Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 The performance of nanoelectronic devices critically depends on the distribution of charge carriers inside such structures. High-vacuum scanning spreading resistance microscopy (HV-SSRM) has established as the method of choice for quantitative 2D-carrier mapping in nanoscale devices during the last decade. However, due to the 3D-nature of these nanoscale device architectures, dopant incorporation and dopant diffusion mechanisms can vary for any of the three dimensions, depending on the particular processes used. Therefore, mapping of carriers in three dimensions with high spatial resolution is inevitable to study and understand the distribution of active dopants in confined 3D-volumes and ultimately to support the process development of next generation devices. In this work, we present for the first time an approach to extend the capabilities of SSRM from an inherent 2D-carrier profiling technique towards a quantitative 3D-characterization technique based on the example of a nanowire (NW)-based heterojunction (SiGe–Si) tunneling transistor. In order to implement a 3D-methodology with a 2D-imaging technique, we acquired 2D-carrier concentration maps on successive cross-section planes through the device of interest. This was facilitated by arranging several devices in a staggered array, allowing to produce a series of cross-sections with incremental offset by a single cleave. A dedicated interpolation algorithm especially suited for structures with rotational symmetry like NWs was developed in order to reconstruct a 3D-carrier distribution map. The validity of the method was assessed by proving the absence of variations in carrier distribution in the third dimension, as expected for NWs etched into a blanket stack. Highlights ► A methodology for 3D carrier mapping in nanowire-based transistors is presented. ► We developed a slice-and-view approach whereby 2D maps are combined to a 3D map. ► A staggered design was used to generate cross-sections at different depth values. ► Dedicated interpolation algorithm for the reconstruction of a 3D map was developed.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 A simple method is described for the accurate and precise measurement of chromatic aberration under electron-optical conditions pertinent to scanning transmission electron microscopy (STEM) and scanning confocal electron microscopy (SCEM). The method requires only the measurement of distances in a coherent CBED pattern and knowledge of the electron wavelength and the lattice spacing of a calibration specimen. The chromatic aberration of a spherical-aberration corrected 300 kV thermal field emission TEM is measured in STEM and SCEM operating modes and under different condenser lens settings. The effect of the measured chromatic aberrations on the 3 dimensional intensity distribution of the electron probe is also considered. Highlights ► A method is presented to measure chromatic aberration ( C c ) using coherent CBED. ► The C c of the probe and imaging lens systems in STEM and SCEM modes is measured in a C 3 -corrected S/TEM. ► The effect of the measured C c on the depth resolution in STEM is simulated for different energy spreads.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 A through-focus series of annular bright field (ABF) images were observed simultaneously with high-angle annular dark field (HAADF) images of very thin lithium manganese oxide (LiMn 2 O 4 ), a typical cathode material used in lithium ion batteries, using a spherical aberration corrected electron microscope with a 50 pm resolution (R005). The ABF images showed dark dips at the positions of Li and Mn–O atomic columns, which reversed to bright peaks when the defocus sign was changed, as commonly observed in phase contrast images. The optimal defocus for the ABF images was about 2 nm of over-focus, while that for the HAADF images was 2 nm of under-focus for an incident probe with a convergent semi-angle of 30 mrad. These experimental results are interpreted based on a weak-phase-object approximation. Highlights ►Focus series of annular bright field images were observed for very thin LiMn2O4. ► Dark dips at the atomic column positions reversed to bright peaks by defocus change. ► The optimum defocus was 2nm over-focus with a convergent semi-angle of 30mrad. ► The experimental results were interpreted based on weak-phase-object approximation.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 A method is presented to use an electron microscope in transmission mode to determine the mis-tilt from a zone axis of a crystalline material. The method involves recording a number of additional diffraction patterns with incident beams tilted over 2 to 3 degrees. It is shown that an accuracy of 0.02 degree can be achieved, which is far better than that of the specimen-stage tilt axes, which is about 0.1 degree for the β-tilt. Highlights ► Beam tilt can be used to determine accurate mis-orientation from the zone-axis. ► The accuracy of the determined tilt is tunable according need down to 0.02 degrees. ► The presented method may lead the way to automatic zone-axis orientation.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 We studied the stability of force-feedback high-speed atomic force microscopy (HSAFM) by imaging soft, hard, and biological sample surfaces at various applied forces. The HSAFM images showed sudden topographic variations of streaky fringes with a negative applied force when collected on a soft hydrocarbon film grown on a grating sample, whereas they showed stable topographic features with positive applied forces. The instability of HSAFM images with the negative applied force was explained by the transition between contact and noncontact regimes in the force–distance curve. When the grating surface was cleaned, and thus hydrophilic by removing the hydrocarbon film, enhanced imaging stability was observed at both positive and negative applied forces. The higher adhesive interaction between the tip and the surface explains the improved imaging stability. The effects of imaging rate on the imaging stability were tested on an even softer adhesive Escherichia coli biofilm deposited onto the grating structure. The biofilm and planktonic cell structures in HSAFM images were reproducible within the force deviation less than ∼0.5 nN at the imaging rate up to 0.2 s per frame, suggesting that the force-feedback HSAFM was stable for various imaging speeds in imaging softer adhesive biological samples. Highlights ► We investigated the imaging stability of force-feedback HSAFM. ► Stable–unstable imaging transitions rely on applied force and sample hydrophilicity. ► The stable–unstable transitions are found to be independent of imaging rate.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 Principal component analysis (PCA) noise filtering is a popular method to remove noise from experimental electron energy loss (EELS) spectrum images. Here, we investigate the statistical behaviour of this method by applying it on a simulated data set with realistic noise levels. This phantom data set provides access to the true values contained in the data set as well as to many different realizations of the noise. Using least squares fitting and parameter estimation theory, we demonstrate that even though the precision on the estimated parameters can be better as the Cramér–Rao lower bound, a significant bias is introduced which can alter the conclusions drawn from experimental data sets. The origin of this bias is in the incorrect retrieval of the principal loadings for noisy data. Using an expression for the bias and precision of the singular values from literature, we present an evaluation criterion for these singular values based on the noise level and the amount of information present in the data set. This criterion can help to judge when to avoid PCA noise filtering in practical situations. Further we show that constructing elemental maps of PCA noise filtered data using the background subtraction method, does not guarantee an increase in the signal to noise ratio due to correlation of the spectral data as a result of the filtering process. Highlights ► We studied the statistical effect of a PCA and weighted PCA noise filter on EELS spectrum images. ► In a PCA scree plot, there are no pure noise or signal components. ► PCA filtering can introduce considerable bias, offsetting the benefits of better SNR. ► An evaluation criterion for estimating the validity of PCA filtering is proposed.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 We have examined the beam spot diameter at the anode of the scanning electron microscopy (SEM) in the near-field mode as a function of the anode-tip distance d . The detector lateral resolution of this type of microscopy is approximately equal to this spot diameter. For our calculations we have simulated the apex region of the tip with an ellipsoid of revolution of radii R 1 and R 2 with R 1 〉 R 2 as suggested by TEM images of the realistic tips. We have then solved the Laplace equation to obtain the electrostatic potential and to this we have added a spherical image potential. The calculated electrostatic field is highly asymmetric, being strong along the tip-axis and weakening quickly towards the sides. When a 3-dimensional WKB approximation is used to calculate the electron paths corresponding to such a potential, the latter are shown to bend significantly towards the vertical (tip-axis) direction producing a beam narrowing effect very similar to the beam narrowing effect we discovered for the traditional SEM case. When the values of R 1 , R 2 are chosen from fittings to the TEM images of the tips used in the experiments, the beam spot diameter W at the anode ( d =25 nm) varies from 12.5 nm to 9 nm depending on the fitted R 1 , R 2 . These values of W are considerably smaller than previously predicted by calculating solid angles of emission from spherical surfaces (41 nm) but also much closer to the detector lateral resolution (6–7 nm) obtained from differentiating the experimental current step. This trend continued at all other d examined. Furthermore the beam width W was found to decrease quickly with increasing sharpness S = R 1 / R 2 of the tip and then saturate. W is also decreasing with decreasing R 1 , R 2 with S kept constant. We deduce that the sharpness of the tip is important not only for creating high extraction fields but also for guaranteeing a very small beam spot diameter. Highlights ► This paper calculates the electrostatic potential around a stack of metallic ellipsoids. ► It then calculates the corresponding electron paths. ► Finally the beam-width of an NFESEM electron beam is obtained. ► Reasonable agreement with experimental data is found.

Description: Available online 22 December 2012 Publication year: 2012 Source: Ultramicroscopy The measurement of accurate and precise structure factors and Debye Waller (DW) factors quantitative convergent-beam electron diffraction (QCBED) permits the experimental determination of the electron density distribution and probing of interatomic bonding in crystals. The three QCBED methods used successfully for high precision measurements of low order structure factors to date, namely the zone axis method, the excited row method and the multi beam off zone axis technique, differ from each other regarding the crystal orientation relative to the incident electron beam. Consequently, the details of their respective dispersion surface representations differ regarding the number, relative amplitudes and phases of excited Bloch wave branches. Under the same experimental setup conditions, the factors most important to the degree of accuracy and precision achievable in electron density determination for crystals with QCBED methods ultimately depend on the sensitivity of the excited Bloch wave branches and the resultant contrast in the respective CBED patterns to changes in both structure and DW factors. In general, a QCBED pattern will be more sensitive to changes in both structure and DW factor, if it contains more and stronger excited Bloch wave branches, as the dynamic interactions of the Bloch waves increase the sensitivity of the pattern. In this work we analyzed Bloch wave excitation and dispersion surfaces for the three most popular QCBED methods. The analysis indicates, that the QCBED patterns obtained using the multi beam off zone axis orientation generally contain more and stronger excited Bloch wave branches. Hence, multi beam off zone axis diffraction patterns are more sensitive than the zone axis and the excited row patterns to changes in both DW and structure factors and therefore allow in differences to the other two methods simultaneous refinements effectively and robustly. Highlights ► We compare sensitivity to structure and DW factor change of zone-axis related CBED methods. ► Increasing number and amplitude of excited Bloch wave branches improves sensitivity. ► Off-zone axis CBED excites more strong Bloch waves than the zone-axis method. ► Off-zone axis CBED excites many anti-bound Bloch wave states that probe bond electrons. ► Off-zone axis methods are more sensitive to structure and DW factor change.

Description: February 2013 Publication year: 2013 Source: Ultramicroscopy, Volume 125 Atom probe tomography (APT) is used to investigate the composition of oxygen rich nanoparticles within a ferritic matrix in Fe-14Cr-2W-0.1Ti oxide-dispersion-strengthened (ODS) steel. This study investigates whether artifacts associated with APT analysis are the cause of a sub-stoichiometric oxide composition measurement. Bulk Y 2 O 3 is analyzed by APT, thus demonstrating the ability of the technique to measure near-stoichiometric composition measurements in insulating oxides. Through analysis of the sequence of ion hits on the detector during APT data acquisition, it is shown that a proportion of yttrium hits are spatially correlated but oxygen hits are not. Y–O based nanoparticles in a ferritic matrix are analyzed by APT using voltage pulsing and a laser pulsing with a range of laser energies from 0.3–0.8 nJ. When the material is analyzed using a high effective evaporation field, this influences the effect of trajectory aberrations, and the apparent size of the nanoparticles is reduced. Some reduction in Y:O ratio is observed, caused by high instances of multiple-ion evaporation events. From a detailed comparison between the results of APT analysis of the bulk Y 2 O 3, the nanoparticles in the ODS material are concluded to have an approximate Y:O ratio of 1:1. Highlights ► First APT analysis of bulk Y2O3 oxide. ► Comparison with the composition of YO based nanoclusters in ODS steel. ► Sequence of ion hits on the detector show that there is a bias towards Y loss. ► Nanoparticles are concluded to have a ∼1:1 Y:O ratio.

Description: Available online 21 December 2012 Publication year: 2012 Source: Ultramicroscopy Understanding the impact of noise and incomplete data is a critical need for using atom probe tomography effectively. Although many tools and techniques have been developed to address this problem, visualization of the raw data remains an important part of this process. In this paper, we present two contributions to the visualization of data acquired through atom probe tomography. First, we describe the application of a rendering technique, ray-cast spherical impostors, that enables the interactive rendering of large numbers (as large as 10 million plus) of pixel-perfect, lit spheres representing individual atoms. This technique is made possible by the use of a consumer-level graphics processing unit (GPU), and it yields an order of magnitude improvement both in render quality and speed over techniques previously used to render spherical glyphs in this domain. Second, we present an interactive tool that allows the user to mask, filter, and colorize the data in real time to help them understand and visualize a precise subset and properties of the raw data. We demonstrate the effectiveness of our tool through benchmarks and an example that shows how the ability to interactively render large numbers of spheres, combined with the use of filters and masks,leads to improved understanding of the three-dimensional (3D) and incomplete nature of atom probe data. This improvement arises from the ability of lit spheres to more effectively show the 3D position and the local spatial distribution of individual atoms than what is possible with point or isosurface renderings. The techniques described in this paper serve to introduce new rendering and interaction techniques that have only recently become practical as well as new ways of interactively exploring the raw data. Highlights ► Application of spherical impostor rendering to atom probe data visualization. ► Presented an interactive tool for visualizing atom probe tomography data. ► Presented a comparison of rendering performance of our tool with commercial tools. ► An example of how our tool can be used to understand atom probe tomography data.