Method and system for electron diffraction

Abstract

The disclosure relates to methods and systems for electron diffraction. A method of generating an electron diffraction pattern includes directing an electron beam onto a sample and detecting electrons scattered from the sample by a particle detector having an array of pixels. The detecting includes, for each detected electron at the pixel, measuring an energy value proportional to the energy of the detected electron. The measured energy value of each detected electron and the identifier of the pixel from which the electron was detected are sent to a processing device. The processing device calculates an energy weighted contribution value from each measured energy value using an energy correlation function. The energy correlation function produces an energy weighted contribution value that varies with electron energy. An energy-weighted electron diffraction pattern is then generated using the pixel position associated with each pixel identifier and the energy-weighted contribution value.

Description

Technical Field

[0001] The present invention relates to a method and system for electron diffraction, and in particular to a method and system for energy-weighted electron diffraction. Background Art

[0002] The electrons incident on the sample interact with the atoms within the sample, causing some electrons to scatter away from the sample at varying angles. Some electrons are backscattered by atoms. When the sample is crystalline or polycrystalline, the electrons can diffract according to Bragg's law, which depends on the energy of the electrons, the angle of incidence relative to the crystal planes, and the spacing of the crystal planes. A particle detector can be used to detect electrons scattered from the sample, including diffracted electrons. The signal provided by the detector can be processed to produce an electron backscatter diffraction (EBSD) pattern that can be used to characterize the structure of the sample. The diffraction pattern is viewed as a series of lines forming a Kikuchi band, which provides information about the lattice and spatial orientation of the sample. The relationship between the diffraction patterns of the sample can be analyzed to determine further information about the sample, such as grain size and texture, phase distribution and strain.

[0003] Conventional electron particle detectors can be operated in a "counting" mode, where the number of electrons incident on a pixel across a time window is counted. Conventional particle detectors may also use a predetermined energy threshold, so that only electrons above the threshold are counted and electrons below the threshold are discarded. The electrons counted for each pixel are provided in a single frame readout of all pixels, which are then used for diffraction pattern processing. For example, the number of electrons counted can be mapped to a display value (such as a color or grayscale value), and an image is presented that maps the display value to the pixel location within the image.

[0004] In one example, U.S. Pat. No. 8,890,065 describes a device for detecting one or each of a Kikuchi diffraction pattern and a Kossel diffraction pattern. The device includes an electron column adapted to provide an electron beam toward a sample, the electron beam having an energy range of 2keV to 50keV. The device also includes a particle detector for receiving and counting electrons from the sample due to the interaction of the electron beam with the sample. The detector includes a pixel array and has a count rate capability of processing at least 1000 electrons per second for each pixel. The particle detector is adapted to provide electron energy filtering of the received electrons so as to count only the received particles representing the diffraction pattern.

[0005] The above-mentioned particle detectors operate cumulatively, where the number of electron impacts across a time frame is summed before being read out. Other particle detectors can operate in a data-driven mode, where electronic data is read out each time the detector records an electron impact. For example, European application No. EP3761016 describes a method for imaging a sample, the method comprising: repeatedly scanning a region of interest (ROI) of the sample by directing a charged particle beam to a plurality of impact points in the ROI and detecting particles scattered from the plurality of impact points; during the repeated scans, calculating the signal quality of each of the plurality of impact points, and after each single scan of the ROI, updating the signal quality of the ROI based on the signal quality of each impact point, wherein the signal quality of a particular impact point is calculated based on the detected particles scattered from the impact point; terminating the repeated scans in response to the signal quality of the ROI being higher than a threshold signal quality; and forming a structural image of the ROI based on the particles detected during the repeated scans.

[0006] In electron diffraction, higher energy electrons generally provide the most information about the sample. However, utilizing the full electron spectrum can also be beneficial, for example to improve signal-to-noise statistics. Conventional electron diffraction methods can apply a set energy threshold to only accept electrons above / below the threshold. However, this results in a loss of information from electrons with energies that do not meet the threshold. Summary of the invention

[0007] In a first aspect, there is a method of producing an electron diffraction pattern. The method includes directing an electron beam incident on a sample, and detecting electrons scattered from the sample by a particle detector having an array of pixels. The detection includes measuring, for each detected electron at a pixel, an energy value proportional to the energy of the detected electron. For example, the measured energy value may include measuring the time that the charge induced in the pixel struck by the electron is above a detector threshold level. The measured time is proportional to the induced charge in the pixel, and therefore also proportional to the actual energy of the detected electron measured by the particle detector.

[0008] The method also includes sending the measured energy value of each detected electron and an identifier of the pixel where the electron was detected to the processing device. These measured energy values ​​are sent each time the electron strikes the pixel array so that information about the energy of the electron is provided to the processing device in real time. The pixel identifier is an identifier from which the processing device can determine the position within the pixel array where the electron was detected. For example, the identifier can identify the pixel position within the pixel array where the electron was detected.

[0009] The method also includes calculating, by the processing device, an energy-weighted contribution from each measured energy value using an energy correlation function. The energy correlation function varies with the electron energy, and thus the contribution of each detected electron depends on its energy. In this way, the electron energy can be used as an additional discriminating parameter when generating the diffraction pattern.

[0010] The method also includes generating, by a processing device, an energy-weighted electron diffraction pattern using the pixel position associated with each pixel identifier and the energy-weighted contribution value.

[0011] Because the energy correlation function varies with the energy of the electron, the energy correlation function can increase or decrease the contribution of the detected electron to the generated electron diffraction pattern depending on the energy of the detected electron. For example, the energy correlation function can increase the contribution of the detected electrons with higher energy relative to the detected electrons with lower energy. The energy-weighted contribution value of the detected electron can be higher than the energy-weighted contribution energy value of all detected electrons with lower energy. The energy correlation function can produce an energy-weighted contribution value that increases with increasing energy value. For example, the energy correlation function can produce an energy-weighted contribution value that increases linearly with increasing electron energy value. The energy correlation function can also be user-settable.

[0012] As part of the method, the detecting step may also include, for each detected electron, measuring a detection time of the detected electron. The detection time may be the time when the electron arrives at the pixel, or may be the time when the signal is sent to a processing device. The sending step may also include sending the measured detection time of the detected electron to the processing device. The measured detection time may then be used for the generation of an energy-weighted electron diffraction pattern.

[0013] Generating an energy-weighted electron diffraction pattern using pixel positions and energy-weighted contributions can include generating an energy-weighted electron diffraction pattern comprising an array of image pixels arranged to coincide with an array of pixels of a particle detector. Each image pixel is displayed with a display value proportional to a combination of energy-weighted contributions of electrons detected to have struck the corresponding pixel of the particle detector. For example, the display value can be the sum of all energy-weighted contributions of electrons detected to have struck the corresponding pixel of the particle detector.

[0014] The generated energy-weighted electron diffraction pattern may be a color image and the displayed values ​​are values ​​on a color gradient, or the generated energy-weighted electron diffraction pattern may be a grayscale image and the displayed values ​​are intensities.

[0015] The quality of the generated energy-weighted electron diffraction pattern is affected by the background signal. To improve the quality, the method may further include performing background subtraction by detecting, by the particle detector, electrons received when the electron beam is not incident on the sample, and, for each image pixel, subtracting an amount proportional to the number of electrons counted for each pixel of the detector from the displayed value of the corresponding image pixel.

[0016] The method may also include applying an energy filter to the measured energy values ​​or to the energy weighted contribution values. For example, the energy filter may be one of a predetermined energy band, one or more electron energy ranges, or a predetermined energy threshold. Applying an energy filter may help reduce processing requirements by discarding unwanted results.

[0017] The method may further include displaying the generated energy-weighted electron diffraction pattern.

[0018] The method may also include interpreting the generated energy-weighted electron diffraction pattern to derive one or more structural characteristics of the sample. The one or more structural characteristics may include crystal orientation, crystal phase, phase composition, texture, and strain location.

[0019] In a second aspect, there is an electron diffraction system. The system includes an electron source configured to provide an electron beam to a sample holder (i.e., toward the sample holder). The sample holder is configured to hold a sample. The system also includes a particle detector configured to detect electrons scattered from the sample, and a processing device communicatively coupled to the particle detector. The processing device is configured to perform all of the method steps as described above.

[0020] In a third aspect, there is a computer program comprising computer program instructions that, when executed by a computer processing system, cause the computer processing system to perform a method for generating an energy-weighted electron diffraction pattern. The method includes receiving data comprising measured energy values ​​of detected electrons and identifiers of pixels at which the detected electrons are detected, wherein the data is obtained by directing an electron beam to be incident on a sample. The method also includes detecting electrons scattered from the sample by a particle detector having an array of pixels, wherein the detection includes measuring an energy value proportional to the energy of the detected electron for each detected electron at the pixel. The method also includes calculating an energy-weighted contribution value from each measured energy value using an energy correlation function, wherein the energy correlation function varies with the energy of the electron. The method also includes generating an energy-weighted electron diffraction pattern using the pixel position associated with each pixel identifier and the energy-weighted contribution value.

[0021] Optionally, the energy correlation function increases or decreases the contribution of the detected electrons to the generated energy-weighted electron diffraction pattern depending on the energy of the detected electrons. The energy correlation function may increase the contribution of detected electrons of higher energy relative to detected electrons of lower energy. The energy-weighted contribution of the detected electron may be higher than the energy-weighted contribution of all detected electrons of lower energy. The energy correlation function may produce an energy-weighted contribution that increases with increasing energy values. The energy correlation function may produce an energy-weighted contribution that increases linearly with increasing energy values. The energy correlation function may be user-settable.

[0022] Optionally, the detecting further comprises measuring, for each detected electron, a detection time of the detected electron, and the transmitting further comprises transmitting the measured detection time of the detected electron to a processing device.

[0023] Optionally, the measured energy value comprises measuring the time that the charge induced in a pixel struck by the electron is above a detector threshold level.

[0024] Optionally, generating an energy-weighted diffraction pattern using pixel positions and energy-weighted contribution values ​​includes generating a diffraction pattern comprising an array of image pixels arranged to coincide with a pixel array of a particle detector, wherein each image pixel is displayed with a display value proportional to a combination of the energy-weighted contribution values ​​of electrons detected to have struck a corresponding pixel of the particle detector.

[0025] Optionally, the energy-dependent electron diffraction pattern is a color image and the displayed values ​​are values ​​on a color gradient, or the energy-dependent electron diffraction pattern is a grayscale image and the displayed values ​​are intensities.

[0026] Optionally, the computer program instructions, when executed by a computer processing system, cause the computer processing system to perform background subtraction by detecting electrons received by the particle detector when the electron beam is not incident on the sample, and, for each image pixel, subtracting an amount proportional to the number of electrons counted for each pixel of the detector from a displayed value of the corresponding image pixel.

[0027] Optionally, the computer program instructions, when executed by a computer processing system, cause the computer processing system to apply an energy filter to the measured energy values ​​or energy-weighted contributions. The energy filter may be one of: a predetermined energy band, one or more electron energy ranges, or a predetermined energy threshold.

[0028] Optionally, the computer program instructions, when executed by a computer processing system, cause the computer processing system to display the generated energy-dependent electron diffraction pattern.

[0029] Optionally, the computer program instructions, when executed by a computer processing system, cause the computer processing system to interpret the generated energy-dependent electron diffraction pattern to derive one or more structural characteristics of the sample. The one or more structural characteristics may include crystal orientation, crystal phase, phase composition, texture, and strain location.

[0030] The computer program may comprise computer program instructions which, when executed by a computer processing system, cause the computer processing system to implement any and all aspects and variations of the method of the first aspect as described above, which are not repeated here for the sake of brevity.

[0031] In a fourth aspect, there is a computer readable medium having stored thereon a computer program as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order that the present invention may be more readily understood, reference will now be made, by way of example only, to the accompanying drawings, in which

[0033] In the figure:

[0034] Figure 1 An exemplary electron diffraction system is illustrated.

[0035] Figure 2 A comparison of the counting mode and the time over threshold (TOT) mode of a particle detector is illustrated.

[0036] Figure 3 An exemplary electron diffraction method is illustrated.

[0037] Figure 4 A histogram of electron energies measured in hourly periods is illustrated.

[0038] Figure 5 The effective contribution of energy-weighted contributions to an electron diffraction pattern is illustrated compared to contributions that are not energy-weighted.

[0039] Figure 6 An image of an exemplary energy-weighted electron diffraction pattern of a sample generated using the exemplary systems and methods is illustrated.

[0040] Figure 7 A comparison with contrast is illustrated, showing the effect of energy-weighted contributions compared to non-weighted contributions.

[0041] Figure 8 An exemplary off-line electron diffraction method is illustrated.

[0042] Fig. 9 An exemplary on-line electron diffraction method is illustrated. DETAILED DESCRIPTION

[0043] Electron diffraction system

[0044] Figure 1 An exemplary electron diffraction system 100 is illustrated for generating an electron backscatter diffraction (EBSD) pattern that can be used to characterize a sample of interest.

[0045] The system 100 includes an electron beam source 105 that provides a monochromatic electron beam or a substantially monochromatic electron beam along a beam axis 115 to a sample 110. The electrons can have an energy selected from a range of energies (e.g., energies between 3keV and 50keV). The electron source 105 can be included in an electron beam column 120 as part of a scanning electron microscope (SEM) setup. The electron beam column 120 is configured to adjust the electron beam to focus on a point on the sample 110 and scan the electron beam across the surface of the sample 110, such as line by line or segment by segment. A controller 125 in communication with the electron beam column 120 controls the movement of the electron beam and the start / stop of the electron beam. The controller 125 can be part of a SEM computer 130.

[0046] The chamber 135 (typically a vacuum chamber) houses a sample holder 140. The sample holder 140 may be part of a manipulator 145 for translating and rotating the sample 110 to a desired position and orientation. In this way, the sample 110 may be placed at a selectable angle relative to the beam axis 115, such as 70 degrees relative to the beam axis 115. In other examples, the angle may be between 0 and 80 degrees. The electron beam column 120 is aligned with an aperture 150 in the chamber housing 155 to allow the electron beam to enter the chamber 135 and contact the sample 110.

[0047] The system 100 also includes a particle detector 160 positioned within the chamber 135, which is configured to detect electrons scattered from the sample 110. That is, the particle detector 160 detects electrons that have interacted with the sample 110. Since electron diffraction methods and systems are described herein, and since electrons are being detected, the particle detector 160 described herein is an electron detector 160. Data about the scattered electrons is collected by the particle detector 160 and sent to a processing device to be processed into a diffraction pattern. In this embodiment, the processing device includes a computer system that includes a computer 165 and / or a field programmable gate array 185. In the following description, the processing device is described as including a computer 165.

[0048] The particle detector 160 includes a front end 170 configured to process the collected data. The front end 170 pre-processes the data before sending it to the computer 165 for further processing. The front end 170 also provides power to the particle detector 160 and handles cooling when necessary. The particle detector 160 also includes a feedthrough 175 through which a communication link that allows data to be transmitted from the chamber 135 to the computer 165 is routed.

[0049] The particle detector 160 also includes a particle detector chip 180, which includes a semiconductor sensor chip (e.g., a silicon chip) bonded to an electronic readout chip. The semiconductor sensor chip includes a pixel array, for example, 256×256 pixels or 512×448 pixels. Examples of such particle detector chips 180 include Timepix3 (see https: / / kt.cern / technologies / timepix3) and Timepix4 (see https: / / indico.cern.ch / event / 591299 / contributions / 2423187 / attachments / 1393307 / 2123191 / Timepix4_specs.pdf and https: / / cds.cern.ch / record / 2825271), available from ADVACAM (https: / / advacam.com / ), ASI (https: / / www.amscins.com / index.html), and Quantum Detectors (https: / / quantumdetectors.com / ).

[0050] Conventional particle detectors operate in a "counting" mode that counts each electron incident on a pixel as 1. In most detectors, the counting mode is frame-based, i.e., the counts for each pixel are collected at predefined time intervals, frames, and the counts for each pixel are read out at the end of the frame. Thus, the readout from each pixel describes how many electrons were counted within the time frame in which the readout was collected.

[0051] The Timepix3 and Timepix4 particle detectors 160 can also operate in a data driven mode. In the data driven mode, as and when the particle detector 160 detects electrons that strike a pixel, information about the detected electrons is read out for each detected electron, rather than accumulating a total count for each pixel to be read out at the end of the frame.

[0052] When operating in data-driven mode, these particle detectors 160 can measure the electron energy deposited in each pixel by each incident electron. When an electron hits a pixel, the particle detector 160 measures a value proportional to the energy of the incident electron. The measured energy value is also referred to as the "over-threshold time" (TOT) value E. The particle detector 160 that measures the over-threshold time operates in "over-threshold time" mode. A threshold value can be used to distinguish electrons according to their energy, that is, so that electrons with energy below a threshold value are not counted. Both the instance of electron detection and the measured value are read out from the pixel and used for subsequent processing.

[0053] In more detail, the electrons that strike the pixel cause a cascade of electron-hole pairs. The number of electron-hole pairs is proportional to the energy released by the electron, and a charge is given to the pixel in proportion to the number of electron-hole pairs. The time it takes for the pixel to discharge back to the detector threshold level after the electron strikes is the super-threshold time, and is proportional to the charge given to the pixel. Therefore, this time is also proportional to the energy of the detected electron measured by the particle detector 160 (i.e., the actual electron energy is the super-threshold time multiplied by a coefficient). In other words, the measured super-threshold time is the time that the signal output from the particle detector 160 is above the detector threshold level, or the number of hour periods (internal clock time) that the charge of the pixel remains above the detector threshold level. The detector threshold level is adjustable and can be set before use and changed during use to provide a desired level of electron energy discrimination.

[0054] Figure 2 A comparison of the counting mode and the overthreshold time mode of a particle detector is shown 200. The top graph shows the charge accumulated in a pixel versus time. The falling curve 205 is the charge signal recorded by the particle detector relative to the detector threshold level 210 during the gradual discharge of the pixel over time.

[0055] The middle graph shows a counting pattern 215. When electrons strike a pixel with sufficient energy to charge the pixel above the detector threshold level 210, as shown at points X and Z, a count is measured by the pixel counter in the particle detector 160. When electrons strike a pixel with sufficient energy to charge the pixel above the detector threshold level 210, as shown at point Y, the pixel counter in the particle detector does not measure a count.

[0056] The bottom graph shows the count of the Overthreshold Time pattern 220. When an electron with energy above the detector threshold level 210 strikes the pixel, the pixel counter counts the number of hour periods that the pixel charge remains above the detector threshold level 210. This is the measured Overthreshold Time E mentioned above. The event at X is a higher energy electron strike, which results in a longer discharge time and an Overthreshold Time E of 4. The event at Z is a lower energy electron strike, which results in a shorter discharge time and an Overthreshold Time E of 2. The event at Y is below the detector threshold level 210 and therefore has an Overthreshold Time of zero.

[0057] While measuring the threshold-exceeding time E, the particle detector 160 may also measure the detection time of the detected electron, such as the arrival time t of the detected electron, i.e., the timestamp when the electron hits the pixel. These measured data may be included in a vector (x, t, E) for each detected electron, which also includes the pixel position x of the pixel where the electron was detected.

[0058] The measured data for each detected electron is sent from the particle detector 160 to the computer 165 for processing. The position of the pixel of each detected electron may be sent, or an identifier of the pixel may be sent, from which the position can be determined in a subsequent processing step. Because the particle detector 160 operates in a data driven mode, this information is sent after each and every electron impact, i.e., information about each and every detected electron (such as the vector x, t, E) is provided to the computer 165 in real time. This is in contrast to the counting mode, in which the information is only sent as a full frame readout after a set measurement time.

[0059] The computer 165 communicates with the particle detector 160 via the feedthrough 175 and receives the collected electronic data for processing to generate an electron diffraction pattern. The computer 165 includes a processor and a memory for storing computer-readable instructions. The computer 165 can be any suitable computer configured to operate software for generating an electron diffraction pattern. Optionally, a field programmable gate array (FPGA) 185 can communicate with the particle detector 160 and the computer 165 for processing the received signals. In some examples, the computer 165 is configured to perform the method steps mentioned below. In other examples, the FPGA 185 is configured to implement the method steps. The controller 125 can also communicate with the computer 165 and / or the FPGA 185 to control the processing. The computer 165 can be a computer separate from the SEM computer 130, or can be a part of the SEM computer 130.

[0060] Electron diffraction method

[0061] Figure 3 An exemplary electron diffraction method 300 is illustrated, which may be used using the methods described above and as described above. Figure 1 The method 300 may also be implemented using other suitable electron diffraction setups including the supra-threshold time mode particle detector 160 as previously described.

[0062] The experimental diffraction conditions may be set up first. For example, the sample of interest 110 is loaded into the sample holder 140 in the chamber 135 and is set to a desired position and orientation by the manipulator 145. The electron beam column 120 is configured to move the electron beam to an appropriate position under the control of the controller 125 so as to align the electron beam and the sample 110 at a fixed angle. Other setup conditions (e.g., electron beam voltage, electron beam current, sample-detector distance, map size, and acquisition time) may also be selected and applied.

[0063] The controller 125 starts the electron beam and the electron beam is directed 305 to be incident on the sample 110. The controller 125 starts data collection and synchronizes the collected data with the beam scanning so that the position of the electron beam on the sample 110 is known for each scattered electron detected. In more detail, since the time of data acquisition and the position of the beam are known, the measured arrival time t values ​​can be used to determine which electrons are scattered from which area of ​​the sample 110. This can be done in one pass (i.e., taking one picture) or in several passes (taking multiple pictures and adding the second picture to the first picture, and so on).

[0064] The electron beam is aimed at a first region of interest of the sample 110. For example, the surface of the sample 110 can be divided into a grid of multiple sections, where each region of interest is one or more sections of the grid. When the electron beam contacts the first region of the sample 110, some electrons from the beam are scattered toward the particle detector 160 at varying angles. The scattered electrons hit the pixels of the particle detector 160, and the particle detector 160 measures 310 the energy E of each detected electron (where E is the overthreshold time proportional to the energy of the electron). The particle detector 160 can also measure the arrival time t of each detected electron. In response to each detected electron, these data about the detected electrons are transmitted 320 to the computer 165 for processing, for example as a vector (x, t, E), which also contains the pixel position x of each detected electron and is sent for processing. When each electron is detected in a continuous manner (the particle detector 160 operates in a data-driven mode), the particle detector 160 continues to transmit the detected electron data to the computer 165. This is in contrast to count mode, where data is sent as a full frame readout, so in data driven mode the total data acquisition time is not limited by the frame readout time.

[0065] After scanning the first area for a set period of time (e.g., 0.5 ms), controller 125 moves the electron beam to a second area of ​​interest of sample 110 and collects data while the electron beam scans over the second area, and so on, until all or at least a portion of sample 110 has been scanned.

[0066] Typically, in electron diffraction, higher energy electrons are the most useful electrons because they provide the most information about the sample 110. Figure 4 A histogram of an exemplary electron energy spectrum 400 is shown in , which illustrates the variation in possible energies of electrons. When generating an electron diffraction pattern, it is desirable to increase the contribution of high energy electrons. Conventional diffraction methods can apply an energy filter to discard electrons with energies below a set threshold. However, because this method operates in a counting mode, conventional particle detectors can only distinguish the number of electrons counted above the threshold. There is no way to know how much the energy of each electron counted is above the threshold, nor how much each electron contributes to the diffraction pattern.

[0067] It is important to be able to include lower energy electrons in the diffraction pattern because they contain information about the sample 110 as well as the background signal detected by the particle detector 160. To better distinguish high energy electrons from low energy electrons, the energy of the detected electrons can be used as a discriminating parameter to produce an energy-weighted diffraction pattern. Figure 3 As shown in step 330, weighting is obtained by using the measured energy / superthreshold time E of the measured electronic data collected by the particle detector 160 in the superthreshold time mode or the value derived from the measured energy / superthreshold time. This is contrary to the conventional technology in which each electron is only counted as the same value (counted as 1). The weighting includes using the measured energy value or a value depending on the measured energy value, such as a scaling value. In all cases, the value used when generating the electron diffraction pattern is referred to as the energy-weighted contribution value in this article, and the energy-weighted contribution value is calculated using an energy-related function. In the case where the measured energy value is directly used as the energy-weighted contribution value, the energy-related function effectively becomes the product of the measured energy value and the unit. Using the calculated energy-weighted contribution value causes each detected electron to have a weighted contribution to the generated diffraction pattern 340, wherein the contribution depends on the energy of the detected electron.

[0068] Weighting relates the measured suprathreshold energy / time to the actual energy of each detected electron (where high energy electrons are considered more important because they carry most of the information). By assigning a weight (and therefore importance) to each individual electron detected, the discernibility of the diffraction pattern is improved. Weighting the data increases the amount of information that can be extracted from the diffraction pattern, for example because it increases the ratio of useful signal to background and provides better contrast in the diffraction pattern compared to conventional electron counting.

[0069] Energy-weighted contribution values ​​are used to give some measured super-threshold energies / times a greater weight (or influence) on the generated electron pattern than other values, and thus reflect the contribution of the electron energy to the diffraction pattern. The energy correlation function can change the nature of this energy dependence of each electron's contribution to the electron diffraction pattern. For example, the energy correlation function can exaggerate the contribution of higher energy electrons to further increase their contribution to the diffraction pattern, and reduce the contribution of lower energy electrons.

[0070] The energy correlation function may be any suitable function that increases or decreases the contribution of the detected electrons according to the energy of the detected electrons. For example, the energy correlation function may be directly proportional or inversely proportional to the electron energy. The energy correlation function may be linear or non-linear, such as exponential, sinusoidal, etc. The energy correlation function may be user selectable. Any suitable function or functions may be selected by the user to be applied to the particle detector data.

[0071] In one example, the energy correlation function is F(E)=0.5E 2 , where E is the measured energy value (e.g., in keV) collected by the particle detector 160 operating in the superthreshold time mode. In another example, the energy correlation function is F(E)=2E-10. The energy correlation function can have the form F(E)=2E-20 to favor electrons above 10 keV and discriminate against electrons below 10 keV (for a 20 keV beam).

[0072] The effect of using energy-weighted contributions is Figure 5 Shown in. Figure 5 The effective contribution 500 to an electron diffraction pattern using an energy-weighted contribution value that increases with electron energy 510 (solid line) is illustrated compared to a contribution that does not vary with electron energy 520 (dashed line). The x-axis shows the energy of the scattered electrons expressed as a fraction of the electron beam energy. The y-axis shows the effective contribution of the electrons to the generated electron diffraction pattern. Figure 5As can be seen in FIG. 1 , using energy-weighted contribution values ​​(solid line) increases the effective contribution based on the energy of the electron, so that higher energy electrons contribute more to the generated electron diffraction pattern than lower energy electrons. As previously described, in the super-threshold time mode, the particle detector 160 measures the super-threshold time E for each detected electron. Since this time depends on the energy of the electron (i.e., since the electron deposits more charge (electron-hole pairs) in the pixel when it strikes, the pixel takes longer to discharge a higher energy electron), the measured time represents a value proportional to the actual energy of the electron.

[0073] The electron diffraction pattern may now be generated 340 as follows. After the detected electron data has been collected and the energy weighted contributions calculated, a matrix of the same size as the pixel array of the particle detector 160 is created, e.g., a 256×256 matrix. Each pixel position x is mapped to a corresponding matrix element position such that the matrix represents the position of each electron detected in the pixel array of the particle detector 160, e.g., pixel position x i,j The matrix element M is mapped to the matrix i,j Each energy-weighted contribution is added to the appropriate matrix element based on the pixel position where the electron was detected, e.g. i,j The calculated energy-weighted contribution is added to the matrix element M i,j Adding each of the energy-weighted contributions associated with a pixel provides a cumulative sum of the energy-weighted contributions of all electrons detected by any particular pixel of particle detector 160. The energy-weighted contributions take into account the energy of each individual electron, so electrons contribute varying counts to their corresponding matrix elements, e.g., 0.5, 3, 0.1, etc., depending on their measured energy E (and the form of the energy correlation function F(E)). This increases the importance of particular portions of the electron energy spectrum.

[0074] In conventional counting mode, each incoming electron always contributes a count of 1 to the corresponding matrix element. Therefore, an electron diffraction pattern is formed by simply adding each of these individual counts. In contrast, the method described herein abandons the use of summing individual counts. Instead, the contribution values ​​from each electron impact are summed. In addition, these contributions are energy-weighted contributions, whose values ​​vary with the energy of the electron, wherein the variation is determined by an energy-related function. Using energy-weighted contributions in this way means that the matrix elements of pixels that experience several low-energy electron impacts can maintain a lower cumulative value than the matrix elements of pixels that experience only a single high-energy electron impact.

[0075] As described above, the energy weighted contribution values ​​are calculated before filling the matrix with the energy weighted contribution values. In other examples, the energy weighted contribution values ​​may be calculated after the matrix has been filled with the measured energy values ​​E (in the same manner).

[0076] The matrix forms an energy-weighted electron diffraction pattern in its raw state. Each element of the matrix can be mapped to a display value, for example, a value in a color gradient or an intensity in a grayscale. An image can then be generated from these display values ​​to illustrate the energy-weighted electron diffraction pattern as a color / grayscale image. An example of such an energy-weighted electron diffraction pattern is shown in Figure 6 The generated energy-weighted electron diffraction pattern can then be interpreted to derive one or more structural characteristics of the sample 110, as described in more detail below. For example, the lattice and spatial orientation of the sample can be derived from the generated electron pattern. From this information and a comparison of diffraction patterns taken across the surface of the sample 110 (i.e., multiple regions of interest / surface portions), additional characteristics of the sample 110 can be determined.

[0077] Further processing may be performed to further improve the quality of the energy-weighted electron diffraction pattern.

[0078] For example, the data collected by the particle detector 160 includes a background signal that can affect the quality of the energy-weighted electron diffraction pattern. The background signal corresponds to a "dark" signal generated by the particle detector 160 when no electron beam is incident on the sample 110. Therefore, the background can be found in the data read out from the particle detector 160 before actual data measurement is performed with the electron beam incident on the sample 110. This background signal can be subtracted / divided from each newly generated electron diffraction pattern.

[0079] Other ways to improve pattern quality include outlier removal and contrast stretching. For example, to remove outliers, the algorithm searches for matrix element values ​​that are much higher or much lower than the median value and replaces them with the median value. In contrast stretching, the lowest and highest matrix element values ​​are found, and then all values ​​are transformed (i.e., stretched) to a larger range of usable display values, thereby improving contrast, for example, improving Figure 6 The contrast in the color / grayscale image shown in .

[0080] The generated images (see, for example, Figure 6 ) shows the Kikuchi bands that are characteristic of a particular sample and they provide information about the structure and orientation of the sample. Figure 7The band profile heights 700 of two flat field patterns taken from the same sample grain are shown. This shows a comparison of band contrast and demonstrates the effect of using energy weighted contributions compared to just counting electron impacts. The solid line 710 is the band profile obtained when each detected electron has been weighted depending on its energy. The dashed line 720 is the band profile when each detected electron has no weighting. By assigning a weight and therefore importance to each individual electron, the contrast and therefore visibility of the bands can be increased. This in turn increases the confidence in resolving the diffraction pattern using an indexing algorithm and produces more accurate results. The higher the contrast in the bands, the more visible they are to the indexing algorithm and therefore the band peaks can be determined with greater accuracy. This allows energy weighted patterns to be acquired in less time than non-weighted patterns. For example, using the above method, an energy weighted electron diffraction pattern can be indexed after 0.7ms of data collection, while using a non-weighted pattern would require much longer, such as over 1ms or more of data collection, to produce the same indexing success rate / confidence.

[0081] To help interpret the energy-weighted electron diffraction pattern, a Hough transform can be applied to help identify the bands. These bands form bright peaks (i.e., the highest intensity peaks) in Hough space. These minima / maxima can be found in Figure 7 , and the minima / maxima are the points to look for in Hough space. Other standard processing methods can also be used to identify bands. As mentioned before, the electron beam is moved over the sample area of ​​interest, for example to cover a grid of the sample surface section by section. At each region of interest, the measured data is collected, weighted, and used to generate an energy-weighted electron diffraction pattern.

[0082] Hundreds to hundreds of thousands of these patterns are generated and subsequently indexed to provide information about the structural properties of the sample 110. For example, the patterns can be interpreted to provide information about the crystal orientation (typically represented by the Euler angles), the crystal phase (e.g., whether the crystal phase in a certain pixel is cubic, hexagonal, orthorhombic, etc.), and the pattern quality (which is proportional to the surface mass or deformation of the sample). From such information, the phase composition (e.g., the percentage of cubic phase, the percentage of hexagonal phase, etc.), texture (the size of the average grain, how the average grain is oriented toward a particular axis, etc.), and the orientation relationship between adjacent phases, as well as the location of strain (e.g., determining sample deformation based on the misorientation of adjacent patterns) can then be calculated.

[0083] Optionally, an energy filter may be applied 350 to the measured data before or after weighting 330 to select one or more energy ranges of interest, such as electrons from 10keV to 15keV, or to set a threshold, such as only electrons exceeding 10keV. Energy filters may also be used to select energy bands. More specifically, in implementation, an energy filter is applied to the measured threshold-exceeding time E of the detected electrons. Additionally, a position filter may also be applied to the pixel position x of the detected electrons. For example, data from pixels located at or near the edge of the pixel array of the particle detector 160 may be excluded.

[0084] In some examples, the energy range of interest is unknown. Therefore, an initial experiment can be performed to determine the initial electron diffraction pattern. Subsequent experiments can be performed with the selected weighting and filtering depending on the initial diffraction pattern to produce improved diffraction pattern results. If the energy range of interest is known (e.g., when the sample is a known material), an energy filter can be applied at the particle detector 160. Therefore, unwanted electrons are discarded before weighting, which saves space and improves processing speed. Alternatively, for example, in the case of unknown sample materials, an energy filter can be applied after weighting. When implementing the electron diffraction method 300 as described above, weighting can be applied at different points. This is useful when it is initially unclear what electron energy should be studied. The initial experiment can be run, and the energy range is narrowed to the region of the energy spectrum of interest in subsequent experiments. For example, in the case where the energy correlation function is still unknown (e.g., in the initial experiment), the energy correlation function is applied after the measured detector data has been saved (i.e., offline). In cases where the energy correlation function is known (e.g., in subsequent experiments that have generated data that gives an idea of ​​the type of energy correlation function to be applied), the energy correlation function can be applied after the measured detector data leaves the particle detector 160 or at the FPGA 185, or the energy correlation function can be applied while the data is being saved or processed by the computer 165 (i.e., online).

[0085] Figure 8An exemplary offline version 800 of the previously described electron diffraction method is shown. The sample is loaded into a sample holder and aligned with an electron beam column 810. Experimental conditions, such as electron beam voltage, electron beam current, area or region of interest, etc., are set, and initial data acquisition 820 is started. The particle detector 160 records the detected electron data and generates an initial electron diffraction pattern 830 (which is not weighted). The quality of the initial pattern is checked, and if its quality is not suitable, the experimental conditions are updated and the initial pattern is regenerated. Once the quality is satisfactory, the system begins to acquire detected electron data 840 for characterizing the sample. An energy correlation function is applied to the detected electron data 850. An energy filter can also be applied at this point. The detected electron data is reprocessed, including weighting, to generate a subsequent energy-weighted electron diffraction pattern 860. The quality of the energy-weighted electron diffraction pattern is checked 870. If the quality is improved due to weighting, the method ends and the experimental conditions are saved for future acquisition 880. If the quality has not improved, the energy correlation function (and optionally the filtering may also be adjusted) is adjusted and reapplied to the detected electron data 890. The detected electron data is again reprocessed, including the adjusted weightings, to generate an updated energy-weighted electron diffraction pattern. This cycle may be repeated until appropriate improvements in the energy-weighted electron diffraction pattern are observed.

[0086] Fig. 9 An exemplary online version 900 of the previously described electron diffraction method is shown. Figure 8In an off-line method, the sample is loaded into a sample holder and aligned with the electron beam column 910. Experimental conditions such as electron beam voltage, electron beam current, area or region of interest, etc. are set, and data acquisition is started 920. The particle detector 160 records the detected electron data, and an energy correlation function is applied when the data is saved to a computer 930 (an energy filter may also be applied). The weighted electron data is processed to generate an energy-weighted electron diffraction pattern 940 and the quality is checked 950. If the quality is improved due to weighting, the system begins to collect detected electron data for characterizing the sample 960. If the quality is not improved, the energy correlation function is adjusted (and optionally the filtering may also be adjusted) and reapplied to the detected electron data 970. The updated weighted electron data is processed to generate an updated energy-weighted electron diffraction pattern, which is again checked for quality. This cycle can be repeated until a suitable improvement in the diffraction pattern is observed. Optionally, the experimental conditions are also updated based on the quality check. As described above, the advantage of weighting the collected data is to increase the contrast in the diffraction pattern. Another advantage of weighting the data is that it reduces the amount of data transmitted from the particle detector 160. Since each electron impact is stored individually when the particle detector 160 is operating in data driven mode, it is possible to discard some entries (so they do not have to be saved). Therefore, less data can be saved without compromising the quality of the results.

[0087] The energy-weighted electron diffraction patterns described herein have a variety of uses, including but not limited to scanning electron microscopy (SEM), transmission electron microscopy (TEM), scanning transmission electron microscopy (STEM), 4D STEM, transmission Kikuchi diffraction (TKD), and pixelated STEM in SEM or TEM. Including electron energy provides an additional dimension for collecting and analyzing information. Compared with simple electron counting, energy-weighted electron diffraction patterns provide better contrast, increasing the amount of information that can be extracted about the sample. The indexing success rate of the electron backscatter diffraction pattern indexer at low illumination can be increased by assigning more weights to higher energy electrons. This provides the same dot pattern, but with information about the energy and arrival time of each electron. This can also be explored to produce energy-resolved virtual dark field imaging, which will carry depth or thickness information (because electrons passing through more or heavier materials are slowed down more). Weighting also enables 2D material sheets (i.e., graphene) to be distinguished based on the intensity and energy of the electrons obtained. In addition, since the data is sent immediately, the diffraction pattern will continuously improve after each scan. The arrival time t of the electron can be used in SEM stroboscopic observation.

[0088] Although a SEM system setup is described, the electron diffraction system may instead be another suitable type of system, such as a transmission electron microscope (TEM) or a scanning transmission electron microscope (STEM) system.

[0089] The above description describes applying an energy correlation function to the measured super-threshold energy / time E. In additional / alternative examples, the energy correlation function may be applied to the pixel position values ​​x. For example, it may be desirable to include only measurements from middle pixels. Thus, a position function may be applied to increase the contribution of middle pixels compared to edge pixels. Alternatively, a band detection algorithm may be adjusted to look for bands in specific areas of a pattern.

[0090] The above description discusses the importance of higher energy electrons in electron diffraction methods. However, in other examples, lower energy electrons may be of more interest, and thus the energy correlation function may be selected to increase the influence of lower energy electrons and discriminate high energy electrons.

Claims

1. A method for generating an energy-weighted electron diffraction pattern, the method comprising: directing the electron beam to be incident on the sample; detecting electrons scattered from the sample by a particle detector having an array of pixels, wherein the detecting comprises measuring, for each detected electron at a pixel, an energy value proportional to the energy of the detected electron; sending the measured energy value of each detected electron and an identifier of the pixel in which the electron was detected to a processing device; calculating, by the processing device, an energy-weighted contribution from each measured energy value using an energy correlation function, wherein the energy correlation function produces an energy-weighted contribution that varies with electron energy; as well as An energy-weighted electron diffraction pattern is generated by the processing device using the pixel position associated with each pixel identifier and the energy-weighted contribution value.

2. The method according to claim 1, wherein: The energy correlation function increases or decreases the contribution of the detected electrons to the generated energy-weighted electron diffraction pattern depending on the energy of the detected electrons.

3. The method according to claim 2, wherein: The energy correlation function increases the contribution of detected electrons of higher energy relative to detected electrons of lower energy such that, optionally, the energy-weighted contribution of a detected electron is higher than the energy-weighted contributions of all detected electrons having lower energies.

4. A method according to any preceding claim, wherein: The energy correlation function produces energy-weighted contribution values ​​that increase with increasing energy values, such that optionally, the energy correlation function produces energy-weighted contribution values ​​that increase linearly with increasing electron energy values.

5. A method according to any preceding claim, wherein: The energy correlation function is user configurable.

6. A method according to any preceding claim, wherein: The detecting further comprises measuring, for each detected electron, a detection time of the detected electron, and wherein the transmitting further comprises transmitting the measured detection time of the detected electron to the processing device.

7. A method according to any preceding claim, wherein: Measuring the energy value includes measuring the time that the charge induced in the pixel struck by the electron is above a detector threshold level.

8. A method according to any preceding claim, wherein: Generating the energy-weighted electron diffraction pattern using the pixel position and the energy-weighted contribution value comprises: An energy-weighted electron diffraction pattern is generated comprising an array of image pixels arranged to coincide with the pixel array of the particle detector, wherein each image pixel is displayed with a display value proportional to a combination of the energy-weighted contribution values ​​of the electrons detected as having struck a corresponding pixel of the particle detector.

9. The method according to claim 8, wherein: The energy-weighted electron diffraction pattern is a color image, and the displayed values ​​are values ​​on a color gradient; or The energy-weighted electron diffraction pattern is a grayscale image, and the displayed value is intensity.

10. The method according to claim 8 or 9, further comprising performing background subtraction by: detecting, by the particle detector, electrons received when the electron beam is not incident on the sample; and For each image pixel, an amount proportional to the electrons counted for each pixel of the detector is subtracted from the displayed value of the corresponding image pixel.

11. A method according to any preceding claim, further comprising applying an energy filter to the measured energy values ​​or to the energy-weighted contribution values, and optionally, the energy filter is one of: a predetermined energy band, one or more electron energy ranges, or a predetermined energy threshold.

12. A method according to any preceding claim, further comprising displaying the generated energy-weighted electron diffraction pattern.

13. A method according to any preceding claim, further comprising interpreting the generated energy-weighted electron diffraction pattern to derive one or more structural properties of the sample, and wherein, Optionally, the one or more structural characteristics include: crystal orientation, crystal phase, phase composition, texture and strain location.

14. An electron diffraction system, comprising: an electron source configured to provide an electron beam to a sample holder configured to hold a sample; a particle detector configured to detect electrons diffracted from the sample; and A processing device communicatively coupled to the particle detector, wherein the processing device is configured to perform a method according to any preceding claim.

15. A computer program comprising computer program instructions which, when executed by a computer processing system, cause the computer processing system to perform a method of generating an energy-weighted electron diffraction pattern, the method comprising: receiving data including measured energy values ​​of detected electrons and identifiers of pixels at which the detected electrons were detected, wherein the data was obtained by directing an electron beam incident on a sample and detecting electrons scattered from the sample by a particle detector having an array of pixels, wherein the detecting includes measuring, for each detected electron at a pixel, an energy value proportional to the energy of the detected electron; calculating an energy-weighted contribution from each measured energy value using an energy correlation function, wherein the energy correlation function produces an energy-dependent contribution that varies with electron energy; as well as generating an electron diffraction pattern using the pixel position associated with each pixel identifier and the energy-weighted contribution value; or A computer readable medium having such a computer program stored thereon.

Images