality of focal spots 5 on the surface 25 of the sample 7 using electrostatic and magnetic fields. In addition to the collective scanning deflector 110, these elements can also contribute to the scanning distortion. Using the method according to the first embodiment, the scanning-induced distortion is minimized overall, regardless of its origin.

[0141] According to a first embodiment, a method for calibrating a multi-beam charged particle scanning electron microscope 1 is provided, thereby reducing distortion caused by maximum scanning. The method comprises:

[0142] a first step of performing a calibration measurement by raster scanning the plurality of primary charged particle beamlets 3 with a first drive signal V1 (p, q) over the surface 25 of the calibration sample with a collective multi-beam raster scanner 110,

[0143] - a second step of deriving a plurality of scanning-induced distortion patterns from the calibration measurements, comprising deriving a scanning-induced distortion pattern for each of the plurality of primary charged particle beamlets 3,

[0144] - The third step is to analyze the subfield distortion patterns caused by multiple scans and derive a correction signal C(p,q),

[0145] - a fourth step of modifying the first drive signal V1(p,q) with the correction signal C(p,q) and deriving a modified drive signal V2(p,q) for driving the collective multi-beam raster scanner 110,

[0146] In one example, the third step of the analysis includes: deriving a reference distortion pattern of the distortion patterns induced by the plurality of scans by a statistical method, the statistical method including any calculation of an average, weighted average, or median; and deriving a correction signal C(p,q) from the reference distortion pattern. By applying a modified drive signal V2(p,q) to the collective multi-beam raster scanner 110, the reference distortion pattern is pre-compensated for each of the distortion patterns induced by the plurality of scans, and the distortion induced by the largest scan is minimized.

[0147] In one example, steps 1 to 4 are repeated until the maximum scan-induced distortion is minimized below a predetermined threshold. In one example, the maximum scan-induced distortion of most of the plurality of primary charged particle beamlets is reduced below the predetermined threshold, and the maximum scan-induced distortion of a few individual primary charged particle beamlets may exceed the predetermined threshold. Individual primary charged particle beamlets exceeding the predetermined threshold may be marked or labeled as exceeding the threshold. In one example, the marked primary charged particle beamlets are not used for high-precision metrology tasks, or digital compensation is applied to digital image data obtained using the marked primary charged particle beamlets.

[0148] In one example, the correction signal C(p,q) is stored in a memory of the control unit of the collective multi-beam raster scanner 110 to pre-compensate for the distortion caused by scanning according to a selected scanning procedure.

[0149] Using the method of the first embodiment, a correction function C(p,q) is generated by which the subfield distortion caused by scanning of the plurality of primary charged particle beamlets 3 is minimized. During use, the correction function C(p,q) is applied to the scanning voltage signal to generate, for example, a nonlinear voltage ramp. A second embodiment of the present invention illustrates an improved method or control of the collective deflection scanner 110 by the imaging control system 820, which enables the generation and application of pre-compensation voltages to be implemented on the collective multi-beam raster scanner 110 to achieve a high-speed raster scanning multi-beam charged particle microscope 1 with increased throughput and reduced scanning-induced distortion. Figure 7A second embodiment of a multi-beam scanning and image acquisition method 707 performed by a control system 820 is illustrated. The method for controlling the collective multi-beam raster scanner 110 by the imaging control module 820 includes a configuration step 710. In the configuration step 710, a scan program 762 is selected and configured and provided to a general scan processing step 720. In the general scan processing step 720, the scan program 762 is received by a bus receiver in a receiving step 722 and provided to a scan command processor, which derives a series of unit scan commands 764 in a scan command processing step 724. During the scan command processing step 724, the scan program 762 is segmented into unit point commands or line commands, which are described by a unit start point (ui, vi) and end point (ue, ve) of a line and a raster interval or step size (du, dv) for a plurality of points between the start point and end point. In one example, a line command is represented by k scan points at coordinates according to u(k+1)=u(k)+du (including u(k=0)=ui) and v(k+1)=v(k)+dv (including v(k=0)=vi. The sequence of unit scan commands 764 is derived in normalized or unit coordinates (u, v), which can be calculated and described, for example, with integer precision. In one example, the normalized or unit coordinates (u, v) represent pixel coordinates of a digital image for each of a plurality of image subfields. The unit scan command 764 is, for example, two synchronized streams of normalized subfield coordinates u and v. A short example of a unit scan command 764 is given by two synchronized streams of two subsequent scan coordinates [..., u(i-1), u(i), u(i+1), ...] and [..., v(i-1), v(i), v(i+1), ...].

[0150] The sequence of unit scan commands 764 is provided to the vertex post-processing step 726, where the sequence of unit scan commands 764 is transformed by a first-order group transform to adjust the scale and rotation of the sequence of unit scan commands 764. In an example of the vertex post-processing step 726, the unit normalized subfield coordinates u, v are scaled to the actual image subfield coordinates P, q with floating point precision. Scaling can generally be achieved by scaling and rotating the unit normalized subfield coordinates u, v to values proportional to the actual image subfield coordinates P, q. In one example, the actual image subfield coordinates (p, q) are obtained as follows:

[0151] p=A10*u+A01*v+A00

[0152] q=B10*u+B02*v+B00

[0153] where the scalar value A mn and B mnis determined in a calibration step and stored in a memory of the scan control unit. In an example of the vertex post-processing step 726, the distortion caused by the average or reference scan is pre-compensated by a correction signal C(p,q), which is determined and stored according to the first embodiment of the present invention and read out from the memory. In one example, a vector or stream of points p(i), q(i) on a curve is generated from a straight line with a start point and an end point using the correction signal C(p,q), thereby pre-compensating the distortion caused by a scan, and the actual scan line of the primary charged particle beamlet follows the straight line on the sample surface. In one example, the correction signal C(p,q) is represented by a power series expansion, and the straight line command with a start point and an end point is transformed into a polygonal line of, for example, a fifth order. In general, (p,q) depends on the power series expansion of (u,v) according to the following:

[0154] p=∑ m，n A mn u m v n

[0155] q=∑ m，n B mn u m v n

[0156] Based on the polyline, a stream or sequence of points p(i) and q(i) on the curve is generated. Thus, a sequence of pre-compensated digital scan coordinates corresponding to the subfield scan coordinates (p, q) modified by the correction signal C(p, q) is generated to minimize the distortion caused by the average scan of multiple image subfields. The output of the vertex post-processing step 726 is a sequence or stream of pre-compensated digital scan commands 766, [..., p(i-1), p(i), p(i+1), ...] and [..., q(i-1), q(i), q(i+1), ...].

[0157] In parallel with the sequence of unit scan commands 764 or pre-compensated digital scan commands 766, during the general scan processing step 720, for example, in the vertex post-processing step 726, synchronization control commands 768 are generated and provided to the scan synchronization control step 718, which controls the operation of the multi-beam charged-particle microscope 1, which is dependent on collective scanning, such as a beam deflection step performed by the beam deflector 350, or other elements configured to operate synchronously with the scanning process. In the scan synchronization control step 718, further scan control signals 768 may be generated and provided back to the general scan processing step 720. For example, during the scan synchronization control step 718, deflection of the plurality of primary charged-particle beamlets 3 is triggered, and the scan control signals 768 are provided to the general scan processing step 720 while the plurality of primary charged-particle beamlets 3 are deflected into the beam dump 130. In another example, a calibration step or system measurement step may be required during the execution of the scan procedure 762. The common scan control signal 768 triggers the general scan processing step 720 to interrupt the scan procedure 762 and insert an additional digital scan signal into the stream of pre-compensated digital scan commands 766. The command is configured, for example, to maintain the current scan position for a longer period of time or to scan a predetermined calibration target. In one example, during the scan command processing step 724, another unit scan command is inserted into the stream of unit scan commands 764, for example, by repeating the unit scan command multiple times and thereby providing a hold signal [... {u(i), u(i), u(i), u(i), ...} ...] and [... {v(i), v(i), v(i), v(i), ...} ...] within the stream of unit scan commands 764. The stream of unit scan commands 764 containing hold signals, [..., u(i-1), u(i), {u(i), u(i), u(i), u(i), ...}, u(i+1), ...] and [..., v(i-1), v(i), {v(i), v(i), v(i), v(i), ...}, v(i+1), ...], is then provided to other steps downstream in the data stream. Hold signals are generated and inserted until the general scan processing step 720 determines from the scan control commands 768 that the scan process 762 can continue and stops inserting hold signals. The synchronized scan control commands 768 provided by the scan synchronization control step 718 and the scan processing step 720 can utilize a common access data bus. The synchronized scan control commands 768 provided to the scan synchronization control step 718 and the scan control commands 769 provided by the scan synchronization control step 718 are described in more detail below using examples.

[0158] The scan processing step 720 provides the sequence of pre-compensated digital scan commands 766 to the specific scan deflection control step 730. In a first conversion step 732 of the specific scan deflection control step 730, the sequence of pre-compensated digital scan commands 766 is converted into digital drive signals 770 to drive each deflection electrode of the collective raster scan deflector 110. In one example, the collective raster scan deflector 110 comprises an octupole deflector having eight electrodes. In this example, during the first conversion step 732, a plurality of streams of eight digital drive signals 700 are generated, one for each of the eight electrodes, corresponding to the pre-compensated sub-field scan coordinates p(i), q(i) provided in the sequence of pre-compensated digital scan commands 766:

[0159] [p(i),q(i)]→[a(i),b(i),c(i),d(i),e(i),f(i),g(i),h(i)]=CM*[p(i),q(i)]

[0160] The first conversion step 732 converts the pre-compensated digital scan command 766 into a digital drive signal 770, for example, using a linear equation or a lookup table. An example linear equation is given above, including a 2x8 conversion matrix CM. The stream of eight digital drive signals 770 includes pre-compensation with the correction signal C(p,q), resulting in a stream of eight pre-compensated voltage signals 772 in the subsequent digital-to-analog conversion step 736a. In the case of a line scan, this stream of voltage signals 772 can be, for example, a voltage ramp with minimal deviation from linearity. An example is described in more detail below.

[0161] The sequence of streams of, for example, 8 non-linear voltage signals 772 from the digital-to-analog conversion step 736a is provided to an amplification step 740. During the amplification step 740, the sequence or stream of 8 voltage signals 772 is amplified, for example, by a set of 8 amplifiers. The resulting 8 amplified sequences of drive voltages 774 are provided to 8 electrodes of the collective multi-beam raster scanning deflector 110 for collective scanning deflection of the plurality of primary charged particle beamlets 3. In the collective deflection step 742, the plurality of primary charged particle beamlets 3 are collectively raster scanned according to the sequence of amplified drive voltages 774 corresponding to the sequence of scanning positions of the selected scanning program 762 within each of the plurality of image subfields. Using the aforementioned method steps, for each electrode of the multi-beam raster scanning deflector 110, for example, a signal such as Figure 6 The nonlinear voltage ramp VC(t) shown is provided to the respective electrodes, thereby achieving a collective raster scan of the plurality of primary charged particle beamlets with high flexibility and speed. In a similar manner as described above, an amplified sequence of drive voltages 774 is generated and provided to the second collective multi-beam raster scan deflector 222 to compensate for the residual scan of the plurality of secondary electrons 9.

[0162] Typically, the amplification step 740 exhibits nonlinearity during use of the amplifiers based on their nonlinear behavior. The nonlinearities of the eight amplifiers can be predetermined, and the eight digital nonlinear amplification transforms can be stored in memory. During the second vertex post-processing step 734, the nonlinearities of the eight amplifiers are pre-compensated by applying the eight digital nonlinear amplification transforms to the sequence of eight digital drive signals 770. Thus, the nonlinearity of each amplifier is compensated separately. The sequence of eight digitally corrected drive signals 776 is then provided to the digital-to-analog converter 736a, pre-compensating the amplifier nonlinearities.

[0163] During the second vertex post-processing step 734, a set of digital offsets 778 may also be generated and provided to the digital-to-analog conversion step 736b. The offset voltages 780 provided by the digital-to-analog conversion step 736b are filtered by the offset voltage filtering step 738. The sequence of, for example, eight nonlinear voltage signals from the digital-to-analog conversion step 736a is combined with at least one offset voltage 782 by the voltage combining step 641 and provided to the amplification step 740. The digital offsets and corresponding offset voltages 782 may be derived, for example, from a calibration step or adjustment step (not shown) of an adjustment or calibration of the multi-beam charged particle microscope system 1 and stored in a memory. Thus, an offset deflection or correction of the multiple primary charged particle beamlets 3 is achieved.

[0164] In one example, the set of digital offsets 778 provided to the digital-to-analog conversion step 736b may include the pattern center offsets of the scanning process 764. Figure 14As shown in the example of FIG, scanning procedure 764 may include multiple scan patterns, which may be separable, for example. Typically, the scanning procedure may be grouped into, for example, a first scan pattern having a first center coordinate and a second scan pattern having a second center coordinate. During the second vertex post-processing step 734, a stream of multiple, i.e., eight, digital drive signals 770 may be derived for scanning the first scan pattern relative to the first center coordinate, and a stream of multiple, i.e., eight, digital drive signals 770 may be derived for scanning the second scan pattern relative to the second center coordinate. During the provision of the stream of multiple, i.e., eight, digital drive signals 770 for scanning the first scan pattern to the digital-to-analog conversion step 736a, a set of digital offsets 778 representing the first center coordinate is provided to the digital-to-analog conversion step 736b. During the provision of the stream of multiple, i.e., eight, digital drive signals 770 for scanning the second scan pattern to the digital-to-analog conversion step 736a, a set of digital offsets 778 representing the second center coordinate is provided to the digital-to-analog conversion step 736b. By this method, the amplitude of the stream of multiple, ie, eight, digital drive signals 770 is reduced. Thus, by this method, the resolution of a particular scan deflection control step 730 can be improved.

[0165] In one example, the response of the voltage amplification step 740, as well as the response of the electrical connections and electrodes, depends on changes in the frequency of the voltage or voltage signal. For example, each amplifier may have low-pass filtering characteristics. Therefore, rapid and large changes in the drive voltage are subject to delays or overshoots in the drive voltage. For example, in a scanning pattern that follows a zigzag path (see below), the actual scan position may deviate from the designed scan position due to low-pass filtering of the drive voltage stream in the time domain. In one example, this deviation from the designed scan position is compensated during the second vertex post-processing step 734. In the second vertex post-processing step 734, the scanning program or scanning pattern is analyzed through time-frequency analysis of the sequence of voltage signals to be generated. The frequency response of the voltage amplifiers, electrical connections, and electrodes is taken into account and pre-compensated, for example, using inverse filtering techniques. In another example, the voltage change to be generated within a given time interval is calculated. For each given voltage change within each time interval, the frequency response of the voltage amplifiers, electrical connections, and electrodes is taken into account and pre-compensated by applying a corresponding correction to the sequence of digital drive signals 776. In another example, a predefined correction pattern is applied to the digital drive signal 776, wherein the predefined correction pattern is stored in memory and corresponds to a selected scan program 762. The frequency response or predetermined correction pattern can be predetermined during system calibration of the multi-beam charged-particle microscope 1 and stored in the memory of the imaging control module 820. In this example, a trigger signal (not shown) indicating the selected scan program 762 can be provided to the vertex post-processing step 734 during the general scan processing step 720, and the selection of the predetermined correction pattern can be initiated by the trigger signal.

[0166] In one example, during the second vertex post-processing step 734, a further drive signal is added to the eight digital drive signals provided by the conversion step 732. Thus, for example, other multi-beam parameters can be corrected in parallel with the collective scanning operation. This is particularly possible for collective multi-beam raster scanners 110 having more degrees of freedom required for collective raster scanning, such as the collective multi-beam raster scanner 110 with eight electrodes as described above.

[0167] Typically, it is preferred to operate at a constant accelerating voltage for the plurality of primary charged beamlets 3. The accelerating voltage for the plurality of primary charged beamlets 3 can, for example, be a constant accelerating voltage between 15 kV and 35 kV, preferably between 20 kV and 30 kV. However, in an example, the accelerating voltage for the plurality of secondary charged beamlets 3 can be selected or varied via the control unit 800. In an example of a variably adjustable accelerating voltage for the plurality of primary charged beamlets 3, the accelerating voltage for the plurality of primary charged beamlets 3 is increased or decreased, for example, by 10%, and an amplification gain corresponding to the selected or adjusted accelerating voltage is selected in the amplification step 740. In this example, a signal corresponding to the desired amplification gain is provided to the amplification step 740, and the amplification of the voltage signal used to deflect the plurality of primary charged beamlets 3 is adjusted according to the selected accelerating voltage for the plurality of primary charged beamlets 3. The accelerating voltage for the plurality of primary charged beamlets 3 can be adjusted by methods known in the art, for example, by adjusting the voltage applied to any accelerating electrode within the multi-beam charged particle microscope system 1.

[0168] The first general scan processing step 720 is further connected to the image data acquisition step 750. During the collective deflection step 742, the plurality of J primary charged particle beamlets 3 are parallelly irradiated onto the corresponding J image subfields of the image tile, thereby generating secondary electrons at the J focus, which form J secondary electron beamlets 9. The plurality of J secondary electron beamlets 9 are collected by the objective lens 102, pass through the collective multi-beam raster scanner 110 and the beam splitter or beam splitter unit 400 in a direction opposite to the propagation direction of the primary charged particle beamlets 3, and are imaged by the projection system 200 onto the detector 207. In the analog data collection step 748, for each of the J secondary beamlets, the secondary electrons are detected and converted into a plurality of J fluctuation voltages representing the amount of secondary electrons for each secondary electron beamlet 9 at each scan position within the plurality of J image subfields. Thus, a stream of J fluctuation voltages 786 is generated and provided to the image data acquisition step 750. In the first image analog-to-digital (AD) conversion step 752 of the image acquisition step 750, the stream of J fluctuating voltages 786 is converted into a stream of digital signals representing the actual values of the stream of J fluctuating voltages 786 at a constant AD conversion frequency and a fixed AD conversion time sequence. In the DA conversion step 736, the sequence of AD conversion times is synchronized with the DA conversion (by Figure 7 Synchronization is achieved by, for example, generating a clock signal 760 during the first general scan processing step 720 by a clock signal generator (see below) and providing the clock signal 760 from the first general scan processing step 720 to the second specific scan deflection control step 730 and the image data acquisition step 750. Figure 7In the example of FIG, the clock signal 760 is directly provided to the DA conversion step 736a and the AD conversion step 752. In one example, the 200 MHz clock signal 760 can be an integer multiple of the digital data stream frequency, such as 100 MHz, and in this case, is equal to the DA conversion frequency of the DA conversion step 736 and the AD conversion frequency during the AD conversion during step 752. The AD conversion during step 752 and the DA conversion during steps 736a and 736b can be synchronized at a time corresponding to every, every two, every three, or every nth signal of the clock signal 760, where n is an integer.

[0169] After the A / D conversion step, J digital sensor data streams 788 are provided to a digital image data selection step 754. Digital image data selection step 754 receives a selection control signal 744 from scan command processing step 720, indicating a scan procedure 762. Using selection control signal 744, during digital image data selection step 754, J digital image data value streams 790 are selected from the J digital sensor data streams 788, thereby skipping, for example, digital sensor data collected during interruptions in scan procedure 762 and corresponding to the aforementioned hold signal. Similarly, digital sensor data collected during scan signals, such as data corresponding to flybacks, is skipped. In one example, selection control signal 744 is equivalent to a sequence of unit scan commands 764. In one example, a sequence of tag signals is generated that provides a tag to each unit scan command in the stream of unit scan commands 764, corresponding to a scan point that contributes to a digital image pixel to be collected by image data acquisition step 750. During the digital image data selection step 754, only the digital sensor data corresponding to the marked scan commands are selected. A stream of J digital image data values 790 is provided to a digital image data addressing and writing step 756. In this step, the stream of J digital image data values 790 is written to a parallel access memory at a plurality of addresses corresponding to the sequence of unit scan commands 764, separated for each of the J image subfields. In a first example, the stream of J digital image data values 790 is pre-processed in the digital image data addressing and writing step 756 and sorted according to the sequence of pixels to be written to the sequence of memory addresses. In an alternative example, the stream of J digital image data values 790 is written directly to the memory addresses calculated from the unit scan commands 764.

[0170] In an example of digital image data addressing and writing step 756, a plurality of sequences of J memory address pointers corresponding to the plurality of J image data for the plurality of J image subfields are generated. In other words, each of the plurality of J image data corresponding to the plurality of J image subfields is written to a separate memory address, corresponding to the selected scan program, and the plurality of J image data is written to the parallel access memory at a plurality of J distinct and non-overlapping sequences of memory address pointers. The plurality of J distinct and non-overlapping sequences of memory address pointers are configured and allocated in the parallel access memory based on the desired number of image pixels, the scan program, and the number J of the plurality of J image subfields. In one example, only segments or portions of the plurality of J image data corresponding to the plurality of J image subfields are immediately written to the parallel access memory and are read out during the parallel reading and image processing step 758 before new segments of other portions of the plurality of J image data corresponding to the plurality of J image subfields are written to the same memory locations within the parallel access memory.

[0171] Using this method, different scanning procedures can be enabled, including scanning procedures with arbitrary or random scanning patterns, separate scanning paths, or scanning patterns with cascaded increased resolution. The different scanning procedures may include scanning patterns of the prior art, with scanning of multiple scan lines in the same direction. In this case, each of the multiple J image data is continuously written to the parallel access memory in a linear sequence synchronized with the scan time controlled by the scan clock interval. In another example of the scanning procedure, the scan lines are scanned in alternating directions, and every two of the multiple J memory address pointer sequences linearly decrease from the memory address corresponding to the line end coordinate. In another example of the separate scanning procedure, for example, if every two or every three scan lines are skipped, the multiple J memory address pointer sequences include memory address jumps or gaps. Several further examples are described below.

[0172] During the parallel readout and image processing step 758, the plurality of image pixel data 792 stored in the parallel access memory is read out and further processed. The reading of the plurality of image pixel data 792 may be performed in a different order than the writing of the stream of J digital image data values 790 during the addressing and writing step 756 and may be delayed relative to the writing of the stream of J digital image data values 790 during the addressing and writing step 756. The multi-beam scanning and image acquisition method 707 may provide a control signal to the control unit 800 to identify when a block of digital image data is achieved and completed during the digital image data addressing and writing step 756. The control unit 800 is configured to receive the control signal. During the parallel readout and image processing step 758, the control signal triggers the reading of the achieved or completed block of digital image data from the image frame memory. Examples of image processing may include at least one of image filtering, image registration, thresholding, object detection, image object size measurement, distortion compensation, contrast enhancement, deconvolution operations, or image correlation. The image processing may further include a stitching operation to form a single image file from a plurality of J image subfields generated by synchronously scanning a plurality of J charged particle beamlets 3 on the sample by the scanning and image acquisition method according to the second embodiment.

[0173] In the multi-beam scanning and image acquisition method 707 according to the second embodiment, a scanning program 762 is selected and provided in a configuration step 710. The multi-beam image scanning and acquisition method 707 is separated into a first general scan processing step 720, a second specific scan deflection control step 730, and a third image data acquisition step 750. The multi-beam scanning and image acquisition method 707 utilizes a common clock signal 760 to synchronize operations and to time and stream the data streams or sequences generated during the multi-beam scanning and image acquisition method 707. The common clock signal 760 is generated, for example, in the general scan processing step 720 and provided to the second and third steps 730 and 750. The parallel read and image step 758 can operate at different or the same clock frequencies. The control unit 800 can operate at different or the same clock frequencies.

[0174] In the multi-beam scanning and image acquisition method 707, for example, the DA conversion step 736 and the AD conversion step 752 are synchronized to provide synchronization of the raster scan deflection with the image pixel data collection. Delays in data processing after the AD conversion step 752, as well as delays in the amplification step 740 and the collective deflection step 742, can be calibrated and accounted for, for example, in the digital image data selection step 754. Thus, digital image pixel data corresponding to the actual scan position in the local subfield p,q coordinates is written to a memory address that corresponds to a unit scan command corresponding to the actual scan position in the p,q coordinates.

[0175] The first general scan processing step 720 includes at least one of a scan command receiving step 722, a scan command processing step 724, and a vertex post-processing step 726. From the scan program 762, the scan processing step 720 generates a sequence of pre-compensated digital scan commands 766, which are provided to the specific scan deflection control step 730. Each scan deflection control step 730 includes at least one of a specific conversion step 732, a vertex post-processing step 734, a digital-to-analog conversion step 736, and an amplification step 740. The third image data acquisition step 750 includes an A / D conversion step 752, a digital image data selection step 754, and a digital image data addressing and writing step 756.

[0176] The module separation in the first and second steps 720 and 730 has the advantage that each specific scan deflection control step 730 can be adjusted or calibrated for a different specific collective multi-beam raster scanning system 110 or second collective multi-beam raster scanning system 222 without having to adapt or modify the general scan processing step 720. Thus, the specific scan deflection control step 730 can be modified, for example, according to the needs of the specific collective multi-beam raster scanning system 110, depending on whether an octopole scanner or a quadrupole scanner sequence, a dipole scanner sequence, or another collective multi-beam raster scanning system 110 is used. By separating the modules in the first and second steps 720 and 730, the systematic aberrations of the multi-beam charged particle system are separated from the specific aberrations or nonlinear effects of the actual collective scanning system. The systematic aberrations of the multi-beam charged particle system 1 are pre-compensated in the first general scan processing step 720. Specific aberrations or nonlinear effects of an actual collective scanning system (e.g., the collective multi-beam raster scanner 110 or the second collective multi-beam raster scanner 222, respectively, including nonlinearities of voltage amplification during the amplification step 740) are pre-compensated in each specific scan deflection control step 730. In one example, the multi-beam scanning and image acquisition method 707 comprises at least a first scan deflection control step 730.1 for operational control of the collective multi-beam raster scanner 110 and a second scan deflection control step 730.2 for operational control of the second collective raster scanner 220. The multi-beam scanning and image acquisition method 707 may comprise further scan deflection control steps 730.3 to 730.n.

[0177] An advantage of module separation in the first and third steps 720 and 730 is that a complex scanning program 762 can be selected and configured for image acquisition of multiple J subfields, synchronized with the collective scanning operation of multiple J charged-particle beamlets. For example, as labeled unit scan commands are generated and provided to the image data acquisition step 750, the stream of J fluctuation voltages 786 from the sensor 207 is converted, selected, sorted, and written to a parallel access memory at a memory address corresponding to the unit scan coordinates. Using this approach, various scanning programs can be enabled, including those with arbitrary or random scan patterns, separate scan paths, or scan patterns with cascaded resolution enhancement. Some examples of scanning programs 762 are described in more detail below.

[0178] In one example, the selected first scanning procedure 762 can be interrupted, and a second scanning procedure can be selected during the interruption of the first scanning procedure 762. Thus, for example, calibration measurements or repeated measurements of multiple J image segments in multiple J image subfields can be performed, and characteristics such as drift of the multi-beam charged-particle microscope can be monitored. By repeatedly measuring multiple J image segments and monitoring image data corresponding to the multiple J image segments, stage movement or stage drift, image rotation, focus drift, or other changes in the multi-beam charged-particle microscope 1 can be detected. In another example, repeated collective scanning of multiple J primary charged-particle beamlets 3 is performed on multiple J image segments in multiple J image subfields, and the charging of the sample 7 is varied.

[0179] Using the method 707 according to the second embodiment, a general digital scan command 766 is generated in real time and provided in real time to the specific scan deflection control step 730 and the image acquisition step 750. During the specific scan deflection control step 730, a plurality of drive voltages 772 and 782 and an amplified drive voltage 774 are generated and provided in real time, for example, to the electrodes of the collective multi-beam raster scanner 110. Simultaneously, other digital control commands 768 are generated and provided in real time to at least the scan synchronization control step 718. According to a third embodiment of the present invention, the method according to the first or second embodiment is implemented in a hardware system comprising at least one field-programmable gate array (FPGA) or an equivalent system (e.g., an ASIC or a complex programmable logic device (CPLD)). These types of practical devices allow for streaming of digital data sequences at clock rates of approximately 1.5 GHz or higher. Figure 8The third embodiment is illustrated by the example of a modified imaging control module 820. The imaging control module 820 is connected to the operation control module 800. The connection between the operation control unit 800 and the imaging control unit 820 can be an Ethernet connection with a high data rate at least in the range of 1Gbit / s. The operation control module 800 includes a scanning program selection module 804, which is configured so that a scanning program 762 can be selected by user input or by other input devices. The scanning program 762 can also be automatically selected by a command provided with an inspection task, for example, provided together with the sample, or automatically selected by an external control command. Different scanning programs 762 can be defined, for example, by an external device and can be stored in the non-volatile memory 806 of the operation control module 800. As mentioned above, the selected scanning program 762 can be, for example, a zigzag scanning program, a zigzag scanning program or any other scanning program. More examples of different scanning programs 762 are described below. During use, a selected scan program 762 is selected from the memory 806 by the scan program selection module 804 and provided to the imaging control module 820 via an Ethernet connection, and within the imaging control module 820, provided to the scan control unit 930 and the image data acquisition unit 810. The scan control unit 930 of the imaging control module 820 is configured to receive and store the first selected scan program 762, and is configured to generate and provide a drive voltage 774 to, for example, the collective multi-beam raster scanning system 110, for example, by the method described in the second embodiment. In one example, the image data acquisition unit 810 is connected to the imaging detector 207 via a plurality of shielded data lines, and the image data acquisition unit 810 is configured to receive a plurality of J-fluctuation voltage streams 786 from the imaging detector 207 and to generate a set of 2D images from the image data streams synchronized with the scanning of the plurality of primary charged particle beamlets 3 according to the first selected scan program 762. Therefore, the image data acquisition unit 810 is connected to the scan control unit 930, and the scan control unit 930 includes a clock signal generator 938, which is configured to provide a clock signal for synchronizing the image data acquisition unit 810 with the scan control unit 930 during use. The image data acquisition unit 810 includes a parallel access memory 1816 and is configured to write a set of 2D images or 2D image data to the parallel access memory 1816 during use. The image data acquisition unit 810 and the synchronized operation of the image acquisition unit 810 are described in more detail below.

[0180] The scanning control unit 930 is connected to the second collective raster scanning system 222 of the projection system 205, to other systems 960, such as the deflector 350 (see Figure 2). The scan control unit 930 is further configured to generate a drive voltage 774 and to provide the drive voltage 774 to the second collective raster scanning system 222 of the projection system 205 and to other systems 960 that need to be controlled synchronously with the scanning operation according to the scanning program 762. An example of such other system is a deflector 350 that is configured to deflect the plurality of primary charged particle beamlets 3 to a position of the beam dump 130 during selective flyback of the plurality of primary charged particle beamlets 3 if the selected scanning program 762 includes flyback or other scanning commands of the plurality of primary charged particle beamlets 3 without parallel and synchronous image acquisition. In one example, the deflector 350 provides a trigger signal back to the scan control unit 930 during use, and the scan control unit 930 is configured to execute a controlled delay in the scanning program 762 that is synchronized with the collective deflection of the plurality of primary charged particle beamlets 3 into the beam dump 130 and back into the optical axis 105. In Figure 8 In the example shown, the scanning control unit 930 is configured to directly provide a driving voltage to the electrodes of the deflector 350. In an alternative example, the scanning control unit 930 may be connected to an external control unit of the beam deflector 350 and configured to provide a trigger signal to the external control unit of the beam deflector 350 for beam deflection by the beam deflector 350.

[0181] The imaging control module 820 further includes a power supply unit 925 configured to provide required voltages to the collective multi-beam raster scanner 110 , the collective deflection system 222 or other systems 960 during use.

[0182] The scanning control unit 930 is further connected to an auxiliary scanning system 950. In one example, the auxiliary scanning system 950 can be a correction system for separately correcting distortion caused by scanning, as described in the German patent application 102020209833.6 filed on August 5, 2020 cited above.

[0183] According to a third embodiment, a multi-beam charged-particle microscope 1 comprises:

[0184] at least one first collective raster scanner 110 for collectively scanning a plurality of J primary charged-particle beamlets 3 over a plurality of J image subfields 31 . 11 to 31 . MN ; and

[0185] a detection system 200 comprising a detector 207 for detecting a plurality of J secondary electron beamlets 9, each beamlet corresponding to one of the J image subfields 31.11 to 31.MN; and

[0186] - Imaging control module 820, the imaging control module 820 includes:

[0187] a scanning control unit 930 connected to the first collective raster scanner 110 and configured to control the scanning operation of the plurality J of primary charged particle beamlets 3 using the first collective raster scanner 110 according to a first selected scanning program 762 during use;

[0188] - an image data acquisition unit 810, which is connected to the scan control unit 930 and the detector 207, and is configured to acquire and select a plurality of J image data from the detector 207 during use, which is synchronized with a clock signal provided by the scan control unit 930, and is configured to write the plurality of J image data into the parallel access memory 1816 at a memory location according to the first selected scan program 762.

[0189] In one example, the scan control unit 930 is configured to generate a series of unit scan commands 764 according to a first selected scan program 762, and the image data acquisition unit 810 is configured to write a plurality of J image data to the parallel access memory 1816 at memory locations according to the sequence of the unit scan commands 764 according to the first selected scan program 762 during use. In one example, the scan control unit 930 includes a clock signal generator 938 configured to provide a clock signal to the scan control unit 930 and the image acquisition unit 810 during use. In one example, the scan control unit 930 may be further connected to at least one further system 960 configured to operate synchronously with the raster scanning operation. In one example, the at least one further system 960 is a collective deflector 350 configured to collectively deflect a plurality of J primary charged particles into the beam dump 130 during use. In one example, the multi-beam charged particle microscope 1 includes a second collective raster scanner 222 in the detection system 200, and the scanning control unit 930 is further connected to the second collective raster scanner 222. In one example, the imaging control module 820 further includes a voltage supplier 925, which is configured to provide voltage to the scanning control unit 930 and the image acquisition unit 810, and is configured to provide a driving voltage to the first collective raster scanner 110 or the second collective raster scanner 222 during use.

[0190] Figure 9The following illustrates further details of a scan control unit 930 according to a third embodiment. The scan control unit 930 includes a scan generator module 932, which is configured to perform the general scan processing steps 720 of the second embodiment during use. The scan generator module 932 is connected to the control unit 800. The control unit 800 is configured to provide a scan program 762 and a trigger signal during use to start or interrupt the operation of the scan generator module 932. The scan generator module 932 is further connected to a general purpose input and output interface (GPIO) 934, which is configured to exchange control data with the scan generator module 932. The control data may be, for example, information from a system calibration method described in the first embodiment, including, for example, pre-compensation C(p,q) for distortion caused by a reference or average scan.

[0191] The scan control unit 930 includes a clock unit 938. The clock unit 938 provides a clock signal 760 during use. The scan generator module 932 is connected to the clock unit 938 and is configured to synchronize the operation of the device controlled by the scan control unit 930. The clock signal may, for example, have a frequency in the range between 100 MHz and 400 MHz, preferably in the range between 150 MHz and 200 MHz.

[0192] The scan control unit 930 includes a plurality of scan signal amplifier modules 936.1 to 936.n, each module being configured to perform a specific scan deflection control step 730.1 to 730.n. The scan generator module 932 is connected to at least a first scan signal amplifier module 936.1 and a second scan signal amplifier module 936.2. The first scan signal amplifier module 936.1 is connected to electrodes, such as electrodes 153.1 and 153.2 of the first collective multi-beam raster scanner 110 of the primary charged particle beam path (see Figure 1 and Figure 4 ). Figure 9 1 illustrates a first collective multi-beam raster scanner 110, referred to as an octagonal octupole scanner. A second scanning signal amplifier module 936.2 is connected to the electrodes of the second collective multi-beam raster scanner 222 in the secondary electron beam path. The second collective multi-beam raster scanner 222 is illustrated as an octagonal octupole scanner. For ease of illustration, instead of eight shielded voltage lines between the first scanning signal amplifier module 936.1 and the eight electrodes of the first collective multi-beam raster scanner 110, a single line representing a high voltage connection 972 is shown. Each voltage connection, including high voltage connection 972, can be electrically shielded to minimize crosstalk.

[0193] Amplifier modules 936.1 or 936.2 are each configured to perform a specific scan deflection control step 730 according to the second embodiment. Thus, each amplifier module 936 includes a conversion unit for performing conversion step 732 and a vertex post-processing unit for performing vertex post-processing step 734; both units can be implemented in an FPGA. Each amplifier module also includes a digital-to-analog converter for converting the sequence of digital scan coordinates into an analog voltage. Each amplifier module 936.i also includes an amplifier connected to voltage source 925 for performing amplification step 740 during use. In one example, the eight amplifiers of first scan signal amplifier module 936.1 are typically configured to provide eight high voltages of approximately + / -80V, for example, + / -50V, to each of the eight electrodes of first collective multi-beam raster scanner 110. These amplifiers typically have a low-pass characteristic. Typically, the amplifier response of scan signal amplifier module 936.i is between 1 MHz and 10 MHz. As previously described, the low-pass characteristic of the amplifiers can be pre-compensated. Therefore, the vertex post-processing unit of the amplifier module used to perform vertex post-processing step 734 is configured to perform scan signal analysis and provide pre-compensation for the low-pass characteristic of first scan signal amplifier module 936.1. In one example, the low-pass characteristic is determined and stored, for example, in a memory of control unit 800, and control unit 800 adjusts the scan routine based on the low-pass characteristic of first scan signal amplifier module 936.1. For example, a delay time may be inserted into a selected scan routine to account for the low-frequency response of the large voltage variations generated by first scan signal amplifier module 936.1.

[0194] Scan generator module 932 is further connected to a third scan signal amplifier module 936.3 and further scan signal amplifier modules 936.i through 936.n. Typically, a plurality of at least first and second scan signal amplifier modules 936.1 ... 936.n are connected to a plurality of scan synchronization systems 960.1 ... 960.n, with the first collective multi-beam raster scanner 110 for the primary charged particle beam path and the second collective multi-beam raster scanner 222 for the secondary electron beam path being two examples of the plurality of scan synchronization systems 960.1 ... 960.n. Examples of scan synchronization systems 960.i may be a scanning multi-aperture array corrector 601 or 602, or a set of correction electrodes provided in a corrected collective multi-beam raster scanner 110, as described in German Patent Application No. 102020209833.6 filed on August 5, 2020, which is incorporated herein by reference. Additionally, for example, the scanning multi-aperture array corrector 601 or 602 may be controlled by an auxiliary scanning system 950 , which is digitally connected to the scan generator module 932 and configured to operate synchronously, for example, via the same clock signal provided by the clock unit 938 .

[0195] The scan control unit 930 is connected to a power supply unit 925. The power supply 925 is configured to supply power to the scan control unit 930 and a plurality of scan signal amplifier modules 936 during use. The power supply 925 is configured to provide low voltage support for the scan control unit 930 and the low voltage and low current portions of the scan signal amplifier modules 936, as well as high current signals required for deflection of, for example, scanning a plurality of primary charged particle beamlets using the first collective multi-beam raster scanner 110. In one example, as described above, the scan generator module 932 is further connected to the deflection unit 350.

[0196] The scan generator module 932 is further connected to the image data acquisition unit 810 and is configured to provide control signals to the image acquisition unit 810 during use, including a clock signal from the clock unit 938, to synchronize scan image acquisition according to the selected scan program 762. More details of the image data acquisition unit 810 are described below.

[0197] According to the third embodiment, the scanning control unit 930 includes:

[0198] - a scan generator module 932 connected to a clock unit 938; and

[0199] a first amplifier module 936 . 1 connected to the first collective raster scanner 110 ; and

[0200] a second amplifier module 936 . 2 connected to the second collective raster scanner 222 ;

[0201] - the scan generator module 932 is configured to generate and provide, during use, a series of pre-compensated digital scan commands to the first amplifier module 936.1 and the second amplifier module 936.2; and - the first amplifier module 936.1 is configured to generate, during use, at least a first amplification sequence of drive voltages to the electrodes of the first collective raster scanner 110;

[0202] - The second amplifier module 936.2 is configured to generate at least a second amplified sequence of drive voltages to the electrodes of the second collective raster scanner 222 during use.

[0203] In one example, the scan generator module 932 is further connected to the image acquisition module 810. In one example, the scan control unit 930 includes at least one further amplifier module 936.3 connected to a system 960.3 configured to operate synchronously with the raster scan deflection during use.

[0204] Figure 10More details of amplifier modules 936.1 ... 936.n are illustrated in the example of a first amplifier module 936.1. Amplifier module 936.1 includes a transformation and vertex post-processing unit 940, which is connected to scan generator module 932 via digital data connection line 974, which includes, for example, two data lines for two sequences of pre-compensated digital scan commands 766, one for each of the two scan directions p and q. Transformation and vertex post-processing unit 940 is further connected to a memory and control unit 942, which provides additional control signals for transformation and vertex post-processing, such as those determined in a previous calibration step of the multi-beam charged-particle microscope 1 and stored in a memory of the memory and control unit 942. Transformation and vertex post-processing unit 940 is configured to perform the transformation step 732 and the vertex post-processing step 734 during use. The conversion and vertex post-processing unit 940 is connected to a plurality of scanning digital-to-analog converters 946.1 to 946.8 (only two are shown) to provide a sequence of digitally corrected drive signals 776 generated by the conversion and vertex post-processing unit 940 during use. Each output line of each scanning digital-to-analog converter 946.1 to 946.8 is connected to a separate amplifier 948.1 to 948.8 (only two are shown), for example, the first output line of the first scanning digital-to-analog converter 946.1 is connected to amplifier 984.1. The amplifiers 948.1 to 948.8 are connected to a power supply 925 to amplify the eight nonlinear voltage signals 772 provided by the scanning digital-to-analog converters 946.1 to 946.8 into eight amplified sequences of drive voltages 774. The output line of each amplifier 948.1 to 948.8 is connected to eight electrodes of the octopole scanner 110 via eight high-voltage connections 972.1 to 972.8 (only two are shown).

[0205] The memory and control unit 942 can further provide a digital offset signal to an offset DAC 944, which in this example generates eight offset voltages that are provided to the input lines of amplifiers 948.1 through 948.8. Thus, the offset DAC 944 is connected to the input lines of amplifiers 948.1 through 948.8. The memory and control unit 942 is further connected to a general purpose input and output device (GIPO) 934 and is configured to receive control commands or control signals to be stored in the memory of the memory and control unit 942.

[0206] Scan generator module 932 can be implemented in a first FPGA. The conversion and vertex post-processing unit 940 and the memory and control unit 942 of the first amplifier module 936.1 can be implemented in a second FPGA. The corresponding conversion and vertex post-processing unit 940 and the corresponding memory and control unit 942 of the amplifier module 936.2 can be implemented in a third FPGA. Image data acquisition unit 810 includes units implemented in a fourth FPGA, as shown below.

[0207] According to the third embodiment, the multi-beam charged particle microscope 1 comprises at least two amplifier modules 936.1 and 936.2, wherein at least the first amplifier module 936.1 comprises:

[0208] a conversion and vertex post-processing unit 940 configured to convert, during use, at least two sequences of pre-compensated digital scan commands 766 into at least first and second sequences of digital corrected drive signals 776 , the conversion and vertex post-processing unit 940 being connected to the scan control unit 930 to receive the at least two pre-compensated digital scan commands 766 ;

[0209] at least one first and second scanning digital-to-analog converter 946.1 and 946.2 connected to the conversion and vertex post-processing unit 940 for receiving, during use, at least a first and a second sequence of digitally corrected drive signals 776 and for converting, during use, said at least first and second sequence of digitally corrected drive signals 776 into at least a first and a second non-linear voltage signal 772;

[0210] at least one first and second amplifier 948.1 and 948.2, each of which is connected to a power supply unit 925;

[0211] wherein the first amplifier 948.1 is connected to the first scan digital-to-analog converter 946.1 and is configured to convert the first non-linear voltage signal 772.1 into a high voltage output during use, and wherein the first amplifier 948.1 is further connected to a first electrode of the collective scan deflector 110 via a first high voltage connection 972.1;

[0212] -Wherein the second amplifier 948.2 is connected to the second scanning digital-to-analog converter 946.2 and is configured to convert the second non-linear voltage signal 772.2 into a high voltage output during use, and wherein the second amplifier 948.2 is further connected to the second electrode of the collective scanning deflector 110 via a second high voltage connection 972.2.

[0213] Figure 11 Having described the architecture and components of the image data acquisition unit 810 , further details of the components and operations during use of the image data acquisition unit 810 will be described below.

[0214] The image data acquisition unit 810 includes a digital-to-analog conversion module or ADC module 1808, which includes a plurality of analog-to-digital (AD) converters. In one example, the image sensor 207 includes a plurality of J photodiodes corresponding to the plurality of J secondary electron beamlets. Each of the J photodiodes, such as an avalanche photodiode (APD), is connected to a separate ADC. The image sensor 207 may also include an electron-to-photon converter, such as that described in DE 102018007455 B4, which is hereby incorporated by reference in its entirety.

[0215] The J photodiode generates a J fluctuating voltage, which corresponds to the electrons generated at the J image subfield. A plurality of amplifiers can be configured between the photodiode and the ADC module 1808. For each of the plurality of J imaging channels of each of the plurality of J secondary electron beamlets, the offset and gain can be adjusted separately in a calibration step, and a plurality of J calibrated and fluctuating sensor voltages can be generated. The details of the configuration and method are described in DE102018007455B4 cited above. The plurality of J calibrated sensor voltages are provided to the ADC unit 1808. The plurality of J AD converters of the ADC module 1808 are configured to generate a plurality of J digital data streams from the calibrated J sensor voltages. The ADC module 1808 is configured to generate and provide a plurality of J digital data streams with a constant ADC clock between 60 MHz and 400 MHz or even higher.

[0216] The multiple ADCs of ADC module 1808 are controlled by an ADC clock signal. ADC module 1808 is connected to scan control unit 930. In one example, the ADC clock signal is synchronized with a scan clock signal generated by clock unit 938 and provided to ADC module 1808 by scan control unit 930. In an alternative example, ADC module 1808 includes a separate clock signal generator configured to generate the ADC clock signal in response to control data provided by scan control unit 930.

[0217] Recorder 1810 is connected to ADC module 1808 and provides an interface to ADC module 1808. Recorder 1810 is configured to provide the multiple J digital data streams to acquisition module (ACQ) 1812. ACQ 1812 is configured to receive a scan clock signal generated by clock unit 938 and provided by scan control unit 930. Therefore, ACQ 1812 is connected to scan control unit 930. ACQ 1812 is configured to synchronize the multiple digital data streams with the scan clock signal and provide the multiple J synchronized digital data streams to image data sorter 1820. Image data sorter 1820 is connected to ACQ 1812 and configured to sort the multiple synchronized digital data streams and write them to parallel access memory 1816. Image data sorter 1820 includes a pixel averaging unit 1822, a line averaging unit 1824, and a pixel addressing unit 1826. The three units 1822, 1824, and 1826 can be configured sequentially or in parallel. Image data classifier 1820 is connected to scan control unit 930 and is configured to receive control data corresponding to the selected scan program 762. Pixel averaging unit 1822 and line averaging unit 1824 are configured to perform processing on the multiple digital data in each digital data stream in response to control data provided by scan control unit 930. This processing may be, for example, averaging or calculating differences. Pixel addressing unit 1826 is configured to receive multiple J-averaged digital data streams from pixel averaging unit 1822 and line averaging unit 1824 and derive and assign pixel addresses to each digital data. Some examples of pixel or line averaging are described in more detail below. Pixel addressing unit 1826 is connected to frame capture memory or parallel access memory 1816 and is configured to write each digital data value in the J digital data stream to the parallel access memory 1816 corresponding to the pixel address, based on the selected scan program 762.

[0218] In one example, the clock frequency provided by scan clock unit 938 or the clock frequency of ADC unit 1808 is adjustable. By, for example, adjusting the scan clock frequency by scan clock unit 938, the dwell time of each of the plurality of image pixels can be variably adjusted. In one example, the scan clock frequency of scan clock unit 938 is 200 MHz, and each scan position is maintained by scan control unit 930 with H=2 clock signals. Therefore, the dwell time at each scan position is 10 ns. As the hold interval changes from H=2 to H=3, the dwell time increases to 15 ns. Fine-tuning of the dwell time can be achieved by adjusting the scan clock frequency generated by scan clock unit 938 from 200 MHz to, for example, 180 MHz, with H=2 having a dwell time of approximately 11 ns, or from 200 MHz to, for example, 220 MHz, with H=3 having a dwell time of approximately 13.6 ns. By fine-tuning the scan clock frequency and optionally averaging the pixel count by the pixel averaging unit 1822, the effective pixel dwell time can be adjusted and a predetermined signal-to-noise ratio (SNR) can be achieved.

[0219] Frame grabber memory column 1816 is connected for parallel readout to control unit 800, which is configured to read out a plurality of J digital images corresponding to J image subfields 31.11 to 31.MN (see Figure 3 The image stitching unit 812 of the control unit 800 is configured to stitch the J digital image into a digital image file corresponding to an image tile (e.g., image tile 17.k). The image stitching unit 812 is connected to the image data processor and the output 814, which is configured to extract information from the digital image file and to write the digital image file to a memory or to provide information from the digital image file to a display.

[0220] According to the third embodiment, the image data acquisition unit 810 includes:

[0221] an ADC module 1808 comprising a plurality of AD converters connected to the image sensor 207 and configured to convert, during use, the plurality of J fluctuating voltages 786 into a plurality of J digital sensor data streams 788 ; and

[0222] an acquisition control unit 1812 connected to the ADC module 1808 and the scan control unit 930 and configured to select, during use, from the plurality of J digital sensor data streams 788 and from the plurality of J digital image data value streams 790 from a selection control signal 744 provided by the scan control unit 930 in accordance with a selected scan program 762 during use; and

[0223] - an image data classifier 1820 connected to the acquisition control unit 1812, the scanning control unit 930, and the parallel access memory 1816;

[0224] The image data classifier 1820 is configured to write a plurality of J digital image data value streams 790 into the parallel access memory 1816 at a plurality of memory addresses corresponding to scan positions of the plurality of J primary charged particle beamlets 3 according to the selected scanning program 762 during use.

[0225] In one example, the ADC module 1808 is connected to the clock unit 938 and is configured to receive a clock signal from the clock unit 938 during use and synchronize the operation of the plurality of AD converters to convert the plurality of J fluctuation voltages 786 into the plurality of J digital sensor data streams 788 during use. In one example, the clock unit 938 is connected to the control unit 800 and is configured to receive a control signal from the control unit 800 and is configured to change the clock frequency of the clock unit 938 during use. Thus, the clock signal and dwell time of the scanning imaging operation are changed during use of the multi-beam charged-particle microscope 1.

[0226] The imaging control module 820 of the multi-beam charged particle microscope 1 may include a plurality of L image data acquisition units 810.n, each comprising at least a first image data acquisition unit 810.1 and a second image data acquisition unit 810.2 arranged in parallel. Each of the image data acquisition units 810.n may be configured to receive sensor data from the image sensor 207 corresponding to a subset of S beamlets of the plurality of J primary charged particle beamlets and to generate a subset of S streams of digital image data values of the plurality of J streams of digital image data values. The number of S beamlets belonging to each of the L image data acquisition units 810.n may be the same, with S x L = J. The number of S may be, for example, between 6 and 10, such as S = 8. The number L of parallel image data acquisition units 810.n may be, for example, 10 to 100 or more, depending on the number of primary charged particle beamlets J. Through the modular concept of the imaging control module 820 , the number J of charged particle beamlets in the multi-beam charged particle microscope 1 can be increased by adding parallel image data acquisition units.

[0227] According to various embodiments, scanning operations and image acquisition can be performed by the multi-beam charged particle microscope 1 according to the selected scanning program 762. In one example, during a wafer inspection task, multiple inspection sites, such as inspection site 33 and inspection site 35 (see FIG. Figure 3) is the object of the inspection task. For the first inspection site 33, a first selected scanning program 762.1 can be selected, and for the second inspection site 35, a different second selected scanning program 762.2 can be selected. In one example, in a first measurement, the first inspection site 33 is inspected according to the first selected scanning program 762.1, and in a second measurement, the first inspection site 33 is inspected according to the third selected scanning program 762.3. During each image acquisition, a selected scanning program may be interrupted and continued, for example, for intermediate monitoring tasks. Utilizing the multi-beam scanning and image acquisition method 707 and the imaging control module 820, flexible scanning control and synchronized image acquisition using a multi-beam charged particle microscope can be achieved, thereby allowing operation according to multiple different scanning programs with increased throughput and high precision. The modular architecture of the imaging control module 820 allows the imaging control module 820 to be dedicated to the configuration of the multi-beam charged particle microscope 1, for example, for a simpler multi-beam charged particle microscope 1 without additional devices such as beam deflection or scanning correction, or a more complex multi-beam charged particle microscope 1, which includes several scanning compensators for compensating for aberrations caused by scanning, such as a multi-aperture corrector 601 or 602, and includes a beam deflector 350 including a beam dump 130.

[0228] Therefore, the multi-beam charged particle microscope 1 according to the present invention comprises:

[0229] - a multi-beam generator 300 for generating a plurality of primary charged particle beamlets;

[0230] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0231] - a detection unit 200 comprising a detector 207; and

[0232] - an imaging control module 820 , which includes a scanning control unit 930 and an image acquisition unit 810 , configured to scan and image the inspection site of the sample using the selected scanning program 762 ;

[0233] - the scanning control unit 930 comprises a universal scanning generator module 932 and at least one first amplifier module 936.1 for supplying at least one sequence of high voltages to the electrodes of the first collective raster scanner 110; and a second amplifier module 936.2 for supplying at least one sequence of high voltages to the electrodes of the second collective raster scanner 220,

[0234] The scanning control unit 930 is adapted to selectively include a third or further amplifier module 936.3 or 936.n for controlling a third or further operating unit 960.3 or 960.n to operate synchronously with the selected scanning program 762 during use.

[0235] In one example, the universal scan generator module 932 includes a vertex post-processing unit that is configured, during use, to pre-compensate for aberrations caused by systematic scanning of the multi-beam charged particle microscope 1. In one example, each amplifier module includes a vertex post-processing unit, a digital-to-analog converter, and an amplifier, so that, in each amplifier module, nonlinearities are pre-compensated individually for each operating unit 960.i when the amplifier module and the operating unit 960.i are working together to perform synchronous operations, such as the first stage and multi-beam raster scanner 110 or the second collective majority raster scanner 222.

[0236] Figures 12 to 15 An example of an image acquisition and scanning procedure according to a selected scanning procedure is illustrated in FIG. The example is illustrated at a single image subfield 31.mn and it is understood that a plurality of N primary charged particle beamlets 3 are simultaneously raster scanned by a collective multi-beam raster scanner 110 (see also FIG. Figure 3 ), and the plurality of N secondary electron beamlets 9 are simultaneously raster scanned by the collective multi-beam raster scanner 110 and the second collective multi-beam raster scanner 222 and focused on the detector array 207 to receive the plurality of N data streams.

[0237] Figure 12 a illustrates a first example of a scanning procedure 712 for a zigzag raster scanning pattern according to a conventional scanning procedure. The scanning procedure 712 comprises a series of line scans, each with a line start and line end at the same p-coordinate, and each line having a p-coordinate that is subsequently increased by a raster distance dq (e.g., dq=1 nm). Between the end of one line and the start of the next line, the plurality of primary charged particle beamlets are rapidly moved back to the starting position of the next line ("flyback").

[0238] Figure 12 b illustrates some of the generated signals received by the scanning control unit 930. The scanning voltage Up represents the voltage used for repeatedly collectively scanning and deflecting a plurality of primary charged particle beamlets in the p direction from the line start to the line end. The figure shows a collective raster scanning deflector having a pair of electrodes for scanning in the x or p direction, and a pair of electrodes for scanning in the y or q direction. For a collective raster scanner having, for example, eight electrodes, the corresponding voltages can therefore be derived. In each time interval Ai, a single scanning line is raster scanned in each image subfield 31.11 to 31.MN. The scanning voltage Uq represents the voltage used for collectively scanning and deflecting a plurality of primary charged particle beamlets in a step-like manner in the q direction. TD illustrates the trigger signal exchanged between the scanning control unit 930 and the deflector 350, and represents the time Ci during which a plurality of primary charged particle beamlets 3 are deflected into the beam dump 130 (see Figure 2TL is an example of a trigger signal provided to the image data acquisition unit 810. Each trigger signal triggers the collection of an image data stream corresponding to one scan line. The following example illustrates this operation.

[0239] At time t1, the next line position in the q direction of the plurality of charged particle beamlets 3 is reached, and the scan control unit 930 triggers the deflector 350 to deflect the plurality of primary charged particle beamlets 3 back from the beam dump 130. At t2, the deflector 350 is in the closed state, and each of the plurality of primary charged particle beamlets 3 reaches the substrate. The trigger TD is set to a low value, and the voltage ramp for scanning the deflection in the p direction is started for t2. At time t3, each of the plurality of primary charged particle beamlets 3 reaches the starting point of image acquisition for the next scan line. At t3, the trigger signal starts recording the digital values corresponding to the image pixels by the image data acquisition unit 810. The image data acquisition unit 810 selects a sequence of S digital image data corresponding to a predetermined number of pixels S in one scan line and writes it to the memory. The predetermined number of pixels S can be, for example, S = 8000 pixels, which are collected at a frequency of, for example, 100 MHz. In one example, the scan control unit 930 operates at a scan frequency that is a multiple of the collection frequency (e.g., 200 MHz), and the drive voltage Up increases for every two clock signals of the scan clock unit 938. Every two scan clock signals, the ADC unit 1808 collects multiple J analog voltages from the J photodetector in the sensor unit 207 and converts them into multiple J digital data values until a data value of S = 8000 is reached for 16,000 clock signals at time t4. At time t4, a trigger signal is provided to the deflector 350. At time t5, when the multiple primary charged particle beamlets 3 are deflected into the beam dump, the deflector 350 sets the trigger TD to a high value. At time t5, the scan control unit 930 triggers a retrace to the start position of the next scan line. Up decreases to the start position of a line in the p direction, and Uq increases to the next deflection voltage corresponding to the next line in the q position.

[0240] Figure 12 The illustrated voltage ramp typically consists of multiple small voltage steps corresponding to multiple scan positions. However, in one example, the focal point 5 of the plurality of J primary charged particle beamlets 3 is continuously moved over the object surface. In this example, the size and shape of the pixel is determined by the movement speed of the focal point 5 of the primary charged particle beamlets divided by the clock frequency used during the A / D conversion step 752 of the image data acquisition step 750.

[0241] Figure 18 This is a further illustration of a first operation of the charged particle microscope 1 using the first scanning procedure 712. Figure 3) is raster scanned with a primary charged particle beamlet and a digital image segment of an object surface region located in the object plane 101 is generated. According to a first scanning procedure 712, a raster scan is performed with horizontal lines 53 having line numbers q=1 to Q. The number of lines can be, for example, Q=8000. During the raster scan according to the scanning procedure 712, the focus 5 of the primary charged particle beam is directed to a series of adjacent pixel positions, which are shown as squares 55. The positions in each line 53 are arranged in columns with column numbers p=1 to p=P. The number P of positions 55 in each line 53 can be equal to the number Q of lines 53, so that, for example, Q=P=8000, but the value P can also be different from the number Q of lines 53.

[0242] In parallel access memory 1816, address region 61.mn is reserved or allocated by image data sorter 1820. Address locations 63 within address region 61.mn are represented by squares. Address region 61.mn may be one of many address regions 61 within parallel access memory 1816. Address locations 63 within address region 61.mn are represented by lines 65, numbered a=1 through A. The number A of lines 65 may be equal to the number Q. Each line 65 contains B memory locations, or columns, numbered b=1 through B. The number B of storage locations in each line 65 is equal to, for example, the number of image pixels P in each line 53.

[0243] According to the number S of primary charged particle beamlets assigned to one image data acquisition unit 810, the number of address areas 61.mn for each parallel access memory 1816 in the image data acquisition unit 810 is selected accordingly, and in each parallel access memory 1816 of each of the multiple L image data acquisition units 810, multiple S address areas 61.1 to 61.S can be allocated.

[0244] In the first operating mode, both raster scanning and image acquisition operate at the same clock frequency and are synchronized by clock signal 760. For example, at a pixel rate of 100 MHz and 8,000 pixels per scan line, a line scan frequency of up to 12.5 kHz can be achieved, corresponding to an image frequency of 1.56 Hz. Each detector signal corresponding to an image pixel 55 of a subfield 31.mn is transmitted and stored in a corresponding single memory location 55 of a corresponding address area 61.mn, as indicated by arrow 67. In the first operating mode, the number P of image pixels 55 within a line 53 is less than or equal to the number B of address locations 63 within the corresponding line of address area 61.mn (Q <= B), and the P detector signal corresponding to the image pixel 55 in a single line 53 is written to the address area 61.mn within a single address line 65, as indicated by arrow 69. The same applies to the number Q of lines 53, where Q <= A. Control unit 800 is then configured to read the digital image data from parallel access memory 1816 and, for example, perform image stitching.

[0245] like Figure 12 As shown in FIG. 1 b, time Ci can be variable and depends on the actual time required for the response of deflector 350. In this example, pixel averaging unit 1822 and line averaging unit 1824 are configured to not average. In another example, the collection frequency of ADC unit 1808 is set to a higher value than the scanning clock, and the multiple digital data streams generated by ADC unit 1808 are operated at, for example, 200 MHz, while the scanning is operated at 100 MHz. In this example, pixel averaging unit 1822 averages each pair of subsequent digital data values in the multiple digital data streams to form a digital pixel data stream provided to pixel addressing unit 1826.

[0246] For pixel or line averaging purposes, pixel averaging unit 1822 or line averaging unit 1824 can each include a separate memory unit. In another example, pixel averaging unit 1822 or line averaging unit 1824 utilizes a frame grabber memory column or parallel access memory 1816. Pixel averaging unit 1822 and pixel addressing unit 1826 then write the digital data values of the multiple digital data streams to a temporary memory location in the frame grabber memory column or parallel access memory 1816. The digital data values are then read from the temporary memory location in the frame grabber memory column 1816, thereby forming a pixel or line average. Further details are described in German Patent Application No. 102020102779.6, filed on February 4, 2020, which is incorporated herein by reference. In a first step, pixel averaging unit 1822 or line averaging unit 1824 writes a first set of digital data values to a first memory address in parallel access memory 1816 via pixel addressing unit 1826. In the second step, pixel averaging unit 1822 or line averaging unit 1824 receives the second set of digital data. In the third step, pixel averaging unit 1822 or line averaging unit 1824 reads the first set of digital values from the first memory address of parallel access memory 1816. In the fourth step, pixel averaging unit 1822 or line averaging unit 1824 processes the first and second sets of digital values and generates a third set of digital values. In the fifth step, the third set of digital values is written to the second memory address of parallel access memory 1816 via pixel addressing unit 1826. In one example, the second memory address is the same as the first memory address. In the second example, the first memory address is a temporary memory address. In these examples, the second memory address is a memory address corresponding to a pixel location within the digital image data.

[0247] An example of processing performed by pixel averaging unit 1822 or line averaging unit 1824 is averaging, where pixel averaging unit 1822 or line averaging unit 1824 generates an average of the first digital value and the second digital value. In another example, pixel averaging unit 1822 or line averaging unit 1824 can be configured and triggered by a control command to perform other processing methods on the first digital value and the second digital value. For example, the difference between the first digital value and the second digital value can be calculated. Thus, for example, an edge can be detected.

[0248] For pixel or line data processing according to the selected scan program 762, the pixel averaging unit 1822, the line averaging unit 1824, and the pixel addressing and memory allocation unit 1826 are configured to receive the selection control signal 744 from the general scan command processing step 724. Accordingly, the pixel addressing and memory allocation unit 1826 receives and allocates a temporary memory location, for example, in the parallel access memory 1816. The pixel averaging unit 1822 or the line averaging unit 1824 receives processing information, such as calculations of average values or differences between data values, according to the processing of the selected scan program 762.

[0249] Figure 19 A second operating mode of scanning and image acquisition of the charged particle microscope 1 is illustrated. The scanning procedure is similar to the first scanning procedure 712. The image subfield 31.mn again has Q lines 53 and P columns of image pixels 55. However, the dwell time at each pixel 55 is increased by a factor of two, and at each image pixel location, the image data acquisition unit 810 generates two digital data values. Figure 19 An example of the operation of pixel averaging unit 1822 is shown. Each time a primary charged particle beam is directed to a pixel location 55, two digital data values are generated and written to two adjacent memory locations within address area 61.nm of parallel access memory 1816 via pixel addressing unit 1826. During a first clock interval, the primary charged particle beam is directed to the first pixel location, and a first digital image value is generated by ADC module 1808 and written to the first memory location. During a second clock interval, the primary charged particle beam is still directed to the first pixel location, and a second digital image value is generated by ADC module 1808 and written to the second memory location. Thus, after each scan line 53, two adjacent memory lines 65 are populated with a sequence of digital image data values. This is also illustrated by arrow 69, indicating that each pixel line 53 is written to at least two memory address lines 65. In this example, the number A of memory address lines 65 must be twice the number Q of image pixel lines 53.

[0250] In one example, the control unit 800 can be configured to derive a digital image segment of the object surface from each address region 61.nm of the parallel access memory 1816, for example, by averaging each of the first and second digital image values from the first and second memory addresses. In another example, the pixel averaging unit 1822 performs the averaging internally, and the first and second digital image values are averaged and combined into one digital image value within the pixel averaging unit 1822 before being written to the memory address locations.

[0251] Figure 20An example of a third operating mode for scanning and image acquisition of the charged particle microscope 1 is illustrated. In the third operating mode, a modified scanning procedure 712.2 is applied, during which each pixel line 53 is raster scanned twice. The modified scanning procedure 712.2 is schematically illustrated. In one example, only the scanning procedure 712 may be repeated. During a first raster scan along the pixel line 53.1 in a first direction, a first number P of pixels are irradiated by the primary charged particle beamlet, and a first sequence of digital image data is generated and written to the address area 61.nm of the parallel access memory 1816 within the first line 65.1. During a subsequent second raster scan along the same pixel line 53.1 in the first direction, the first number P of pixels are again irradiated by the primary charged particle beamlet, and a second sequence of digital image data is generated and written to the address area 61.nm of the parallel access memory 1816 within the first line 65.2. In this example, each pixel position is irradiated at least twice by the primary charged particle beamlet. This is also shown by arrow 69, indicating that each pixel line 53 is written to at least two memory address lines 65. As in the second example, the number A of memory address lines 65 must be twice the number Q of image pixel lines 53.

[0252] In one example, the control unit 800 can be configured to derive a digital image segment of the object surface from each address region 61.nm of the parallel access memory 1816, for example, by averaging each of the first and second digital image values from the first and second memory locations. As previously described, other processing operations are also possible.

[0253] exist Figure 19 and Figure 20 In the example of , each pixel position on the object surface is raster scanned twice, and the address area 61 needs to be twice as large as Figure 18Memory locations of the example. Therefore, the sequence of digital data values generated by illuminating, detecting, and converting the data sequence corresponding to J subfield 31 may exceed the capacity of parallel access memory 1816. In one example, the raster scanning and image acquisition method can be divided into at least two stages. Thus, in a first stage, a first set of lines is raster scanned and a first set of digital image data is written to memory location 61 of memory 1816. The first set of digital image data is then read out by control unit 800. In a second stage, a second set of lines is raster scanned and the second set of digital image data is written to memory location 61 of memory 1816. The second set of digital image data is then read out by control unit 800 and processed, for example, by image stitching unit 812, along with the first set of digital image data. During the transition from the first stage to the second stage, the first set of digital image data is relocated from memory address area 61 to temporary storage in control unit 800. After relocation, address area 61 is available for the second set of digital image data. This process can also be performed continuously, with a greater number of stages.

[0254] In an example scanning sequence, the scanning operations of each subsequent scan line, such as the first scanning sequence 712, are repeated in reverse order, and each line of each image subfield is subsequently scanned twice in both the positive and negative p-directions. In this example, a line averaging unit 1824 is provided with a trigger signal to perform line averaging of the two subsequent digital data value streams and generate a plurality of streams of pixel data values representing image pixel values for a line within the digital image of each subfield. By averaging the line scan signals generated in opposite scan directions and averaging them by the line averaging unit 1824, a symmetrical image signal, such as a line or edge of a semiconductor feature, is generated. In another example, a first signal from a first scan of a line in a first positive direction and a second signal from a second scan of a line in a second negative direction are subtracted to detect and enhance edges.

[0255] In the example shown, the first scan and the second scan of the first pixel line 53.1 are performed sequentially, and the address lines 65.1 and 65.2 are arranged sequentially, but this is not necessarily the case, and other configurations are possible. For example, a first digital data sequence according to a first scan operation can be written to a first memory area 61.mn.1, and a second digital data sequence according to a second scan operation can be written to a second memory area 61.mn.2 (not shown), which is separate from the first memory area 61.mn.1. Other examples are shown in FIG. Figure 21 In this example, line averaging unit 1824 and pixel addressing unit 1826 are configured to write a first sequence of digital data values to temporary memory location 75. The digital data values collected and converted from pixel location 55.1 in first scan line 53.1 are assigned to temporary memory location 75. Arrow 67.1 illustrates this assignment. Figure 21 In the second step, shown in step b, a second sequence of digital data values is collected and converted from pixel position 55.2 in first scan line 53.1. Simultaneously, line averaging unit 1824 reads the first sequence of digital data values from temporary memory location 75 and performs processing 73 on the first and second digital data values to calculate a third digital data value. Finally, the third digital data value is written to address area 61.mn in parallel access memory 1816 (indicated by arrows 67.2 and 71). As described in the above example, address area 61.mn is assigned to subfield 31.mn. In another example, line averaging unit 1824 can also be configured to calculate an average value over the sequence of digital data values for a selected set of scan positions and can be configured to perform statistical analysis on the sequence of digital data values.

[0256] Explained in the example of a single primary charged particle beamlet Figures 19 to 21 It will be appreciated that in a multi-beam charged particle microscope 1 having a plurality of J primary charged particle beamlets 3, Figures 18 to 21 The diagram is applicable to a plurality of J primary charged particle beamlets.

[0257] In the example of the first scanning procedure 712, Figures 19 to 21 Different pixel or line averaging methods may be used, but examples of pixel or line averaging are not limited to the first scanning procedure 712. In the example, each primary charged particle beamlet is directed at each pixel location on the object surface at least twice, and the cumulative time intervals during which the primary charged particle beamlets are directed to the pixels are increased. The cumulative time intervals correspond to the dwell time.

[0258] In the above example, the dwell time is increased by a factor of two. However, using the pixel averaging unit 1822 or the line averaging unit 1824, the dwell time can be increased by even numbers. For example, to increase the dwell time by a factor of three, each pixel is irradiated three times with the primary charged particle beamlet, and the average of the three digital data values can be calculated equivalently to the above example.

[0259] The multi-beam scanning and image acquisition method 707 comprises, according to an operation mode including pixel averaging or line averaging units 1822, 1824, selecting a selected scanning procedure 762, such as scanning procedure 712 or 716, collectively deflecting J primary charged particle beamlets 3 on a plurality of J image subfields on the surface 25 of the sample 7 according to the selected scanning procedure 762, acquiring a J fluctuating voltage stream 786, converting the J fluctuating voltage stream 786 at a clock rate 760 to form a J digital image data value stream 790, processing at least two digital image data values from each of the J streams of digital image data values 790 to form a sum, an average digital image data value or a difference between digital image data values, and writing the J digital image data value stream 790 including the sum, the average digital image data value or the difference between digital image data values to a shared access memory 1816, at a predefined memory location. Figure 19 In the example of , each pixel position 55 within each of the plurality of image subfields 31 is illuminated by a primary charged particle beamlet with an extended dwell time. For example, at a sampling clock frequency of 100 Hz, the data sampling frequency corresponds to a sampling rate of 10 ms. In the above example, the dwell time is, for example, 20 ms, which is doubled compared to the inverse of the sampling clock frequency 760. In general, the dwell time may correspond to G multiplied by the inverse of the clock rate, where G is an integer of G=2, 3, 4, or a larger integer. Figure 19 In the example of the embodiment, the scanning procedure 762 includes a scanning pattern, such as the scanning pattern 712, 714, 716.1, 716.2 or 716.3, and further includes repeatedly scanning illumination of each pixel position 55 using at least one of the scanning patterns 712, 714, 716.1, 716.2 or 716.3, wherein each of the plurality of J primary charged beamlets 3 is in each of the plurality of image subfields 31. Using this method, the signal to noise ratio (SNR) of the digital image data collected from the surface area of the object can be increased. Using the third example of the scanning procedure 716 described below (see Figure 14 ), it is also possible to increase the cumulative dwell time for only one selected scanning pattern from among the plurality of scanning patterns 716.1 to 716.3, for example, scanning pattern 716.3, within each subfield 31.mn. This allows, for example, to combine images of selected areas with higher SNRs to obtain an overview image of the object surface area.

[0260] Figure 13 a illustrates a second example of a scanning procedure 714. In this example, each scanning path follows a zigzag shaped path, as shown for an image subfield 31.mn. Figure 13b illustrates an example of corresponding signals. The second example of scanning procedure 714 is faster than the first scanning procedure 712 and does not require the beam deflector 350 to deflect the beam into the beam dump 130 during the flyback period. At time t6, corresponding to the line scan from the end position to the start position, new image data acquisition by the image data acquisition unit 810 begins during the falling phase of Up. The trigger signal TL provided to the image data acquisition unit 810 is set to a second value, e.g., a low trigger signal, and the pixel addressing unit 1826 provides pixel addresses in reverse order to account for the reverse scan direction. After a time interval Ai (e.g., S = 8000 collected pixels), the trigger TL is set to zero to stop image acquisition. During time Ci, the scanning process resumes and steps into the next scan line by switching Uq to the next deflection voltage to deflect the multiple J primary charged particle beamlets into the J next scan line in each of the J image subfields. At time t7, the start position of the next scan line is reached, and the scan voltage ramp Up for scanning in the positive p direction to the end of the line is initiated. The trigger TL switches to a first trigger signal level, which in this example is a high value, and the pixel addressing unit 1826 provides pixel addresses in a linear sequence to take into account the scanning direction in the positive p-coordinate.

[0261] Figure 14 A third example of scanning procedure 716 is illustrated in the example of image subfield 31.mn. This third example of scanning procedure 716 includes multiple scan patterns in each image subfield, such as a first scan pattern 716.1, a second scan pattern 716.2, and a third scan pattern 716.3. Scan patterns 716.1 through 716.3 are configured separately from each other. Scan pattern 716.3 is rotated relative to the pq coordinate system. The spacing between scan patterns 716.1 through 716.3 can be configured to minimize or mitigate charging effects on the sample. Rotation of scan pattern 716.3 can be used to measure structures of interest in different scanning directions. For example, semiconductor structures are typically configured in the horizontal (H) or vertical (V) direction. In such cases, performing additional scans in directions deviating from the HV direction may be advantageous. In this example, the image stitching unit 812 can be skipped, and the image data is used only for image data processing in the image data processor and output unit 814, such as line width measurement, line edge roughness measurement, and the like.

[0262] Using the aforementioned apparatus and method, a complex scanning program, such as the third example of a scanning program with scan patterns 716.1 through 716.3, can be implemented. In a first step, the scanning program is described by control unit 800. In a second step, the scanning program is divided into line or point commands by scan generator module 932 in a scan command processing step 724, as described above. The line or point commands are processed by vertex post-processing step 726 and provided to amplifier modules 936.1 through 936.n to perform a specific scan deflection control step 730. A plurality of drive voltages 972 are provided to, for example, electrodes of collective multi-beam raster scanners 110 and 222. Another trigger signal is provided to image data acquisition unit 810 for synchronous data collection at memory address values corresponding to unit pixel coordinates and for writing pixel data to parallel access memory 1816.

[0263] Other examples of scanning procedures may include interlaced scanning techniques that subsequently scan image subfields using separate scan lines. For example, in a first scanning pattern, every three rows, e.g., rows 1–4–7 and thereafter, are scanned using a scanning pattern similar to that of the first or second scanning procedures, and in a second scanning pattern, every next three rows, e.g., rows 2–5–8 and thereafter, are scanned using a scanning pattern similar to that of the first or second scanning procedures. In another example, the scanning procedure follows a raster scanning strategy of increasing resolution, e.g., starting with a first raster scanning pattern of the first 512 x 512 pixels of each image subfield, and a second raster scanning pattern of the second 512 x 512 pixels arranged between the first 512 x 512 pixels of the first raster scanning pattern. This operation is continued with a third raster scanning pattern of 1024 x 1024 pixels arranged between the first and second plurality of pixels of each image subfield, and additional raster scanning patterns, e.g., 2048 x 2048 pixels, are added until image acquisition is achieved at the desired resolution (e.g., 8000 x 8000 pixels) for each image subfield.

[0264] In another example of a scanning procedure, the scanning procedure includes a first small scanning pattern at a predetermined position in each image subfield, for example, similar to scanning pattern 716.3. The scanning procedure includes a second scanning pattern similar to, for example, the second scanning procedure 714. During the execution of the second scanning pattern, the second scanning pattern is interrupted at least once, and the first scanning pattern is repeated, for example, two or three times, and the drift of the multi-beam charged-particle microscope 1 is monitored. Based on the results of the repeated scanning operation and the image acquisition of the first scanning pattern, drift compensation of the active compensator can be triggered, or calibration or adjustment of the multi-beam charged-particle microscope 1 can be triggered. Methods and solutions for providing active drift control are disclosed, for example, in German Patent Application No. 102020206739.2 filed on May 28, 2020, which is incorporated herein by reference.

[0265] When only surface sections of the wafer surface need to be inspected, such as for metrology tasks, or what are often referred to as die-to-die comparisons or die-to-database comparisons, a scanning program with a scan pattern comprising at least a smaller area than subfield 31.mn offers a further advantage in this particular inspection task. In this inspection task, only individual sections of the wafer surface are inspected and compared, for example, to one another or to data derived from CAD data of an ideal wafer surface. Thus, the scanning and image acquisition method according to the present invention increases the throughput of wafer inspection tasks.

[0266] Figure 15 Another example of the scanning procedure 762 is described. In one example, the plurality of J primary charged particle beamlets 3 of the multi-beam charged particle microscope 1 are arranged in a hexagonal raster configuration, and the centers of the image subfields 29 are distributed at the hexagonal array. Figure 15 An example of 10 centers 29.ij of a primary charged particle beamlet 3 represented by a cross symbol is shown (beamlets not shown). Figure 3 As shown, the corresponding image subfield 31 is constructed as a rectangular image subfield. However, using the image data acquisition unit 810 and the multi-beam scanning and image acquisition method 707, and according to an embodiment of the present invention, it is also possible to raster scan a plurality of primary charged particle beamlets 3 in a hexagonal pattern 27h. In the hexagonal scanning procedure 762 according to the hexagonal pattern 27h, the scan line length in the p direction is variable and changes according to the q coordinate during the scanning of different lines. Therefore, the corresponding image subfield 31.ij that is raster scanned according to the hexagonal scanning pattern 27h has a hexagonal shape (only two of them are illustrated without the scan line 27h). Therefore, compared to the image subfield 31.ij according to the hexagonal scanning pattern 27h, the scan line length in the p direction is variable and changes according to the q coordinate during the scanning of different lines. Figure 3 With the rectangular scanning pattern, the maximum value of the scanning deflection of each primary charged particle 3 is reduced by about 10%, and the distortion caused by the third-order scanning is reduced by at least 20%, for example, 25% to 30%.

[0267] Therefore, a multi-beam charged particle microscope 1 according to one embodiment comprises:

[0268] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3 in a hexagonal grating configuration;

[0269] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0270] - a detection unit 200 comprising a detector 207; and

[0271] An imaging control module 820 , comprising a scanning control unit 930 and an image acquisition unit 810 , is configured to scan and image the inspection site of the sample through a plurality of J image subfields 31 , each of the J image subfields 31 having a hexagonal shape.

[0272] Figure 16 Description based on Figure 15 Example of operation of a multi-beam charged particle microscope 1 with a hexagonal grating configuration of a plurality of primary charged particle beamlets. Only seven center coordinates 29.ij of a plurality of image subfields 31 are shown with crosses. Each center coordinate 29.ij represents a focus 5 of a plurality of primary charged particle beamlets 3 when the collective multi-beam deflector 110 is in the off state. Figure 16 In a, it shows Figure 15 The hexagonal subfield 31.ij has a diameter D1 and is raster-scanned using a scanning pattern 27h according to a predefined scanning program. During operation of the multi-beam charged-particle microscope 1, the raster configuration or the primary charged-particle beamlets 3 may be rotated 37, for example, due to working distance or image plane adjustments or changes in the operating conditions of the objective lens 102 (which may be configured as a magnetic immersion lens). For conventional scanning operations with an unchanged scanning program 27h, the orientation of the scan lines will rotate by the same rotation angle 37, and the scanning operation will not be parallel to or perpendicular to the xy coordinate system, which is consistent with the coordinate system of the sample stage 500 and the sample surface 25. Scanning operations using the scanning pattern 27h using an unmodified scanning program will produce scan lines, such as scan lines along the p-axis, that are rotated by angle 37 relative to the x-axis. In this case, the imaging results of scanning imaging using the multi-beam charged-particle microscope 1 may depend on the working distance of the focus position. However, using the image data acquisition unit 810 and multi-beam scanning and image acquisition method 707 according to the above embodiment, the multi-beam charged-particle microscope 1 is configured to maintain a scanning direction parallel to, for example, the x-direction, even when the grating configuration having the center coordinates 29.ij and the pq coordinate system is rotated by an angle 37. In response to the rotation 37 of the grating configuration, the first scanning program having the scan pattern 27h is changed to a second scanning program 27h2, which covers a slightly larger area corresponding to a slightly larger image subfield 31 having a diameter D2 and scan lines parallel to the x-direction. By varying at least one of the length or number of scan lines between the first scanning program 27h and the second scanning program 27h2, the size of the plurality of J image subfields 31 is altered, and the image tile is covered by multiple image subfields, but at the expense of a slight increase in the overlap area 39. Furthermore, during the scan command processing step 724, the scanning program 762 includes a rotation of the sequence of unit scan commands 764 to compensate for the rotation of the scan coordinate systems p and q.

[0273] Variations in the scanning procedure or scanning pattern are not limited to hexagonal grating configurations, but can also be applied to one-dimensional grating configurations, circular grating configurations, or rectangular grating configurations, e.g. Figure 3 shown.

[0274] By means of the image data acquisition unit 810 and the multi-beam scanning and image acquisition method 707 according to an embodiment, the multi-beam charged particle microscope 1 is configured for performing image scanning with an arbitrary scanning orientation of the scan lines, for example parallel to the x-direction, parallel to the y-direction, or parallel to a direction at an arbitrary angle, thereby maintaining complete coverage of an image tile with a plurality of appropriately sized image subfields. For example, an xy coordinate system represents the orientation of a sample such as a wafer, and the orientation of the scan lines is oriented to the orientation of the wafer structure. This structure may, for example, comprise horizontal and vertical lines or edges (a so-called HV structure), and in some examples, the lines or edges are preferably raster scanned at a predefined, constant angle (e.g., perpendicular to the lines or edges). Thus, the multi-beam charged particle microscope 1 according to an embodiment comprises:

[0275] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3,

[0276] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0277] - a detection unit 200 comprising a detector 207; and

[0278] - objective lens 102;

[0279] - an imaging control module 820 comprising a scanning control unit 930 and an image acquisition unit 810, for controlling the scanning of the plurality of J primary charged particle beamlets 3 by configuring a plurality of scanning lines in the plurality of J image subfields 31, and for controlling the acquisition of image patches of the sample surface,

[0280] The imaging control module 820 is configured to change the orientation of the plurality of scan lines relative to the sample orientation and to change at least the length of the scan lines or the number of scan lines to change the size of the plurality of J image subfields 31 used to cover the image tile. In one example, the orientation of the plurality of scan lines is changed based on a rotation of the raster configuration of the plurality of J primary charged particle beamlets 3 by changing the operating conditions of the objective lens 102.

[0281] In all illustrations, the illustrated scanning procedure 762 is simplified by illustrating only a small number of scan lines. It will be appreciated that the number of image pixels may be, for example, 8000 x 8000 or more, encompassing 8000 scan lines in this example.

[0282] It will be clear from the description that combinations of examples and embodiments and various modifications are possible and can be applied similarly to the embodiments or examples. The charged particles of the primary beam can be, for example, electrons, but can also be other charged particles, such as helium ions. Secondary electrons include secondary electrons in a narrow sense, but also include any other secondary charged particles generated by the interaction of the primary charged particle beamlet with the sample, such as backscattered electrons generated by backscattered electrons, or secondary electrons of the second level. In another example, secondary ions can be collected instead of secondary electrons.

[0283] The present invention is described in the following terms:

[0284] Item 1: A method for calibrating a multi-beam charged particle scanning electron microscope 1, comprising:

[0285] - a first step of performing a calibration measurement by raster scanning the plurality of primary charged particle beamlets 3 with a collective multi-beam raster scanner 110 over the surface 25 of the calibration sample with a first drive signal V1 (p,q);

[0286] a second step of deriving a plurality of scanning-induced sub-field distortion patterns from the calibration measurements, comprising deriving a scanning-induced distortion pattern for each of the plurality of primary charged particle beamlets 3,

[0287] - A third step, analyzing the subfield distortion patterns caused by the plurality of scans and deriving a correction signal C(p,q);

[0288] - a fourth step of modifying the first drive signal V1 (p, q) with the correction signal C (p, q) and deriving a modified drive signal V2 (p, q) for driving the collective multi-beam raster scanner 110;

[0289] - This reduces the distortion caused by the maximum scan.

[0290] Clause 2: The method of clause 1, wherein the third analyzing step comprises:

[0291] - deriving a reference distortion pattern of the distortion patterns caused by the plurality of scans by a statistical method, the statistical method comprising calculation of an average value, a weighted average value, or a median value; and

[0292] - deriving the correction signal C(p,q) from the reference distortion pattern.

[0293] Clause 3: The method of clause 1 or 2, wherein the first to fourth steps are repeated until a sub-field distortion caused by a maximum scan of each of the plurality of primary charged particle beamlets 3 is minimized below a predetermined threshold.

[0294] Clause 4: The method of clause 1 or 2, wherein the distortion caused by the majority of maximum scans of the plurality of primary charged particle beamlets 3 falls below a predetermined threshold and the subfield distortion caused by the maximum scans of a few individual primary charged particle beamlets exceeds a predetermined threshold.

[0295] Item 5: The method of any one of items 1 to 4, wherein the correction signal C(p,q) is stored in a memory of a control unit of the collective multi-beam raster scanner 110 to pre-compensate for distortion caused by scanning according to a selected scanning procedure.

[0296] Item 6: A multi-beam scanning and image acquisition method 707 for controlling collective scanning of a plurality of J primary charged particle beamlets 3 over a plurality of J image subfields and acquiring a plurality of J digital image data corresponding to the plurality of J image subfields using a multi-beam charged particle microscope 1, comprising:

[0297] - a configuration step 710 for providing a plurality of scanning programs 762 and for selecting a selected scanning program;

[0298] a general scan processing step 720 in which the selected scan program 762 is received and at least a first sequence of pre-compensated digital scan commands 766 and selection control signals 744 are generated from the selected scan program 762;

[0299] a specific scan deflection control step 730, in which at least one first amplification sequence of drive voltages 744 is generated from the at least first sequence of pre-compensated digital scan commands 766 and in which the at least one first amplification sequence of drive voltages 744 is supplied to a collective deflection step 742 for collectively deflecting the plurality of J primary charged particle beamlets 3 over the plurality of J image subfields on the surface 25 of the sample 7;

[0300] - An image data acquisition step 750, in which the J fluctuating voltage stream 786 collected from the image sensor unit 207 in the analog data collection step 748 is converted and selected to form a J digital image data value stream 790, which has been written into the common access memory at multiple J memory locations to form the multiple J digital image data corresponding to the multiple J image subfields, whereby the selection and writing are controlled by the selection control signal 744 generated and provided by the general scanning processing step 720.

[0301] Clause 7: The method of clause 6, further comprising a parallel reading and image processing step 758, in which the plurality of J digital image data corresponding to the plurality of J image subfields are read from the common access memory and image processing is performed.

[0302] Item 8: A method as described in Item 7, wherein the image processing step includes one of the following: image filtering, image registration, threshold operation, object detection, size measurement of image objects, distortion compensation, contrast enhancement, deconvolution operation, or image association applied to each of the multiple J digital image data corresponding to the multiple J image subfields.

[0303] Clause 9: The method of clause 7, wherein the image processing step comprises a stitching operation to form a single digital image file from the plurality of J digital image data.

[0304] Item 10: A method as described in any of items 6 to 9, wherein during the general scan processing step 720, at least a first sequence of unit scan commands 764 are generated in normalized subfield coordinates (u, v), and the first sequence of unit scan commands 764 are converted into a sequence of pre-compensated digital scan commands 766 in image subfield coordinates (p, q) by applying an operation including rotation, scale change, or consideration of one of a predetermined correction function C(p, q).

[0305] Item 11: A method as described in Item 10, wherein the selection control signal 744 generated during the general scan processing step 720 includes a first sequence of unit scan commands 764, and during the image data acquisition step 750, the J digital image data value stream 790 is written to the common access memory at multiple J memory locations corresponding to the first sequence of unit scan commands 764.

[0306] Clause 12: The method of any of clauses 6 to 11, wherein the method further comprises a scan synchronization control step 718, and wherein during the general scan processing step 720, synchronization control commands 768 are exchanged with the scan synchronization control step 718.

[0307] Item 13: A method as described in any of items 6 to 12, wherein the plurality of J primary charged particle beamlets 3 are configured in a grating configuration which is rotated by a rotation angle relative to the orientation of a coordinate system of a stage or a sample mounted on the stage during use, and wherein in the general scan processing step 720, the sequence of the unit scan commands 764 is adjusted to compensate for the rotation of the grating configuration.

[0308] Item 14: A multi-beam charged particle microscope (1), comprising:

[0309] at least one first collective raster scanner 110 for collectively scanning a plurality of J primary charged-particle beamlets 3 over a plurality of J image subfields 31 . 11 to 31 . MN ; and

[0310] a detection system 200 comprising a detector 207 for detecting a plurality of J secondary electron beamlets 9, each corresponding to one of said J image subfields 31.11 to 31.MN; and

[0311] - Imaging control module 820, the imaging control module 820 includes:

[0312] a scanning control unit 930 connected to the first collective raster scanner 110 and configured to control the raster scanning operation of the plurality J of primary charged particle beamlets 3 using the first collective raster scanner 110 according to a first selected scanning program 762 during use;

[0313] - At least one image data acquisition unit 810, which is connected to the scanning control unit 930 and the detector 207, and is configured to acquire and select multiple S image data from the detector 207 during use to be synchronized with the clock signal provided by the scanning control unit 930, and is configured to write the multiple S image data at corresponding memory locations in the parallel access memory 1816 according to the first selected scanning program 762, where S<=J.

[0314] Item 15: The multi-beam charged particle microscope 1 as described in Item 14, wherein the scanning control unit 930 includes a clock signal generator 938, which is configured to provide the clock signal to the scanning control unit 930 and the image acquisition unit 810 during use.

[0315] Item 16: The multi-beam charged particle microscope 1 according to Item 15, wherein the scanning control unit 930 can be further connected to at least one further system 960, which is configured to operate synchronously with the raster scanning operation.

[0316] Item 17: The multi-beam charged particle microscope 1 of Item 16, wherein the at least one further system 960 may be a collective deflector 350 configured to collectively deflect the plurality of J primary charged particles into the beam dump 130 during use.

[0317] Clause 18: The multi-beam charged particle microscope 1 according to any one of clauses 14 to 16 further comprises a second collective raster scanner 222 in the detection system 200 , and the scanning control unit 930 is further connected to the second collective raster scanner 222 .

[0318] Item 19: A multi-beam charged particle microscope 1 as described in Item 18, wherein the imaging control module 820 further includes a voltage supplier 925, which is configured to provide voltage to the scanning control unit 930 and the image acquisition unit 810, and is configured to provide a driving voltage to the first collective grating scanner 110 or the second collective grating scanner 222 during use.

[0319] Clause 20: The multi-beam charged particle microscope 1 according to any one of clauses 14 to 19, wherein the scanning control unit 930 further comprises:

[0320] - a scan generator module 932 connected to a clock unit 938; and

[0321] a first amplifier module 936 . 1 connected to the first collective raster scanner 110 ; and

[0322] a second amplifier module 936 . 2 connected to the second collective raster scanner 222 ;

[0323] - the scan generator module 932 is configured to generate and provide a series of pre-compensated digital scan commands to the first amplifier module 936.1 and the second amplifier module 936.2 during use; and

[0324] the first amplifier module 936 . 1 being configured to generate, during use, at least a first amplified sequence of drive voltages to the electrodes of the first collective raster scanner 110 ;

[0325] - The second amplifier module 936.2 is configured to generate at least one second amplified sequence of drive voltages to the electrodes of the second collective raster scanner 222 during use.

[0326] Item 21: The multi-beam charged particle microscope 1 according to any one of items 14 to 20, wherein the scan generator module 932 is further connected to the image acquisition module 810.

[0327] Item 22: A multi-beam charged particle microscope 1 as described in any of items 14 to 21, wherein the scanning control unit 930 includes at least one further amplifier module 936.3, which is connected to a system 960.3 that is configured to operate synchronously with the raster scanning deflection during use.

[0328] Clause 23: The multi-beam charged particle microscope 1 according to any one of clauses 14 to 22, wherein the at least one image data acquisition unit 810, or each of the image data acquisition units 810, comprises:

[0329] an ADC module 1808 comprising a plurality of AD converters connected to the image sensor 207 and configured to convert the plurality of S undulating voltages 786 into a plurality of S digital sensor data streams 788 during use; and

[0330] an acquisition control unit 1812 connected to the ADC module 1808 and the scan control unit 930 and configured to select, during use, a plurality of S digital image data value streams 790 from the plurality of S digital sensor data streams 788 and from a selection control signal 744 provided by the scan control unit 930 in accordance with a selected scan program 762 during use; and

[0331] An image data classifier 1820 connected to the acquisition control unit 1812, the scan control unit 930, and the parallel access memory 1816;

[0332] The image data classifier 1820 is configured to write the plurality of S digital image data value streams 790 into the parallel access memory 1816 at a plurality of memory addresses corresponding to the scanning positions of the plurality of J primary charged particle beamlets 3 according to the selected scanning program 762 during use.

[0333] Item 24: A multi-beam charged particle microscope 1 as described in Item 23, wherein the ADC module 1808 can be connected to the clock unit 938 and is configured to receive a clock signal from the clock unit 938 during use and synchronize the operation of the multiple AD converters to convert the multiple S fluctuating voltages 786 into multiple S digital sensor data streams 788 during use.

[0334] Item 25: The multi-beam charged particle microscope 1 according to Item 24, wherein the clock unit 938 is connected to the control unit 800 and is configured to receive a control signal from the control unit 800 and is configured to change the clock frequency of the clock unit 938 during use.

[0335] Item 26: The multi-beam charged particle microscope 1 according to any one of Items 14 to 25, wherein the imaging control module 820 comprises a plurality of L image data acquisition units 810, wherein L is L=8, 10 or greater.

[0336] Item 27: The multi-beam charged-particle microscope 1 according to Item 26, wherein the number S is given by S=6, 8, 10 or 12.

[0337] Item 28: A multi-beam charged particle microscope (1), comprising:

[0338] - a multi-beam generator 300 for generating a plurality of primary charged particle beamlets;

[0339] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0340] - a detection unit 200 comprising a detector 207; and

[0341] - an imaging control module 820 , which includes a scanning control unit 930 and an image acquisition unit 810 , configured to scan and image the inspection site of the sample using the selected scanning program 762 ;

[0342] The scanning control unit 930 comprises a universal scanning generator module 932 and at least one first amplifier module 936.1 for supplying at least one sequence of high voltages to the electrodes of the first collective raster scanner 110; and a second amplifier module 936.2 for supplying at least one sequence of high voltages to the electrodes of the second collective raster scanner 220,

[0343] The scan control unit 930 is adapted to selectively include a third or further amplifier module 936.3 or 936.n for controlling a third or further operating unit 960.3 or 960.n to operate in synchronism with the selected scan program 762 during use.

[0344] Item 29: The multi-beam charged particle microscope 1 as described in Item 28, wherein the universal scan generator module 932 includes a vertex post-processing unit, which is configured to pre-compensate for aberrations caused by system scanning of the multi-beam charged particle microscope 1 during use.

[0345] Item 30: A multi-beam charged particle microscope 1 as described in Item 28, wherein each amplifier module includes a vertex post-processing unit, a digital-to-analog converter and an amplifier, so that in each amplifier module, each operating unit 960.i is pre-compensated separately for the nonlinearity of the operation of the amplifier module combined with the operating unit 960.i for synchronous operation.

[0346] Item 31: The multi-beam charged-particle microscope 1 according to Item 30, wherein the operating unit 960.i is the first collective multi-beam raster scanner 110 or the second collective multi-beam raster scanner 222.

[0347] Item 32: The multi-beam charged-particle microscope 1 is configured to perform any one of the methods described in items 1 to 13.

[0348] Item 33: A multi-beam charged particle microscope (1), comprising:

[0349] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3 in a hexagonal grating configuration;

[0350] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0351] - a detection unit 200 comprising a detector 207; and

[0352] An imaging control module 820 , comprising a scanning control unit 930 and an image acquisition unit 810 , is configured to scan and image the inspection site of the sample through a plurality of J image subfields 31 , each of the J image subfields 31 having a hexagonal shape.

[0353] Item 34: A multi-beam charged particle microscope (1), comprising:

[0354] - a multi-beam generator 300 for generating a plurality of J primary charged particle beamlets 3;

[0355] - a first collective raster scanner 110 and a second collective raster scanner 222; and

[0356] - a detection unit 200 comprising a detector 207; and

[0357] - objective lens 102,

[0358] - an imaging control module 820 comprising a scanning control unit 930 and an image acquisition unit 810, for controlling the scanning of the plurality of J primary charged particle beamlets 3, and for controlling the acquisition of image patches of the sample surface via a plurality of scanning lines arranged in a plurality of J image subfields 31,

[0359] The imaging control module 820 is configured to change the orientation of the plurality of scan lines relative to the sample orientation.

[0360] Item 35: The multi-beam charged particle microscope 1 as described in Item 34, wherein the imaging control module 820 is further configured to change at least one length of the scanning line or the number of scanning lines to change the size of the multiple J image subfields 31 used to cover the image block.

[0361] Item 36: The multi-beam charged particle microscope 1 as described in Item 34 or 35 further includes a control unit 800 for controlling the working conditions of the objective lens 102, and wherein the orientations of the multiple scanning lines change according to the grating configuration rotation of the multiple J primary charged particle beamlets 3 caused by changes in the working conditions of the objective lens 102.

[0362] Clause 37: A multi-beam scanning and image acquisition method 707 for controlling collective scanning of a plurality of J primary charged particle beamlets 3 over a plurality of J image subfields and acquiring a plurality of J digital image data corresponding to the plurality of J image subfields using a multi-beam charged particle microscope 1, comprising:

[0363] - Selection of a selected scanning program 762;

[0364] - collectively deflecting the plurality of J primary charged particle beamlets 3 on the plurality of J image subfields on the surface 25 of the sample 7 according to the selected scanning procedure 762;

[0365] - Get J fluctuation voltage stream 786;

[0366] - converting J fluctuating voltage stream 786 at clock rate 760 to form J digital image data value stream 790;

[0367] - processing at least two digital image data values from each of J digital image data value streams 790 to form a sum, an average digital image data value, or a difference of digital image data values;

[0368] - Writing the J digital image data value stream 790 comprising the sum, average digital image data value, or difference of digital image data values into the shared access memory 1816 at a predetermined memory location.

[0369] Clause 38: The method according to Clause 37 further comprises the following steps:

[0370] - generating at least a first sequence of pre-compensated digital scan commands 766 and selection control signals 744 from the selected scan program 762;

[0371] - generating at least a first amplified sequence of drive voltages 774 from the at least first sequence of pre-compensated digital scan commands 766;

[0372] - providing at least a first amplified sequence of drive voltages 774 to the collective deflection step 742 .

[0373] Item 39: A method as described in item 37 or 38, wherein the scanning procedure 762 includes scanning illumination of the pixel position 55, and the residence time of each of the plurality of J primary charged particle beamlets 3 within each of the plurality of J image subfields 31 corresponds to G multiplied by the inverse of the clock rate 760 applied in the step of converting the J fluctuating voltage stream 786 to form the J digital image data value stream 790, and wherein G is an integer of G = 2, 3, 4 or greater.

[0374] Item 40: A method as described in item 37 or 39, wherein the scanning procedure 762 includes a scanning pattern 712, 714, 716.1, 716.2 or 716.3, and further includes repeatedly scanning and illuminating each pixel position 55 using at least one of the scanning patterns 712, 714, 716.1, 716.2 or 716.3, wherein each of the multiple J primary charged beamlets 3 is in each of the multiple image subfields 31.

[0375] A list of reference numerals is provided below:

[0376] 1 Multi-beamlet charged particle microscope and detection system

[0377] 3. One or more primary charged particle beamlets forming a plurality of primary charged particle beamlets

[0378] 5 Primary charged particle beam or focus

[0379] 7 Objects, such as wafers

[0380] 9. Forming a plurality of secondary electron beamlets

[0381] 11 Secondary electron beam path

[0382] 13 Primary charged particle beam path

[0383] 15 Secondary charged particle image spots

[0384] 17 Image Tiles

[0385] 19 Overlapping areas of image tiles

[0386] 21 Center position of image tile

[0387] 25 chip surface

[0388] 27 Scanning path of the primary beamlet

[0389] 29 Center of image subfield

[0390] 31 One or more image subfields

[0391] 33 First inspection area

[0392] 35 Second inspection area

[0393] 37 Rotation of the grating configuration

[0394] 39 Overlapping area of subfield 31

[0395] 53 pixel line

[0396] 55 pixels or lighting point location

[0397] 61, 61.mn Address area assigned to a subfield, for example, subfield 31.mn

[0398] 63 Address Location

[0399] 65 address lines

[0400] 67 pixel allocation

[0401] 69 line allocation

[0402] 71 Temporary memory reconfiguration

[0403] 73 Processing Operations

[0404] 75 Temporary memory address

[0405] 100 Object Illumination Unit

[0406] 101 Object or Image or Focal Plane

[0407] 102 objective lens

[0408] 103 field lens group

[0409] 105 Optical axis of the multi-beamlet charged particle microscope system

[0410] 108 The First Cross

[0411] 109 Intersection Plane

[0412] 110 Collective multi-beam raster scanner

[0413] 130 beam dump

[0414] 153 Deflector Electrode

[0415] 157 Off-axis or field beamlets

[0416] 189 Crossover Volume of Traveling Beamlets

[0417] 200 detection units

[0418] 205 Projection System

[0419] 206 Electrostatic Lens

[0420] 207 Image Sensor

[0421] 208 Imaging Lens

[0422] 209 Imaging Lens

[0423] 210 Imaging Lens

[0424] 212 Second Cross

[0425] 214 pore size filter

[0426] 216 Active Components

[0427] 218 Third Deflection System

[0428] 220 multi-aperture corrector

[0429] 222 Second Deflection System

[0430] 300 Charged Particle Multi-Beamlet Generator

[0431] 301 Charged Particle Source

[0432] 303 Collimating Lens

[0433] 305 Primary Multi-Beamlet Forming Unit

[0434] 306 Active Multi-Aperture Plate

[0435] 307 First Scene Lens

[0436] 308 Second Field Lens

[0437] 309 Electron Beam

[0438] 311 Primary Electron Beamlet Spot

[0439] 321 Intermediate Image Surface

[0440] 350 First Deflector

[0441] 351 Second Deflector

[0442] 390 Beam Steering Perforated Plate

[0443] 400 beam splitter units

[0444] 420 Magnetic Components

[0445] 500 sample stage

[0446] 503 Sample Voltage Supply

[0447] 601 Scanning Distortion Compensator Array

[0448] 602 Telecentric Aberration Scanning Compensator Array

[0449] 641 Voltage Combiner

[0450] 707 Multi-beam scanning and image acquisition method

[0451] 710 Configuration Steps

[0452] First example of a 712 scanning procedure

[0453] 714 Second example of scanning procedure

[0454] 716 Third Example of Scanning Procedure

[0455] 718 Scan synchronization control steps

[0456] 720 General Scan Processing Steps

[0457] 722 Receiving Steps

[0458] 724 Scan command processing steps

[0459] 726 Vertex Post-Processing Steps

[0460] 730 Specific Scan Deflection Control Steps

[0461] 732 Conversion Steps

[0462] 734 Vertex Post-Processing Steps

[0463] 736 DA Conversion Steps

[0464] 738 Offset Voltage Filtering Steps

[0465] 740 Amplification Steps

[0466] 742 Collective Deflection Step

[0467] 744 Select control signal

[0468] 748 Simulation Data Collection Steps

[0469] 750 Image Data Acquisition Steps

[0470] 752 AD conversion steps

[0471] 754 Digital Image Data Selection

[0472] 756 Digital Image Data Addressing and Writing Steps

[0473] 758 Parallel reading and image processing steps

[0474] 760 clock signal

[0475] 762 Scanner

[0476] 764 Unit Scan Command Sequence

[0477] 766 Pre-compensation Digital Scan Command Sequence

[0478] 768 Synchronous Control Command

[0479] 770 digital drive signal

[0480] 772 Nonlinear Voltage Signal

[0481] 774 Amplification sequence of driving voltage

[0482] 776 Digital Correction Drive Signal Sequence

[0483] 778 digital offset

[0484] 780 Offset Voltage

[0485] 782 Filtered offset voltage

[0486] 784 Image Data Stream

[0487] 786 Fluctuating Voltage Stream

[0488] 788 Digital Sensor Data Streaming

[0489] 790 Digital Image Data Value Stream

[0490] 792 multiple image pixel data

[0491] 800 Control Unit

[0492] 804 Scanner selection module

[0493] 806 Non-volatile Memory

[0494] 810 Image Data Acquisition Unit

[0495] 812 Image Stitching Unit

[0496] 814 Image Data Processor and Output

[0497] 820 Imaging Control Module

[0498] 830 Primary Beam Path Control Module

[0499] 925 Power Supply Unit

[0500] 930 Scanning Control Unit

[0501] 932 Scan Generator Module

[0502] 934 General Purpose Input / Output (GPIO)

[0503] 936 Amplifier Module

[0504] 938 Clock

[0505] 940 Transformation and Vertex Post-Processing Unit

[0506] 942 Memory and Control Unit

[0507] 944 Offset DAC

[0508] 946 Scan DAC

[0509] 948 High Voltage Amplifier

[0510] 950 Auxiliary Scanning System

[0511] 960 Scan Synchronous System

[0512] 972 High Voltage Connection

[0513] 974 Digital Data Cable

[0514] 1808 ADC module

[0515] 1810 Recorder

[0516] 1812 Get Module (ACQ)

[0517] 1814 Test Mode

[0518] 1816 Frame Grabber Memory Column

[0519] 1820 Image Data Classifier

[0520] 1822 Pixel Averager

[0521] 1824 Line Averager

[0522] 1826 Pixel Addressing and Memory Allocation Unit

Claims

1. A multi-beam scanning and image acquisition method (707) for controlling the collective scanning of a plurality of J primary charged particle beamlets (3) on a plurality of J image subfields and acquiring a plurality of J digital image data corresponding to the plurality of J image subfields using a multi-beam charged particle microscope (1), the method comprising: - a configuration step (710) for providing a plurality of scanning programs and for selecting a selected scanning program (762); - a general scan processing step (720) in which the selected scan program (762) is received and at least a first sequence of pre-compensated digital scan commands (766) and selection control signals (744) are generated from the selected scan program (762); a specific scan deflection control step (730) in which at least a first amplification sequence of drive voltages (774) is generated from at least a first sequence of pre-compensated digital scan commands (766) and in which the at least first amplification sequence of drive voltages (774) is supplied to a collective deflection step (742) for collectively deflecting a plurality of J primary charged particle beamlets (3) over said plurality of J image subfields on the surface (25) of the sample (7); - an image data acquisition step (750) comprising an analog data collection step (748) during which a J fluctuating voltage stream (786) is collected from an image sensor unit (207), wherein in the image data acquisition step (750), the J fluctuating voltage stream (786) is converted into a J digital sensor data stream (788), and a J digital image data value stream (790) is selected from the J digital sensor data stream (788) using a selection control signal (744) generated and provided by the general scan processing step (720), and the J digital image data value stream (790) is written into at least one parallel access memory (1816) at a plurality of J memory locations (61) to form the plurality of J digital image data corresponding to the plurality of J image subfields.

2. The method of claim 1 , further comprising a parallel readout and image processing step (758) in which the plurality of J digital image data corresponding to the plurality of J image subfields are read out from a common access memory and image processing is performed.

3. The method of claim 2, wherein the image processing comprises one of: image filtering, image registration, a threshold operation, object detection, size measurement of image objects, distortion compensation, contrast enhancement, a deconvolution operation, or image correlation applied to each of the plurality of J digital image data corresponding to the plurality of J image subfields.

4. The method of claim 2 or 3, wherein the image processing further comprises a stitching operation to form a single digital image file from the plurality of digital image data.

5. A method as claimed in claim 2 or 3, wherein during the general scan processing step (720), at least a first sequence of unit scan commands (764) is generated in normalized subfield coordinates (u, v), and the first sequence of unit scan commands (764) is converted into a sequence of pre-compensated digital scan commands (766) in image subfield coordinates (p, q) by applying one of an operation including rotation, rescaling, or taking into account a predetermined correction function C(p, q).

6. The method of claim 5 , wherein the select control signal (744) generated during the common scan processing step (720) includes a first sequence of unit scan commands (764), and during the image data acquisition step (750), the stream (790) of J digital image data values is written to the common access memory at a plurality of J memory locations corresponding to the first sequence of unit scan commands (764).

7. The method of any one of claims 1 to 3, wherein the method further comprises a scan synchronization control step (718), and wherein the synchronization control command (768) is communicated to the scan synchronization control step (718) during the general scan processing step (720).

8. A method as claimed in any one of claims 1 to 3, wherein a plurality of J primary charged particle beamlets (3) are configured in a grating configuration which is rotated by a rotation angle relative to the orientation of a coordinate system of a gantry or a sample mounted on the gantry during use, and wherein in a general scan processing step (720), the sequence of unit scan commands (764) is adjusted to compensate for the rotation of the grating configuration.

9. A multi-beam charged particle microscope (1) configured to perform the method according to any one of claims 1 to 8.

Images