WO2021046118A1

WO2021046118A1 - Direct detection electron energy loss spectroscopy system

Info

Publication number: WO2021046118A1
Application number: PCT/US2020/049052
Authority: WO
Inventors: Mitra L. TAHERI; James L. HART
Original assignee: Drexel University
Priority date: 2019-09-03
Filing date: 2020-09-02
Publication date: 2021-03-11

Abstract

An electron energy loss spectroscopy (EELS) system has an electron gun directing an electron beam towards a sample, a dispersing magnet dispersing the electron beam after passing through the sample, and one or more planar detectors. If there is only one planar detector, then the planar detector is a gradient detector having a pixel array with multiple regions designed to detect electrons in different ranges of intensities, such that the dispersing magnet directs (i) high-intensity electron-beam portions towards one or more regions designed to detect high-intensity ranges and (ii) low-intensity electron-beam portions towards one or more regions designed to detect low-intensity ranges. If there are two or more planar detectors, then the two or more planar detectors are selectable detectors including (i) zero, one, or more monolithic detectors, each designed to detect electrons in a single intensity range, and (ii) zero, one, or more gradient detectors.

Description

DIRECT DETECTION

ELECTRON ENERGY LOSS SPECTROSCOPY SYSTEM

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application No. 62/895,180, filed on September 3, 2019, the entire disclosure of which is hereby incorporated by reference as if set forth fully herein.

STATEMENT OF GOVERNMENT INTEREST

This invention was made with government support under contract no. 1429661 awarded by the National Science Foundation. The Government has certain rights in the invention.

BACKGROUND

Field of the Disclosure

The invention relates to transmission electron microscopy (TEM) and electron energy loss spectroscopy (EELS).

Description of the Related Art

This section introduces aspects that may help facilitate a better understanding of the disclosure. Accordingly, the statements of this section are to be read in this light and are not to be understood as admissions about what is prior art or what is not prior art.

A first attempt to solve the problems solved by certain embodiments of the invention involved a conventional indirect detection sensor and a direct detection monolithic active pixel sensor mounted on an imaging filter, as described in “Direct Detection Electron Energy Loss Spectroscopy: A New Method to Push the Limits of Resolution and Sensitivity” (Reference 2 cited below). With this initial system, data on signal-to-noise ratio (SNR) and spectral range was collected. However, this indirect detection sensor had poor resolution and a poor SNR in the low loss EELS region.

SUMMARY

This invention includes but is not limited to the use of monolithic active pixel sensors, hybrid pixel sensors, as well as proposed gradient detectors which incorporate multiple detection schemes into a single detector platform. This includes both the use of multiple detector technologies, e.g., a monolithic pixel sensor and a hybrid pixel sensor on the same device, as well as a single detector technology operated with different conditions across the detector area, e.g., a monolithic active pixel sensor where the electron counting algorithm varies from one side of the detector to the next. These differing electron sensors will be housed within a single electron spectrometer and/or imaging filter, and it will be possible to switch back and forth between the different sensors for different experiments. See, e.g., FIG. 5.

Certain embodiments of the present invention facilitate the recording of the entire EELS signal with high fidelity, which is currently believed to be impossible. The EELS signal can vary by many orders of magnitude as a function of energy loss, with low loss data (0 - 50 eV) being very intense and containing information related to phonons and plasmons, and the ultra-high energy loss data (> 3 keV) being very low intensity and containing information related to short range order and coordination chemistry. To record the low loss data, having a detector with a high dynamic range is most important, but to record the ultra- high energy loss data, having a detector with a high SNR at low doses is most important. In both cases, the electron detector must provide high resolution. No single detector has all of these attributes. In certain embodiments, the present invention combines multiple detector technologies, taking advantage of the unique advantages of the different detectors to record the full EELS signal. For instance, a hybrid pixel array detector (e.g., the EMPAD or the Medipix3) may be used to record the high signal of the low loss data, and then a monolithic active pixel (e.g., Gatan K3 camera or the Direct Electron DE-16) sensor may be used to record the high energy loss data.

The impact of certain embodiments of this invention will be significant for the materials science community. By recording the full EELS spectrum in a single electron microscope, it will be possible to locally correlate information related to atomic structure, coordination chemistry, chemical bonding, electronic properties, and vibrational properties. No other instrument can measure all of these signals. In certain embodiments, the present invention enables the simultaneous capture of this information with nanoscale spatial resolution.

In certain embodiments, the invention can be applied to the study of various disordered, multicomponent, and low dimensional material systems. This includes, but is not limited to, the study of high entropy alloys - where the present system offers an unprecedented ability to measure local chemical order/disorder, 2D materials (for example, MXenes), catalytic nanoparticles, bulk metallic glasses, and others.

Traditional crystallography approaches struggle in the determination of local structure and coordination chemistry for disordered, multicomponent materials. Such materials are best studied through extended fine structure analysis of core level excitations, generally measured via synchrotron radiation and extended X-ray absorption fine structure (EXAFS). However, the EXAFS spatial resolution is only ~1 pm, preventing localized measurements of materials with nanoscale heterogeneity. In principle, (scanning) transmission electron microscopy ((S)TEM) and extended energy loss fine structure (EXELFS) provide the same benefits of EXAFS but with an ~ 1000-fold increase in spatial resolution and improved low energy capability. Unfortunately, the promise of EXELFS has not been realized owing to signal-to- noise ratio (SNR) limitations. Here, using direct detection spectroscopy, certain embodiments of the present invention provide localized EXELFS measurements up to a record 12 keV, rivaling the SNR of a synchrotron experiment, but maintaining the spatial resolution of a (S)TEM. The unique advantages offered by EXELFS are showcased using candidate low dimensional (2D) and composite materials.

For many materials, subtle differences in short range order (SRO) - defined here to include bond lengths, coordinating species, and coordination numbers - can have profound effects on functional properties. For example, in correlated oxides, slight changes in bond lengths are associated with resistivity changes of up to 6 orders of magnitude [1]. Rational materials design and device fabrication efforts are thus dependent on the ability to accurately measure SRO. Methods exist for structural determination in perfectly crystalline materials and/or disordered materials that are homogenous on pm length scales. However, real materials always have some degree of atomic disorder, and increasingly, engineering materials possess nanoscale heterogeneity. For such samples, no robust method for SRO determination exists. Here, EXELFS measurements are demonstrated with nanoscale spatial resolution, providing a much-needed local probe of SRO.

Currently, SRO is commonly measured with EXAFS, despite the limited spatial resolution and energy range of this technique [3,4] EXAFS (and EXELFS) analysis focuses on subtle oscillations in the X-ray absorption (or inelastic electron scattering) cross-section of a core-level excitation. The oscillations arise from the interaction of the excited photoelectron and neighboring atoms, allowing SRO measurements in crystalline, amorphous, and even liquid specimens. However, X-ray spot sizes for EXAFS are on the pm scale, preventing localized measurements. In contrast, the electron probe of a STEM can be focused to sub-A dimensions, with clear benefits for localized measurements of interfaces, secondary phases, and individual nanoparticles [5]. EXELFS offers an additional benefit in energy range. EXAFS beamlines have a low energy cutoff of ~ 2 keV, preventing, for instance, the measurement of low Z elements. Conversely, EELS can measure down to less than 1 eV, enabling the correlation of SRO measurements with lower energy core-loss edges and the low loss EELS signal. See FIG. 1A. [6]. While the benefits of EXELFS were recognized in the 1970s [7,13], this technique is not widely used due to prohibitive SNR limitations. Extended edge analysis is best performed on -edges (or A-edges if Z > 58) which are usually in the range of 2 - 20 keV. EELS measurements, however, are usually performed at < 2 keV. At higher energies, the cross section for inelastic scattering decreases, leading to inherently low SNR. As a result, prior EXELFS measurements have been restricted to relatively light elements with low energy edges. Recently, through optimization of electron optics, MacLaren et al. demonstrated EELS up to a record 10 keV, though the SNR was far too low for EXELFS analysis ( See Fig. 1A ) [15]. To enable localized SRO measurements, EXELFS must be extended to higher energies.

Thus, certain embodiments of the present invention are directed to nanoscale EXELFS measurements up to 12 keV with unprecedented SNR through use of advanced electron detection technology. Specifically, direction detection (DD) and electron counting provide significant increases in SNR compared to conventional electron detectors. By coupling DD with an improved electron optical set up, high spatial resolution EXELFS measurements are possible with a SNR comparable to synchrotron EXAFS measurements. The unique advantages of (S)TEM-EXELFS are showcased in two key examples. First, demonstrating the ability to locally probe SRO, structural variation in an amorphous/crystalline nanolaminate is measured and preferential bonding is detected within the amorphous phase. Second, taking advantage of the increased (low energy) spectral range of EELS, a comprehensive understanding of surface chemistry in a 2D material is provided by correlating EXELFS measurements at 6 keV with multiple core-loss edges below 1 keV.

Additional details and advantages of the disclosure will be set forth in part in the description which follows, and/or may be learned by practice of the disclosure. The details and advantages of the disclosure may be realized and attained by means of the elements and combinations particularly pointed out in the appended claims. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 A shows a plot of spatial resolution versus energy range for conventional EELS, EXAFS, and the examples herein.

FIG. IB shows a Ni A-edge EELS and XAS comparison.

FIG. 1C shows a/(A) comparison of EXELFS and EXAFS.

FIG. ID shows a fit to the Ni A-edge EXELFS data shown in FIG. IB. FIG. 2A shows a schematic of the structure of the laminate and the improved spatial resolution of STEM-EXELFS compared to EXAFS.

FIG. 2B shows STEM annular dark field (ADF) image showing the laminate. The magnified region shows a Ni A-edge intensity map, taken from a STEM-EXELFS spectrum image (SI).

FIG. 2C shows a comparison of the Ni AT-edge EXELFS \c{K)\ data for crystalline Ni and the BMG. The EXAFS simulation is the average of the EXELFS Ni and BMG measurements.

FIG. 2D shows a comparison of the EXELFS \/_{R)\ at different locations within the BMG layer, see FIG. 2B.

FIG. 3A shows a schematic showing the CxiTiCiT_x structure in cross-section (left), and a magnified view of the local Cr environment (right).

FIG. 3B shows a comparison of the EELS/EXELFS energy range with the EXAFS energy range.

FIG. 3C shows the evolution in the Cr AT-edge \/_{R)\ with annealing. Two areas of the sample were measured, differentiated by the solid and dashed lines.

FIG. 3D shows an example fit to an initial state dataset from FIG. 3C.

FIG. 3E shows the evolution of the F AT-edge intensity with annealing.

FIG. 3F shows the evolution of the Cr A3-edge fine structure with annealing.

FIG. 4. Shows a fit to the Ni-Zr data.

FIG. 5 shows a schematic of a hybrid TEM/EELS system, according to one embodiment.

FIG. 6 shows a schematic diagram of a gradient detector, according to one embodiment.

DETAILED DESCRIPTION

First, the ability to obtain synchrotron quality data at high energy is demonstrated using a focused electron probe. Using the Ni AT-edge (8.3 keV) of pure Ni as an example, the normalized X-ray adsorption spectroscopy (XAS) signal is compared with the EELS signal after background subtraction and the removal of plural scattering (Fig. IB). Excellent agreement is observed between the two techniques. To compare the extended edge signals, the general procedure of setting the edge onset to 0 eV and then converting from energy to wavenumber, k, of the excited electron with units of A ¹ was followed. The momentum dependent extended edge fine structure, x(k), is shown in Fig. 1C, and again, excellent agreement is observed between EXAFS and EXELFS. The Fourier transform of c(K), referred to as x(R), is closely related to the radial distribution function of the excited atom. In Fig. ID, the EXELFS /(R) is shown along with a fit using ab initio multiple scattering calculations; good agreement is observed out to the 4^th nearest neighbor bond distance. The measurement of the Ni A-edge is the highest energy EXELFS measurement reported to date, and moreover, the highest reported EXELFS SNR was obtained, with /(k) oscillations out to k ~ 14 A ¹. Supplemental EXELFS data for the Ti A-edge at 5 keV, the Ga AT-edge at 10.3 keV, and the Au A-edge at 11.9 keV is described below.

There are several advantages offered by these high quality EXELFS at record energies. With respect to EXAFS measurements - which are inherently ensemble-averaging over pm length scales - EXELFS provides an ~1000-fold improvement in spatial resolution. In comparison to other localized (S)TEM techniques such as imaging, diffraction, tomography, and elemental mapping, EXELFS is uniquely able to quantify SRO in disordered, multi- component materials. To illustrate this ability, local structure in a laminate of crystalline Ni and the bulk metallic glass (BMG) ZresCun.sNiioAE.s (See Figs. 2 A - 2D) [21] were studied. Such laminates offer high strength with improved ductility compared to single-phase BMGs [22] However, crystallization within the BMG during processing can hinder mechanical properties [21], and as such, it is important to quantify SRO within the BMG after the laminate is formed. With EXAFS, Ni AT-edge \/_(R)\ measurements would yield ensemble averaged data, mixing together the crystalline Ni and BMG structures [23]. Alternatively, with STEM-EXELFS, it is possible to selectively probe both phases (See Figs. 2A - 2C). In contrast to the crystalline Ni, the BMG Ni AT-edge \/_(R)\ shows no structure beyond ~ 3 A, indicative of an amorphous material. No significant differences in the BMG structure are observed at interfaces versus the phase interior (See Fig. 2D). Interestingly, while the BMG is primarily composed of Zr, EXELFS fitting suggests that Ni is mostly bonded with Ni and Cu (Supplementary Information). Such preferential bonding can strongly influence the BMG’s mechanical properties [24] This result could not have been directly obtained with EXAFS or conventional STEM techniques. These nanoscale STEM-EXELFS measurements have implications for a number of emerging material systems which possess some degree of disorder, e.g. high entropy alloys (which possess chemical disorder) [25], amorphous magnets [26], and amorphous transparent oxides [27].

Next, the improved spectral range of EXELFS is considered. While EXAFS beamlines have a low energy cut off of ~ 2 keV, EELS can measure down to the meV scale, enabling simultaneous measurement of both high Z and low Z elements. To demonstrate this advantage, surface chemistry evolution in CnTiCaT* (T = surface termination) was studied by measuring from ~ 1 eV out to 6 keV in a single experiment ( See Fig. 3 A - 3B). This material belongs to the family of 2D Mxenes, which show excellent performance in fields such as energy storage and catalysis [28]. Mxene surfaces are terminated with a mix of -OH, -F, and -O species. Variations ins these terminations can strongly alter the Mxene properties [16], and as such, it is critical to quantify and control the surface chemistry. In situ annealing of CnTiCiT_x was performed to thermally induce surface termination loss, which was detected via Cr K-e dge EXELFS (the surfaces of this Mxene are entirely composed of Cr) (Fig. 3A). In the \/_(R)\ data, the first peak at R ~ 1.4 A is related to Cr-C and Cr-T bonding, and the decrease in this peak with annealing indicates termination loss (Fig. 3C). EXELFS analysis indicates a 45% decrease in the termination concentration after 600 °C annealing (See Fig. 3D and Supplementary Information). However, due to the similar weight of F and O, the modeling cannot differentiate between -OH, -F or -O loss. At lower energies, measurement of the O K- and F AT-edges clearly reveals that the measured decrease in termination concentration is due to loss of-F (See Fig. 3E and Supplementary Information). Additionally, measurement of the Cr Z.3-edge at 577 eV - which is a better probe of the Cr Fermi level orbitals than the AT-edge - shows changes in the near edge fine structure, suggesting a change in chemical bonding (Fig. 3F). A complete understanding of the Mxene surface chemistry evolution was only gained by correlating the Cr AT-edge at 6 keV with multiple edges below 1 keV which are inaccessible to EXAFS beamlines. Beyond this example, the improved (low) energy range of EXELFS holds wider potential. For instance, in bimetallic plasmonic nanoparticles, local structure could be measured via EXELFS and then correlated with the material permittivity as measured with low loss EELS [29].

In conclusion, direct detection EELS and an improved TEM/spectrometer coupling were employed to perform nanoscale EXELFS measurements with a record SNR out to an unprecedented 12 keV. This approach allows highly localized measurements of SRO, which were not previously possible with EXAFS or existing (S)TEM-EELS techniques. This ability is applicable to the characterization of disordered, nanostructured materials, which include many important material systems ranging from catalytic nanoparticles to metal organic frameworks and materials for spintronics.

Methods:

All electron microscopy and spectroscopy were performed on a JEOL 21 OOF instrument with an accelerating voltage of 200 kV and a Schottky emitter. EELS was collected with a Gatan Imaging Filter Quantum equipped with a Gatan K2 IS camera operated in electron counting mode [2] A custom post-specimen lens configuration was used for all EXELFS measurements, with a collection semi-angle that varied from 75 mR for the Ti A-edge at 5 keV to 58 mR for the Au A-edge at 11.9 keV (Supplementary Information). Plural scattering was removed with a Fourier ratio deconvolution. The dispersion was nominally set to 0.5 eV/channel, but, after EXELFS calibration on the Ti, Ni and Au reference samples, the dispersion for all edges was reset to 0.485 eV/channel. The absolute energy calibration for EELS at high energy is poor, and all edges were aligned based on known edge onsets from EXAFS data. For all EXELFS measurements, multiple acquisitions of ~ 30 s were performed, and individual spectra were then aligned and summed. Prior to summation, hot pixels were removed through use of a modified median filter (Supplementary Information). For the Mxene experiment, the microscope was operated in TEM mode to spread the electron dose, and for all other measurements, the microscope was operated in STEM mode with a convergence semi angle of 16 mR. For the Ni A-edge measurement shown in Fig. 2A - 2D, the beam current was 6 nA and the total acquisition time was 1200 s, giving a total dose of 4.5 x 10¹³ electrons. The sample had a thickness of t/l = 1.03, where t is the sample thickness and l is the inelastic mean free path. For the spatially resolved mapping of the crystalline / amorphous laminate, the STEM-EXELFS SI had a step size of 2 nm and total dwell time per pixel of 0.3 s, given a multi-frame SI acquisition with 30 passes and 0.01 s per pass. Seven Sis were acquired and summed. The STEM probe current was 8 nA, and the average sample thickness across the SI area was t/k = 0 55 For the Mxene Cr AT-edge measurements, the TEM beam current was 7 nA, and each EXELFS measurement had a total acquisition time of 600 s, giving a total dose of 2.6 x 10¹³ electrons for each measurement. For the 8 areas sampled, the thicknesses ranged from t/l = 0.7 up to 1.1. Annealing was performed with the Gatan 626 hot stage. All EXELFS fitting was performed with the DEMETER software package [20].

Information relating to EXAFS data acquisition, EXELFS fitting, and sample preparation can be found in the Supplementary Information.

Supplementary Information:

EXELFS Fitting of data in manuscript:

The Ni AT-edge data of pure Ni (Fig. 2B - 2C) was fit using the Ni FCC crystal structure.

For the BMG fitting, we focused on the nearest neighbor bonding and only considered single scattering paths. The fit was performed on a dataset which merged the BMG 1, 2, 3, and 4 data shown in the manuscript Fig. 2A - 2D. We began with a model where the excited Ni atom is bonded to neighboring Ni, Cu, Zr, and A1 atoms, with the starting bond distances determined by the sum of the metallic radii of Ni and the neighboring species. With this model, we sought to fit So² and a bond distance for each scattering path, with So² effectively giving the likelihood of Ni being bonded to the given neighboring species. Owing to the similar Z of Ni and Cu, the EXELFS signal arising from the Ni-Ni and Ni-Cu scattering paths is similar. In an effort to simplify the fit, we thus removed Cu from the model. The fitted So² for the Ni-Ni path is then assumed to correlate with the likelihood Ni-Ni or Ni-Cu bonding. Initial fits consistently showed the So² for the Al-Ni path to be ~ 0, and this path was therefore removed from the dataset. Our final fit then included 4 parameters: So² for Ni-Ni, So² for Ni-Zr, R for Ni-Ni, and R for Ni-Zr. We assumed that the in the amorphous structure s² was 0.01 A², and we did not allow for an energy offset.

The large values of So² obtained here are unphysical, which is the result of setting the bond multiplicity of Ni-Ni and Ni-Zr to 1. In this way, the fitted So² is correlated with the number of Ni-Ni (or Ni-Zr) bonds. The main result from this analysis is that the fitted So² for Ni-Ni is greater than that of Ni-Zr, indicating that Ni is more likely to be bonded with Ni and/or Cu than with Zr, despite the atomic concentration of Zr being higher than the combined concentration of Ni and Cu. There are large errors in the fitted values So²; however, regardless of the set value of so or the set value of AEo (which were not fit but were manually varied) the So² for Ni-Ni was always greater than that of Ni-Zr. We also manually adjusted this to the starting Ni-Ni and Ni-Zr R values, and while the final fitted structure was sensitive to the initial conditions, the Ni-Ni So² was always greater than that of Ni-Zr.

For the Cr2TiC2T_* MXene EXELF S fitting, all termination species were treated as O atoms (given the similar Z of O and F), and the surfaces were assumed to be fully terminated. With annealing, it was assumed that the only change in structure is the loss of surface terminations (meaning a decrease in the Cr coordination), which was captured by allowing So² for the Cr-T scattering path to vary between each dataset. This was achieved by defining an So² reduction factor (RF) for each Cr AT-edge measurement. All of the other fitted parameters were kept constant across all datasets: So², R_J, AR, s , and <t , where R is the Cr-T bond distance, AR describes the percent change in the Cr-C, Cr-Cr, and Cr-Ti bond lengths from their nominal values, s² t is the bond length mean square deviation for Cr-T bonding, and <T²L is the ‘lattice’ bond length mean square deviation assigned to Cr-C, Cr-Cr, and Cr-Ti bonds. Additionally, energy offsets ( \Eo) were fit to each dataset.

Additional EXELFS data and fitting:

For the Ti AT-edge measurement shown in Figs. 2A - 2D, the beam current was 1.9 nA and the total acquisition time was 200 s, giving a total dose of 2.4 x 10¹² electrons. The sample had a thickness of t/l = 1.1, where t is the sample thickness and l is the inelastic mean free path.

Apparatus:

FIG. 5 is a schematic diagram of a hybrid TEM/EELS system 500 that can be configured for either TEM imaging or EELS detection according to one embodiment. As shown in FIG. 5, the hybrid system 500 includes an electron gun 5A, a magnetic prism 5E, a slit filter 5G, one or more electrostatic lenses 5H, and three different planar detectors: a high- intensity PAD detector 51, a low-intensity MAPS detector 5J, and a gradient detector 5K. As indicated in FIG. 5, the slit filter 5G can be selectively moved into or out of the path of the electron beam 5F, and one of the three planar detectors 5I-5K can be selectively moved into the electron beam path.

In operation, the electron gun 5A directs an electron beam 5B at a sample 5C. The resulting electron beam 5D is bent and dispersed by the magnetic fields generated by magnetic prism 5E of a conventional EELS detector with electrons of different energy levels being deflected at different angles by the magnetic fields. The resulting dispersed electron beam 5F is directed towards a selected one of the planar detectors 5I-5K via the slit filter 5G (if selected) and the electrostatic lens(es) 5H.

For energy-filtered TEM imaging, the slit filter 5G and either the PAD detector 51 or the MAPS detector 5 J are moved into the electron beam path, and the electrostatic lens(es) are configured for energy-filtered TEM imaging. Those skilled in the art will understand that the slit filter 5G is designed and positioned to selectively filter out electrons of specific energies, allowing improved imaging and chemical analysis.

For EELS detection, the slit filter 5G is moved out of the electron beam path, any selected one of the three planar detectors 5I-5K is moved into the electron beam path, and the electrostatic lens(es) are configured for EELS detection. Note that the PAD detector 51 is mode appropriate for high-intensity EELS detection, the MAPS detector 5J is appropriate for low- intensity EELS detection, and, as described further below, the gradient detector 5K is appropriate for both high- and low-intensity EELS detection.

FIG. 6 is a schematic diagram of a gradient detector 600 that can be used for the gradient detector 5K of FIG. 5 according to certain embodiments. In these particular embodiments, the gradient detector 600 is a pixel array having at least four different pixelated regions 6A-6D, where each region has one or more rows of identical pixels that are designed for a specific range of electron intensities and where region 6A is designed for the highest intensity range, region 6B for the second highest intensity range, region 6C for the second lowest intensity range, and region 6D for the lowest intensity range. Depending on the particular implementation, there may be one or more regions (not explicitly shown) between regions 6B and 6C for intermediate intensity ranges that decrease in intensity from top to bottom. In other embodiments, a gradient detector used the gradient detector 5K of FIG. 5 may have only two or three regions for different ranges of electron intensities. Depending on the implementation, each pair of adjacent regions may have overlapping or non-overlapping intensity ranges.

Those skilled in the art will understand that the pixels in the high-intensity regions 6A and 6B may be based on the same or similar technologies used in conventional PAD detectors, such as the PAD detector 51 of FIG. 5, while the pixels in the low-intensity regions 6C and 6D may be based on the same or similar technologies used in conventional MAPS detectors, such as the MAPS detector 5J of FIG. 5.

In some implementations, the different regions 6A-6D are integrated on a single substrate, such as a semiconductor substrate. In other implementations, the gradient detector 600 is formed from two more different integrated substrates, each having one or more regions.

As used in the claims, the PAD detector 51 and the MAPS detector 5J are referred to as “monolithic detectors” because they have the same type of pixel throughout their single pixel array corresponding to a single intensity range, as opposed to the gradient detector 5K, which has two or more different types of pixels throughout its pixel array that vary in intensity range in a gradient manner (i.e., from relatively high intensity to relatively low intensity). Those skilled in the art will understand that all three detectors 5I-5K are types of direct detectors that directly detect electrons in the electron beam as opposed to indirect detectors that detect photons generated by directing the electron beam onto, for example, a phosphorescent screen.

Referring again to FIG. 5, in a typical EELS detection mode, most of the electrons in the dispersed electron beam 5F are the higher-energy electrons, which are deflected the least by the magnetic prism 5E, with the lower-energy electrons being fewer while being deflected more. As such, when the gradient detector 600 is used for the gradient detector 5K, the gradient detector 600 can be positioned and oriented such that the highest-energy, highest-intensity electrons are directed to the highest-intensity region 6A, the second-highest-energy, second- highest- intensity electrons are directed to the second-highest-intensity region 6B, and so on with the lowest-energy, lowest-intensity electrons being directed to the lowest-intensity region 6D. In this way, high-accuracy, low-noise EELS detection can be performed when the electron gun 5A is configured to generate a high-intensity electron beam 5B without risk of damaging the planar detector.

Those skilled in the art will understand that electron counting can be performed with different suitable voltage threshold levels being applied to the signals generated by the different regions 6A-6D as appropriate for the electron intensity levels and different operating characteristics of those different regions. Such hybrid electron counting can also be applied across a monolithic detector to tune different areas for detecting different electron energies and/or different electron intensities.

As understood by those skilled in the art, the EELS collection angle is an experimental parameter that is determined by the post-specimen lens of the TEM (not shown in FIG. 5), after the beam passes through the sample but before the beam enters the spectrometer. The collection angle is also determined by the size of the spectrometer entrance aperture. In an EELS experiment, the user will manually adjust the TEM lens and/or aperture size to select the right collection angle. For this technology, it is recommended to use a very large collection angle ( > 50 mRad) in order to collect the high-energy (> 3 keV) EELS data. For the low-loss EELS data (< 50 eV or so), it may be beneficial to use a relatively small collection angle. For the gradient detector, the collection angle is optimized for simultaneous low-energy and high- energy EELS detection (corresponding to high-intensity and low-intensity regions of the EELS spectrum, respectively), through the spectrometer geometry design.

Although the system of FIG. 5 is the hybrid TEM/EELS system 500 having the three detectors 5I-5K, in other embodiments, a hybrid TEM/EELS system may have additional detectors, including (without limitation) one or more other types of PAD detectors, one or more other types of MAPS detectors, and/or one or more other types of gradient detectors. In other embodiments, a hybrid TEM/EELS system may have only two of the three detectors 5I-5K, possibly with one or more other types of detectors. In still other embodiments, the system is an EELS-only system with no slit filter 5G and two or more different types of detectors, including any one, any two, or all three of the three detectors 5I-5K. In still other embodiments, the system is either a hybrid TEM/EELS system or an EELS-only system with only one or more gradient detectors.

Although the system 500 of FIG. 5 has a magnetic prism 5E, which is a type of dispersing magnet that both bends the electron beam (e.g., by about 90 degrees in this particular implementation) and disperses the electron beam, in alternative embodiments, the magnetic prism 5E can be replaced by a different suitable type of dispersing magnet that disperses, but does not bend the electron beam. Furthermore, in some implementations, the electrostatic lenses may be optional. In addition, those skilled in the art will understand that the system 500 may include other suitable elements and components not shown in FIG. 5.

Certain embodiments are a system for electron energy loss spectroscopy EELS. The system comprises an electron gun capable of directing an electron beam towards a sample, a dispersing magnet capable of dispersing the electron beam after passing through the sample, and one or more planar detectors capable of detecting the dispersed electron beam. If there is only one planar detector in the system, then the planar detector is a gradient detector having a pixel array comprising two or more regions designed to detect electrons in two or more different ranges of intensities, such that the dispersing magnet directs (i) relatively high- intensity portions of the electron beam towards one or more of the regions designed to detect one or more relatively high-intensity ranges and (ii) relatively low-intensity portions of the electron beam towards one or more of the regions designed to detect one or more relatively low-intensity ranges. If there are two or more planar detectors in the system, then the two or more planar detectors are selectable detectors comprising (i) zero, one, or more monolithic detectors, each designed to detect electrons in a single intensity range, and (ii) zero, one, or more gradient detectors.

According to at least some of the previous embodiments, the system further comprises a selectable slit filter such that the system supports energy- filtered transmission electron microscopy TEM when the slit filter is selected to be in the electron beam’s path.

According to at least some of the previous embodiments, the one or more planar detectors comprise at least one gradient detector.

According to at least some of the previous embodiments, the one or more planar detectors further comprise at least one monolithic detector.

According to at least some of the previous embodiments, the one or more planar detectors comprise a PAD detector and a MAPS detector.

According to at least some of the previous embodiments, each planar detector is a direct detector. According to at least some of the previous embodiments, the system does not include any indirect detectors.

According to at least some of the previous embodiments, the system is configured to measure a low- loss spectrum of an EELS spectrum which is in the range of from 0 to 100 eV, or from 0 to 75 eV, or from 0 to 50 eV in energy loss.

According to at least some of the previous embodiments, the system is configured to measure a high-loss spectrum of an EELS spectrum which is in the range of greater than 1 keV, or greater than 2 keV, or greater than 3 keV in energy loss.

Certain embodiments are methods of performing EELS using the system according to any of the previous embodiments.

According to at least some of the previous embodiments, the system measures one or more extended energy loss fine structure EXELFS measurements with nanoscale spatial resolution.

According to at least some of the previous embodiments, the one or more EXELFS measurements are selected from locally correlated information related to atomic structure, coordination chemistry, chemical bonding, electronic properties, and vibrational properties.

According to at least some of the previous embodiments, the system is used to perform TEM.

Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the embodiments disclosed herein. As used throughout the specification and claims, “a” and/or “an” may refer to one or more than one. It is intended that the specification and examples be considered as exemplary only, with the scope of the disclosure being indicated by the following claims.

The foregoing embodiments are susceptible to considerable variation in practice. Accordingly, the embodiments are not intended to be limited to the specific exemplifications set forth hereinabove. Rather, the foregoing embodiments are within the scope of the appended claims, including the equivalents thereof available as a matter of law. Other suitable modifications and adaptations of the variety of conditions and parameters normally encountered in the field, and which are obvious to those skilled in the art, are within the scope of the disclosure.

All patents and publications cited herein are fully incorporated by reference herein in their entirety or at least for the portion of their description for which they are specifically cited or relied upon in the present description. The patentees do not intend to dedicate any disclosed embodiments to the public, and to the extent any disclosed modifications or alterations may not literally fall within the scope of the claims, they are considered to be part hereof under the doctrine of equivalents.

References

The following references may be useful in understanding some of the principles discussed herein:

[1] Y. Zhou and S. Ramanathan, Crit. Rev. Solid State Mater. Sci. 38, 286 (2013).

[2] J. L. Hart, A. C. Lang, A. C. Leff, P. Longo, C. Trevor, R. D. Twesten, and M. L. Taheri, Sci. Rep. 7, 8243 (2017).

[3] P. Lee, P. H. Citrin, P. Eisenberger, and B. M. Kincaid, Rev. Mod. Phys. 53, 769 (1981).

[4] A. I. Frenkel, Chem. Soc. Rev. 41, 8163 (2012).

[5] Y. Li, D. Zakharov, S. Zhao, R. Tappero, U. Jung, A. Elsen, P. Baumann, R. G. Nuzzo, E. A. Stach, and A. I. Frenkel, Nat. Commun. 6, 7583 (2015).

[6] R. F. Egerton, Electron Energy-Loss Spectroscopy in the Electron Microscope (Springer US, Boston, MA, 2011).

[7] P. E. Batson and A. J. Craven, Phys. Rev. Lett. 42, 893 (1979).

[8] Y. Tabira, Mater. Sci. Eng. B 41, 63 (1996).

[9] I. MacLaren, R. B. Cummings, F. Gordon, E. Frutos-Myro, S. McFadzean, A. P. Brown, and A. J. Craven, in Adv. Imaging Electron Phys. (2019), p. 345.

[10] F. M. Alamgir, H. Jain, D. B. Williams, and R. B. Schwarz, Micron 34, 433 (2003).

[11] Y. Ito, F. M. Alamgir, H. Jain, D. B. Williams, and R. B. Schwarz, in Bulk Met. Glas. (Boston, MA, 1999).

[12] G. Blanche and G. Hug, Ultramicroscopy 50, 141 (1993).

[13] R. D. Leapman and V. E. Cosslett, J. Phys. D. Appl. Phys. 9, L29 (1976).

[14] R. F. Egerton, P. Li, and M. Malac, Micron 35, 399 (2004).

[15] I. Maclaren, K. J. Annand, C. Black, and A. J. Craven, Microscopy 67, 78 (2018).

[16] J. L. Hart, K. Hantanasirisakul, A. C. Lang, B. Anasori, D. Pinto, Y. Pivak, J. T. Van Omme, S. J. May, Y. Gogotsi, and M. L. Taheri, Nat. Commun. 10:522, (2019).

[17] D. H. Howe, J. L. Hart, R. M. Mcdaniel, M. L. Taheri, and A. J. D. Magenau, ACS Macro Lett. 7, 1503 (2018). [18] V. Naut, J. L. Hart, M. Sokol, H. Chiang, M. L. Taheri, and M. W. Barsoum, Under Rev. Angew. Chemie (2019).

[19] A. J. Craven, H. Sawada, S. Mcfadzean, and I. Maclaren, Ultramicroscopy 180, 66 (2017). [20] B. Ravel and M. Newville, J. Synchrotron Radiat. 12, 537 (2005).

[21] S. Shahrezaei, D. C. Hofmann, and S. N. Mathaudhu, J. Mater. 71, 585 (2019).

[22] J. Kim, D. Jang, and J. R. Greer, Adv. Funct. Mater. 21, 4550 (2011).

[23] A. I. Frenkel, A. V Kolobov, I. K. Robinson, J. O. Cross, and C. E. Bouldin, Phys. Rev. Lett. 89, 285503 (2002). [24] T. Rouxel, J. Am. Ceram. Soc. 90, 3019 (2007).

[25] F. X. Zhang, S. Zhao, K. Jin, H. Xue, G. Velisa, H. Bei, R. Huang, J. Y. P. Ko, D. C. Pagan, J. C. Neuefeind, W. J. Weber, and Y. Zhang, Phys. Rev. Lett. 205501, 1 (2017).

[26] R. Streubel, C. Lambert, N. Kent, P. Ercius, A. T. N. Diaye, C. Ophus, S. Salahuddin, and P. Fischer, Adv. Mater. 1800199 (2018).

[27] T. Kamiya and H. Hosono, NPG Asia Mater. 2, 15 (2010).

[28] M. Naguib, M. Kurtoglu, V. Presser, J. Lu, J. Niu, M. Heon, L. Hultman, Y. Gogotsi, and M. W. Barsoum, Adv. Mater. 23, 4248 (2011).

[29] S. Kadkhodazadeh, F. Anggoro, A. Nugroho, C. Langhammer, M. Beleggia, and J. B. Wagner, ACS Photonics 6, 779 (2019).

[30] J. C. Idrobo, J. Rusz, J. Spiegelberg, M. A. Mcguire, C. T. Symons, R. R. Vatsavai, C. Cantoni, and A. R. Lupini, Adv. Struct. Chem. Imaging 2, (2016).

Each of the foregoing references is hereby incorporated by reference herein, in its entirety.

Claims

WHAT IS CLAIMED IS:

1. A system for electron energy loss spectroscopy EELS, the system comprising: an electron gun capable of directing an electron beam towards a sample; a dispersing magnet capable of dispersing the electron beam after passing through the sample; and one or more planar detectors capable of detecting the dispersed electron beam, wherein: if there is only one planar detector in the system, then the planar detector is a gradient detector having a pixel array comprising two or more regions designed to detect electrons in two or more different ranges of intensities, such that the dispersing magnet directs (i) relatively high-intensity portions of the electron beam towards one or more of the regions designed to detect one or more relatively high-intensity ranges and (ii) relatively low- intensity portions of the electron beam towards one or more of the regions designed to detect one or more relatively low- intensity ranges; and if there are two or more planar detectors in the system, then the two or more planar detectors are selectable detectors comprising (i) zero, one, or more monolithic detectors, each designed to detect electrons in a single intensity range, and (ii) zero, one, or more gradient detectors.

2. The system of claim 1, further comprising a selectable slit filter such that the system supports energy-filtered transmission electron microscopy TEM when the slit filter is selected to be in the electron beam’s path.

3. The system of any one of claims 1-2, wherein the one or more planar detectors comprise at least one gradient detector.

4. The system of claim 3, wherein the one or more planar detectors further comprise at least one monolithic detector.

5. The system of any one of claims 1-4, wherein the one or more planar detectors comprise a PAD detector and a MAPS detector.

6. The system of any one of claims 1-5, wherein each planar detector is a direct detector.

7. The system of any one of claims 1-6, wherein the system does not include any indirect detectors.

8. The system of any one of claims 1-7, wherein the system is configured to measure a low- loss spectrum of an EELS spectrum which is in the range of from 0 to 100 eV, or from 0 to 75 eV, or from 0 to 50 eV in energy loss.

9. The system of any one of claims 1-8, wherein the system is configured to measure a high-loss spectrum of an EELS spectrum which is in the range of greater than 1 keV, or greater than 2 keV, or greater than 3 keV in energy loss.

10. A method of performing EELS using the system of any one of claims 1-9.

11. The method of claim 10, wherein the system measures one or more extended energy loss fine structure EXELFS measurements with nanoscale spatial resolution.

12. The method of claims 11 , wherein the one or more EXELFS measurements are selected from locally correlated information related to atomic structure, coordination chemistry, chemical bonding, electronic properties, and vibrational properties.

13. A method of performing TEM using the system of claim 2.