Scanning transmission electron microscopy (STEM) allows atomic scale characterization of solid–solid interfaces, but has seen limited applications to solid–liquid interfaces due to the volatility of liquids in the microscope vacuum. Although cryo-electron microscopy is routinely used to characterize hydrated samples stabilized by rapid freezing, sample thinning is required to access the internal interfaces of thicker specimens. Here, we adapt cryo-focused ion beam (FIB) “lift-out,” a technique recently developed for biological specimens, to prepare intact internal solid–liquid interfaces for high-resolution structural and chemical analysis by cryo-STEM. To guide the milling process we introduce a label-free in situ method of localizing subsurface structures in suitable materials by energy dispersive X-ray spectroscopy (EDX). Monte Carlo simulations are performed to evaluate the depth-probing capability of the technique, and show good qualitative agreement with experiment. We also detail procedures to produce homogeneously thin lamellae, which enable nanoscale structural, elemental, and chemical analysis of intact solid–liquid interfaces by analytical cryo-STEM. This work demonstrates the potential of cryo-FIB lift-out and cryo-STEM for understanding physical and chemical processes at solid–liquid interfaces.

In recent years, an increasing number of studies utilizing in situ liquid and/or gaseous cell scanning/transmission electron microscopy (S/TEM) have been reported. Because of the difficulty in the preparation of suitable specimens, these environmental S/TEM studies have been generally limited to studies of nanoscale structured materials such as nanoparticles, nanowires, or sputtered thin films. In this paper, we present two methodologies which have been developed to facilitate the preparation of electron-transparent samples from conventional bulk metals and alloys for in situ liquid/gaseous cell S/TEM experiments. These methods take advantage of combining sequential electrochemical jet polishing followed by focused ion beam extraction techniques to create large electron-transparent areas for site-specific observation. As an example, we illustrate the application of this methodology for the preparation of in situ specimens from a cold-rolled Type 304 austenitic stainless steel sample, which was subsequently examined in both 1 atm of air as well as fully immersed in a H2O environment in the S/TEM followed by hyperspectral imaging. These preparation techniques can be successfully applied as a general procedure for a wide range of metals and alloys, and are suitable for a variety of in situ analytical S/TEM studies in both aqueous and gaseous environments.

A method is proposed to determine the effective detector area for energy-dispersive X-ray spectrometers (EDS). Nowadays, detectors are available for a wide range of nominal areas ranging from 10 up to 150 mm2. However, it remains in most cases unknown whether this nominal area coincides with the “net active sensor area” that should be given according to the related standard ISO 15632, or with any other area of the detector device. Moreover, the specific geometry of EDS installation may further reduce a given detector area. The proposed method can be applied to most scanning electron microscope/EDS configurations. The basic idea consists in a comparison of the measured count rate with the count rate resulting from known X-ray yields of copper, titanium, or silicon. The method was successfully tested on three detectors with known effective area and applied further to seven spectrometers from different manufacturers. In most cases the method gave an effective area smaller than the area given in the detector description.

We report on a robust method for chemical element-sensitive imaging by scanning electron microscopy (SEM). The commercial Auriga FE-SEM microscope (Carl Zeiss, Oberkochen, Germany), equipped with an energy-selective grid detector (EsB) as a part of the experimental setup, was applied for generation of chemical contrast at low accelerating voltages, which is gentle for sensitive samples. The EsB-grid detector, conceptually adapted by us as an energy retarding field analyzer (RFA), was used to detect the two-dimensional (2D) energy spectrum for the first time. The electron energy spectrum measured by sweeping the retarding grid potential revealed thresholds corresponding to electronic transitions in the specimen, followed by 2D-derivation treatment applied just at the observed thresholds. This allowed chemical mapping by SEM. In this report the 273 eV Auger transition in carbon deposited onto the Si(100) sample was chosen as a source for chemical contrast in the SEM image. In addition to Auger electrons, we expect analogous energy-selective contrast enhancement for inelastically scattered electrons, for example, in plasmonic contrast and elastically scattered electrons, for example in phase contrast, our method, proved for carbon, is expected to apply to a broader list of elements as a general capability of chemical mapping, at several-fold better lateral resolution when compared with energy dispersive spectroscopy (EDS).

In electron probe microanalysis where materials are coated with a thin conductive carbon coat before analysis, the X-ray intensity detected from a specimen may be affected to various degrees by the thickness of the carbon coating. Differences in the carbon film thickness between specimens and standards may lead to errors in analytical results, particular for lower energy X-rays. In this study, we demonstrate that the location and the distance of the specimen relative to the carbon tip in the coating chamber can affect the thickness of the carbon film produced on the specimen surface during carbon coating. The closer the specimen is to the carbon tip contacting point, the thicker is the carbon film deposited. A mathematical model to calculate the carbon film thickness at different locations on the coater plate is established, based on the assumption that carbon atoms evaporate from the carbon tip equally in all directions during the coating process. In order to reduce the differences in the carbon coating thickness, we suggest moving the carbon rod to a higher position, moving the thinner samples to the center and thicker samples to the edge of the coater plate, and using a rotating circular coater plate during coating.

When the value of the dispersion coefficient is “greatly different from 0.5,” as is the case for “free-electron” materials such as sodium (Na), the approximate expression for the volume plasmon critical wave vector (PCV) used by Hadji stops being valid and a different, more precise, expression must be used. Here a more accurate PCV formula is used to get a more accurate expression for plasmon scattering cross-section per atom (PSCA) species. This PSCA is then employed to calculate some physical quantities for several “free-electron” materials and together with the techniques from the quoted paper to determine values for physical quantities from amorphous silicon (a-Si) experimental data. The program source used to obtain these values is supplied. Any valid formula for the PSCA species is, in fact, relevant for use together with the two quoted techniques. The PCV and the dispersion coefficient have upper limits. Negative dispersion coefficient values are allowed. A PCV-related dimensionless universal function that can represent all ideal “free-electron” materials is given. “Not greatly different from 0.5” is mathematically expressed.