Semiconductor charged particle detector for microscopy

Abstract

A detector can be provided with an array of sensing elements. The detector can include a semiconductor substrate including the array, and circuitry configured to count a number of charged particles incident on the detector. The circuitry of the detector can be configured to process output from the plurality of sensing elements, and to increase a counter in response to a charged particle arrival event on a sensing element in the array. Various counting modes can be used. Counting can be based on an energy range. A number of charged particles in a certain energy range can be counted, and an overflow flag can be set when an overflow is encountered in a sensing element. The circuitry can be configured to determine a timestamp of a respective charged particle arrival event occurring at each sensing element. A size of the sensing elements can be determined based on a standard for implementing charged particle counting.

Description

[0001] Cross Reference to Related Applications

[0002] This application claims priority to U.S. Application 62 / 682,730, filed June 8, 2018, U.S. Application 62 / 787,066, filed December 31, 2018, and U.S. Application 62 / 852,816, filed May 24, 2019, which are incorporated by reference in their entirety. TECHNICAL FIELD

[0003] The description herein relates to charged particle detection, and more particularly to systems and methods that can be applied to charged particle beam detection. BACKGROUND

[0004] Detectors can be used to sense physically observable phenomena. For example, charged particle beam tools such as electron microscopes can include detectors that receive charged particles projected from a sample and output detection signals. The detection signals can be used to reconstruct images of the structure of the sample under examination, and can be used, for example, to reveal defects in the sample. The detection of defects in a sample becomes increasingly important in the manufacture of semiconductor devices that can include a large number of densely packed small integrated circuit (IC) components. Specialized inspection tools can be provided for this purpose.

[0005] In some applications in the field of inspection, such as microscopy using a scanning electron microscope (SEM), an electron beam can be scanned over a sample to derive information from backscattered or secondary electrons generated from the sample. In related art, an electron detection system in a SEM tool can include a detector configured to detect electrons from the sample. Existing detectors in SEM tools can only detect the intensity of the beam. The sensitivity in conventional detection systems can be limited by poor signal-to-noise ratio (SNR), especially when the beam current is reduced to, for example, the picoampere range. In some detection methods, a large-area semiconductor detector or a set of small-area semiconductor detectors with an area equal to, less than, or greater than the area of the beam spot can be used. The current induced by the incoming electron beam can be generated within the detector and then amplified by an amplifier behind the detector.

[0006] As semiconductor devices continue to be miniaturized, inspection systems can use increasingly lower electron beam currents. As the beam current is reduced, it becomes more difficult to maintain the SNR. For example, the SNR can drop sharply when the probe current is reduced to 200 pA or lower. Poor SNR can require measures such as image averaging or extending the integration time of the signal corresponding to each pixel in the image of the sample, which can increase the electron dose on the sample surface, causing surface charge artifacts or other detrimental effects. Such measures can also reduce the overall throughput of the inspection system.

[0007] In the related art, particle counting can be useful in low current applications. Particle counting can be used in detectors such as Everhart-Thornley detectors (ETD), which can use a scintillator and a photomultiplier tube (PMT). In a probe current range of some applications, such as 8 pA to 100 pA, an ETD can show good SNR. However, the light yield of the scintillator can decrease with the accumulation of electron dose, and thus has a limited useful life. Aging of the scintillator can also cause performance drift at the system level, and can contribute to generating non-uniform images. Thus, an ETD can not be suitable for use in inspection tools, especially when used in semiconductor manufacturing facilities that can need to operate 24 hours a day, 7 days a week.

[0008] There is a need for a charged particle detector that can achieve a higher SNR and can be used with low probe currents, such as probe currents below 200 pA. At the same time, the detector should guarantee stable quantum efficiency and long useful life with low performance drift, for example, even when using a probe current of 1 nA or more in continuous operation.

[0009] Detection systems that employ related art methods can face limitations in detection sensitivity and SNR, especially at low electron dose. In addition, in some applications, additional information other than beam intensity can be needed. Some related art systems can employ an energy filter, such as a filter electrode, to filter out some charged particles having a certain energy level. This can help derive additional information from the sample. However, an energy filter can add additional complexity to the system, and can cause SNR degradation due to loss introduced by the energy filter. Thus, improvements in detection systems and methods are needed. SUMMARY

[0010] Embodiments of the present disclosure provide systems and methods for charged particle detection. In some embodiments, a detector for a charged particle beam apparatus can be provided, the detector comprising: a semiconductor substrate comprising an array of sensing elements; and a circuit configured to count a number of charged particles incident on the detector. The circuit of the detector can be configured to process outputs from the plurality of sensing elements and to increment a counter in response to a charged particle arrival event on a sensing element in the array. The circuit can comprise a plurality of circuitry, each corresponding to a sensing element. The circuit can be configured to determine a timestamp of a respective charged particle arrival event occurring at each sensing element.

[0011] The detector can be configured to operate in various counting modes. For example, the detector can be configured to count, without discriminating the energy of the one charged particle arrival event, at most one charged particle arrival event in the sensing element before a reset. The reset of the sensing element can refer to a reset of the sensing element itself or circuitry associated therewith. The detector can also be configured to count, without discriminating the energy of the charged particle arrival event, the number of charged particle arrival events and set an overflow flag when an overflow is encountered in the sensing element. The detector can also be configured to count, for a first energy range, at most one charged particle arrival event in the sensing element before a reset. The detector can also be configured to count the number of charged particle arrival events for the first energy range and set an overflow flag when an overflow is encountered in the sensing element.

[0012] The detector can be configured to reset the sensing elements in the array in various reset modes. For example, the detector can be configured to reset all of the sensing elements in the array of sensing elements simultaneously. The detector can also be configured to reset all of the sensing elements of a region of the array of sensing elements simultaneously. The detector can also be configured to individually reset each of the sensing elements in the array of sensing elements. The detector can also be configured to simultaneously reset some of the sensing elements in the array of sensing elements and individually reset some of the sensing elements in the array of sensing elements.

[0013] Some embodiments of the present disclosure can provide a method comprising processing outputs from a plurality of sensing elements in an array of sensing elements of a detector, counting a number of charged particles incident on the detector, wherein the counting comprises incrementing a counter in response to a charged particle arrival event on a sensing element in the array, and determining a timestamp of the charged particle arrival event.

[0014] Some embodiments of the present disclosure can provide a non-transitory computer readable medium storing instructions for implementing a method such as the method described above.

[0015] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosed embodiments, as can be claimed. BRIEF DESCRIPTION OF DRAWINGS

[0016] The above and other aspects of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings in which:

[0017] Figure 1 is a schematic diagram illustrating an exemplary electron beam inspection (EBI) system consistent with embodiments of the present disclosure.

[0018] Figure 2A 、 Figure 2B andFigure 2C is a schematic diagram of an exemplary electron beam tool that can be part of an exemplary electron beam inspection system consistent with embodiments of the present disclosure. Figure 1

[0019] Figure 3A is a representation of an exemplary structure of a detector consistent with embodiments of the present disclosure.

[0020] Figure 3B and Figure 3C is a diagram illustrating a cross-sectional view of a detector consistent with embodiments of the present disclosure.

[0021] Figure 3D and Figure 3E is a diagram illustrating a cross-sectional view of various detector elements consistent with embodiments of the present disclosure.

[0022] Figure 3F is a diagram of a detector consistent with embodiments of the present disclosure.

[0023] Figure 3G is a representation of an exemplary structure of a detector consistent with embodiments of the present disclosure.

[0024] Figure 4A is a view of a portion of Figure 2B showing the projection of secondary electrons from a sample toward a detector.

[0025] Figure 4B illustrates an example of a secondary electron landing site distribution on a detector surface consistent with embodiments of the present disclosure.

[0026] Figure 5 is a schematic representation of an electron arrival event and its relationship to a current signal.

[0027] Figure 6 is a representation of a detection system architecture using analog signals.

[0028] Figure 7 is a schematic representation of an electron arrival event and its relationship to a current signal.

[0029] Figure 8 illustrates an example of a detector element sized relative to a secondary electron landing site distribution on a detector surface consistent with embodiments of the present disclosure.

[0030] Figure 9A and 9B illustrates an example of a detector including an array of detector elements consistent with embodiments of the present disclosure.

[0031] Figure 10 illustrates a detector and a region of high electron arrival rate consistent with embodiments of the present disclosure.​

[0032] Figure 11A and 11B A schematic representation of an electron arrival event and its relationship to a detector output signal consistent with embodiments of the present disclosure is shown.

[0033] Figure 12 is a table showing the relationship between the number of events to be counted and the number of events missed in a detector consistent with embodiments of the present disclosure.

[0034] Figure 13A A plot that can represent the confidence level of an electron count at a particular count buffer consistent with embodiments of the present disclosure is illustrated.

[0035] Figure 13B A plot that can represent the loss of detection rate due to a maximum count at a particular count buffer consistent with embodiments of the present disclosure is illustrated.

[0036] Figure 14A A schematic representation of an electron arrival event and its relationship to an output signal consistent with embodiments of the present disclosure is shown.

[0037] Figure 14B A schematic representation of an electron arrival event and an overflow cutoff consistent with embodiments of the present disclosure is shown.

[0038] Figure 15 A plot of an estimated miscount rate plotted against an average incoming electron arrival rate consistent with embodiments of the present disclosure is illustrated.

[0039] Figure 16A and 16B A schematic representation of a signal flow in a detection system consistent with embodiments of the present disclosure is illustrated.

[0040] Figure 17 An exemplary representation of a dead time miscount consistent with embodiments of the present disclosure is shown.

[0041] Figure 18 A detection system consistent with embodiments of the present disclosure is shown that is configured to detect a charged particle arrival event with an output that includes event markers.

[0042] Figure 19A and 19B A schematic representation of an exemplary architecture of a detection system consistent with embodiments of the present disclosure is illustrated.

[0043] Figure 20A , Figure 20B and Figure 20C is a plot that illustrates an exemplary representation of an output signal from a sensing element consistent with embodiments of the present disclosure.

[0044] Figure 21A 、 Figure 21B and Figure 21C is a graph illustrating an exemplary representation of an output signal from a sensing element with respect to a threshold value, consistent with embodiments of the present disclosure.

[0045] Figure 22 is a graph illustrating an energy spectrum, consistent with embodiments of the present disclosure.

[0046] Figure 23 is a graph illustrating a plurality of pixels and corresponding sensing element detection signal outputs, consistent with embodiments of the present disclosure.

[0047] Figure 24 is a flowchart representing an exemplary method, consistent with embodiments of the present disclosure.

[0048] Figure 25 is a flowchart representing an exemplary method, consistent with embodiments of the present disclosure.

[0049] Figure 26 is a flowchart representing an exemplary method, consistent with embodiments of the present disclosure.

[0050] Figure 27 is a flowchart representing an exemplary method, consistent with embodiments of the present disclosure.

[0051] Figure 28A is a graph illustrating a charged particle beam scanned in a raster pattern, consistent with embodiments of the present disclosure.

[0052] Figure 28B 、 28C , 28D, and 28E are graphs illustrating a sensing element array at a particular scan time point, consistent with embodiments of the present disclosure.

[0053] Figure 29A and Figure 29B illustrate a first mode of operating a detector, consistent with embodiments of the present disclosure.

[0054] Figure 30A and Figure 30B illustrate a second mode of operating a detector, consistent with embodiments of the present disclosure.

[0055] Figure 31A and Figure 31B illustrate a third mode of operating a detector, consistent with embodiments of the present disclosure.

[0056] Figure 32A and Figure 32B illustrate a fourth mode of operating a detector, consistent with embodiments of the present disclosure.

[0057] Figure 33A andFigure 33B FIG. 1 illustrates detection signals according to a first reset mode, consistent with embodiments of the present disclosure.

[0058] Figure 34A and Figure 34B FIG. 1 illustrates detection signals according to a first reset mode, consistent with embodiments of the present disclosure.

[0059] Figure 35A and Figure 35B FIG. 1 illustrates detection signals according to a first reset mode, consistent with embodiments of the present disclosure.

[0060] Figure 36 FIG. 1 illustrates detection signals according to a first reset mode, consistent with embodiments of the present disclosure. DETAILED DESCRIPTION

[0061] Reference will now be made in detail to the example embodiments, examples of which are illustrated in the accompanying drawings. The following description refers to the accompanying drawings in which the same numbers represent the same or similar elements between the several drawings. The implementation set forth in the following description of example embodiments is not meant to be an all inclusive description of all implementations consistent with the present disclosure. Rather, it is a description of one or more implementations consistent with the subject matter claimed and that the changes in, or combinations of, the illustrations schematically disclosed can be made by those skilled in the art, and that such changes, combinations, and / or alternations are readily implemented, to accomplish ever-larger processes and / or systems.

[0062] Various aspects of the present application relate to systems and methods for charged particle beam detection. The systems and methods can employ counting of charged particles, such as electrons, and can be useful in inspection tools, such as scanning electron microscopes (SEMs). The inspection tools can be used in the manufacturing process of integrated circuit (IC) components. To achieve enhanced computing power of modern electronic devices, the physical size of the devices can be scaled down, while the packing density of circuit components, such as transistors, capacitors, diodes, etc., on IC chips can be greatly increased. For example, in a smart phone, an IC chip, which can be the size of a thumbnail, can include over 2 billion transistors, each smaller than one thousandth of a human hair. Not surprisingly, semiconductor IC manufacturing is a complex process requiring hundreds of individual steps. Even an error in one step can greatly affect the functioning of the final product. Even one “killer defect” can cause the device to fail. The goal of the manufacturing process is to improve the overall yield of the process. For example, to have a 75% yield for a process of 50 steps, the yield of each individual step must be greater than 99.4%, and if the yield of a single step is 95%, the overall process yield drops to 7%.

[0063] While maintaining high throughput (e.g., defined as the number of wafer processes per hour), the ability to ensure detection of defects with high accuracy and high resolution becomes increasingly important. The presence of defects can impact high process yields and high wafer throughput, especially when operator intervention is involved. Thus, detection and identification of micron and nanometer sized defects by inspection tools, such as SEMs, is important to maintain high yields and low costs.

[0064] In some inspection tools, a sample can be inspected by scanning a high-energy electron beam over a surface of the sample. Due to interactions at the sample surface, secondary electrons or backscattered electrons can be generated by the sample, which can then be detected by a detector.

[0065] As mentioned above, detectors of the related art can have limitations, such as poor signal-to-noise ratio (SNR) or poor durability. Various aspects of the present disclosure can address some of these limitations by providing a detector with an array of detector elements, each detector element including a sensing element, and an area of each sensing element such that no more than a certain number of charged particles are received at each sampling period of the sensing element. The detector can include circuitry coupled to each sensing element, which can enable charged particle counting. Charged particle counting can allow for simpler and smaller components to be packaged on a chip relative to, for example, analog signal detection, allowing for robust and reliable detection of charged particles with good SNR. Although some example embodiments are discussed in the context of electrons, it is to be understood that the present disclosure can be applicable to other types of charged particles, such as ions.

[0066] To help ensure accurate electron counting, the time interval between subsequent electron arrival events can be an important parameter. If the distance between electron arrival events is too close, the detector can be overwhelmed and discrimination of individual electron arrival events can be hindered. Similarly, the signal pulse width can be another important parameter that limits electron counting, which can be related to the pulse width of a signal generated in response to an electron arrival event at the detector. If the signal generated by the detector is too weak or too wide (as opposed to a sharp spot), the signal from a subsequent electron arrival event can merge into one signal. Additionally, the sampling rate of the detector should be high enough so that individual electron arrival events can be captured. That is, the detector should be fast enough so that electron arrival events can be detected. Another consideration for electron counting can be to achieve accuracy, with a level of miscounting that can not exceed a certain degree. Miscounting can be based on a dead time of the detector element. Thus, a number of criteria can be relevant in configuring a detector for electron counting.

[0067] As one example of a sensing element, a PIN diode can be provided. A PIN diode can be well suited for electron counting. A PIN diode can have a relatively high natural internal gain, and thus, even in the case of a single electron arrival event, a strong measurable signal can be generated that can be easily distinguished from the relatively low floor of background noise. The need to provide an amplifier or complex system on the chip, such as an avalanche diode, to boost the signal can be reduced or eliminated. Instead, the signal generated by the PIN diode itself or with a relatively low gain amplifier can be well suited for electron counting because it is generated quickly in response to an electron arrival event and stands out from the background noise.

[0068] However, a single detector element including a PIN diode with one output can not be able to handle counting of a full range of beam currents. For example, for a 1 nA electron beam, it is known that approximately 64 electrons can be incident on the detector in a typical 10 ns sampling period. In some SEM systems, the detector can be operated at a sampling rate of 100 MHz, thus corresponding to a sampling period of 10 ns. In one sampling period of 10 ns, 64 electron arrival events can occur, and thus, the signals generated by the individual electron arrival events can not be easily discriminated. Even in high speed detectors, such as one operated at a sampling rate of 800 MHz, there can be approximately 8 incident electrons per sampling period, which can overload the detector.

[0069] In some embodiments of the present disclosure, the size of the sensing elements in the array can be determined such that each sampling period receives no more than a certain number of charged particles in the area of the individual sensing elements. The certain number can be one. The size of the sensing elements can be less than the geometric spread of the charged particles incident on the detector. Thus, the individual sensing elements can be configured to receive fewer charged particles than the total number of charged particles incident on the detector. Various aspects of the detector can be set in accordance with various criteria in order to enable charged particle counting, such as the size of the sensing elements, the sampling rate, and other characteristics.

[0070] Without limiting the scope of the present disclosure, some embodiments can be described in the context of a detector and detection method provided in a system utilizing an electron beam. However, the present disclosure is not limited thereto. Similarly, other types of charged particle beams can be applied. Furthermore, the systems and methods for detection can be used in other imaging systems, such as optical imaging, photon detection, x-ray detection, ion detection, etc.

[0071] As used herein, the term “or” encompasses all possible combinations except where infeasible, unless otherwise specifically stated. For example, if a component is stated as including A or B, then unless infeasible or otherwise specifically stated, the component can include A or B or A and B. As a second example, if a component is stated as including A, B, or C, then unless infeasible or otherwise specifically stated, the component can include A or B or C or A and B or A and C or B and C or A and B and C.

[0072] As used throughout the present disclosure, the expression “a first number of sensing elements” shall mean a group of a first number of sensing elements from a super set of sensing elements. For example, a plurality of sensing elements can be provided. The first number of sensing elements can refer to a subset of the plurality of sensing elements. The first number can be from one to the total number of sensing elements in the plurality. Similarly, the expression “a second number of sensing elements” shall mean a group of a second number of sensing elements from the super set of sensing elements. The second number of sensing elements can refer to a further subset of the plurality of sensing elements, e.g., a subset within the first number of sensing elements.

[0073] Additionally, the term “detector element” can include or cover “sensing element,” “sensor element,” “detection unit,” or “detector segment,” among others. The sensing element can be a diode configured to have a depletion region, and in some embodiments discussed herein, the term “sensing element” can exclude avalanche diodes operating in Geiger mode. The detector element can include diodes, interconnections, and circuitry, e.g., it can include front-end electronics. Further, the term “frame” can include or cover “sampling period,” “SEM image pixel period,” or “pixel period,” among others. The SEM image frame can refer to a frame of pixels that can be refreshed on a frame-by-frame basis, while the data frame can refer to a set of data acquired by the detection system over a specified time period.

[0074] Embodiments of the present disclosure can provide a detection method. The detection method includes charged particle counting. For example, in some embodiments, a charged particle detection method for electron microscopy can be provided. The method can be applied to a SEM detection system. The charged particle detection method can be based on electron counting. By counting the number of electrons received within a predefined period, the intensity of the incoming electron beam can be determined. The term “incoming electron” can include or cover incident electrons, such as electrons that impinge on a surface of the detector. According to some embodiments, noise from the charged particle detection process can be reduced. However, merely improving SNR can not meet the ever-growing demands of various SEM applications.

[0075] In some embodiments, a semiconductor electron detector can be provided to not only detect the intensity of an incoming electron beam, but also to identify the proportion of electrons having different energies. For example, in addition to intensity information, the energy spectrum of an incoming electron beam can be acquired.

[0076] Some embodiments can provide display methods that accompany detection methods. For example, in some embodiments, additional dimensions of information can be provided, such as energy spectrum information. In some applications, a method can be applied to generate a color SEM image.

[0077] Electron counting can involve determining individual electron arrival events that occur at a detector. For example, an electron can be detected one-to-one as it arrives at a detector. In some embodiments, an electron that is incident on a detector can generate an electrical signal that is routed to signal processing circuitry and then read out to an interface such as a digital controller. The detector can be configured to resolve the signal generated by the incident electron and distinguish individual electrons as discrete counts.

[0078] In some embodiments, electron counting can be applied in situations where the beam current is very small. For example, an electron beam can be set to irradiate a sample at a low dose. A low current can be used to prevent an electron counting detector from becoming oversaturated by a large current. For example, a large current can have the effect of introducing non-linearity in the detection results. At the same time, for a detector that can be used in an industrial setting, the detector should also be able to handle situations where a large beam current is present.

[0079] Some embodiments can address the above issues. For example, some embodiments can provide a plurality of relatively small sensing elements that can be used to detect an electron beam. Isolation can be provided between adjacent sensing elements so that the probability of one incoming electron reaching its adjacent sensing element from one sensing element can be reduced. In this way, cross-talk between adjacent sensing elements can be reduced.

[0080] In some embodiments, the data frame rate can be set based on a first parameter. The data frame rate can be a rate at which the sensing elements collect incoming electrons from the electron beam for a data frame for imaging. The data frame rate can be set such that a predefined proportion (e.g., A%) of the sensing elements receive at least one incoming electron. The data frame rate can also be expressed by a period (e.g., duration) of the data frame. In addition, the data frame rate can be set based on a second parameter. For example, among the sensing elements that receive at least one incoming electron, only a second predefined proportion (e.g., B%) of the sensing elements can receive more than one electron. In this way, a predefined detection linearity can be maintained while the electron beam with large beam current can be handled. The data frame rate can be a constant value for a particular SEM setting, or can be a varying value that is set to accommodate the signal strength of the electron beam to be detected even at the same SEM setting. As a result, adjacent data frame periods in the time domain can be the same or can be different at the same SEM setting.

[0081] In addition to the adaptive frame, each frame can include information about when the frame starts and when the frame stops. The information about the frame start time and stop time (e.g., frame start time point and frame stop time point) can be used when generating a pixel in a SEM image. For example, each pixel in a SEM image can be generated using the frames acquired during a particular time period. The period (or rate) of SEM image pixel acquisition can be based on a predefined parameter set according to particular requirements. During the period of each SEM image pixel acquisition, one or more frames can be acquired. The number of frames acquired in adjacent SEM image pixel periods can be the same or different.

[0082] In addition to the frame rate adjustment, the systems and methods for charged particle detection can employ adjustments to the structure or settings of the SEM system. For example, to ensure that a predefined A% of a group of sensing elements receive only one electron during the period of each frame, the SEM system can be adjusted such that the electron density within each electron beam spot is more evenly distributed. One such adjustment can be to defocus the projection system in the secondary SEM column in a multi-beam inspection (MBI) system. The projection system can be configured to defocus the beam to a certain extent. In addition, the magnification of the SEM system can be changed to enlarge the beam spot size of the electron beam or beams. The size of each beam spot can be enlarged. The magnification setting can be configured taking into account the crosstalk between beam spots.

[0083] In some embodiments, a statistical analysis can be performed at each frame. For example, after each frame, for each electron beam, in addition to the total number of electrons received during the frame, a statistical result of the received electron energy plotted for a number of electrons for each energy level within the frame can be obtained. The overall numerical output can be used to generate one pixel in a SEM image, such as a grayscale image in a conventional SEM. The total number of electrons can correspond to the grayscale of the pixel. In addition, one pixel of a color SEM image can also be generated. In a color SEM image, color information such as red green blue (RGB) values for each pixel can be determined by the statistical result of the corresponding frame generated in the foregoing manner. In this way, additional degrees of freedom can be added to SEM imaging. Thus, for example, the analysis of a sample under investigation can be enhanced by elucidating other aspects of the sample, such as material properties, microstructure, and alignment between layers.

[0084] In some embodiments, a detection method can be applied to grayscale SEM imaging. The method can include determining a series of thresholds. Instead of or in addition to generating a statistical result of the received electron energy plotted for a number of electrons for each energy level within the frame, information about the thresholds can be generated. For example, three thresholds can be set in a way that the electron energy increases from low to high. The first threshold at the lowest electron energy can be used to identify whether a sensing element has received an electron or its output is caused by interference or dark current, etc. The second threshold with an intermediate electron energy can be used to identify whether an electron received by a sensing element is a secondary electron from a sample or a scattered electron from a sample. The third threshold with the highest electron energy can be used to identify whether a sensing element has received more than one electron during a particular frame. The number of secondary electrons received during a particular frame, the number of scattered electrons received, and the total number of electrons received can be determined. By accumulating the above information pixel by pixel for a SEM image, one or more of the following can be obtained: a SEM image based on all received electrons, a secondary electron SEM image, and a scattered electron SEM image. Such images can be obtained with improved signal-to-noise ratio without the aid of an energy filter.

[0085] In some embodiments, a detector can be formed by using digital circuitry instead of an implementation that requires a large amount of analog circuitry. Thus, various aspects of the implementation of the detector, such as design and manufacturing, can be improved.

[0086] Reference is now made to Figure 1 , Figure 1 An exemplary electron beam inspection (EBI) system 10 that can include a detector consistent with embodiments of the present disclosure is illustrated. The EBI system 10 can be used for imaging. As Figure 1As shown, the EBI system 10 includes a main chamber 11, a load / lock chamber 20, an e-beam tool 100, and an equipment front end module (EFEM) 30. The e-beam tool 100 is located within the main chamber 11. The EFEM 30 includes a first load port 30a and a second load port 30b. The EFEM 30 can include additional load port(s). The first load port 30a and the second load port 30b house a front opening unified pod (FOUP) containing a wafer (e.g., a semiconductor wafer or one or more wafers made of other material(s)) or a specimen (wafers and specimens can be collectively referred to herein as “wafers”) to be inspected.

[0087] One or more robotic arms (not shown) in the EFEM 30 deliver the wafer to the load / lock chamber 20. The load / lock chamber 20 is connected to a load / lock vacuum pump system (not shown) that removes gas molecules in the load / lock chamber 20 to reach a first pressure that is lower than atmospheric pressure. After reaching the first pressure, the one or more robotic arms (not shown) can deliver the wafer from the load / lock chamber 20 to the main chamber 11. The main chamber 11 is connected to a main chamber vacuum pump system (not shown) that removes gas molecules in the main chamber 11 to reach a second pressure that is lower than the first pressure. After reaching the second pressure, the wafer is inspected by the e-beam tool 100. The e-beam tool 100 can be a single-beam system or a multi-beam system. A controller 109 is electrically connected to the e-beam tool 100, and can also be electrically connected to other components. The controller 109 can be a computer configured to perform various controls of the EBI system 10. Although the controller 109 is shown in FIG. 1 as being external to the structure that includes the main chamber 11, the load / lock chamber 20, and the EFEM 30, it is to be understood that the controller 109 can be part of the structure. Figure 1

[0088] Figure 2A FIG. 1 illustrates a charged particle beam device in which the inspection system can include a multi-beam inspection tool that uses multiple primary electron sub-beams to simultaneously scan multiple locations on a specimen.

[0089] As shown in FIG. 2A, the e-beam tool 100A can include an electron source 202, a gun aperture 204, a condenser lens 206, a primary electron beam 210 emitted from the electron source 202, a source conversion unit 212, multiple beamlets 214, 216, and 218 of the primary electron beam 210, a primary projection optical system 220, a wafer stage 222, a wafer 224, a secondary electron detector 226, a secondary electron detector 228, and a controller 230. Figure 2A Figure 2A ​​The electron source 202 can generate primary particles, such as electrons of the primary electron beam 210. A controller, image processing system, or the like can be coupled to the electron detection device 244. The primary projection optics 220 can include a beam separator 222, a deflection scanning unit 226, and an objective lens 228. The electron detection device 244 can include detection sub-areas 246, 248, and 250.

[0090] The electron source 202, the gun aperture 204, the condenser lens 206, the source conversion unit 212, the beam separator 222, the deflection scanning unit 226, and the objective lens 228 can be aligned with a primary optical axis 260 of the electron beam tool 100A. The secondary optical system 242 and the electron detection device 244 can be aligned with a secondary optical axis 252 of the electron beam tool 100A.

[0091] The electron source 202 can include a cathode from which primary electrons can be emitted and an extractor or an anode to extract or accelerate the primary electrons to form the primary electron beam 210 with a crossover (virtual or real) 208. The primary electron beam 210 can be considered to be emitted from the crossover 208. The gun aperture 204 can block peripheral electrons of the primary electron beam 210 to reduce the size of the probe spots 270, 272, and 274.

[0092] The source conversion unit 212 can include an image forming element array (not shown in the Figure 2A The source conversion unit 212 can include an image forming element array (not shown in the Figure 2A Examples of the source conversion unit 212 can be found in U.S. Patent No. 9,691,586; U.S. Publication No. 2017 / 0025243; and International Application No. PCT / EP2017 / 084429, all of which are incorporated by reference herein in their entireties. The image forming element array can include an array of micro-deflectors or micro-lenses. The image forming element array can form a plurality of parallel images (virtual or real) of the crossover 208 with the plurality of beamlets 214, 216, and 218 of the primary electron beam 210. The beam limiting aperture array can limit the plurality of beamlets 214, 216, and 218.

[0093] Condenser lens 206 can focus primary electron beam 210. Changes in the current of beamlets 214, 216, and 218 downstream of source conversion unit 212 can be made by adjusting the focusing power of condenser lens 206 or by changing the radial size of the corresponding beam limiting aperture within the beam limiting aperture array. Condenser lens 206 can be a movable condenser lens that can be configured such that the position of its first principal plane is movable. The movable condenser lens can be configured to be magnetic, which can cause off-axis beamlets 216 and 218 to land on the beam limiting aperture at a rotation angle. The rotation angle varies with the focusing power of the movable condenser lens and the position of the first principal plane. In some embodiments, the movable condenser lens can be a movable anti-rotation condenser lens, which involves an anti-rotation lens with a movable first principal plane. Movable condenser lenses are also described in U.S. Pub. No. 2017 / 0025241, which is incorporated by reference in its entirety.

[0094] Objective lens 228 can focus beamlets 214, 216, and 218 on wafer 230 for inspection and can form a plurality of probe spots 270, 272, and 274 on the surface of wafer 230.

[0095] Beam separator 222 can be a Wien filter type beam separator that generates an electrostatic dipole field and a magnetic dipole field. In some embodiments, the force exerted on the electrons of beamlets 214, 216, and 218 by the electrostatic dipole field can be equal in magnitude and opposite in direction to the force exerted on the electrons by the magnetic dipole field, if they are applied. As a result, beamlets 214, 216, and 218 can pass directly through beam separator 222 at a zero deflection angle. However, the total dispersion of beamlets 214, 216, and 218 generated by beam separator 222 can also be non-zero. Beam separator 222 can separate secondary electron beams 236, 238, and 240 from beamlets 214, 216, and 218 and direct secondary electron beams 236, 238, and 240 toward secondary optical system 242.

[0096] The deflection scanning unit 226 can deflect the beamlets 214, 216, and 218 so that the probe spots 270, 272, and 274 scan over the surface area of the wafer 230. In response to the incidence of the beamlets 214, 216, and 218 at the probe spots 270, 272, and 274, secondary electron beams 236, 238, and 240 can be emitted from the wafer 230. The secondary electron beams 236, 238, and 240 can include electrons having an energy distribution, including secondary electrons and backscattered primary electrons. The secondary optics system 242 can focus the secondary electron beams 236, 238, and 240 on the detection sub-areas 246, 248, and 250 of the electron detection device 244. The detection sub-areas 246, 248, and 250 can be configured to detect the corresponding secondary electron beams 236, 238, and 240 and generate corresponding signals for reconstructing an image of the surface area of the wafer 230.

[0097] Although Figure 2A The electron beam tool 100 is shown as an example of a multi-beam tool that uses multiple beamlets, but embodiments of the present disclosure are not limited thereto. For example, the electron beam tool 100 can also be a single-beam tool that uses only one primary electron beam to scan one location on a wafer at a time.

[0098] As Figure 2B shown, the electron beam tool 100B can be a single-beam inspection tool used in the EBI system 10. The electron beam tool 100B includes a wafer holder 136 supported by a motorized stage 134 to hold a wafer 150 to be inspected. The electron beam tool 100B includes an electron emitter, which can include a cathode 103, an anode 121, and a gun aperture 122. The electron beam tool 100B also includes a beam limiting aperture 125, a condenser lens 126, a cylindrical aperture 135, an objective lens assembly 132, and a detector 144. In some embodiments, the objective lens assembly 132 can be a modified SORIL lens that includes a pole piece 132a, a control electrode 132b, a deflector 132c, and a kick coil 132d. During imaging, an electron beam 161 emitted from a tip of the cathode 103 can be accelerated by the anode 121 voltage, pass through the gun aperture 122, the beam limiting aperture 125, the condenser lens 126, and be focused by the modified SORIL lens into a probe spot 170 and impinge onto a surface of the wafer 150. The probe spot 170 can be scanned over the surface of the wafer 150 by a deflector, such as the deflector 132c in the SORIL lens or other deflector. The detector 144 can collect secondary or scattered primary particles, such as secondary electrons or scattered primary electrons, emitted from the wafer surface to determine the intensity of the beam so that an image of an area of interest on the wafer 150 can be reconstructed.

[0099] An image processing system 199 can also be provided that includes the image acquirer 120, the storage device 130, and the controller 109. The image acquirer 120 can include one or more processors. For example, the image acquirer 120 can include a computer, a server, a mainframe, a terminal, a personal computer, any kind of mobile computing device, etc., or a combination thereof. The image acquirer 120 can be connected to the detector 144 of the e-beam tool 100B through a medium such as an electrical conductor, a fiber optic cable, a portable storage medium, IR, Bluetooth, the Internet, a wireless network, a radio station, or a combination thereof. The image acquirer 120 can receive signals from the detector 144 and can construct an image. The image acquirer 120 can thus acquire an image of the wafer 150. The image acquirer 120 can also perform various post-processing functions such as generating a contour, superimposing an indicator on the acquired image, etc. The image acquirer 120 can be configured to perform adjustment of brightness and contrast, etc., of the acquired image. The storage device 130 can be a storage medium such as a hard disk, a random access memory (RAM), a cloud storage device, other types of computer-readable storage, etc. The storage device 130 can be coupled with the image acquirer 120 and can be used to save the scanned raw image data as a raw image, as well as post-processed images. The image acquirer 120 and the storage device 130 can be connected to the controller 109. In some embodiments, the image acquirer 120, the storage device 130, and the controller 109 can be integrated together as one electronic control unit.

[0100] In some embodiments, the image acquirer 120 can acquire one or more images of the sample based on the imaging signals received from the detector 144. The imaging signals can correspond to a scanning operation for charged particle imaging. The acquired image can be a single image that includes a plurality of imaging areas that can contain various features of the wafer 150. The single image can be stored in the storage device 130. Imaging can be performed based on imaging frames.

[0101] The condenser and illumination optics of the e-beam tool can include or be supplemented by electromagnetic quadrupole electron lenses. For example, as shown in FIG. 1B, the e-beam tool 100B can include a first quadrupole lens 148 and a second quadrupole lens 158. In some embodiments, the quadrupole lenses are used to control the electron beam. For example, the first quadrupole lens 148 can be controlled to adjust the beam current, and the second quadrupole lens 158 can be controlled to adjust the beam spot size and beam shape. Figure 2B

[0102] Figure 2B A charged particle beam device is illustrated in which the inspection system can use a single primary beam that can be configured to generate secondary electrons by interacting with the wafer 150. The detector 144 can be placed along the optical axis 105, as in FIG. 1A.​Figure 2B As in the illustrated embodiment, the primary electron beam can be configured to travel along the optical axis 105. Therefore, the detector 144 can include a hole at its center, allowing the primary electron beam to pass through and reach the wafer 150. Figure 3G An example of a detector 144 with an opening 145 at its center is shown. However, some embodiments may use a detector positioned off-axis relative to the optical axis along which the primary electron beam travels. For example, as in Figure 2A As in the illustrated embodiment, a beam splitter 222 can be provided to direct the secondary electron beam toward an off-axis detector. The beam splitter 222 can be configured to shift the secondary electron beam by an angle α.

[0103] Now refer to Figure 2C Another example of a charged particle beam device is discussed. Electron beam tool 100C can be an example of electron beam tool 100, and can be similar to... Figure 2A The electron beam tool 100A shown is shown.

[0104] like Figure 2C As shown, beam splitter 222 can be a Wien filter type beam splitter that generates electrostatic dipole fields and magnetic dipole fields. In some embodiments, if this beam splitter is applied, the force exerted by the electrostatic dipole field on the electrons of beam waves 214, 216, and 218 on the electrons may be equal in amplitude and opposite in direction to the force exerted by the magnetic dipole field on the electrons. Therefore, beam waves 214, 216, and 218 can pass directly through beam splitter 222 with zero deflection angle. However, the total dispersion of beam waves 214, 216, and 218 generated by beam splitter 222 can also be non-zero. Regarding the dispersion plane 224 of beam splitter 222, Figure 2C A beam 214 with a nominal energy V0 and energy dispersion ΔV is shown dispersed into three portions: a beam portion 262 corresponding to energy V0, a ​​beam portion 264 corresponding to energy V0 + ΔV / 2, and a beam portion 266 corresponding to energy V0 - ΔV / 2. The total force exerted by the beam splitter 222 on the electrons of the secondary electron beams 236, 238, and 240 can be non-zero. The beam splitter 222 can separate the secondary electron beams 236, 238, and 240 from beams 214, 216, and 218, and direct the secondary electron beams 236, 238, and 240 toward the secondary optical system 242.

[0105] A semiconductor electron detector (sometimes referred to as a "PIN detector") can be used in the electron beam tool 100 of the EBI system 10. The EBI system 10 can be a high-speed wafer imaging SEM including an image processor. The electron beam generated by the EBI system 10 can illuminate the surface of a sample or penetrate the sample. The EBI system 10 can be used to image the surface of a sample or the structure beneath it, such as for analyzing layer alignment. In some embodiments, for example by comparing the SEM image with an SEM image of a device layout pattern or the same pattern at other locations on the wafer being inspected, the EBI system 10 can detect and report process defects related to the manufacture of a semiconductor wafer. The PIN detector can include a silicon PIN diode that can operate with a negative bias. The PIN detector can be configured such that incoming electrons generate relatively large and distinct detection signals. In some embodiments, the PIN detector can be configured such that incoming electrons can generate multiple electron-hole pairs, while photons can generate only one electron-hole pair. PIN detectors for electron counting can have many differences compared to photodiodes used for photon detection, as will be discussed below.

[0106] Now refer to Figure 3A The illustration shows a schematic representation of an exemplary structure of detector 300. Detector 300 may be provided as a reference. Figure 2A , Figure 2B and Figure 2C Detector 144 or electronic detection device 244. Although in Figure 3A An array is shown, but it is important to understand that detector 300 may include multiple arrays, such as one array for each secondary electron beam.

[0107] Detector 300 may include an array of sensing elements, including sensing elements 311, 312, and 313. The sensing elements may be arranged as a planar two-dimensional array, the plane of which is substantially perpendicular to the incident direction of the incoming charged particles. In some embodiments, detector 300 may be arranged to be tilted relative to the incident direction.

[0108] Detector 300 may include a substrate 310. Substrate 310 may be a semiconductor substrate that may include a sensing element. The sensing element may be a diode. The sensing element may also be a diode-like element that converts incident energy into a measurable signal. The sensing element may include, for example, a PIN diode, an avalanche diode, an electron multiplier tube (EMT), or combinations thereof. A region 325 may be provided between adjacent sensing elements. Region 325 may be an isolation region to isolate the sides or corners of adjacent sensing elements from each other. Region 325 may include an insulating material that is a different material from the material of other areas of the detection surface of detector 300. Region 325 may be provided as... Figure 3AThe area 325 can be provided as a cross-shaped area seen in plan view of the sensing element. The area 325 can be provided as a square. In some embodiments, the area 325 can not be provided between adjacent sides of the sensing element. For example, in some embodiments, an isolation area can not be provided on the detection surface of the detector.

[0109] The sensing element can generate an electrical signal corresponding to the charged particles received in the active area of the sensing element. For example, the sensing element can generate a current signal corresponding to the energy of the received electrons. The pre-processing circuitry can convert the generated current signal to a voltage, which can represent the intensity of the electron beam spot or a portion thereof. The pre-processing circuitry can comprise, for example, preamplifier circuitry. The preamplifier circuitry can comprise, for example, a charge transfer amplifier (CTA), a transimpedance amplifier (TIA), or an impedance conversion circuit coupled with the CTA or TIA. In some embodiments, signal processing circuitry can be provided, which provides an output signal in arbitrary units in time. One or more substrates, such as dies, can be provided, which can form circuit layers for processing the output of the sensing elements. The dies can be stacked together in the thickness direction of the detector. Other circuitry can also be provided for other functions. For example, switch actuation circuitry can be provided, which can control the switching elements for connecting the sensing elements to each other.

[0110] Reference is now made to Figure 3B which shows a schematic illustration of a cross-sectional structure of a substrate 310, which can be an example of a structure included in a PIN detector. The substrate 310 can comprise one or more layers. For example, the substrate 310 can be configured to have multiple layers stacked in a thickness direction, which is substantially parallel to the direction of incidence of the electron beam. In some embodiments, the substrate 310 can have multiple layers stacked in a direction perpendicular to the direction of incidence of the electron beam. The substrate 310 can be provided with a sensor surface 301 for receiving the incident charged particles. Sensing elements, such as sensing elements 311, 312, and 313, can be provided in a sensing layer of the substrate 310. An area 325 can be provided between adjacent sensing elements. For example, the substrate 310 can comprise trenches or other structures made of or filled with insulating material. In some embodiments, the area 325 can extend completely or partially through the substrate 310.

[0111] As shown in Figure 3C In some embodiments, the area 325 can not be provided between the sensing elements. For example, in cross-sectional view, no insulating material can be provided between the sides of adjacent sensing elements. The multiple sensing elements can be continuous in cross-sectional view. Isolation between adjacent sensing elements can still be achieved by other means, such as by controlling the electric field. For example, the electric field can be controlled between each sensing element.

[0112] Although the figures can show the sensing elements 311, 312, and 313 as discrete units, such a division can not actually exist. For example, the sensing elements of a detector can be formed from a semiconductor device that constitutes a PIN diode device. A PIN diode device can be fabricated as a substrate having multiple layers including a p-type region, an intrinsic region, and an n-type region. One or more of such layers can be continuous in a cross-sectional view. However, in some embodiments, the sensing elements can be provided with a physical separation between them. For example, in addition to a sensor layer, other layers such as a circuit layer and a readout layer can be provided.

[0113] As one example of yet another layer, the detector 300 can be provided with one or more circuit layers adjacent to the sensor layer. The one or more circuit layers can include wiring conductors, interconnects, and various electronic circuit components. The one or more circuit layers can include a processing system. The one or more circuit layers can include signal processing circuitry. The one or more circuit layers can be configured to receive output currents detected from the sensing elements in the sensor layer. For example, the one or more circuit layers and the sensor layer can be provided in the same or separate dies.

[0114] Figure 3D and 3E A schematic illustration of an individual sensing element is shown, which can be an example of one of the sensing elements 311, 312, and 313. For example, in Figure 3D , a sensing element 311A is shown. The sensing element 311A can include a semiconductor structure of a p-type layer 321, an intrinsic layer 322, and an n-type layer 323. The sensing element 311A can include two terminals such as an anode and a cathode. The sensing element 311A can be reverse biased, and a depletion region 330 can form and can span a portion of the length of the p-type layer 321, substantially the entire length of the intrinsic layer 322, and a portion of the length of the n-type layer 323. In the depletion region 330, charge carriers can be removed, and new charge carriers generated in the depletion region 330 can be swept away according to their charge. For example, when an incoming charged particle reaches the sensor surface 301, an electron-hole pair can be created, and holes 351 can be attracted toward the p-type layer 321, while electrons 352 can be attracted toward the n-type layer 323. In some embodiments, a protective layer can be provided on the sensor surface 301.

[0115] As Figure 3EAs shown, in addition to changing orientation, sensing element 311B can operate in a similar manner as sensing element 311A. For example, p-type layer 321 can include sensor surface 301. P-type layer 321 can be exposed to incident charged particles. As a result, incident charged particles can interact with p-type layer 321 and depletion region 330 and can generate electron-hole pairs. In some embodiments, a metal layer can be provided on top of p-type layer 321.

[0116] In operation, the depletion region of a detection element can act as a capture region. Incoming charged particles can interact with the semiconductor material in the depletion region and generate new charges. For example, a detection element can be configured such that a charged particle having an amount or greater energy can cause an electron of a lattice of the semiconductor material to be ejected, creating an electron-hole pair. The resulting electron and hole can be caused to travel in opposite directions due to, for example, an electric field in the depletion region. The generation of charge carriers traveling toward a terminal of a sensing element can correspond to a current flow in the detection element.

[0117] In a comparative example, a photodiode can be configured to generate a charge in response to receiving a photon. A photon can have an energy corresponding to its wavelength or frequency. Generally, a photon in the visible spectrum can have an energy of about leV. However, in a semiconductor photodiode, approximately 3.6eV can typically be required to generate one electron-hole pair. Thus, a photodiode can encounter difficulties in detecting current generation, such as the following.

[0118] Generally, the energy level of a photon can be similar to that required to generate an electron-hole pair in a semiconductor photodiode. Thus, to stably and reliably generate a current, a photon of high energy can need to be incident on a semiconductor photodiode. When the frequency of a photon is equal to or higher than a certain level, its energy can be sufficient to generate one electron-hole pair.

[0119] Further, the current generated by an electron-hole pair in response to a photon arrival event can be relatively low. The current generated in response to a photon arrival event can not be sufficient to overcome background noise. Some diodes, such as photodiodes biased into avalanche or Geiger counting mode, can employ amplification to generate a higher level of current such that a useful detection signal can be generated. In some embodiments, a photodiode can be biased into an avalanche mode of operation. In some embodiments, amplification can be provided by a gain block attached to the photodiode. Avalanche effects can be generated due to an internal strong electric field caused by a bias voltage. Avalanche effects can be used to achieve amplification due to impact ionization.

[0120] Background noise in a detector can be caused by dark current in a diode, among other factors. For example, imperfections in the crystal structure of a semiconductor device acting as a diode can cause fluctuations in current. Dark current in a detector can be due to defects in the material forming the detector, which can produce current even in the absence of incident radiation. The “dark” current can refer to fluctuations in current that are independent of any incoming charged particles.

[0121] A diode can be configured to generate electron-hole pairs when a particle (e.g., a photon) having not less than a certain energy level enters the diode. For example, a photodiode can generate only electron-hole pairs when a photon having not less than a certain energy level enters the photodiode. This can be due to, for example, the bandgap of the material forming the photodiode. A photon having an energy equal to the certain level can only generate one electron-hole pair, and even if the photon has more energy beyond the certain level, it can still only generate one electron-hole pair. Additional electron-hole pairs can not be generated. Meanwhile, an electron detector can be configured such that every time an electron enters a depletion region of a detector sensing element, which can include a diode, an electron-hole pair can begin to be generated as long as the electron has not less than a certain amount of energy (e.g., approximately 3.6 eV). If the electron has more energy than the certain amount, then more electron-hole pairs can be generated during the arrival event of the incoming electron.

[0122] In a diode configured for photon detection, defects in the diode can cause random generation of electron-hole pairs in the diode due to, for example, imperfections in the lattice of the semiconductor structure. Dark current can be amplified by amplification effects such as avalanche amplification. Signals produced by the dark current can continue to be input to a counting circuit, where they can be recorded as arrival events. Such events can be referred to as “dark counts.” In addition, the amplifier itself can produce noise. Thus, various sources of noise, such as dark current, thermal energy, extraneous radiation, etc., can all contribute to unexpected fluctuations in current in the detector output.

[0123] In contrast to photons, electrons can have significantly more energy that can be used to generate signals in a diode. An incident electron on a sensing element of a detector can have a much greater energy than the threshold energy level required to generate an electron-hole pair in the sensing element. Thus, an incident electron can generate many electron-hole pairs in the sensing element.

[0124] Reference is now made to Figure 3Fwhich shows an exemplary view of a charged particle beam spot 500 (also referred to as beam spot 500) received on a detector 300. The beam spot 500 can have a circular shape without trajectory deviations, as illustrated. In some embodiments, the beam spot can have a shape different from a circular shape. For example, in a single beam system, the beam spot can have a shape deviating from a circular shape due to aberrations. Further, in some embodiments, multiple beam spots can be incident on the detector, such as in a multi-beam system. The beam spots can deviate from a circular shape in terms of, for example, position, shape, and grid step (e.g., spacing between beam spots when multiple beam spots are formed). For example, the deviations can be caused by aberrations, dispersion, drifts in the electron optical system, or imperfections of components.

[0125] In some embodiments, the detection system can comprise a controller, which can be configured to determine charged particle incidence on the detector. The controller can be configured to determine a number of charged particles incident on a sensing element of the detector within a frame. For example, the controller can perform a charged particle count, such as an electron count. The charged particle count can be performed frame by frame. The detector can be configured such that individual sensing elements, such as sensing elements 311, 312, and 313 of Figure 3A the detector can output a detection signal over time. The detection signal can be transmitted to the controller. The detection signal can be a signal in amperes, volts, or arbitrary units, for example, corresponding to the energy of the electrons received at the respective sensing element. The controller can determine a discrete number of charged particles to have reached the sensing element based on the detection signal. The number of charged particles can be resolved as an integer.

[0126] The controller can be configured to determine a first group of sensing elements of a plurality of sensing elements provided in the detector based on a first grouping criterion. The first grouping criterion can comprise, for example, a condition that at least one charged particle is incident on each sensing element of a first number of sensing elements of the detector. The first number can be a raw number or a proportion of the sensing elements. The controller can be configured to determine the first group over time within a period of one frame. The determination can be performed repeatedly over a plurality of frames such that the controller has a frame rate for performing the processing, such as a charged particle count determination within each frame. The controller can further determine a boundary line. For example, as shown in Figure 3F the boundary line 350 can be determined. The boundary line 350 can be provided to enclose the sensing elements that received at least one charged particle. The sensing elements contained within the boundary line 350 can be at least partially covered by the same charged particle spot.

[0127] The beam spot 500 can have a well-defined center or locus. Near the center of the beam spot 500, the intensity can be higher than near the periphery. The difference in intensity can be attributed to various factors, including tip size of the electron source 202, aberrations of the electron-optical system, electron dispersion, and other parameters of the electron beam tool 100A, etc. Also, in some embodiments, the variation in intensity can be caused by the sample topography, material (e.g., in the case of backscattered electrons), charge conditions on the sample surface, landing energy, etc. of the scattered electrons. Thus, the area of high intensity can not necessarily be at the center of the beam spot 500.

[0128] In the area of the beam spot 500 where the intensity is higher, there can be more than one electron incident on a sensing element of the detector. Thus, the controller can be configured to determine a second group of sensing elements based on a second grouping criterion. The second grouping criterion can include a condition that more than one charged particle is incident on each sensing element of a second number of sensing elements. The second group of sensing elements including the second number of sensing elements can be determined in the first group including the first number of sensing elements. That is, the second group can be a subset of the first group. The determination of the second group can be concurrent with the determination of the first group. Thus, the determination of the first group and the second group can be for the same frame. The controller can also determine a second boundary line 360 that encloses the sensing elements that receive more than one charged particle.

[0129] The controller can be configured to determine or adjust a frame rate (or period) at which processing is performed. The processing can correspond, for example, to image processing for generating a SEM image based on the output from the detector. As discussed above, the processing can also include determining the first group of sensing elements and the second group of sensing elements. The period of the first frame can be determined based on the first parameter as follows. The period can be set such that a first predetermined number of sensing elements receive at least one incident charged particle in the first frame. The first predetermined number can be a proportion of the sensing elements in all of the sensing elements of the detector, e.g., A%. The first predetermined number can also be a proportion of the sensing elements in those sensing elements in a particular region of the detector, without necessarily being a proportion of the sensing elements in all of the sensing elements. For example, the first predetermined number can be a proportion of the sensing elements in a first quadrant of the detector. The first predetermined number can also be a raw number, such as X sensing elements.

[0130] Alternatively, the period can be set based on a second parameter. The second parameter could be that a second predetermined number of sensing elements each receive more than one incident charged particle in the first frame. For example, the second parameter could be that among the sensing elements receiving at least one incident charged particle, only a second proportion (e.g., B%) of the sensing elements receive more than one charged particle. The second predetermined number could also be an original number, such as Y sensing elements. The parameters can be adjusted such that the first parameter is satisfied before the second parameter is satisfied.

[0131] The first and second parameters can define the boundary conditions used to determine the period of the first frame. Either the first or second parameter can be used. The first and second parameters can be used together. In addition to determining the period of the first frame, the frame rate of multiple frames can also be determined. The frame rate can be a constant value that can be set based on, for example, a specific SEM setting. Therefore, the frame rate can be the reciprocal of the period of the first frame. The frame rate can also be adaptive, i.e., it can have varying values. An adaptive frame rate can be set to adapt to the signal strength of the detected charged particle beam.

[0132] In some embodiments, the electron beam tool 100 can be configured to distribute the electron density more uniformly within the electron beam spot. For example, the controller 109 can control the electron optics to defocus the electron beam or beam wave. The electron optics can adjust the electron beam (or beam wave) so that its focal point does not coincide with the surface of the detector 144 or the electronic detection device 244. Furthermore, the projection system in the secondary SEM column can be configured to defocus the secondary beam (or beam wave) to a certain extent. Additionally, the magnification of the projection system in the secondary SEM column can be changed to increase the beam spot size of the electron beam or beam wave(s). The size of each beam wave spot can be increased. Crosstalk between beam waves can be configured in the magnification settings.

[0133] Figure 3G An example of a surface that may include a PIN detector 144 is shown. Detector 144 may include a sensor surface 301 arranged to receive charged particles generated from a sample. In some embodiments of this disclosure, the PIN detector may be used as an in-lens detector in a hysteresis objective SEM column of an EBI system 10. The PIN detector may be placed between the cathode for generating the electron beam and the objective. The electron beam emitted from the cathode may be potentialized to -BE keV (typically about -10kV). Electrons in the electron beam may be immediately accelerated and travel parallel through the column. The column may be at a ground potential. Therefore, electrons may travel with a kinetic energy of BE keV as they pass through the opening 145 of detector 144. The electrodes of the objective (such as...) Figure 2BThe electrons of the pole piece 132a) of the objective assembly 132 can be sharply decelerated to a landing energy LE keV, as the wafer surface potential can be set to -(BE-LE) keV.

[0134] The secondary electrons emitted from the wafer surface by the impact of the primary beam electrons can be accelerated by an acceleration field (e.g., the retarding electric field in the vicinity of the wafer can act as an acceleration field for the secondary electrons) and travel back towards the surface of the PIN detector. For example, as shown in Figure 4A due to the interaction with the wafer 150 at the probe point 170, secondary electrons can be generated that travel back towards the detector 144. The secondary electrons emitted from the wafer surface that travel along the optical axis 105 can arrive at the surface of the detector 144 with a position distribution. The landing positions of the secondary electrons can be within a generally circular area with a radius of, for example, a few millimeters. The geometric spread of the landing positions of the secondary electrons can be due to the electrons having different trajectories, which can depend on, for example, the initial kinetic energy and the emission angle of the electrons.

[0135] Figure 4B An example of a secondary electron landing spot distribution on the surface of a detector is illustrated. The electrons 300a can land at different points on the surface of the detector 144, while typically most of the electrons can be clustered around the center portion of the detector 144. The landing spot distribution can shift depending on the secondary emission position and the SEM deflection field (e.g., the scan field). Thus, in some applications, the required size of an in-lens PIN detector can be large if a particular field of view (FOV) of the SEM image is needed. Typically, for example, the diameter of the detector can be 10 mm or more. In some embodiments, the diameter of the detector can be about 4 to 10 mm.

[0136] The electrons that are incident on the detection surface of the PIN detector can be converted to charge. The charge can be collected at the terminals of the PIN detector and used as a detection signal, which can be proportional to the rate of incoming electrons. In an ideal PIN detector, the kinetic energy of the incoming electrons with energy (BE-LE) keV can be completely consumed by creating many electron-hole pairs at a rate of approximately 3.61 eV per pair. Thus, for incoming electrons with an energy of 10,000 eV, approximately 2,700 electron-hole pairs can be created. In contrast to a photon arrival event, which can generate only a single electron-hole pair, an electron arrival event can generate significantly more electron-hole pairs.

[0137] The sensing element can be configured to generate a number of electron-hole pairs in response to an electron arrival event. In some embodiments, the current generated in response to an electron arrival event in the sensing element can be used as a detection current signal. The output of the sensing element in response to an electron arrival event can be used as is or can be relatively small amplified. The need to provide amplification can be reduced or omitted. Omitting or providing reduced amplification can be beneficial to reduce noise. Furthermore, an amplifier can indiscriminately apply amplification to all signals generated in the diode. Thus, even so-called “dark counts” can be amplified and can lead to false detection signals.

[0138] In some embodiments according to the present disclosure, a dark current can generate only a small output compared to a charged particle that the sensing element is configured to detect. For example, a dark current can be caused by a dislocation in the lattice of the semiconductor structure of the diode, which can allow an electron to be displaced. Thus, in some cases, a dark current can cause only a single electron-hole pair to be generated in the sensing element. However, as discussed above, in a sensing element configured to generate a number of electron-hole pairs in response to the arrival of an equally charged particle, such as a secondary electron, approximately 3,000 electron-hole pairs can be generated. Thus, the signal to dark current noise ratio can be approximately 3,000: 1.

[0139] A semiconductor diode, such as a diode having a PIN structure, can be operated in various modes. For example, in a first mode, the diode can be operated with a normal reverse bias. In this mode, each incoming photon having a sufficiently high energy can generate only one electron-hole pair. When the external radiation, e.g. incoming photons, disappear, the current flow in the diode can stop immediately.

[0140] In a second mode of operating the diode, the diode can be operated with a higher reverse bias than in the first mode. The second mode can introduce impact ionization. This can also be referred to as an avalanche photodiode mode. In this mode, each incoming photon having a sufficiently high energy can generate one electron-hole pair. This one pair can then be multiplied with an avalanche gain due to internal impact ionization, such that eventually several electron-hole pairs can be generated. Thus, each incoming photon can cause several electron-hole pairs to be generated. When the external radiation disappears, the current flow in the diode can stop immediately. The second mode can include a linear region and a non-linear region.

[0141] In a third mode of operation of the diode, the diode can be operated at an even higher reverse bias than in the second mode. The third mode can introduce stronger impact ionization. The third mode can enable photon counting. The third mode can include a Geiger counting mode. In the third mode, each incoming photon with a high enough energy can generate one electron-hole pair. This one pair can then be multiplied by an avalanche gain due to internal impact ionization, such that several electron-hole pairs can eventually be generated. Thus, each incoming photon can result in the generation of several electron-hole pairs. The multiplication process can continue due to a strong internal electric field from the high reverse bias. The multiplication can be self-sustaining. When the external radiation disappears, the current flow in the diode can not necessarily stop. The current in the diode can be stopped by disconnecting the diode from the power supply. After the disconnection, the current in the diode can then decay. The current output of the diode operated in the third mode can exhibit behavior including a long tail. For example, the output can decrease gradually after an initial peak. In the third mode, the diode can be provided with a quenching circuit. The quenching circuit can include a passive or active quenching circuit. Actuating the quenching circuit can allow the diode to be turned off after each photon arrival event. Quenching can be used to reset the diode.

[0142] The diode can be configured to operate at a gain level. For example, the diode can be configured to operate at a gain lower than 100. This can refer to the gain imparted by the operation of the diode through the application of a voltage. The gain can amplify a signal by a factor of up to, for example, 100 with respect to its initial strength. It is to be understood that other specific levels of gain can also be used.

[0143] Using a gain effect, such as by a diode biased into an avalanche mode or a Geiger counting mode, can involve time-dependent phenomena. For example, a diode biased into an avalanche mode can impart a gain through avalanche multiplication. There can be a finite time associated with the gain effect. The speed of the diode can be related to the time it takes for the gain effect to occur. The speed of a diode biased into an avalanche mode but not into a Geiger counting mode can be at least equal to the speed of the diode under normal biasing conditions. The speed of a diode biased into an avalanche mode can also be higher than the speed of the diode under normal biasing conditions. In some cases, there can be a recovery time after an arrival event of a charged particle at the diode. A diode operated in a Geiger counting mode can have an associated recovery time. The recovery time can limit the ability of the diode to detect discrete signals in succession. A diode operated in a Geiger counting mode can need to be quenched after an arrival event of a charged particle in order to accurately detect the next event.

[0144] For example, if detectable events occur in succession, problems can be encountered when applying a gain effect to amplify the signals of subsequent events following a first event, as the initial avalanche and its associated effects are still ongoing. In contrast to conventional diodes operated in avalanche mode, detectors according to some embodiments of the present disclosure can address issues related to recovery time. For example, as should be discussed in more detail below, a PIN detector can be configured to generate electron-hole pairs at high gain without requiring a reverse bias to, for example, avalanche mode or Geiger mode. The gain provided in a PIN detector can be related to the kinetic energy of incoming charged particles, such as electrons. The detector can include a sensing element having a PIN structure and circuitry. The need to provide quenching circuitry can be omitted. For example, the detector can be configured to generate electron-hole pairs corresponding to pulses lasting about 3 to 5 ns or less.

[0145] In an example PIN detector, holes can be excited in a depletion region in the intrinsic region of the PIN detector, and can drift toward the anode due to a field created in the PIN detector by a reverse bias. The holes can then be collected at the anode. Electrons generated in the depletion region can drift in a direction opposite to the holes. Thus, the electrons can be collected at the cathode, which can be grounded. The holes and electrons created in the depletion region can recombine with opposite charges within the PIN detector. The recombination rate outside the depletion region can be high. The depletion region can include a portion of the P+ region, which can act as the anode due to the reverse bias. Recombination of holes or electrons on a side of the P+ region on which incident electrons enter the detector can result in energy loss without affecting the detector current at the anode terminal. Thus, it can be desirable to configure an electrode located on a side on which incident electrons enter the detector to be thin, for example, to reduce energy loss. For example, in a PIN detector, it can be desirable to configure the P+ layer thickness to be as thin as possible.

[0146] The reverse bias applied to the PIN detector can involve a voltage application. The diode can be configured to operate at a certain amount or less of reverse voltage. In some embodiments, the certain amount can be 100 volts. The diode can operate within a linear region.

[0147] In some embodiments, both secondary electrons and backscattered electrons can reach the detector. For example, in a comparative example, about 20% to 30% of the incoming electrons on a PIN detector can be backscattered electrons with an energy approximately equal to the energy of the electrons in the primary beam (e.g., BE). The backscattered electrons can be the same as the electrons included in the primary beam generated by the electron source, except that they are backreflected from the sample without losing a substantial amount of energy.

[0148] Furthermore, some of the electrons that have not been backscattered can lose their kinetic energy by causing the lattice atoms in the PIN detector (e.g. Si atoms in a silicon substrate) to emit their characteristic X-ray photons. Other excitations, such as phonons, can also be generated. Thus, the number of charges created by a single incoming electron with a fixed kinetic energy can vary. That is, the electron gain (e.g. the number of charges collected at the terminals of the diode per incoming electron) can vary between incoming electrons. However, in a typical PIN detector, even though the electron gain varies, as discussed above, it should not exceed the electron gain of an ideal PIN detector. Typically, the distribution of the actual electron gain has a distinct peak at gain 0, indicating a loss of detection due to electron scattering of the Si crystal.

[0149] The charges collected at the terminals of the PIN detector can form a current signal. When the electron beam is scanned over the wafer surface, the current signal can follow the modulation of the incoming electron rate.

[0150] Figure 5 A schematic representation of an electron arrival event and its relation to the current signal is shown. In each of the three graphs of Figure 5 the t-axis represents time. In the top graph of Figure 5 a single electron 501 can be incident on the surface of the PIN detector at a certain point in time. As Figure 5 indicated in the middle graph of a number of theoretical signal pulses 502 can be generated in the PIN detector in response to the electron arrival event. The y-axis can represent signal strength in arbitrary units. The signal pulses can rise and fall over time, and the pulse width can be determined, for example, based on the characteristics of the PIN detector and the circuitry connected to the PIN detector. As discussed above, the signal pulses can have an amplitude that can vary depending on how many electron-hole pairs are generated in response to the electron arrival event. The signal pulses 502 can represent, for example, an ideal signal pulse, which corresponds to the generation of an electron-hole pair at a rate of 3.61 eV per pair in response to the electron arrival event. Since electrons can arrive at the detector surface continuously, the signal pulses can overlap each other. It is to be understood that the illustrated signal pulses are only schematic.

[0151] In one sampling period τ s there can be multiple electrons incident on the detector continuously, and thus, the signal pulses of the individual electrons can substantially overlap when the current signal is read. The output signal when the signal is read can correspond to the current generated in response to the multiple electron arrival events. As Figure 5As shown in the bottom plot, the single output signal at time T2 can correspond to the electron arrival rate at that moment in time. As the electron beam scans across the wafer surface, the signal output will be read at different points in time (e.g., times T1, T2, T3, etc.), each receiving a different electron rate. Thus, a time-dependent signal can be generated that can be used to reconstruct an image.

[0152] In some embodiments, the raw detector current signal can be fed to a preamplifier. The preamplifier can include a current buffer and a transimpedance amplifier (TIA). The signal can then also be amplified by a main amplifier. The fully amplified signal is a signal that can be sampled at each pixel period and can be converted by an analog-to-digital converter (ADC). The pixel period can correspond to a time period associated with data for one pixel of a SEM image. The digital signal (typically an 8-bit signal) can be sent to an image processor. The image processor can generate a frame image (e.g., a 2D array of pixel data) when the primary electron beam completes one cycle of raster scanning over a region of the specimen. The image processor can have an aggregation (or accumulation) function that can use multiple frame images obtained over the same scan area to improve the SNR of the acquired image. In some applications, a certain level of SNR can be needed to perform defect detection or critical dimension (CD) measurement.

[0153] Reference is now made to Figure 6 , Figure 6 A representation of a detection system architecture is shown that can be configured for handling analog signals from a detector. A detection system 400 can be provided that includes a detector 144, a signal conditioning circuit 410, an analog signal processing path 420, an ADC 430, and a digital interface 440. The signal conditioning circuit 410 can include circuitry configured to process the output of the detection element of the detector 144, such as a PIN diode. The signal conditioning circuit 410 can include a current buffer and a transimpedance amplifier. The analog signal processing path 420 can include a main amplifier. The ADC 430 can convert the analog signal to an 8-bit digital signal. The digital interface 440 can communicate with components of the EBI system 10 (e.g., a deflection and image control (DIC) unit) via a transceiver that can include a transmitter TX and a receiver RX. The digital interface 440 can also include a digital switch, a digital control unit, or a controller configured to perform image processing, etc. Other circuitry can be provided in the architecture, such as a signal processing path configured to provide gain and offset control, for example.

[0154] At high electron beam currents, such as where a 4 nA secondary electron beam is incident on the detector, it is known that the average number of electrons reaching the detector can be a certain number, such as about 250 electrons per 10 ns pixel sampling period. That is, in a 10 ns pixel period (corresponding to a 100 MHz sampling rate), on average about 250 electrons can reach the detector. As discussed above, the detector can generate a current in response to receiving an electron on the detection surface. The width of the current pulse can be related to the speed of the detector, the signal path, or a combination of the detector and signal conditioning circuitry. The speed of the detector can also be determined in part by the energy of the incoming electrons. Individual current pulses induced by single electron arrival events can overlap closely, and thus a relatively smooth current signal can be generated at the terminals of the detector. The modulation of the current signal can follow the rate of incoming electrons. In this way, an output signal can be generated and fed into a detection system, such as detection system 400, and can be used for image reconstruction.

[0155] However, when the electron beam current is reduced to lower values, it is less likely that a smooth current signal of closely overlapping electron pulses can be generated. For example, at 40 pA, on average about 2.5 electrons can reach per 10 ns sampling period. When the current pulse generated at the PIN detector by a single electron arrival event has a pulse width of, for example, 3 to 5 ns, the electron pulses can not substantially overlap. Instead, there can be a gap between subsequent electron pulses such that the excitation has almost dissipated before a subsequent electron arrives. Thus, the electrical signal generated in the PIN detector measured at the time of sampling can not accurately reflect the number of electrons received during the entire sampling period. Shot noise of the electron beam can cause fluctuations in the signal current from the detector. Due to statistical differences in the nature of the secondary electron and backscattered electron generation processes, it is less likely that a consistent signal can be generated. Furthermore, there can be a fixed electrical noise floor that is reduced proportionally with respect to the signal current from the PIN detector, which can cause the SNR to drop sharply around, for example, 200 pA. Thus, the SNR can degrade at lower electron beam current values. The signal fluctuations from the detector can also represent a decrease in SNR due to the random arrival times of the signal pulses. Additionally, the detector speed, signal path, or combination of the detector and signal conditioning circuitry that can determine the total bandwidth of the detection system can also affect the SNR.

[0156] Figure 7 A schematic representation of electron arrival events and their relationship to the current signal when the electron beam current is relatively low is shown. For example, as discussed above, at a beam current of 40 pA, only two or three electrons can be received at the PIN detector during a sampling period. The sampling period τ smay be 10 ns. A signal pulse generated by a single electron arrival event can have a pulse width of about 5 ns. Thus, the signal read at time T1 or time T2 can not account for all electrons arriving during the respective sampling period. Moreover, the value of the measured signal can be close to the noise threshold Th N .

[0157] To address the fluctuations in the current signal, the detection system can be configured to extend the integration period to cancel the fluctuations. Alternatively, the bandwidth of the detection system can be reduced. Within one imaging pixel period, e.g. 10 ns, the current signal can only occur at certain points in time. The total period of the signal can be less than the entire pixel period. Thus, the integration period can be extended to cover the signal. However, extending the integration time can involve other drawbacks. For example, noise from the detector sensing element and associated circuitry can be present throughout the pixel period. Integrating over the pixel period can cause noise energy to also be captured, which can result in a reduced overall SNR. Noise present during the sub-periods between electron arrival event signals can be included in the pixel period together with the noise present during the electron arrival event signal pulses. The method of counting particles can address this problem, e.g. by measuring the signal only when an electron arrival event is determined to have occurred.

[0158] In applications requiring accuracy at lower electron beam currents, such as critical dimension SEM (CD-SEM), other types of detectors, such as Everhart-Thornley detectors (ETD), can be useful. ETDs can combine a scintillator and a photomultiplier tube (PMT) together and can effectively detect individual electron arrival events as they occur without substantial overlap. ETDs can have good SNR in a probe current range of about 8 to 100 pA. However, the light yield of the scintillator can decrease over time as the electron dose accumulates, thus having a limited useful lifetime. Thus, if an ETD is used with an instrument that operates using higher probe currents or continuously, especially in inspection tools that operate continuously at high throughput, such as in semiconductor manufacturing facilities (e.g. fabs), its service life can be shorter. Moreover, ETDs coupled with a scintillator can exhibit poor performance due to the energy conversion step and signal loss associated with the detection signal path.

[0159] In contrast to ETDs, PIN detectors can exhibit excellent robustness against radiation damage. Moreover, in some embodiments, semiconductor detectors, such as PIN detectors, can achieve higher SNR under high performance signal paths.

[0160] In comparative embodiments, a PIN detector can have a reduced SNR at low electron beam currents. A primary cause of the reduced SNR at probe currents below, for example, 100 pA can be electrical noise in the front end of the amplifier before entering the ADC. For example, there can be large capacitances, thermal noise, or dark current generated by various electrical components. Additionally, some background noise sources can be proportional to the detection surface area of the detector. Furthermore, capacitance-based noise can increase with the cube of the frequency, and thus there can be many limitations when designing a high-bandwidth detector due to the reduced SNR. For example, this limitation can be addressed by providing a detection system that utilizes charged particle counting with a detector that includes a sensing element such as a PIN diode.

[0161] Some embodiments of the present disclosure can provide a detector that includes a sensing element sized such that an area of the sensing element is configured to receive no more than a predetermined number of charged particles, such as electrons, per sampling period. The area of the sensing element can be based on a geometric spread of charged particles incident on the detector. The area of the sensing element can be determined considering a region of the detector where the highest density of charged particles is incident.

[0162] Some embodiments of the present disclosure can provide a detector that includes a sensing element sized smaller than a size of a secondary electron spread. Referring now to Figure 8 which illustrates an example of a sensing element sized relative to a distribution of secondary electron landing points on a surface of the detector.

[0163] As Figure 8 illustrated, a natural geometric spread of secondary electrons can be determined based on, for example, initial kinetic energy and emission angle of the electrons. Although a primary electron beam can be focused to a relatively small beam spot size on a sample surface, secondary electrons emitted from the sample can produce a beam spot on the detector that is larger than the sample beam spot. The size of the beam spot on the detector can depend on the energy and angle of the electrons emitted from the sample surface. The size of the beam spot on the detector can be several times larger than the sample beam spot. A secondary optical system can be provided that projects the secondary electron beam onto the detector. Design or operating conditions of the secondary optical system, for example including magnification parameters, can affect the size of the beam spot on the detector. For a given set of imaging conditions, a geometric spread of secondary electrons incident on the detector can be determined. For example, dimensions X1 and Y1 can represent limits of an expected geometric spread of secondary electrons on the detector surface. Dimensions X1 and Y1 can be based on a statistical confidence that a predetermined percentage of the electrons will land in a region. For example, X1 and Y1 can represent a 99.5% confidence interval. In some embodiments, a 95% confidence interval can be used. In some embodiments, a 99% confidence interval can be used.

[0164] The size of the sensing element 701 can be determined to be smaller than the geometrical distribution of the secondary electrons. The sensing element 701 can have a side length of D. x and D y A rectangular shape. D x It can be set to be less than X1, and D y It can be set to be less than Y1. The size of the sensing element 701 can be determined such that the number of electrons received in the region of the sensing element 701 is less than the total number of electrons incident on the detector surface.

[0165] In some embodiments, the beam spot on the detector surface can be larger than the beam spot on the sample surface. Therefore, the overall size of the detector's detection surface can be configured to be large enough to accommodate a wider beam spot. The diameter of the beam spot on the detector surface can be approximately a few millimeters. However, increasing the detector size can lead to noise effects. For example, the detector's capacitance can be proportional to the area of ​​the detector surface. Some noise sources, such as those caused by components coupled to the detector (e.g., amplifiers), may be capacitance-related.

[0166] In some embodiments, a signal conditioning circuit with very low input impedance may follow the detector. Most of the charge generated in response to an electron arrival event can be extracted with low loss. However, in other embodiments, some losses that may be associated with capacitance may be introduced. For example, the detector may exhibit diode-like behavior and may have a corresponding capacitance for generating a specific voltage. The voltage from the detector may be fed to another component in the detection system, such as an amplifier. The detector may need to generate a voltage greater than the background noise level seen by the component. As an example, the component may have a certain level of background noise due to thermal noise. To overcome the thermal noise seen by the component, and thus input a meaningful signal into the component, a voltage greater than a certain amount should be input. Therefore, if the capacitance of the sensing element increases with the area of ​​the sensing element, the current required to achieve a voltage greater than a certain amount also increases. Therefore, with the increase of capacitance, the detector may need to generate more current to deliver the appropriate voltage.

[0167] In some embodiments, the area of ​​the sensing element can be reduced by configuring the sensing element to be smaller than the geometrical distribution of secondary electrons incident on the entire detector surface. When the sensing element is completely covered by the beam spot, the signal and capacitance of the detector sensing element can be proportional to the area of ​​the sensing element. In some cases, the sensing element may only be partially covered by the beam spot. When the sensing element is not completely covered by the beam spot, the reduction in SNR due to capacitance can be mitigated by reducing the area of ​​the sensing element.

[0168] Further, dark current can also be related to the area. The larger the area of a sensing element, the larger the dark current noise. For example, with a larger area sensing element, it is more likely that there can be a large number of imperfections in the crystal structure of the semiconductor diode device, and thus, there can be a greater chance of dark counts occurring. A sensing element configured to generate a number of electron-hole pairs in response to the arrival of a charged particle such as a secondary electron and having a relatively small area can be advantageous for reducing the impact of dark current events, as the relatively small noise signal generated by the dark current can be dwarfed by the signal generated in response to a charged particle arrival event. In some embodiments, where a sensing element is not fully covered by a beam spot, the SNR can be improved by reducing the size of the sensing element, such that the noise impact is reduced at a greater rate than the signal strength.

[0169] The sensitivity of a sensing element with a reduced area can be enhanced. For example, a sensing element configured to detect a single electron arrival event can have the following advantages. Considering two sensing elements with different areas, each configured to receive no more than one electron per sampling period, the sensing element with the smaller area can have a lower capacitance and be less impacted by dark current noise. Thus, the amount of current needed to generate a signal for feeding into another component of a detection system and overcome, for example, the thermal noise of that component can be less. Further, in some cases, a beam spot of a secondary electron beam on a detector surface can only partially cover a sensing element. Some of the detector area can not effectively receive incident electrons. However, all of the material under the detection surface of a sensing element can contribute to generating a current signal in response to incident electrons ionized by impingement. The number of electron-hole pairs generated in response to an electron arrival event can be similar regardless of whether a sensing element is fully covered by a beam spot. Considering two sensing elements of different sizes, the sensing element with the smaller area can be more likely to generate a signal that overcomes noise, as it can require fewer electrons to generate a particular voltage in response to individual electron arrival events.

[0170] In some embodiments, the area ratio, which can be the ratio of the area of an individual sensing element to the area of the entire surface of the detector, can be varied. The area ratio can have a correspondence with the SNR. For example, in some embodiments, reducing the size of a sensing element to 1 / 1,000 of the area of the detector can correspond to a 1,000-fold increase in the SNR.

[0171] In some embodiments, a detector can include an array of sensing elements. The array can include a plurality of sensing elements, each having, for example, a D x and D yor smaller. The sensing elements can be arranged in a planar two-dimensional array, the plane of the array being substantially perpendicular to the direction of incidence of the incoming charged particles. In some embodiments, the detector can be arranged to be tilted with respect to the direction of incidence.

[0172] Figure 9A An example of a detector 800 comprising an array of sensing elements is shown. The size of the sensing elements 801 of the detector 800 can be determined such that a certain number of electrons received in the sensing elements 801 is less than the total amount of electrons incident on the surface of the detector 800. The size of the sensing elements in the array can be uniform. The total size of the detector can be based on the geometric spread of the secondary electrons such that all or substantially all of the electrons can be captured by the detector. Thus, the detector 800 can have a total area corresponding to a predetermined FOV. For example, the detector 800 can comprise a circular plate having a diameter of 4 to 10 mm.

[0173] The detection area of the detector 800 can be divided into an array of smaller area PIN diode elements. Each PIN diode element can correspond to a discrete detection cell. The PIN diode can be pixelated into individual detection cells in various forms. For example, the semiconductor detection cell can be divided by means of an internal field generated due to the internal structure. Further, in some embodiments, there can be a physical separation between adjacent sensing elements. That is, in some embodiments, the detector array can be provided with sensing elements that are physically spaced apart from each other. Some isolation regions can be provided between adjacent sensing elements.

[0174] In other embodiments, when the separate detection cells are formed by the internal field, a plurality of doped semiconductor structures having different conductivities (such as P+ and N+ semiconductor regions) can be formed at the bottom of the detector. Meanwhile, the top surface of the detector can be provided with a detection surface and a cathode formed of a single layer of doped semiconductor. The intrinsic semiconductor region across the plurality of semiconductor structures having different conductivities can be substantially continuous. Thus, in some embodiments, the dead area between adjacent sensing elements can be reduced. Such a detector can have a structure capable of forming an internal electric field that can be configured to direct the charge carriers generated due to the electrons incident on each detection sub-region to the corresponding detection cell.

[0175] In contrast to ETDs, which can not be easily pixelated, PIN diodes can be configured with segmented sensing element arrays in a variety of forms. Also, PIN diodes can have other advantages over ETDs, including small capacitance and low background noise.

[0176] Figure 9BAnother example of a detector 850 including an array of sensing elements is shown. The detector 850 can include a plate 851 with a plurality of sensing elements 861 formed thereon. The plate 851 can include an opening 895 to allow a primary electron beam to pass through the plate 851.

[0177] When individual sensing elements of a detector are made smaller than the geometric spread of secondary electrons incident on the detector, electron counting can become more manageable. For example, each sensing element can have its own counting unit including circuitry configured to measure an output signal from the sensing element. When the sensing elements are made smaller, the electron rate on each sensing element becomes smaller, thus making it possible to count electrons at each sensing element. Electron counting will be discussed with reference to the following example validation study.

[0178] As discussed above with reference to Figure 4B Electrons incident on a detector can spread over an area. When a single sensing element is used to measure the output current of the entire detector surface, a large number of incident electrons in a short time frame can overwhelm the single sensing element. For example, for an electron beam with a beam current of 1 nA, a certain number (e.g., 6.4 billion) of secondary electrons can be incident on the detector in one second. This corresponds to 64 electrons per 10 ns sampling period for a detection system running at a 100 MHz pixel rate. For a detector with a single sensing element, it can be difficult to handle such a high electron rate, thus counting individual electrons can be impractical while maintaining reasonable circuit complexity, miscount rate, and power consumption. However, when a detector is subdivided, individual sensing elements can be made with a size such that the number of incident electrons in a sampling period is low enough to enable electron counting. In addition, a detector divided into multiple sensing elements can have many advantages such as those discussed above with respect to reducing noise, reducing capacitance, and improving SNR.

[0179] Figure 4B The distribution of electrons shown can represent 10,000 simulated landing positions of electrons. Assuming an opening 145 can be formed in the center of the detector 144, the densest distribution of electrons landing on the detector can be in an annular region immediately around the opening 145. For example, Figure 10 An annular space 720 around the opening 145 is shown. It can be simulated that, for example, in an annular space 720 with an inner diameter of 0.5 mm and an outer diameter of 1.0 mm, a first number of the total 10,000 simulated electrons can land in the region. The first number can be 1,669. The total area of the annular space 720 can be determined to be 0.589 mm 2Thus, the average number of electrons incident on the detector in the area of the annular space 720 can be 2,834 per unit area. Further, the proportion of electrons received in the annular space 720 can be 16.69% of the total.

[0180] In some embodiments, individual sensing elements can have a square shape of 50 pm x 50 pm. For example, Figure 10 A sensing element 731 having a square shape is shown inside the area of the annular space 720. Such a sensing element has an area of 0.0025 mm 2 The number of electrons per unit area in the annular space 720 multiplied by this area gives 7.084 electrons per square shaped sensing element.

[0181] Although the simulated 10,000 electrons are less than the number of electrons in a typical 1 nA beam (e.g., 6.4 billion per second), these results can be scaled to such levels of beam current. That is, the distribution of secondary electrons from a 1 nA electron beam can be substantially similar to that discussed above, but with higher absolute numbers. Thus, in the case of a beam of 1 nA, 16.69% of the total 6.4 billion electrons per second incident on the detector 144 in the annular space 720 can land in the annular space 720. Thus, approximately 1.8 billion electrons per second can land in the area of the annular space 720. Figure 10

[0182] If the size of the sensing elements is as discussed above, e.g., like the sensing element 731, then approximately 4.5 million electrons per second can land in the 50 pm x 50 pm square area of the sensing element. This is three orders of magnitude less than a detector using a single electron sensing element that can have all 6.4 billion electrons per second arrive at the single electron sensing element. By dividing the detector into a segmented array, the rate of electron arrival at the respective sensing elements can be significantly reduced. This can facilitate the implementation of electron counting. For example, when 4.5 million electrons per second are incident in a sensing element, the average interval time between subsequent electron arrival events (which should be discussed in detail later) can be approximately 200 ns. By comparison, a 100 MHz detector has a sampling period of 10 ns. With an average interval time of 200 ns, it is unlikely that two electrons will arrive within a 10 ns sampling period. Thus, the detector can have an area that is configured such that no more than a predetermined number of electrons are received in the sensing element during a sampling period. The predetermined number can be one.

[0183] ​Further, although the above examples are discussed with reference to a detector 144 having an opening 145 at its center, other configurations of the detector can be similarly applied. For example, in the case of an off-axis detector, all of the electrons in the central region can be incident on the detector surface. In this case, no hole can be provided at the center of the detector plate. Thus, the region of the highest density of electron arrival rates can be the center. However, even at the center, the electron arrival rate can not be significantly greater than the annular space near the center. For example, when simulating 10,000 electrons, a sensing element of a square area of 50 pm x 50 pm placed at the center of the detector can receive a second number of electrons. The second number can be approximately 8.9 electrons. Thus, extending to the case of a 1 nA beam, approximately 5.7 million electrons can land in the 50 pm x 50 pm square area of the sensing element in one second. Even at this rate of electron arrival events, the average interval time can be approximately 175 ns, which is still much higher than the sampling period of 10 ns.

[0184] Figure 11A A schematic representation of electron arrival events and their relation to the output of a detector element is shown. The detector can be provided with an array of detector elements. The array of detector elements can comprise an array of sensing elements, each sensing element having a size configured such that no more than one electron is received in the sensing element during a sampling period τ s of the corresponding detector element. As in the above examples, the interval time between two adjacent electron arrival events can be, for example, 200 ns. The detector can have a sampling rate of, by way of example only, 20 MHz, which corresponds to a sampling period of 50 ns. Thus, as Figure 11A shown in the lower plot of FIG. 6B, because the interval time between electron arrival events is greater than the sampling period, no more than one electron is received per sampling period.

[0185] In Figure 11A plot (A) can represent a time axis. It can be indicated that at a time point T1, a first electron arrives at the detector, and at a time point T2, a second electron arrives at the detector. Plot (B) can represent an event signal of a circuit associated with a sensing element of the detector. The pulse height of the event signal can correspond to the energy of the incident electron. The pulse height of the event signal can be compared to a threshold value, as indicated by the dashed line in plot (B). As shown in plot (C), when the pulse height exceeds the threshold value, a detection pulse can be registered. A count can be output in the next sampling period, as shown in plot (D).

[0186] Figure 11B Another example of a schematic representation of electron arrival events according to a time axis in plot (A) is shown. As in the above examples, the interval time between two adjacent electron arrival events can be, for example, 200 ns. The detector can have a sampling rate of, by way of example only, 20 MHz, which corresponds to a sampling period of 50 ns. Thus, as Figure 11BAs shown in graph (B), the pulse height of the event signal corresponding to different electron arrival events can be different. Detection pulses can be registered based on a first threshold or a second threshold. For example, as... Figure 11B As shown in the curve (D), when the pulse height of the event signal exceeds the first threshold, a detection pulse can be registered. Figure 11B As shown in graph (C), a detection pulse can be registered when the pulse height of the event signal exceeds the second threshold. A count can be output based on the registered detection pulses, such as in... Figure 11B In the curve (E) or (F).

[0187] Electron arrival detectors can be stochastic processes. Therefore, while some deterministic properties can be determined to describe electron arrival events, electron arrival events on a detector may inherently involve some randomness. For example, although the average interval between adjacent electron arrival events can be determined as a value, some electron arrival events may have intervals less than or greater than the average interval. Therefore, the sensing element can have an area configured to receive no more than a predetermined number of electrons in each sampling period with at least a confidence level. The confidence level can be based on statistical parameters. For example, the confidence level may correspond to the statistical probability of receiving no more than a predetermined number of charged particles in the sensing element in each sampling period. In some embodiments, the confidence level may be, for example, 90%. The sensing element can have an area configured to receive no more than a predetermined number (e.g., 1 or 2) of electrons in each sampling period with at least a 90% confidence level. In some embodiments, at least 90% of electron arrival events may occur such that no more than a predetermined number of electrons are received in the area of ​​the sensing element in each sampling period.

[0188] Statistical parameters describing the distribution of electron arrival events may include the mean or median of the time intervals between electron arrival events, the standard deviation of the time intervals between electron arrival events, the variance, the probability density, the cumulative distribution function, the skewness, the entropy, etc. In some embodiments, electron arrival events can be modeled using a Poisson distribution. In some embodiments, other types of distributions can be used to model electron arrival events.

[0189] In some embodiments, the average interval time can be related to the average current I entering the detector element l. d,l Related, where I d,l This is the average incoming electron current incident on detector element l. Although it is assumed here that the total current incident on the detector and its spatial distribution are constant throughout the SEM image frame, it is important to understand that the total current incident on the detector can be modulated based on the pixel positions within the SEM image frame. Given the sampling period τ... SAt time t, the average number of electrons h entering the detector element 1 can be described as follows, where e is the unit charge of an electron:

[0190] h = l d，l τ S / e (1)

[0191] Assuming that each electron arrival is a random event and is independent of other events, a Poisson distribution can be applied. According to the Poisson distribution, the probability of no electron arrivals during a sampling period can be given by the following equation:

[0192] P(0) = e -h (2)

[0193] The probability of one electron arrival during a sampling period can be given by the following equation, where λ is the average number of events in a sampling period:

[0194] P(1) = he -h (3)

[0195] The probability of two electron arrivals during a sampling period can be given by the following equation:

[0196]

[0197] The probability of three electron arrivals during a sampling period can be given by the following equation:

[0198]

[0199] The probability of k electron arrivals during a sampling period can be given by the following equation:

[0200]

[0201] Summing the probabilities of all possible independent cases of electron arrival events results in 1, as follows:

[0202] P(0) + P(1) + P(2) + P(3) +... = 1 (7)

[0203] The sum of the products kP(k) should equal h, as follows:

[0204] 1P(1) + 2P(2) + 3P(3) +... = h (8)

[0205] A counter implemented in a circuit can have its own maximum count, which represents the number of electronic arrival events that occur in its count buffer. If the count buffer consists of only one bit, the maximum count of events is one. When one or more additional events occur after the first event within the same sampling period, the count buffer value can remain one, and the additional events can be miscounted. If the count buffer consists of two bits, the maximum count of events can be two or three, depending on its implementation. Figure 12 The relationship between the number of events to be counted and the number of missed events is summarized, depending on the configuration of the counter and the number of events during sampling.

[0206] Based on the above, the confidence level that no more than a predetermined number of electrons (e.g., one or two) arrive within a sampling period can be expressed as follows:

[0207]

[0208] And

[0209]

[0210] The “confidence level that no more than a predetermined number of electrons (e.g., one or two) arrive in a sampling period” is expressed here as the ratio of the average number of electrons counted to the average number of entire electrons that impinge on the sensing element within a sampling period, e.g., the capture rate. The examples discussed herein can differ in terms of specific counter functionality, limitations, or maximum number of events that can be counted in a sampling period (e.g., one or two).

[0211] Figure 13A A graph illustrating the confidence level that can represent the count of electrons at a particular count buffer is shown. Figure 13A The x-axis in the graph can represent the average number of electrons that arrive at the sensing element per sampling period. Figure 13A The y-axis in the graph can represent the confidence level of the detected count of electrons. Figure 13A The solid line in the graph can represent the confidence level for a maximum count of one. Figure 13A The dashed line in the graph can represent the confidence level for a maximum count of two.

[0212] Loss rate of event count due to count buffer limitations L h (n) can be expressed as follows, where n is the maximum count of the count buffer:

[0213] For h « 1

[0214]

[0215] And

[0216] For h « 1

[0217]

[0218] Figure 13B A graph illustrating a curve that can represent a detection loss rate due to a maximum count at a particular count buffer. Figure 13B The x-axis in can represent an average number of electrons arriving at a sensing element per sampling period. Figure 13B The y-axis in can represent a detection loss rate. Similar to Figure 13A , Figure 13B The solid line in can represent a confidence level for a maximum count of 1. Figure 13B The dashed line in can represent a confidence level for a maximum count of 2.

[0219] Reference is now made to Figure 14A which shows a schematic representation of electron arrival events and their relationship to an output signal, consistent with embodiments of the present disclosure. In Figure 14A each of the three graphs of, the t-axis represents time. In the top graph of Figure 14A a single electron can be incident on the surface of a detector at a particular point in time. For example, electrons can arrive at time Tl, time T2, time T3, and time T4, respectively. The electron arrival events at time Tl, time T2, time T3, and time T4 can occur at one of the detector elements in the detector. The detector elements can include sensing elements. The electron arrival events can have statistical variance, and thus, the electrons can arrive at the surface of the sensing elements at uneven intervals of time. As shown in the middle graph of Figure 14A a signal pulse 1011 can be generated in the detector element in response to the electron arrival event. As discussed above, a PIN diode can be used in the detector element, and the signal pulse can resemble a waveform. The amplitude of the waveform can be related to the number of electron-hole pairs generated in response to the electron arrival event. Figure 14A The Y-axes of the middle and lower graphs of can represent signal strength in arbitrary units.

[0220] The time between subsequent electron arrival events can be given by T i -T i-1 , where i is an index representing the order of arrival of the respective electron. In some cases, the time between subsequent electron arrival events can be large enough that the signals do not interfere with each other. For example, the time between the electron arrival events at Tl and T2 given by (T i -T i-1 )1 can be long enough that the electrical signal generated in the sensing element rises and falls before the next electron arrival event occurs. In some cases, the time between subsequent electron arrival events can be short, such that the signals can merge. For example, the time between the electron arrival events at T2 and T3 given by (T i -Ti-1 The time between the electron arrival events T3 and T4, given in point 2, may be very short, causing the electrical signals generated in the detector to potentially overlap. The merged signals can be obtained from... Figure 14A The shaded area in the middle curve graph is an example.

[0221] Figure 14A The bottom curve can demonstrate the process related to the counting buffer operation. When the rising edge of the incoming analog signal crosses the threshold Th... e At that time, a count corresponding to the arrival of an electron can be recorded. When the corresponding input channel drops below the threshold Th... e The value changes to be higher than the threshold Th e When the value is less than the threshold Th, the function block can generate a pulse (e.g., a detection pulse) in the output channel (e.g., a value of 1). When the corresponding input channel does not receive a pulse below the threshold Th... e Transform into a value above the threshold Th e At this time, the block can hold the value at 0. Detection pulses can be counted over a given time period, and the count can then be placed into the count buffer for the next sampling period, such as... Figure 14A The bottom curve illustrates this. The detector's sampling period can be determined by τ. S Provided. In some embodiments, the detector may use a sampling rate of 400 MHz, corresponding to a sampling period of 2.5 ns. As mentioned above, in detectors using PIN diodes, the pulse width of the signal (e.g., an event pulse) generated in response to an electron arrival event can be approximately 3 to 5 ns. In this embodiment, two detection pulses may not be generated within a time period equal to the event signal pulse width, which may even be greater than the sampling period τ. S The event count for any sampling period can be 0 or 1. Therefore, even if two or more electrons may arrive at the sensing element within a sampling period, configuring the count buffer with 1 bit may be effective.

[0222] When T i -T i-1 Given a sufficiently long duration, the event signal waveform can rise and fall corresponding to a single electron arrival event. A discriminator block can be configured to detect the leading edge of the event signal pulse when an electron arrival event occurs. The discriminator block can be configured to make the determination based on a reference value, which may include a reference level. For example, a circuit system can be provided that compares the incoming event signal to a reference level and sets the detection signal to 1 when the incoming signal crosses the reference level. The detection signal may return to 0 after a short period, or hold that value and return to 0 at the end of the sampling period. This circuit system may, for example, include logical operators. Figure 14A In the example, a threshold Th can be set. eThis corresponds to the expected energy level of the secondary electron. For example, when an electron with a kinetic energy of approximately 9 keV penetrates into the detection unit of a PIN diode, this event may excite approximately 2,000 electron-hole pairs in the depletion region. These electron pairs can be collected at the cathode of the diode. The electrons collected at the cathode can form a small current pulse signal, which is fed into a circuit including front-end electronics, where the current pulse signal can be converted into a voltage pulse and amplified to an appropriate voltage level for use as a discriminator input. This processing can form an event signal pulse. The expected event signal pulse can be considered to set the threshold Th. e For example, the threshold Th e This can be a predetermined percentage of the expected peak value of an event signal pulse, which is initiated by secondary electrons arriving at a given energy level of the sensing element. In some embodiments, the threshold Th e It can be 60% of the expected peak value of an event signal pulse initiated by, for example, a 9keV secondary electron.

[0223] However, when T i -T i-1 In shorter cases, the event signal pulse waveform may represent electrons from multiple electrons arriving at the event, generating electron-hole pairs, even though the waveform from the initial current pulse shape may be significantly broadened and shaped due to the bandwidth and characteristics of the front-end electronics. For example, in Figure 14A At time T4, an electron arrival event may occur, and another event signal pulse is initiated before the event signal from the previous electron arrival event at T3 declines. Therefore, after time T4, the signal reading can rise relative to the previous reading and may continue to rise thereafter before fading to the threshold Th. e Therefore, based on the signal reading and the threshold Th e The comparative discrimination operation may not accurately explain the arrival of two separate electron events, and signal overlap and miscounting may occur.

[0224] In some embodiments, to address this miscounting, a comparison with another threshold may occur. For example, a further comparison with a different threshold may take place. This other threshold can be set to be different from threshold Th. e Furthermore, in some embodiments, this can represent an overflow state. When an overflow state is detected, it can be determined that more than one electron has been incident on the sensing element within a short interval.

[0225] In some embodiments, an overflow state may correspond to a situation where the sensing element has reached the limit of signal output generation. Overflow may occur when two electron arrival events occur consecutively. In some embodiments, overflow may occur when the combined event pulses build up to a level higher than the dynamic range of the analog signal of a particular detector element. The output signal may be determined by an overflow threshold Th. o Limitation. In some embodiments, the overflow threshold Th o This could be a theoretical limitation. In some embodiments, the overflow threshold Th o It can be set to a predetermined percentage that limits the output of the sensing element. In some embodiments, the threshold Th o It can be 90% of the expected limit. In some embodiments, the sensing element may be able to generate output signals corresponding to more than one electron arrival event (e.g., two, three, four electron arrival events, etc.) within a short interval. The overflow threshold Th can be... o Set as a multiple of the expected event signal pulse initiated by a secondary electron at a given energy level. In some embodiments, for values ​​exceeding Th... e The detection of the signal can correspond to a single electron arrival event, while for events exceeding Th... e The detection of the signal can correspond to two (or more) electron arrival events.

[0226] In some embodiments, the detector element may not be equipped to handle event pulse overlap, and this can lead to false counts of event pulse overlap. For example, the detector element may be configured to generate an event signal with an amplitude that is less than the sum of the amplitudes of the event signals from two separately occurring electron arrival events. Detector elements that continuously receive more than one electron may experience overflow cutoff.

[0227] Figure 14A A schematic representation of electronic arrival events and overflow cutoff consistent with embodiments of this disclosure is shown. Figure 14B As shown in the upper graph, two electron arrival events can occur rapidly and consecutively. A detector with detector elements can be provided, each detector element including a sensing element and circuitry. The circuitry associated with the sensing element can include front-end electronics, such as a preamplifier. When an electron arrival event occurs at the sensing element after the sensing element and its associated circuitry have fully recovered, theoretical signal pulses 1021 and 1022 can represent the event signals, respectively. For example, the detector element can generate signal pulse 1021 in response to an electron arrival event, and then generate signal pulse 1022 after signal pulse 1021 dissipates. In the case where the sensing element or its circuitry does not have a recovery time, signal pulses 1021 and 1022 can closely overlap and may appear to be superimposed on top of each other. Figure 14BAs shown in the middle plot, the output signal generated in the sensing element can resemble a signal pulse 1031 with a flat top. The signal pulse 1031 can be generated when the energy generated in the detector element reaches a limit in response to two or more electron arrival events. The signal pulse 1031 can represent the electrons from electron-hole pairs generated by incident electrons arriving on the sensing element.

[0228] In Figure 14B , time T1 can correspond to the start of a first electron arrival event, and time T2 can correspond to the start of a second electron arrival event that occurs before the detector element has completed generating a signal in response to the first electron arrival event. The signal can rise from time T1 and can be urged to continue rising at time T2. However, at time T2, the output limit of the sensing element can have been reached, and the output of the detector element can not rise further, but can remain constant. As shown in the bottom plot, when another signal is read at or after time T2, the signal does not exceed a cutoff threshold that can equal the threshold Th Figure 14B . o Thus, a discrimination operation based solely on a comparison of a signal reading to a threshold Th e may not accurately account for two separate electron arrival events, and signal overlap miscounting can occur.

[0229] A detector element can be configured to receive no more than a predetermined number of charged particles at a confidence level for a period. The confidence level can be expressed as a capture rate of events. The capture of an event can refer to the fact that an electron arrival event is recorded by the detector. In some embodiments, the confidence level can depend on a miscount rate. The miscount rate can be defined as one minus the capture rate. As discussed above, the miscount rate can include event signal overlap miscounting. A detector element of a detector array can be configured to have a miscount rate that satisfies a condition related to a first miscount threshold. The detector element can include a sensing element and a front-end circuit. The condition can be satisfied when the miscount rate is less than or equal to the first miscount threshold. The first miscount threshold can be a value corresponding to an acceptable level of miscounting. For example, in some embodiments, the detector element can be configured such that no more than a predetermined number of electrons are received in a region of the sensing element for a particular period of time, resulting in no more than a predetermined miscount rate corresponding to 10% or less of total electron arrival events. In this example, no more than 10% of total electron arrival events (counted as single electron arrival events) can in fact be a case of two or more electrons being received in the area of the sensing element.

[0230] While rapidly successive electron arrival events can cause miscounts, such as miscounts due to event signal overlap, the level of miscounts can be below a statistically acceptable amount. For example, electron arrival events can generally follow a Poisson distribution, and a miscount rate can be expressed with a Poisson distribution term, such as P(k).

[0231] In some embodiments, it can be assumed that two successive electrons arrive at the sensing element of the detector at, for example, T i-1 and T i successive arrival time points. A dead time τ D may be defined so that an electron arriving at T i is not counted (e.g., electron miscount) when the following expression is satisfied:

[0232] T i - T i-1 ≤ τ D (13)

[0233] As should be discussed in more detail herein, the dead time can be determined primarily by characteristics of the detector element. The detector element can include a front-end circuit. The dead time can be affected by the design or characteristics of the event detector, or other characteristics, such as a threshold level set to determine that an arrival event has occurred. In addition to the expression of equation (13), the dead time can also cause an electron arriving at T i to be counted when the following expression is satisfied:

[0234] T i - T i-1 > τ D (14)

[0235] In some implementations of the circuit, detection of an event can not be so clear-cut. Circuit noise or other random factors can assist or inhibit detection. For example, an electron arriving slightly earlier than the condition expressed in equation (13) can be detected (e.g., an arrival event can be counted). Similarly, an electron arriving slightly later than the condition expressed in equation (14) can be missed. In some embodiments, when T i - T i-1 = τ D , the probability of detection can be interpreted as 50%.

[0236] In some embodiments, the condition for an electron arriving at time T i to be counted can be that no electron arrived at the detector sensing element during the immediately preceding time period [(T i - τ D ), T i ].

[0237] Further, λ can be determined as follows, where Id,l is the average incoming electron current incident on the sensing element I, and e is the electronic unit charge:

[0238] λ = I d,l τ D / e (15)

[0239] The Poisson distribution can be a function of the time period τ, and does not depend on the position of the time period on the time axis. The Poisson distribution can be used to express the electron arrival events within the time period τ D . The probability that no electron arrives during the time period [(T i - τ D ), T i ] can be given as:

[0240] P(0) = e -λ (16)

[0241] The probability that an electron is counted can be P(0), which can correspond to the confidence level. The miscount rate (examples of which can include miscount rates due to event signal overlap or dead time) P Miscount can be given as:

[0242] P Miscount = 1 - P(0) (17)

[0243] = 1 - e ～λ (18)

[0244] For very small amounts of λ, the following condition can hold:

[0245]

[0246] Figure 14B A plot that can represent the miscount rate P Miscount as a function of λ, P Miscount , can initially increase linearly with increasing λ. As the saturation limit is approached, for example, values approaching λ = 1, P Miscount may begin to increase at a lower rate. Figure 15 The x-axis of the plot can represent the average number of electrons arriving at the sensing element within a time period τ D equal to the dead time. Figure 15 The y-axis of the plot can represent the miscount rate.

[0247] In Figure 15It can be seen that, in order to keep the miscount rate within, for example, 10% of the true count rate, given a dead time and a given distribution of electron arrival locations on the detector surface, the sensing elements can be configured to have an area that causes the incoming current to any sensing element to be limited to a particular lambda value. For example, the average number of incoming electrons on a sensing element can be set to remain at 0.1 electron / τ D The following.

[0248] In some embodiments of the disclosure, the detector can be provided with an array of sensing elements, each having a predetermined area. For example, in one embodiment, the size of the individual sensing elements can be 100 pm x 100 pm. However, other sizes of sensing elements can also be used. The sensing elements can be segmented. By providing an array of multiple sensing elements, the rate of electrons incident on each sensing element can be significantly reduced compared to the case of a PIN detector where the entire detection area consists of a single element (as discussed above) or a limited number of detection segments (e.g. 8 segments). For example, even when an electron beam with a current of 1 nA arrives at the detector at a rate of, for example, about 6.4 billion electrons per second, the number of electrons incident on a 100 pm x 100 pm sensing element of the array can be on the order of 20 million per second, even in a high density electron arrival region. In comparison, all 6.4 billion electrons would be incident on one sensing element in one second in the case of a single element detector. Incidentally, the number of electrons incident on a sensing element having an area of 1 / 4 of a 100 pm x 100 pm sensing element (e.g. a 50 pm x 50 pm sensing element) can be 1 / 4 of that, e.g. about 5 million per second, which can correspond to about 0.025 electrons per 5 ns.

[0249] There can be many criteria for determining the area of the individual sensing elements. For example, a first area determination criterion can be based on avoiding miscounts. In some embodiments, a miscount can be related to the average interval time (T i -T i-1 ) of electron arrival events on a given size of sensing element. The size of the sensing element (e.g. the area of the surface configured to receive electrons) can affect the number of electrons received on the sensing element in a given sampling period. The sampling period associated with a sensing element can be determined based on parameters of the detector element that can include the sensing element (e.g. using a sampling rate of 400 MHz to provide the detector). The size of the sensing element can be determined such that it is unlikely that a predetermined number of electrons is received in the sensing element within the sampling period. The sampling period can be related to a dead time that can be associated with the sensing element. For example, the sampling period can be set to be less than the dead time. The dead time can be independent of the sensing element size.

[0250] In some cases, the average electron interval time can be short, resulting in a high probability of miscounting. The greater the average electron interval time, the lower the probability of miscounting can be. The area of the sensing element can be set such that the average electron interval time is not greater than a particular value. The average electron interval time can be set to be greater than a value such as a multiple of the detector element sampling period, for example, τ S . Thus, the sampling period of the detector can be set to be a factor (e.g., a number between 0 and 1) of the average electron interval time. In some embodiments, the average electron interval time can be set to be greater than twice the sampling period τ S . In other words, the sampling period τ S can be set to be less than half of the average electron interval time (T i -T i-1 ). As discussed above, the average electron interval time (T i -T i-1 ) can be controlled by adjusting the electron beam current or by changing the area or position of the sensing element. Further, the sampling period τ S may be defined by the sampling frequency (e.g., sampling rate) of the detector. For a given electron beam current, sampling rate, and maximum electron arrival rate (in the region of the highest electron arrival density) of a detector, the average electron interval time (T i -T i-1 ) per unit area can be determined.

[0251] In some embodiments, a detection system can include a sensing element and circuitry that can be used to count charged particle arrival events. Figure 15 Signal flow in a detection system 1300 is illustrated. The detection system 1300 can include a detector element 1330, a first circuit 1340, and a second circuit 1350. The detector element 1330 can include a transimpedance amplifier (TIA). In some embodiments, the detector element 1330 can consist of a sensing element and a front-end circuit. The front-end circuit can be referred to as a signal conditioning circuit and can include a TIA. The first circuit 1340 can include an event detector. The first circuit 1340 can be configured to implement the function of event detection. The first circuit 1340 can include a discriminator, such as a pulse height analyzer. The second circuit 1350 can include a count buffer. The second circuit 1350 can be configured to implement the function of count buffering. The second circuit 1350 can include a scaler.

[0252] The detection system 1300 can be configured to perform a detection operation that can include generating a pulse signal by the detector element in response to the electron 1301 reaching the detector element 1330. The pulse signal can include an event signal 1305 that can be output to the first circuit 1340. The detection operation can allow the detection system 1300 to determine that an electron arrival event has occurred. The detection operation can begin with the detector element 1330 generating the event signal 1305 in response to the electron 1301 reaching the detector element 1330. The event signal 1305 output from the detector element 1330 can be fed into the first circuit 1340 and analyzed to determine that an electron was received at the detector element 1330. The determination can include comparing a pulse height of the event signal 1305 to a predetermined value. If the height of the event pulse signal exceeds the predetermined value, the electron event can be detected. In response to detecting the electron event, a counter can be incremented. The first circuit 1340 can include a unit counter. The first circuit 1340 can include an event marker detector. The first circuit 1340 can set a marker when an electron arrival event is detected. The first circuit 1340 can output a detection pulse 1309. The detection pulse 1309 can be fed to the marker.

[0253] The output from the first circuit 1340 can include unit information. For example, in response to determining that an electron occurrence event has occurred, such as by the pulse height analyzer determining that the input event signal exceeds a predetermined value, the first circuit 1340 can output a signal of 1. In some embodiments, the first circuit 1340 can be configured to output data in multiple bits. For example, the first circuit 1340 can include a 2-bit or 3-bit counter.

[0254] In some embodiments, a detection system 1300A can be provided, as shown in Figure 16A The detection system 1300A can include a first circuit 1340 and a second circuit 1350 similar to the detection system 1300. Figure 16B The detection system 1300A can also include a sensing element 1332 and a third circuit 1334. The third circuit 1334 can include a front-end electronics, such as a preamplifier. The third circuit 1334 can include a transimpedance amplifier. The signal 1302 output from the sensing element 1332 can be processed by the third circuit 1334. The signal 1302 can include a current pulse signal. The current pulse signal can be converted to a voltage pulse signal and amplified to form the event signal 1305. The event signal 1305 can be output from the third circuit 1334.

[0255] In some embodiments, a dead time can be associated with the event signal 1305. The event signal 1305 can have a width over which the pulse height rises and falls. Characteristics of the diode forming the sensing element can affect the shape and width of the event signal 1305. The dead time can be associated with the event signal 1305 and other processing involving the detection system 1300 or 1300A. The dead time can refer to a time in which a subsequent electron arrival event can be hindered from detection. In some cases, an electron 1301 arriving at the sensing element 1332 can trigger the generation of the signal 1302. As a result of the arrival of the electron 1301, the event signal 1305 can be output from the third circuit 1334. The event signal 1305 can be processed by the first circuit 1340. For example, when a second signal (a signal after the signal 1302) is generated at the sensing element 1332, there can be a situation in which two separate event signals cannot be reliably generated and a dead time effect can occur. As a result of the dead time effect, a continuous occurring electron arrival event can not be reliably counted. This effect can refer to an event detector dead time. For example, an event detector included in the first circuit 1340 can be configured to determine that an electron arrival event has occurred when the pulse height of the event signal 1305 rises above a threshold value. After determining that the pulse height of the event signal 1305 falls below the threshold value, detection of another electron arrival event can occur. However, when the subsequent electron arrival event prevents the pulse height of the event signal 1305 from falling, a miscount can occur. The event detector dead time can be associated with the width of the event signal 1305 and the threshold value used to determine an electron arrival event. For example, the event detector dead time can be associated with a distance between the left and right sides of the event signal 1305 at a pulse height equal to the threshold value. Other factors and characteristics of the detection system can also affect the dead time. For example, stray capacitance of interconnects from the sensing element to the circuit can cause a dead time.

[0256] The dead time can be associated with a particular sensing element. For example, there can be a dead time associated with the sensing element 1332. This dead time can be related to characteristics of the sensing element 1332 or its associated circuitry. The dead time can be associated with the first circuit 1340 including the event detector.

[0257] Circuitry included in the detection system 1300 or 1300A can be configured to make various determinations. The first circuit 1340 can be configured to determine a characteristic of the signal input therein. Similarly, the second circuit 1350 and the third circuit 1334 can be configured to determine a characteristic of the input signal. The third circuit 1334 can be configured to determine a first characteristic of the signal 1302, which can be a current pulse. The first characteristic can be an amplitude of the current pulse. Other determinations can be based on the determined first characteristic. For example, based on the amplitude of the current pulse thereof, it can be determined whether the electron incident on the sensing element is a first type of electron or a second type of electron. Such a determination can involve a comparison to a threshold energy level, as should be discussed in more detail herein. It will be understood that such a determination can also be based on the event signal 1305.

[0258] In SEM systems, a generated image of a sample can include a plurality of pixels. Each pixel can show information, such as a grayscale, that can correspond to information derived from the detector. The information can include a count of the number of electrons received in a sampling period. The count of the number of electrons can be summed from a plurality of sensing elements. In some embodiments, a pixel can use data from multiple sampling periods from the sensing elements. A pixel period can be defined that uses information derived from the detector over multiple sampling periods at individual sensing elements. The pixel period can be determined by an operator. For example, the pixel period can be a user-defined value that can be selected from a list of available values. In some embodiments, a user can pick a number D of sampling periods from a list of available integers, such as 1, 2, 4, 8, etc. The pixel period can be set to D times the sampling period. For each pixel, a value (e.g., grayscale) can be determined based on the consecutive sampling data of the D sampling periods. The grayscale can be determined by, for example, summing or averaging the data of the D sampling periods. In some embodiments, the pixel period can be predetermined based on parameters of the inspection system instrument. For example, the pixel period can be preset for a particular machine.

[0259] The SEM system can scan the primary electron beam over the sample in a pattern such as a raster pattern. Pixels to be generated in the image can be associated with locations of the scanned sample surface. Pixels can be associated with particular scan times. Data from the detector can be associated with scan times. A scaler can be used to associate the data received from the detector with particular scan times. The scaler can be configured to associate the detector data with scan times (and thus, pixels) based on a delay time. For example, the scaler can record the time at which data associated with an electron arrival event is received and apply a correction based on a delay from the trigger. The trigger can correspond to an electron of the primary electron beam impacting the sample. The delay can account for the time for an electron (e.g., a secondary electron) to travel from the sample to the detector, or the time for a signal pulse to be generated and output in a sense element. In some embodiments, the detector can include an array of sense elements. The scaler can include multiple channels, such that data from each sense element can be assigned to its own channel.

[0260] The output of the first circuit 1340 in 16A or 16B can be fed into a second circuit 1350. The second circuit 1350 can include a scaler. The scaler can be a multi-channel scaler. The second circuit 1350 can be configured to determine a count of electron arrival events to be associated with a pixel of an image. The second circuit 1350 can account for a time channel width, which can be based on a pixel period and a number of channels. The number of channels can be based on a number of pixels per line of a scan. The base time channel width can be set to be less than the dead time. Figure 16A

[0261] A detection system that records discrete events, such as electron arrival events, can experience a dead time. The dead time can refer to a time after an event in which the detection system is unable to record another event. For a detector element in a detector that can be configured to detect electron arrival events, there can be a period of time immediately after an electron arrival event in which a subsequent electron arrival event can not be accurately captured and recorded.

[0262] The dead time can include a collapsible or non-collapsible dead time. The non-collapsible dead time can refer to a situation in which an event occurring in the dead time is not recorded. For example, the event can simply be ignored. The collapsible dead time can refer to a situation in which an event occurring in the dead time is not recorded, in addition, a subsequent event causes the dead time period to restart. The collapsible dead time can also be referred to as an extended dead time.

[0263] ​In some embodiments, a sensing element can be configured to generate electron-hole pairs in response to an incident electron (e.g., a secondary electron) reaching the sensing element. As the energy of the incident electron is consumed in the sensing element, an electron arrival event can trigger a cascading generation of electron-hole pairs in the sensing element. The time behavior of the cascade can resemble a pulse. During the cascade of electron-hole pairs generated as a result of one electron arrival event, a subsequent electron arrival event can cause the cascade to continue, thereby extending the time it takes for the sensing element to return to a normal state. Sensing elements in a detector array can be subject to a paralyzable dead time. For example, a diode biased into Geiger mode can require a recovery time or can require quenching before a subsequent electron arrival event can be accurately detected. Some embodiments of the present disclosure can avoid a paralyzable dead time associated with a sensing element. For example, a detector can be configured such that a dead time associated with a sensing element does not include a paralyzable dead time. A sensing element can be provided with a higher internal gain such that it is not necessary to bias the sensing element into a region where the sensing element can be subject to a paralyzable dead time. The sensing element can be operated with a reverse bias in a linear region. The kinetic energy of an incoming electron can provide a high gain while maintaining the speed of the detector element. The speed of the detector element can be at least as high as the speed of a detector element operating under normal biasing conditions. In some embodiments, the speed of the detector element can be higher than the speed of a detector element operating under normal biasing conditions due to an avalanche effect. Long tail behavior of the detector element can be avoided.

[0264] Considerations of dead time can be used to determine whether a detector can detect a subsequent charged particle arrival event. For example, when electron arrival events occur in rapid succession, although there can have been two or more electron arrivals during a sampling period, only one electron arrival event can be recorded in the sampling period. Thus, an electron arrival event can be miscounted. If the sampling period is set to be shorter than the dead time, miscounting can occur based on a dead time arrival event. Miscounting can be acceptable to occur in a detector when the level of miscounting is below a certain level. In some embodiments, the certain level can be 10%. In some embodiments, the certain level can be 5%. In some embodiments, the certain level can be 2.5%.

[0265] In some embodiments, a dead time can correspond to a pulse width of an event signal initiated by an electron arrival event. In some embodiments, a dead time can be different from a total pulse width of an event signal generated in response to an electron arrival event. For example, a dead time can be measured from the time an electron arrival event is recorded until an output level of the event signal subsides below a detection threshold. The detection threshold can be set to be above a certain level of background noise. An electron arrival event can be recorded when a level of the event signal has exceeded the detection threshold.

[0266] Figure 16A An exemplary representation of dead-time miscounting is shown. Figure 17 The graph (A) in the figure can show the exemplary output of detector element 1330 over time (as shown in the figure about...). Figure 17 (As discussed). Event signal pulses may be relatively dispersed. Figure 16A The signal shown in graph (A) can be fed into a pulse height analyzer. The event signal pulse can be compared with a detection threshold 1410. Signals below the detection threshold 1410 can be filtered out as noise. Figure 17 The graph (B) in the figure shows the curve with Figure 17 The curve (A) shows the input signal and the corresponding output signal from the pulse height analyzer. Figure 17 There are no miscounts in the curves (A) and (B).

[0267] like Figure 17 As shown in graph (C), miscounting can occur when event signal pulses overlap closely. For example, when an electron arrival event is configured to be detected when the signal level exceeds a fixed threshold (such as threshold 1410), the signal can be counted as a single pulse because it does not decrease below the threshold before rising back up. Figure 17 The graph (C) illustrates the two types of dead-time miscounting. Figure 17 As shown in the graph (D), multiple arrival events can be counted as a single event.

[0268] As discussed above, miscounting can include dead-time miscounting. In some embodiments, dead-time miscounting can be considered as follows. In a first dead-time miscounting analysis, it can be assumed that the i-th electron at time t... i The electron reaches the sensing element. Whether the i-th electron is countable may depend on whether any other electrons arrived during the previous dead time period. For example, the electron may have already arrived from (t... i -τ D ) to t i It arrives within the period, where τ D This is the dead time of the event detector. This period can be expressed as [(t...]. i -τ D ),t i ].

[0269] In the first dead-time miscount analysis, the Poisson distribution can be used to model electron arrival events on the detector. The probability P of k events occurring during the time interval Δt can be given as follows, where λ is the average number of events per time interval Δt:

[0270]

[0271] for [(t i -τ D ),t i ] can be equal to τ D . Substituting the value of k into equation (20) gives the following results:

[0272] P(0) = e -λ (21)

[0273] P(1) = λe -λ (22)

[0274]

[0275] P(0) can represent the probability that no electron arrives during the period [(t i -τ D ),t i ], and thus the ith electron should be countable because its arrival event will not be affected by the dead time. P(1) can represent the probability that one electron arrives during the period [(t i -τ D ),t i ], in which case the ith electron can not be countable. P(2) can represent the probability that two electrons arrive during the period [(t i -τ D ),t i ], in which case the ith electron can not be countable. In the first dead time miscounting analysis, it can be assumed that the ith electron can be countable only when no other electron arrives in the period [(t i -τ D ),t i ] (i.e., immediately before the arrival event of the ith electron). In the first dead time miscounting analysis, the sampling period of the sensing element can be less than the event detector dead time τ D . In terms of the Poisson assumption, P(k) can be independent of the arrival time of the electron as long as λ remains constant in time.

[0276] The probability that an electron is counted within a given dead time can be expressed by P(0). Since the sum of the probability of miscounting and the probability of successful counting should be 1, the miscounting probability of an electron should be 1-P(0). According to the first dead time miscounting analysis, P miscount can be given as follows due to the dead time:

[0277] for λ << 1

[0278]

[0279] As can be seen from equation (24) above, the probability P of a miscount due to the dead time according to the first dead time miscount analysis miscount For small values of λ, λ can be approximated to equal λ. As mentioned above, λ can be the average number of electron arrival events in a time interval. The relevant time interval can be the event detector dead time τ D .

[0280] As can be seen from equation (24) above, the probability P of a miscount due to the dead time according to the first dead time miscount analysis miscount may also be given as:

[0281] P miscount = 1 - e d,l τ D / e (25)

[0282] As an illustration, consider the following scenario. A primary electron beam can illuminate a sample. A secondary electron beam can be generated that is directed towards a detector. The beam current incident on the detector can be I Det . For simplicity, it can be assumed that I Det is uniformly distributed over the sensing area S Det of the detector. S Det may comprise a circular area with a diameter of 4 mm, thus S Det = 4π mm 2 . Assuming that the detector comprises an array of segmented sensing elements, each with a size of 100 pm x 100 pm, the sensing element area can be given as S Seg = 0.01 mm 2 .

[0283] With exemplary values of S Det = 4π mm 2 , S Seg = 0.01 mm 2 , I Det = 1 nA and τ D = 5 ns, the probability P of a miscount due to the dead time according to the first dead time miscount analysis miscount may be given as:

[0284]

[0285] Therefore, a miscount rate of 2.5% or less can be achieved. Even when the beam current level incident on the detector is relatively high, such as 1 nA in the example described above, the miscount rate can be limited to a relatively low level of 2.5%. It should be noted that equation (26) can be based on an assumption that the beam density of the beam incident on the detector is uniformly distributed. However, such a miscount rate can be acceptable in various applications. Further, in some embodiments, a statistical correction of the electron counts can be used when the miscount rate is less than 10%. Therefore, even when miscounts exist, they can be corrected by the statistical correction. Compared to the beam current level of 1 nA incident on the detector, some applications, such as high throughput CD (critical dimension) measurements, can use a primary beam current of 250 pA or less, leaving up to 4 times the spontaneous peak secondary yield margin at the edge line of a printed pattern on a wafer surface. In some applications, the standard beam current of the system can be set to approximately 10 pA. Therefore, the miscount rate can be reduced even further in some applications. At the same time, the detector can be configured to handle situations of higher secondary electron local yield, which can increase the electron current reaching the detector, such as at a metal pattern edge (e.g., edge enhancement signal).

[0286] In some embodiments, the beam current can not be uniformly distributed, but rather distributed non-uniformly across all sensing elements on the detector. Even when the beam current is non-uniformly distributed, the situation of maximum beam current in the area where the electrons are expected to arrive can be considered. For example, the area of higher density of electron arrival events can be around the center of the detector. The detector can be configured such that the miscount rate in the sensing elements of the area of higher electron arrival does not exceed a predetermined level.

[0287] In some embodiments, the size of each sensing element can be set to a uniform square shape, where S Det / 3,125 is approximately equal to 0.009 mm 2 This can correspond to a sensing element of approximately 100 x 100 pm. Therefore, even when the beam current is high, a sensing element size of 100 x 100 pm can be suitable for obtaining an acceptable miscount rate. It will be apparent that other sizes of sensing elements can be used while achieving an acceptable miscount rate. Further, other parameters, such as the sampling period, can be adjusted to achieve an acceptable miscount rate.

[0288] Because the miscount rate can be low when the sampling period is less than the sensing element dead time, it can be desirable to set the sampling period to be less than the dead time. In some embodiments, the sampling period of the sensing element can be set to be less than 5 ns. When the sampling time is less than the dead time, the counter can be configured to count only one electron arrival event at a time.

[0289] The detector can be configured such that no more than a predetermined number of charged particles are received on the sensing element within a sampling period at no more than a predetermined miscount rate. The sampling period can be less than or equal to the sensing element dead time. The sensing element can be one of an array of sensing elements that make up the detector. The predetermined number can be one. In some embodiments, the predetermined number can be, for example, two, three, or more. Various parameters of the detection system can be set based on the predetermined miscount rate. For example, assuming the detection system is used with a primary charged particle beam having a maximum beam current, parameters of the sensing element size and the number of sensing elements can be determined based on the predetermined miscount rate. The beam current of the primary charged particle beam can be related to the beam current of a secondary charged particle beam that is incident on the detector. For example, a yield factor can indicate a collection rate of charged particles of the secondary beam, and can be used to determine the beam current of the secondary beam from the primary beam. The yield factor can be influenced by the material of the specimen under inspection, settings of the charged particle beam system including the detection system, and the structure of the charged particle beam system. Further, assuming the array of sensing elements of the detector includes sensing elements of a fixed size, operating conditions of the detection system, including the sampling rate and the beam current, can be set. In some embodiments, the miscount rate of the detector can be kept at a predetermined level even when the beam current is increased to levels that can be desirable for some applications. As an example, when high resolution SEM is desired, the beam current can be reduced. When throughput is desired, the beam current can be increased. Electron beam wafer inspection systems can use a relatively high beam current. In some embodiments, subdividing the detector into an array of sensing elements, where the sensing elements are smaller than the entire detector, can reduce the miscount rate. The smaller the sensing elements, the smaller the miscount rate can become.

[0290] The dead time can be a characteristic of the sensing elements and other components of the detector, which can be determined by its construction. For example, as discussed herein in some embodiments, a semiconductor diode used as a sensing element can influence the dead time due to phenomena related to the electron-hole pair generation mechanism. The dead time can be fixed when the circuitry and array of sensing elements are designed and manufactured. Once the detection system is constructed, the dead time can be a non-adjustable parameter. In contrast, the dead time can be constant for a given array of sensing elements and accompanying circuitry.

[0291] The sensing element can be connected to a circuit. The circuit can include a first circuit, a second circuit, or a third circuit, such as discussed above with reference to Figure 17 and 16B The circuit can be configured to determine a number of charged particles incident on the sensing element within a sampling period. The circuit can include an event pulse detector. The event pulse detector can include a discriminator, a comparator, or a pulse height analyzer, among others. The circuit can include a canceller. The canceller can be configured to filter out noise signals. The circuit can be configured to compare an input signal to a threshold value and produce an output when the input signal exceeds the threshold value. The output can include a bit indicating that a charged particle has been received at the sensing element in the sampling period. In some embodiments, the output can include a multi-bit signal. The sampling period can include a time channel width of a multi-channel scaler. In some embodiments, the output can include an event flag and an overflow flag.

[0292] Figure 17 A detection system 1501 configured to detect charged particle arrival events with an output including an event flag is shown, consistent with embodiments of the present disclosure. The detection system 1501 can include a detector element 1530, a first circuit 1540, a second circuit 1550, and a third circuit 1560. The detector element 1530 can include a diode and a preamplifier. The detection system 1501 can be configured to perform a detection operation that can include the detector element generating a pulse signal 1511 in response to an electron arriving at the detector element 1530. As discussed above, the pulse signal 1511 can include an event signal similar to the event signal 1305 shown in FIGS. Figure 16A or 16B. In some embodiments, the pulse signal 1511 can include overlapping or superimposed pulses, such as those shown in the graph (C) of FIG. Figure 18 The pulse signal 1511 can be input to the first circuit 1540. The first circuit 1540 can include a pulse height analyzer. The first circuit 1540 can generate an event flag 1541 that can be transmitted to the second circuit 1550. The second circuit 1550 can include an incrementing counter. The second circuit 1550 can include an event flag detector. Meanwhile, the first circuit 1540 can generate an overflow flag 1542 that can be transmitted to the third circuit 1560. The overflow flag can be generated in response to an energy of the pulse signal exceeding a threshold value. The third circuit 1560 can include an overflow flag detector. For example, it can be advantageous to configure the circuit as an event flag detector in terms of system simplicity.

[0293] In some embodiments, a detection system can include multiple timing scales. A detector element sampling period can be set to be less than an event detector dead time. The detector element sampling period can correspond to a time channel width of a multi-channel scaler. An electronic arrival event can be detected by circuitry including an event detector such as a discriminator, comparator, pulse height analyzer, etc. The detection of an electronic arrival event can occur asynchronously from the timing of a channel boundary. The circuitry can be configured to increment a value in a specified time channel that is within a time window of an event time after the detection of an electronic arrival event. The overall system timing can be configured to be faster than the timing of the individual detector elements (e.g., detector element sampling period). For example, a time channel can be configured as a one-bit counter (e.g., using an event flag), while another circuitry can be configured to implement the number of event flags set during a time channel period at a rate faster than the rate at which event flags are pushed out of components of the detection system. When the time channel width is greater than the dead time, the circuitry can be configured to run faster than the individual event component detectors corresponding to the detector elements. In some embodiments, the circuitry can be configured to run faster than the individual event flag detectors when the time channel width » the dead time.

[0294] In some embodiments, a one-bit counter can be provided with a detector element, where a detector element sampling period can be set to be less than an event detector dead time. When the detector element sampling period is set to be less than the event detector dead time, it can be possible to achieve a situation where there is no miscounting due to the event detector dead time. In one sampling period, there can be no multiple event signals input to the event detector. Electronic arrival events can occur consecutively, and in some cases can result in a merged event signal, but the event detector corresponding to the detector element can count no more than one charged particle arrival event in one sampling period. This configuration can enhance system simplicity and allow for high bandwidth. The detector can be manufactured using readily available technology, and can be implemented to run at 400 MHz or higher. The detector element can not need to be configured to detect more than one electronic arrival event at a time. For example, when multiple electrons arrive at the sensing element in a time period that is less than the event detector dead time, a mis-detection can occur. However, such mis-detection can be extremely rare. The statistical probability of multiple electronic arrival events occurring within the event detector dead time can be below a predetermined level. For example, the detector can be configured to have a size such that there can be at least a predetermined confidence level that two or more electrons will not arrive within the event detector dead time.

[0295] Circuitry connected to the sensing elements can be provided on a per-sensing element basis. For example, in a detector comprising an array of sensing elements, each sensing element in the array can be connected to a circuit. The circuit can be configured to count the number of charged particles incident on the sensing element over a sampling period. In some embodiments, the number can be one or zero. The detection system can comprise a global circuit that can be configured to sum the counts of a plurality of sensing elements in the array. The total number of charged particles received in the array can be determined by adding the number of charged particles received on the plurality of sensing elements.

[0296] Reference is now made to Figure 16A which illustrates a schematic representation of an exemplary architecture of a detection system 901. The detection system can be provided with a detector 900 comprising an array of sensing elements. As seen in cross-section, the detector 900 can have a structure similar to a back-illuminated CMOS image sensor. Electrons can be incident on the front (incident) surface side of the detector 900 that is exposed. At the front side, there can be a P+ layer that can act as an anode of a diode included in the array of sensing elements. The anode can be negatively biased by a voltage source having zero source impedance. For example, when an electron having an energy of about 9 keV penetrates into a sensing element of the detector 900, the event can excite about 2,000 electron-hole pairs in the depletion region of the sensing element. The electrons of those pairs can be collected at the cathode of the sensing element and can contribute to a detection signal. For example, a current pulse having an amplitude substantially proportional to the energy of the incident electron can be generated.

[0297] The raw detection signal from the sensing element in the form of a sharp current pulse can be fed into electronic circuitry. For example, front-end electronics 910 can be provided. The front-end electronics 910 can comprise individual circuits 911, 912, 913, 914, and so on, for example, up to an nth circuit 919. Each individual circuit of the front-end electronics 910 can correspond to one sensing element in the array of the detector 900. Each individual circuit of the front-end electronics 910 can comprise a current buffer and an amplifier, such as a transimpedance amplifier, a charge amplifier, or a charge transfer amplifier.

[0298] In some embodiments, components can be integrated with the detector 900. The detector 900 can be provided as a semiconductor substrate comprising an array of sensing elements. The array of sensing elements can comprise diodes. A low-gain amplifier can be established at the backside of the semiconductor substrate comprising the PIN diodes. The amplifier can be configured to operate in a linear mode and can avoid regions in which the amplifier can exhibit a recovery time. The semiconductor substrate can comprise a plurality of segmented diodes. A plurality of amplifiers can be configured such that a respective amplifier of the plurality of amplifiers is attached to each diode.

[0299] In some embodiments, the sensing element array can be configured with a low-gain avalanche diode (LGAD) segment. LGADs can provide relatively small current gain, such as about 10 to about 20 times, and operate very fast while maintaining linear mode. This differs from conventional avalanche diodes, which may enter an avalanche current after a trigger event and require a mechanism to reset the current in preparation for detection of the next trigger event. The additional current gain supplied to the LGAD can contribute to accurate detection timing or simplify the front-end electronics 910. LGADs may not require bias current. Furthermore, LGADs can help reduce the power consumption of the detection system.

[0300] After amplification, the signal output from each circuit in the front-end electronics 910 can be fed into the event detector 930. The event detector 930 may include individual circuits 931, 932, 933, 934, etc., for example, at most an nth circuit 939. Each individual circuit of the event detector 930 may correspond to a sensing element in the array of detectors 900. Each individual circuit of the event detector 930 may include a discriminator, a discriminator block, a pulse height analyzer, etc.

[0301] Figure 17 The detection system and Figure 19A The difference between this detection system (such as a CMOS sensor) and others is that it can provide an event detector 930 instead of an ADC block. The event detector 930 can be used to count electrons.

[0302] return Figure 6 The event detector 930 may include a discriminator block. The discriminator block may include logic operations that can be implemented using hardware such as electronic circuitry. The discriminator block may compare an incoming signal (e.g., an event signal) with a reference level and output a signal such as a detection pulse or an event detection flag when the rising edge of the incoming signal crosses the reference level. For example, as mentioned above... Figure 19A The discussed approach involves comparing the measured current signal with a threshold Th. e Comparison. Th can be used. e The level is set sufficiently above the background noise level. In some embodiments, the discriminator block may be implemented by a controller. The event detector 930 may also include other types of circuitry, such as comparators configured to compare incoming voltage, current, or other types of signals with a reference level.

[0303] Figure 14AThe detection system of FIG. 9 can also include a pixel count buffer 950. The pixel count buffer 950 can include individual circuits 951, 952, 953, 954, etc., for example, up to an nth circuit 959. Each individual circuit of the pixel count buffer 950 can correspond to one sensing element in the array of the detector 900. The output from each circuit of the event detector 930 can be fed into the corresponding circuit of the pixel count buffer 950.

[0304] Figure 19A The detection system of FIG. 9 can also include a count summing unit 960. The output from each circuit of the pixel count buffer 950 can be fed into the count summing unit 960. The counts of received electrons from each sensing element in the detector 900 can be combined, for example, to obtain a gray scale of an image to be used in image processing. The combined counts of received electrons can include summing counts across multiple sensing elements, multiple sampling time periods, or multiple frames. The data from the sensing elements can be combined based on properties of the charged particle arrival events. For example, the data from the sensing elements can be combined based on timestamps of the charged particle arrival events, sensing element locations, scanning actions of the primary beam of the charged particle beam device, or properties of the charged particle beam device. The combined data can be used to reconstruct an image of the sample. For example, the pixel counts and 970 can be fed into an image processing system 199. The count summing unit 960 can include logic operation processing, which can be implemented using hardware, such as electronic circuits, or for example, by a controller.

[0305] Various signal lines, etc., can be provided to connect Figure 19A the various elements in the detection system of FIG. 9. For example, there can be one pixel clock 920 connected to each individual circuit of the event detector 930. In addition, a detection control 925 can be connected to each individual circuit of the event detector 930. Similarly, there can be another pixel clock 940 connected to each individual circuit of the pixel count buffer 950. In addition, a detection control 945 can be connected to each individual circuit of the pixel count buffer 950. Also, a delayed pixel clock 965 can be connected to the count summing unit 960. The pixel clocks 920, 940, and 965 can run at the same or different speeds.

[0306] Figure 19AThe detection system of FIG. 1 can use relatively simple electronic components and can achieve high speed while achieving good packaging flexibility. For example, components can be integrated on a semiconductor chip. The front-end electronics 910, event detector 930, pixel count buffer 950, or count summing unit 960 can be provided in a semiconductor package such as a single monolithic semiconductor chip. For example, the front-end electronics 910, event detector 930, pixel count buffer 950, and count summing unit 960 can be provided as layers in a semiconductor chip that can be connected to the detector 800 and image processing system 199. In some embodiments, the front-end electronics 910, event detector 930, pixel count buffer 950, and count summing unit 960 can be provided as separate modules. Figure 19A The dashed lines shown between the image processing system 199 and the count summing unit 960 in FIG. 1 can represent a division between on-chip and off-chip structures. In some embodiments, the front-end electronics 910, event detector 930, pixel count buffer 950, and count summing unit 960 can be provided as separate modules.

[0307] Counting electrons can have a number of advantages compared to detecting analog signals. For example, as discussed in embodiments herein, a semiconductor chip including a detector and a detection system can achieve higher speed and can avoid processing bottlenecks. As discussed above, overlap of signal pulses in adjacent electron arrival events can impede detection accuracy. Thus, it can be important for a detector to have high speed. In some devices, front-end electronics can limit bandwidth. However, as discussed with respect to embodiments herein, high speed can be achieved when front-end electronics or other components are integrated in a semiconductor chip. In some embodiments, components configured to count multiple charged particles, such as by providing an output in the form of event markers, can be relatively simple and can allow a system to be simplified and high speed. Rather than sampling analog signals to represent beam intensity, it can be advantageous to detect individual events and count the number of individual events that occur over a period of time.

[0308] In some types of applications, such as CD SEM, high resolution high throughput inspection, or metrology of manufactured semiconductor devices, electron counting can be particularly effective. In some embodiments discussed herein, a detection system can enable electron counting of electron beams of, for example, 100 pA or more.

[0309] In a comparative example, some detectors can be provided in high energy particle physics analysis instruments. For example, a particle accelerator can propel charged particles at high speed and high energy. The charged particles can collide with material or other particles, and collision products can be detected by a detector. The detector can be configured to receive the particles. However, in high energy physics instruments, the detector can be configured such that the particles pass through the detector. For example, a calorimeter can measure energy lost by a particle as it passes through the calorimeter.

[0310] In contrast to, for example, a calorimeter, a sensing element configured to count the number of charged particles can receive a charged particle and, in response to the charged particle reaching the sensing element, generate an electron-hole pair in the sensing element until the energy of the charged particle is exhausted. Further, a circuit connected to the sensing element can be configured to discriminate the energy level of the incident charged particle, as should be discussed herein. Still further, a circuit connected to the sensing element can be configured to sum the counts of charged particle arrival events occurring at multiple sensing elements in an array. The detector can derive information from the sensing elements, including the timing of receiving a charged particle and its energy level.

[0311] In a scanning electron microscope (SEM), secondary electrons (SE) and backscattered electrons (BSE) can be emitted from a specimen. It can be useful for a SEM system to distinguish between SE and BSE and to generate two separate images (SE image and BSE image) at the same time. The two images can show different properties of the specimen in two-dimensional space because of the two different mechanisms of SE and BSE generation processes at the specimen in response to electron impact on the specimen.

[0312] In some embodiments, multiple thresholds can be set to acquire both SE and BSE images simultaneously. The event signal pulses entering the event detector can be proportional to the energy of the incoming electrons that caused the event signal pulses of interest. The average energy of BSEs is higher than the average energy of SEs. There can be a distinct energy distribution gap between SEs and BSEs. In this case, for example, a first threshold can be set to a value equal to 60% of the average height of the event signal pulses caused by SEs. A second threshold can be set between the average height of SE event signal pulses and the average height of BSE event signal pulses. Corresponding to the two thresholds, two detection flags, for example, flagLow and flagHigh, can be defined. Each flag can be set when an event signal crosses the corresponding threshold on its rising edge is determined. Thus, both flagLow and flagHigh can be set by event signal pulses caused by BSEs, and flagLow can be set only by SE-caused event signal pulses while flagHigh can remain reset. Through simple logic, the event detector can determine which type of electron is detected. The logic can be implemented, for example, in the event detector 930 or the count buffer 950. In some embodiments, the count buffer can be an advantageous location, for example, where the count buffer functionality has less stringent requirements on timing and can perform summation over multiple clock cycles in a pipelined architecture. Adding one cycle of execution logic before the series of clock cycles used for summation can not compromise the overall performance of the detector system. With this functionality in a detector system such as the detection system 901, summation can be performed independently between SE counts and BSE counts, and the summation unit 960 can output two sums per delayed pixel clock cycle. To accurately discriminate between SEs and BSEs, additional thresholds can be added.

[0313] Reference is now made to Figure 19A which illustrates a schematic representation of another exemplary architecture of a detection system, consistent with embodiments of the present disclosure. Figure 19B The detection system 902 of Figure 19BThe detection system 901 of FIG. 1 can be modified to include an energy storage device 1310 coupled to the detector 900 including the array of sensing elements. The energy storage device 1310 can include individual energy storage units 1311, 1312, 1313, 1314, etc., for example, up to an nth energy storage unit 1319. Each individual energy storage unit can correspond to a sensing element in the array of sensing elements of the detector 900. The energy storage units 1311-1319 can be configured to accumulate energy in response to an output signal from a corresponding sensing element in the array of sensing elements of the detector 900 being fed to the energy storage units 1311-1319. Each energy storage unit can include, for example, a micro-capacitor.

[0314] The raw detection signal from the sensing element can be fed into the energy storage unit. The energy accumulated in the energy storage unit can be stored until the energy storage unit is reset. For example, when the energy storage device 1310 includes a capacitor, the energy level can be reset when the capacitor is discharged. The stored energy level can be read when energy is fed into the energy storage device 1310. The energy level can be maintained for a period of time until reset.

[0315] In some embodiments, the energy associated with the sensing element output can be reset by active reset or passive reset. Active reset can include positively causing the element to lower its energy level. For example, active reset can include discharging a capacitor. Passive reset can include passively allowing the energy level to be lowered. For example, in some devices, the energy level can decay over time. In the sensing element itself, the energy generated in response to an electron arrival event can dissipate over time. In some embodiments, as discussed above, the current pulse generated by a single electron arrival event at a PIN detector can have a pulse width of, for example, 3-5 ns, and thus, the sensing element can be passively reset after approximately 3-5 ns.

[0316] In some embodiments, the energy storage device 1310 can be omitted. For example, the sensing element itself can function as an energy collection unit, and can be inherently periodically reset as the electrical surge of the electron-hole pair generated in response to an electron arrival event gradually dissipates.

[0317] The detection system 902 can also include a detection circuit array 1320. The detection circuit array 1320 can include individual circuits 1321, 1322, 1323, 1324, and so on, up to an nth circuit 1329. Each individual circuit of the detection circuit array 1320 can correspond to a sensing element of the detector 900 array. Each individual circuit of the detection circuit array 1320 can include an event pulse detector. The event pulse detector can be configured to detect an arrival event of a charged particle on the sensing element. For example, the event pulse detector can be configured to detect a charged particle arrival event by detecting an amount of energy in the sensing element, and can be configured to increment a counter, thereby indicating that a charged particle has been counted. In some embodiments, other ways can be used to detect an output signal from a sensing element. For example, an output from a sensing element can be read by sampling at a predetermined time. The detection circuit array 1320 can include a clock. In some embodiments, a global clock can be provided. Using a global clock, the operation of a sensing element and its associated circuitry can be synchronized. In some embodiments, individual sensing elements and their associated circuitry can have their own clock.

[0318] The circuitry of an event pulse detector can include a comparator. In addition, various other circuit components can be provided, such as a voltage reference. In some embodiments, the energy captured in an energy storage unit can exceed an overflow limit, and can hinder the counting of charged particles. For example, a miscount can occur when more than one charged particle is received in a sensing element and the received energy can not indicate the total number of received charged particles. The energy accumulated in the energy storage unit can remain the same even after a subsequent charged particle is received. It can therefore be beneficial to provide circuitry for processing the data of a sensing element in time. The detection circuit array 1320 can include storage for storing sensing element data that can be associated with a time stamp.

[0319] The detection system 902 can also include a count summing unit 960. The output from each circuit of the detection circuit array 1320 can be fed into the count summing unit 960. The counts of charged particles received from each sensing element in the detector 900 can be summed to obtain a gray scale of an image to be used in image processing. For example, the pixel count sum 970 can be fed into an image processing system 199. The count summing unit 960 can include a logic operation process, which can be implemented using hardware, such as electronic circuitry, or for example, by a controller.

[0320] Various signal lines, etc. can be provided to connect various elements in the detection system 902. For example, the detection system 902 can include other components, such as amplifiers, signal processing circuitry, etc. It is to be understood that various signal lines, etc. can be provided, such as Figure 19AVarious connections and other elements shown can be added to the detection system 902.

[0321] In some embodiments, the sampling period τ S of the detector can be set to be less than the dead time τ D , where the electron count in any sampling period must be 0 or 1, since any two consecutive detection pulses will be separated by one sampling period or more. Thus, the count buffer can be configured with one bit, for example, as a flag. In some embodiments, when applied to a SEM, the fastest pixel rate of the detector can be configured to be 1 / τ S Hz. While the sampling period is fixed, a SEM image with a lower pixel rate (e.g., 1 / (Nτ S ) where N = 2, 3,...) can be constructed in the image processor by grouping and summing each consecutive count result and representing one pixel with the sum of all counts of the group. If 1 / τ S is set to, for example, 400 MHz, then SEM images with variable but discrete pixel rates 400 / N MHz (N = 1, 2, 3,...) can be supported.

[0322] In some embodiments, the sampling period τ S may be set to be greater than the dead time τ D , where the event detector can generate multiple detection pulses within a single sampling period. Thus, the count buffer can be configured with multiple bits. The count buffer can be configured with multiple bits so that additional false counts can be avoided and the overall false count rate can be maintained at an acceptable level. As mentioned above, a SEM image is formed at the image processor end at a lower pixel rate while it can be advantageous for the detector end to run at a fixed sampling period τ S . For example, since modern foundry chips can achieve clock operation at 400 MHz, it can be rare that τ S > τ D .

[0323] Reference is now made to Figure 19Awhich shows an exemplary representation of output signals of sensing elements coupled to a circuit, where the sensing elements can be connected to the circuit on a per sensing element basis. As used herein, the term "output signal," "sensing element output," or "output of a sensing element," and the like can refer to the output of a sensing element or its associated circuit. The output signal can be generated by the sensing element, and its output can be coupled to a corresponding circuit. A controller, such as a central electronic control unit, can then receive the output from the circuit, rather than directly from the sensing element. The controller can receive the output from multiple circuits. In some embodiments, the sensing element can not only include a PIN diode, but can include some other components that can form a circuit within the sensing element. The output of the sensing element can then be directly connected to the controller. The controller can be configured to receive the output of the sensing element through the respective circuit coupled to each sensing element.

[0324] The sensing elements can output a signal based on time. The circuit can include a functional block that can process the signal from the sensing elements to derive a detection signal. Figure 20A is a plot that can represent the detection signal (intensity) in arbitrary units on the vertical axis plotted against time on the horizontal axis. The detector can be configured to have a speed such that individual charged particle arrival events at each sensing element can be distinguished. For example, as shown in Figure 20A a sudden increase in the signal can correspond to a charged particle arrival event. The charged particle arrival events can have corresponding energy levels, such as El, E2, E3, E4, and the like.

[0325] Figure 20A A plurality of frames Fl, F2, and F3 at a constant data frame rate are shown. The frames Fl, F2, and F3 can be data frames. A data frame can refer to a set of data acquired by the detection system over a specified period of time. A frame can include one or more sampling periods of the sensing elements. The frame rate can be the frame rate of the controller performing processing such as making charged particle count determinations, determining the first group, determining the second group, processing data for image processing, and the like within each frame. The frame rate can be related to, for example, the system clock rate. As discussed above, the frame rate can be set according to the first parameter and the second parameter. The first parameter or the second parameter can be based on predetermined criteria, and can include or override the first predetermined frame criteria or the second predetermined frame criteria, which will be discussed below.

[0326] Figure 20A Another representation of a graphical display of exemplary output signals of sensing elements is shown. As discussed above, the output signals can be generated by the sensing elements, and their outputs can be coupled to a corresponding circuit. A controller, such as a central electronic control unit, can then receive the output from the circuit, rather than directly from the sensing element. The controller can receive the output from multiple circuits. In some embodiments, the sensing element can not only include a PIN diode, but can include some other components that can form a circuit within the sensing element. The output of the sensing element can then be directly connected to the controller. The controller can be configured to receive the output of the sensing element through the respective circuit coupled to each sensing element. Figure 20BDifferent, the output signal of the sensing element can be represented as constant for the entire frame. Thus, in some embodiments, for one frame, the sensing element can be considered to have one value. When multiple charged particle arrival events occur during one frame, the signal strength can be higher than when only one charged particle arrival event occurs.

[0327] The frame rate can be variable. The period of the frame can vary over time. For example, Figure 20A Another graph showing the output signal of the sensing element and circuit is shown. In Figure 20C In some embodiments, the adaptation can occur within the same frame. For example, a first frame can be determined and can be adjusted concurrently. In some embodiments, real-time adjustments to the frame can be used, which can help reduce the error rate in the detection of the output signal.

[0328] The period can be set so that one charged particle is counted within one period. The period can be initially set based on the time period for collecting the charged particles. The period can also be initialized in other ways. Thereafter, the period of the next frame can be the same or different. In some embodiments, an iterative approach can be used. For example, in Figure 20C In some embodiments, the adaptation can occur within the same frame. For example, a first frame can be determined and can be adjusted concurrently. In some embodiments, real-time adjustments to the frame can be used, which can help reduce the error rate in the detection of the output signal.

[0329] Reference is now made to Figure 20C which shows a representation of the output signal of the sensing element that can be coupled to the circuit relative to a first threshold Tl. The first threshold Tl can be set to filter out noise. For example, the first threshold Tl can be set to a value greater than 0 and less than or equal to a value corresponding to a predetermined energy of a charged particle arrival event. The first threshold Tl can be used to determine that the output signal is a detection signal, and not noise caused by, for example, interference, dark current, etc.

[0330] Reference is now made to Figure 21AThe graph shows a representation of output signals of sensing elements that can be coupled to a circuit relative to a first threshold Tl and a second threshold T2. The first threshold Tl can be set similarly as above to filter out noise. The second threshold T2 can be set to distinguish different types of charged particles having different energy characteristics. For example, the second threshold T2 can be set to distinguish secondary electrons and scattered electrons. The value of the second threshold T2 can be predetermined. When Figure 21B the vertical axis of the graph is in electron volts, the second threshold T2 can be set to a value of the acceleration energy imparted by the SEM system (e.g., E_acc) plus 50 eV. Thus, electrons having an energy less than or equal to the second threshold T2 can be determined to be secondary electrons, and electrons having an energy higher than the second threshold T2 can be determined to be scattered electrons and counted accordingly.

[0331] Referring now to Figure 21B , a representation of output signals of sensing elements that can be coupled to a circuit relative to a first threshold Tl, a second threshold T2, and a third threshold T3 is shown. In the representation of Figure 21C , the output detection signals can take a value per frame. The first threshold Tl can be set similarly as above to filter out noise. The second threshold T2 can be set to distinguish different types of charged particles. The third threshold T3 can be set to determine whether a sensing element received more than one charged particle. The value of the third threshold T3 can be twice the value of the second threshold T2.

[0332] Referring now to Figure 21C , an exemplary representation of a statistical result of a frame is shown. In some embodiments, a controller of the detection system can be configured to perform a statistical analysis at each frame. For example, sensing elements of the detector can be output to the controller. The sensing elements can be connected to a circuit. The controller can be configured to receive the output from each sensing element and organize the output of each sensing element frame by frame. After one frame, the controller can collect the output from a plurality of sensing elements associated with a beam spot. For example, as discussed above with respect to Figure 22 , the controller can collect the output from a first group of sensing elements contained within the boundary line 350. The controller can perform a statistical analysis based on the information of the output of each sensing element at the frame. As an example, Figure 3F a charged particle count on the vertical axis and an energy level on the horizontal axis are shown for one frame. The energy level can correspond to all different energy levels of charged particles received by the first group of sensing elements in one frame. For example, the energy level can correspond to the energy levels El, E2, E3, E4 discussed above with respect to Figure 22 . Although Figure 20AFour such energy levels are illustrated, but it's important to understand that different numbers of individual energy levels can be provided and analyzed. The count of charged particles can correspond to all charged particles detected by the first set of sensing elements in a single frame.

[0333] Other types of statistics can be determined. For example, the following can be determined: the proportion of charged particles at a specific energy level to all received charged particles; the proportion of charged particles received in a sensing element to the total number of charged particles received in the first group, etc.

[0334] although Figure 22 The example statistical results are shown as a histogram, but various other representations can be used. For example, statistical analysis can provide a scatter plot.

[0335] In some embodiments, the controller can be configured to perform statistical analysis on a pixel-by-pixel basis. The controller can be configured to receive output from each sensing element and organize the output of each sensing element corresponding to a pixel in the SEM image. For example, when a charged particle beam is scanned across a sample, sensing element outputs corresponding to different pixels can be generated. Each pixel in the SEM image can be generated using frames acquired during the dispatched time period of the scanning operation. For example, one or more frames of data can be acquired during the period associated with the pixel. Thus, information corresponding to a pixel can be included in one or more frames. As used herein, the term "pixel" can refer to a unit area on the imaged sample surface. Therefore, an SEM image can include a pixel map, with each pixel corresponding to a location on the sample surface. The higher the imaging resolution (or, for example, the larger the field of view on the sample surface), the greater the number of pixels.

[0336] Figure 22 The diagram illustrates a representation of multiple pixels on a line. Pixels can be organized according to a scanning pattern, such as a raster scan of an area of ​​interest on a sample surface. Each pixel can be associated with data output from a sensing element at a specific time. Each pixel can be associated with data from one or more sensing elements. As discussed above, a pixel can include one or more frames of data associated with the sensing element output.

[0337] For multiple pixels to be used in an SEM image, statistical analysis can be performed on a pixel-by-pixel basis based on information associated with each pixel. For example, a controller can be configured to receive the output of a sensing element via corresponding circuitry coupled to each sensing element. For a single pixel, the controller can collect output from a sensing element associated with a beam spot and within a time period associated with the pixel. For example, the timestamp of a pixel can correspond to the time when the primary electron beam of a charged particle beam imaging tool scans a location on the sample surface. For example, as referenced above... Figure 23As discussed, the controller can collect output from the sensing elements of the first group that are contained within the boundary line. The controller can perform a statistical analysis based on information of the output of each sensing element in one or more frames associated with the pixel. As an example, Figure 3F Sensing element output associated with one pixel X1 is shown on the left. The sensing element output can be represented with information such as charged particle counts plotted against energy levels based on the statistical analysis. The information of charged particle counts plotted against energy levels can be based on information of multiple frames. For example, in Figure 23 there are two frames of information associated with pixel X1. Sensing element output associated with another pixel can be provided at different time points of the scan. For example, Figure 23 Sensing element output associated with another pixel X2 is shown on the right. For pixel X2, there can be only one frame of information associated with the pixel.

[0338] In the representation of information of charged particle counts plotted against energy levels, the scale can be determined based on data of all multiple pixels collected in the scan. The energy levels of the horizontal axis can include all energy levels of the charged particles collected in the scan. The scale for each pixel in the image can remain constant. For example, the scale can be determined so that gray scale distortion can be avoided. In addition, all images acquired can be set to have the same scale under the same settings of the charged particle inspection tool. Thus, gray scale distortion between images can be avoided.

[0339] According to some embodiments of the present disclosure, information indicative of the energy spectrum of the incident charged particles can be generated. The information can be used to generate pixels of an image in charged particle beam imaging, such as a SEM image.

[0340] For example, the overall numerical output can be used to generate one pixel in a SEM image, such as a gray scale image (a type that can be used in conventional SEMs). That is, the total count of the incident charged particles can correspond to the gray scale of the pixel. The total count of the charged particles can correspond to the intensity. In addition, color information can be added. For example, visible light is the portion of the electromagnetic spectrum that is visible to the human eye, which corresponds to wavelengths of approximately 390 to 700 nm or photon energies of approximately 1.63 to 3.26 eV. In a similar manner that visible colors such as red, orange, yellow, green, blue, and violet correspond to certain energy levels, color values can be assigned to pixels in a SEM image. For example, the RGB value of each pixel in the SEM image can be determined based on the statistical analysis of the sensing element output analyzed on a pixel-by-pixel basis. The RGB value can be based on the number of charged particles received at a particular energy level.

[0341] In some embodiments, the spectrum of radiation received from a sample can be represented by the count of incident charged particles at discrete frequency bands of energy levels. The energy level spacing can be determined, for example, predetermined or based on statistical analysis. In some embodiments, the spacing can be based on thresholds including, for example, thresholds T1, T2, and T3.

[0342] In some embodiments, statistical analysis of the sensing element output may include providing an approximation of the radiation spectrum via a function such as F(x), where x is an energy level. Color specifications can be directly transformed from the function into color information, such as RGB values ​​for the color gamut.

[0343] Although the RGB color model has been discussed above, in some implementations, other representations of color, such as HSL (hue, saturation, brightness) or HSV (hue, saturation, value), may be used.

[0344] In some embodiments, additional degrees of freedom can be obtained in information acquisition and display. Color can be added to SEM images to represent additional properties, such as material properties or microstructure.

[0345] Examples of controllers consistent with the above description include Figure 23 The controller 109 may be part of an image processing system 199, which also includes an image acquirer 120 and a storage device 130 (see [link]). Figure 1 In some embodiments, the controller 109 may be a separate control unit capable of performing image acquisition functions.

[0346] Now refer to Figure 2B The diagram illustrates a flowchart of an exemplary period determination method. This can be applied to a controller (e.g., Figure 24 The controller 109 is programmed to implement Figure 1 The flowchart includes one or more blocks. In step S101, charged particle imaging can be started. In step S102, the processing period is set to an initial value. The processing may include one or more of the following: determining charged particle counts, determining a first set of sensing elements, determining a second set of sensing elements, and performing image processing, etc.

[0347] In step S103, the controller may receive a data frame associated with the output of a sensing element. This frame may include the outputs of multiple sensing elements from the detector, which can be connected to the circuit on a per-sensing-element basis. Therefore, the controller may receive multiple detection signals in a single frame, each signal corresponding to the output of a respective sensing element.

[0348] In step S104, a number of charged particles incident on the detector within the frame can be determined. Step S104 can include determining a number of charged particles incident on each of a plurality of sensing elements of the detector within the frame. The number of charged particles can be a total number.

[0349] In step S105, a determination can be made as to whether the number of sensing elements receiving at least one charged particle is greater than or equal to a first number. The first number can be based on a first predetermined frame criterion. For example, the first number can be a number of sensing elements corresponding to a standard size of an electron beam spot based on parameters of a charged particle beam tool such as Figure 24 the apparatus of FIG. 1, etc. Figure 1 The determination of step S105 can also be in the form of a ratio. For example, step S105 can include determining whether at least A% of the sensing elements received at least one charged particle.

[0350] When a negative determination is made in step S105, the process can proceed to step S107. In step S107, the period can be adjusted. For example, the period can be made longer so that more sensing elements can receive at least one charged particle in one frame. The period can be increased by a predetermined amount. Thereafter, the process can return to step S103.

[0351] On the other hand, when a positive determination is made in step S105, the process can proceed to step S106. In step S106, the process can end. In some embodiments, another process can follow.

[0352] A method can include elements of the flowchart of FIG. 1 with or without various modifications. Figure 24 For example, concurrently with step S104, a boundary line can be determined, which can be provided to enclose and group sensing elements receiving at least one charged particle. The sensing elements receiving at least one charged particle can be associated with the same charged particle beam spot. For example, the sensing elements associated with the same beam spot can be adjacent to one another.

[0353] In some embodiments, the determination made in step S105 can occur once per frame. In some embodiments, the determination can occur once for a plurality of frames. For example, instead of determining whether to adjust the period of the frame after each frame, the determination of step S105 can be made after a predetermined number of frames. Thus, for the predetermined number of frames, the frame can have the same period, after which a determination to adjust the period can be made.

[0354] Reference is now made to Figure 24 which illustrates a flowchart of an exemplary period determination method. A controller (e.g., controller 109 of FIG. 1) can be programmed to implement the method of FIG. 2. Figure 25 Figure 1 ​one or more blocks of the flowchart. Figure 25 The process of FIG. 2 can begin in step S201. The process can follow directly from Figure 25 step S106 of FIG. 1.

[0355] In step S202, the controller can receive a frame of sensor element output. The frame can include output from a plurality of sensor elements of the detector. Thus, the controller can receive a plurality of detection signals in one frame, each signal corresponding to output of a respective sensor element.

[0356] In step S203, a number of charged particles incident on the detector in the frame can be determined. Step S203 can include determining a number of charged particles incident on each sensor element of the plurality of sensor elements of the detector in the frame.

[0357] In step S204, a determination can be made as to whether the number of sensor elements receiving more than one charged particle is less than or equal to a second number. The number can be in the form of a percentage. For example, step S204 can include determining whether no more than B% of the sensor elements receive more than one charged particle. The second number can be based on a second predetermined frame criterion. For example, the second number can be based on a predetermined requirement for linearity of the measurement. In some embodiments, the second number can be based on linearity of the output image signal corresponding to the incident charged particle count. For example, when the incident charged particle beam is too strong for the charged particle counting capability to handle, the sensor element output can exhibit non-linear behavior. The sensor elements can exhibit non-linear output behavior when more than one charged particle is incident on the sensor element in one frame. The predetermined criterion can be based on a desired level of linearity of the measurement, such that the number of sensor elements receiving more than one charged particle is limited.

[0358] When a negative determination is made in step S204, the process can proceed to step S206. In step S206, the period can be adjusted. For example, the period can be made shorter, such that fewer sensor elements will receive more than one charged particle in one frame. The period can be reduced by a predetermined amount. Thereafter, the process can return to step S202.

[0359] On the other hand, when a positive determination is made in step S204, the process can proceed to step S205. In step S205, the process can end. In some embodiments, other processing can follow.

[0360] A method can include the process of FIG. 2 with or without various modifications. Figure 24of the flowchart of FIG. 2. For example, concurrently with step S203, a boundary line can be determined, which can be provided to enclose and group sensing elements that receive more than one charged particle. Sensing elements that receive more than one charged particle can be associated with a high intensity region of a charged particle beam spot.

[0361] It will be appreciated that Figure 25 and Figure 24 The processes of FIG. 2 and FIG. 3 can be incorporated into a control routine. For example, the flowchart of the processes of FIG. 2 can be modified such that blocks S204 and S206 of FIG. 2 operate in parallel with blocks S105 and S107. Figure 25 Figure 24 In addition, the control routine can include additional processing, such as determining to adjust a setting or structure of the charged particle beam device based on the determination relating to the first predetermined criterion or the second predetermined criterion. Reference is now made to FIG. 4, which illustrates a flowchart of an exemplary determination method.

[0362] The process of FIG. 4 can begin in step S301. Figure 25 Figure 26 In step S302, a controller (e.g., controller 109 of FIG. 1) can receive a data frame relating to sensing element outputs. The frame can include outputs from a plurality of sensing elements of a detector that are connectable to a circuit on a per-sensing element basis. Thus, the controller can receive a plurality of detection signals in one frame, each signal corresponding to an output of a respective sensing element.

[0363] In step S303, a number of charged particles incident on the detector within the frame can be determined. Step S303 can include determining a number of charged particles incident on each sensing element of the plurality of sensing elements of the detector within the frame. Figure 26 In step S304, the controller can determine whether a first criterion has been met. For example, a determination can be made as to whether a number of sensing elements that receive more than one charged particle is less than or equal to a first number. Step S304 can include determining whether no more than B% of the sensing elements receive more than one charged particle.

[0364] When a positive determination is made in step S304, the process can proceed to step S305. In step S305, the controller can determine whether a second criterion is met. For example, a determination can be made as to whether a number of sensing elements that receive at least one charged particle is greater than or equal to a second number. Step S305 can include determining whether at least A% of the sensing elements receive at least one charged particle.

[0365]

[0366]

[0367] ​​​​When a negative determination is made in step S304, the process can proceed to step S307. A negative determination in step S304 can mean that the first criterion and the second criterion should be adjusted. For example, a negative determination in step S304 can mean that the number B% can be set too low or the number A% can be set too high. It can also mean that the distribution of charged particles within the beam spot is not uniform enough. In step S307, adjustments can be made. Adjustments can be made to make the distribution of charged particles within the beam spot more uniform. The adjustments can include defocusing the projection system in the charged particle beam device. For example, the projection system can be configured to defocus the beam to a certain extent. Furthermore, the magnification of the charged particle beam device can be changed to enlarge the beam spot. Further adjustments can also be made. For example, the frame period can be made shorter. Thereafter, the process can return to step S302.

[0368] When a negative determination is made in step S305, the process can proceed to step S308. In step S308, the numbers A% and B% can be adjusted. Further adjustments can also be made. For example, the frame period can be made longer. Thereafter, the process can return to step S302.

[0369] When a positive determination is made in step S305, the process can proceed to step S306. In step S306, the process can end.

[0370] Frame adaptation can occur after the current frame period has ended. For example, adjustments can be made in a subsequent frame in response to determining that the sensor output does not meet the criterion for the current frame. For example, the subsequent frame can be shortened or lengthened as in the example discussed above. However, in some embodiments, frame adaptation can occur in real time. A current frame period can be determined. A determination can be made to shorten the current frame or to lengthen the frame. For example, the determination can be made when the sensor output is unable to meet the criterion due to the period of the current frame.

[0371] Reference is now made to Figure 1 which illustrates a flowchart of an example period determination method. A controller (e.g., the controller 109 of Figure 27 ) can be programmed to implement one or more blocks of the flowchart of Figure 1 . The process can begin in step S401. Step S401 can follow processing such as Figure 27 , Figure 24 or Figure 25 . From step S401, charged particle imaging can begin or can be ongoing. The period of the frame can be set to a value such as from a previous frame processing or has been initialized to a predetermined initial value.

[0372] Figure 26 or Figure 24The processing of the data can result in an adjustment that can be applied to a next data frame after the current data frame. In such processing, the current frame can end after the controller receives the data even with the adjustment. However, in the processing of the data, the current frame can not necessarily end after the data is received. For example, if the data does not meet the criteria, an adjustment can be made to the current frame and updated data reflecting the adjustment can then be transmitted to the controller. When the data transmitted to the controller meets the criteria, then the current frame can end. Figure 25

[0373] Continuing the process of FIG. 4A, in step S402, the controller can determine properties of the current frame. The properties of the current frame can include a frame period. Step S402 can include reading a value from a storage device. Step S402 can also include determining properties related to the current frame, such as imaging conditions of the SEM system operating during the current frame. Figure 27 In step S403, the controller can receive a frame of data related to the sensing element outputs. The frame can include outputs from the plurality of sensing elements of the detector connectable to the circuit on a per sensing element basis. Thus, the controller can receive a plurality of detection signals in one frame, each signal corresponding to an output of a respective sensing element.

[0374] In step S404, a number of charged particles incident on the detector within the frame can be determined. Step S404 can include determining a number of charged particles incident on each sensing element of the plurality of sensing elements of the detector within the frame.

[0375] In step S405, a determination can be made as to whether the criteria can be met based on the received output signals, properties of the current frame, or other properties determined previously. The criteria can include the first predetermined frame criteria or the second predetermined frame criteria as discussed above with respect to FIGS. 3A and 3B. For example, step S405 can include determining whether at least A% of the sensing elements received at least one charged particle or whether no more than B% of the sensing elements received more than one charged particle. In

[0376] Figure 27 Figure 24 Figure 25 ​​​​In step S405, it can be determined based on the current frame properties that the output from the sensing elements in the array will not satisfy the criteria. For example, it can be determined that the current frame has such a short period that there will not be multiple sensing elements receiving more than the first number of at least one charged particle within the current frame. The current frame period can be determined to be too short based on the average number of electron arrival events determined for a particular level of beam current. It can be determined that, under the current imaging conditions using, for example, a 4nA electron beam, it is unlikely that more than one electron will be received by more than the first number of sensing elements because the electron arrival events will be too sparse. Step S405 can include a determination based on a confidence level. In some embodiments, the determination in step S405 can be based on information gathered in steps S402, S403, or S404. Step S405 can include a determination based on a first predetermined frame criteria (e.g., at least A% of the sensing elements receiving at least one charged particle) and a second predetermined frame criteria (e.g., no more than B% of the sensing elements receiving more than one charged particle). The determinations can be made in parallel or in separate process flows. For example, it can be determined together in one step whether the first and second predetermined frame criteria can be met, or it can be determined separately whether the first predetermined frame criteria can be met and then whether the second predetermined frame criteria can be met, or vice versa.

[0377] When a negative determination is made in step S405, the process can proceed to step S406. In step S406, the period of the current frame can be adjusted, or other adjustments can be made to the frame. For example, the period can be made longer so that more sensing elements can receive at least one charged particle in one frame. The amount of adjustment made can be proportional to the amount of deficiency anticipated, for example, from the determination made in step S405. After adjusting the current frame, the process can return to step S403.

[0378] When a positive determination is made in step S405, the process can proceed to step S410 without adjustment of the period, and the process can end. In some embodiments, other processing can follow. For example, the process can end with respect to the current frame and can be repeated for a new frame.

[0379] A method can include elements of the flowchart of Figure 27 with or without various modifications, such as those discussed above with respect to Figure 27 , Figure 24 and Figure 25 . Moreover, although the above examples discuss shortening of the current frame period, it is also possible that there is lengthening of the current frame period. In some embodiments, there can also be a step of determining whether the number of sensing elements receiving more than one charged particle is less than or equal to a second number, similar to step S204 discussed above with respect to Figure 26 or step S404 discussed above with respect to Figure 25The step S305 is discussed. In some embodiments, the steps S403 to S406 can occur iteratively.

[0380] In some detectors, the detection process can comprise two steps with a fixed sequence. The two steps can be used to determine the intensity of the secondary electron beam landing on the detector. Such steps can be used together with or instead of a period determination process, such as the processes discussed above. For example, there can be a first step that identifies the boundaries of the beam spot on the detector surface. And there can be a second step that determines the intensity of the electron beam based on the grouped sensing elements corresponding to the beam spot. Information indicative of the sample surface can be derived from the determined beam intensity.

[0381] The above-mentioned two steps with a fixed sequence can be useful in determining the beam intensity if the electron beam is incident on the detector with a fixed projection pattern. However, complications can arise if there is a variation or fluctuation in the projection pattern of the electron beam on the detector. For example, it can become difficult to determine the intensity of the electron beam when the electron distribution of the electron beam is on multiple non-adjacent sensing elements of the sensing element array of the detector. This can occur, for example, when the beam current is low and a relatively small number of electrons are incident on a large area of the sensing element array. With a low rate of electron arrival, individual electron arrival events on the sensing elements can be sparse and can be spread over a wide area. As discussed above, the geometric spread of the landing positions of the electrons in the secondary electron beam can be large due to the different trajectories of the electrons, which can depend on, for example, the initial kinetic energy and the angle of emission of the electrons.

[0382] Reference is now made to Figure 26 which illustrates a case where the charged particle beam device can scan a primary beam of charged particles over a sample surface in a raster pattern. The beam can be continuously scanned over the sample surface. The beam scan path can correspond to the pixels X i where i is an index. As Figure 28A illustrated in the sensing element array 2100 of the detector, there can be a sensing element 2101 that receives incident charged particles at a scan time T of T1. The detector can be configured for charged particle counting, and thus, the rate of charged particle arrival at a particular instant can be relatively low. As the primary charged particle beam is scanned over the sample surface, secondary or scattered charged particles can be generated that are incident on the detector. The distribution of the landing positions of the charged particles in the secondary charged particle beam can be relatively wide. Thus, at time T1, the sensing element 2101 can receive incident charged particles. However, at a next detection frame, for example, at time T2 as Figure 28B illustrated, charged particles can be incident on a different region of the sensing element array 2100. AsFigure 28C As shown, at scan time T of T2, there can be a sensing element 2102 and a sensing element 2103 that received an incident charged particle during the sampling period, respectively. Next, as shown, at scan time T of T3, there can be a sensing element 2104 that received an incident charged particle in the next frame. Further, as shown, at scan time T of T4, there can be a sensing element 2105 and a sensing element 2106 that received an incident charged particle in yet another frame, respectively. Although the sensing elements 2101, 2102, 2103, 2104, 2105, and 2106 are not necessarily adjacent to each other, they can all be associated with the same secondary charged particle beam, and thus, some grouping can be required to determine the intensity of the secondary beam incident on the detector. Figure 28C Figure 28D As shown, at scan time T of T4, there can be a sensing element 2105 and a sensing element 2106 that received an incident charged particle in yet another frame, respectively. Although the sensing elements 2101, 2102, 2103, 2104, 2105, and 2106 are not necessarily adjacent to each other, they can all be associated with the same secondary charged particle beam, and thus, some grouping can be required to determine the intensity of the secondary beam incident on the detector.

[0383] In some detection systems, it can be difficult to associate sensing elements with one charged particle beam spot when a charged particle arrival event occurs across a wide area including multiple sensing elements. Further, in some detection systems, there can be multiple sub-beams of a primary charged particle beam that simultaneously scan a sample, and thus, there can be multiple sub-beams of secondary charged particles incident on the detector. Thus, further complexity can arise in distinguishing between secondary charged particle arrival events and different beams, which can require individual grouping of sensing elements. Additionally, the flexibility of the detector can be limited due to the requirement of grouping sensing elements. If the projection pattern of the secondary beam of charged particles is not constant on the detector surface, for example, when a beamlet deflection unit is not functioning properly, which can cause the projection pattern to move rapidly on the detector surface, it can be difficult to track the beam. The requirement of processes such as grouping of sensing elements can hinder the ability to track the movement of the projection pattern and can introduce detection errors.

[0384] Some embodiments of the present disclosure can provide a detection process in which sensing elements do not need to be grouped prior to counting of charged particles. Rather, grouping can be performed after the charged particles have been counted and data representing the detector condition at a particular time has been stored. In some embodiments, a time stamp can be associated with the sensing element data. Sensing element data corresponding to a particular time stamp can be associated with a scan time to associate the sensing element data with a scan position on the sample surface. For example, sensing element data including a particular time stamp can be associated with a pixel of a SEM image of the sample surface.

[0385] ​A detector comprising a plurality of sensing elements can be operated in various counting modes. Circuitry can be connected to the sensing elements, and the circuitry can be configured to perform signal and data processing. The circuitry can be built-in in the sensing elements. The counting method can be based on signals output from the circuitry connected to the sensing elements. Multiple circuitry can be provided, where each of the multiple circuitry is connected to a respective one of the sensing elements in the array.

[0386] As an example of counting modes, there can be a first mode where the sensing elements can count at most one incident charged particle before a reset, without discriminating the energy of the incident charged particle. The reset can involve resetting the sensing elements or their associated circuitry. As used herein, the phrase "resetting the sensing elements" can refer to resetting the sensing elements or their associated circuitry. In a second mode, the sensing elements can count the number of incident charged particles before a reset, without discriminating the energy of the incident charged particle, until the counter is full. The numbers can include a predefined number. The predefined number can be one. After the counter is full, the sensing elements can stop counting charged particles. If another charged particle arrives after the counter is full but before the next reset, an overflow flag can be set. In a third mode, the sensing elements can count charged particles according to predefined energy levels. The energy ranges can include, for example, zero to a first predefined energy level, the first predefined energy level to another predefined energy level (and so on) and above the highest predefined energy level. Only at most one charged particle per range can be counted before the sensing elements are reset. In a fourth mode, the sensing elements can count charged particles according to predefined energy levels, and an overflow flag can be set for an energy range if more than a predetermined number of charged particles are received in that energy range. In some embodiments, the predetermined number for a corresponding energy range can be zero or one. The above modes and other modes will be described as follows.

[0387] Referring to Figure 28E and Figure 29A which show a first counting mode for operating a detector or detection system, consistent with embodiments of the present disclosure. The sensing elements can detect the energy of incident charged particles. Circuitry connected to the sensing elements can process the output of the sensing elements. As shown in Figure 29B , the sensing elements and circuitry can output detection signals based on time. Figure 29Ais a plot of the intensity of a detection signal in arbitrary units on the vertical axis plotted against time on the horizontal axis. A charged particle arrival event at a sensing element can occur at time points Tl, T2, and T3. The detector can have sensing elements and circuitry configured to detect charged particle arrival events. For example, a sensing element can be configured to generate a signal pulse in response to an incident charged particle arriving at the sensing element, which can be due to an electron-hole pair being generated in the sensing element, and can be fed to the circuitry. Upon determining that a charged particle has arrived at the sensing element, the circuitry can record a charged particle arrival event. Figure 29A It is shown that at each time point Tl, T2, and T3, a charged particle (such as an electron) arrival event can be recorded. Each event (e.g. Figure 29B The events shown) can correspond to a“count.” A count of charged particles can include an indication that a charged particle arrival event has occurred at a sensing element. The count can indicate that a number of charged particles have arrived at the sensing element, the number being a total number. Data associated with a charged particle arrival event can include a time at which the charged particle arrival event was recorded. The time can be a system time. In some embodiments, the time can be based on a global clock cycle of the detection system. In some embodiments, the time can be based on a local time running at each individual sensing element. The time can be related to a scan time of a primary electron beam of an electron beam tool. Data associated with a charged particle arrival event can also include an identifier indicating which sensing element in an array of sensing elements received the charged particle. Thus, a location on the surface of the detector can also be recorded.

[0388] In a first counting mode, in some embodiments, the detector can be configured to count only up to one charged particle arrival event in a sensing element before a reset. The detector can be configured to count one charged particle arrival event without discriminating the energy of the one charged particle arrival event. For example, the detector can count a charged particle arrival event in response to detecting an amount of energy in a sensing element. The reset of a sensing element can occur at the end of a detection frame. The reset can also include resetting circuitry connected to the sensing element.

[0389] In the first counting mode, in some embodiments, miscounting may occur if charged particle arrival events occur rapidly and consecutively. For example, a charged particle may arrive at the sensing element shortly after the previous particle and shortly before the sensing element or its associated circuitry is reset, so subsequent charged particles may not be counted. That is, when two particles have actually arrived at the sensing element, the number of charged particles counted may be determined to be one. This miscounting can be addressed by using, for example, a second counting mode. It should be noted that in some embodiments, the sensing element array may be designed and sized such that the subsequent arrival time of adjacent charged particles for the intended application (e.g., the available beam current range) is sufficiently long that the probability of such miscounting may be very low.

[0390] Figure 29B and Figure 30A The illustration depicts a second counting mode of an operational detector or detection system consistent with embodiments of this disclosure. For example... Figure 30B As shown, the sensing element and circuitry can output a detection signal based on time, and the arrival of a charged particle at the sensing element can occur at times T1, T2, T3, and T4. At time T2, a charged particle may arrive before the sensing element or its associated circuitry resets. However, the detector can set an overflow flag instead of ignoring subsequent charged particle arrival events. The overflow flag indicates that another charged particle has arrived at the sensing element. The increase in signal generated at the sensing element and circuitry due to subsequent charged particle arrival events can trigger the overflow flag.

[0391] An overflow can include a first type of overflow. The first type of overflow can be encountered when a charged particle arrives at a sensing element, and a subsequent charged particle arrives at the sensing element and the circuit is unable to give an appropriate response to the subsequent charged particle arrival event. This can be due to the circuit being in the middle of processing the first charged particle arrival event when the subsequent charged particle arrives. The circuit can be unable to process the output for the subsequent charged particle arrival event until the processing of the first charged particle arrival event is complete, and can only record the subsequent charged particle arrival event as an overflow event by, for example, setting an overflow flag. The first charged particle and the subsequent charged particle can be in the same or different energy level ranges. Processing the output of a charged particle arrival event can include determining the energy level of the charged particle. Thus, when a subsequent charged particle arrival event is determined to be an overflow event, the energy level of the subsequent charged particle can be ignored. For example, if a subsequent charged particle arrival event occurs before the processing of the first charged particle arrival event begins, an overflow flag can not be set, and two incoming charged particles that still arrive in close succession within a very short time period can still be identified and counted. For example, an event signal can be generated that has an amplitude corresponding to the amplitude of a single secondary charged particle arrival event plus a single backscatter charged particle arrival event, and thus it can be determined that a secondary particle and a backscatter particle have arrived.

[0392] An overflow can also include a second type of overflow. The second type of overflow can be encountered when a counter of a circuit connected to a sensing element is in an overflow state, such as in a second counting mode. In some embodiments, the second type of overflow can be encountered when a counter of a circuit connected to a sensing element for counting charged particles in a particular energy range is in an overflow state, such as in a fourth counting mode as should be discussed below. An overflow flag can be based on the second type of overflow in, for example, the second counting mode or the fourth counting mode. A detector can be configured to reset a sensing element and its associated circuit based on the second type of overflow.

[0393] In the second counting mode, in some embodiments, the detector may be configured to count the number of charged particles incident on the sensing elements in the sensing element array of the detector. The detector may be configured to count the number of charged particles as discrete charged particle arrival events. The detector may be configured to count the number of charged particles without discriminating the energy levels of the charged particle arrival events. For example, the detector may count the charged particle arrival events in response to detecting a certain amount of energy received by the sensing element. The detector may detect an overflow state in response to an increase in the energy received by the sensing element relative to the energy at the time of detecting the charged particle arrival event. Resetting of the sensing element may occur at the end of a detection frame. The detector may set an overflow flag, which may indicate that another charged particle has arrived during the same detection frame in which a charged particle arrival event has been detected.

[0394] In the first and second counting modes, the charged particles may be counted without considering the energy of the incident charged particles. Thus, regardless of whether the charged particles are backscattered electrons or secondary electrons, the charged particle arrival events may be simply recorded as arrival events.

[0395] Figure 30A and Figure 31A FIG. illustrates a third counting mode of operating a detector or a detection system in accordance with an embodiment of the present disclosure. As Figure 31B shown, the sensing element and the circuit may output a detection signal based on time, and charged particle arrival events at the sensing element may occur at time points T1, T2, T3, and T4. The charged particles may have different energy levels. Separate energy level thresholds may be set. Thresholds for the separate energy level thresholds may be set to distinguish noise from the detection signal. For example, the detector may be configured not to count charged particles with energy greater than zero but less than a first energy threshold E1. The detector may be configured to count charged particles with energy greater than or equal to the first energy threshold E1 and less than a second energy threshold E2. The detector may also be configured to count charged particles with energy greater than or equal to the second energy threshold E2 and less than a third energy threshold E3. The detector may also be configured to count charged particles with energy greater than or equal to the third energy threshold E3. It will be understood that the boundaries of the energy ranges may be modified, for example, the range may be defined as greater than the first energy threshold E1 and less than or equal to the second energy threshold E2. That is, in some embodiments, the range may be set as E1 < x ≤ E2. In some embodiments, the range may be set as E1 < x < E2. In some embodiments, the range may be set as E1 ≤ x ≤ E2, etc.

[0396] In the third counting mode, in some embodiments, the detector may be configured to: for a specific energy range before reset, count at most one charged particle arrival event in the sensing element only. The detector may: in response to detecting an energy level corresponding to the energy range in the sensing element, count the charged particle arrival events within that energy range. The detector may: in response to detecting an output signal of the sensing element corresponding to the energy range, count the charged particle arrival events within that energy range. Reset of the sensing element may occur at the end of the detection frame.

[0397] In some embodiments, the detector may be configured to count only charged particle arrival events within one or more energy ranges. For example, the circuit connected to the sensing element may be configured to count only charged particle arrival events corresponding to an energy range of E1 < x ≤ E2 and ignore events corresponding to other energy ranges.

[0398] Figure 31A It is shown that at each time point T1, T2, T3, and T4, electron arrival events can be recorded. Data associated with the electron arrival events may include the energy level and the time when the electron arrival event is recorded. Multiple recordings can be made for different energy levels. For example, the detector may count a first charged particle arrival event at a first time, where the first charged particle has a first energy, and may also count a second charged particle arrival event at or near the first time, where the second charged particle has a second energy within a different energy range compared to the first charged particle.

[0399] Figure 31B and Figure 32A [[ID=I3]]FIG. illustrates a fourth counting mode of operating a detector or a detection system consistent with embodiments of the present disclosure. As [[ID=I4]] Figure 32B [[ID=I5]]shown, the sensing element and the circuit may output a detection signal based on time, and charged particle arrival events at the sensing element may occur at time points T1, T2, T3, T4, T5, and T6. The charged particles may have different energy levels. [[ID=I6]] [[ID=I7]]

[0400] [[ID=I8]] Figure 32A [[ID=I9]]It is shown that at each time point T1, T3, and T4, electron arrival events can be recorded. Data associated with the electron arrival events may include the energy level and the time when the electron arrival event is recorded. Multiple recordings can be made for different energy levels. For example, the detector may count a first charged particle arrival event at a first time, where the first charged particle has a first energy, and may also count a second charged particle arrival event at or near the first time, where the second charged particle has a second energy within a different energy range compared to the first charged particle.

[0401] At time T2, a charged particle in the same energy range as the previous charged particle can arrive before the sensing element or its associated circuitry is reset. The sensing element can set an overflow flag in response to the charged particle arrival event occurring at time T2. The overflow flag can indicate that yet another charged particle has arrived at the sensing element. The overflow flag can be specific to an energy range. The overflow flag can be triggered when a charged particle is detected when a previous charged particle in the same energy range has already been detected before the sensing element or its associated circuitry is reset.

[0402] Similarly, at time T6, a charged particle in the same energy range as the charged particle that has already been received at the sensing element in the detection frame can arrive before the sensing element or its associated circuitry is reset. However, at time T5, a charged particle in a different energy range than the previous charged particle can arrive before the sensing element or its associated circuitry is reset. In this case, the detector can still record that a charged particle has arrived without triggering an overflow flag. As Figure 32B shown, the charged particle arrival event occurring at time T6 can trigger the overflow flag. The foregoing can include an example in which the detector is configured to count the number of charged particle arrival events in the first energy range and set the overflow flag when an overflow is encountered in the sensing element. In the foregoing example, the number of charged particles can be one. However, in some embodiments, the number can be zero or can be greater than one. Further, the detector can be configured to count the same or different number of charged particles in different energy ranges.

[0403] Using different energy level thresholds can be helpful in distinguishing between different types of charged particles. For example, in some embodiments, it can be useful to distinguish between backscattered electrons and secondary electrons from one another.

[0404] The comparison of the incident charged particle energy can be based on a reference value. Circuitry including, for example, a comparator can be provided. The comparator can compare the output signal from the sensing element to a reference provided in the circuitry. As one example, in cases where the sensing element is connected to an element that converts the output current signal to a voltage corresponding to the energy of the incident charged particle, the comparator can compare the detected voltage to a reference voltage. The comparator can determine whether the voltage signal is above or below the reference voltage without storing the detected voltage value.

[0405] In some embodiments, the data associated with the charged particle arrival event can include a recorded energy level corresponding to the incident charged particle. For example, the sensing element can be connected to an analog-to-digital converter that can convert the output signal of the sensing element to a value that can be stored. The energy of the incident charged particle can be directly measured and recorded.

[0406] In the fourth counting mode, in some embodiments, the detector can be configured to count the number of charged particles incident on the sensing elements in the array of sensing elements in the detector. The detector can be configured to count the number of charged particles as discrete charged particle arrival events. The detector can be configured to count the number of charged particles in a particular energy range before a reset. The detector can count a charged particle arrival event in the energy range in response to detecting an energy level corresponding to the energy range received by a sensing element. The detector can count a charged particle arrival event in the energy range in response to detecting an output signal from a sensing element corresponding to the energy range. The detector can detect an overflow condition in response to an increase in energy received by a sensing element relative to the energy at which a charged particle arrival event in a particular energy range is detected. The reset of a sensing element can occur at the end of a detection frame. The reset of a sensing element can occur with the reset of its associated circuitry. The detector can set an overflow flag, which can indicate that another charged particle has arrived during the same detection frame as a detection frame in which a charged particle arrival event in a particular energy range has been detected.

[0407] The detector can include an array of sensing elements and circuitry, and can operate in various ways consistent with the modes discussed above. The circuitry can include multiple circuits, each of which can be connected to a respective sensing element in the array. All of the sensing elements in the array can be operated in the same mode. In some embodiments, the sensing elements in the array can be operated in different modes simultaneously. Data associated with a charged particle arrival event can include an indicator of which operating mode was used.

[0408] In some embodiments, a timestamp can be recorded when each incident charged particle is counted by a sensing element. If only a maximum of one charged particle is counted without discriminating the energy of the particle, the timestamp can indicate the time at which the first charged particle arrived before the next reset. If only a maximum of one charged particle is counted in each predetermined energy range, the timestamp can indicate the time at which the first charged particle in its associated energy range arrived before the next reset. An overflow flag can also be set with its corresponding timestamp.

[0409] The reset of a sensing element can be performed in a variety of ways. For example, in a first reset mode, all of the sensing elements in the array can be reset simultaneously, with a fixed or varying period. In a second reset mode, the sensing elements can be reset at different points in time based on a predetermined region. In a third reset mode, the sensing elements can be individually reset at different points in time, with a fixed or varying period. In a fourth reset mode, one or more of the above modes can be combined. These modes will be discussed in more detail below.

[0410] The first reset mode can be based on a common reset. Some or all of the sensing elements in the sensing element array can be reset at once. For example, all sensing elements in the sensing element array can be reset simultaneously. The reset rate can be determined by a period. This period can be predetermined. In some embodiments, the period can be a fixed period. In some embodiments, the period can be varied. For example, as referenced above. Figure 32B The frames discussed may have different lengths. The frame period can be set to suit the signal strength of the detected charged particle beam. In some embodiments, the period may include a mixture of a fixed period and a variable period.

[0411] In the first reset mode, in some embodiments, data from each sensing element can be stored before each reset. The data can be stored immediately before the reset. Therefore, snapshots of the surface of the sensing element array can be acquired at intervals. The intervals can be fixed or variable. In this way, the detector can output data at a detection frame rate over time. The detection frame rate can be based on conditions that might be required for a specific SEM imaging frame rate.

[0412] In the first reset mode, in some embodiments, the reset may occur based on predetermined conditions. Predetermined conditions may include criteria, such as a condition that at least A% of the sensing elements in the sensing element array receive at least one charged particle. Another example of predetermined conditions may include a condition that at least A% of the sensing elements in the sensing element array encounters an overflow. A timestamp may be recorded each time a reset occurs. The timestamp can be used to mark the corresponding detection frame.

[0413] Figure 20C and Figure 33A The illustration shows a detection signal according to a first reset mode, consistent with an embodiment of this disclosure. Figure 33B The exemplary embodiments shown are similar, in Figure 29A In this system, the arrival of charged particles at the sensing element can occur at time points T1, T2, and T3. The sensing element can detect these arrival events at these time points. Figure 33A Multiple frames F1, F2, F3, and F4 are also shown at a constant frame rate. These frames may correspond to a reset of the sensing element. In some embodiments, the reset may be initiated by the controller. At the end of each frame, the sensing element and its associated circuitry may be reset. Therefore, the reset can occur at time point T. F1 T F2 T F3 and T F4 This occurs. The data from the sensing element can be stored at time point T. F1 T F2 T F3 and T F4 Or occurred before. Although Figure 33AThe output of only one sensing element can be shown, but it is to be understood that the resetting according to the first reset mode can occur simultaneously for all sensing elements in the array.

[0414] Referring now to the second reset mode, the second reset mode can be based on a common resetting of a portion of the sensing elements in the array. Sensing elements in different regions of the array can be reset at different points in time. Thus, not all sensing elements on the array can be reset at the same time. The regions can be predetermined. For example, the regions can be determined based on quadrants of the detector. In some embodiments, for example in case a multi-beam apparatus is used, the regions can correspond to the portions of the detector where the secondary charged particle beams are expected to land. In some embodiments, these regions can correspond to the sub-regions, such as 246, 248 and 250 of the electron detection device 244 (see Fig. 2). In some embodiments, the regions can be determined in real-time. The regions can also change during operation of the detector. Figure 33A ) In some embodiments, the regions can be determined in real-time. The regions can also change during operation of the detector.

[0415] In the second reset mode, the sensing elements in one region of the array of sensing elements can be reset at a time. The rate of resetting can be determined by a period. The period can be a fixed period, or the period can vary, for example as discussed above. Between different regions, the resetting can be done at different points in time.

[0416] In the second reset mode, in some embodiments, data from each sensing element in a region can be stored prior to each reset. The data can be associated with other sensing elements in the same region. The data can be stored immediately prior to the reset. Thus, a snapshot of the surface of a particular region of the array of sensing elements can be taken at intervals. The intervals can be fixed or varying. In this way, the detector can output data at a detection frame rate based on the different regions in time. The detection frame rate can be based on conditions that can be required for a particular SEM imaging frame rate.

[0417] In the second reset mode, in some embodiments, the resetting can occur based on predetermined conditions. The predetermined conditions can include criteria, such as a condition that at least A% of the sensing elements in a region of the array of sensing elements receive at least one charged particle. Another example of a predetermined condition can include a condition that at least A% of the sensing elements in a region of the array of sensing elements encounter an overflow. A time stamp can be recorded each time a reset occurs. The time stamp can be used to mark a corresponding detection frame from a particular region. In some embodiments, the number of regions on the detector surface can be related to the number of beams to be detected. In some embodiments, the number of regions can be independent of the number of beams to be detected. The size and shape of the regions can be the same or different from each other.

[0418] Although Figure 2A and Figure 33AThe detection signal of a sensing element according to the first reset mode can be illustrated, but it is to be understood that a plurality of sensing elements associated with one region can be reset in a similar manner according to the second reset mode. Thus, although only one sensing element output can be shown in Figure 33B and Figure 33A , other sensing element outputs can be represented in a similar manner.

[0419] Reference is now made to a third reset mode, which can be based on individual reset of sensing elements in the array. Each sensing element in the array can be reset at a different point in time. The sensing elements can be reset at fixed or varying periods, e.g. as discussed above.

[0420] In the third reset mode, in some embodiments, data from each sensing element can be stored prior to each reset. The data can be stored immediately prior to the reset. Thus, a snapshot of the surface of a particular sensing element of the array of sensing elements can be taken at intervals. The intervals can be fixed or varying.

[0421] The reset of a sensing element in the third reset mode can be based on a predetermined condition. The predetermined condition can include, for example, that a particular sensing element receives at least one charged particle. Another example of a predetermined condition can include that a particular sensing element encounters a condition of overflow. A timestamp can be recorded each time a reset occurs.

[0422] Figure 33B and Figure 34A An example of resetting a sensing element and its associated circuit in response to a charged particle arrival event is illustrated. The reset can occur at a predetermined time after the charged particle arrival event. In some embodiments, the reset can be initiated directly after the charged particle arrival event occurs or is detected. The charged particle arrival event on the sensing element can initiate a feedback loop in which a counter can be increased and a reset operation can be started with, for example, Figure 34B and Figure 33A In contrast, the reset of a sensing element and its associated circuit according to Figure 33B Figure 34A and FIG. 34B may not take into account a predetermined frame.

[0423] FIG. 35A and FIG. 35B An example of a reset that can be made by a passive method is illustrated. As FIG. 35B shown, the detection signal can naturally decay in response to the charged particle arrival event without an active reset operation. The gradual decay of the signal can include a long tail behavior.

[0424] In some embodiments, the reset of the sensing element or its associated circuitry can be passive or active. The reset of the counter detector or the sensing element itself can include quenching. In some embodiments, a quenching circuit can be provided. The quenching circuit can control the quenching of the respective sensing element. The reset in the detector can include resetting another circuit connected to each sensing element in the array, such as a counting circuit. The counting circuit can include an analog portion, such as a front-end electronics. The reset of the analog portion of the circuit can be passive or active. The counting circuit can also include a digital portion, such as a pulse height analyzer and a scaler. In some embodiments, the reset of the digital portion of the circuit can be active. Active reset of the circuit or sensing element can be advantageous for achieving high speed.

[0425] In some embodiments, electronic counting can be enabled, where it is not necessary to bias the diode of the sensing element into an avalanche mode or even a so-called "Geiger counting" mode to achieve gain. Gain can be achieved due to the kinetic energy of the incoming electron. Providing gain due to the kinetic energy of the electron can help maintain high speed of the detector. Additionally, in some embodiments, the diode can be biased into an avalanche mode, where the diode can have high speed. In the avalanche mode, long tails in the signal can be reduced or avoided. Due to the avalanche effect, the speed can even be higher than the diode under normal biasing conditions. As a result, the detector with the sensing element can be fast enough so that there are no long tails in the signal from the sensing element. In some embodiments, a reset mechanism can be provided for the sensing element as well as the associated signal conditioning and processing circuitry to return the detector to an initial state after each electron arrival event.

[0426] Referring now to a fourth reset mode, the fourth reset mode can use a combination of the reset modes discussed above. For example, the fourth reset mode can include resetting the sensing element according to a combination of the second reset mode and the third reset mode.

[0427] In some embodiments, the reset of the sensing element can occur uniformly among all sensing elements in the array. The reset interval of a single sensing element can vary from frame to frame. When all sensing elements in the array are reset uniformly, the same reset interval can be used for all sensing elements. However, in some embodiments, the reset interval can be different among the sensing elements.

[0428] The first through fourth reset modes can be used with any of the first through fourth modes of operating the detector discussed above. A variety of combinations can be employed.

[0429] The detector can be configured such that, over the course of time, data related to the number of charged particles counted at the sensing elements is continually generated. In some embodiments, data can be generated from the sensing elements with a time stamp at the time of a charged particle arrival event. Further, in some embodiments, data can be generated from the sensing elements with a time stamp at the time of a sensing element reset. The time stamps of these events can differ from one another. Based on the information derived from the sensing elements, including information about the location of each sensing element, a moving picture of the incoming charged particles received by the sensing elements in the array can be generated. The raw data about the charged particle arrival events can be converted to a desired format with a specified frame rate and pixel resolution. Information about the intensity of the charged particle beam received on the detector can be derived.

[0430] Reference is now made to FIG. 36 which illustrates the impact of a charged particle arrival event on a detection surface. In FIG. 36 the detector can include a plurality of sensing elements, including sensing element 36a, sensing element 36b, sensing element 36c, and sensing element 36d. A charged particle can impact (e.g., hit) the detector surface in a region near the boundary between sensing element 36a and sensing element 36c. When an electron hits the detector, it can create a charge in the volume of the detector. This volume can span two or more sensing elements. For example, an ionized region 37 can be generated that can enter multiple sensing elements. The charge developed in the multiple sensing elements can cause the electron to be counted multiple times or not at all.

[0431] In some embodiments, the miscounting of electrons arriving at the detection surface of the detector can also be reduced. Some miscounting can be based on electron arrival events that occur near the boundaries between sensing elements. When electrons from the electron optical column of the SEM arrive at the sensing surface of the detector, the location at which the electrons enter the detection apparatus can be randomly distributed. Due to the ionization process following each electron entering the detector, a temporary ionized region can be created within the detector. The temporary ionized region can expand outside the depletion region of the sensing elements. Each ionized region can have a volume related to the energy of each incoming electron and the material forming the detector. The randomness of the location at which each electron enters the detection apparatus can cause the temporary ionized region induced by each incoming electron to cross the boundary of an adjacent sensing element. As a result, multiple sensing elements in the detector can have an output signal corresponding to a particular incoming electron. In an electron counting apparatus based on each sensing element, this can result in miscounting.

[0432] To address the above issues, some embodiments can employ the following. The sensing elements can be configured to have a predetermined size and shape. The array of sensing elements can include sensing elements arranged in a pattern such as a grid. The size of each sensing element in the detector can be chosen in such a way that in any direction, the size of the sensing element is not smaller than the maximum penetration depth of an incoming electron. The sensing elements can be configured such that the dimensions (e.g. length, width, height) of the sensing elements are not smaller than the maximum penetration depth. The maximum penetration depth can be based on the material used to form the detector. For example, the penetration of an electron can be deeper in one material than in another material. In this way, each incoming electron can only hit four sensing elements at a time. This can help to further simplify the signal post-processing for miscount correction.

[0433] Furthermore, in each sensing element, if the strength of the signal due to an incoming electron is higher than a predefined reference value, e.g. a threshold value TH A , then a time stamp can be left for the event. The threshold value TH A may correspond to a value that is a predetermined proportion of the average event pulse signal, or some other parameter such as a measure of the energy of the electron that caused the event. For example, the threshold value TH A may be set to a value equal to 20% of the average amplitude of the event signal pulse corresponding to the case of one incoming secondary electron hitting one sensing element. The case of one incoming secondary electron hitting one sensing element can correspond to a temporary ionized region formed within the volume of a single sensing element. The threshold value TH A may be based on parameters of secondary electrons that can have a relatively consistent energy level. Furthermore, under fixed SEM imaging settings, the energy variation of secondary electrons can be less than, for example, the energy variation of backscattered electrons.

[0434] Events with a signal strength exceeding the threshold value TH A may indicate that the sensing element has been at least partially hit by an incoming electron. Then, information from any two adjacent sensing elements can be analyzed. If both of the two adjacent sensing elements encounter an electron hit event at the same time stamp, the signals from the two sensing elements can be added in an analog manner and the result can be sent to one of the two sensing elements. The signal stored in the other one of the two sensing elements can be cleared (e.g. deleted).

[0435] For example, the signal caused by an incoming electron from each sensing element that has been hit (e.g. impacted) can be presented in the form of a voltage on a capacitor. In each sensing element, the capacitor used for generating and storing the signal can have the same capacitance. Then, during the analog signal processing (e.g. analog signal addition), the charge from the capacitor in one of the two adjacent sensing elements can be transferred to the capacitor in the other one of the two adjacent sensing elements. In this way, the signal addition can be performed and then the signal in the first one of the two adjacent sensing elements can be deleted.

[0436] In some configurations, in each sensing element, only one capacitor and one memory are used for time stamping. Whenever there are two adjacent sensing elements that are hit by an incoming electron at the same point in time, the signal processing procedure can start. This procedure can be asynchronous to the system clock.

[0437] In other configurations, in each sensing element, there are more than one capacitor used for signal generation and storage. In addition, each capacitor can have its corresponding time stamp memory. These capacitors and their corresponding time stamp memories can be used in a way that each of them is used only for generating and storing the signal of a single event of an electron hitting the sensing element. For example, during each event, there can be one or more than one electron hitting the sensing element in a sufficiently short period of time such that the circuit in the sensing element considers this period as one point in time. In other words, from the circuit’s perspective, these electrons hit the sensing element at the same time. Then, if a predefined number M or percentage B% of capacitors are used in the sensing element and there is at least one event quantity involving at least two adjacent sensing elements, the signal processing procedure for counting correction can start. The procedure in that particular sensing element can be stopped until all the electron hit events involving at least two sensing elements are processed. During the signal processing procedure, the sensing elements involved in the procedure can still count new events of incoming electrons until all the capacitors and time stamp memories in the sensing element are used. Then, if there is a new incident electron hitting the same sensing element, an overflow can be recorded. The reset of the sensing element can be performed after the signal processing procedure mentioned above is completed and the data stored in or associated with the sensing element is issued.

[0438] In a signal processing procedure such as consistent with the above, each time a boundary hit event occurs, the combined signal can be stored in one of the two adjacent sensing elements that are hit by the same incoming electron or electrons, where the signal level from this element is higher. In addition, the signal in the other one of the two sensing elements can be deleted. Then, this capacitor and its corresponding time stamp memory will be immediately available for the next electron hit event of the sensing element.

[0439] A detector can be provided in which the rate of false counts due to an incoming electron hitting more than one sensing element is reduced. Such an event can be referred to as a border hit event. A border hit event can occur when the ionization region 37 is generated in more than one sensing element, for example as shown in FIG. 36

[0440] In some embodiments, each type of electron (e.g., secondary electrons, backscattered electrons, etc.) has a characteristic energy range and creates a characteristic amount of charge when hitting the detector. When multiple adjacent sensing elements detect charge at substantially the same time, the charge can be compared to these characteristic energies to determine whether they correspond to multiple electron hits or a single hit that generated charge in multiple sensing elements. A method for reducing false counts in an electron counting detector can be provided. The method can include detecting charge in two or more adjacent sensing elements at substantially the same time; comparing the charge detected in each of the adjacent sensing elements to one or more reference values; and based on the comparison, determining whether the charge was caused by one electron hitting one of the sensing elements or multiple electrons hitting the two or more sensing elements.

[0441] A charged particle beam device can be configured to image a sample using a charged particle beam. A detector can be configured to determine a count of charged particles that reach the detector. The count of charged particles can be based on individual counts of charged particles received by individual sensing elements in an array. Based on the count, an image of the sample can be reconstructed. Producing an image such as a SEM image can include summing the counts of the plurality of sensing elements and correlating the count to a particular time corresponding to a location on the surface of the sample being scanned. The summed count can be used to determine a grayscale of a pixel image. In some embodiments, other degrees of freedom in the information display can be used. For example, the information included in the count can include a time stamp, a location (e.g., a location of the sensing element on the array of sensing elements), and a number of charged particles counted. The counts of the sensing elements can be associated with the sample surface location by correlation with the scan time. The counts of individual sensing elements can be grouped together based on, for example, their time stamps. Sensing element data whose time stamps are the same in value or close to each other can be grouped together. The grouped sensing element data can be associated with the charged particle beam incident on the detector. The number of charged particles counted can be taken as a sum from a plurality of sensing elements that received a charged particle within a time window based on their time stamps.

[0442] ​In some embodiments, the sensing elements can be grouped without pre-processing. Instead, the sensing elements can be grouped with post-processing methods. Sensing elements that determined to receive a charged particle at a certain time based on their time stamp can be grouped together and can be associated with a beam spot. Thus, it can not be necessary to determine the beam boundaries, such as boundary lines 350 and 360 (see FIG. 3F ). Sensing elements that are not necessarily adjacent to each other can be grouped, or further the information from these sensing elements can be grouped, based on their time stamp, position, scanning action of the primary beam, or properties of the SEM system. Thus, for example, referring to FIGS. 28A-28E , the detector can determine that all sensing elements that received a charged particle at a particular scan time (e.g., sensing elements 2102 and 2103 at T=T2) can be associated with the same charged particle beam incident to a particular point on the detector.

[0443] According to some embodiments, real-time high linearity (high fidelity) or high dynamic range charged particle beam detection can be implemented. The charged particle beam detection system can be able to track slow and fast movements of the projection pattern of the secondary charged particle beam incident to the detector. In some applications, the SEM system can be made more robust and more fault tolerant. As a result, system downtime can be reduced.

[0444] In some embodiments, the anti-deflection system in the projection system can be omitted, which can help to simplify the system. This can also improve reliability and reduce system downtime.

[0445] Various results can be obtained based on the same set of acquired raw data, depending on the specific application requirements. The results can be used for post-processing, instead of pre-processing. In some embodiments, pre-processing can include sensing element grouping. The data can also be customized to optimize for specific purposes. For example, there can be a trade-off between optimizing detection parameters (such as cross-talk) and secondary charged particle collection efficiency. However, when using post-processing to adjust the data, the trade-off can be adjusted without losing information. This can provide more flexibility for the application layer, and can reduce the risk of having to redo certain operations just to obtain data with different detection parameter settings.

[0446] In some embodiments, the detection system can include analog, mixed-signal, and digital circuitry. In comparative embodiments, the percentage of analog circuitry portion can be relatively high, which can make device implementation more difficult and can reduce compatibility of the detector design with the most advanced semiconductor process nodes. Some embodiments of the present disclosure can provide a detection system that uses or consists only of a relatively high percentage of mixed-signal and digital circuitry, which can reduce the difficulties associated with analog circuit design and fabrication in ASICs (application-specific integrated circuits). Some embodiments can help enable high compatibility with advanced semiconductor nodes. This can help reduce overall power consumption.

[0447] Embodiments can also be described using the following clauses:

[0448] 1. A method comprising:

[0449] determining a number of charged particles incident on a detector within a first frame of a plurality of frames; and

[0450] determining a period of the first frame based on a first criterion, the first criterion comprising a condition that a first number of the charged particles is incident on each of a first number of sensing elements of the detector.

[0451] 2. The method of clause 1, further comprising:

[0452] determining the period of the first frame based on the first criterion and a second criterion, the second criterion comprising a condition that more than one of the charged particles is incident on no more than a second number of the first number of sensing elements.

[0453] 3. The method of clause 1 or 2, further comprising:

[0454] determining a frame rate of the plurality of frames based on the first criterion.

[0455] 4. The method of clause 2, further comprising:

[0456] determining the frame rate of the plurality of frames based on the first criterion and the second criterion.

[0457] 5. The method of any one of clauses 1 to 4, further comprising:

[0458] determining a statistical quantity of the amount of incident charged particles at a plurality of energy levels within the first frame.

[0459] 6. The method of clause 5, wherein the statistical quantity comprises a total amount of incident charged particles in the first number of sensing elements at each of a corresponding number of energy levels, the incident charged particles corresponding to the charged particles incident on each of the first number of sensing elements.

[0460] 7. The method of any one of clauses 1 to 6, further comprising:

[0461] determining a third number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the first threshold and less than the second threshold.

[0462] 8. The method of clause 7, further comprising:

[0463] determining a fourth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the second threshold and less than the third threshold.

[0464] 9. The method of clause 8, further comprising:

[0465] determining a fifth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the third threshold.

[0466] 10. The method of any of clauses 1 to 9, further comprising:

[0467] generating a pixel of a grayscale image based on the at least one frame.

[0468] 11. The method of any of clauses 1 to 10, further comprising:

[0469] generating a pixel of a color image based on the at least one frame.

[0470] 12. The method of any of clauses 1 to 11, further comprising:

[0471] generating a beam of charged particles; and

[0472] scanning the beam of charged particles across a sample,

[0473] wherein the detector is configured to receive charged particles projected from the sample.

[0474] 13. A non-transitory computer-readable medium comprising a set of instructions executable by one or more processors of a controller to cause the controller to perform a method comprising:

[0475] determining a number of charged particles incident on the detector within a first frame of a plurality of frames; and

[0476] determining a period of the first frame based on a first criterion, the first criterion comprising a condition of the first number of charged particles incident on each sensing element of the sensing elements of the detector.

[0477] 14. The non-transitory computer-readable medium of clause 13, wherein the set of instructions executable by the at least one processor of the controller causes the controller to further perform:

[0478] determine a period of the first frame based on the first criterion and a second criterion, the second criterion comprising a condition that more than a second number of the charged particles incident on no more than a second number of the sensing elements of the first number of sensing elements.

[0479] 15. The non-transitory computer-readable medium of clause 13 or 14, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a frame rate of the plurality of frames based on the first criterion.

[0480] 16. The non-transitory computer-readable medium of clause 14, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a frame rate of the plurality of frames based on the first criterion and a second criterion.

[0481] 17. The non-transitory computer-readable medium of any one of clauses 13 to 16, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a statistical quantity of an amount of incident charged particles at a plurality of energy levels within the first frame.

[0482] 18. The non-transitory computer-readable medium of clause 17, wherein the statistical quantity comprises a total amount of incident charged particles at each of the respective energy levels, the incident charged particles corresponding to charged particles incident on each of the first number of sensing elements.

[0483] 19. The non-transitory computer-readable medium of any one of clauses 13 to 18, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a third number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to a first threshold value and less than a second threshold value.

[0484] 20. The non-transitory computer-readable medium of clause 19, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a fourth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the second threshold value and less than a third threshold value.

[0485] 21. The non-transitory computer-readable medium of clause 20, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform determining a fifth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the third threshold value.

[0486] 22. The non-transitory computer-readable medium of any one of clauses 13 to 21, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform generating a pixel of a grayscale image based on the at least one frame.

[0487] 23. The non-transitory computer readable medium of any of clauses 13-22, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform generating pixels of a color image based on the at least one frame.

[0488] 24. The non-transitory computer readable medium of any of clauses 13-23, wherein the set of instructions executable by the at least one processor of the controller cause the controller to further perform:

[0489] causing a charged particle source to generate a charged particle beam; and

[0490] causing a deflector to effectuate scanning of the charged particle beam across a sample,

[0491] wherein the detector is configured to receive charged particles projected from the sample.

[0492] 25. A charged particle beam device comprising a detector and a controller having at least one processor and a non-transitory computer readable medium comprising instructions that, when executed by the processor, cause the device to:

[0493] determine a number of charged particles incident on the detector within a first frame of a plurality of frames; and

[0494] determine a period of the first frame based on a first criterion, the first criterion comprising a condition of a first number of the charged particles being incident on each of a number of sensing elements of the detector.

[0495] 26. The device of clause 25, wherein the instructions further cause the device to:

[0496] determine the period of the first frame based on the first criterion and a second criterion, the second criterion comprising a condition of more than one of the charged particles being incident on no more than a second number of the sensing elements of the first number of sensing elements.

[0497] 27. The device of clause 25 or 26, wherein the instructions further cause the device to:

[0498] determine a frame rate of the plurality of frames based on the first criterion.

[0499] 28. The device of clause 26, wherein the instructions further cause the device to:

[0500] determine the frame rate of the plurality of frames based on the first criterion and the second criterion.

[0501] 29. The device of any of clauses 25-28, wherein the instructions further cause the device to:

[0502] determine a statistical quantity of an amount of the incident charged particles at a plurality of energy levels within the first frame.

[0503] 30. The apparatus of clause 29, wherein the statistical quantity comprises a total number of incident charged particles in the first number of sensing elements at each of the respective energy levels, the incident charged particles corresponding to charged particles incident on each of the first number of sensing elements.

[0504] 31. The apparatus of any one of clauses 25 to 30, wherein the instructions further cause the apparatus to:

[0505] determine a third number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the first threshold value and less than the second threshold value.

[0506] 32. The apparatus of clause 31, wherein the instructions further cause the apparatus to:

[0507] determine a fourth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the second threshold value and less than the third threshold value.

[0508] 33. The apparatus of clause 32, wherein the instructions further cause the apparatus to:

[0509] determine a fifth number of charged particles incident on the sensing elements of the detector at an energy level greater than or equal to the third threshold value.

[0510] 34. The apparatus of any one of clauses 25 to 33, wherein the instructions further cause the apparatus to:

[0511] ...

Claims

1. A detector for a charged particle device, comprising: a plurality of sensor elements, each of the sensor elements having a corresponding circuit, the corresponding circuit including an event detector circuit configured to detect an electron event based on a current pulse generated by the corresponding sensor element in response to an electron hitting the corresponding sensor element, and a pixel count buffer circuit configured to count electron events, each of the plurality of sensor elements including a diode configured to operate in a linear mode; and a second circuit configured to sum a number of the electron events counted by the pixel count buffer circuit of each of the plurality of sensor elements during a predetermined time period.

2. The detector of claim 1, wherein the diode is a PIN diode or an avalanche diode operating in the linear mode.

3. The detector of claim 2, wherein the avalanche diode operating in the linear mode is a low gain avalanche diode.

4. The detector of claim 1, wherein the diode configured to operate in a linear mode includes the diode configured to avoid operating in a Geiger mode or a photon counting mode.

5. The detector of claim 1, wherein the diode configured to operate in a linear mode includes a diode configured to generate a current pulse having an amplitude distribution that is substantially proportional to a kinetic energy of a hitting electron at a time the hitting electron enters the diode.

6. The detector of claim 5, wherein the corresponding circuit includes circuitry to determine a first characteristic of the current pulse and determine a second characteristic of an incoming electron based on the first characteristic.

7. The detector of claim 6, wherein the first characteristic is an amplitude of the current pulse and the second characteristic is a determination of whether the electron is any one of a secondary electron, a backscattered electron, or an electron from a noise source.

8. The detector of claim 7, wherein the determination that the electron is from a noise source is based on the amplitude of the current pulse being below a first threshold, wherein the determination that the electron is a secondary electron is based on the amplitude of the current pulse being above the first threshold and below a second threshold, and wherein the determination that the electron is a backscattered electron is based on the amplitude of the current pulse being above the second threshold.

9. The detector of claim 6, wherein the second characteristic is a determination that the electron event corresponding to the electron is one of a plurality of electron events caused by a plurality of electrons hitting the corresponding sensor element during the predetermined time period.

10. The detector of claim 9, wherein the current pulse generated by the corresponding sensor element is in response to the plurality of electrons hitting the corresponding sensor element. ​ 11. A method of operating a detector for a charged particle device, comprising: The detector includes a plurality of sensor elements, each of the sensor elements having a corresponding circuit including an event detector circuit configured to detect an electron event based on a current pulse generated by the corresponding sensor element in response to an electron hitting the corresponding sensor element, and a pixel count buffer circuit configured to count electron events, each of the plurality of sensor elements including a diode configured to operate in a linear mode; and a second circuit configured to sum a plurality of sensor element counts of the plurality of sensor elements during a predetermined time period, the method comprising: counting a number of charged particles incident on a sensor element of the plurality of sensor elements, and summing each sensor element count of the plurality of sensor elements; wherein an area of the sensor element of the plurality of sensor elements is configured to receive no more than a predetermined number of charged particles per sampling period.

Images