Physics

We discuss several ways of using Fourier-ratio deconvolution to process low-loss spectra. They include removal of the tail arising from the zero-loss peak, extraction of the spectrum of a particle from data recorded from the particle on a substrate, separation of the bulk and surface components in spectra recorded from samples of the same composition but different thickness, and investigation of interface energy-loss modes. We also demonstrate the use of a Bayesian-equivalent procedure based on the Richardson-Lucy algorithm.

Although the theory of high-angle elastic scattering of fast electrons is well developed, accurate calculation of the incident-energy threshold and cross section for surface-atom sputtering is hampered by uncertainties in the value of the surface-displacement energy E d and its angular dependence. We show that reasonable agreement with experiment is achieved by assuming a non-spherical escape potential with E d = (5/3) E sub , where E sub is the sublimation energy. Since field-emission sources and aberration-corrected TEM lenses have become more widespread, sputtering has begun to impose a practical limit to the spatial resolution of microanalysis for some specimens. Sputtering can be delayed by coating the specimen with a thin layer of carbon, or prevented by reducing the incident energy; 60 keV should be sufficiently low for most materials.

Contamination is a long standing problem in electron microscopy . It can arise from poor vacuum, from sample handling and storage, or from the nature of the sample nature. We are systematically investigating candidate sources of contamination using a common TEM substrate: 10-nm thick Si 3 N 4 membrane, with typical storage methods and practices. We use a dry-pumped Hitachi H 9500 ETEM whose base pressure is1.4x10 -5 Pa without liquid nitrogen (LN 2 ). The measured residual gas is composed mainly of water vapor (8x10 -7 Pa) and N2 (about 3x10 -7 Pa), both decreasing to about 2x10 -7 Pa with LN 2 . We use electron energy-loss spectroscopy (EELS), energy-filtering thickness maps and bright-field TEM imaging in the H9500 to characterize the contamination build up, and SEM imaging in Hitachi S 5500 to image the surface morphology of the contamination dots. The Si 3 N 4 membranes were examined in the TEM as received, after cleaning using a ZONE cleaner [4] with LN 2 trap of the microscope, either cooled (ZONE-LN 2 ) or at room temperature (ZONE-RT), or without LN 2 in the cold trap and after heating the Si 3 N 4 membrane on a hot plate at 50 ºC (Hot-50C) prior to inserting in the microscope. A new Si 3 N 4 membrane was used for each experiment and the variation of the contamination rate was evaluated for 5-min increments of the cleaning process. The membrane was irradiated by a probe of 37 nm FWHM with 0.83 nA total current for 300 seconds, resulting in a 250 C/cm 2 irradiation dose. Five areas were measured for each condition and a thickness map and a bright-field image was then collected from the irradiated area using a broad illumination. For each condition, the experiment was performed on five locations of the Si 3 N 4 membrane and the volume of resulting contamination measured in nm 3 by summing over a 134x134 nm 2 thickness map, subtracting the contribution of the SiN membrane.

Often it is radiation damage of the specimen rather than the ultimate resolution of a microscope that determines whether an object can be imaged at a desired resolution. For that reason, individual organic molecules cannot be currently imaged in a TEM. The radiation sensitivity is determined by the sample, current density and microscope acceleration voltage [1] and cannot be significantly improved but it is possible to adjust instrument parameters such that the information per unit electron dose is maximized. In the example below, we assume a "molecule" suspended between tips of an in-situ STM holder, as illustrated in . The hydrogen atoms were substituted by iodine, which decreases radiation sensitivity of the molecule and increases image contrast at the iodine sites.

We have investigated the mechanism of radiation damage (RD) in pentacene thin films by diffraction ring fading (loss of crystallinity). Pentacene has application in large area flexible organic electronics . Pentacene films can be studied by TEM if artifacts from RD are avoided.

Spatially resolved energy-loss near-edge structure (ELNES) reveals geometric and electronic structure of materials at high spatial resolution [1]. In scanning transmission ele ctron microscope (STEM), ELNES allows to study the valence state of atoms at interfaces of a nanocomposite system. The obtainable signal-to-noise ratio (SNR), set by specimen drift and radiation damage, is an obstacle in obtaining interpretable ELNES data at high spatial resolution. The white line ratio (WLR) is often used as a "fingerprinting" technique in identifying the oxidation of transition metals (TMs) [2], but hard to extract when the SNR is low. We recently found that a broad postpeak following the L-edge, observed in thin specimens where plural scattering is negligible, is characteristic of TMs in an oxidized state [3]. The postpeak was applied to identify surface oxidation of nanoparticles (NP) by examining average oxidation state of many NPs [4]. Here we show that the postpeak can be a convenient and sensitive indicator of oxidation of single iron NP by forming point analysis using a STEM with a small (0.5 nm) probe. The measurements were performed on isolated Fe NPs embedded in silica, shown in the high-angle annular dark-field (HAADF) image . The data were obtained using a JEOL 2200 FS (a 200 kV Schottky field emission instrument equipped with an W filter). shows the ELNES of Fe L-edges recorded at different sites of a selected particle, as labelled in HAADF image in . The signal comes mostly from the surface of the NP when the beam is focused at its edge, and mainly from the interior when the beam is focused at the center. A broad postpeak, about 40 eV above the onset of Fe L 3 , appears in the spectrum recorded from the edge (location "2" in ), but not in the spectrum taken from the center (location "1"). Comparison of "1" and "2" suggests a very thin oxide layer at the interface between the NP and silica matrix. A spectrum acquired in between nanoparticles (location "3") shows a prominent postpeak despite that SNR is much worse than the spectrum recorded at "2". This shows that residual iron left behind after an annealing step is in oxidized state [5].

Magnesium oxide (MgO) has been of interest for several decades as a promising tunable broadband laser due to its vacancy defects (color centers). In this work we introduced color centers into MgO nanocube by electron irradiation in a transmission electron microscope (TEM). Square nano-holes were formed from the electron-exit face using 100 and 300 keV electrons, and a broad O-vacancy (color-center) absorption peak around 4.1-6.6 eV was observed by valence-electron energy-loss spectroscopy (VEELS). We investigated the mechanism of MgO damage by high-energy electron beams. The hole formation is believed to involve a mixed removal of diatomic MgO molecules as well as Mg and O species in stoichiometric proportion. Observations using high-resolution transmission electron microscopy (HRTEM) and VEELS suggest that bulk O-vacancies are generated near the electron-exit face, due to the forward momentum transferred from fast-electron collisions and the Coulomb attraction of negative O-ions by the positively charged MgO surface.

We observe bandlike transport in pentacene and functionalized pentacene thin films using time-resolved terahertz pulse spectroscopy. The measured transient photoconductivity exhibits fast ͑Ͻ400 fs͒ photogeneration of mobile charge carriers and reveals a transient carrier mobility that increases as the temperature decreases from 300 K down to 10 K, indicative of bandlike transport over subpicosecond time scales. A wavelength-independent photoconductive signal is observed. The transient photoconductivity in the thin-film samples exhibits a single-exponential decay, whereas a power-law decay is seen in single-crystal samples.

We discuss several ways of using Fourier-ratio deconvolution to process low-loss spectra. They include removal of the tail arising from the zero-loss peak, extraction of the spectrum of a particle from data recorded from the particle on a substrate, separation of the bulk and surface components in spectra recorded from samples of the same composition but different thickness, and investigation of interface energy-loss modes. We also demonstrate the use of a Bayesian-equivalent procedure based on the Richardson-Lucy algorithm.

Negative tone patterns were created by polymer-contamination writing in a transmission electron microscope. Typical line edge roughness of the polymer lines was found to be below 1 nm. The patterns were transferred to bismuth films using an anisotropic ion etch, resulting in final bismuth line edge roughness less than 2 nm. This low roughness was achieved by growing the bismuth film such that the same crystallographic planes were exposed to the etching flux of ions.

We report detailed investigation of high-resolution imaging using secondary electrons (SE) with a sub-nanometer probe in an aberration-corrected transmission electron microscope, Hitachi HD2700C. This instrument also allows us to acquire the corresponding annular dark-field (ADF) images both simultaneously and separately. We demonstrate that atomic SE imaging is achievable for a wide range of elements, from uranium to carbon. Using the ADF images as a reference, we studied the SE image intensity and contrast as functions of applied bias, atomic number, crystal tilt, and thickness to shed light on the origin of the unexpected ultrahigh resolution in SE imaging. We have also demonstrated that the SE signal is sensitive to the terminating species at a crystal surface. A possible mechanism for atomic-scale SE imaging is proposed. The ability to image both the surface and bulk of a sample at atomic-scale is unprecedented, and can have important applications in the field of electron microscopy and materials characterization.

We have used electron energy-loss spectroscopy to search for differences in the energy-loss near-edge structure of SnO, SnO 2 , and an intermediate oxide, with a view to distinguishing them unambigously. We have found that the oxygen K edge exhibits clear differences that can be used for fingerprinting each phase. The oxygen edge appears at the same position for each phase whereas a chemical shift of the Sn M 4,5 edge of about 3.5 eV was observed between phases with Sn in 2+ and 4+ oxidation states. Both observations can be used to distinguish between the three phases, allowing their on-line identification within nanostructured materials.

The electronic structures of cubic and hexagonal phases of GaN have been investigated by high-resolution electron energy-loss spectroscopy in a monochromated transmission electron microscope. Both the Ga-L 2,3 and N K-edges were measured. The data are compared to the latest versions of two different calculation schemes: band structure and multiple-scattering calculations. We have found that both methods are capable of giving results that can be compared quantitatively to the experiment. Small discrepancies with experiment could be eliminated by future developments in the implementation of these methods.

We have optimized a bright-field transmission electron microscope for imaging of high-resolution radiation-sensitive materials by calculating the imaging dose n 0 needed to obtain a signal–to-noise ratio (SNR)=5. Installing a Zernike phase plate (ZP) decreases the dose ...