How to Examine a Sample Using a Charged Particle Microscope
The method enhances charged particle microscopy by combining measured and estimated spectral data with machine learning to quickly and accurately determine the chemical composition of samples, addressing the time constraints of EDS methods.
Patent Information
- Application Number
- JP2020091005
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2019-05-29
- Filing Date
- 2020-05-26
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2040-05-26
AI Technical Summary
Existing charged particle microscopy methods, particularly Energy Dispersive X-ray Spectroscopy (EDS), are time-consuming in determining the chemical composition of samples due to the need for extensive X-ray spectrum accumulation at discrete analysis points, limiting the speed and accuracy of phase identification.
A method utilizing spectral information of detected radiation to rapidly and accurately assign phases to sample locations by combining measured and estimated data, employing a first detector for primary radiation and optionally a second detector for secondary radiation, and utilizing machine learning estimators to predict dense spectra from sparse data.
This approach significantly speeds up the EDS mapping process by allowing rapid phase determination, even with sparse data, improving the accuracy and efficiency of chemical composition analysis.
Smart Images

Figure 0007582590000001 
Figure 0007582590000002 
Figure 0007582590000003
Abstract
Description
[Technical field]
[0001] explanation The present invention relates to a method of inspecting a sample using a charged particle microscope, comprising the steps of providing a charged particle beam and a sample, scanning the charged particle beam over a number of sample locations on the sample, and detecting a first type of radiation from the sample in response to the beam being scanned over the number of sample locations using a first detector. [Background technology]
[0002] Charged particle microscopy is a well-known and increasingly important technique for imaging microscopic objects, especially in the form of electron microscopes.So far, the basic type of electron microscope has evolved into a number of well-known devices, such as transmission electron microscopes (TEM), scanning electron microscopes (SEM) and scanning transmission electron microscopes (STEM), and into various variants, such as so-called "dual beam" devices (e.g., FIB-SEM), which additionally use a "machining" focused ion beam (FIB) to enable supporting actions, such as ion beam milling or ion beam induced deposition (IBID).Those skilled in the art will be familiar with the different types of charged particle microscopes.
[0003] Irradiation of a sample with a scanning electron beam stimulates the emission of "auxiliary" radiation from the sample in the form of secondary electrons, backscattered electrons, X-rays, and cathodoluminescence (infrared, visible, and / or ultraviolet photons). One or more components of this emitted radiation can be detected and used for sample analysis.
[0004] Typically in an SEM, backscattered electrons are detected by a solid-state detector, and each backscattered electron is amplified to create many electron-hole pairs in the semiconductor detector. The backscattered electron detector signal is used to form an image as the beam is scanned, and the brightness of each image point is determined by the number of backscattered electrons detected at the corresponding point on the sample as the primary beam moves across the sample. The image only provides information about the topology of the sample being inspected.
[0005] In a process called "energy dispersive X-ray spectroscopy" (also referred to as "EDS" or "EDX"), the energy of the X-rays coming from a sample in response to an electron beam is measured and plotted in a histogram to form a material-specific spectrum. The measured spectrum can be compared to known spectra of various elements to determine which elements and minerals are present in the sample.
[0006] One drawback of EDS is that it takes a significant amount of time to accumulate an X-ray spectrum of a sample. Typically, a grid with discrete analysis points is used. The electron beam dwells at each analysis point while the EDS detector records the X-rays. Once a sufficient number of X-ray counts have been recorded, the beam moves to the next analysis point. The signal from the EDS detector is fed to a signal processing unit which constructs an X-ray spectrum curve for each analysis point, which may then be queried against an extensive library of known mineral phases to select the one that best matches that analysis point. This known method is relatively time consuming to determine the phases present in a sample, i.e. the chemical composition. Summary of the Invention
[0007] With the above in mind, it is an object of the present invention to provide an improved method of examining a sample using a charged particle microscope, where spectral information of detected radiation is used to examine said sample. In particular, it is an object of the present invention to provide a method and apparatus for obtaining information about a sample more quickly and / or more accurately.
[0008] To this end, the present invention provides a method for inspecting a sample using a charged particle microscope, as defined in claim 1. The method comprises: Providing a charged particle beam and a sample; scanning the charged particle beam across a plurality of sample locations on the sample; detecting a first type of radiation from the sample in response to the beam scanned over a plurality of sample locations using a first detector; using spectral information of the detected first type of radiation to assign a plurality of mutually distinct phases to the sample at the plurality of sample locations; and providing, by a control unit, a data representation of the samples comprising information regarding at least the plurality of sample locations and the phase.
[0009] As defined herein, the method comprises: establishing an estimated phase for at least one other of the plurality of sample locations using information regarding the at least one previously assigned phase and its respective sample location; and assigning the estimated phase to the other sample locations.
[0010] As mentioned above, the method includes detecting a first type of radiation from the sample in response to the beam scanned over an area of the sample using a first detector. The method further includes collecting spectral information of the detected radiation of said first type. The radiation detected by the first detector may be associated with a particular scanning beam position, i.e., with a particular location on the sample. This means that spectral information may be collected and / or determined for corresponding locations on the sample as well. The spectral information obtained for different locations may be compared with each other, and one or more specific phases may be assigned to these different locations. The already assigned phases may be used to establish estimated phases for other locations, for example, based on proximity to these already assigned phases, based on partially obtained spectral profiles obtained for other locations, or based on other parameters.
[0011] Thus, in the methods described herein, measurement data for at least one sample location may be used to establish extrapolated data for at least one further sample location. In one embodiment, measurement data for several sample locations may be used to establish extrapolated data for at least one further sample location. It is believed that extrapolated data is established for several further sample locations.
[0012] The measured data (i.e. the phase assigned where the spectral information of the detected first type of radiation was used) and the estimated data (i.e. at least one estimated phase) may be combined. The measured data and the estimated data may then be combined into the data representation of the sample. This makes it possible to generate a data representation of the sample more quickly, especially since the process of estimating the phase based on an easily established phase is relatively quick. It is therefore possible to provide a user with information regarding the phase of the sample in a shorter time. Thus, the objective as defined in this specification is achieved.
[0013] In particular, the methods defined herein enable the established phase, i.e., chemical composition, at one (or more) sample location(s) to be assigned to the phase, i.e., chemical composition, at one (or more) further sample location(s). This enables an improvement in speed.
[0014] This is particularly applicable when the first detector is an EDS detector, i.e., in EDS mapping. The inventors have discovered that during X-ray mapping, the EDS spectra from a total of N individual pixels are not completely independent. In most EDS scenarios, it can be assumed that in the sample there are several (K << N) different chemical substances (also called "phases"), each having its characteristic spectrum. Thus, each spectrum originating from a given phase provides additional information about that phase and, by extension, about all pixels belonging to that phase. This additional information enables the prediction of a dense spectrum from a sparse spectrum and thus can speed up the entire EDS X-ray mapping process.
[0015] Advantageous embodiments are described below.
[0016] In one embodiment, the spectral information of the detected first type of radiation can be used to divide at least a portion of the scanned area of the sample into a plurality of segments. The first type of radiation at different positions along the scan in at least one of the plurality of segments can be combined to generate a combined spectrum of the sample in that one of the plurality of segments. Using a plurality of segments enables, for example, an improved estimated phase to be established by using the proximity of a new position to a previously determined position within the range of a particular segment. Using the combined spectrum enables the partial data of the new position to be compared with one or more previously obtained combined spectra, thereby facilitating the establishment of an estimated phase for that new position.
[0017] In one embodiment, a second detector is used to detect a second type of radiation from the sample in response to the beam being scanned over the region of the sample, the second type of radiation being used at least in part to establish an estimated phase for at least one other of the plurality of sample locations. In this embodiment, an additional detector is used to detect the second type of radiation. These second type of radiation are different from the first type of radiation. The second type of radiation may be used directly or indirectly in establishing an estimated phase for at least one other of the plurality of sample locations.
[0018] In an embodiment, the second detector is configured to detect charged particles, in particular electrons, such as backscattered electrons. Thus, the second detector may be a BSE detector. It is further conceivable that the first detector is configured to detect particles, in particular photons, such as X-ray photons.
[0019] The backscattering of electrons depends on the atomic number of the elements at the surface and the geometric relationship between the surface, the primary beam, and the detector. Hence, the backscattered electron image shows contour information, i.e., the boundaries between regions of different composition, and cross-sectional information. Acquiring a backscattered electron image is much faster than acquiring enough X-rays to collect a complete spectrum at each point, since it requires collecting only a sufficient number of electrons at each point to produce a reasonable contrast between points with different properties. Also, the probability of an electron being backscattered is higher than the probability of an electron producing a characteristic X-ray emission of a particular frequency. Acquiring sufficient backscattered electron image data at a single dwell point typically requires less than one microsecond, whereas acquiring enough X-rays to obtain an analyzable spectrum at a single dwell point typically requires more than one millisecond.
[0020] In one embodiment, the second type of radiation is used to define a plurality of segments of the sample. In this embodiment, the method includes: - detecting a second type of radiation from the sample in response to the beam being scanned over the area of the sample using a second detector; - dividing the scanned area of the sample into a plurality of segments using a second type of radiation.
[0021] As already mentioned above, the use of segments can be advantageous.
[0022] In one embodiment, an image is first acquired using a backscattered electron detector, and then the image is processed to identify regions (or segments) that appear to have the same elemental composition based on contrast. The beam is then scanned over the sample, preferably multiple times, over at least one or more of the identified regions, and x-ray spectra representative of the regions are collected. The x-rays generated during the identified regions and / or the backscattered electron detector scan can be advantageously used to obtain information that can be used to establish probable phases for other points at an early stage.
[0023] In one embodiment, the beam may be used to scan the sample and obtain spectral information for multiple analysis points simultaneously. Rather than collecting complete spectral information for each individual analysis point before moving on to the next analysis point, the sample may be scanned continuously, obtaining data for all required sample locations. The scanning of the beam may be relatively fast, and the beam may be scanned continuously or semi-continuously over the area of the sample being examined. After one or more scans, the spectral information obtained may be sparse, i.e., some analysis points may actually provide complete information, while others do not. In this way, the data obtained for some analysis points may be used to establish inferential data for other points.
[0024] The method may include the steps of additionally scanning at least a portion of an area of the sample to be examined and detecting the first type of radiation in response to the additional scanning using the first detector. By repeatedly scanning continuously or semi-continuously, more information about the first type of radiation can be obtained, further improving the acquired information and improving the estimation phase.
[0025] In one embodiment, the method includes the step of associating a phase with the measured and / or predicted radiation of the first type.
[0026] It is conceivable that the measurement data of at least one further sample position is used in establishing a phase for said further sample position. For example, the measurement data of said further sample position, i.e. the spectral information of the first type of radiation detected for said position, may still be relatively sparse. In that case, the measurement data can be compared with previously obtained data. For example, based on the previously assigned phases, it can be determined already whether the further sample position contains one or more known phases. In that case, the information on the at least one previously assigned phase can be used to compare the data of the further sample position and an estimated phase can be assigned to said further sample position.
[0027] In one embodiment, the establishing step comprises the use of a machine learning estimator. The machine learning estimator is able to predict the most likely dense spectrum for each sparse spectrum in the input. The proposed machine learning estimator can maintain an internal model of the phases and their spectra. With each input (sparse) spectrum, the model is updated and the spectrum can be evaluated to obtain a similarity score with the modeled phases. Finally, the spectra of the modeled phases can be combined according to these similarity scores to generate an estimate of the most likely dense spectrum given the input and the current state of the model.
[0028] The machine learning estimators may include one or more estimators selected from the group including Non-negative Matrix Factorization (NMF), Singular Value Decomposition (SVD), Independent Component Analysis (ICA), Latent Dirichlet Allocation (LDA) and K-Means, all of which are unsupervised learning techniques capable of identifying and characterizing components / clusters.
[0029] In one embodiment, one or more probabilistic phases are provided incrementally. In other words, the method may include using information about at least one previously assigned phase and its respective sample location to establish an improved probabilistic phase for at least one other of the plurality of sample locations.
[0030] In one embodiment, the machine learning estimator has an online learning function. This means that the machine learning estimator can learn from new examples and improve with further estimation. One way to achieve this is by means of a feedback loop, which can be provided by a full scan of samples and establishing a phase based on measurement data. The measurement data can be checked with previously estimated data, which allows for online learning of the machine learning estimator. This allows for a significant speedup, especially when the user generally investigates similar samples.
[0031] The machine learning estimator may include one or more cues for initialization. In one embodiment, an analysis of backscattered electrons is used. For this purpose, a second detector is provided for detecting a second type of radiation. In this embodiment, the first detector may be an EDS detector and said second detector is a BSE detector.
[0032] In one embodiment, the quantification of the integrated spectrum results is used as a cue for the initialization of the machine learning estimator. Other initialization means are also contemplated.
[0033] In one embodiment, the ratio between the number of measured sample locations and the number of estimated sample locations is in the range of 10:1 to 1:10, in particular the ratio is in the range of 1:2 to 1:10. Experimental measurements confirmed that the proposed method allows high quality estimates to be obtained from very sparse spectra. For example, estimates calculated from spectra with only 15 counts were comparable to raw spectra with 150 counts. This allows, for example, to significantly increase the mapping speed of EDS. This is especially true in the early stages of data collection. Moreover, even for larger data sets collected over longer acquisition periods, the quality of EDS maps constructed using the proposed estimator is significantly improved. This is because the regularization inherent in the estimator reduces the effect of stochastic noise in the input data.
[0034] In this regard, it is noted that in one embodiment, the sample locations may include a first group of sample locations each including a substantially complete spectral profile and a second group of sample locations each including a substantially incomplete spectral profile. A substantially incomplete spectral profile means that the spectral data is sparse in the sense that no spectral data exists or no final conclusion can be made about a particular facies. The method may include estimating the facies for at least one sample location from the second group of sample locations using sample locations from the first group of sample locations. The method may additionally or alternatively include estimating the facies for at least one sample location from the second group using partial data of at least one sample location of the second group of sample locations. The estimating step may include comparing the partial data of the second group with the spectral data from the sample locations of the first group. The comparison may be performed by a machine learning estimator as described herein. It is this embodiment in particular that allows the above ratio to be achieved, which in particular allows for fast and accurate data acquisition. The information regarding the at least one previously assigned phase at each sample location may include spectral information of a first type of radiation detected at the sample location.
[0035] According to one aspect, there is provided a charged particle microscope for inspecting a sample using a method according to one or more of the preceding claims, said charged particle microscope comprising: an optical column for focusing a beam of charged particles emitted from the charged particle source onto a sample, the optical column including a charged particle source, a final probe forming lens, and a scanner; a sample stage positioned downstream of the final probe forming lens and configured to hold the sample; and a first detector for detecting a first type of radiation arising from the sample in response to the incidence of charged particles emitted from the charged particle source; a control unit and a processing device connected to the first detector;
[0036] According to this aspect, the charged particle microscope is configured to carry out the method defined herein, the advantages of which have already been explained above.
[0037] The invention will now be explained in more detail on the basis of exemplary embodiments and the accompanying schematic drawings. [Brief description of the drawings]
[0038] [Figure 1] 1 shows a longitudinal section of a charged particle microscope according to a first embodiment of the invention;
[0039] [Diagram 2] 2 shows a longitudinal section of a charged particle microscope according to a second embodiment of the invention;
[0040] [Diagram 3] 1 shows a schematic diagram of one embodiment of the method according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0041] FIG. 1 (not to scale) is a very schematic depiction of an embodiment of a charged particle microscope M according to an embodiment of the present invention. More specifically, an embodiment of a transmission microscope M is shown, which in this case is a TEM / STEM (although in the context of the present invention it can usefully be an SEM (see FIG. 2) or, for example, an ion-based microscope). In FIG. 1, within a vacuum enclosure 2, an electron source 4 generates an electron beam B propagating along an electron-optical axis B′ and traversing an electron optical illuminator 6, which serves to direct / focus the electrons onto a selected portion of a sample S (which can, for example, be (locally) thinned / flattened). Also shown is a deflector 8, which can be used (among other things) to effect a scanning movement of the beam B.
[0042] The sample S is held in a sample holder H that can be positioned in multiple degrees of freedom by a positioning device / stage A, which moves a cradle A' to which the holder H is (removably) fixed; for example, the sample holder H can be equipped with fingers that can move (among other things) in the XY plane (see the Cartesian coordinate system shown; typically, movements parallel to Z and tilts relative to X / Y are also possible). Such movements allow different parts of the sample S to be illuminated / imaged / inspected by the electron beam B traveling along an axis B' (in Z direction) (and / or allow a scanning movement to be performed, as an alternative to beam scanning). If desired, an optional cooling device (not depicted) can be in intimate thermal contact with the sample holder H to maintain the sample holder H (and the sample S thereon) at, for example, a cryogenic temperature.
[0043] The electron beam B interacts with the sample S in such a way that it emits various types of "stimulated" radiation, including (for example) secondary electrons, backscattered electrons, X-rays, and light radiation (cathodoluminescence). If desired, one or more of these types of radiation can be detected with the aid of an analysis device 22, which can be, for example, a combined scintillator / photomultiplier or EDS (energy dispersive spectroscopy) module, in which case an image can be constructed using essentially the same principles as in an SEM. However, alternatively or supplementarily, the electrons that traverse (pass) the sample S, emerge / emit from the sample, and continue to propagate (substantially, but generally, deflected / scattered to some extent) along the axis B' can be investigated. Such a transmitted electron bundle is incident on an imaging system (projection lens) 24, which typically comprises various electrostatic / magnetic lenses, deflectors, correctors (such as stigmators), etc. In normal (non-scanning) TEM mode, this imaging system 24 can focus the transmitted electron beam onto a phosphor screen 26, which can be retracted / pulled out of axis B', if desired (as shown diagrammatically by arrow 26'). An image (or Fourier transform diagram) of (a part of) the sample S is formed on the screen 26 by the imaging system 24, which image can be viewed through a viewing port 28 located in a suitable part of the wall of the housing 2. The retraction mechanism of the screen 26 can be, for example, mechanical and / or electrical in nature, but is not depicted in the drawings.
[0044] As an alternative to viewing an image on screen 26, one can instead take advantage of the fact that the focal depth of the electron beam leaving imaging system 24 is typically quite deep (e.g., about one meter). As a result, various other types of analytical equipment can be used downstream of screen 26, e.g. -TEM camera 30. At the location of camera 30 the electron beam can form a still image (or Fourier transform diagram) which can be processed by controller / or processor 20 and displayed on a display device (not shown), such as a flat panel display. When not needed, camera 30 can be retracted / retracted (as shown diagrammatically by arrow 30') to move the camera off axis B'. - STEM camera 32. The output from camera 32 can be recorded as a function of the (X,Y) scanning position of beam B on sample S, and an image can be constructed that is a "map" of the output from camera 32 as a function of X,Y. Camera 32 differs from the pixel matrix characteristically present in camera 30 and may include a single pixel, e.g., 20 mm in diameter. Furthermore, camera 32 will typically have a much higher acquisition rate (e.g., 106 points / sec) than camera 30 (e.g., 102 images / sec). Again, when not required, camera 32 can be retracted / withdrawn (as shown diagrammatically by arrow 32') to move the camera off axis B' (such retraction is not required, for example, in the case of a doughnut-shaped annular dark field camera 32, where a central hole allows the electron flux to pass when the camera is not in use). As an alternative to imaging using the cameras 30 or 32, a spectroscopic device 34 can also be invoked, which can be, for example, an EELS module.
[0045] It should be noted that the order / position of items 30, 32, and 34 is not strict and many possible variations are contemplated. For example, spectroscopic device 34 could be integrated with imaging system 24.
[0046] In the embodiment shown, microscope M further comprises a retractable X-ray computed tomography (CT) module, generally indicated at 40. In computed tomography (also referred to as tomographic imaging), a source and a detector (at the opposite pole) are used to interrogate the specimen along different lines of sight to obtain transmitted views of the specimen from various viewpoints.
[0047] It should be noted that a controller (computer processor) 20 is connected to the various components shown via control lines (bus) 20'. This controller 20 may provide various functions such as synchronizing operations, providing set points, processing signals, performing calculations, and displaying messages / information on a display device (not shown). Of course, the controller 20 (schematically depicted) may be located (partially) inside or outside the housing 2 and may have a unitary or composite structure, as desired.
[0048] The skilled person will understand that the interior of the housing 2 does not have to be kept in an airtight vacuum. For example, in so-called "environmental TEM / STEM", a background atmosphere of a given gas is deliberately introduced / maintained in the housing 2. The skilled person will also understand that in practice, it may be advantageous, if possible, to limit the volume of the housing 2 so that it essentially encloses the axis B', taking the form of a small diameter tube (e.g. of the order of 1 cm in diameter) through which the electron beam used passes, but widening to accommodate structures such as the source 4, the specimen holder H, the screen 26, the camera 30, the camera 32, the spectrometer 34, etc.
[0049] The charged particle microscope M according to the invention, an embodiment of which is shown in Fig. 1, thus comprises a charged particle source 4, a final probe-forming lens 6 and a scanner 8, and comprises an optical column O for focusing a beam B of charged particles emitted from said charged particle source 4 onto a sample. The device further comprises a sample stage A, H, positioned downstream of said final probe-forming lens 6 and configured to hold said sample S. The device further comprises a first detector 22 for detecting a first type of radiation arising from said sample in response to the incidence of the charged particles B emitted from said charged particle source 4. In the embodiment shown, the first detector 22 is an analysis device 22, which may be a combined scintillator / photomultiplier or EDS (Energy Dispersive X-ray Spectroscopy) module, as described above. In a preferred embodiment, said first detector is an EDS. Furthermore, the device according to the invention comprises a control unit (i.e. controller / processor) 20 connected (by line 20') to said first detector 22 (schematically shown). According to the invention, said charged particle microscope M is adapted to carry out the method according to the invention which will be explained in more detail later.
[0050] Reference is now first made to Figure 2, which shows another embodiment of the apparatus according to the invention. Figure 2 (not to scale) is a very schematic representation of a charged particle microscope M according to the invention, and more specifically shows an embodiment of a non-transmission microscope M, in this case an SEM (although in the context of the invention, for example, an ion-based microscope may be equally useful). In the figure, parts corresponding to those in Figure 1 are indicated using the same reference numbers and are not described separately here. The following parts (among others) are added to Figure 1: -2a: a vacuum port that can be opened to introduce / remove parts (components, samples) into / from the interior of the vacuum chamber 2 or onto which, for example, auxiliary devices / modules can be mounted. The microscope M can be equipped with multiple such ports 2a, if necessary. - 10a, 10b: lenses / optical elements within the illuminator 6 shown diagrammatically. -12: Optionally, a voltage source that allows the sample holder H or at least the sample S to be biased (floated) to a certain potential with respect to ground. -14: Display such as FPD or CRT. 22a, 22b: a segmented electron detector 22a, comprising a number of independent detection segments (e.g., quadrants) arranged around a central aperture 22b (allowing the passage of the beam B). Such a detector can be used, for example, to investigate the (angular dependence of) the output (secondary or backscattered) electron flux emanating from the sample S.
[0051] Thus, the charged particle microscope M shown in FIG. 2 comprises an optical column O, including a charged particle source 4, a final probe-forming lens 6, 10a, 10b, and a scanner 8, for focusing a beam B of charged particles emitted from said charged particle source 4 onto a sample S. The device further comprises a sample stage A, H, positioned downstream of said final probe-forming lens 6 and configured to hold said sample S. The device further comprises a first detector 22 for detecting a first type of radiation arising from said sample in response to the incidence of the beam B of charged particles emitted from said charged particle source 4. In the illustrated embodiment, the first detector 22 is said analysis device 22, which may be a combined scintillator / photomultiplier or EDS (Energy Dispersive X-ray Spectroscopy) module, as described above. In an alternative embodiment, the first detector 22 may be a segmented detector 22a, 22b. In a preferred embodiment, said first detector is an EDS. Furthermore, the device according to the invention comprises said control unit 20 which is connected to said first detector 22 (by a line 20 ′).
[0052] The apparatus shown in Figures 1 and 2 may be used to test a sample using the method according to the invention. An embodiment of the method according to the invention is shown in Figure 3. In general, these methods all involve: Providing a charged particle beam B and a sample S, and scanning the charged particle beam B over a plurality of sample positions on the sample; detecting a first type of radiation from the sample in response to the beam B scanned over a plurality of sample locations using a first detector 22; using the spectral information G of the detected first type of radiation to assign a plurality of mutually distinct phases to the sample at the plurality of sample locations; providing, by a control unit 20, a data representation of said sample S comprising information about at least said plurality of sample positions and said phase.
[0053] According to the method defined herein, the general method further comprises: establishing an estimated phase for at least one other of the plurality of sample locations using information regarding the at least one previously assigned phase and its respective sample location; and assigning the estimated phase to the other sample locations.
[0054] Referring now to figure 3, a first embodiment of the method according to the invention is shown diagrammatically, which shows, on the left, a region 50 of a sample inspected by acquired data 54a, 54b and, on the right, a representation 150 of the acquired data.
[0055] An area 50 of the sample is scanned with a beam of charged particles. A first detector, for example an EDS detector as described with reference to Figures 1 and 2, is used to detect a first type of radiation from the sample in response to the beam scanned over the area 50 of the sample. These radiations are shown diagrammatically in Figure 3 by data points 54a (white dots) and 54b (black dots). Data point 54a represents a different radiation than data point 54b, meaning that the detector detects a different signal at these different locations. For example, this can be the result of a white dot 54a representing a first material, and a black dot 54b representing a second material (different from said first material). In this example, the sample contains only these two materials, i.e. the sample is considered to have only a first material (white) and a second material (black). Also shown are non-white and non-black dots 54c, 54d, which represent sample points that are still non-deterministic, for example because they contain sparse spectral information. As shown diagrammatically in the top left diagram, point 54c appears to resemble the spectrum of the first material (54a) and point 54d appears to resemble the spectrum of the second material (54b). Those skilled in the art will understand that the acquired data 54a, 54b can in principle have many values (not just black or white) and the representation values can be chosen arbitrarily depending on the desired application. The illustrated example is merely illustrative of a single possibility and is not intended to be limiting.
[0056] As shown from top to bottom on the left, the data information 54a, 54b is relatively sparse at the start and increases as the scan continues. Some particular points are relatively sparse in terms of the spectral information about them and therefore belong to an undetermined material, i.e., neither black nor white (54c, 54d). In the top step, the acquired data of the sample area 50 (left) is used to establish an estimated phase 154a, 154b of the sample 150 (right). The spectral information of the detected first type of radiation 54a, 54b is used to establish an estimated phase of the multiple sample locations. In the illustrated embodiment, the image representation 150 uses the measured and estimated points to represent the entire sample area 50 of the sample with data points. In other words, where no data exists, estimated data points are used. Where non-deterministic data is measured (points 54c, 54d), an estimated phase is also taken in the image representation 150, and thus the phase is associated with the measured first type of radiation. Where no data is present, the image representation has a phase associated with the expected first type of radiation. Note that while the acquired and estimated points are shown in a sparse pattern, more regular matrix-like patterns are also contemplated. In the illustrated embodiment, the data representation 150 (right) is an image representation, although other representations, i.e., data representations, are of course also contemplated.
[0057] FIG. 3 shows that the scanning of the sample can continue (left side, from top to bottom) and that more data of the sample 50 is coming in. As more data comes in, this means that the image representation 150 (right side) can be updated. Again, based on the measurement data already present, the sparse data points can be assigned a particular phase (in this example, either black or white). In this regard, it is conceivable that adjacent sample points are compared with non-deterministic sample points and, based on the comparison, a phase is assigned to the non-deterministic sample points (54c, 54d). With increasing time, and therefore input data, a more deterministic and reliable image representation 150 can be obtained.
[0058] During data acquisition and / or processing, it is conceivable to divide the scan area 50 of the sample S into a number of segments 51. As shown in FIG. 3, the scan area 50 is divided into four regular rectangular segments 51, although one skilled in the art will appreciate that the number, shape and regularity may vary. These segments can be used in establishing a putative phase for one or more of the multiple sample locations. In a particularly advantageous embodiment, a second detector is used to establish the segments. The second detector can be a BSE detector, as previously described, and contours identified in the EM image can be used for segmentation. This makes it possible to assign similar phases within a particular segment, or to ensure that the putative phase has a higher probability for a given phase measured within a particular segment.
[0059] It should be noted that further divisions into further segments are also contemplated, in particular subdivisions based on a larger number of input data points allowing for a more accurate image representation of the sample.
[0060] It is conceivable that the scanned area 50 is scanned multiple times to obtain a desired amount of data. Scanning multiple times may include scanning only a portion of the area 50 of the sample. For example, it is conceivable that based on a first scan (or a first set of scans), a region of interest and a region of non-interest are defined, and only the region of interest is scanned in a second scan (or a second set of scans). This increases the efficiency of the method. In particular, the region of interest may be defined using data obtained from a second type of radiation, i.e., the EM data may be used to define the region of interest that is scanned to obtain EDS data in particular.
[0061] To establish the estimated phase, it is advantageous to use a machine learning estimator. The use of a machine learning estimator makes it possible, in particular, to obtain a high ratio of measured sample positions to estimated sample positions. As can be seen in the top right diagram of Fig. 3, the ratio between the number of measured sample positions and the number of estimated sample positions is about 1:10, and for each measured sample position, it is possible to estimate a total of 10 other positions. The ratio can be in the range of 10:1 to 1:10, more specifically in the range of 1:2 to 1:10. The machine learning estimator allows predicting the most likely dense spectrum for each sparse spectrum in the input. The proposed machine learning estimator is able to maintain an internal model of the phases and their spectra. With each input (sparse) spectrum, the model is updated and the spectrum can be evaluated to obtain a similarity score with the modeled phase. In other words, during data acquisition (top to bottom in Fig. 3), each newly acquired data point (left side) gives information about the accuracy of the previously determined image representation (right side). It is therefore possible to update the machine learning estimator during the use of the charged particle device, which makes it very effective. As previously mentioned, the machine learning estimators may include one or more estimators selected from the group including Non-negative Matrix Factorization (NMF), Singular Value Decomposition (SVD), Independent Component Analysis (ICA), Latent Dirichlet Allocation (LDA) and K-Means, all of which are unsupervised learning techniques that may identify and characterize components / clusters.
[0062] The method has been described above by means of several embodiments. The desired protection is given by the appended claims.
Claims
1. 1. A method of inspecting a sample using a charged particle microscope, comprising: Providing a charged particle beam and a sample; scanning the charged particle beam across a plurality of sample locations on the sample; detecting a first type of radiation from the sample in response to the beam scanned over a plurality of sample locations using a first detector; assigning a plurality of mutually distinct phases to the sample at the plurality of sample locations using spectral information of the detected first type of radiation; providing, by a control unit, a data representation of the samples comprising information regarding at least the plurality of sample locations and the phase; The method comprises: establishing an estimated phase for at least one other of the plurality of sample locations using information regarding the at least one previously assigned phase and its respective sample location; assigning the estimated phase to the other sample locations, comprising comparing adjacent sample points to non-deterministic sample points and assigning a phase to the non-deterministic sample points based on the comparison; The method, wherein the establishing step includes the use of a machine learning estimator to predict a most likely dense spectrum for each sparse spectrum in the input, the sparse spectrum is used to update a model, and the spectrum is evaluated to obtain a similarity score with the modeled phase.
2. The method of claim 1 , comprising the step of associating a phase with the measured and / or predicted first type of radiation.
3. 3. The method of claim 2, wherein the machine learning estimators comprise one or more estimators selected from the group comprising non-negative matrix factorization (NMF), singular value decomposition (SVD), independent component analysis (ICA), latent Dirichlet allocation (LDA) and K-means.
4. 4. The method of claim 1, wherein the ratio of the number of measured sample locations to the number of estimated sample locations is in the range of 10:1 to 1:10, or said ratio is in the range of 1:2 to 1:
10.
5. The method of any one of claims 1 to 4, wherein the information relating to the at least one previously assigned phase at each sample location comprises spectral information of a first type of radiation detected at the sample location.
6. The method according to any one of claims 1 to 5, wherein the first detector is an EDS detector.
7. detecting a second type of radiation from the sample in response to the beam scanned over the plurality of sample locations using a second detector; A method according to any preceding claim, further comprising using the detected second type of radiation to establish the estimated phase for the at least one other of a plurality of sample locations.
8. The method according to any one of claims 1 to 7, wherein the detected first type of radiation and / or the detected second type of radiation is / are used to divide at least a portion of the scanned area of the sample into a plurality of segments.
9. The method of claim 8 , further comprising using the segment to establish the estimated phase for at least one other of a plurality of sample locations.
10. A charged particle microscope for examining a sample using the method according to any one of claims 1 to 9, comprising: an optical column for focusing a beam of charged particles emitted from the charged particle source onto a sample, the optical column including a charged particle source, a final probe forming lens, and a scanner; a sample stage positioned downstream of the final probe forming lens and configured to hold the sample; and a first detector for detecting a first type of radiation arising from the sample in response to the incidence of charged particles emitted from the charged particle source; a control unit and a processing device connected to the first detector, Charged particle microscope, wherein the charged particle microscope is configured to carry out the method according to any one of claims 1 to 9.
11. A charged particle microscope as described in claim 10, comprising a second detector for detecting a second type of radiation from the sample in response to the beam scanned over a plurality of sample locations, the charged particle microscope being configured to perform the method of any one of claims 1 to 9.
Citation Information
Patent Citations
Electron beam analyzer
JP2002323463A
Clustering multi-modal data
JP2013019900A
Method and system for spectroscopic data analysis
US20060028643A1
