Real-time detection and correction of system response

Abstract

This application relates to real-time detection and correction of system response. Embodiments of the invention can include methods, systems, and apparatuses for correcting a response function of an electron beam tool. The correction can include modulating an electron beam parameter having a frequency; emitting an electron beam toward a sample based on the electron beam parameter, thereby scattering electrons, wherein the electron beam is described by a source wave function having a source phase and a landing angle; detecting a portion of the scattered electrons at an electron detector, thereby generating electron data including an electron wave function having an electron phase and an electron landing angle; determining, using a processor, a phase delay between the source phase and the electron phase, thereby generating a latency; and using the processor, using the latency and a difference between the source wave function and the electron wave function to correct the response function of the electron beam tool.

Description

[0001] Related Application

[0002] This application is a divisional application. The parent application of this divisional application is Chinese Invention Patent Application Case No. 201980079627.1 entitled “Real-time Detection and Correction of System Response” filed on December 12, 2019.

[0003] Cross-Reference to Related Applications

[0004] This application claims priority to provisional patent application U.S. Application No. 62 / 779,485 filed on December 14, 2018 and assigned to, provisional patent application U.S. Application No. 62 / 781,412 filed on December 18, 2018 and assigned to, and provisional patent application U.S. Application No. 62 / 785,164 filed on December 26, 2018 and assigned to, the disclosures of which are hereby incorporated by reference. TECHNICAL FIELD

[0005] The present invention relates generally to improvements in semiconductor wafer metrology. More particularly, the present invention relates generally to methods and systems for reducing systematic errors in electron beam measurements. BACKGROUND

[0006] Evolution of the semiconductor manufacturing industry places higher demands on yield management and, in particular, on metrology and inspection systems. Critical dimensions continue to shrink, but the industry needs to reduce the time for achieving high yield, high value products. Minimizing the total time from detecting a yield problem to solving the problem determines the return on investment for semiconductor manufacturers.

[0007] Measurement of overlay between different semiconductor layers is critical to ensuring functionality of integrated circuits (ICs). As the size of structures scales down, electron beam (EB) systems (e.g., EB inspection, re-inspection, and metrology systems) based on the principles of scanning electron microscopy become increasingly attractive due to their high resolution capabilities. However, EB systems can suffer from systematic error contributions, reducing the theoretically possible accuracy, and thus perturbing the measurement results.

[0008] One previous method relies on a single offline measurement of the internal latency and correction for the internal latency. Changes in these latencies are not covered by the measurement. Internal latencies can cause shifts or asymmetries in the detected image, thus causing error contributions to distance measurements. Changes in the latency on a short time scale can act as an additional noise source. Changes in the latency on a long time scale can exhibit effects similar to drift.

[0009] Detection of secondary electrons (SE) is beneficial because secondary electrons are generated in the sample with a small interaction volume close to the surface, thereby resulting in precise measurements with high spatial resolution. However, the SE signal is perturbed by other sources of electrons, primarily backscattered electrons (BSE) with a broad band energy distribution. Current bandpass energy filters can discriminate BSE with high energy. Low and medium energy BSE are typically passed through the bandpass filter, thereby adding additional background signal to the SE signal. Thus, the measured detector signal represents a combination of SE and spurious low to medium energy backscattered electron signals. In such prior art solutions, the spurious BSE contribution can cause image blurring, resulting in significant systematic errors.

[0010] Some previous methods rely on static pre-alignment of the beam axis relative to the optical axis by aligning the image symmetry of specialized target structures or dedicated sensors, as illustrated in FIGS. 1 and 2A-2C. However, beam alignment has a direct impact on distance or position measurements in different heights from the sample. In addition, the orientation of the sample can change slightly from load to load, requiring a new alignment of the beam. Keeping the old orientation results in tilt-induced errors.

[0011] Therefore, there is a need for improved methods and systems for reducing systematic errors of EB measurements. SUMMARY

[0012] Embodiments of the invention serve the following purposes, including: (1) reducing systematic errors of high resolution EB measurements by determining, tracking, and correcting for tool internal responses; (2) reducing systematic errors of high resolution EB measurements by improved methods of secondary electron (SE) detection - embodiments are not limited to overlay measurements and can be used for both semiconductor metrology and inspection; and (3) reducing systematic errors of EB measurements by measuring and aligning the beam orientation relative to the surface normal of the sample.

[0013] Embodiments can correct for response functions by modulating beam parameters, detector or filter parameters, or using asymmetric detection to correct for beam tilt.

[0014] Embodiments methods and systems can detect a response of a system due to perturbations of the system, determine a correction metric for the response, and apply the correction metric. Measurement devices to detect a response of a system can include metrology, inspection, and re-inspection electron beam tools. The response of the system due to perturbations can be defined as imparted by the system in nature and detected by the metrology, can be induced by purposeful perturbations of the beam parameters, or can be induced by external sources. The perturbations of the beam parameters can be periodic at a defined frequency. The perturbations can be intrinsic, for example due to beam tilt. The perturbations can be external, for example due to magnetic fields.

[0015] Embodiments can include a method comprising correcting a response function of an electron beam tool. Such a method can comprise modulating an electron beam parameter having a frequency; scattering electrons based on the electron beam parameter being emitted toward a sample, wherein the electron beam is described by a source wave function having a source phase and a landing angle; scattering electrons based on the electron beam parameter being emitted toward a sample, wherein the electron beam is described by a source wave function having a source phase and a landing angle; detecting a portion of the scattered electrons at an electron detector, thereby producing electron data comprising an electron wave function having an electron phase and an electron landing angle; determining, using a processor, a phase delay between the source phase and the electron phase, thereby producing a latency; and using the processor, using the latency and a difference between the source wave function and the electron wave function to correct the response function of the electron beam tool.

[0016] The scattered electrons can comprise backscattered electrons, and the method can further comprise detecting a portion of the backscattered electrons at a backscattered electron detector with the electron beam tool, thereby producing backscattered electron data.

[0017] Correcting the response function can comprise subtracting the backscattered electron data from the electron data.

[0018] Correcting the response function can comprise: extrapolating, using the processor, a spectral distribution of the backscattered electron data toward lower energies, thereby producing extrapolated backscattered electron data; and subtracting, using the processor, the extrapolated backscattered electron data from the electron data.

[0019] Correcting the response function can comprise: modeling, using the processor, an energy distribution function of the backscattered electron data, thereby producing a modeled energy distribution function; and calibrating, using the processor, the modeled energy distribution function using one or more measurements taken with the electron beam tool.

[0020] Correcting the response function can comprise determining a relative fraction of a measurement parameter to a sample property based on the backscattered electron data as a sum of the backscattered electron data and the electron data.

[0021] The relative fraction can be stored on an electron data storage unit.

[0022] Correcting the response function can comprise using a machine learning model, and the machine learning model can be trained using the stored relative fraction stored on the electron data storage unit.

[0023] The method can further include detecting a second portion of the backscattered electrons at an opposing electron detector with the electron beam tool, thereby generating opposing electron data; using the processor to compare the electron data to the opposing electron data to generate tilt data; and using the tilt data to change a tilt of the electron beam.

[0024] Embodiments can include a system comprising an electron beam tool and a controller in electronic communication with the electron beam tool.

[0025] The electron beam tool can include an electron beam emitter, a stage, and an electron detector. The electron beam emitter can be configured to emit electrons in an electron beam. The electron beam can be described by a source wave function having a source phase and a landing angle. The stage can be configured to hold a sample in a path of the electron beam. The electron detector can be configured to detect a portion of the electrons scattered by the electron beam impacting the sample, and thereby generate electron data including an electron wave function having an electron phase and an electron landing angle.

[0026] The controller can have a processor. The controller can be configured to instruct the electron beam emitter to modulate an electron beam parameter of the electron beam, the electron beam parameter having a frequency; determine a phase delay between the source phase and the electron phase, thereby generating a latency; and correct a response function of the electron beam tool using the latency and a difference between the source wave function and the electron wave function.

[0027] The scattered electrons can include backscattered electrons, and the system can further comprise a backscattered electron detector configured to detect a portion of backscattered electrons and thereby generate backscattered electron data.

[0028] The controller can be configured to correct the response function by subtracting the backscattered electron data from the electron data.

[0029] The controller can be configured to extrapolate a spectral distribution of the backscattered electron data toward lower energies, thereby generating extrapolated backscattered electron data; and subtract the extrapolated backscattered electron data from the electron data to correct the response function.

[0030] The controller can be configured to model an energy distribution function of the backscattered electron data, thereby generating a modeled energy distribution function; and calibrate the modeled energy distribution function using one or more measurements taken with the electron beam tool to correct the response function.

[0031] The controller can be configured to correct the response function by determining a relative fraction of a measurement parameter to a sample property as a sum of the backscattered electron data and the electron data based on the backscattered electron data.

[0032] The system can further comprise an electron data storage unit configured to store the relative fraction.

[0033] The controller can be configured to correct the response function by using a machine learning model. The machine learning model can be trained using the stored relative fraction stored on the electron data storage unit.

[0034] The system can further comprise a back electron detector configured to detect a second portion of the backscattered electrons so as to generate back electron data. The controller can be further configured to compare the electron data to the back electron data so as to generate tilt data; and use the tilt data to change an angle of the electron beam emitter, thereby changing a tilt of the electron beam.

[0035] Embodiments can include a non-transitory computer readable storage medium comprising one or more programs for execution on one or more computing devices to perform steps including correcting a response function of an electron beam tool. Correcting the response function can include modulating an electron beam parameter having a frequency; instructing the electron beam tool to emit an electron beam toward a sample based on the electron beam parameter, thereby scattering electrons, wherein the electron beam is described by a source wave function having a source phase and a landing angle; receiving, from an electron detector that detects a portion of the scattered electrons, electron data including an electron wave function having an electron phase and an electron landing angle; determining a phase delay between the source phase and the electron phase, thereby generating a latency; and correcting the response function of the electron beam tool using the latency and a difference between the source wave function and the electron wave function.

[0036] The steps can include training a machine learning algorithm using the electron beam parameter, the electron data, and the response function. BRIEF DESCRIPTION OF DRAWINGS

[0037] For a more complete understanding of the nature and objects of the application, reference should be made to the following detailed description taken in connection with the accompanying drawings in which:

[0038] FIG. 1 illustrates a beam correction method according to the prior art;

[0039] FIGS. 2A-2C illustrate a metrology correction method according to embodiments of the application;

[0040] Figure 3illustrating a metrology correction method according to an embodiment of the application;

[0041] Figure 4 illustrating a metrology correction method according to an embodiment of the application;

[0042] Figure 5 illustrating a metrology correction system according to an embodiment of the application;

[0043] Figure 6 illustrating a metrology correction system according to an embodiment of the application;

[0044] Figures 7A-7C illustrating an embodiment dark field metrology correction;

[0045] Figure 8 illustrating a system embodiment of the application; and

[0046] Figure 9 illustrating another system embodiment of the application. DETAILED DESCRIPTION

[0047] Although the claimed subject matter will be described in terms of some embodiments, other embodiments encompass modifications and / or improvements to the examples described herein without departing from the scope of the claimed subject matter. Various structural, logical, process steps, and electronic changes can be made without departing from the scope of the claimed subject matter. Accordingly, the scope of the claimed subject matter is only to be defined by the appended claims.

[0048] Embodiments disclosed herein include methods, systems, and apparatuses for correcting a response function of an electron beam inspection tool. Such embodiments can provide real-time detection and correction of potential latency and reduce systematic errors of measurements. As a result, measurements can have higher accuracy.

[0049] In an example, Figure 3A first method of depicting a response function (warped sinusoid) for correcting an electron beam inspection tool is described. Method 1 can determine and correct for inherent static or dynamic responses, for example, latency, by modulating a source property with one or more defined frequencies at a defined amplitude. Method 1 can begin by generating a frequency 2 that can be modulated. A frequency generator can be used to generate the frequency 2. The frequency generated at 2 can be an input to generate a beam parameter 3 (e.g., electron beam current). The beam parameter 3 can be a parameter of a primary beam 4. The beam parameter 3 can be modulated. The primary beam 4 can be directed toward a sample 5, which can be a wafer or a die on a wafer. A portion of the primary beam 4 can impact the wafer 5 and cause a portion of the electrons near the surface of the wafer 5 to scatter, resulting in electrons that can include secondary and backscattered electrons. A portion of the electrons can be detected by a detector 6, and modified raw data 7 can be generated. The modulation of the electrons is due to the modulation of the beam parameter 3. This modulation can be analyzed with respect to an excitation modulation by a phase delay representing latency, for example. The modulation can be filtered from the detected image, and does not perturb the measurement. The modified raw data 7, which can represent the signal received at the detector 6, can be used in conjunction with the beam parameter 3 to determine the phase delay in order to give the latency 9. A filter 10 can be used to filter the modified raw data 7 using the latency 9 to generate a corrected response function 11.

[0050] The primary beam 4 can be turned on and off, or it can be moved to a buncher structure to modulate beam deflection. The buncher structure can be a beam buncher, for example, and can include a blocking panel or a metal foil.

[0051] Analyzing the modified raw data 7 can include comparing a sinusoid that controls the beam parameter 3 to the measured warped sinusoid.

[0052] In another example, a system can include an electron beam inspection tool. The electron beam inspection tool can include an electron beam emitter configured to emit electrons in an electron beam. The electron beam can be described by a source electron wave function having a source electron phase.

[0053] The electron beam inspection tool can further include a stage. The stage can be configured to hold a sample. The sample can be a wafer, for example. The wafer can contain one or more dies. The sample can be held in the path of the electron beam by the stage.

[0054] The electron beam inspection tool can further include a secondary electron detector, such as an Everhart-Thornely detector or a solid-state detector, and can incorporate a low-pass optical filter. The secondary electron detector can be configured to detect a portion of secondary electrons generated due to the electron beam impinging the sample. The secondary electron detector can then generate electron data representing a signal received at the detector 6 (primary, secondary, or backscattered) as an output after detecting the electrons. The electron wave function can include an electron phase.

[0055] The controller can be in electronic communication with the inspection tool. The controller can be configured to instruct the electron beam emitter, for example, to modulate the frequency of the electron beam parameter of the electron beam. The controller can determine a phase delay between the source electron phase and the electron phase. This can result in a latency. Using this latency, in conjunction with the difference between the source wave function and the electron wave function, the controller can correct the response function of the electron beam inspection tool. The way in which the response function is corrected can depend on which aspect of the response function needs to be corrected. In the case where beam tilt modulation is needed to correct the response function, beam position stitching can be performed. In the case where beam current modulation is needed to correct the response function, image intensity correction can be performed.

[0056] Another embodiment is a non-transitory computer-readable storage medium that can include one or more programs for executing steps including those of Method 1.

[0057] In other embodiments, the electron detectors in various inspection tools can be improved. Emitting an electron beam toward a sample can additionally cause a portion of the electrons in the sample to backscatter, thus generating backscattered electrons. The method can further include detecting a portion of the backscattered electrons at a backscattered electron detector, thus generating backscattered electron data, which can represent a signal received from the backscattered electrons.

[0058] In Figure 4In the depicted embodiment, the method 100 can be used to improve electron detectors in various inspection tools. Spurious backscattered electron contributions can cause image blurring, resulting in significant systematic errors. The method 100 can clean the electron data 101 from spurious backscattered electron contributions. The method can determine and correct for spurious backscattered electron data by using a high frequency modulation 103 to scan or cycle the bias voltage of an existing high pass filter 102 from near zero (high pass off) to tens or hundreds of volts (high pass on). Whether the high pass filter 102 is in high pass off or high pass on mode depends on the strength of the filter voltage. The detected data in high pass off mode corresponds to backscattered electron and secondary electron data 106, while high pass on represents the remaining backscattered electron portion 104. An optional spectral extrapolation 105 can be performed. The backscattered electron and secondary electron data 107 is corrected by the backscattered electron data to achieve backscattered electron corrected pure secondary electron data 108.

[0059] In this way, a secondary electron detector can be improved. The secondary electron detector can be used to receive data primarily from the top layer for overlay applications. A separate detector, such as a backscattered electron detector, can be used to receive data from the hidden layers.

[0060] Such embodiments can operate to correct data in real-time or offline.

[0061] The correction response function can include subtracting the backscattered electron data from the secondary electron data.

[0062] The correction response function can include extrapolating a spectral distribution of the backscattered electron data (similar to an optical spectrum) toward lower energies. For example, the extrapolation can be performed by fitting a function to the spectrum and using the function to extrapolate to other spectral locations that lack measured data. This can result in extrapolated backscattered electron data. The extrapolated backscattered electron data can be subtracted from the secondary electron data to produce a corrected response function.

[0063] The correction response function can include modeling an energy distribution function of the backscattered electron data, thereby producing a modeled energy distribution function. The modeled energy distribution function can be calibrated using one or more measurements made with an electron beam inspection tool.

[0064] The correction response function can include determining a relative fraction. A measured parameter can be related to a relative fraction of a sample property based on the backscattered electron data to the sum of the backscattered electron data and the secondary electron data.

[0065] The relative fraction can be stored on an electron data storage unit. One or more determined relative fractions can be stored in a lookup table for later use.

[0066] The correction response function can include using a machine learning model. The machine learning module can be trained using stored relative fractions stored on an electronic data storage unit. For example, the machine learning model can be an artificial neural network, such as a convolutional neural network, and the artificial neural network can be applied to new measurements.

[0067] Such embodiments provide for detection by measuring backscatter electron data and backscatter electron and secondary electron data pixel by pixel. Other such embodiments provide for detection by measuring backscatter electron data and backscatter electron and secondary electron data by pixel group, such as line by line or frame by frame. The same location can be measured for backscatter electron and backscatter electron plus secondary electron data. Calibration between different detectors can not be needed.

[0068] Embodiments can include hardware that controls the scan or cycling of the filter bias voltage and synchronizes it with other system time scales, such as the pixel clock or frame rate.

[0069] The method can further include detecting a second portion of secondary electrons at an opposing secondary electron detector. Such detection can produce opposing secondary electron data. The secondary electron data can be compared to the opposing secondary electron data to produce tilt data, which can refer to the alignment of the beam axis relative to the optical axis. The tilt data can be used to change the tilt of the electron beam. The tilt of the electron beam can be changed by changing the electron beam emitter.

[0070] The secondary electron data can be compared to the opposing secondary electron data with a dark field image. Dark field images can be used for contrast enhancement and / or topography measurement. Such images can be used to measure and control beam tilt.

[0071] Embodiments can determine and correct the beam tilt of an electron beam inspection system. The tilt related values can be measured by comparing opposing channels using a multi-channel detector. The comparison of data, for example, difference data, can be used to control the beam tilt in real time with a feedback loop. Such a feedback loop can be used to determine and correct the beam tilt in real time by using opposing channels of a multi-channel detector and a beam tilt calibration target to determine the tilt related values.

[0072] One embodiment can implement a specialized target using a high-Z substrate and aperture array. The high-Z substrate can be selected so that it produces backscatter electrons. For example, this can be a Pelco silicon nitride support film.

[0073] A system or non-transitory computer readable storage medium according to the present invention can implement such various embodiments of the present invention.

[0074] In Figure 5An embodiment system 12 is illustrated. In system 12, primary beam components 41-43 can be suitably deflected by deflection unit 13 toward sample 5. Portions of secondary and backscattered electrons 44-45 resulting from primary beam components 41-43 impinging sample 5 can be detected at detector 6. Detector 6 can have one or more portions (e.g., portions 61-65) configured to detect secondary or backscattered electrons at various angles depending on the flatness or tilt of the surface. Note that there can be embodiments with varying numbers of portions (in addition to portions 61-65). Detector 6 can be configured to supply one or more dark field images 14 to analysis unit 15. Examples of dark field images 14 include those images depicted in Figures 7A-7C , including images 201 and 202.

[0075] Figure 6 Still another embodiment system 16 is depicted, where a membrane with holes is placed between sample 5 and detectors 66-68. Note that there can be embodiments with varying numbers of detectors (in addition to detectors 66-68). For example, the membrane with holes can be 200 nm thick and / or can be made of silicon. However, another thickness or material can be used depending on process needs. Membrane with holes 17 can have one or more holes, and those holes can be circular or other shapes. For example, membrane with holes 17 can have circular holes with diameters ranging (for example) from 2500 nm to 5000 nm. Sample 5 can be a high-Z target of strong backscatter data. Membrane with holes 17 can be placed at a distance h above sample 5. For example, distance h can range from 2 μιη to 10 μιη, however, another distance h can be used depending on process needs. Portions of backscattered electrons 47 and 48 can result from primary beam 4 impinging sample 5.

[0076] Referring back to Figure 5 , analysis unit 15 can analyze dark field images (for example) by using image processing techniques (e.g., by comparing a number of side channels) to determine a tilt of the sample. For example, referring to Figure 7A , dark field image 201 can show that the surface from which scattered electrons emanate is flat. For yet another example, referring to Figure 7B , dark field image 202 can show that the surface from which scattered electrons emanate is tilted. In this manner, a tilt corresponding to a tilted surface of the sample can be measured and can be fed back into the system for correction.

[0077] Figure 8An embodiment of system 800 is shown. System 800 includes an optically based subsystem 801. Generally, the optically based subsystem 801 is configured to generate an optically based output for sample 802 by directing light to sample 802 (or scanning light across sample 802) and detecting the light from sample 802. In one embodiment, sample 802 includes a wafer. The wafer may include any wafer known in the art. In another embodiment, the sample includes a photomask. The photomask may include any photomask known in the art.

[0078] In Figure 8 In the embodiment of system 800 shown, the optical-based subsystem 801 includes an illumination subsystem configured to direct light to sample 802. The illumination subsystem includes at least one light source. For example, such as... Figure 8 As shown, the illumination subsystem includes a light source 803. In one embodiment, the illumination subsystem is configured to direct light to the sample 802 at one or more incident angles (which may include one or more tilt angles and / or one or more normal angles). For example, such as Figure 8 As shown, light from light source 803 is guided to sample 802 through optical element 804 and then lens 805 at an oblique angle of incidence. The oblique angle of incidence can include any suitable oblique angle of incidence, which can vary depending on (for example) the characteristics of sample 802.

[0079] The light source 803 or beam source may include a broadband plasma source, a lamp, or a laser. In some embodiments, the beam source may also emit light or photons, which may be in the form of infrared light, visible light, ultraviolet light, or X-ray light.

[0080] The optical-based subsystem 801 can be configured to guide light to the sample 802 at different times and at different incident angles. For example, the optical-based subsystem 801 can be configured to change one or more characteristics of one or more elements of the illumination subsystem, so that it can differ from... Figure 8 The incident angle shown guides light to sample 802. In one such example, an optically based subsystem 801 can be configured to move light source 803, optical element 804, and lens 805 such that light is guided to sample 802 at different tilted incident angles or normal (near normal) incident angles.

[0081] In some examples, the optical-based subsystem 801 may be configured to direct light to the sample 802 at more than one incident angle at the same time. For example, the illumination subsystem may include more than one illumination channel, one of which may include, for instance, […]. Figure 8The light source 803, optical element 804, and lens 805 shown in FIG. 8, and another one of the illumination channels (not shown), can include similarly configured elements, which can be different or the same, or can include at least a light source and possibly one or more other components, such as those described further herein. If such light is directed to the sample at the same time as other light, one or more properties of the light directed to the sample 802 at different angles of incidence (e.g., wavelength, polarization, etc.) can be different, such that light resulting from illumination of the sample 802 at different angles of incidence can be distinguished from one another at the detector.

[0082] In another example, the illumination subsystem can include only one light source (e.g., Figure 8 The light source 803 shown in FIG. 8), and light from the light source can be separated into different optical paths (e.g., based on wavelength, polarization, etc.) by one or more optical elements (not shown) of the illumination subsystem. Light in each of the different optical paths can then be directed to the sample 802. Multiple illumination channels can be configured to direct light to the sample 802 at the same time or at different times (e.g., when different illumination channels are used to sequentially illuminate the sample). In another example, the same illumination channel can be configured to direct light having different properties to the sample 802 at different times. For example, in some examples, the optical element 804 can be configured as a spectral filter, and the properties of the spectral filter can be changed in a variety of different ways (e.g., by swapping spectral filters) such that different wavelengths of light can be directed to the sample 802 at different times. The illumination subsystem can have any other suitable configuration known in the art for sequentially or simultaneously directing light having different or the same properties to the sample 802 at different or the same angles of incidence.

[0083] In one embodiment, the light source 803 can include a broadband plasma (BBP) source. In this way, light generated by the light source 803 and directed to the sample 802 can include broadband light. However, the light source can include any other suitable light source, such as a laser or a lamp. The laser can include any suitable laser known in the art and can be configured to generate light at any one or several suitable wavelengths known in the art. In addition, the laser can be configured to generate light that is monochromatic or nearly monochromatic. In this way, the laser can be a narrowband laser. The light source 803 can also include a polychromatic light source that generates light at a plurality of discrete wavelengths or wavelength bands.

[0084] Light from the optical element 804 can be focused onto the sample 802 by the lens 805. Although the lens 805 is shown in FIG. 8 as a single refractive optical element, it should be understood that in practice the lens 805 can include several refractive and / or reflective optical elements that combine to focus light from the optical element to the sample. Figure 8 Although the lens 805 is shown in FIG. 8 as a single refractive optical element, it should be understood that in practice the lens 805 can include several refractive and / or reflective optical elements that combine to focus light from the optical element to the sample. Figure 8The illumination subsystem shown in and described herein can include any other suitable optical elements (not shown). Examples of such optical elements include, but are not limited to, polarizing components, spectral filters, spatial filters, reflective optical elements, apodizers, beam splitters (e.g., beam splitter 813), apertures, and the like, which can include any such suitable optical elements known in the art. Additionally, the optical-based subsystem 801 can be configured to vary one or more of the elements of the illumination subsystem based on the type of illumination that will be used to generate the optical-based output.

[0085] The optical-based subsystem 801 can also include a scanning subsystem configured to cause light to be scanned over the sample 802. For example, the optical-based subsystem 801 can include a stage 806 on which the sample 802 is disposed during generation of the optical-based output. The scanning subsystem can include any suitable mechanical and / or robotic assembly (including the stage 806) that can be configured to move the sample 802 so that light can be scanned over the sample 802. Additionally or alternatively, the optical-based subsystem 801 can be configured so that one or more optical elements of the optical-based subsystem 801 perform some scanning of light over the sample 802. Light can be scanned over the sample 802 in any suitable manner, such as in a serpentine path or in a spiral path.

[0086] The optical-based subsystem 801 further includes one or more detection channels. At least one of the one or more detection channels includes a detector configured to detect, by the subsystem, light from the sample 802 due to illumination of the sample 802 and generate an output in response to the detected light. For example, Figure 8 The optical-based subsystem 801 shown in includes two detection channels, one formed by the collector 807, element 808, and detector 809 and the other formed by the collector 810, element 811, and detector 812. As Figure 8 As shown in, the two detection channels are configured so as to collect and detect light at different collection angles. In some examples, the two detection channels are configured to detect scattered light, and the detection channels are configured to detect light scattered from the sample 802 at different angles. However, one or more of the detection channels can be configured to detect another type of light (e.g., reflected light) from the sample 802.

[0087] As Figure 8As further shown in FIG. 8, the two detection channels are shown positioned in the plane of the paper, and the illumination subsystem is also shown positioned in the plane of the paper. Thus, in this embodiment, the two detection channels are positioned in (e.g., centered in) the plane of incidence. However, one or more of the detection channels are positioned out of the plane of incidence. For example, the detection channel formed by the collector 810, the element 811, and the detector 812 can be configured to collect and detect light scattered out of the plane of incidence. Thus, such a detection channel can generally be referred to as a "side" channel, and such a side channel can be centered in a plane that is generally perpendicular to the plane of incidence.

[0088] Although Figure 8 Although an embodiment of the optical-based subsystem 801 including two detection channels is shown, the optical-based subsystem 801 can include a different number of detection channels (e.g., only one detection channel or two or more detection channels). In one such example, the detection channel formed by the collector 810, the element 811, and the detector 812 can form one side channel as described above, and the optical-based subsystem 801 can include an additional detection channel (not shown) formed as another side channel positioned on an opposite side of the plane of incidence. Thus, the optical-based subsystem 801 can include detection channels including the collector 807, the element 808, and the detector 809 and centered in the plane of incidence and configured to collect and detect light scattered at normal or near-normal angles of scatter from the surface of the sample 802. Thus, this detection channel can generally be referred to as a "top" channel, and the optical-based subsystem 801 can also include two or more side channels configured as described above. As such, the optical-based subsystem 801 can include at least three channels (i.e., one top channel and two side channels), and each of the at least three channels has its own collector, each of which is configured to collect light at a different angle of scatter than each of the other collectors.

[0089] As further described above, each of the detection channels included in the optical-based subsystem 801 can be configured to detect scattered light. Thus, Figure 8 The optical-based subsystem 801 shown in FIG. 8 can be configured for dark-field (DF) output generation of the sample 802. However, the optical-based subsystem 801 can also or alternatively include detection channels configured for bright-field (BF) output generation of the sample 802. In other words, the optical-based subsystem 801 can include at least one detection channel configured to detect light specularly reflected from the sample 802. Thus, the optical-based subsystem 801 described herein can be configured for DF imaging only, BF imaging only, or both DF and BF imaging. Although each of the collectors is in the plane of incidence in the embodiment shown in FIG. 8, the collectors can be positioned out of the plane of incidence in other embodiments. Figure 8The collection optics are shown as single refractive optical elements, but it should be understood that each of the collection optics can include one or more refractive optical elements and / or one or more reflective optical elements.

[0090] The one or more detection channels can include any suitable detectors known in the art. For example, the detectors can include photomultiplier tubes (PMTs), charge-coupled devices (CCDs), time delay integration (TDI) cameras, and any other suitable detectors known in the art. The detectors can also include non-imaging detectors or imaging detectors. In this way, if the detectors are non-imaging detectors, each of the detectors can be configured to detect some characteristic (e.g., intensity) of the scattered light but can not be configured to detect such a characteristic as a function of position within the imaging plane. As such, the output generated by each of the detectors included in each of the detection channels of the optics-based subsystem can be a signal or data, but not an image signal or image data. In such examples, a processor such as the processor 814 can be configured to generate an image of the sample 802 from the non-imaging output of the detectors. However, in other examples, the detectors can be configured as imaging detectors configured to generate an imaging signal or image data. Thus, the optics-based subsystem can be configured to generate optical images or other optics-based outputs described herein in a number of ways.

[0091] It should be noted that the systems described herein Figure 9 To generally illustrate a configuration of the optics-based subsystem 801, which can be included in or can generate the optics-based output used by the system embodiments described herein. The optics-based subsystem 801 configuration described herein can be varied to optimize the performance of the optics-based subsystem 801, as is typically performed when designing a commercial output acquisition system. In addition, the systems described herein can be implemented using existing systems (e.g., by adding the functionality described herein to an existing system). For some such systems, the methods described herein can be provided as optional functionality of the system (e.g., in addition to other functionality of the system). Another option is that the systems described herein can be designed as a completely new system.

[0092] In examples, the processor 814 is in communication with the system 800.

[0093] Figure 9 is a block diagram of an embodiment of a system 900. The system 900 includes a tool, e.g., a wafer inspection tool, configured to generate an image of a sample 904, which can include a wafer or a reticle, including an e-beam column 901.

[0094] The tool includes an output acquisition subsystem, which comprises at least an energy source and a detector. The output acquisition subsystem can be an electron beam-based output acquisition subsystem. For example, in one embodiment, the energy directed to sample 904 contains electrons, and the energy detected from sample 904 contains electrons. In this way, the energy source can be an electron beam source. Figure 9 In one such embodiment shown, the output acquisition subsystem includes an electron column 901 coupled to a computer subsystem 902. A stage 910 holds a sample 904.

[0095] Also Figure 9 As shown, electron column 901 includes electron beam source 903 configured to generate electrons that are focused onto sample 904 by one or more elements 905. For example, electron beam source 903 may include a cathode source or an emitter tip. For example, one or more elements 905 may include a gun lens, anode, beam-limiting aperture, gate valve, beam current selection aperture, objective lens, and scanning subsystem; all of these elements may include any such suitable elements known in the art.

[0096] Electrons (e.g., secondary electrons) returning from sample 904 can be focused onto detector 907 by one or more elements 906. One or more elements 906 may include (for example) a scanning subsystem, which may be the same scanning subsystem included in element 905.

[0097] The electronic column 901 may also contain any other suitable elements known in this art.

[0098] Despite the electron column Figure 9 The diagram illustrates a configuration where electrons are guided to sample 904 at an angle of incidence and scattered from sample 904 at another angle. However, the electron beam can be guided to and scattered from sample 904 at any suitable angle. Furthermore, the electron beam-based output acquisition subsystem can be configured to generate images of sample 904 using multiple modes (e.g., with different illumination angles, collection angles, etc.). These multiple modes of the electron beam-based output acquisition subsystem can differ in any image generation parameters of the output acquisition subsystem.

[0099] Computer subsystem 902 may be coupled to detector 907 as described above. Detector 907 may detect electrons returning from the surface of sample 904, thereby forming an electron beam image of sample 904. The electron beam image may contain anything suitable for an electron beam image. Computer subsystem 902 may be configured to use the output of detector 907 and / or the electron beam image to perform any of the functions described herein. Computer subsystem 902 may be configured to perform any additional steps described herein. Figure 9The system 900, which exhibits an output acquisition subsystem as shown in

[0100] It is noted that the systems described herein Figure 9 A configuration of an electron beam-based output acquisition subsystem that can be used in the embodiments described herein is generally illustrated. The electron beam-based output acquisition subsystem configuration described herein can be altered to optimize the performance of the output acquisition subsystem, as is typically performed when designing a commercial output acquisition system. In addition, the systems described herein can be implemented using existing systems (e.g., by adding the functionality described herein to an existing system). For some such systems, the methods described herein can be provided as optional functionality of the system (e.g., in addition to other functionality of the system). Another option is that the systems described herein can be designed as a completely new system.

[0101] Although the output acquisition subsystem is described above as an electron beam-based output acquisition subsystem, the output acquisition subsystem can be an ion beam-based output acquisition subsystem. Such an output acquisition subsystem can be configured as shown in Figure 8 In addition, the output acquisition subsystem can be any other suitable ion beam-based output acquisition subsystem, such as an ion beam-based output acquisition subsystem included in a commercially available focused ion beam (FIB) system, a helium ion microscope (HIM) system, and a secondary ion mass spectrometry (SIMS) system.

[0102] The computer subsystem 902 includes a processor 908 and an electronic data storage unit 909. The processor 908 can include a microprocessor, microcontroller, or other device.

[0103] The processor 814 or 908, or the computer subsystem 902, respectively, can be coupled to the components of the system 800 or 900, respectively, in any suitable manner (e.g., via one or more transmission media, which can include wired and / or wireless transmission media), such that the processor 814 or 908 can receive the output, respectively. The processor 814 or 908 can be configured to use the output to perform a number of functions. The system 800 or 900 can receive instructions or other information from the processor 814 or 908, respectively. Optionally, the processor 814 or 908 and / or the electronic data storage unit 815 or 909 can be in electronic communication with a wafer inspection tool, a wafer metrology tool, or a wafer re-inspection tool (not illustrated) to receive additional information or send instructions, respectively. For example, the processor 814 or 908 and / or the electronic data storage unit 815 or 909 can be in electronic communication with a scanning electron microscope (SEM), respectively.

[0104] The processor 814 or 908 is in electronic communication with a wafer inspection tool, such as the detector 809 or 812 or the detector 907, respectively. The processor 814 or 908 can be configured to process images generated using measurements from the detector 809 or 812 or the detector 907, respectively. For example, the processor can perform the method 100 or embodiments of the schematic 1 or portions of the systems 12 and 16.

[0105] The processor 814 or 908 or computer subsystem 902, other system, or other subsystem described herein can be part of various systems, including a personal computer system, image computer, mainframe computer system, workstation, network appliance, internet appliance, or other device. The subsystem or system can also include any suitable processor known in the art, such as a parallel processor. In addition, the subsystem or system can include a platform with high speed processing and software, either as a standalone or networked tool.

[0106] The processor 814 or 908 and electronic data storage unit 815 or 909, respectively, can be disposed in or otherwise part of the system 800 or 900 or another device, respectively. In an example, the processor 814 or 908 and electronic data storage unit 815 or 909 can be part of a separate control unit or in a centralized quality control unit. Multiple processors 814 or 908 or electronic data storage units 815 or 909, respectively, can be used.

[0107] The processor 814 or 908 can be implemented in fact by any combination of hardware, software and firmware. Moreover, functions described herein can be performed by one unit, or allocated among different components, each of which can be implemented by any combination of hardware, software and firmware. Program code or instructions for implementing various methods and functions of the processor 814 or 908 can be stored in a readable storage medium such as a memory or other memory in the electronic data storage unit 815 or 909, respectively.

[0108] If the system 800 or 900 includes more than one processor 814 or 908 or computer subsystem 902, the different subsystems can be coupled to each other so that images, data, information, instructions, etc. can be sent between the subsystems. For example, one subsystem can be coupled to additional subsystems by any suitable transmission medium, which can include any suitable wired and / or wireless transmission medium known in the art. Two or more of such subsystems can also be coupled by a shared computer readable storage medium (not shown).

[0109] The processor 814 or 908 can be configured to perform several functions using the output of the system 800 or 900, respectively, or other output. For example, the processor 814 or 908 can be configured to send the output to an electronic data storage unit 815 or 909, respectively, or another storage medium. The processor 814 or 908 can be further configured as described herein.

[0110] The processor 814 or 908, or the computer subsystem 902, can be part of a defect review system, an inspection system, a metrology system, or some other type of system. Thus, embodiments disclosed herein describe some configurations that can be tailored to several ways for systems having different capabilities more or less suitable for different applications.

[0111] If the system includes more than one subsystem, the different subsystems can be coupled to one another such that images, data, information, instructions, etc. can be sent between the subsystems. For example, one subsystem can be coupled to an additional subsystem by any suitable transmission medium, which can include any suitable wired and / or wireless transmission medium known in the art. Two or more of such subsystems can also be effectively coupled by sharing a computer-readable storage medium (not shown).

[0112] The processor 814 or 908 can be configured according to any of the embodiments described herein. The processor 814 or 908 can also be configured to perform other functions or additional steps using the output of the system 800 or 900, respectively, or using images or data from other sources.

[0113] The processor 814 or 908 can be communicatively coupled to any of the various components or subsystems of the system 800 or 900, respectively, in any manner known in the art. Furthermore, the processor 814 or 908 can be configured to receive and / or acquire data or information from other systems (e.g., inspection results of an inspection system such as a review tool, a remote database containing design data, etc.) through a transmission medium, which can include wired and / or wireless portions. In this manner, the transmission medium can serve as a data link between the processor 814 or 908 and other subsystems of the system 800 or 900, respectively, or between systems external to the system 800 or 900, respectively.

[0114] In embodiments, the processor 814 or the processor 908 can be configured to perform steps according to embodiments of the method 100 or portions of the schematic 1 or the systems 12 and 16.

[0115] In an embodiment, processor 814 or processor 908 may be further configured to perform fine alignment including partition translation, wherein the partition translation includes: dividing a reference image into one or more reference image sub-segments; dividing a test image into one or more test image sub-segments, each test image sub-segment corresponding to a reference image sub-segment; and translating each test image sub-segment to align with its corresponding reference image sub-segment.

[0116] The various steps, functions, and / or operations of system 800 or system 900, and the methods disclosed herein, are performed by one or more of the following: electronic circuits, logic gates, multiplexers, programmable logic devices, ASICs, analog or digital controllers / switches, microcontrollers, or computing systems. Program instructions for implementing methods such as those described herein may be transmitted via or stored on a carrier medium. The carrier medium may include storage media, such as read-only memory, random access memory, magnetic disks or optical disks, non-volatile memory, solid-state memory, magnetic tape, etc. The carrier medium may include transmission media, such as wires, cables, or wireless transmission links. For example, the various steps described throughout the invention may be performed by a single processor 814 or a single processor 908 (or computer subsystem 902), or alternatively, multiple processors 814 or multiple processors 908 (or multiple computer subsystems 902). Furthermore, different subsystems of system 800 or system 900 may comprise one or more computing or logic systems. Therefore, the above description should not be construed as a limitation of the invention but merely as illustrative.

[0117] Additional embodiments relate to a non-transitory computer-readable medium storing program instructions executable on a controller for performing a computer-implemented method for determining the height of an illuminated area on the surface of sample 802 or 904, as disclosed herein. Specifically, as ​ As shown in Figure 9, electronic data storage unit 815 or 909 or other storage media may contain non-transitory computer-readable media containing program instructions executable on processor 814 or 908, respectively. The computer-implemented method may include any step of any of the methods described herein, including portions of embodiments of method 100 or schematic diagram 1 or systems 12 and 16.

[0118] Program instructions for implementing methods such as those described herein may be stored on a computer-readable medium, such as electronic data storage unit 815, electronic data storage unit 909, or other storage media. The computer-readable medium may be a storage medium, such as a magnetic disk or optical disk, magnetic tape, or any other suitable non-transitory computer-readable medium known in the art.

[0119] The program instructions can be implemented in any of a wide variety of ways including program-based techniques, component-based techniques, and / or object-oriented techniques, among other techniques. For example, the program instructions can be implemented using ActiveX controls, C++ objects, JavaBeans, Microsoft Foundation Classes (MFC), Streaming SIMD Extensions (SSE), or other technologies or methods, as desired.

[0120] The components executed by the processor can include a deep learning module (e.g., a convolutional neural network (CNN) module). The deep learning module can have one of the configurations described further herein. Rooted in neural network technology, deep learning is a probabilistic graphical model with many layers of neurons, often referred to as a deep architecture. Deep learning techniques process information such as images, text, speech, and so on in a hierarchical manner. During use of deep learning in the present disclosure, feature extraction is achieved automatically using learning from data. For example, the deep learning module can be used to extract features referenced in determining rotational and translational offsets based on one or more extracted features.

[0121] Generally, deep learning (also referred to as deep structured learning, hierarchical learning, or deep machine learning) is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data. In a simple case, there can be two groups of neurons: one group receives input signals and one group sends output signals. When the input layer receives input, it passes a modified version of the input to the next layer. In a deep network, there are many layers between the input and the output, allowing the algorithm to use multiple layers of processing consisting of linear and non-linear transformations.

[0122] Deep learning is part of a wider family of machine learning methods based on learning data representations. Observations (e.g., features to be extracted for reference) can be represented in many ways, such as vectors of per-pixel intensity values, or in more abstract ways as a set of edges, regions of particular shapes, and so on. Some representations are better than others at simplifying the learning task (e.g., face recognition or facial expression recognition). Deep learning can provide efficient algorithms for unsupervised or semi-supervised feature learning and hierarchical feature extraction.

[0123] Research in this area attempts to make better representations and create models to learn these representations from large-scale data. Some of the representations are inspired by advances in neuroscience and loosely based on interpretations of information processing and communication patterns in the nervous system, such as trying to define relationships between various stimuli and associated neuron responses in the brain, e.g., neural coding.

[0124] There are many variants of neural networks that have deep architectures depending on the probabilistic specification and network architecture, including but not limited to deep belief networks (DBN), restricted Boltzmann machines (RBM), and autoencoders. Another type of deep neural network (CNN) can be used for feature analysis. The actual implementation can vary depending on the size of the input image, the number of features to be analyzed, and the nature of the problem. Other layers can be included in the deep learning module in addition to the neural networks disclosed herein.

[0125] In an embodiment, the deep learning model is a machine learning model. Machine learning can generally be defined as a type of artificial intelligence (AI) that endows computers with the ability to learn without being explicitly programmed. Machine learning focuses on the development of computer programs that can learn to grow and change when exposed to new data. Machine learning explores the study and construction of algorithms that can learn from data and make predictions about data. Such algorithms overcome the rigidity of strictly following static program instructions by making data-driven predictions or decisions by building models from sample inputs.

[0126] In some embodiments, the deep learning model is a generative model. Generative models can generally be defined as models that are probabilistic in nature. In other words, a generative model is one that performs forward simulation or rule-based methods. A generative model can be learned based on a suitable training data set, where its parameters can be learned. In one embodiment, the deep learning model is configured as a deep generative model. For example, the model can be configured to have a deep learning architecture, where the model can include multiple layers that perform several algorithms or transformations.

[0127] In another embodiment, the deep learning model is configured as a neural network. In yet another embodiment, the deep learning model can be a deep neural network with a set of weights that model the world from data that has been fed to it to train on. Neural networks can generally be defined as a computational method based on a relatively large set of neural units that loosely model the way the biological brain solves problems with relatively large clusters of biological neurons connected by axons. Each neural unit is connected to many others, and links can force or inhibit their effect on the activation state of connected neural units. These systems are self-learning and trained rather than explicitly programmable, and are good at areas where traditional computer programs have difficulty expressing solutions or feature detection.

[0128] Neural networks are generally composed of multiple layers, and the signal path goes through from front to back. The goal of a neural network is to solve problems in the same way that the human brain does, although several neural networks are more abstract. Modern neural network projects typically use thousands to millions of neural units and millions of connections. The neural network can have any suitable architecture and / or configuration known in the art.

[0129] In one embodiment, a deep learning model for semiconductor inspection applications disclosed herein is configured as AlexNet. For example, AlexNet includes several (e.g., 5) convolutional layers followed by several (e.g., 3) fully connected layers that are configured or trained in combination to analyze features for determining rotational and translational offsets. In another such embodiment, a deep learning model for semiconductor inspection applications disclosed herein is configured as GoogleNet. For example, GoogleNet can include convolutional layers, pooling layers, and fully connected layers, such as those described further herein, that are configured and trained to analyze features for determining rotational and translational offsets. While GoogleNet architectures can include a relatively high number of layers, particularly as compared to some other neural networks described herein, some of the layers can operate in parallel and groups of layers that function in parallel with each other are commonly referred to as inception modules. Other of the layers can operate sequentially. Thus, GoogleNet differs from other neural networks described herein in that not all layers are arranged in a sequential structure. Parallel layers can be similar to Google's inception network or other structures.

[0130] In yet another such embodiment, a deep learning model for semiconductor inspection applications described herein is configured as a Visual Geometry Group (VGG) network. For example, VGG networks are created by increasing the number of convolutional layers while fixing other parameters of the architecture. Convolutional layers are added to increase depth by using roughly small convolutional filters in all layers. Like other neural networks described herein, VGG networks are created and trained to analyze features for determining rotational and translational offsets. VGG networks also include convolutional layers followed by fully connected layers.

[0131] In some such embodiments, a deep learning model for semiconductor inspection applications described herein is configured as a deep residual network. For example, like some other networks described herein, a deep residual network can include convolutional layers followed by fully connected layers that are configured and trained in combination for feature property extraction. In a deep residual network, the layers are configured to learn a residual function with respect to layer inputs rather than learning an un-referenced function. In particular, rather than hoping that each individual stack of layers directly fits the desired base mapping, these layers are explicitly allowed to fit a residual mapping, which is achieved by having a feed-forward neural network with shortcut connections. A shortcut connection is a connection that skips one or more layers. A deep residual net can be created by taking an ordinary neural network structure that includes convolutional layers and inserting shortcut connections, which thereby takes the ordinary neural network and converts it into a residual learning copy thereof.

[0132] In yet another such embodiment, a deep learning model for semiconductor inspection applications disclosed herein includes one or more fully connected layers configured for analyzing features for determining rotational and translational offsets. A fully connected layer can generally be defined as a layer in which each of the nodes is connected to each of the nodes in the previous layer. The fully connected layer can perform classification based on features extracted by a convolutional layer, which can be configured as further described herein. The fully connected layer is configured for feature selection and classification. In other words, the fully connected layer selects features from a feature map and then analyzes the input image based on the selected features. The selected features can include all of the features in the feature map (if appropriate) or just some of the features in the feature map.

[0133] In some embodiments, the information determined by the deep learning model includes feature properties extracted by the deep learning model. In one such embodiment, the deep learning model includes one or more convolutional layers. The convolutional layers can have any suitable configuration known in the art. In this way, the deep learning model (or at least a portion of the deep learning model) can be configured as a CNN. For example, the deep learning model can be configured as a CNN (which is typically a stack of convolutional layers and pooling layers) to extract local features. The embodiments described herein can employ deep learning concepts (e.g., CNNs) to solve representation inversion problems that are typically difficult to solve. The deep learning module can have any CNN configuration or architecture known in the art. One or more pooling layers can also have any suitable configuration known in the art (e.g., max pooling layers) and are generally configured for reducing the dimensionality of the feature map produced by the one or more convolutional layers while preserving the most important features.

[0134] In general, the deep learning models described herein are trained deep learning models. For example, the deep learning model can be pre-trained by one or more other systems and / or methods. The deep learning model has been created and trained and then the functionality of the model can be determined as described herein, which can then be used to perform one or more additional functions of the deep learning model.

[0135] As stated above, although CNNs are used herein to illustrate the architecture of the deep learning system, the present disclosure is not limited to CNNs. Other variants of deep learning architectures can be used in embodiments. For example, autoencoders, DBNs, and RBMs can be used. Random data forests can also be used.

[0136] The training data can be input for model training (e.g., CNN training), which can be performed in any suitable manner. For example, model training can include inputting the training data to a deep learning model (e.g., a CNN); and modifying one or more parameters of the model until the output of the model is the same (or substantially the same) as the external validation data. The model training can result in one or more trained models, which can then be sent to model selection performed using the validation data. The results produced by the one or more trained models used on the validation data (input to the one or more trained models) can be compared to the validation data to determine which of the models is the best model. For example, the model that produces results that most closely match the validation data can be selected as the best model. The test data can then be used for model evaluation of the selected model (e.g., the best model). The model evaluation can be performed in any suitable manner. The best model can also be sent to model deployment, where the best model can be sent to a semiconductor inspection tool for use (post-training mode).

[0137] The steps of the methods described in the various embodiments, as well as the examples disclosed herein, are sufficient to perform the methods of the present invention. Thus, in an embodiment, the method consists essentially of a combination of the steps of the methods disclosed herein. In another embodiment, the method consists of such steps.

[0138] While the present invention has been described with respect to one or more particular embodiments, it is understood that other embodiments of the present invention can be made without departing from the scope of the invention.

Claims

1. A method of calibration comprising: emitting an electron beam toward a sample; modulating the electron beam to obtain a beam signal, wherein the electron beam is described by a source wave function having a source phase and a landing angle; detecting, using an electron detector, a portion of scattered electrons including secondary electrons and backscattered electrons emitted by the sample to obtain electron data, wherein the electron data includes an electron wave function having an electron phase and an electron landing angle; determining, using a processor, a phase delay between the source phase and the electron phase; and correcting, using the processor, a response function of the electron beam tool based on the phase delay and a difference between the source wave function and the electron wave function.

2. The correction method of claim 1, wherein, the method further comprising: applying a high pass filter to detect backscattered electron data and secondary electron data in a high pass off mode; and applying the high pass filter to detect a remaining backscattered electron portion in a high pass on mode.

3. The method of calibration of claim 2, further comprising subtracting the backscattered electron data from the electron data to obtain the secondary electron data.

4. The method of calibration of claim 2, further comprising determining, based on the backscattered electron data, a relative fraction of a measurement parameter and a sample property as a sum of the backscattered electron data and the electron data.

5. The method of calibration of claim 2, further comprising: detecting, using a counter electron detector, the secondary electrons and backscattered electrons emitted by the sample to obtain counter electron data; comparing, using the processor, the electron data and the counter electron data to produce tilt data; and using the tilt data to change a tilt of the electron beam.

6. The method of calibration of claim 1, wherein the modulating comprises changing a beam current of the electron beam.

7. The method of calibration of claim 1, wherein the modulating comprises switching the electron beam between on and off.

8. The method of calibration of claim 1, wherein the modulating comprises directing the electron beam between a beamlet stack and the sample.

9. The method of calibration of claim 1, wherein the modulating comprises deflecting the electron beam toward the sample.

10. A calibration system comprising: an electron beam tool, the electron beam tool comprising: an electron beam emitter configured to emit electrons in an electron beam toward a sample, wherein the electron beam is described by a source wave function having a source phase and a landing angle; a stage configured to hold the sample; an electron detector configured to detect a portion of scattered electrons including secondary electrons and backscattered electrons emitted by the sample to obtain electron data, wherein the electron data includes an electron wave function having an electron phase and an electron landing angle; and a controller in electronic communication with the electron beam tool, the controller having a processor; wherein the controller is configured to: instruct the electron beam emitter to modulate the electron beam to obtain a beam signal; determine a phase delay between the source phase and the electron phase; and correct a response function of the electron beam tool based on the phase delay and a difference between the source wave function and the electron wave function. correcting a response function of the electron beam tool based on the phase delay and a difference between the source wave function and the electron wave function.

11. The correction system of claim 10, wherein the controller is further configured to: apply a high pass filter to detect backscattered electron data and secondary electron data in a high pass off mode; and apply the high pass filter to detect a remaining backscattered electron portion in a high pass on mode.

12. The correction system of claim 11, wherein the controller is configured to further subtract the backscattered electron data from the electron data to obtain the secondary electron data.

13. The correction system of claim 11, wherein the controller is configured to: determine, based on the backscattered electron data, a relative fraction of a measurement parameter and a sample property as a sum of the backscattered electron data and the electron data.

14. The correction system of claim 10, further comprising: a back electron detector configured to detect the secondary electrons and backscattered electrons emitted by the sample to obtain back electron data; and wherein the controller is further configured to: compare the electron data and the back electron data to produce tilt data; and use the tilt data to change an angle of the electron beam emitter to change a tilt of the electron beam.

15. The correction system of claim 10, wherein the controller is configured to modulate the electron beam by instructing the electron beam emitter to change a beam current of the electron beam.

16. The correction system of claim 10, wherein the controller is configured to modulate the electron beam by instructing the electron beam emitter to turn on and off the electron beam.

17. The correction system of claim 10, wherein the controller is configured to modulate the electron beam by instructing the electron beam emitter to direct the electron beam between a beamlet stack and the sample.

18. The correction system of claim 10, wherein the controller is configured to modulate the electron beam by instructing the electron beam emitter to deflect the electron beam toward the sample.

19. A non-transitory computer-readable storage medium comprising one or more programs for execution on one or more computing devices to: instruct an electron beam tool to emit an electron beam toward a sample, wherein the electron beam is described by a source wave function having a source phase and a landing angle; modulate the electron beam to obtain a beam signal; receive electron data from an electron detector that detects a portion of scattered electrons including secondary electrons and backscattered electrons emitted by the sample, wherein the electron data includes an electron wave function having an electron phase and an electron landing angle; determine a phase delay between the source phase and the electron phase; and correct a response function of the electron beam tool based on the phase delay and a difference between the source wave function and the electron wave function.

20. The non-transitory computer-readable storage medium of claim 19, wherein the one or more programs are further configured to, on one or more computing devices: train a machine learning algorithm using a beam current of the electron beam, the electronic data, and the response function.

Images