TEM Techniques Summary The TEM is an analytical instrument in which a thin membrane (typically < 100nm) is placed in the path of an energetic and highly coherent beam of electrons. Typical operating voltages are between 100-300kV. As the electrons transit the thin membrane, they interact with the atoms of specimen giving rise to a number of observable phenomena or signals. The resultant signals may be collected by the appropriate detector and analyzed to obtain information about the morphology, microstructure, composition, and chemistry of the specimen. The TEM may be used to form images of the specimen, produce electron diffraction patterns, or collect spectroscopic data. The TEM may be operated in a number of different modes, conventional imaging (CTEM), scanning transmission electron microscopy (STEM), diffraction mode, either selected area diffraction (SAD), or convergent beam diffraction (CBED), energy filtered imaging (EFTEM) and spectroscopy mode, X-ray energy-dispersive spectrometry (X-EDS) and electron energy loss spectrometry (EELS). The contrast observed in conventional TEM images is the result of three mechanisms; diffraction contrast, phase contrast, and mass-thickness contrast. Diffraction contrast occurs in crystalline specimens when the crystallographic planes that are oriented nearly parallel to the incident beam act as minute diffraction gratings that divert the post specimen electrons away from the optic axis. Diffraction contrast in a bright field (BF) TEM image causes the strongly diffracting regions of the image to appear dark in the image. BF diffraction contrast is enhanced by inserting an aperture into the back focal plane of the objective lens. Diffraction contrast is also the operative mechanism in dark field (DF) TEM imaging. A centered dark field (CDF) TEM image of a crystalline specimen is created by tilting the incident beam so that the incident beam approaches the specimen at an angle equal and opposite to a selected diffraction angle. The action of tilting the incident beam causes the transmitted beam to be laterally shifted off the optic axis of the TEM while selected diffracted beams are simultaneously directed down the optic axis. If an aperture is concurrently inserted into the back focal plane of the objective lens, only the rays passing through the central opening of the annulus will contribute to the image. By precisely varying and recording the tilt and rotation angles of the incident beam, reflections from specific crystallographic planes can be directed down the optic axis and thus contribute to the dark field images. Crystals with lattice planes that are diffracting down the optic axis at a given tilt angle will appear bright on the image. The second contrast mechanism, phase contrast, occurs due to the wave nature of the electrons. As the beam passes through various portions of the specimen the relative phase of the electron waves is modified. The superposition or recombination of the post specimen waves causes spatially varying regions of constructive and destructive interference that results in an observable intensity deviation or contrast. The third contrast mechanism, massthickness contrast occurs due to variations in specimen thickness and/or atomic composition. Electron scattering increases with increasing specimen thickness and increasing atomic number (Z). Thus, regions that are thicker or are composed of higher Z elements will strongly scatter the beam electrons and will appear darker on the image. Mass thickness contrast is an important contrast mechanism in amorphous specimens. STEM imaging is performed when the incident electron probe is focused to a few nanometers in diameter and is rastered over the specimen. The image is formed digitally as the intensity recorded at the each of the raster coordinates is mapped to the corresponding pixels in the image. Both bright field and dark field STEM imaging is possible. Diffraction contrast and mass-thickness contrast are observed in STEM images. Contrast that arises primarily due to atomic composition (Z contrast) is due to Rutherford scattering. Rutherford scattering is characterized by a large angular deviation from the optic axis. By using a high angular dark field detector (HAADF) and concurrently selecting a small camera length, Z-contrast becomes the dominant contrast mechanism in the observed HAADF STEM image.
Energy filtered EFTEM imaging, spectrum imaging, and electron energy loss spectroscopy EELS are possible because certain electron beam/atomic interactions cause the incident electrons to lose energy. Such interactions are collectively called inelastic scattering processes. As the incident highenergy electrons pass through the thin membrane they have the potential to interact with the core electrons of the constituent atoms. If an incident electron causes the ejection of a core electron from the specimen, then the incident electron loses a quantum of energy equal to the characteristic ionization energy of the ejected core electron. Thus, by using an electron spectrometer, which consists of several lenses, a magnetic prism and an energy-selecting slit, the emerging electrons can be filtered according to their post specimen energy. Energy filtered imaging and EELS are closely related techniques both based on the inelastic interactions between the electron beam and the specimen. The same system, the Gatan Image Filter (GIF), acts as both a parallel EEL-spectrometer as well as an energy-filtering system. The difference between the two techniques is a function of user-selected microscope operating parameters. The back focal plane (BFP) of the last projector lens in the TEM column serves as the object plane for the spectrometer. Selecting TEM lens settings that will place a diffraction pattern in the object plane and consequently the dispersion plane of the spectrometer allows the post slit lens assembly of the spectrometer to form an image on the square 1024 x 1024 pixel charged coupled device (CCD) camera. The microscope is said to be in image mode but the spectrometer is diffraction coupled. If the energy-selecting slit is inserted into the dispersion plane of the spectrometer then an energy filtered (EFTEM) image is formed. By adjusting the strength of the magnetic prism and the width of the energy-selecting slit (typically 10-50eV), only electrons of a specifically determined energy are allowed to contribute to the image. A zero loss EFTEM image is formed when only electrons that have lost no energy while transiting the specimen are selected to contribute to the image. If the spectrometer is operated as described above with no energy slit inserted then an unfiltered image is formed. Although an unfiltered image is collected with the GIF camera, the primary contrast mechanisms are the same as in any BF TEM image. The GIF system acts as a spectrometer when the TEM lens settings are adjusted so that an image is present in the object and dispersion planes of the spectrometer. The post slit lens assembly then projects a dispersed energy spectrum on a linear 1024 x 1 array CCD. The microscope is said to be in diffraction mode but the spectrometer is said to be image coupled. Because the entire spectrum is collected simultaneously, the technique is specifically referred to as parallel electron energy loss spectroscopy (PEELS). A PEELS spectrum from a single spot may be collected either with the TEM in conventional imaging mode, diffraction mode or in STEM mode. A PEELS line scan or PEELS areal scan is only possible when operating in STEM mode. A PEELS areal scan is more accurately referred to as a spectrum image because a complete spectrum is recorded from each (x,y) coordinate in the raster. Thus, a data cube is generated and elemental maps can be subsequently generated from the stored data.
Figure 1. Schematic diagram of a Gatan imaging filter energy (GIF) showing collection planes for an energy loss spectrum J(E) and an energy filtered image, respectively. PEELS and EFTEM imaging are tools for qualitative and quantitative elemental analysis in very thin specimens. Detection limits can be as low as 0.1% by weight. Because PEELS is characterized by exceptionally high energy resolution it can with very careful parameterization be used to obtain chemical or bonding information. Both EFTEM and PEELS provide a means for elemental mapping based on the characteristic ionization energies of the elements.
Inelastic scattering events occur when the electrons transiting the specimen lose energy due to interactions with the atoms of that specimen. Such events are completely characterized by recording the scattering intensity as a function of the variables J(x, y, z, q x, q y, E), where x, y and z refer to position in real space, q x, q y are scattering vectors and E is energy. It would be cumbersome to record and manipulate this large amount of data simultaneously. Thus, as a matter practicality, the energy loss data acquisition is restricted into categories. An energy loss spectrum J(E) which is a single PEELS spectrum where the number of electrons (intensity) having lost a given amount of energy is plotted as a function of energy. Figure 2. An energy loss spectrum J(E)
A line plot J(y, E) or J(q y, E) is created as the electron beam is advanced along a spatially defined path on the specimen. A single spectrum is collected at each position of the beam. Post acquisition, the individual spectra are analyzed and the number of electrons (intensity) having lost energy within a range specified by the analyst is plotted as a function of position on the specimen. The energy ranges selected for analysis typically correspond to known characteristic energy loss ranges corresponding to specific elements. Figure 3. A single energy loss spectrum J(E) in which an energy range has been defined by the green energy window. A line spectrum J(y, E) (figure 4) is generated point by point from integrated intensity within the limits of the energy window for each spectrum collected along the path of the beam. The intensity is plotted as a function of beam position on the TEM specimen. Figure 4a. High angle annular dark field image with red line showing path of the beam along the TEM specimen; b line plot showing intensity of carbon signal as a function of beam position.
Figure 5. Line plot shown in figure 4 overlaid on high angle annular dark field image.
An energy selected or EFTEM image J(x, y,) or filtered diffraction pattern J(qx, qy,) is created when the lens configurations are adjusted so that an image is projected on the GIF camera. In a fashion that is conceptually similar to the placement of the energy window over a desired range in the line spectrum shown above, an energy range is selected by placing a physical slit in the path of the spectrum and the subsequent image is formed with only electrons that have passed through the slit, i.e., electrons that have lost a specified amount of energy. Figure 6a. EFTEM image formed with only the electrons that have undergone no energy loss to elastic scattering events; b. EFTEM image formed with only the electrons that lost energy via momentum transfer with the nitrogen K-shell electrons; b. EFTEM image formed with only the electrons that lost energy via momentum transfer to the nitrogen K-shell electrons; c. EFTEM image formed with only the electrons that lost energy via momentum transfer to the titanium L-shell electrons; d. EFTEM image formed with only the electrons that lost energy via momentum transfer to the oxygen K-shell electrons e. multicolor overlay of EFTEM images each formed with only the electrons that lost energy via momentum transfer to the elements listed in the key. NanoSpective, Inc. 3267 Progress Drive Suite 137 Orlando, Florida 32826
A spectrum image J(x, y, E) an energy loss spectrum is recorded at each pixel which generates a cube of data. From this cube either a PEELS spectrum or EFTEM image may be retrieved post analysis. Figure 7. Cube of data resulting from spectrum imaging. The x and y axes represent the position coordinates on the TEM specimen. The ΔE axis represents the distribution of energy lost by the incident electrons while transiting the TEM specimen, where the energy scale is increasing from top to bottom. Each xy plane contains an energy filtered image J(x,y) where the energy loss to be mapped is a coordinate on the vertical energy axis. Each vertical column contains a PEELS spectrum J(E) where the xy coordinates correspond to the collection location on the TEM specimen.