TY - JOUR
AU - Zhu,, Yimei
AB - Abstract We report the performance of the first aberration-corrected scanning transmission electron microscope (STEM) manufactured by Hitachi. We describe its unique features and versatile capabilities in atomic-scale characterization and its applications in materials research. We also discuss contrast variation of the STEM images obtained from different annular dark-field (ADF) detectors of the instrument, and the increased complexity in contrast interpretation and quantification due to the increased convergent angles of the electron probe associated with the aberration corrector. We demonstrate that the intensity of atomic columns in an ADF image depends strongly on a variety of imaging parameters, sample thickness, as well as the nuclear charge and the deviation from their periodic position of the atoms we are probing. Image simulations are often required to correctly interpret the atomic structure of an ADF-STEM image. aberration correction, Hitachi STEM, Z-contrast and ADF-image analysis Introduction The last decade witnessed the rapid development and implementation of aberration correction in electron optics, realizing a more-than-70-year-old dream of aberration-corrected electron microscopy with a spatial resolution below 1 Å [1–9]. With sophisticated aberration correctors, modern electron microscopes now can reveal local structural information unavailable with neutrons and x-rays, such as the local arrangement of atoms, order/disorder, electronic inhomogenity, bonding states, spin configuration, quantum confinement and symmetry breaking [10–17]. Aberration correction through multipole-based correctors, as well as the associated improvement of stability in accelerating voltage, lens supplies and goniometers in electron microscopes now enables medium-voltage (200– 300 kV) microscopes to achieve image resolution below 0.1 nm. Aberration correction not only improves the instrument's spatial resolution but also, equally importantly, allows larger objective lens pole-piece gaps to be employed, thus realizing the potential of the instrument as a nanoscale property-measurement tool. That is, while retaining a high spatial resolution, we can use various sample stages to observe the materials’ behavior under various temperatures, electric and magnetic fields, and atmospheric environments. Such capabilities afford tremendous opportunities to tackle challenging science and technology issues in physics, chemistry, materials science and biology. In this article, we will briefly report the performance of the dedicated scanning transmission electron microscope (STEM), Hitachi HD-2700C, recently installed at the Center for Functional Nanomaterials (CFN), Brookhaven National Laboratory. It is the first aberration-corrected electron microscope manufactured by Hitachi, which has recently attracted considerable attention. In the following, we will first describe the instrument and then present some application data to show the performance of the microscope. Finally, we will discuss issues in STEM contrast interpretation. The microscope and the microscope lab The Hitachi HD-2700C was designed based on Hitachi HD-2300A [18], a dedicated STEM developed more than 10 years ago aiming at element-sensitive high-resolution imaging and spectroscopy with user-friendly operation. The HD-2700C is equipped with a cold-field-emission electron source with high brightness and small energy spread, ideal for atomically resolved STEM imaging and EELS. The microscope has two condenser lenses and an objective lens with a 3.8 mm gap, compared to the 5-mm-gap objective lens in HD-2300A, with the same ±30° sample tilt capability and various holders for heating and cooling (−170 to 1000°C). The projector system consists of two lenses that provide considerable flexibility in choosing various camera lengths and collection angles for imaging and spectroscopy (see Table 1). Figure 1a is an image of the microscope with its cover off. During normal operation, the entire instrument is covered with a telephone-booth-like metal box (Fig. 2a) to reduce acoustic noise and thermal drift. The assignment of the optical lenses and detectors are schematically shown in Fig. 1b. Besides the optical lenses similar to most of the TEM/STEM configuration, there are seven fixed or retractable detectors in the microscope. Above the objective lens is the secondary electron (SE) detector for imaging the sample's surface morphology. Below are the Hitachi analog high-angle annular dark-field (HAADF) and bright-field (BF) detectors for STEM, and a Hitachi TV-rate Live-Diffraction CCD camera (30 frames s−1, 8 bits and 480 × 480 pixels) for fast and low-magnification observations and alignment. The Gatan 2.6 k × 2.6 k 14-bit CCD camera located further down is for diffraction (both convergent and parallel illumination) and Ronchigram analysis. The Gatan analog medium-angle annular dark-field (MAADF) detector and EELS spectrometer (a 16-bit, 100 × 1340 pixel CCD) are sited at the bottom of the instrument. The spectrometer (Enfina ER) is a high-vacuum compatible high-resolution device that Gatan designed particularly for BNL. Since the ADF image contrast, or ‘Z-contrast’, is very sensitive to the ratio of the convergent and collection angles, for quantitative analysis of the STEM images, knowledge of the detector setting is of significant importance. Table 1 summarizes the collection angles for various settings of the HD-2700C (both Hitachi and Gatan detectors), including those for simultaneous acquisition of STEM and EELS. For reference, we also listed the energy resolution and probe current for the four most commonly used modes. Additional settings can be realized by using free-lens control. Fig. 1 Open in new tabDownload slide (a) The Hitachi aberration-corrected HD-2700C STEM at BNL. (b) The schematics of the lens and detector assignment of the instrument. Fig. 1 Open in new tabDownload slide (a) The Hitachi aberration-corrected HD-2700C STEM at BNL. (b) The schematics of the lens and detector assignment of the instrument. Fig. 2 Open in new tabDownload slide The high-precision microscope lab that hosts the HD-2700C. The lab was built with a room-in-room concept to achieve the environmental requirement that is compatible with the performance of the instrument. (a) The instrument room behind the window and (b) the operation room. Fig. 2 Open in new tabDownload slide The high-precision microscope lab that hosts the HD-2700C. The lab was built with a room-in-room concept to achieve the environmental requirement that is compatible with the performance of the instrument. (a) The instrument room behind the window and (b) the operation room. Table 1 Various calibrated settings for collection angles of Hitachi and Gatan ADF detectors in the Hitachi HD-2700C . . HD ADF . Gatan ADF . . Entrance aperture size . collection angle . collection angle . . . . . . Inner . Outer . Inner . Outer . Mode . 5 mm . 3 mm . 2 mm . 1 mm . (mrad) . (mrad) . (mrad) . (mrad) . ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) HAADF 114 608 46 101 MAADF 53 280 21 47 CL1 0.60 205 0.55 146 0.45 83 0.40 26 64 341 26 57 CL2 0.55 180 0.50 105 0.40 48 0.35 13 45 242 18 40 CL3 0.60 53 0.55 29 0.45 14 0.35 5 24 126 10 21 CL4 0.60 25 0.55 13 0.45 7 0.35 3 16 85 6 14 . . HD ADF . Gatan ADF . . Entrance aperture size . collection angle . collection angle . . . . . . Inner . Outer . Inner . Outer . Mode . 5 mm . 3 mm . 2 mm . 1 mm . (mrad) . (mrad) . (mrad) . (mrad) . ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) HAADF 114 608 46 101 MAADF 53 280 21 47 CL1 0.60 205 0.55 146 0.45 83 0.40 26 64 341 26 57 CL2 0.55 180 0.50 105 0.40 48 0.35 13 45 242 18 40 CL3 0.60 53 0.55 29 0.45 14 0.35 5 24 126 10 21 CL4 0.60 25 0.55 13 0.45 7 0.35 3 16 85 6 14 The energy resolution and beam current of the probe measured on the Enfina spectrometer are also shown. The energy resolution was measured at an emission current 4.0 μA and electron probe 1.4 Å with dispersion 0.05 eV/ch, and was determined using the full width of half maximum (FWHM) of the zero-loss peak that was adjusted to the maximum height for each setting. For the 5 mm aperture setting, there is a significant tail of ∼4 eV at the full width of one-tenth maximum (FWTM). The resolution at mode CL2 is slightly better than the other modes because the spectrometer is mostly optimized. The beam current was recorded at an emission current 20 μA and electron probe 1.4 Å. In this measurement, we connected the drift-tube to a picoampere meter and used it as a Faraday cup. Ip is the maximum beam current we could obtain for each setting with no incident electrons being scattered from the sample. As high as 500 pA beam current can be obtained at the electron probe 2.0 Å. Please note that the simultaneous acquisition of STEM-EELS is available only for modes CL1–4 since in HAADF and MAADF, the electron beam cannot be focused well at the project-lens crossover, i.e. onto the back-focal plane of the Enfina spectrometer. Open in new tab Table 1 Various calibrated settings for collection angles of Hitachi and Gatan ADF detectors in the Hitachi HD-2700C . . HD ADF . Gatan ADF . . Entrance aperture size . collection angle . collection angle . . . . . . Inner . Outer . Inner . Outer . Mode . 5 mm . 3 mm . 2 mm . 1 mm . (mrad) . (mrad) . (mrad) . (mrad) . ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) HAADF 114 608 46 101 MAADF 53 280 21 47 CL1 0.60 205 0.55 146 0.45 83 0.40 26 64 341 26 57 CL2 0.55 180 0.50 105 0.40 48 0.35 13 45 242 18 40 CL3 0.60 53 0.55 29 0.45 14 0.35 5 24 126 10 21 CL4 0.60 25 0.55 13 0.45 7 0.35 3 16 85 6 14 . . HD ADF . Gatan ADF . . Entrance aperture size . collection angle . collection angle . . . . . . Inner . Outer . Inner . Outer . Mode . 5 mm . 3 mm . 2 mm . 1 mm . (mrad) . (mrad) . (mrad) . (mrad) . ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) ΔE (eV) Ip (pA) HAADF 114 608 46 101 MAADF 53 280 21 47 CL1 0.60 205 0.55 146 0.45 83 0.40 26 64 341 26 57 CL2 0.55 180 0.50 105 0.40 48 0.35 13 45 242 18 40 CL3 0.60 53 0.55 29 0.45 14 0.35 5 24 126 10 21 CL4 0.60 25 0.55 13 0.45 7 0.35 3 16 85 6 14 The energy resolution and beam current of the probe measured on the Enfina spectrometer are also shown. The energy resolution was measured at an emission current 4.0 μA and electron probe 1.4 Å with dispersion 0.05 eV/ch, and was determined using the full width of half maximum (FWHM) of the zero-loss peak that was adjusted to the maximum height for each setting. For the 5 mm aperture setting, there is a significant tail of ∼4 eV at the full width of one-tenth maximum (FWTM). The resolution at mode CL2 is slightly better than the other modes because the spectrometer is mostly optimized. The beam current was recorded at an emission current 20 μA and electron probe 1.4 Å. In this measurement, we connected the drift-tube to a picoampere meter and used it as a Faraday cup. Ip is the maximum beam current we could obtain for each setting with no incident electrons being scattered from the sample. As high as 500 pA beam current can be obtained at the electron probe 2.0 Å. Please note that the simultaneous acquisition of STEM-EELS is available only for modes CL1–4 since in HAADF and MAADF, the electron beam cannot be focused well at the project-lens crossover, i.e. onto the back-focal plane of the Enfina spectrometer. Open in new tab The CEOS probe corrector, located between the condenser lens and the objective lens, has two hexapoles and five electromagnetic round lenses, seven dipoles for alignment and one quadrupole and one hexapole for astigmatism correction. Other features of the instrument include remote operation, double shielding of the high-tension tank and anti-vibration system for the field emission tank. The instrument was installed in July 2007. Within the first 2 weeks, we had achieved a 0.1 nm resolution of the HAADF-STEM image [19]. The instrument was accepted in November 2007. A residual gas analyzer (RGA-200, Stanford Research Systems, Sunnyvale, CA, USA) to monitor vacuum quality, especially associated with specimen contamination, has been added to the instrument recently. The HD-2700C laboratory sits on a thick slab located in the new building of the CFN. It occupies one of the four high-accuracy electron microscopy labs in the building. The high-accuracy labs were designed with a room-in-room concept. Each laboratory consists of an Instrument Room, an Equipment Room and a Control Room (Fig. 2). Double walls, double windows, refrigerator-type doors and cooling panels are used to reduce mechanical vibration, and temperature and air pressure fluctuation. The design criteria for the high-accuracy laboratories include floor vibrations below 0.25 μm s−1 (rms) in all directions and frequencies, acoustic noise below 40 dB, stray AC magnetic fields <0.1 mG (p-p) at 60 Hz and at lower frequencies scaled by f/60, stray DC field below 1 mG vertical and below 0.01 mG horizontal above earth ambient field, airflow <1 cm min−1 vertically and no horizontal air current permitted. Temperature and humidity were set at 21.1°C (with drift < ±0.1°C h−1) (70 ± 0.18°F h−1) and 40–60%, respectively. Some of the room performance specifications satisfy the NIST-A criterion, i.e. 25 nm (1 μin) between 1 and 20 Hz and 3.1 μm s (125 μin s−1) between 20 and 100 Hz; electromagnetic field below 0.01 mG in all directions; temperature fluctuation below ±0.05°F day−1. For details on the room design and performance, readers are referred to reference [20]. The performance of the instrument Hitachi microscopes have a reputation for stability of high tension, as once again demonstrated by the HD-2700C with 0.5 ppm routinely achievable. Figure 3 illustrates a time trace of the zero-loss peak of the 200 kV electron beam over 130 s; it indicates a stability of <8 × 10−7 for this duration. The energy resolution derived from this experiment is 0.35 eV (FWHM) for a 10 s acquisition. The energy in EELS, determined by the FWHM of the zero-loss peak is ∼0.4 eV at normal operation mode with an emission current of 9.0 μA. For low-emission currents, 0.3–0.35 eV energy resolution can be reached. Fig. 3 Open in new tabDownload slide (a) Voltage plot showing the high-tension stability of 0.5 ppm over a 130 s period. (b) Energy resolution (HWFM) as a function of emission current. Fig. 3 Open in new tabDownload slide (a) Voltage plot showing the high-tension stability of 0.5 ppm over a 130 s period. (b) Energy resolution (HWFM) as a function of emission current. There are many unique and useful features in the HD-2700C. The SE detector can be easily used to observe surface morphology with extremely high spatial resolution, and the results will be published separately. The Gatan CCD camera can be used to record spot pattern in diffraction with parallel illumination, which is extremely useful for structural determination. STEM imaging in the HD-2700C is very user friendly, compared with the TEM/STEM instruments. Aberration correction is done through the CEOS tuning system and software. When a Ronchigram is used for aberration correction, 43 mrad of flat-phase region (half angle) is attainable. Figure 4 is a HAADF-STEM image of Si 110 dumbbells taken at a collection angle of 114–608 mrad with a scan speed of 40 s at 15 Mx magnification. The separation of the neighboring Si atom columns with a distance of 1.36 A is clearly visible with ∼40% contrast dip between nearest neighbors. Fourier analysis of the STEM image in Fig. 4a suggests that higher frequencies (0.96 A) of Si(044) can be transferred. Fig. 4 Open in new tabDownload slide (a) High-resolution HAADF-STEM image of Si 110 dumbbells (raw image with a probe convergence angle of 28 mrad, an ADF collection angle of 114–608 mrad and a scan speed of 40 s at 15 Mx magnification, showing the clear separation of the two neighboring Si atom columns with a distance of 1.36 A. (b) Intensity profile of the line scan in (a) showing ∼40% dumbbell contrast. (c) Power spectrum of the STEM image in (a). Fig. 4 Open in new tabDownload slide (a) High-resolution HAADF-STEM image of Si 110 dumbbells (raw image with a probe convergence angle of 28 mrad, an ADF collection angle of 114–608 mrad and a scan speed of 40 s at 15 Mx magnification, showing the clear separation of the two neighboring Si atom columns with a distance of 1.36 A. (b) Intensity profile of the line scan in (a) showing ∼40% dumbbell contrast. (c) Power spectrum of the STEM image in (a). Figure 5a shows an example of the MAADF-STEM image of BaTiO3 taken with an inner collection angle of 53 mrad. The image shows a beautiful square lattice with little noticeable distortion. Figure 5b is a spectroscopy image from the line scan (A–B) in Fig. 5a with an EELS spectrum of the Ti L3 and L2 edge in BaTiO3, showing a clear separation of the eg and the t2g peaks using the Gatan Enfina spectrometer. Fig. 5 Open in new tabDownload slide (a) High-resolution MAADF-STEM image of BaTiO3 recorded along the 100 direction (raw image with a collection angle of 53–280 mrad at 10 Mx magnification. (b) The EELS spectroscopy image and spectrum of the Ti L3 and L2 edge in BaTiO3 along the line marked A–B in (a), showing a clear separation of the eg and the t2g peaks. Fig. 5 Open in new tabDownload slide (a) High-resolution MAADF-STEM image of BaTiO3 recorded along the 100 direction (raw image with a collection angle of 53–280 mrad at 10 Mx magnification. (b) The EELS spectroscopy image and spectrum of the Ti L3 and L2 edge in BaTiO3 along the line marked A–B in (a), showing a clear separation of the eg and the t2g peaks. Currently, the spatial resolution of an electron microscope is often tested and defined by the periodic structure of a test sample with certain atomic spacing. This is not ideal since the results will strongly depend on the sample including its thickness, surface condition and channeling effect. We select uranium atoms on a thin (<2 nm) carbon film for high Z and easy availability and characteristic core-loss spectrum for atomic EELS and STEM, especially for measuring the probe size from single individual atom images, where the single atom's electrostatic potential can be considered as a delta function. The sample is typical of negative staining employed in biological studies except that the uranyl acetate is 100 times more dilute. Tobacco mosaic virus (TMV) was included to give a thickness gradient with higher concentration of uranium atoms near the TMV, sometimes forming small clumps. The UO2 species observed on such a specimen has a nearest neighbor spacing of 0.34 nm, but the atom size ‘seen’ by the electrons scattered onto the dark-field annular detectors is much smaller and can be considered as a delta function [20]. Figure 6 shows the sequential images excerpted from a movie of uranium single-atom motion. Single atoms move due to irradiation by the electron beam. During this experiment, the specimen was cooled to −160°C with a Gatan 636 liquid nitrogen cooling stage [21]. Uranium atom images show clear bright spots with a high signal-to-noise ratio and quantized intensity. In an electron microscope, the detection of a single atom is accompanied by the passage of high-energy electrons within 0.5 Å of the atom [4,5]. Therefore, there is a significant probability that the atom will gain enough energy to escape from its binding site; hence, a sequence of images will contain information about the movement of individual atoms, limited only by the time resolution of the image acquisition. This behavior is determined by the balance of several bonding energies, including Van der Waals forces, molecular orbital- and bonding-valence electrons’ state, surface energy and electric attraction and repulsion. Roughly two-thirds of atoms move <2 Å on subsequent scans and have symmetrical profiles suitable for probe profile measurement. Occasionally a spot will have a flat side or a gap, indicating that the atom moved between successive scan lines. However, enough atoms remain stationary on sequential scan lines and from frame to frame to permit reliable measurement of the probe diameter. Fig. 6 Open in new tabDownload slide Sequential image acquisition of uranium single-atom motion. Circles in the figure mark the corresponding individual atoms. Specimen was cooled down to −160°C (image size 256 × 256 pixels, 0.04 nm/pixel, 64 ms/pixel.) Fig. 6 Open in new tabDownload slide Sequential image acquisition of uranium single-atom motion. Circles in the figure mark the corresponding individual atoms. Specimen was cooled down to −160°C (image size 256 × 256 pixels, 0.04 nm/pixel, 64 ms/pixel.) The challenge in STEM performance is the simultaneous acquisition of atomically resolved ADF imaging and EELS. Since spectroscopy imaging necessitates a lengthy acquisition period (a 100 × 100 pixel map requires 104 times longer acquisition compared to a point acquisition), the quality of the data largely depends on the stability of the instrument that may be perturbed by various factors, including mechanical vibration, electrical noise, electromagnetic field fluctuations, thermal drift in lens and emission current and specimen charging. For each category, a monitoring protocol has to be established to detect any deterioration in performance, trace it to its source and then correct it; otherwise the instrument will not perform as expected. Figure 7a and b show an HAADF-STEM image (probe convergence angle of 28 mrad, collection angle of 53– 280 mrad) and an EELS spectrum of the U–O and C–K edges from the same uranium sample in Fig. 6 with the beam on a U-atom clump. Uranyl ions have a tendency to form chains or clumps with 3.4 Å spacing as well as vertical stacks that look like single atoms but with higher signal values. In the clump, the atoms appear to stack nearly on top of each other, so the substrate signal level is visible between bright columns and profile measurements of columns give nearly the same FWHM as single atoms. The simultaneous acquisition of the HAADF intensity map and uranium concentration map is shown in Fig. 7d and e. Figure 7f–h show the spectroscopy imaging of single atoms with an HAADF image (Fig. 7f), a line scan of the STEM image intensity (Fig. 7g) that corresponds to the line in Fig. 7f and the U–O4,5 edge intensity (Fig. 7h) of the same location after the background subtraction. Fig. 7 Open in new tabDownload slide (a) HAADF-STEM image of uranium atoms and clumps (probe convergence angle of 28 mrad, collection angle of 53–280 mrad). (b) EELS spectrum showing the U–O edge and the C–K edge. (c) Intensity profile along line ‘c’ in (a) showing 0.32 nm spacing. (d) HAADF intensity map of uranium clump (25 × 25 pixels, 0.4 nm pixel spacing). (e) Simultaneous spectrum imaging of area (d) using a Gatan Enfina spectrometer set for 96–150 eV loss (U–O4,5 edge) with a spectrometer acceptance angle of 30 mrad. (f) HAADF-STEM image that had single-atom spectroscopy imaging measurement. (g) HAADF intensity profile of a single uranium atom in (f) showing an intensity profile (0.03 nm/pixel). (h) line scan of the U–O4,5 edge intensity of the same area in (g) after background subtraction. Fig. 7 Open in new tabDownload slide (a) HAADF-STEM image of uranium atoms and clumps (probe convergence angle of 28 mrad, collection angle of 53–280 mrad). (b) EELS spectrum showing the U–O edge and the C–K edge. (c) Intensity profile along line ‘c’ in (a) showing 0.32 nm spacing. (d) HAADF intensity map of uranium clump (25 × 25 pixels, 0.4 nm pixel spacing). (e) Simultaneous spectrum imaging of area (d) using a Gatan Enfina spectrometer set for 96–150 eV loss (U–O4,5 edge) with a spectrometer acceptance angle of 30 mrad. (f) HAADF-STEM image that had single-atom spectroscopy imaging measurement. (g) HAADF intensity profile of a single uranium atom in (f) showing an intensity profile (0.03 nm/pixel). (h) line scan of the U–O4,5 edge intensity of the same area in (g) after background subtraction. On the application side, using atomically resolved STEM-EELS we have worked on various samples in the past year, ranging from carbon nanotubes to interfaces of multilayers of strongly correlated oxide thin films. As a demonstration, we show here two examples on energy-related materials. The first one is nano-particles that have a core-shell structure (Pd core and Pt shell) used as catalyst in fuel cells. Their activities for the oxygen reduction reaction in fuel cells were correlated to the thickness and the uniformity of the Pt shell, and can be used as a guide for making active, durable and low-cost catalysts. Structure of the nano-catalysts with Pt shell thickness ranging from one to five monolayers was determined by the STEM-EELS measurements. Figure 8a shows a STEM image of the sample with nano-particles on a carbon support. EELS spectra of the C–K edge and Pd–M5,4 as well as Pt–M5,4 edges from the marked particle (∼5 nm in size) and its carbon support in Fig. 8a are shown in Fig. 8b and c. Figure 8d–g are from the simultaneous acquisition of STEM-EELS of the particle. The 2D chemical mapping using the Pd and Pt edges clearly reveals the Pd-core and Pt-shell structure of the particle and its carbon support. The shell thickness was estimated to be four monolayers. Please note that the color in Fig. 8e–g corresponds to the spectrum intensity (Fig. 8e–g). Fig. 8 Open in new tabDownload slide (a) A STEM image of Pd/Pt catalyst with a core-shell (Pd core and Pt shell) structure used in the fuel cell. (b and c) The EELS spectra of the C, Pd and Pt edges from the marked particle and its surrounding in (a). (d–g) Simultaneous acquisition of STEM-EELS of the core-shell particle (∼5 nm in size). (d) The HAADF-STEM image. (e–g) 2D chemical mapping of the core-shell particle using the Pd–M, Pt–M and C–K edges, respectively. Fig. 8 Open in new tabDownload slide (a) A STEM image of Pd/Pt catalyst with a core-shell (Pd core and Pt shell) structure used in the fuel cell. (b and c) The EELS spectra of the C, Pd and Pt edges from the marked particle and its surrounding in (a). (d–g) Simultaneous acquisition of STEM-EELS of the core-shell particle (∼5 nm in size). (d) The HAADF-STEM image. (e–g) 2D chemical mapping of the core-shell particle using the Pd–M, Pt–M and C–K edges, respectively. Another example is the study of the thermoelectric material, (Ca2CoO3)0.62(CoO2), which has a complex misfit-layered structure with alternating Ca2CoO3 layers having the rock-salt structure and CoO2 layers having the CdI structure. The two types of layers are stacked along the c-direction with a structural incommensurate modulation along the b-direction. Since the key to the advancement of the thermoelectric materials is to develop a structure that has high electron conductivity and poor thermal conductivity, in other words, it can scatter electrons and phonons differently, the ability to measure the atomic and chemical disorder at atomic resolution is the key to the understanding of the unique structure–property relationship of the material. Figure 9a shows a high-resolution HAADF-STEM image of (Ca2CoO3)0.62(CoO2) viewed along the [010] direction. Figure 9b is an HAADF-STEM survey image of the area with the green-box region indicating where spectroscopy imaging was carried out. Figure 9c shows the simultaneous acquisition of the HAADF-STEM image and chemical mapping of Ca, O and Co. Experiment conditions include an emission current of 15 μA with a collection angle of 64–341 mrad at Mag 8 Mx, energy dispersion of 0.5 eV/ch, scanning area 70 × 70 pixel (0.32 A(Amstrong)/pixel) at a dwelling time of 0.05 s per pixel. Drift correction was also made using the area (yellow-box region in Fig. 9b) as a reference. The overlay of the Ca, O and Co maps of the area is shown in Fig. 9d. The EELS spectrum of the area, showing the Ca–L3,2 and O–K and Co–L3,2 edges, is presented in Fig. 9e. Fig. 9 Open in new tabDownload slide (a) An HAADF-STEM image of a (Ca2CoO3)0.62(CoO2) thermoelectric sample with a misfit layered structure viewed along the [010] direction. (b) An HAADF-STEM survey image of the area. (c) ADF-STEM image and chemical maps of Ca, O and Co from the same area [green-box in (b)], showing atomically resolved spectroscopic imaging capability of the instrument. Emission current: 15 μA with a collection angle of 64–341 mrad at Mag 8 Mx, 0.5 eV/ch, 70 × 70 pixels (0.32 A(Amstrong)/pixel) at an acquisition speed of 0.05 s/pixel. Drift correction (using the yellow box as a reference) was implemented. (d) The overlay of the Ca, O and Co maps. (e) The EELS spectrum of the area. Fig. 9 Open in new tabDownload slide (a) An HAADF-STEM image of a (Ca2CoO3)0.62(CoO2) thermoelectric sample with a misfit layered structure viewed along the [010] direction. (b) An HAADF-STEM survey image of the area. (c) ADF-STEM image and chemical maps of Ca, O and Co from the same area [green-box in (b)], showing atomically resolved spectroscopic imaging capability of the instrument. Emission current: 15 μA with a collection angle of 64–341 mrad at Mag 8 Mx, 0.5 eV/ch, 70 × 70 pixels (0.32 A(Amstrong)/pixel) at an acquisition speed of 0.05 s/pixel. Drift correction (using the yellow box as a reference) was implemented. (d) The overlay of the Ca, O and Co maps. (e) The EELS spectrum of the area. STEM contrast analyses STEM imaging has gained significant popularity in the recent years due to its straightforward contrast interpretation (Z-contrast) and the development of probe aberration correctors that greatly reduce the probe size and enhance the probe current to make atomically resolved EELS a routine. In contrast, due to the large increase in convergent angles of the electron beam to gain current in aberration-corrected microscopes, care must be taken in selecting the collection angles of the ADF detector for STEM imaging. Otherwise, the ADF imaging does not generate pure amplitude contrast but a mixture of amplitude and phase of the electron wave, thus, making interpretation of the contrast difficult. In some situations, for example, when the collection angle of the ADF detector is not sufficiently large, and/or there exists a significant local static or dynamic displacement, it is possible to encounter contrast reversal where heavy-atom columns show low intensity than the light-atom columns. As we know in quantum mechanical approach, the magnitude of an incident electron being scattered per unit solid angle Ω by a given atom can be described by the differential cross section dσ/dΩ, which is a complex quantity. The product of dσ/dΩ and its complex conjugate is the probability of the electron being scattered. The phase component of the differential cross section is important in conventional high-resolution phase-contrast (HREM) imaging, while in ADF-STEM very often only the amplitude is used to calculate the scattered intensity. For elastic scattering, where fx(q) is the x-ray scattering amplitude at scattering vector q, a0 is the first Bohr radius and γ is a relativistic factor. The atomic number Z represents the nuclear charge and denotes the fact that incident electrons are scattered by the entire electrostatic field or potential in crystal, while x-rays interact with the entire atomic electron clouds. For sufficiently large q, the fx(q) becomes negligible. The differential cross section dσ/dΩ can then be described by Rutherford scattering: which is the earliest model for elastic scattering of charged particles based on the unscreened electrostatic field of a nucleus [22]. When the collection angle of the ADF detector is large enough, the intensity of the STEM image can therefore be considered to be dependent on the composition through the Z2 relation of the cross section within the scattering angle ranges of the detector [23]. This is the simplest description about the ADF intensity of the STEM image based on a single isolated atom. For the crystal oriented in a zone axis, the intensity detected by the HAADF detector is considered to be dominated by incoherent thermal diffuse scattering (TDS) with each atom being an independent incoherent center [24]. Analysis of image formation for thin specimens based on Bloch wave approximation further shows that the HAADF geometry could destroy many of the coherent interference effects; thus, the HAADF image represents a direct incoherent structure image even in the thin area as well [24,25]. The incoherent nature of the HAADF image means that the image intensity is given by a simple convolution between an object function, O(R) and a probe function, P2(R) [26]: Because the HAADF detector filters the Bloch states, the 1s states dominate the contrast and the columnar intensity of the HAADF image is proportional to the mean square atomic number, Z2 [26,27]. Intensity calculation for 11 different elements with a range of Z values for residual object function (ROF) gives a power law Z-dependence (AZn) with a value of n = 1.77 ± 0.05 and n = 1.88 ± 0.03 for the 1s-type Bloch states and all Bloch states, respectively [28]. During the acceptance tests of the HD-2700C at the factory as well as at BNL, we noticed that the ADF image contrast does not follow the simple I∼Z2 or I∼Z1.8 power rule. When we used (Ca2CoO3)0.62(CoO2) (approximant Ca3Co4O9) as a test sample for simultaneous column-by-column STEM-EELS spectroscopy imaging experiment with the Gatan ADF detector (see Table 1), we noticed that the high Co signal in EELS came from the atomic columns that had low ADF intensity (based on crystallography analysis, they are highly packed Co4+ ions in the CoO2 columns), while the low Co signal came from the columns that had high ADF intensity (they are Co2+ ions in CoO columns with 0.62 packing density compared with Co4+, see Fig. 9a). The discrepancy motivated us to conduct a systemic study on STEM image contrast analysis. Since the detailed discussion is beyond the scope of the article, in the following we briefly discuss how ADF-STEM contrast varies with different imaging parameters as well as sample thickness. The intensities of ADF-STEM images were calculated using the Brookhaven computer codes based on the multislice method with frozen phonon approximation, which has been demonstrated to have excellent quantitative agreement with experiments ref. 29. Here, we use SrTiO3 as an example. The structure parameter of SrTiO3 was based on reference [30] with lattice parameter a = 0.3901 nm, and Debye–Waller factor BSr = 0.6214, BTi = 0.4398 and BO = 0.7323. In the multislice calculation, the slice dimensions were chosen to be 2 × 2 unit cells, corresponding to 7.802 × 7.802 A in a 1024 × 1024 pixel setting. These choices yield a maximum scattering angle of 1097 mrad after taking the bandwidth limit into account. Although in STEM, the position of the ADF detector is fixed, the collection angles can be varied via changing the camera length. The inner and outer collection angles of the detector are thus defined by βinner = Rinner/L and βouter = Router/L, where Rinner and Router are the inner and outer radii of the detector and L is the camera length. Figure 10 compares experimental and calculated ADF-STEM images of SrTiO3 in (001) projection with two different thickness (∼11 nm and ∼54 nm) at a collection angle of 64–341 mrad and a convergent angle of 27 mrad. In our calculations, a Gaussian point-spread function with FWHM = 0.09 nm was used. A line scan of the intensity profiles of Sr and Ti/O is also included in Fig. 10 to demonstrate the change in the intensity ratio of Sr and Ti/O columns, and the intensity profiles (Fig. 10e and f) were normalized by the Sr atom columns. It is clear that the relative difference in intensities between Sr and Ti/O columns decreases with the increase in the sample thickness. Our calculations agree very well with our experiments, and the measured difference between the two is similar to what was reported in ref. 29. Fig. 10 Open in new tabDownload slide (a and b) STEM images with a convergent angle α = 27 mrad and collection angle β = 64–341 mrad (a) in thin and (b) thick areas. (c and d) Simulated STEM images with (a) thickness = 11 nm and (d) thickness = 54 nm, under the image condition of α = 27 mrad and β = 64–341 mrad. The simulated images are convoluted with a Gaussian point spread function (FWHM = 0.09 nm). (e and f) Intensity profiles from (a)–(d) with black lines for experiments and red lines for calculations. (g) The crystal model of SrTiO3 where blue dots represent Sr, green dots represent Ti and yellow dots represent O. Fig. 10 Open in new tabDownload slide (a and b) STEM images with a convergent angle α = 27 mrad and collection angle β = 64–341 mrad (a) in thin and (b) thick areas. (c and d) Simulated STEM images with (a) thickness = 11 nm and (d) thickness = 54 nm, under the image condition of α = 27 mrad and β = 64–341 mrad. The simulated images are convoluted with a Gaussian point spread function (FWHM = 0.09 nm). (e and f) Intensity profiles from (a)–(d) with black lines for experiments and red lines for calculations. (g) The crystal model of SrTiO3 where blue dots represent Sr, green dots represent Ti and yellow dots represent O. Figure 11a and b show the calculated intensity of Sr and Ti/O columns as a function of collection angle and thickness, respectively (the inner collection angle changes from 2 to 200 mrad, while the outer collection angle is about five times larger, ranging from 10.5 to 1050 mrad). Note that for inner angles <28 mrad, the incident beam strikes the annular detector giving partial bright-field imaging. The thickness was varied from 0.39 to 62.42 nm with a step of 0.3901 nm which is one unit cell thick along the beam direction. In multislice calculation, we divide one unit cell into two slices: one is SrO slice and the other TiO slice. Qualitatively, for inner angles >10 mrad, the intensity at Sr columns is stronger than that at TiO columns, and for both cases the intensity increases with the increase in the sample thickness but decreases with the increase in the collection angle (Fig. 11c and d). When the collection inner angle is in the range of 5.1– 27 mrad, the annular detector partially collects the transmitted beam. The intensity rises first, and then falls with the increase in the collection angle. The maximum intensity, depending on the thickness, occurs at 12 and 8 mrad for Sr and TiO columns, respectively, at t = 10.14 nm. The intensity versus thickness plots (Fig. 11d) show strong oscillation when the transmitted beam strikes a portion of the annular detector (blue and green lines). The thickness oscillation remains even when the collection inner angle is larger than the convergent angle (black and red lines). The thickness oscillation can be attributed to the strong dynamical effects in the low scattering angles. When the collection angle is large enough, say >36 mrad, the intensity is dominated by the incoherent inelastic scattering and the thickness oscillation disappears (dark green and pink lines). Fig. 11 Open in new tabDownload slide (a and b) Calculated intensity map of Sr (a) and Ti (b) columns with the collection angle increasing from left to right and thickness from top to bottom. The inner and outer collection angles vary from 2 to 200 mrad and 10.5 to 1050 mrad, respectively, and the thickness from 0.39 to 62.42 nm. The images are shown in color for clarity. (c) Intensity profiles versus inner collection angle with thickness = 10.14 nm. (d) Intensity profiles of thickness with collection angles of 20–105, 30–158 and 64–336 mrad. Fig. 11 Open in new tabDownload slide (a and b) Calculated intensity map of Sr (a) and Ti (b) columns with the collection angle increasing from left to right and thickness from top to bottom. The inner and outer collection angles vary from 2 to 200 mrad and 10.5 to 1050 mrad, respectively, and the thickness from 0.39 to 62.42 nm. The images are shown in color for clarity. (c) Intensity profiles versus inner collection angle with thickness = 10.14 nm. (d) Intensity profiles of thickness with collection angles of 20–105, 30–158 and 64–336 mrad. In order to better understand the relationship between the image intensity in this simulation and the atomic number Z of the elements, we calculate the intensity ratio of the Sr and TiO columns, as shown in Fig. 12. Based on the Z2 dependence, I∼Z2, ISr/ITiO should be 382/(222+82) = 2.64 (ISr/ITi = 382/222 = 2.98 if we ignore O for simplicity). Surprisingly, for the entire regime of the collection angles and thicknesses shown in Fig. 12a, i.e. in the most commonly used conditions, we do not obtain the relation I∼Z2. We also note that the intensity ratio increases with the increase in the inner collection angles and after they reach to >40 mrad; however, the change in the intensity ratio is significantly reduced (Fig. 12c). While we see that the intensity ratio decreases with the thickness, strong thickness oscillation is observed when the collection inner angle is ≤30 mrad (black and red lines), as shown in Fig. 12d, similar to the observations in Fig. 11d. The Z2 dependence [26] was derived without considering atoms’ thermal vibration that reduces atomic cross section by a factor of exp(−2Bs2). We noted that the Debye–Waller factor of Sr is larger than that of Ti, which may result in the lower intensity ratio ISr/ITiO. To eliminate the effect of different Debye–Waller factors of Sr, Ti and O in the SrTiO3 crystal, we calculated the intensity ratio map by using the same Debye–Waller factor B = 0.5 for all Sr, Ti and O (Fig. 12b). In comparison with Fig. 12a, which uses the experimentally determined Debye–Waller factors for the elements, the Z2-dependence occurs in a very small region with a large collection angle and small thickness. A power law Z-dependence (Zn) with n ranging from 1.77 to 2 gives a slightly larger valid region, as indicated by the red arrows. In summary, ADF images in STEM indeed show Z-dependence contrast. Nevertheless, only true HAADF images yield strong intensity in high-Z atom columns and weak intensity in low Z. In general, the ADF image intensity varies with detector collection angle, sample thickness and Debye–Waller factor of the atomic species. To correctly interpret the ADF intensity in STEM images, the effect of atomic thermal vibration of the atoms must be taken into account. A power law Z-dependence is only valid for very large collection angles on a very thin specimen. Fig. 12 Open in new tabDownload slide (a) Intensity ratio ISr/ITiO with the collection angle increasing from left to right and thickness from top to bottom (BSr = 0.6214, BTi = 0.4398 and BO = 0.7323). (b) Intensity ratio ISr/ITiO with all Debye–Waller factors being BSr = BTi = BO = 0.5. The area with collection angle ≥64 mrad and thickness ≤4.3 nm is considered to comply with a power law Z-dependence (Zn) with n ranging from 1.77 to 2, as outlined by the contour. (c) Intensity ratio profiles of collection angle with the thickness being 10.1, 30.4 and 50.3 nm. (d) Intensity ratio profiles of thickness with the inner collection angle being 20, 30, 64 and 100 mrad. Fig. 12 Open in new tabDownload slide (a) Intensity ratio ISr/ITiO with the collection angle increasing from left to right and thickness from top to bottom (BSr = 0.6214, BTi = 0.4398 and BO = 0.7323). (b) Intensity ratio ISr/ITiO with all Debye–Waller factors being BSr = BTi = BO = 0.5. The area with collection angle ≥64 mrad and thickness ≤4.3 nm is considered to comply with a power law Z-dependence (Zn) with n ranging from 1.77 to 2, as outlined by the contour. (c) Intensity ratio profiles of collection angle with the thickness being 10.1, 30.4 and 50.3 nm. (d) Intensity ratio profiles of thickness with the inner collection angle being 20, 30, 64 and 100 mrad. Concluding remarks It is our hope that in this short presentation we have demonstrated the outstanding performance of the Hitachi HD-2700C aberration-corrected STEM due to its superior design and configuration, including the stability of the high tension and lens current, the cold-field-emission electron source, the aberration corrector and the energy-loss spectrometer. It is a versatile and user-friendly instrument that can be routinely used for atomically resolved STEM imaging and EELS with a wide range of applications in physics, chemistry, materials science and biology. Although aberration correction provides much needed improvement in spatial resolution in microscopy and ample opportunities for materials research, it does not automatically make interpretation of experimental data easier, even in HAADF-STEM imaging. Care must be taken in interpreting the image intensities and relating them to atomic number Z of the elements, especially when lattice displacement (channeling) and strong thermal effects are expected in the materials. The real challenge in electron-beam-based structural characterization is quantitative electron microscopy, which in our view is the future direction of the field. We wish to acknowledge help and advice of many colleagues: Shigeto Isakozawa, Kuniyasu Nakamura and Mitsuru Konno of Hitachi High Technologies Corporation, Naka, Japan; Ray Twesten of Gatan and V.V. Volkov of BNL. We would also like to thank Jia Wang and Qiang Li for providing samples for our test of the instrument. The work at BNL was supported by the Office of Basic Energy Science, Division of Materials Science, U.S. Department of Energy, under contract no. DE-AC02-98CH10886. References 1 Scherzer O . Uber einige Fehler von Elektronenlinsen (some defects of electron lenses) , Optik , 1936 , vol. 101 (pg. 593 - 603 ) OpenURL Placeholder Text WorldCat 2 Scherzer O . Sparische and chromatische korrektur vonelektronen-linsen , Optik , 1947 , vol. 2 (pg. 114 - 132 ) OpenURL Placeholder Text WorldCat 3 Haider M , Uhlemann S , Schwan E , Rose H , Kabius B , Urban K . Electron microscopy image enhanced , Nature , 1998 , vol. 392 (pg. 768 - 769 ) Google Scholar Crossref Search ADS WorldCat 4 Batson P E , Dellby N , Krivanek O L . Sub-angstrom resolution using aberration corrected electron optics , Nature , 2002 , vol. 418 (pg. 617 - 620 ) Google Scholar Crossref Search ADS PubMed WorldCat 5 Batson P E , Dellby N , Krivanek O L . Sub-angstrom resolution using aberration corrected electron optics , Nature , 2002 , vol. 419 (pg. 94 - 94 ) Google Scholar Crossref Search ADS WorldCat 6 Jia C L , Lentzen M , Urban K . Atomic-resolution imaging of oxygen in perovskite ceramics , Science , 2003 , vol. 299 (pg. 870 - 873 ) Google Scholar Crossref Search ADS PubMed WorldCat 7 Jia C L , Urban K . Atomic-resolution measurement of oxygen concentration in oxide materials , Science , 2004 , vol. 303 (pg. 2001 - 2004 ) Google Scholar Crossref Search ADS PubMed WorldCat 8 Borisevich A Y , Lupini A R , Pennycook S J . Depth sectioning with the aberration-corrected scanning transmission electron microscope , Proc. Natl Acad. Sci. USA , 2006 , vol. 103 (pg. 3044 - 3048 ) Google Scholar Crossref Search ADS WorldCat 9 Sawada H , Tomita T , Naruse M , Honda T , Hambridge P , Hartel P , Haider M , Hetherington C , Doole R , Kirkland A , Hutchison J , Titchmarsh J , Cockayne D . Experimental evaluation of a spherical aberration-corrected TEM and STEM , J. Electron Microsc. , 2005 , vol. 54 (pg. 119 - 121 ) Google Scholar Crossref Search ADS WorldCat 10 Kimoto K , Asaka T , Nagai T , Saito M , Matsui Y , Ishizuka K . Element-selective imaging of atomic columns in a crystal using STEM and EELS , Nature , 2007 , vol. 450 (pg. 702 - 704 ) Google Scholar Crossref Search ADS PubMed WorldCat 11 Klie R F , Buban J P , Varela M , Franceschetti A , Jooss C , Zhu Y , Browning N D , Pantelides S T , Pennycook S J . Enhanced current transport at grain boundaries in high-T-c superconductors , Nature , 2005 , vol. 435 (pg. 475 - 478 ) Google Scholar Crossref Search ADS PubMed WorldCat 12 Wu L J , Zhu Y M , Tafto J . Picometer accuracy in measuring lattice displacements across planar faults by interferometry in coherent electron diffraction , Phys. Rev. Lett. , 2000 , vol. 85 (pg. 5126 - 5129 ) Google Scholar Crossref Search ADS PubMed WorldCat 13 Zuo J M , Kim M , O’Keeffe M , Spence J C H . Direct observation of d-orbital holes and Cu-Cu bonding in Cu2O , Nature , 1999 , vol. 401 (pg. 49 - 52 ) Google Scholar Crossref Search ADS WorldCat 14 Zhu Y , Zheng J C , Wu L , Frenkel A I , Hanson J , Northrup P , Ku W . Nanoscale disorder in CaCu3Ti4O12: a new route to the enhanced dielectric response , Phys. Rev. Lett. , 2007 , vol. 99 pg. 037602 Google Scholar Crossref Search ADS PubMed WorldCat 15 Jooss C , Wu L , Beetz T , Klie R F , Beleggia M , Schofield M A , Schramm S , Hoffmann J , Zhu Y . Polaron melting and ordering as key mechanisms for colossal resistance effects in manganites , Proc. Natl Acad. Sci. USA , 2007 , vol. 104 (pg. 13597 - 13602 ) Google Scholar Crossref Search ADS WorldCat 16 Volkov V V , Zhu Y . Phase imaging and nanoscale currents in phase objects imaged with fast electrons , Phys. Rev. Lett. , 2003 , vol. 91 pg. 043904 Google Scholar Crossref Search ADS PubMed WorldCat 17 Wu L , Klie R F , Zhu Y , Jooss C . Experimental confirmation of Zener-polaron-type charge and orbital ordering in Pr1−xCaxMnO3 , Phys. Rev. , 2007 , vol. B76 pg. 174210 Google Scholar Crossref Search ADS WorldCat 18 Nakamura K , Konno M , Yaguchi T , Kamino T , Terauchi D , Inada H , Tanaka H , Taniguchi Y , Isakozawa S . Development of a Cs-corrected dedicated STEM , 2006 In: Proceedings of the 16th International Microscopy Congress, pp 633–634 (Sapporo, Japan) 19 Inada H , Zhu Y , Wall Y , Volkov V , Nakamura K , Konno M , Jarausch K , Twesten R D . The newly installed aberration corrected and dedicated STEM (Hitachi HD2700C) at Brookhaven National Laboratory , Proc. Microsc. Microanal. , 2008 , vol. 14 Suppl 2 (pg. 1374 - 1375 ) Google Scholar Crossref Search ADS WorldCat 20 Zhu Y , Wall J . Hawkes P W . Aberration-corrected electron microscopes at Brookhaven National Laboratory , Aberration-Corrected Electron Microscopy, a Thematic Volume of Advances in Imaging & Electron Physics , 2008 Elsevier/Academic Press, New York Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 21 Inada H , Wall J , Zhu Y , Volkov V , Nakamura K , Konno M , Jarausch K . Uranium single atom imaging and EELS mapping using aberration corrected scanning transmission electron microscope and LN2 cold stage , Proc. Microsc. Microanal. , 2008 , vol. 14 Suppl 2 (pg. 1386 - 1387 ) Google Scholar Crossref Search ADS WorldCat 22 Egerton R F . , Electron Energy-Loss Spectroscopy in Electron Microscopy , 1996 2nd edn New York Plenum Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 23 Pennycook S J . Z-contrast transmission electron microscopy – direct atomic imaging of materials , Annu. Rev. Mater. Sci. , 1992 , vol. 22 (pg. 171 - 195 ) Google Scholar Crossref Search ADS WorldCat 24 Pennycook S J , Jesson D E . High resolution incoherent imaging of crystals , Phys. Rev. Lett. , 1990 , vol. 64 pg. 938 Google Scholar Crossref Search ADS PubMed WorldCat 25 Jesson D E , Pennycook S J . Incoherent imaging of thin specimens using coherently scattered electrons , Proc. R. Soc. A , 1993 , vol. 441 pg. 261 Google Scholar Crossref Search ADS WorldCat 26 Pennycook S J , Rafferty B , Nellist P D . Z-contrast imaging in an aberration-corrected scanning transmission electron microscope , Microsc. Microanal. , 2000 , vol. 6 (pg. 343 - 352 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat 27 Nellist P D , Pennycook S J . Incoherent imaging using dynamically scattered coherent electrons , Ultramicroscopy , 1999 , vol. 78 (pg. 111 - 124 ) Google Scholar Crossref Search ADS WorldCat 28 Rafferty B , Nellist P D , Pennycook S J . On the origin of transverse incoherence in Z-contrast STEM Journal of electron microscopy , 2001 , vol. 50 (pg. 227 - 233 ) OpenURL Placeholder Text WorldCat 29 LeBeau J M , Findlay S D , Allen L J , Stemmer S . Quantitative atomic resolution scanning transmission electron microscopy , Phys. Rev. Lett. , 2008 , vol. 100 pg. 206101 Google Scholar Crossref Search ADS PubMed WorldCat 30 Abramov Yu A , Tsirelson V G , Zavodnik V E , Ivanov S A , Brown I D . The chemical bond and atomic displacements in SrTiO3 from x-ray diffraction analysis , Acta Crystallogr. , 1995 , vol. B51 pg. 942 Google Scholar Crossref Search ADS WorldCat © The Author 2009. Published by Oxford University Press on behalf of Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org Oxford University Press
TI - Performance and image analysis of the aberration-corrected Hitachi HD-2700C STEM
JF - Journal of Electron Microscopy
DO - 10.1093/jmicro/dfp011
DA - 2009-06-01
UR - https://www.deepdyve.com/lp/oxford-university-press/performance-and-image-analysis-of-the-aberration-corrected-hitachi-hd-00L02HTnjA
SP - 111
EP - 122
VL - 58
IS - 3
DP - DeepDyve
ER -