Material analysis using multiple detectors

Abstract

A detector module for use in an apparatus for analyzing a sample is provided, comprising a plurality of X-ray sensor elements and one or more electron sensor elements, and adapted to be positioned below a pole piece of an electron beam assembly of the apparatus, from which an electron beam generated by the assembly emerges, in use, toward the sample, and the detector module receives X-rays and backscattered electrons generated by interaction between the electron beam and the sample. Each of the plurality of X-ray sensor elements is configured to monitor the energy of an individual received X-ray photon, and the plurality of X-ray sensor elements have a total active area greater than 20 mm2. The radial length of the detector module relative to the electron beam axis is less than 10 mm over at least a first portion of the detector module. Apparatus and methods for analyzing a sample are also provided.

Description

[Technical Field]

[0001] The present invention relates to an apparatus and method for the analysis of materials in a scanning electron microscope that combines information from multiple X-ray and electron detectors to improve navigation of a sample at high speed and spatial resolution and to provide its elemental analysis. [Background technology]

[0002] Referring to FIG. 1 , for X-ray analysis in an electron microscope (EM) 100, an X-ray spectrum is measured by sensing and measuring the energy of individual X-ray photons emitted by a sample 101 when it is struck by a focused electron beam 102. (Note that, by convention, the electron beam travels vertically downward toward the sample, and this is the context for words like “downward” and “upward.” In practice, the electron beam can be directed in any direction, including vertically upward.) Each X-ray photon is a high-energy particle, and the energy is converted to an electric charge, typically using a solid-state detector 105. The charge is measured so that counts can be recorded, and a histogram of the recorded measurements represents a digitized X-ray energy spectrum. Peaks characteristic of chemical elements are identified in the X-ray energy spectrum, and the intensities of those peaks can be used as the basis for determining the elemental content of the material directly beneath the electron beam 102.

[0003] The X-ray detector 105, the final pole piece 104 of the electron microscope 104, and the sample 101 are typically all located within the same vacuum chamber. The vacuum is primarily required so that electrons can be accelerated to energies of several keV and focused into a narrow beam without scattering off gas molecules. However, alternative configurations exist in which the electron beam can be focused within a vacuum region while the sample is in a region of higher pressure. The X-ray detector can be located in the same vacuum region as the electron beam or in a region of higher pressure. In addition to the X-ray signal, signals derived from backscattered electrons (BSE) from the sample are also useful for distinguishing between different materials because the percentage of electrons backscattered from a material increases with the average atomic number (Z) of that material. Consequently, a backscattered electron detector (BSED) is often positioned above the sample 101 and below the pole piece 104. The BSED detector 106 typically comprises one or more sensor segments arranged around a central hole through which the focused beam 102 passes to reach the sample. This positioning is designed to maximize the collection solid angle subtended by the BSED segments at the point where the focused beam strikes the sample, i.e., the "probe spot," thereby maximizing the BSE signal. An additional detector, typical of the "Everhart-Thornley" type and mounted on one side of the chamber, is used to detect secondary electrons generated in the sample and emerging from the surface, i.e., generating the "SE" signal. The SE signal is typically much stronger than the BSE signal and is very sensitive to surface orientation relative to the incident beam.

[0004] If a focused incident beam is rastered or scanned in some other pattern by magnetically or electrostatically deflecting the beam to sequentially position it over a 2D grid of pixel locations on the sample surface, the SE or BSE signal at each location can be used to construct a digital image that can be displayed on a monitor and provides an expanded view of the sample. This is a well-known operating principle for scanning electron microscopes (SEMs), and the SE image is particularly useful for navigating around the sample because it shows the surface topography. If any topography is present on the sample, tilted surface facets will also generate a BSE signal that depends on the facet orientation and will be stronger in certain directions. The BSE signal is stronger for regions of the sensor in the direction of the facet tilt away from the normal to the electron beam and weaker for regions in the opposite direction. This sensitivity to surface tilt creates a "topographical contrast" in the image that interferes with the "atomic number contrast" or "Z contrast" caused by changes in the material's composition within the scanned area. To minimize the directional effect of topography on the BSE signal, it is essential that the entire sensitive area of ​​the BSE detector is symmetrically positioned with respect to the incident electron beam. Ideally, the sensitive area would be a disk with perfect rotational symmetry about the central hole, but there are some benefits to using multiple independent sensor segments to make up the entire sensitive area while maintaining rotational symmetry. A "four-quadrant" arrangement with four-fold rotational symmetry is shown, for example, in the "Micron Semiconductor" catalog: http: / / 79.170.44.80 / micronsemiconductor.co.uk / wp-content / uploads / 2017 / 03 / MSL-OEM-Catalogues.pdf It is common, as can be seen in [1]. It is known that when all segments of a multi-segment BSE detector are used to collect electrons symmetrically with respect to the beam, the "atomic number" contrast becomes dominant over the topographic contrast. See, for example, the Wikipedia entry: https: / / en.wikipedia.org / wiki / Scanning_electron_microscope#Detection_of_backscattered_electrons. That is, positioning the BSE detector 106 directly below the pole piece not only provides good collection efficiency for the BSE signal, but also allows a signal to be collected that is more representative of the average atomic number of the material than the local surface tilt. The BSE signal at each pixel location can be used to construct an image where each pixel intensity indicates the atomic number of the material at that location.

[0005] The same position 106 can also be used to maximize the collection solid angle for the annular X-ray detector. However, replacing the BSED with an X-ray detector at position 106 would remove the ability to detect BSE signals, which are insensitive to topography. Soltau et al. (Microsc. Microanal. 15 (Suppl. 2), 2009, 204 5) suggest a way to overcome this problem by surrounding the ring of sensor segments for BSE with an outer ring of X-ray sensor segments, as seen in Figure 2.

[0006] The BSE detector segments are closest to and symmetrically arranged around the central hole, while the X-ray detector segments, also symmetrically arranged, are further from the central hole.

[0007] While this arrangement places both the BSED and the X-ray sensor segments close to the sample and maintains four-fold rotational symmetry for the entire sensitive area of ​​the BSED, the collection solid angle for the X-ray sensor segments is compromised by being positioned farther from the central hole to accommodate the BSED sensor. Furthermore, if the individual X-ray sensor segments are of the "silicon drift detector" (SDD) type (see, for example, https: / / en.wikipedia.org / wiki / Silicon_drift_detector), for a given area, the optimum response time is achieved for circular sensor segments, and therefore the elongated shape of the X-ray sensor segments in the arrangement of Figure 2 is not optimal with respect to response time for a given sensitive area.

[0008] The X-ray detector segment will also be sensitive to high-energy backscattered electrons, and since there are orders of magnitude more BSEs than X-rays impinging on the sensor segment, the BSE signal will likely overwhelm the normally small X-ray signal and prevent the acquisition of a useful X-ray spectrum. Therefore, a filter material with an appropriate thickness must be inserted between the sample and the sensor to prevent the transmission of the highest-energy BSEs while allowing low-energy X-rays to pass to the sensor. Liebel et al. (Microsc. Microanal. 20 (Suppl. 3), 2014, pp. 1118-9) suggest an arrangement using a combined BSE and X-ray sensor as shown in Figure 2, with a filter placed in front of the X-ray sensor as shown in Figure 3.

[0009] In a scanning electron microscope, incident electrons are typically accelerated to energies of 20 keV, and a filter such as a 6-micron thick Mylar foil may be required to block BSEs up to this energy. The X-ray transmission through such a foil as a function of X-ray photon energy is shown in Figure 4.

[0010] With such a filter in place, the detector would be insensitive to x-rays below about 1 keV in energy, such that elements with characteristic line emissions below 1 keV would be difficult to distinguish in the x-ray spectrum. For example, the elements Be, B, C, N, O, F, Ne, and Na would be very difficult to detect with such a filter in place.

[0011] Yet another problem with this arrangement is that because the module enclosing the combined BSE and X-ray sensor has a larger diameter than a typical BSE detector, the module may partially obscure the line of sight 103 to the sample for a detector or any other accessory device mounted on one side of the microscope, as shown in FIG. 5.

[0012] Examples of accessory devices that could be mounted in a side position such as 105 within an electron microscope and require line-of-sight to the sample include X-ray detectors, electron detectors, cathodoluminescence detectors, mass spectrometers, micromanipulators, gas injection devices, and lasers. An example of an existing X-ray detector module held below the pole piece is the "Rococo Preamplifier Module" product from PNDetector (https: / / pndetector.com / w / wp-content / uploads / 2018 / 08 / Rococo_2017.pdf). This module does not have any BSE sensors, but when in position for analysis, its perimeter extends 10.5 mm away from the electron beam axis even when it is oriented to minimize shielding of side-mounted accessories.

[0013] To maximize the collection solid angle for the segments, they need to be as close as possible to the electron beam axis, which ideally means that the central hole should be as small as possible. However, if the central hole of the module is made too small, this makes it difficult to align the hole to ensure that the electron beam passes through unobstructed. Furthermore, for example, if the material surrounding the hole is magnetic or if dirt accumulates on the inner surface of the hole, causing a buildup of static charge, this may interfere with the focusing of the beam or the deflection of the focused beam as it is scanned over the area to be imaged on the surface of the sample. Therefore, the central hole must be large enough to avoid any interference with the focusing or deflection of the incident electron beam, which could distort the image.

[0014] When a sample is observed in an SEM, the area scanned by deflecting a focused electron beam to obtain material analysis mapping data is typically less than 0.1 mm wide. The SEM operator typically sees a magnified image of the scanned area based on SE signal data acquired while scanning the beam. However, a typical sample support stub is at least 10 mm in diameter, making it difficult for the operator to know exactly where the small scan area is located on the much larger sample support. The walls of an SEM are opaque, and when the SEM is operating, the sample is inside a vacuum region; therefore, the operator cannot see the entire sample support using the observation window. A so-called "chamber scope," i.e., an infrared-sensitive TV camera with an infrared light source, is typically mounted on one side of the sample tube and is used to provide the operator with an optical image of the inside of the SEM chamber. While the chamber scope is very useful for determining the position of the sample stage relative to other detectors and accessories in the chamber, it does not provide a view of the sample surface at a suitable orientation and magnification to help the operator locate the area scanned by the electron beam relative to the surrounding sample surface. [Prior art documents] [Patent documents]

[0015] [Patent Document 1] US 8,049,182 B2 [Patent Document 2] US 7533000 [Patent Document 3] PCT / GB2011 / 051060 [Patent Document 4] US5357110 [Non-patent literature]

[0016] [Non-Patent Document 1] Soltau et al., Microscopy and Microanalysis, Vol. 1, 5 (Suppl. 2), 2009, pp. 204-205 [Non-patent document 2] Liebel et al., Microscopy and Microanalysis, Vol. 2, No. 0 (Suppl. 3), 2014, pp. 1118-1119 [Non-patent document 3] D. Samak et al., "3D Reconstruction and Visualization of Microstructure Surfaces from 2D Images," CIRP Annals, Vol. 56, No. 1, 2007, pp. 149-152 [Non-patent document 4] Statham et al., Microscopy and Microanalysis, Vol. 1, No. 9 (Suppl. 2), 2013, p. 752 [Non-Patent Document 5] Statham, Journal of Research of the National Institute of Standards and Technology, Vol. 1, No. 7, pp. 531-546 (2002) [Non-patent document 6] Goldstein et al., "Scanning Electron Microscopy and X-ray Microanalysis," ISBN: 0-306-47292-9 [Non-Patent Document 7] Manjon et al., "Multispectral MRI de-noising using non-local means," Proc. MIUA'07, pp. 41-45, Aberystwyth, Wales, 2007 Summary of the Invention [Problem to be solved by the invention]

[0017] What is needed is a solution that provides a large solid angle for both BSE and X-ray detection, does not interfere with the electron optics, provides good material atomic number contrast for BSE images while minimizing topographical contrast, avoids obscuring the line of sight for accessory devices mounted on one side of the electron beam, has a fast response time and the ability to distinguish between materials of different compositions, and helps the operator navigate to areas of interest on the sample. [Means for solving the problem]

[0018] According to a first aspect of the present invention there is provided a detector module for use in an apparatus for analysing a sample, the detector module comprising a plurality of X-ray sensor elements and one or more electron sensor elements, the detector module adapted to be positioned below a pole piece of an electron beam assembly of the apparatus from which, in use, an electron beam generated by the assembly emerges towards the sample, such that the detector module receives X-rays and backscattered electrons generated by interaction between the electron beam and the sample. Each of the plurality of X-ray sensor elements is configured to monitor the energy of an individual received X-ray photon, and the plurality of X-ray sensor elements are spaced apart by 20 mm 2and the radial extent or length of the detector module relative to the electron beam axis is less than 10 mm over at least a first portion of the detector module.

[0019] It has been found that using detector modules having this configuration and including sensor elements arranged in this manner can provide significant improvements in sample navigation and analysis in electron beam instruments. This is achieved by the above-described arrangement, which allows a large total collection solid angle to be maintained for both the X-ray and backscattered electron (BSE) detectors, while ensuring that the modules comprising these detectors do not block the line of sight to the spot where the electron beam strikes the sample (the "probe spot") for any accessory devices that may be positioned to one side of the electron beam assembly.

[0020] The term "module" as used in disclosing the present invention is intended to refer to a device, particularly one adapted for use with an apparatus, typically with an electron beam instrument such as an electron microscope. The term "module" can therefore be used interchangeably with the term "device" in the context of the present disclosure. Preferably, a module is provided as a self-contained set of components suitable for inclusion in such an instrument. The module is preferably removably mountable within the apparatus or adapted to be removably mounted on a support or mounting structure or mechanism thereon. The term "detector module" can be understood as a reference to a device or set of components adapted for the purpose of detecting particles in an electron beam instrument, particularly by means of sensor elements.

[0021] It will be understood that in the context of this disclosure, the electron beam assembly can be considered as configured to generate a focused electron beam, and the detector module can be described as a module for detecting particles, the expression "particles" being used in this disclosure to refer to both electrons and X-ray photons.

[0022] Typically, each of the x-ray sensor elements is configured to detect and be sensitive to, or output a signal in response to, incident x-rays, and similarly, each of the electron sensor elements is typically configured to similarly respond to incident electrons (preferably backscattered electrons).

[0023] A module adapted to be positioned below the pole piece of an electronic assembly of an analytical instrument can be understood as a module suitable for being positioned in such a position during use. In other words, the module typically has a size, shape, and sensor element configuration that allows the sensor to receive x-rays and electrons generated by a sample in the instrument during use. Typically, the module is therefore adapted to operate while positioned between the pole piece and the sample, particularly closer to the former than to the latter. The module is also typically shaped to allow passage of an electron beam from the pole piece to the sample during use. In preferred embodiments, this configuration involves the module's shape defining a space through which the electron beam can pass unimpeded. Typically, the module is shaped to encompass some or all of this space, thereby also encompassing the electron beam during use. In exemplary embodiments, positioning a detector module below the pole piece can be understood as the module being in a position below the pole piece that allows the electron beam to pass to the sample. This arrangement can be achieved even in embodiments in which the sensor elements are positioned with at least two-fold rotational symmetry about the beam axis, as disclosed in detail below. Typically, in embodiments where a space is defined between the pole piece and the detector module in use, that space is as small as is practical for the mechanical mechanisms holding the module. The term "lower" as used in this context typically refers to being farther along the direction of travel of electrons in the electron beam in use. The term does not necessarily refer to being further down vertically or at any particular orientation relative to gravity.

[0024] Typically, the electron beam assembly takes the form of an electron beam column. The pole piece below which the detector module is mounted or mountable is typically the final lens pole piece, which can be understood as the final lens pole piece of the electron beam assembly or column through which the electron beam passes before striking the sample.

[0025] It will be appreciated that each of the plurality of X-ray sensor elements can be used to monitor X-ray photon energy as described above. The sensors can be adapted to measure or otherwise provide or output a signal or data representative of the detection or energy of an received X-ray photon. In this manner, the sensor element can be configured to detect each received X-ray photon and monitor the energy of each received X-ray photon. The total active area of ​​the plurality of X-ray sensor elements can be considered to be the combination of all of the plurality of X-ray sensor elements. The active area will typically be understood as the area of ​​the sensor where the sensor element is responsive or sensitive to particles incident on its surface or a portion thereof. In a preferred embodiment, the plurality of X-ray sensor elements has an area of ​​30 mm 2 greater than 40 mm, preferably 2 has a total effective area greater than

[0026] The radial extent or length of the detector module can be considered, for example, as the extent or length of the module in a direction perpendicular to the electron beam axis. The radial extent or length of the module of the first portion relative to the electron beam axis, and any configuration or shape typically defined in terms of the electron beam axis, can be understood to be defined relative to the position and / or orientation of the electron beam while the module is mounted or otherwise disposed in an operating position or position within an apparatus that directs the beam onto a sample during use. Typically, the module and / or analytical apparatus, such as an electron microscope, are configured so that the module can be mounted in a defined orientation and position relative to the apparatus, or at least relative to the sample and / or the final pole piece of the electronic assembly. As alluded to above, this defined relative arrangement of the module and the apparatus that uses it to detect X-rays and electrons from a sample is preferably one in which the module is mounted below the pole piece to receive the particles. Thus, it is typically the position and orientation of the electron beam axis when the module is mounted within an apparatus for analysis, relative to which the above-mentioned radial extent or length of the first portion is defined. It will be appreciated that this radial extent or length typically defines the extent to which the module obscures the line of sight from the part of the sample on which the beam impinges, in use, to a location around the electron beam assembly where further detectors or accessory devices may be mounted. Providing a first part of the detector module whose extension is less than 10 mm advantageously ensures that some line of sight is maintained by the module and is not obscured.

[0027] Alternatively or in addition to the above-described features regarding the first portion of the detector module and its radial length and the total active area of ​​the plurality of X-ray sensor elements, the total active area of ​​all X-ray sensor elements is 30 mm 2and the minimum radial distance from the electron beam axis to the protruding periphery of the module on the side facing the sample may be less than 10 mm. A specific application of such a limit to the extension of the module on the side facing the sample is that in typical embodiments, it is the side or main surface of the detector module that is closest to the sample that determines the degree to which the detector module obscures the probe spot from accessory devices disposed thereon.

[0028] The electron beam axis can be considered to be an axis that is collinear with the electron beam, or an axis that coincides with at least two centroids of the electron beam that are typically spaced apart along the path of the beam.

[0029] The aforementioned first portion of the detector module can typically be defined as a portion of the module where the radial length of the module is less than a predetermined radial length value. The required maximum of 10 mm for the radial length of this portion advantageously provides a module shape or configuration that allows unrestricted visibility or access to the probe spot. As described in more detail below, in some embodiments, the minimum radial length of the modules within the first portion can be less than this predetermined value for at least a portion of the first portion, typically the central portion.

[0030] The first portion can also be defined as the portion of the module that lies between a first plane and a second plane, the intersection of which is collinear with the electron beam axis in use. The first portion, and in particular its size, can be defined by its angular or circumferential extension, e.g., the angle between the first and second planes, or the angle between their respective normal vectors, or the length of the outer periphery, edge, or peripheral surface (i.e., distal to the beam) of the first portion of the detector module. The angular or circumferential extension of the first portion, which is necessary in some embodiments to avoid obstruction of the line of sight to the accessory device, can be set depending on the type of accessory device intended for use with the module in the electron beam instrument. That is, some accessory devices may only require a small "cutout" that provides an unobscured line of sight to the probe spot. In such cases, the "width" or angular extension of the first portion can be minimized. On the other hand, some devices may benefit from a wider portion, in which the radial length of the module is smaller than the outward extension of the first portion. In some embodiments, this angular extension is insignificant for portions where the radial length of the module is less than the maximum radial length value required within the first portion, i.e., 10 mm. This value is typically a local maximum in that it represents a "maximum" value only for the first portion, but it will be understood that in most embodiments, this portion forms a portion of the module having a radial length that is smaller than the rest of the module. A first portion having an insignificant angular extension can be understood as one in which the first and second planes are coplanar or substantially so.

[0031] In such an arrangement, the first portion can thus correspond to a radial minimum on any side where the radial length increases. The shape of the module can be such that this increase follows some continuous function, so that the periphery of the module can have a linear or curved section where this minimum exists. Examples of such shapes are provided later in this disclosure. In another embodiment, the first portion can be bounded on one or both sides by an acute or discontinuous periphery shape, beyond which the radial length of the module increases, for example, defining a slot-shaped gap. In these various embodiments, the defined maximum radial length value can preferably be 7 mm or more preferably 5 mm. In some embodiments, on one or both (circumferential) sides of the first portion, another portion can be defined where the radial length of the module is greater than the radial length value for the first portion but does not exceed a second radial value that is less than the maximum and / or average radial length value for the entire periphery or periphery of the module.

[0032] In addition to the first portion, the detector module may further comprise a second portion in which the radial length of the detector module is also minimized, in particular also less than 10 mm.

[0033] In some embodiments, the detector module comprises an outer edge, typically between an outer surface of the detector module and a surface of the detector module that faces the sample in use. In such embodiments, the minimum radial distance between the electron beam axis and the outer edge is preferably less than 10 mm.

[0034] In some preferred embodiments, the radial length of the first portion of the detector module, i.e., the smallest radial distance from the electron beam axis to the periphery of the module on the side facing the sample, is preferably less than 7 mm, more preferably less than 5 mm.

[0035] Typically, at least half, and preferably more than half, of the total active area of ​​the plurality of X-ray sensor elements is less than 6 mm from the electron beam axis when in use. It is advantageous to provide a predetermined proportion of the active X-ray sensor area within a predetermined radius relative to the beam axis. More preferably, within the predetermined proportion of the active area, which may be half or more than half as described above, the radial distance from the electron beam axis may be less than 6 mm, preferably less than 5 mm, more preferably less than 3 mm.

[0036] In addition to the minimum overall active area mentioned above for the X-ray sensor element, in a preferred embodiment, the one or more electronic sensor elements are 2 In a preferred embodiment, the module includes a plurality of electronic sensor elements. However, it will be appreciated that sensor arrangements can be provided in which two or more electronic sensor elements are connected together by a connecting sensor portion or are fabricated as a single sensor element.

[0037] It has been found to be advantageous to arrange the sensor elements in the detector module so that the X-ray sensor has limited rotational symmetry, while the electronic sensor elements have the same or greater rotational symmetry. Thus, preferably, the X-ray sensor elements are arranged with no more than two-fold rotational symmetry about the electron beam axis, and the electronic sensor elements are arranged with at least two-fold rotational symmetry about the electron beam axis. This symmetry is typically predetermined in terms of the orientation and position of the electron beam axis relative to the module in use, i.e., when the module is installed in an analytical device. These requirements will be understood to refer to the appearance of multiple X-ray sensor elements being identical in two or fewer different orientations about the beam axis, and to the appearance of multiple electronic sensor elements being identical in two or more different orientations about the beam axis.

[0038] In some embodiments, the detector module's performance can be improved by positioning a portion of the active area of ​​the electronic sensor element farther from the electron beam axis than the farthest portion of the X-ray sensor element relative to the electron beam axis, and by positioning a portion of the active area of ​​the electronic sensor element closer to the electron beam axis than the closest portion of the X-ray sensor element relative to the electron beam axis. Thus, in such preferred embodiments, the radial distance from the electron beam axis to a first location within the active area of ​​the electronic sensor element is greater than the maximum radial length of the active area of ​​the X-ray sensor element relative to the electron beam axis, and the radial distance from the electron beam axis to a second location within the active area of ​​the electronic sensor element is less than the minimum radial distance between the electron beam axis and the active area of ​​the X-ray sensor element. Similar to the symmetry disclosed above, these radial distances are typically predetermined in terms of the orientation and position of the electron beam axis relative to the module during use, i.e., when the module is installed in an analyzer.

[0039] In such an embodiment, the detector module is preferably configured to amplify signals from a first portion of the active area of ​​the plurality of electronic sensor elements positioned furthest from the electron beam axis before adding the signals to signals from a second portion of the active area of ​​the plurality of electronic sensor elements, which may be considered as other electronic sensor elements different from the first portion, where the portion furthest from the beam axis may refer to being located at a radial distance from the beam axis at least as large as the maximum radial length of the X-ray sensor element area.

[0040] As shown in the examples provided later in this disclosure, detector modules are typically shaped to at least partially surround a focused electron beam in use, with a gap or, more preferably, an opening in the module through which the beam can pass unimpeded. As explained in more detail later in this disclosure, arranging this gap or opening to have a smaller diameter may result in a greater degree of unwanted contamination of the detector module through use. Therefore, it may be advantageous to apply a minimum size to this gap or opening. Its value is preferably selected to be small enough to avoid an undesired reduction in the solid angle subtended by the detector area predetermined by the probe spot. Preferably, the detector module includes an opening through which the electron beam is arranged to pass in use, with an inner diameter greater than 1.0 mm, or preferably greater than 1.5 mm, or more preferably greater than 2.5 mm. In some embodiments, the gap or opening is non-circular; in such cases, the inner diameter referred to above can be defined as the minimum diameter of the opening. Typically, the opening or hole is centered about the center of gravity of the detector module or its major surface. A beam passing through an aperture in a detector module can be understood as a beam passing from one side of the module to the other.

[0041] When an accessory device is included, in such embodiments the accessory device includes or is one or more of the following: an X-ray sensor, a backscattered electron sensor, a cathodoluminescence sensor, a micromanipulator, a gas injection device, a laser, and an X-ray sensor equipped with or fitted with an electron trap.

[0042] The embodiments described in this disclosure can advantageously include a filter material covering to prevent backscattered electrons from adversely affecting the X-ray sensor during use. Preferably, therefore, the detector module further includes at least one material layer disposed over at least a portion or all of the active areas of the plurality of X-ray sensor elements. The at least one material layer can be adapted to partially or completely block transmission of any one or more of electrons, visible light, and infrared light, while allowing partial or complete transmission of X-rays having energies within a first energy range or energy band. This band can correspond to a range of X-ray photon energies identified or determined to be useful for analysis. Typically, the layer material and / or its thickness is configured to have the required X-ray transparency and the required opacity to unwanted radiation or particles. The first range can correspond to X-ray energies useful for elemental analysis. This can be a predetermined range of energies. The numerical boundaries of this range are not necessarily known. However, the layer is preferably configured to allow X-rays in at least the desired energy band or sub-bands thereof.

[0043] In other words, in some embodiments, the X-ray sensors in the module can be covered with one or more layers of material to block the transmission of electrons or visible or infrared light to the respective sensors, while still allowing X-rays with energy useful for analysis to pass through.

[0044] In some embodiments, the layer of material, or at least one of the layers of material in those embodiments having multiple layers, is in the form of a coating applied to a surface of the portion of the active area of ​​the multiple x-ray sensor elements. The surface can be considered to be the top surface, i.e., the surface configured to receive x-rays to detect and monitor x-ray photons and their energy. In various embodiments, one or more of the layers of material can alternatively or additionally be provided as a self-supporting film or sheet.

[0045] A coating applied to the sensor element surface can prevent light or some electrons from reaching the sensor. However, in some such embodiments, one or more material layers can further comprise a separate filter material in front of the sensor to block more electrons. If the filter material is an electrical insulator, it is beneficial to provide an additional conductive coating that is substantially grounded to prevent this insulating layer from becoming charged due to incident electrons. The material layer, or at least one of the layers in embodiments with multiple layers, for example, stacked layers, is preferably electrically grounded. Preferably, it is also made conductive to prevent the potential of the material layer, layer, or stack from increasing due to electrostatic charging caused by incident electrons, i.e., electrons irradiating or impacting the layer. The aforementioned conductivity of a layer typically refers to the material being sufficiently conductive to allow the layer to dissipate charge and prevent static charge buildup within the layer.

[0046] The detector module, in some embodiments, may further comprise a cooling controller configured to maintain the plurality of electronic sensor elements within a predetermined temperature range during use. The controller may more preferably be configured to maintain the plurality of electronic sensor elements at a predetermined temperature. Maintaining the temperature of these elements within a desired temperature range may be advantageous for certain operating regimes of the analytical instrument.

[0047] Typically, in such embodiments, the cooling controller is configurable to maintain the plurality of electronic sensor elements at a temperature range selected from a plurality of predetermined temperature ranges, although alternatively, each of the temperature ranges may be a predetermined temperature value rather than a finite range. Each of the ranges or values ​​may correspond to an optimal operating temperature range or value for a respective operating mode of the electron beam assembly. Preferably, they may correspond to a respective mode of an electron microscope having the device or detector module.

[0048] Thus, the temperature of the sensor can be configured in some embodiments to optimize performance and suit different operating modes of the electron microscope.

[0049] In some embodiments, each of the plurality of X-ray sensor elements can be positioned behind or aligned with an aperture in a conductive plate configured to function as a first electrode of a capacitive sensor adapted to monitor the proximity of the sample and / or sample holder. Such plates are typically configured as such by monitoring the capacitance between the plate and the sample or holder to derive information indicative of the separation or distance therebetween or its rate of change. In these arrangements, an aperture is typically provided to allow X-rays generated at the sample by interaction with the electron beam to pass through the aperture and be received by the sensor element.

[0050] Typically, the detector module further comprises an optical camera arranged to have a field of view of the sample of at least 10 mm across the sample plane when the camera-sample distance is less than 10 mm, and optionally the depth of field of the camera may be sufficient such that the width of the field of view can be increased to at least 20 mm, preferably more than 60 mm, by increasing the camera-sample distance. An optical camera may for example be understood as an image sensor configured to operate in or use the visible part of the electromagnetic spectrum.

[0051] In some embodiments, the detector module may be further configured to perform a method for displaying the location of the area being analyzed under the electron beam in an optical image of the sample to be inspected, by positioning the surface of the cathodoluminescent sample at a particular working distance from the final lens pole piece of the electron microscope, acquiring a digitized optical image with the illumination light source turned off while the focused electron beam is incident on the sample, determining the position coordinates in the optical image of the spot center of the light emitted by cathodoluminescence, positioning the surface of the sample to be inspected at the same particular working distance as the cathodoluminescent sample and acquiring the digitized optical image with the light source illuminating the sample, and the optical image of the sample to be inspected is shown on a visual display in which the location of the electron beam analysis area is highlighted centered on the position coordinates obtained from the optical image obtained from the cathodoluminescent sample.

[0052] In another embodiment, the detector module may be further configured to perform a method for displaying the location of the area being analyzed under the electron beam within an optical image of the sample to be inspected, by focusing the face of a small, recognizable feature at a particular working distance from the final lens pole piece of the electron microscope and acquiring an electron image with the feature centered in the field of view, acquiring a digitized optical image with the illumination source on to determine the position coordinates of the center of the same recognizable feature in the image, positioning the face of the sample to be inspected at the same particular working distance as used to image the small, recognizable feature and acquiring the digitized optical image with the light source illuminating the sample, and the optical image of the sample to be inspected is shown on a visual display centered on the position coordinates of the recognizable feature with the location of the electron beam analysis area highlighted.

[0053] In any of these embodiments, the detector module includes a first optical camera and a second optical camera, which may be the same or different from those described above, and which may have respective first and second fields of view of the sample that at least partially overlap. The first and second optical cameras may be positioned such that the first and second images captured thereby, respectively, can be used to provide a stereoscopic representation of the sample and / or such that data from or output by the cameras can be used to generate a topographical map of the sample surface.

[0054] The detector module may in some embodiments be configured to perform a data acquisition and processing step or method, in which spectral data from an associated X-ray detector, spectral data from a plurality of X-ray sensor elements and optionally data from a plurality of electronic sensor elements are recorded while the electron beam is positioned at a series of points covering a region on the sample, and the spectral data from the module and optionally data from the plurality of electronic sensor elements and / or the spectral data from the associated X-ray detector are used to identify sub-regions in which the recorded signals are similar for points within a given sub-region, and combining spectral data from the associated X-ray detector, preferably from the associated X-ray detector, for a series of positions, preferably all positions, to generate a single spectrum representative of the material within that sub-region; processing the representative spectrum of the sub-region to determine intensity values ​​of one or more characteristic elemental X-ray emissions, and optionally the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning the elemental intensity or concentration values ​​derived from the representative spectrum of the sub-region to points, preferably every image point, within the sub-region; and providing a visual representation of the elemental distribution across a region of the sample using the image data collected for the elements for the identified sub-regions.

[0055] Typically, in such an embodiment, the data at each point from the X-ray sensors in the module is scaled by a factor that is inversely proportional to the sum of the electronic sensor data at that point.

[0056] Data for points within the identified sub-regions is typically inspected before the spectral data from the attached X-ray detector is combined to exclude points from sub-regions where the point data is outside the expected range of variation for the data for points within the sub-region.

[0057] By using a weighted combination of spectra, where the weight of a spectrum depends on a measure of the difference between that spectrum and the average spectrum for the entire sub-region, spectral data from the attached X-ray detector for points within the identified sub-region can be aggregated to generate a single spectrum for the sub-region.

[0058] The detector module may be configured to perform a data acquisition and processing step or method, wherein, while the electron beam is positioned at a series of points covering an area on the sample, spectral data from an associated X-ray detector, spectral data from an X-ray sensor in the module and optionally data from an electronic sensor in the module are recorded, and the spectral data from the module, optionally data from the electronic sensor in the module and optionally data from the associated X-ray detector are used to identify sub-areas in which the recorded signals are similar for points in a given sub-area, and to perform a data acquisition and processing step or method for the sub-areas, preferably all sub-areas. combining spectral data from the associated X-ray detector for each point within the sub-region with data from other points, preferably all other points, within the sub-region by weighted averaging to produce a denoised version of the spectrum at that point; processing the denoised spectrum for each point to determine intensity values ​​of one or more characteristic elemental X-ray emissions and, optionally, the concentrations of the corresponding elements responsible for those emissions; assembling image data for one or more elements by assigning the elemental intensity or concentration values ​​derived from the denoised spectrum to the image point, preferably every image point; and using the image data collected for the elements to provide a visual representation of the elemental distribution across the region of the sample.

[0059] In such an embodiment, preferably the weighted average for the denoised spectrum at a point uses weighting factors for each of the other points in the sub-region that depend on a measure of the difference between the spectral data or module vector data for that point and the corresponding spectral data or module vector data for each of the other points in the sub-region.

[0060] The data at each point from the multiple x-ray sensor elements contained in the detector module is preferably scaled by a factor inversely proportional to the sum of the electronic sensor data values ​​at that point. The data for points within the identified sub-regions can be examined to exclude points from sub-regions where the point data is outside the expected range of variation for points within the sub-region.

[0061] According to a second aspect of the present invention there is provided an apparatus for analysing a sample, the apparatus comprising an electron beam assembly for generating a focused electron beam and a detector module according to the first aspect.

[0062] In some embodiments, the detector module can be oriented such that the smallest radial distance from the electron beam axis to the module periphery facing the sample is in the direction of an accessory device mounted on one side of the electron beam assembly. Therefore, in such embodiments, the apparatus typically further comprises an accessory device mounted on the electron beam assembly, and the orientation of the detector module relative to the electron beam assembly and / or the accessory device, particularly relative to the electron beam axis, is such that at least a portion of the first part of the detector module, preferably a central portion, more preferably a portion that is annularly / circumferentially centered so as to be equidistant or substantially equidistant from the two planes that define the first part, and at least a portion of the accessory device, preferably a portion that coincides with its central portion or the center of a sensor or working component of the accessory device, coincides with the plane in which the electron beam axis lies. The plane in which the axis lies can be considered a third plane in relation to the above-mentioned first and second planes that can define or surround the first part of the module as described above. This arrangement advantageously means that the smallest radial length of the module is aligned with the accessory device with respect to rotation about the electron beam axis. That is, it can provide an unobstructed line of sight from the sample, particularly the probe spot, to an accessory device mounted above the module near the electron beam assembly. Typically, the first portion is positioned between the electron beam axis and the accessory device in use so that the accessory device is oriented to observe or access the sample by, or at least in a manner facilitated by, minimal extension of the module at the first portion.

[0063] As mentioned above, the apparatus may include an accessory device, which, or in some embodiments, a further accessory device, may include or be an X-ray detector fitted with an electron trap supported on a mechanical assembly that also supports the detector module.

[0064] According to a third aspect of the present invention, there is provided a method of analysing a sample, preferably carried out using a detector module according to an embodiment of the first aspect or an apparatus according to an embodiment of the second aspect. Typically, the method is carried out in an electron beam instrument such as an electron microscope, comprising the steps of generating a focused electron beam using an electron beam assembly; providing a detector module having a plurality of X-ray sensor elements and a plurality of electron sensor elements, the detector module being positioned below a pole piece of the electron beam assembly from which the focused electron beam emerges towards the sample, the detector module being adapted to receive X-rays and backscattered electrons generated by interaction between the electron beam and the sample; and monitoring the energy of each received X-ray photon using the plurality of X-ray sensor elements, the plurality of X-ray sensor elements being spaced apart by 20 mm. 2 and a radial length of the detector module relative to the electron beam axis is less than 10 mm over at least a first portion of the detector module.

[0065] The method may further include, for example, recording a respective signal for each point within the region, the signal including spectral data from a plurality of X-ray sensor elements and optionally data from a plurality of electronic sensor elements, while impinging the electron beam on a series of points covering the region; using the spectral data obtained by the detector module, optionally data from a plurality of electronic sensor elements within the module, and optionally spectral data from an associated X-ray detector to identify sub-regions where the recorded signals are similar for points within a given sub-region; combining spectral data from the associated X-ray detector for some, preferably all, positions within each sub-region to generate a single spectrum representative of the material within that sub-region; processing the representative spectrum for each sub-region to generate data representing intensity values ​​of one or more characteristic elemental X-ray emissions and optionally the concentrations of the corresponding elements associated with those emissions; collecting image data for one or more elements by assigning the elemental intensity values ​​or concentration data derived from the representative spectrum for each sub-region to a point within the sub-region, preferably every image point; and generating a visual representation of the elemental distribution across the region of the sample using the image data collected for the elements for the identified sub-regions. Typically, the spectral data are obtained by X-ray sensor elements of the detector module. Identifying the subregions can be considered as defining the subregions as parts of regions in which the recorded signals for all or at least several points are similar. This similarity can be determined, for example, by comparing the signals for pairs or several points according to a required similarity threshold. Spectral data can be combined for some or all positions within at least one, preferably several or all of the identified subregions. Representative spectra can be processed for at least one, preferably several or all of the identified subregions. Data representing the concentrations of the corresponding elements can be understood to represent the concentrations of elements that are responsible for the radiation in question, in particular due to their presence and interaction with the incident electron beam at the probe spot.

[0066] Generating the visual representation can be thought of as providing that representation and can take several different forms. The aggregated image data used to do so can relate to at least one set, and preferably all, of the identified sub-regions.

[0067] Typically, the data from the X-ray sensors in the module at each point is scaled by a factor inversely proportional to the sum of the electronic sensor data at that point. Preferably, in such an embodiment, the data for points within the identified sub-regions is examined before the spectral data from the associated X-ray detectors is combined to exclude points from sub-regions where the point data is outside the expected range of variation for the data for points within the sub-region. In another embodiment, the spectral data from the associated X-ray detectors for points within the identified sub-regions is aggregated to generate a single spectrum for the sub-region using a weighted combination of spectra, where the weight of a spectrum depends on a measure of the difference between that spectrum and the average spectrum for the entire sub-region.

[0068] In a method of acquiring and processing data using a detector module according to an embodiment of the first aspect or an apparatus according to an embodiment of the second aspect, or alternatively in an embodiment of a method according to the third aspect, typically signals comprising spectral data from an associated X-ray detector, spectral data from an X-ray sensor in the module and optionally data from an electronic sensor in the module are recorded while the electron beam is positioned at a series of points covering a region on the sample, and the spectral data from the module, optionally data from the electronic sensor in the module and optionally data from the associated X-ray detector are used to identify sub-regions in which the recorded signals are similar with respect to points within a given sub-region, and or for any sub-region, combining spectral data from the associated X-ray detector for each point within the sub-region with data from other points, preferably all other points, within the sub-region by weighted averaging to produce a denoised version of the spectrum at that point; processing the denoised spectrum for each point to determine intensity values ​​of one or more characteristic elemental X-ray emissions, and optionally the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning the elemental intensity or concentration values ​​derived from the denoised spectrum to the image point, preferably every image point; and using the image data collected for the elements to provide a visual representation of the elemental distribution across the region of the sample.

[0069] Preferably, the weighted average for the denoised spectrum at a point uses a weighting factor for each other point in the sub-region that depends on a measure of the difference between the spectral data or module vector data for that point and the corresponding spectral data or module vector data for each other point in the sub-region. More preferably, the data at each point from the X-ray sensor elements that the module contains is scaled by a factor that is inversely proportional to the sum of the electronic sensor data values ​​at that point. More preferably, the data for points in identified sub-regions is examined to exclude points from sub-regions where the point data is outside the expected range of variation for points in the sub-region.

[0070] There is provided a method of displaying a location of an area being analyzed under the electron beam together with an optical image of a sample to be inspected using a detector module according to an embodiment of the first aspect or an apparatus according to an embodiment of the second aspect, comprising: positioning the face of a cathodoluminescent sample at a particular working distance from a final lens pole piece of an electron microscope; acquiring a digitized optical image with the illumination light source turned off while a focused electron beam is incident on the sample; determining a position coordinate in the optical image of a spot center of light emitted by cathodoluminescence; positioning the face of the sample to be inspected at the same particular working distance as the cathodoluminescent sample and acquiring a digitized optical image with the light source illuminating the sample; and showing the optical image of the sample to be inspected on a visual display in which the location of the electron beam analysis area is highlighted centered on the position coordinate obtained from the optical image obtained from the cathodoluminescent sample.

[0071] There is provided a method of displaying the location of an area being analyzed under the electron beam together with an optical image of a sample to be inspected using a detector module according to an embodiment of the first aspect or an apparatus according to an embodiment of the second aspect, comprising: acquiring an electron image by focusing a face of a small, recognizable feature at a particular working distance from the final lens pole piece of the electron microscope and centering the feature on the field of view; acquiring a digitized optical image with the illumination source on and determining the position coordinates of the center of the same recognizable feature in the image; positioning the face of the sample to be inspected at the same particular working distance as used to image the small, recognizable feature and acquiring the digitized optical image while illuminating the sample with the light source; and showing the optical image of the sample to be inspected on a visual display centered on the position coordinates of the recognizable feature with the location of the electron beam analysis area highlighted.

[0072] An embodiment of the present invention will now be described with reference to the accompanying drawings. [Brief explanation of the drawings]

[0073] [Figure 1]1 shows a schematic diagram of a first configuration for X-ray analysis in an electron microscope according to the prior art; FIG. [Figure 2] FIG. 1 is a plan view of a prior art detector arrangement showing the layout of a central hole for an electron beam surrounded by a ring of backscattered electron sensor segments and an outer ring of X-ray sensor segments. [Figure 3] FIG. 1 is a schematic cross-sectional view of a combined backscattered electron and X-ray detector arrangement. [Figure 4] 1 is a graph showing the X-ray transmittance as a function of X-ray light energy for a 6 micron thick Mylar filter. [Figure 5] FIG. 10 is a schematic diagram illustrating a second configuration for X-ray analysis in an electron microscope according to the prior art, showing that the line of sight of the side-mounted detector is partially blocked by a large diameter detector module below the microscope pole piece. [Figure 6] 1 is a plan view of a first exemplary detector module arrangement according to the present invention including two circular X-ray sensors on either side of a BSE sensor surrounding a central hole. FIG. [Figure 7] FIG. 10 is a plan view of a second exemplary detector module arrangement according to the present invention including two circular X-ray sensors on either side of a central hole with two BSE segments on either vertical side of the central hole. [Figure 8] FIG. 10 is a plan view of a third exemplary detector module arrangement according to the present invention including two circular X-ray sensors and two BSE sensor segments on either side of a central hole with two-fold rotational symmetry. [Figure 9] FIG. 10 is a side view of a portion of the sample showing radiation from the base of an internal cavity being received by a portion of the sensor. [Figure 10] 1 is a schematic cross-sectional view of a portion of an exemplary apparatus according to the present invention illustrating how the collection solid angle of different sensors changes when a focused electron beam is deflected away from a central position. [Figure 11]10 is a data table showing the total response of both BSE sensors relative to the center position for different positions of a focused beam incident on a 1.5 mm x 1.5 mm grid on a sample when the sample is positioned 5.8 mm from the plane of the BSE sensors of the third exemplary module. [Figure 12] FIG. 10 is a plan view of a fourth exemplary detector module arrangement according to the present invention including two circular X-ray sensors and four BSE sensor segments on either side of a central hole with two-fold rotational symmetry. [Figure 13] 10 is a data table showing the total response of all BSE sensors relative to the center position for different positions of a focused beam incident on a 1.5 mm x 1.5 mm grid on the sample when the sample is positioned 5.8 mm from the plane of the BSE sensors of the fourth exemplary module. [Figure 14] 10 is a data table showing the response of [BSE sensor (1 + 2) + 3.3 × sensor (3 + 4)] relative to the center position to different positions of a focused beam incident on a 1.5 mm × 1.5 mm grid on the sample when the sample is positioned 5.8 mm from the plane of the BSE sensor of the fourth exemplary module. [Figure 15] FIG. 10 is a perspective view of a third exemplary detector module showing two circular X-ray sensors and two BSE sensors facing the sample in use, showing connectors for providing bias voltages for the sensors, power for thermoelectric cooling, and taking the sensor signal out to external electronics through a vacuum feedthrough. [Figure 16] 1 is a schematic side view of an exemplary arrangement according to the present invention including an optical camera directly below the SEM pole piece. [Figure 17] FIG. 12 is a perspective view of an exemplary arrangement according to the present invention including a retractable side arm supporting a downward-looking detector module below the pole piece with an additional accessory X-ray detector with an integral electron trap mounted in a position that does not obstruct the line of sight to the probe spot on the sample. [Figure 18] 18A and 18B are side and top views of the exemplary arrangement shown in FIG. 17 in a position below the pole piece of the last lens. [Figure 19]Figure 19a is a cross-sectional view through an exemplary arrangement according to the invention in which an additional X-ray sensor fitted with a BSE-blocking material filter is mounted at the end of the electron trap to prevent electrons from reaching the X-ray sensor which does not have a BSE-blocking filter and is more sensitive to low energy X-rays, and Figure 19b is a view of the module from the direction of the sample in use showing the X-ray sensors fitted with filters on either side of the entrance aperture of the electron trap. [Figure 20] 3 is a flow chart illustrating a first exemplary data acquisition process according to the present invention. [Figure 21] 4 is a flow chart illustrating a second exemplary data acquisition process according to the present invention. [Figure 22] This graph shows an overlay of the X-ray total spectrum of the module data (solid black area) and the total spectrum of the attached detector data (lighter line). [Figure 23] FIG. 1 shows an element distribution map taken from an example of module X-ray data. [Figure 24] FIG. 1 shows an elemental distribution map obtained from an exemplary device including an attached detector. [Figure 25] FIG. 10 illustrates the "phase" distribution of pixels with similar spectra in an exemplary modular dataset. [Figure 26] FIG. 24 shows elemental distributions calculated from concentration results derived from attached detector data for the phases shown in FIG. 23. [Figure 27] FIG. 1 shows a "phase" map depicting the distribution of different materials, illustrating that an individual color can typically be assigned to each phase and displayed overlaid on a corresponding electronic image obtained from the same region of the sample. DETAILED DESCRIPTION OF THE INVENTION

[0074] An exemplary apparatus employing the advantageous approach described above uses a module housing both the BSE and X-ray detector sensors, positioned below the pole piece of the final lens (i.e., position 106 in FIG. 5 ) with the active face of the sensor facing the “probe spot” where the electron beam strikes the sample. This module is designed to maintain rotational symmetry with respect to the BSE sensor placement to provide the X-ray detector sensor with a high total collection solid angle for X-rays emitted from the sample. The module can be cooled to reduce electronic noise from the sensor, or it can be at ambient temperature. When electronic cooling, such as thermoelectric / Peltier cooling, is used, the temperature can be sensed and adjusted by feedback, with the benefit of maintaining constant gain and offset of the BSE detector sensor. As noted above, a key aspect of these embodiments is maintaining a high total collection solid angle for both the X-ray and BSE detectors while ensuring that the module does not block line-of-sight to the probe spot for accessory devices positioned on either side of the electron beam. For conventional SEM designs, it is known that the bottom edge of the module (the side facing the specimen) should not protrude more than 10 mm from the electron beam axis, preferably observing a radial distance of 7 mm, and ideally a radial distance of 4.5 mm, to avoid shielding typical accessory devices such as X-ray detectors mounted on the sides of the SEM when the module is positioned near the pole pieces. Reducing the radial shielding by reducing the overall dimensions requires a compromise in sensitivity. For example, if a known module design such as that shown in Figures 2 and 3 were scaled down to a module outer diameter of 9 mm (with a radial distance to the periphery of 4.5 mm), the total area of ​​the four X-ray sensors would be approximately 30 mm. 2 The total area of ​​the four BSE sensors is 8 mm 2 Exemplary devices according to the present disclosure preferably achieve a 30 mm 2 The overall area of ​​the X-ray sensor is significantly larger than 30mm 2This allows for a significantly larger overall BSE sensor area than the conventional BSE sensor. It has been recognized that it is beneficial to be able to rotate and fix the module at any angular orientation in order to reduce the radial distance to the exterior of the module radially toward the attached device. Thus, if a rectangular module with a sensor arrangement such as that shown in FIG. 6 could be rotated about the central beam axis, it could be positioned so that its periphery is only 7 mm radially from the beam axis to minimize shielding of side-mounted attached devices while providing a high solid angle for x-ray signal collection.

[0075] While the BSE sensor in Figure 6 is shown schematically as an octagon with a central hole, it could also be circular, consisting of a set of four segments, for example, as in the prior art arrangement of Figure 2. This achieves a high degree of rotational symmetry of the active area, minimizing the effect of topography on the overall BSE signal, as discussed above. Whether a single segment or multiple segments are used, provided that the segments cover most of the area, the overall active area approximates a disk (or "annulus") with a hole and perfect rotational symmetry about the electron beam axis.

[0076] In the exemplary arrangement of FIG. 6 , the X-ray sensor is positioned further away from the beam axis than the BSE sensor, thereby compromising the solid angle of the X-ray detector. The inventors have found that, contrary to previous field experience, even if the detector has only two segments that do not cover the entire annulus, the effect of topography on the BSE signal can still be dramatically reduced as long as the segments exhibit two-fold rotational symmetry. The arrangement of FIG. 7 takes advantage of this property by positioning the two X-ray sensors closer to the electron beam axis, increasing the collection solid angle while maintaining the two-fold rotational symmetry of the BSE sensor segments. Similar to the arrangement of FIG. 6 , if the exemplary oval module of FIG. 7 could be rotated around the central beam axis, it could be positioned to minimize shielding of the side-mounted accessory device by reducing the radial distance to the module's periphery in the direction toward the side-mounted accessory device to 6 mm.

[0077] While the arrangement of Figure 7 improves the collection solid angle for X-rays, it compromises the overall effective area and overall solid angle for collection of BSE signals. The exemplary module arrangement shown in Figure 8 extends the BSE segment while still maintaining two-fold rotational symmetry. The outer boundary of the approximately oval module is shaped to provide two regions that reduce the radial distance from the beam axis to the periphery, reducing shielding of side-mounted accessory devices in all directions. In the arrangement of Figure 8, the 21.2 mm long module accommodates two SDD sensors and extends to a 40 mm 2 or greater overall effective area, with a minimum radial distance from the beam axis to the periphery of less than 4.5 mm.

[0078] In this embodiment, the central hole of the module is necessary to allow the focused electron beam to propel toward the sample. Over time, the aperture can accumulate contaminant material, causing electron charge to build up on the material. If the edge of the hole is too close to the focused electron beam, it can distort the beam profile and degrade microscope image quality. When inserting the module support arm through a side port in the electron microscope, the hole must be carefully aligned with the electron beam axis to maximize the distance from the hole's sidewall to the focused beam. The smaller the hole, the more difficult this alignment becomes. Furthermore, a small percentage of electrons in the incident beam can stray from the main focused beam and include an intense "beam tail" that extends far beyond the extent of the main beam. If part of this beam tail strikes the side of the aperture, it can exacerbate contamination accumulation. Therefore, to minimize such potential problems, it is beneficial to provide as large a central hole as possible. However, a compromise is required because the solid angle to the sample is reduced when the X-ray and BSE sensor elements are moved away from the axis to accommodate a larger central hole. The inventors have determined that to avoid these effects in most SEMs, the inner diameter of the central hole should be at least 1.5 mm, and ideally greater than 2.5 mm.

[0079] An additional benefit of placing the X-ray sensor below the pole piece is that it can detect X-rays emitted from materials that are not visible to side-mounted detectors. Figure 9 shows a side view of a sample with a focused electron beam impinging on material at the base of a cavity. X-rays emitted from the probe spot can reach any portion of the active sensor area that is visible and not obscured by the sidewalls of the cavity. Therefore, it is desirable to have as much of the active sensor area as possible within a small radial distance from the electron beam axis. If the sensor face is 6 mm from the probe spot on the sample, X-rays emitted at an elevation angle θ of 45° relative to the horizontal will impinge on the sensor at a radial distance of 6 mm from the axis. To provide some analytical capability for materials at the base of the cavity, it is desirable to have at least half of the X-ray sensor's total active area within 6 mm of the axis.

[0080] As mentioned above, arranging the BSE sensors with at least two-fold rotational symmetry about the focused beam axis is important for reducing the effect of topography on the overall BSE signal, making it more representative of the material composition rather than the local orientation of the surface. Similarly, arranging the X-ray sensors symmetrically about the focused beam axis is advantageous, especially when observing samples with severe surface topography. If the incident electron beam strikes an object at the base of a "cliff," the cliff itself may block the X-rays from passing to one of the X-ray sensors. However, the other X-ray sensor, diametrically opposed, can still clearly see the object and detect the X-ray emission from the object. Furthermore, for samples with locally flat surfaces where both X-ray sensors can clearly see the probe spot, the emitted X-rays will be subject to so-called "matrix" effects, such as sample self-absorption, which depends on the surface orientation. If the surface is tilted from the horizontal and the X-ray sensors have two-fold rotational symmetry about the focused electron beam axis, the signals from the two sensors will be averaged to reduce the impact of the matrix effect of the surface tilt.

[0081] Although the sensor's two-fold rotational symmetry helps mitigate topography effects, deflecting the incident electron beam to different locations on the sample plane changes the sensor's overall response to the excitation signal. This is illustrated in Figure 10, which shows a side view of the microscope and two sensors equally spaced about the central axis. When a focused electron beam is deflected to an off-axis location on the sample, the collection solid angle for electrons and X-rays is larger for the sensor closer to the beam position on the sample. Along the line between the sensors projected onto the sample surface, the increase in the collection solid angle of one sensor is nearly compensated for by the decrease in the solid angle of the opposite sensor. However, for beam positions perpendicular to this line, the collection solid angle of both sensors decreases as the distance from the central axis increases.

[0082] The higher the magnification of the SEM, the smaller the maximum scanning deflection on the sample required to cover the field of view, and the smaller the variation in signal detection efficiency at different locations within the field of view. However, in a typical SEM with a fairly low magnification of 200X, the field of view is approximately 1.5 mm wide, and the response across the field of view can vary by several percent due to variations in the collection solid angle with position, for example, for the design shown in Figure 8. While BSE signals can distinguish between materials of different compositions and therefore signal thresholds can be established to define regions of different materials within the field of view, any spatial variation in signal collection efficiency can make this delineation unreliable when the materials have similar average atomic numbers. Therefore, the degree of this response nonuniformity is particularly important for BSE signals. Because the physics of BSE emission is known and can be simulated using Monte Carlo techniques, the distribution of backscattered electrons as a function of both angle and energy can be predicted. For a detector module with the dimensions shown in Figure 8, the relative response of the sum of the two BSE sensors to a 20 keV incident electron beam is shown in Figure 11 for various positions within a 1.5 mm wide square field of view when the BSE sensor's incident surface is positioned 5.8 mm above the sample surface. The positions are from the same direction as in Figure 8, and the field of view is from the sample. When the beam position is at the top left or bottom right of the field of view, it is closest to one of the two BSE sensors, and its sum response is 3.3% larger than when the beam position is on the central axis. When the beam position is at the top right or bottom left, it is farther from both sensors, and the sum response is 3% smaller than when it is at the central position.

[0083] For applications that cannot tolerate non-uniform response over a wide field of view, a conventional solution would be to use a full-disk or four-fold symmetric segmented BSE detector, such as the type shown in Figures 2 or 6. However, this requires moving the SDD X-ray sensor further away from the beam axis, sacrificing the collection solid angle for X-ray detection. An alternative approach of placing a ring of BSE sensors outside the X-ray sensor reduces the BSE signal per unit area of ​​the sensor, because the intensity of BSE electrons is known to fall approximately as the cosine of the emission angle relative to the surface normal. Increasing the module diameter to improve either BSE or X-ray sensitivity increases the likelihood of blocking the line of sight of other instruments in the SEM chamber or detectors mounted on one side. To avoid these compromises, we have devised a way to achieve a more uniform BSE response over a wide field of view with an oval module with only two-fold rotational symmetry, as exemplified by the 26.5 mm long module shown in Figure 12. BSE sensor elements 1 and 2 are closest to the central axis and are therefore most efficient at detecting BSE electrons. The combined response of sensors 1 and 2 will exhibit non-uniformity across a wide field of view, with the weakest response at the farthest ends of the line perpendicular to the vertical line connecting the sensors. Adding BSE sensors 3 and 4 to the outside of the SDD X-ray sensor helps compensate for this loss of sensitivity. As a result, if the BSE sensor entrance faces are again positioned 5.8 mm above the sample surface for a module with the dimensions shown in Figure 12, the BSE response for the sum of all sensors is calculated, resulting in improved uniformity of response across a 1.5 mm square area on the sample, as shown in Figure 13. Here, the maximum response is only 0.2% larger and the minimum response is only 0.7% smaller than at the center position.

[0084] The uniformity of the overall BSE response can be further improved by increasing the size of BSE sensors 3 and 4 to compensate for the reduced efficiency due to the cosine emission response compared to sensors 1 and 2. However, the modifications necessary to improve uniformity necessarily increase the length of the module, which increases the potential for interference with other auxiliary devices in the SEM chamber. Instead of increasing the diameter, the responses of BSE sensors 3 and 4 can be amplified by electronic, digital, or software algorithms to compensate for the signal loss due to the cosine emission response. For the module of Figure 12, results for the same 1.5 mm square area are shown in Figure 14, using an amplification that produces a result equivalent to the sum of sensors 1 and 2 plus 3.3 times the sum of sensors 3 and 4. Here, the minimum response is the same as at the center position, and the maximum response is only 0.2% larger.

[0085] The factor of 3.3 required to minimize response nonuniformity across a 1.5 mm x 1.5 mm field of view with respect to beam position on the sample is a function of the sensor geometry, position, and distance from the sample on the module. A factor of 1.0 gives the same response versus position as shown in Figure 13, while factors between 1 and 3.3 provide intermediate improvements in uniformity across that same area. Amplifying the signal from sensors 3 and 4 in this way improves uniformity, but also amplifies the noise in that portion of the signal. The minimum noise is essentially limited by the electron flux impinging on the sensor, and the smaller the sensor, the greater the noise. Therefore, a trade-off exists between improved response uniformity with respect to beam position and a reduction in the signal-to-noise ratio of the overall response, as increasing the amplification factor increases the noise in the overall response. To further improve spatial response uniformity, it may be possible to divide the BSE sensor into more segments and apply different levels of amplification before summing the signals.

[0086] The same design principles used to improve the uniformity of the BSE spatial response for a combined X-ray and BSE detection module can be used for any BSE detector requiring an elliptical geometry. Whereas the conventional approach is to have segments cover the full 360 degrees around the central axis or to have at least four-fold rotational symmetry, in the case of an elliptical detector with two-fold rotational symmetry, the non-uniformity of the spatial response for the sensor areas on either side of the center at the narrowest region of the detector is compensated for by including signals with different amplification levels for the portions of the sensor areas on either side of the center at the widest part of the detector, thereby providing a more uniform BSE response for different beam positions on the sample.

[0087] The best BSE signal response per unit area is obtained when the sensor's active area is close to the central beam axis. However, to maximize the X-ray sensor's response, it is also desirable to position the X-ray sensor's active area close to the central beam axis to maximize the collection solid angle. When BSE sensors are fabricated from silicon wafers, individual sensors are detached from the wafer, which can damage the surrounding surface. X-ray sensors encounter a similar problem, requiring the incorporation of a dead boundary around each sensor's active area to prevent electrical connection failure or leakage. Figure 8 shows the dead boundary of a BSE sensor. If this dead boundary is close to the module's central axis, useful signal is wasted, so it is desirable to reduce the extent of the dead area. One way to reduce the dead area is to fabricate both the BSE and X-ray sensors on the same semiconductor wafer, eliminating the need for separate dicing. This achieves a cross-section similar to that shown in Figure 3, but without the gap between the BSE and X-ray sensors.

[0088] An alternative method for correcting response nonuniformity due to beam position is to calculate the response for a particular design using knowledge of the sensor-to-sample distance. A response surface like that shown in Figure 11 can then be calculated and used to adjust the measured response for position, so that the adjusted signal response values ​​can be stored in a digital image of a field of view with good uniformity. Yet another alternative is to use a flat, homogeneous sample, such as a polished piece of copper or nickel, to measure the response when the module is positioned at a specific distance from the sample. Once a digital image is recorded over the desired field of view, the values ​​of the individual pixels provide a calibration of the relative signal response under these conditions. Subsequently, when the sample of interest is examined using the same conditions as for calibration, the values ​​from the calibration image can be used to adjust the signal response to correct for nonuniformity.

[0089] Similar nonuniformity problems arise for X-ray sensors, which also have simple two-fold rotational symmetry. The X-ray signal does not fall off as the cosine of the emission angle relative to the surface normal, but is also affected by the collection solid angle relative to the emission from the beam position. As with BSE sensors, nonuniformity in response across the field of view can be corrected by calculating the response or by measuring the response on a flat calibration sample of homogeneous composition under the same microscope conditions to calibrate the response at different beam positions. While BSE sensors generate a single value representing the BSE emission from a beam position on the sample, X-ray sensors generate a histogram equivalent to the energy spectrum of the X-rays emitted from that position. Typically, X-ray maps showing the spatial distribution of a chemical element are obtained by recording, at each beam position, the contribution to the spectrum that comes only from the energy emission specific to that element. When such a map is recorded over a wide field of view, even if the sample is homogeneous, the change in the total collection solid angle with position will produce a nonuniform map. If it is not possible to calculate or calibrate the response to correct for this non-uniformity, another option is to use the ratio of the intrinsic emission count to the total spectral emission count, rather than just the intrinsic emission count at each pixel location in the map. Because the total spectral emission count is subject to the same collection solid angle, this ratio is not affected by variations in collection solid angle with beam position, and therefore the ratio map will be more uniform than a map of the intrinsic emission of that element.

[0090] In any of the sensor configurations of the present invention, the X-ray sensor can be covered with a filter material to prevent BSE from saturating the X-ray sensor while allowing low-energy X-rays to reach the sensor. The filter can be composed of multiple layers of different materials to optimize the transmission of important X-ray characteristic emission lines while blocking the transmission of BSE. If the sample is cathodoluminescent, or if the sample or other devices nearby emit visible or infrared (IR) radiation, it may also be advantageous to prevent these additional radiation sources from reaching the X-ray sensor and adversely affecting the measurement of the X-ray energy spectrum emitted by the sample. This may require additional layers if the BSE-blocking material does not block light or infrared radiation. At least one layer of material must be conductive and connected to ground or a current sink to prevent the filter from accumulating charge and reaching a potential that could interfere with the operation of the electron beam optics and scanning system.

[0091] The filter material can include a self-supporting foil, such as a thin layer of Mylar or polyimide with a conductive coating of aluminum or carbon. Typical filters, several microns thick, can be removed and repaired if damaged. Alternatively, a filter that blocks BSE can include a coating applied directly to the face of the X-ray sensor. For example, a 150-nm aluminum or 34-nm palladium coating would be sufficient to block BSE up to 3 keV in energy, but would still pass low-energy X-rays from carbon, nitrogen, and oxygen. A 350-nm aluminum coating would block BSE up to approximately 5 keV in energy. Such a coating would be much more transparent to low-energy X-rays than the 6-micron Mylar filter of Figure 4, which significantly attenuates X-rays below 1 keV. If a coating sufficient to block BSE is applied directly to the X-ray sensor face when the SEM is operated at low incident beam voltages, a removable self-supporting foil can be added to allow the module to be used when the SEM is operated at high incident beam voltages. Thus, the module can include a foil filter with or without to optimize X-ray detection at different SEM beam voltages. When operating the module without a foil filter to maximize X-ray sensitivity to low-energy X-rays, it is beneficial to operate the X-ray sensor at a lower temperature to improve the energy resolution of the X-ray spectrum for low-energy X-rays. When using a foil filter, the X-ray sensor can be operated at temperatures closer to room temperature, since the energy resolution for high-energy X-rays detected with the foil filter in place is less important. Operation at a higher temperature is advantageous in that it reduces the power required to cool the module and reduces the likelihood of condensation or icing on the sensor. Condensation or icing is more likely to occur in SEMs that can operate in "low vacuum" or "ambient" modes, which are suitable for analyzing hydrated samples. Therefore, the ability to set the degree of cooling allows the sensor temperature to be adjusted to optimize performance for different operating modes of the electron microscope.

[0092] With the module in its home position 106, there is a risk that the sample stage will be accidentally actuated, resulting in a collision with the stage or sample 101 and damaging the sensors within the module. Therefore, each sensor has a protective opening closer to the sample than the sensor's sensitive incidence surface to prevent the sample from damaging this surface if the sample stage accidentally gets too close to the sensor. Each sensor is typically located behind an opening in a conductive plate, which can be used as one electrode of a capacitive sensor that monitors the proximity of the sample and / or sample holder. The other plate can ideally be formed from a conductive sample or a sample holder connected to a conductive coating on the sample. This arrangement allows the proximity sensor to be calibrated when the sample is positioned at a safe minimum distance from the module, issuing a warning or limiting stage movement with an interlock to prevent accidents while the user is using the stage to examine the sample.

[0093] The module must be connected to external electronics to obtain voltages for energizing the sensor elements, i.e., power for the thermoelectric cooling stack, and to extract signals from the BSE and X-ray sensors and any temperature sensors. Figure 15 shows a 3D representation of the module, illustrating examples of where connectors can be attached to the module for these purposes. Additionally, the module will need some means for transferring heat away from the module to a heat sink. The module can either be attached to the pole piece of the final lens, conducting heat to the metal of the pole piece, or to a support with a thermal conductor, transferring heat to the metal wall of the microscope or to a heat sink or cooling device outside the microscope. Alternatively, the module can be attached to a support arm extending from a side port of the electron microscope, allowing it to be inserted into an operating position below the pole piece or retracted to the side, allowing other devices to be brought into position below the pole piece or the sample itself to be closer to the pole piece. In this case, the support arm can also be a conduit for conducting heat and receiving any electrical wiring or conductors connecting the module to external electronics.

[0094] As mentioned above, SEM images obtained by deflecting the beam cover a very narrow area on the sample, making it difficult for the operator to accurately determine where that area is located on the sample without a wide field of view. If the camera can be positioned near vertical, pointing directly at the sample to avoid excessive perspective distortion, the camera image of the sample surface can help the operator locate the area scanned by the electron beam relative to the surrounding sample surface, as long as the field of view is large enough. To achieve good imaging performance in an SEM, it is desirable to keep the distance between the final lens pole piece and the sample less than 10 mm. Therefore, the camera must not only be very small to maximize the camera-to-sample distance, but also have a short focal length and excellent depth of field to maximize the field of view (e.g., Omnivision OVM6946 / 8 wafer-level camera, external dimensions 1.1 × 1.1 × 2.2 mm). If the camera-to-sample distance is less than 10 mm, a field of view of at least 10 mm wide can be achieved with a single camera, and the "fisheye" distortion of the camera image can be corrected using known image correction methods. Furthermore, the camera's depth of field is sufficient, and the vertical height of the sample stage can be adjusted to increase the camera-to-sample distance, providing a much wider field of view—at least 20 mm, ideally greater than 60 mm. Sensor modules that fit under the pole piece and accept only BSE sensors can be equipped with one or more miniature cameras, with electrical connections provided through the same conduits used to support the sensor module and insert it into position under the pole piece. Similarly, new detector modules combining X-ray and electronic sensors, such as those shown in Figures 8 and 10, can also be equipped with one or more miniature cameras. Because visible light can damage the sensitive photomultiplier tubes (PMTs) used in typical Everhart-Thornley SE detectors, an infrared (IR) light source (typically around 900 nm) is required to illuminate the sample when the PMTs are activated. This light source can be provided by either a small IR LED within the module or a large IR light source mounted on the support arm.When the PMT is known to be inoperative, a visible light LED can be used to illuminate the sample. The illumination source, whether IR or visible, is preferably turned on only for the time necessary to obtain an optical image of the sample, because photons can also interfere with the operation of sensors for X-rays and BSE, for example, causing excessive noise in those sensors. Figure 16 shows a cross-sectional view illustrating the wide field of view achieved with a miniature optical camera mounted within the sensor module compared to the very small field of view contained in the SEM electron beam scanning area.

[0095] If the working distance (WD) between the pole piece of the final lens and the probe spot where the focused electron beam strikes the sample surface is kept constant, the position coordinates of the probe spot in the digitized camera image will remain constant as long as the camera position is fixed relative to the pole piece. These position coordinates can be determined from a sample with a small feature, particle, or fiducial mark that is easily recognizable in both the SEM electron image and the optical image. The feature is brought to the center of the field of view of the SEM electron image (typically an SE image), and the stage height is adjusted to focus the feature plane at a specific WD. A digital optical image is acquired from the camera, and the position coordinates of the recognizable feature in the image are determined. Once these position coordinates are determined, an optical image can be recorded for any sample under analysis, as long as the SEM beam is focused on the sample plane at the same WD. The location of the sample plane at the center of the field of view of the SEM image will be at the same position coordinates in the optical image. Highlighting the pixel at these coordinates in the visual display of the optical image (e.g., using a crosshair or box centered on this pixel adjusted to the size of the SEM electron beam scanning field) allows the operator to confirm where the analysis location is located within the wide field of view covered by the optical image. An alternative method for determining the location of the probe spot is to make the probe spot visible to the camera while the focused beam strikes the sample. When a cathodoluminescent sample, such as ZnS, MgO, or a fluorescent material, is placed under the focused electron beam with the illumination source turned off, the point where the beam strikes the sample becomes visible to the optical camera as a bright spot centered on the probe spot. Similarly, some samples produce enough IR radiation that a stationary focused electron beam can be detected by the camera. For such special samples, the coordinate location of the probe spot within the wide field of view image of the optical camera can be determined. For other samples, as long as the SEM beam is focused to the same WD on the sample surface, when an optical image is acquired, highlighting the analysis location in the visual display of the optical image as described above can help the operator confirm that the desired area of ​​the sample is being analyzed.Furthermore, when the optical image is corrected for spatial distortion, the stage can be moved so that a set of recognizable features that serve as fiducials on the optical image are each under the electron beam, establishing a relationship between the digital stage position coordinates and the pixel coordinates of the corrected optical image. With such a relationship between the optical system and the stage coordinate system established, an operator can use the digital stage controls to position the electron beam over any corresponding point visible on the optical image without capturing a new optical image.

[0096] Typically, the sample surface is positioned at the optimal working distance for SEM imaging or analysis. Because an attached X-ray detector mounted on the side of the SEM is typically positioned to point at a specific working distance in space on the electron beam axis, the sample surface is positioned at this working distance (typically 10 mm or less) to obtain accurate results in X-ray analysis. Lowering the sample stage (e.g., by adjusting the stage Z control) to increase the working distance allows the optical camera to capture a much wider field of view. For example, a compact optical camera with a 120° field of view can capture a 69 mm wide area on the sample at a 20 mm camera-to-sample distance. Optical images recorded with the sample lowered can be calibrated using the above-described technique to establish a relationship between stage position coordinates and optical image pixel coordinates. An operator can then select a feature of interest in the optical image and use the pixel coordinates in the image to determine the XY stage position that will bring that feature directly under the electron beam. Using the stage Z control, the sample surface can then be raised back to the optimal working distance for analysis, and the operator can use the optical image to drive the stage to bring any desired feature into the optimal position for analysis.

[0097] As shown in Figure 16, two simultaneously acquired images can be used to form a "stereo pair" by using two optical cameras positioned symmetrically with respect to the electron beam axis, but adjusting the camera orientation and / or distance between the cameras so that there is significant overlap in the fields of view of the two cameras. When these images are displayed separately to the operator's left and right eyes, the operator experiences the illusion of depth through binocular stereopsis, allowing them to view the surface topography of the sample in 3D. Alternatively, the two digital images can be processed with a stereophotogrammetry algorithm (e.g., D. Samak et al., "3D Reconstruction and Visualization of Microstructure Surfaces from 2D Images," CIRP Annals, Vol. 56, No. 1, 2007, pp. 149-152) that identifies pixel coordinates in each image for the same feature and uses parallax to determine the feature's relative distance from the camera. If a feature on the sample surface is first positioned to be at a specific WD from the final lens (e.g., by moving the stage height to focus the electron beam on the feature when the SEM lens is set to focus at a specific WD), and images are acquired from two cameras and then processed with a stereogrammetry algorithm, the WD of the feature from the pole piece of the final lens can be used to calibrate the results so that the distance from the pole piece can be determined for any other feature visible in both optical images. After this calibration, the surface of a sample introduced into the SEM can be examined by acquiring stereopair optical images and processing these images to confirm that the surface is at the proper WD and orientation (face tilt) for analysis. If the sample is not properly oriented, the orientation can be corrected by tilting the stage.

[0098] Combining an optical camera with a BSE and / or X-ray sensor in a module below the pole piece offers the unique benefit of assisting in the navigation and locating areas of the sample surface that contain specific materials.

[0099] A common example of an accessory device is a conventional X-ray detector mounted on the side port of an electron microscope, as shown at 105 in FIG. 5 . The accessory detector is located far enough from the sample to include an electron trap that blocks backscattered electrons from impinging on the detector while allowing low-energy X-rays to pass unattenuated. The electron trap typically consists of one or more pairs of permanent magnets or a circular Halbach magnet array to generate a strong magnetic field to deflect any BSEs traveling toward the detector, and a soft-iron enclosure to limit stray magnetic fields so as not to interfere with the focusing optics of the electron microscope (see, e.g., U.S. Pat. No. 8,049,182 B2). An accessory detector with such an electron trap will be much more sensitive to low-energy X-rays than an X-ray sensor in a module below the pole piece, which requires a filter material to block BSEs. When the module below the pole piece is installed on a support arm extending from the side port, an accessory X-ray detector with an integrated electron trap can also be mounted on the same support arm. The advantage of this arrangement is that only one electron microscope port is required, and conduits in the support arm accommodate the electrical and cooling requirements of both the under-pole module and the attached X-ray detector. An example of this arrangement is shown in Figure 17. The additional device with a clear line of sight to the sample probe spot can be an X-ray detector as shown, but could also be a cathodoluminescence detector, a laser, a micromanipulator, or other device small enough to be mounted on the side support arm, such as a gas injector.

[0100] When the associated detectors are mounted in fixed positions relative to the detector module, the module can be designed to ensure that all sensors have a clear line of sight to the probe spot on the sample. Figure 18 shows a side view and a top view of one configuration according to Figure 17, where a symmetric cutout is used to maintain the two-fold rotational symmetry of the BSE sensors, and the cutout allows a clear line of sight for the associated X-ray detectors with electron traps.

[0101] The under-pole module may be limited in terms of spectral quality when required to map and identify regions of different elemental composition. The presence of filter material in front of each X-ray sensor to block backscattered electrons in the module reduces sensitivity to low-energy X-rays. The module may be operated at high temperatures or other conditions that can result in the addition of electronic noise, which reduces spectral resolution and broadens characteristic spectral peaks. The processors used to analyze pulse trains from the X-ray sensor require very short processing times per photon to accommodate high counting rates, which can also cause increased electronic noise, which broadens spectral peaks. If the filters used to block BSE are too thin, the energy spectrum from the module X-ray sensor may exhibit increased background contributions due to BSE. "Pile-up" or "coincidence" artifacts may be present in the spectrum, resulting from corrupted photon energy measurements due to photon-induced pulses occurring too close in time to be detected as separate events. Furthermore, additional spectral artifacts may arise from X-rays generated outside the sample or from electrons that penetrate the filter material and reach the X-ray sensor.

[0102] If the auxiliary X-ray detector is mounted on one side of the module (e.g., as at 105 in FIG. 5 ), the solid angle subtended by the detector's active area at the point where the incident electron beam strikes the sample is likely to be much smaller than the solid angle subtended by an X-ray sensor in the module mounted below the pole piece (e.g., at position 106 in FIG. 5 ). As a result, the count rate of X-rays incident on the auxiliary X-ray detector may be much lower than the count rate obtained from the module. The auxiliary X-ray detector may itself contain multiple X-ray sensors and may not be able to accommodate high photon count rates. However, the sensitivity of the auxiliary detector to low-energy X-rays may be much higher than the module when it is mounted on the electron trap, thus avoiding the need for a material filter to block BSE from reaching the sensor. The auxiliary X-ray detector may also use a collimator to limit its field of view and reduce artifacts due to X-rays outside the sample area. Therefore, there is reason to believe that spectra obtained from accessory detectors are better indicators of the detailed elemental composition of a sample than spectra obtained from modules that are more susceptible to artifacts and less sensitive to low energy X-ray emissions. In some configurations, in addition to modules, there may be more than one accessory detector with a field of view of the sample.

[0103] A scanning electron microscope uses beam deflection to position the beam over a 2D grid on the sample surface, acquiring data while the beam is stationary at each point. These points correspond to pixel locations in a digital image of the scanned area on the sample. At each pixel, an X-ray sensor in the module provides a vector of data corresponding to the X-ray energy spectrum generated at that pixel. Additional signals may be recorded from electronic sensors in the module. Similarly, vectors of data are recorded from each of one or more attached detectors.

[0104] The pixel data obtained from the module represents a "hyperspectral image" in which a "data vector" exists at each pixel location. The data vector can be a histogram of X-ray photon energies representing the X-ray spectrum specific to the material excited by the incident electron beam at that point. This histogram can typically include 2000 bins or channels, where each channel value represents the photon counts in a 10 eV energy range, thus spanning the entire 20 keV energy range. While the full data vector consisting of this 2000 values ​​can be used, smaller data vectors can be obtained by summing the counts in successive energy regions. For example, aggregating the counts as a group of 10 channels spanning 100 eV reduces the data vector representing the spectrum to 200 values, where each value represents the number of photons recorded in a 100 eV energy range. The number of channels aggregated is not fixed but can be varied so that the aggregated counts correspond to an energy range that varies with energy. For example, each range can be proportional to the energy resolution of the spectrometer at that energy, providing a reduced data vector containing sufficient information to distinguish between spectra from different materials. Alternatively, an even smaller data vector can be derived simply by summing data from a series of energy ranges, not necessarily contiguous, with the limits of each energy range selected to maximize the likelihood of obtaining information that will aid in detecting regions of different composition, for example, using techniques described in U.S. Pat. No. 7,533,000. Similarly, energy ranges can be positioned to accumulate signals dominated by bremsstrahlung or energies for important characteristic elemental emission lines. Data vector values ​​corresponding to X-ray photon counts from energy regions dominated by bremsstrahlung emission will depend strongly on the weighted average atomic number of the sample, while data vector values ​​from energy ranges spanning the characteristic elemental emission lines will depend strongly on the mass concentration of the elements in the sample. Typically, the sum of signals from electron sensors positioned symmetrically around the electron beam axis represents the weighted average atomic number of the material, which for pure elements is a monotonically increasing function of atomic number.However, because some multi-element materials with different chemical compositions have similar weighted average atomic numbers and give similar backscattered electron signals, the electron signal alone may not always be able to distinguish between materials. However, a data vector constructed from a combination of X-ray data and electron signal measurements is more likely to be able to distinguish between pixels in regions of different chemical composition than simply the X-ray data or the electron signal data alone.

[0105] In an electron microscope, the intensities of the backscattered electron and X-ray signals are directly affected by the current of the electron beam impinging on the sample. During data acquisition, the beam current may drift or fluctuate. Because both X-rays and BSE are affected by changes in beam current to the same extent, the data vector can be made independent of the beam current by scaling the values ​​obtained from the X-ray spectrum by a factor inversely proportional to the BSE signal. Thus, for example, if the X-ray photon count from a certain energy range is X and the BSE signal is B, the scaled data vector value is CX / B, where C is an appropriate constant. By scaling all acquired X-ray count values ​​in this way, data vectors from anywhere in the scan region will be similar when the same material is under the beam, even if the beam current drifts during acquisition. Furthermore, because the BSE signal varies with the average Z of the sample material, the scaled data vector values ​​will be sensitive to the effect of low-Z elements on the average Z of the material, even if the module's X-ray sensor does not see X-rays from low-Z elements.

[0106] Attaching a material filter to the module's X-ray sensor to prevent BSE from reaching it results in a very poor response to low-energy X-rays. For example, if a filter blocking BSE up to 20 keV is used, the X-ray energy spectrum acquired by the module will have almost no useful information content at X-ray energies below 1 keV. Although the attached X-ray detector captures a much smaller percentage of the generated X-rays than the module, useful spectral information will be present for X-ray energies below 1 keV. Therefore, the X-ray spectrum acquired by the attached X-ray detector for energies below 1 keV can be used to augment the data vectors obtained from the module data. Data vectors containing this additional information obtained from the attached detector for low-energy X-rays can better distinguish materials containing low atomic number elements, such as boron, carbon, nitrogen, oxygen, and fluorine.

[0107] When the electron beam is scanned over the area of ​​the sample, the field of view on the sample is determined by the coordinate x i ,y i The image can be thought of as being covered by a grid of n locations, denoted by (i=1,n), where i is the pixel number in the digital image. The hyperspectral data array M is a set of n data vectors m for pixel numbers i ranging from 1 to n, corresponding to the digital image of the field of view. i Each data vector m i L contains the data obtained while the beam is positioned at the ith pixel position on the sample. m values, where the values ​​typically represent component intensities obtained from an X-ray spectrum derived from the module, although as noted above, some values ​​may be obtained from an X-ray spectrum from an attached X-ray detector, and one of the values ​​may be a backscattered electron signal measurement. Preferably, the L m For each of the values, the measurement uncertainty is estimated. i The kth value of m ik , and the associated measurement uncertainty σik If we name the vectors ρ and q ρ, where p and q are pixel numbers for two different locations on the sample, then we can derive a metric that indicates the difference between the vectors at these two locations. This metric can be, for example, a standardized Euclidean distance metric using statistical weighting: D pq =Σ[(m pk -m qk ) 2 / (σ pk 2 +σ qk 2 )] (1) where the sum is from 1 to L m for all values ​​of k up to D pq A small value of indicates that the material composition at pixel locations p and q on the sample is similar. The metric in equation (1) is the m It is appropriate when there is a diagonal covariance matrix for the values ​​(as is the case for X-ray counts accumulated from separate energy bands) and is a special case of the Mahalanobis distance. Including measurement uncertainty in this metric gives more weight to vector values ​​with higher precision, for example because they represent a larger overall number of X-ray counts totaled over a certain energy range.

[0108] A grouping algorithm (e.g., Statham et al., Microscopy and Microanalysis, Vol. 1, No. 9 (Suppl. 2), 2013, p. 752) can be used to group pixels with similar data vectors, thereby identifying regions where the material under the beam at each pixel is the same. The algorithm identifies groups of pixels with similar data vectors and assigns them to "phases," thus segmenting the image field into regions corresponding to "phases" with different material compositions. Phase membership may require, for example, that the Mahalanobis distance from the mean vector for all elements of a phase for a data vector be less than a threshold. The signal-to-noise ratio (S / N) for X-ray data depends on the counts governing Poisson statistics recorded in the energy spectrum; the counts recorded in spectra from the module may be much higher than those recorded in spectra from an attached detector, which has a smaller solid angle at the sample due to the proximity of the sensor in the module. Therefore, the signal-to-noise (S / N) of module data is expected to be much better than that of data collected from any attached X-ray detector, and a high S / N ratio improves the ability to reduce overlapping data vector clouds from materials with similar compositions. However, while module data with a high S / N ratio is very useful for segmenting the field of view into pixels corresponding to different materials, the spectral quality of X-ray data from partial polepiece modules may not be sufficient to identify the exact chemical elements present in different materials. As mentioned above, spectra may contain artifacts that are mistaken for elements not present in the material, low-energy characteristic X-rays may be missing due to absorption in the filter, spectral resolution may not be sufficient to resolve some overlapping characteristic spectral peaks, and BSE transmitted by the filter may add to the spectral background. Nevertheless, data vectors from modules will serve as "fingerprints" for pixels of the same material, so spatial regions of similar chemical composition can be represented using data from one module from another.

[0109] Preferably, the X-ray spectrum vectors obtained from the attached detector are summed for all pixels within a phase when similar regions (or "phases") of chemical composition are all represented by grouping pixels with similar data vectors. Some pixels, for example, those adjacent to boundaries, may need to be excluded because they may not be as representative of the true material composition. This can occur, for example, when the region of X-ray emission extends beyond the dimensions of a single pixel. Thus, aggregation of many spectra obtained from pixel locations on the same material improves the signal-to-noise ratio of the spectrum obtained by the attached X-ray detector for that "phase." Because the attached detector is highly sensitive to low-energy X-rays and optimized for removing spectral artifacts, conventional spectral processing techniques can be applied to the aggregated spectrum to identify elements contributing to characteristic emission peaks in the spectrum and determine the intensity of those peaks. Thus, data from one or more attached detectors can be used to provide a more accurate estimate of the elemental composition of a phase than can be obtained from partial polepiece module data alone. By combining the high signal-to-noise data of the partial pole piece module, which distinguishes regions of different composition, with more accurate elemental composition data from one or more additional detectors, accurate elemental content can be determined for the entire scanned field of view on the sample.

[0110] The principle of dividing the field of view into regions of similar X-ray spectral emission can also be utilized in other types of modules that are not necessarily positioned below a pole piece or that do not necessarily include an electronic sensor. The key requirement is that one or more X-ray sensors in the module subtend a total solid angle at the probe spot on the sample that is much larger than the total solid angle subtended by an attached X-ray detector with better spectroscopic performance. An electron trap, which uses a magnetic field to deflect electrons from reaching the X-ray sensor, must be inserted between the sample and the X-ray sensor, preventing the X-ray sensor from being positioned too close to the sample to achieve a high solid angle for X-ray signal collection. The advantage of the electron trap is that it does not prevent low-energy X-rays from reaching the sensor. However, if a filter material is used in front of the sensor face to block electrons, the sensor can still operate effectively even when positioned close to the sample. Therefore, a module incorporating an X-ray sensor with an electron-blocking material filter can achieve a much higher solid angle than a detector with an electron trap. Such a detector module can be positioned, for example, below the pole piece, as in Figures 1 and 17, but can alternatively be mounted in front of the detector equipped with the electron trap. Figure 19 shows an example of a module with two X-ray sensors mounted on the end of a detector equipped with an electron trap. The two X-ray sensors would be covered with a material filter to block BSE and achieve a higher collection solid angle than the detector behind the electron trap. To achieve an even higher collection solid angle, an additional sensor equipped with a material filter can be mounted on the end of the trap closest to the sample. Furthermore, one or more sensors with material filters can be mounted below or on either side of the entrance aperture of the electron trap. To benefit from this module, the statistical accuracy of the data must be significantly better than that of the attached detector. Therefore, the total collection solid angle subtended by the probe spot by the module sensors must be at least three times, preferably more than five times, and ideally more than ten times that of the attached detector.

[0111] As noted above, X-ray spectra obtained from a sensor with a material BSE filter close to the sample are subject to stray radiation and artifacts and will not exhibit characteristic X-ray emission peaks for low-energy X-ray photons that are strongly absorbed by the material BSE filter. However, such X-ray spectra will still vary depending on the material composition beneath the probe spot on the sample due to the composition's effect on the relative intensities of bremsstrahlung and high-energy characteristic radiation. Furthermore, such X-ray spectra will have higher counts and lower statistical noise than typical spectra from an attached detector equipped with an electron trap. Thus, a hyperspectral image formed by X-ray spectral data collected at an array of pixel locations about the sample from one or more sensors covered by a BSE filter can be used to partition the field of view into regions of similar composition, where each value in the data vector is a count recorded for a specific range of X-ray energies. For a set of pixels with similar data vectors (arising from materials of similar composition), simultaneously recorded spectra from an attached X-ray detector with an electron trap can be summed to provide a spectrum with improved count statistics that can be analyzed to determine the elemental composition of the materials corresponding to those pixels. In effect, the module data is used to improve the S / N of the x-ray elemental map or image obtained from the attached detector, and there are alternative ways to achieve this improvement.

[0112] The flow diagram in Figure 20 summarizes one procedure for acquiring and combining data from the modules and attached detectors to achieve an image showing how material composition varies across the field of view. The steps within the dashed box describe the data acquisition process. As the electron beam is scanned across an area of ​​the sample, the field of view on the sample is determined by coordinates x i ,y i The image can be thought of as being covered by a grid of n locations, denoted by (i=1,n), where i is the pixel number in the digital image. The hyperspectral data array M is a set of n data vectors m for pixel numbers i ranging from 1 to n, corresponding to the digital image of the field of view. iEach data vector m i contains the data acquired while the beam is positioned at the ith pixel location on the sample, and L m n values, where the values ​​typically represent component intensities obtained from an X-ray spectrum derived from the module, although as noted above, some values ​​may be obtained from an X-ray spectrum from an attached X-ray detector, and one of the values ​​may be a backscattered electron signal measurement. Similarly, a hyperspectral data array A is made up of n data vectors a i (i=1,n), and each data vector a i is a function L whose values ​​represent the intensity of an energy band or channel for the X-ray spectrum obtained from the attached detector when the beam is positioned at the ith pixel position on the sample. a (The attached X-ray detector typically obtains spectra with much fewer counts than those obtained from the module, but with better spectral fidelity, resulting in more accurate estimates of chemical element abundances, even if the statistical precision is much poorer.) If more than one attached detector is available, the X-ray spectral data from all attached detectors can be combined.

[0113] A processing algorithm is used to identify k groups of pixels, and the group g p (p=1,k) is the data vector m j contains all pixels for which m meets some similarity criterion. For example, a data vector m in n-dimensional space jThere are many different algorithms available for dividing an image into groups of similar pixels (see, for example, the Wikipedia entry https: / / en.wikipedia.org / wiki / Cluster_analysis), where a pixel can be a member of a group if it is within a predetermined Mahalanobis distance from the centroid of all vectors in the group and / or if its centroid is the closest centroid in n-dimensional space. A group of pixels comprises a subregion of the overall area on the sample covered by the scan, and the grouping technique divides the area into subregions that are not necessarily spatially contiguous. Groups of similar pixels identified using module data M may contain "bad" pixels whose material content is located at a location quite different from the average for the group. Such "bad" data may appear, for example, at a pixel located at the boundary of two materials; the observed signal is a mixture of emissions from the two different materials, and the measured signal vector from the module may accidentally cause these "bad" pixels to be considered part of the group. An optional approach to this problem is to identify similar regions by looking at signal variations near each pixel or by grouping before considering the pixel data, and then exclude data from pixels at or near boundaries between dissimilar materials as a post-processing step after the boundaries can be identified.

[0114] In the method of Figure 20, the group g p For pixel j in data vector a j The spectrum a that represents the group by aggregating p If the "grouping" procedure fails to properly assign pixels to the appropriate group, resulting in "bad" pixels being included in the group, a p may be biased and not truly representative. To reduce this bias, an optional step is to p Each data vector in a jThis may involve testing all pixels in the group to see if g is within the expected range of variation from the mean or median vector for the entire group of data vectors. Any pixels that appear to be group outliers are removed from the group before calculating the overall group sum or aggregate spectrum. The test for outliers may involve a threshold test according to the similarity distance measure used for grouping, or a statistical test based on Poisson counting statistics for X-ray spectra. The removed pixels may be omitted from the final display to avoid misleading the observer. Alternatively, the group g may be first p Overall average data vector JPEG0007744405000001.jpg63, then a j but If it is significantly different from JPEG0007744405000002.jpg63, j The effect of "bad" pixel data can be reduced by recalculating a new weighted average that strongly reduces the weighting of a. j The appropriate weighting of JPEG0007744405000003.jpg8150 is proportional to the sum of 1 and L a Over k up to σ jk 2 is the spectral vector a j The expected variance for the k-th component of a is given by: A different method of weighting uses the same principle, but uses a data vector from M, and the weighting of aj is JPEG0007744405000004.jpg8150, where the sum is from 1 to L m Over r up to σ jr 2 is the data vector m j is the expected variance for the r-th component of JPEG0007744405000005.jpg64 is the mean data vector of group gp. If the grouping algorithm incorporates "bad" pixels within the group, this weighting will penalize outliers that cannot be explained by statistical variation alone, giving a more appropriate mean spectrum of the group.

[0115] Spectrum a p can be processed to correct for artifacts, bremsstrahlung background, and peak overlap, so that it is possible to determine the area for a peak characteristic of a particular element q (e.g., using the techniques described in Statham, Journal of Research of the National Institute of Standards and Technology, Vol. 1, No. 07, pp. 531-546 (2002)). Furthermore, once the peak areas for all elements have been determined, known procedures of electron probe microanalysis (e.g., the methods described in Goldstein et al., "Scanning Electron Microscopy and X-ray Microanalysis," ISBN: 0-306-47292-9) can be used to determine the material composition in terms of the mass fractions of the constituent elements. Thus, the area of ​​the group g associated with element q corresponds to either the intensity of the characteristic X-ray emission of that element or the mass fraction of that element. p Intensity values ​​representing all pixels in I q can be determined, and this intensity value can then be used to generate the output image E for element q q Group g in p , and preferably every pixel location.

[0116] It is contemplated that the optional step of handling "bad" pixels may have been unused or incomplete, potentially leaving image pixels that are inconsistent with the input data. Rather than displaying misleading information, it may be preferable to filter out such pixels before display. For example, for an element q that has a high-energy X-ray emission sensed by the module and appears in the data M, the data vector m for each pixel j is j An elemental map can be constructed directly from M by processing M to extract the X-ray emission intensity of element q at pixel j. Although this intensity is considered insignificant (e.g., because it is in the range of what is considered random or systematic error), the output image E qIf,shows a large contribution at the same pixel j, this inconsistent pixel,E, q can be excluded by setting the intensity to 0 in m. A similar consistency check can be applied using the auxiliary data A, but is less effective because the random statistical error is much larger than for M, except for elements with low-energy X-ray emission. The statistical error of A can be reduced by summarizing data vectors in the vicinity of pixel j. This "data smoothing" or "local averaging" process spatially blurs the data, so there is a trade-off between spatial resolution and statistical noise, which depends on how large a neighborhood is used for the summation.

[0117] The entire output is c digital images E q (q=1,c), each image having n pixels corresponding to a corresponding chemical element q, and image E q The intensity value of pixel i at corresponds to either the X-ray emission intensity of element q or the material concentration (e.g., mass fraction). An additional image can be generated in which pixels in a particular group, preferably all pixels, are assigned a specific color representing a phase, and this color image can be combined with the black and white electronic image as an overlay, highlighting regions of different chemical composition with different colors. Similarly, the digital image E q (q=1,c) can be assigned colors suitable for overlay with electronic images using techniques such as those described in PCT / GB2011 / 051060 or US5357110, for example.

[0118] An alternative procedure is described in Figure 21. The data acquisition and grouping steps are the same as in Figure 20. Instead of obtaining a single aggregate spectrum for a group of similar pixels, the principle of non-local means (NLM) averaging is to denoise all the individual pixel spectra within the group gp (see, for example, Manjon et al., "Multispectral MRI de-noising using non-local means," Proc. MIUA'07, pp. 41-45, Aberystwyth, Wales, 2007). This averaging involves weighting the spectra according to a similarity function based on the differences between the spectra and is applied only to pixels within the same group obtained by the grouping algorithm. For example, for a group g p Spectrum a in j A new "denoised" version of a j To find ', we use a i The weighted average including all other spectra such as exp(-Σ[(a jk -a ik ) 2 / (2(σ jk 2 +σ ik 2 ))]) using weighting factors proportional to the sum of 1 to L a Over k up to σ jk 2 is the spectral vector a j is the expected variance for the kth component of . An alternative weighting scheme for noise removal is one where the weights are exp(-Σ[(m jr -m ir ) 2 / 2(σ jr 2 +σ ir 2 ))) where the sum is from 1 to L m Over r up to σ jr 2 is the data vector m jis the expected variance of the r-th component of

[0000] . Similar to the procedure in Figure 18, an additional step can be added to reduce the influence of outlier data due to "bad" pixels that are mistakenly included in groups as a result of the grouping process.

[0119] The following illustrates an example of the results achieved with this invention. The modules are equipped with X-ray sensors, each covered with a filter composed of 6-micron-thick Mylar. The sensors subtend a total solid angle of 0.438 steradians at the sample. An attached X-ray detector equipped with an electron trap is mounted on a port on the side of the electron column and subtends a solid angle of 0.044 steradians at the sample. A focused 20 keV electron beam is raster-scanned across the sample over a roughly square area approximately 200 microns wide, and the resulting spectra from both the module sensor and the attached detector at each pixel location are recorded to construct a corresponding 256 x 256 resolution hyperspectral image dataset. The module sensor receives an average photon count rate of 1250 kcps, while the attached detector only has a photon count rate of 125 kcps due to its smaller solid angle.

[0120] The data was acquired in 2 seconds, with the beam covering all 256 x 256 pixel locations. Each spectrum is effectively an energy histogram with 1024 energy bins, each 20 eV wide. The "total spectrum" is formed by summing the counts in each bin for every pixel in the field of view; Figure 22 shows a comparison of the total spectrum from the module data with the total spectrum from the ancillary detector data. The module data has about 10 times the counts of the ancillary data, but almost no counts below 1 keV, whereas the ancillary data has a significant characteristic peak corresponding to oxygen (OCa) at energy around 0.5 keV, due to the presence of electron traps and the elimination of the need for a material filter to block BSE.

[0121] In conventional X-ray mapping, for each element of interest, counts from all channels falling within the energy range spanning the main characteristic emission peak in the spectrum are summed. The total count for the element at the pixel location determines the intensity of the displayed pixel. Figure 23 shows conventional X-ray maps obtained from the module's hyperspectral dataset for a set of elements that contribute peaks to the sum spectrum. The map intensity is brightness-adjusted so that low-count level map details remain visible. Regions of high oxygen content appear in the oxygen map (OKα1), even though the module sensor is insensitive to such low-energy X-rays and the map is expected to be essentially free of counts. This observed intensity is due to an artifact called the "silicon escape effect," in which a small percentage of photons absorbed by a silicon-based sensor do not convert all of the photon energy into charge because SiKα photons escape the sensor. That is, a small fraction of the 2.31 keV sulfur Kα radiation, which easily passes through the Mylar filter, appears as a spectral artifact at 0.57 keV, thereby falling within the energy range used to map OKα1 intensity. The resulting typically blank OKα1 map exhibits artifacts that follow the same distribution as SKα1-emitting materials and therefore appears to be an SKα1 map. Figure 24 shows the corresponding elemental distribution maps derived from the hyperspectral dataset of the attached detector. Most elemental maps have much fewer counts than Figure 23 and high levels of statistical (Poisson) noise, resulting in poor signal-to-noise. The exception is the OKα1 map, which, while still noisy, represents the true distribution of oxygen-containing materials.

[0122] The X-ray map data from the modules is used as input to a grouping algorithm (Statham et al., Microscopy and Microanalysis, Vol. 1, No. 9 (Suppl. 2), 2013, p. 752), which finds similar groups of data vectors, in this case X-ray spectra, from the modules. The algorithm finds six distinct groups, and the pixels within each of these groups are shown as the "phase" distribution in Figure 25. The image labeled "Unassign." shows boundary pixels identified by the algorithm that cannot be assigned to any group.

[0123] For each "phase," preferably for all pixel locations within the phase, X-ray spectra from the attached detector's hyperspectral dataset are summed to generate a single spectrum representing the phase. Some pixels, such as those adjacent to boundaries, may need to be excluded because they may not be as representative of the true material composition. This can occur, for example, when the region of X-ray emission extends beyond the dimensions of a single pixel. The single spectrum is processed to correct for artifacts, background, and peak overlap and to determine the intensities of the characteristic X-ray emission. These intensities are then used as inputs to an algorithm to correct for various effects, such as electron scattering, X-ray generation, X-ray absorption within the sample, and the efficiency of X-ray detection, and to estimate the mass fraction of each element in the material that contributed to the spectrum. Once the chemical content of a "phase" has been quantified, the mass fractions of the elements within that "phase" are used to control the intensity values ​​used for pixel locations for that "phase" in the elemental output distribution image.

[0124] Figure 26 shows the final elemental distribution images generated by this procedure. The signal-to-noise ratio of these images is much higher than that of the original attached detector map shown in Figure 24. The representation of the different compositional regions is much sharper than possible using attached detector data alone, reflecting the reduction in statistical noise in the data obtained from the module. Figure 27 shows a "phase" map representing the distribution of different materials obtained by combining the individual phase distributions from Figure 25. A separate color is typically assigned to each phase for overlay on a corresponding electron image obtained from the same region of the sample.

[0125] The present invention can be further understood by reference to the following numbered clauses:

[0126] Article 1. An apparatus for analyzing a sample, comprising: an electron beam assembly for generating a focused electron beam; a first detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a first total solid angle at a point where the electron beam strikes the sample; a second detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a second total solid angle at the point where the electron beam strikes the sample; Equipped with Both detectors receive X-rays generated by the interaction between the electron beam and the sample and can use X-ray sensor elements to detect individual photons and measure their energy; the first total solid angle is at least three times the second total solid angle; spectral data from the first and second detectors are recorded as the electron beam is scanned over an area of ​​the sample; using the spectral data from the first detector and optionally the data from the second detector to identify sub-regions in which the recorded signals are similar with respect to points within the given sub-region; combining the spectral data from the second detector for some, preferably all, of the locations within the sub-region to generate a single spectrum representative of the material within the sub-region; processing the spectra representative of the subregions to determine intensity values ​​of one or more characteristic elemental X-ray emissions and, optionally, the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning elemental intensity or concentration values ​​derived from the representative spectrum of the subregion to points, preferably every image point, within the subregion; providing a visual representation of the element distribution across the region of the sample using the image data collected for the elements for the identified sub-regions; Device.

[0127] Clause 2. The apparatus of clause 1, wherein data for points within the identified sub-regions is inspected before the spectral data from the second X-ray detector is combined to exclude points from the sub-regions whose point data is outside an expected range of variation for data for points within the sub-region.

[0128] Clause 3. An apparatus as described in clause 1 or clause 2, wherein spectral data from the second X-ray detector for points within the identified sub-region is aggregated to generate a single spectrum for the sub-region using a weighted combination of spectra, where the weight of the spectrum depends on a measure of the difference between that spectrum and an average spectrum for the entire sub-region.

[0129] Clause 4. An electron beam assembly for generating a focused electron beam; a first detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a first total solid angle at a point where the electron beam strikes the sample; a second detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a second total solid angle at the point where the electron beam strikes the sample; 1. A method of analyzing a sample using an apparatus comprising: Both detectors receive X-rays generated by the interaction between the electron beam and the sample and can use X-ray sensor elements to detect individual photons and measure their energy; the first total solid angle is at least three times the second total solid angle; spectral data from the first and second detectors are recorded as the electron beam is scanned over an area of ​​the sample; using the spectral data from the first detector and optionally the data from the second detector to identify sub-regions in which the recorded signals are similar with respect to points within the given sub-region; combining the spectral data from the second detector for some, preferably all, of the locations within the sub-region to generate a single spectrum representative of the material within the sub-region; processing the spectra representative of the subregions to determine intensity values ​​of one or more characteristic elemental X-ray emissions and, optionally, the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning elemental intensity or concentration values ​​derived from the representative spectrum of the subregion to points, preferably every image point, within the subregion; providing a visual representation of the element distribution across the region of the sample using the image data collected for the elements for the identified sub-regions; How to prepare for this.

[0130] Clause 5. The method of clause 4, wherein data for points within the identified sub-regions is inspected before the spectral data from the second X-ray detector is combined to exclude points from sub-regions where the point data is outside an expected range of variation for data for points within the sub-region.

[0131] Clause 6. A method according to clause 4 or clause 5, wherein spectral data from the second X-ray detector for points within the identified sub-region is aggregated to generate a single spectrum for the sub-region using a weighted combination of spectra, where the weight of the spectrum depends on a measure of the difference between that spectrum and an average spectrum for the entire sub-region.

[0132] Article 7. An apparatus for analyzing samples, comprising: an electron beam assembly for generating a focused electron beam; a first detector for detecting photons including an x-ray sensor element, the x-ray sensor element subtending a first total solid angle at the point where the electron beam strikes the sample; a second detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a second total solid angle at the point where the electron beam strikes the sample; Equipped with Both detectors receive X-rays generated by the interaction between the electron beam and the sample and can use X-ray sensor elements to detect individual photons and measure their energy; the first total solid angle is at least three times the second total solid angle; spectral data from the first and second detectors are recorded as the electron beam is scanned over an area of ​​the sample; using the spectral data from the first detector and optionally the data from the second detector to identify sub-regions in which the recorded signals are similar with respect to points within the given sub-region; for a sub-region, preferably any sub-region, combining the spectral data from the second X-ray detector for each point within the sub-region with data from other points, preferably all other points, within the sub-region by weighted averaging to produce a denoised version of the spectrum at that point; processing the denoised spectrum for each point to determine intensity values ​​of one or more characteristic elemental X-ray emissions and, optionally, the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning elemental intensity or concentration values ​​derived from the denoised spectrum to the image points, preferably every image point; using the image data collected for the elements to provide a visual representation of the element distribution over an area of ​​the sample; Device.

[0133] Clause 8. The apparatus of clause 7, wherein the weighted average for the denoised spectrum at a point uses weighting factors for each other point in the sub-region that depend on a measure of the difference between the spectral data from the second X-ray detector or the spectral data from the first X-ray detector for that point and the corresponding spectral data from the second detector or the first detector for each other point in the sub-region.

[0134] Clause 9. The apparatus of clause 7 or clause 8, wherein data for points within the identified sub-regions is examined to exclude points from sub-regions where the point data is outside an expected range of variation for points within the sub-region.

[0135] Clause 10. An electron beam assembly for generating a focused electron beam; a first detector for detecting photons including an x-ray sensor element, the x-ray sensor element subtending a first total solid angle at the point where the electron beam strikes the sample; a second detector for detecting photons including an X-ray sensor element, the X-ray sensor element subtending a second total solid angle at the point where the electron beam strikes the sample; 1. A method of analyzing a sample using an apparatus comprising: Both detectors receive X-rays generated by the interaction between the electron beam and the sample and can use X-ray sensor elements to detect individual photons and measure their energy; the first total solid angle is at least three times the second total solid angle; spectral data from the first and second detectors are recorded as the electron beam is scanned over an area of ​​the sample; using the spectral data from the first detector and optionally the data from the second detector to identify sub-regions in which the recorded signals are similar with respect to points within the given sub-region; combining, by weighted averaging, the spectral data from the second X-ray detector for each point within a sub-region, preferably for any sub-region, with data from other points, preferably all other points, within the sub-region to generate a denoised version of the spectrum at that point; processing the denoised spectrum for each point to determine intensity values ​​of one or more characteristic elemental X-ray emissions and, optionally, the concentrations of the corresponding elements responsible for those emissions; aggregating image data for one or more elements by assigning elemental intensity or concentration values ​​derived from the denoised spectrum to the image points, preferably every image point; using the image data collected for the elements to provide a visual representation of the element distribution over an area of ​​the sample; How to prepare for this.

[0136] Clause 11. A method according to clause 10, wherein the weighted average for the denoised spectrum at a point uses weighting factors for each other point in the sub-region that depend on a measure of the difference between the spectral data from the second X-ray detector or the spectral data from the first X-ray detector for that point and the corresponding spectral data from the second detector or the first detector for each other point in the sub-region.

[0137] Clause 12. The method of clause 10 or clause 11, wherein data for points within the identified sub-regions is examined to exclude points from sub-regions where the point data is outside an expected range of variation for points within the sub-region.

[0138] Clause 13. Apparatus for analyzing samples in a scanning electron microscope, comprising: a detector module for detecting backscattered electrons; Equipped with a detector module positioned directly below a pole piece for a final lens of the electron beam assembly through which the electron beam passes before striking the specimen; An apparatus fitted with an optical module, wherein the camera has a field of view of the sample that is at least 10 mm wide when the camera-to-sample distance is less than 10 mm, and optionally the depth of field of the camera is sufficient such that the width of the field of view can be increased to at least 20 mm or preferably greater than 60 mm by increasing the camera-to-sample distance.

[0139] Clause 14. Apparatus according to clause 13, wherein two optical cameras are fitted into the module, the camera fields of view relative to the sample overlapping, and optionally the two images can be used to provide a stereoscopic view of the sample and / or data from the cameras can be used to calculate a topographical map of the sample surface. [Explanation of symbols]

[0140] BSE backscattered electrons SDD Silicon Drift Detector

Claims

1. 1. A detector module for use in an apparatus for analyzing a sample, comprising: the detector module comprises a plurality of x-ray sensor elements and one or more electronic sensor elements; the detector module is positioned below a pole piece of an electron beam assembly of the apparatus from which, in use, an electron beam generated by the assembly emerges towards a specimen, the detector module being shaped to at least partially surround the electron beam, the detector module being adapted to receive x-rays and backscattered electrons generated by interaction between the electron beam and the specimen; each of the plurality of X-ray sensor elements is configured to monitor the energy of an individual received X-ray photon; The plurality of X-ray sensor elements are 20 mm 2 and has a total effective area greater than A detector module, wherein the radial length of the detector module relative to the electron beam axis in use is less than 10 mm in at least a first portion of the detector module, so as to provide an accessory device positioned to the side of the electron beam assembly in use with an unobstructed line of sight to the spot where the electron beam impinges on the sample.

2. The detector module of claim 1 , wherein the radial length of the first portion of the detector module is less than 7 mm.

3. 3. The detector module of claim 1, wherein more than half of the total active area of ​​the plurality of X-ray sensor elements is less than 6 mm from the electron beam axis.

4. The one or more electronic sensor elements are 30 mm 2 4. A detector module according to claim 1, having a total active area greater than 1000 mV.

5. 5. A detector module according to claim 1, wherein the X-ray sensor elements are arranged with no more than two-fold rotational symmetry about the electron beam axis in use, and the one or more electronic sensor elements are arranged with at least two-fold rotational symmetry about the electron beam axis in use.

6. a radial distance from the electron beam axis to a first location within an active area of ​​the one or more electronic sensor elements is greater than a maximum radial length of the active area of ​​the plurality of X-ray sensor elements relative to the electron beam axis when in use; 6. A detector module according to claim 1, wherein a radial distance from the electron beam axis to a second location within an active area of ​​the one or more electronic sensor elements is shorter than a minimum radial distance between the electron beam axis and an active area of ​​the plurality of X-ray sensor elements in use.

7. 7. The detector module of claim 6, configured such that a signal from a first portion of the active area of ​​the one or more electronic sensor elements that is positioned furthest from the electron beam axis in use is amplified before it is added to a signal from a second portion of the active area of ​​the one or more electronic sensor elements that is different from the first portion.

8. 8. A detector module according to claim 1, wherein the inner diameter of an opening included in the detector module and arranged to allow the electron beam to pass through in use is greater than 1.0 mm.

9. A detector module as described in claim 1, wherein the accessory device comprises any of an X-ray sensor, a backscattered electron sensor, a cathodoluminescence sensor, a micromanipulator, a gas injection device, a laser, and an X-ray sensor with an electron trap.

10. 10. The detector module of claim 1, further comprising one or more material layers disposed on or over at least a portion of an active area of ​​the plurality of X-ray sensor elements and blocking transmission of any one or more of electrons, visible light, and infrared light while allowing transmission of X-rays having energies within a first energy range.

11. The detector module of claim 10 , wherein one of the one or more layers of material is a coating applied to a surface of the portion of the active area of ​​the plurality of X-ray sensor elements.

12. 12. A detector module as described in claim 10 or 11, wherein one of the one or more material layers is electrically grounded or connected to a fixed voltage supply and is conductive to prevent the potential of the material layer from rising due to electrostatic charging caused by incident electrons.

13. 13. A detector module according to any preceding claim, further comprising a cooling controller configured to maintain the one or more electronic sensor elements within a predetermined temperature range in use.

14. 14. The detector module of claim 13, wherein the cooling controller is configurable to maintain the one or more electronic sensor elements within a selected one of a plurality of predetermined temperature ranges, each of the temperature ranges corresponding to an optimal operating temperature range for a respective operating mode of the electron beam assembly.

15. 15. The detector module of claim 1, wherein each of the plurality of X-ray sensor elements is arranged behind an aperture in a conductive plate configured to function as a first electrode of a capacitive sensor configured to monitor the proximity of the sample and / or sample holder.

16. 16. A detector module according to any one of claims 1 to 15, further comprising an optical camera arranged to have a field of view of the sample spanning at least 10 mm with a camera-sample distance of less than 10 mm, the depth of field of the camera being sufficient to allow the width of the field of view to be increased to at least 20 mm by increasing the camera-sample distance.

17. configured to perform a method for displaying the location of an area being analyzed under the electron beam in an optical image of the sample being inspected, the surface of the cathodoluminescence sample is positioned at a specific working distance from the final lens pole piece of the electron microscope; a digitized optical image is acquired with the illumination source turned off while a focused electron beam is incident on the sample; determining the position coordinates within the optical image of the center of the spot of light emitted by cathodoluminescence; the surface of the sample to be inspected is positioned at the same specified working distance as the cathodoluminescence sample, and a digitized optical image is acquired while illuminating the sample with a light source; 17. The detector module of claim 16, wherein the optical image of the sample being inspected is shown on a visual display highlighting the location of an electron beam analysis area centered on the position coordinates obtained from the optical image obtained from the cathodoluminescence sample.

18. configured to perform a method for displaying the location of an area being analyzed under the electron beam in an optical image of the sample being inspected, a surface of a small, recognizable feature is focused at a specific working distance from the final lens pole piece of the electron microscope and an electron image is acquired with the feature at the center of the field of view; a digitized optical image is acquired with the illumination source on and the position coordinates of the center of the same recognizable feature in the image are determined; the surface of the sample to be inspected is positioned at the same specific working distance used to image the small recognizable features, and a digitized optical image is acquired while illuminating the sample with a light source; 17. The detector module of claim 16, wherein the optical image of the inspected sample is shown on a visual display with the location of an electron beam analysis area highlighted and centered on the position coordinates of the recognizable feature.

19. a first optical camera and a second optical camera having at least partially overlapping first and second fields of view of the sample, respectively; 19. A detector module according to any one of claims 1 to 18, wherein first and second optical cameras are arranged such that first and second images acquired by them, respectively, can be used to provide a stereoscopic representation of the sample and / or data from the cameras can be used to generate a topographical map of the sample surface.

20. configured to perform a method for obtaining and processing data, spectral data from an associated X-ray detector, spectral data from the plurality of X-ray sensor elements, and data from the one or more electronic sensor elements are recorded as the electron beam is positioned at a series of points covering an area of ​​the sample; spectral data from the module and data from the one or more electronic sensor elements and / or spectral data from an associated X-ray detector are used to identify sub-regions in which recorded signals are similar at points within a given sub-region; Spectral data from the associated X-ray detector for a set of locations within the sub-region are combined to generate a single spectrum representative of the material within the sub-region; the spectra representing the subregions are processed to determine intensity values ​​for one or more characteristic elemental X-ray emissions and the concentrations of the corresponding elements responsible for those emissions; image data is collected for one or more elements by assigning the intensity values ​​or concentrations for the elements derived from the spectrum represented by the subregion to image points within the subregion; 20. A detector module according to any preceding claim, wherein the image data collected for elements for identified sub-regions is used to provide a visual representation of element distribution across a region of the sample.

21. configured to perform a method for obtaining and processing data, spectral data from an attached X-ray detector, spectral data from the X-ray sensor elements in the module, and data from the one or more electronic sensor elements in the module are recorded as the electron beam is positioned at a series of points covering an area of ​​the sample; spectral data from the module and data from the one or more electronic sensor elements within the module and spectral data from an associated X-ray detector are used to identify sub-regions in which recorded signals are similar at points within the given sub-region; For each sub-region, spectral data from the associated x-ray detector for each point within the sub-region is combined by weighted averaging with data from other points within said sub-region to produce a denoised version of the spectrum at that point; the denoised spectrum for each point is processed to determine intensity values ​​for one or more characteristic elemental X-ray emissions and the concentrations of the corresponding elements responsible for those emissions; image data is collected for one or more elements by assigning the intensity values ​​or concentrations of the elements derived from the denoised spectrum to image points; 21. A detector module according to any preceding claim, wherein the image data collected for an element is used to provide a visual representation of element distribution over an area of ​​the sample.

22. 1. An apparatus for analyzing a sample, comprising: an electron beam assembly for generating a focused electron beam; A detector module according to any one of claims 1 to 21; An apparatus comprising:

23. the attachment device mounted on the electron beam assembly; an orientation of the detector module relative to the electron beam assembly such that at least a portion of the first portion of the detector module and at least a portion of the attachment device are aligned with a plane in which the electron beam axis lies; 23. The apparatus of claim 22, wherein the first portion is positioned, in use, between the electron beam axis and the accessory device.

24. 24. Apparatus according to claim 22 or 23, further comprising an accessory device comprising an X-ray detector fitted with an electron trap.

25. 1. A method for analyzing a sample, comprising: generating a focused electron beam using an electron beam assembly; providing a detector module comprising a plurality of x-ray sensor elements and one or more electron sensor elements, the detector module being positioned below a pole piece of the electron beam assembly from which the focused electron beam emerges toward the sample, the detector module being shaped to at least partially surround the electron beam, the detector module receiving x-rays and backscattered electrons generated by interaction between the electron beam and the sample; monitoring the energy of each received X-ray photon using the plurality of X-ray sensor elements; Equipped with The plurality of X-ray sensor elements are 20 mm 2 and wherein the radial length of the detector module relative to the electron beam axis is less than 10 mm in at least a first portion of the detector module so as to provide an accessory device positioned laterally of the electron beam assembly in use with an unobstructed line of sight to the spot where the electron beam impinges on the sample.

26. recording a signal comprising spectral data from an associated X-ray detector, spectral data from the plurality of X-ray sensor elements, and data from the one or more electronic sensor elements while impinging the electron beam at a series of points covering an area on the sample; using spectral data obtained by the detector module and data from the one or more electronic sensor elements therein and spectral data from an associated X-ray detector to identify sub-regions in which the recorded signals are similar at points within a given sub-region; combining spectral data from the associated X-ray detector for several locations within each sub-region to generate a single spectrum representative of the material within that sub-region; processing the spectrum representative of each subregion to generate data representing intensity values ​​for one or more characteristic elemental X-ray emissions and concentrations of corresponding elements associated with the one or more characteristic elemental X-ray emissions; collecting image data for one or more elements by assigning the intensity or concentration data for the element derived from the spectrum represented by each sub-region to image points within the sub-region; 26. The method of claim 25, further comprising using the image data collected for elements for the identified sub-regions to generate a visual representation of element distribution across the region of the sample.

Images