《[研究生用]X射线衍射及电子显微分析(下).pdf》由会员分享，可在线阅读，更多相关《[研究生用]X射线衍射及电子显微分析(下).pdf（226页珍藏版）》请在三一文库上搜索。
1、4. TRANSMISSION ELECTRON MICROSCOPY 4.1. Transmission Electron Microscope- INSTRUMENT 4.1.1. Electron Sources 4.1.1.1. THE PHYSICS OF DIFFERENT ELECTRON SOURCES We use two kinds of electron sources in TEMs: the first kind is called a thermionic source, which, as the name suggests, produces electrons
2、 when heated, and the second type is a field-emission source, which produces electrons when an intense electric field is applied to it. These sources are part of an assembly which we refer to as the “electron gun.“ Now, from a physics standpoint, it is really quite interesting to know the details of
3、 how electron sources work and theres a great deal of active research into new and improved sources. However, from a practical standpoint, you dont have to know too much of the physics, and we can summarize the essential points very briefly, using a few simple equations. Keep in mind two points as y
4、ou read about sources: Your TEM will use a thermionic source or a field-emission source and the two cannot be interchanged. Field-emission sources give “monochromatic“ electrons; thermionic sources are less monochromatic and give “whiter“ electrons. The analogy here is to X-rays or visible light. Yo
5、u dont always want to use “monochromatic“ electrons, even if the field-emission TEM did cost twice as much as a “conventional microscope would with a thermionic source. 4.1.1.1.A. Thermionic Emission If we heat any material to a high enough temperature, we can give the electrons sufficient energy to
6、 overcome the natural barrier that prevents them from leaking out. This barrier is termed the “work function“ () and has a value of a few electron volts. The physics of thermionic emission can be summarized in Richardsons Law, which relates the current density from the source, J, to the operating te
7、mperature, T in Kelvin kT eATJ = 2 4.1.1.1 where k is Boltzmanns constant (8.610-5 eV/K) and A is Richardsons “constant“ (A/m2 K2), which depends on the source material. From this equation then you can see that we need to heat the source to a temperature T such that energy greater than is given to t
8、he electrons; then they will escape from the source and be available to form an electron beam. Unfortunately, when we put a few eV of thermal energy into most materials they either melt or vaporize. So the only viable thermionic sources are either refractory (high melting point) materials or those w
9、ith an exceptionally low work function. In practice we use both types: tungsten has the necessary high melting temperature (3660 K) and lanthanum hexaboride (LaB6) has a low work function. If you look ahead to Table 4.1.1.1, youll see the relative values of J, T, and for tungsten and LaB6. We use se
10、veral different words to describe the sources. We sometimes call tungsten sources “filaments,“ because tungsten can be drawn into fine “thread“ which is about 0.1 mm in diameter and is similar to the filament used in an incandescent light bulb. The wire is bent into a V shape so theyre also called “
11、hairpin“ filaments, or they may be sharpened to a fine point. For decades these have been the standard source in most electron-beam instruments. LaB6, or other rare-earth boride crystals (which should not be called filaments) are usually grown with a orientation to enhance emission. Sometimes we cal
12、l both tungsten and LaB6 sources “cathodes“ because, as well see, the complete gun assembly acts as a triode system in which the source is the cathode. Thermionic sources: W hairpin Pointed W LaB6 and other low- materials, e.g., CeB6 So all you need to know from the physics is that heating up a ther
13、mionic source gives you a higher J. But there is a limit because higher temperatures shorten the source life through evaporation and/or oxidation. So we seek a compromise operating temperature, and we achieve this by operating under a condition called “saturation“. 4.1.1.1.B. Field Emission Field-em
14、ission sources operate on a fundamentally different principle than thermionic sources. The principle behind field emission is that the strength of an electric field E is considerably increased at sharp points, because if we have a voltage V applied to a (spherical) point of radius r then r V E = 4.1
15、.1.2 The technique of field-ion microscopy is another well established experimental tool. It requires specimens with a very fine needle shape, and so theres a lot of expertise available to help produce field-emission electron sources. One of the easiest materials to produce with a fine tip is tungst
16、en wire, which can readily be given a tip radius of orientation is found to be best. To allow field emission, the surface has to be pristine, that is, free of contaminants and oxide. We can achieve this by operating in UHV conditions (1000 4.1.2. How to “See” Electrons 4.1.2.1. ELECTRON DETECTION AN
17、D DISPLAY Images and diffraction patterns are different two-dimensional electron-intensity distributions which can be produced by scattering by the same object. We detect and display them in different ways depending on whether we are using a TEM or STEM, as well explain in Section 4.1.4. In a conven
18、tional TEM the images and diffraction patterns are static, because the incident beam is fixed, and so we can easily project them onto a viewing screen within the microscope column. TEM images, for example, are analog images of electron density variations in the image plane of the objective lens. We
19、cannot manipulate the image or its contrast in any way between the electrons leaving the image plane and being projected onto the viewing screen. We will briefly discuss the properties of the viewing screen. The manufacturer controls the initial choice of screen materials so you might think theres n
20、ot much need to understand this aspect in any depth. You might be surprised by the limitations you dont need to accept or the improvements which could be made. When we operate our TEM as a STEM, or we use a dedicated STEM, the image is not static; it is built up over time as the small probe is scann
21、ed across the area of interest. Under these circumstances, we detect the electron signals by various types of electronic detection. If we are seeking secondary electron (SE) or backscattered electron (BSE) signals, then these detectors sit in the specimen stage area. If we are seeking to image the s
22、ame forwardscattered electrons that we view on the TEM screen, the detectors are in the viewing chamber of the TEM. After weve detected any one of these signals, it is usually digitized and digital scanning images are presented on a fluorescent screen as an analog image. We often refer to this fluor
23、escent screen as the CRT, which is the acronym for “cathode-ray tube“ and a relic from the early days of electron physics. We should point out that the sequential or serial nature of the scanning image makes it ideal for on-line image enhancement, image processing, and image analysis. The signal fro
24、m any electronic detector can be digitized and electronically manipulated prior to display on the CRT, in a way that is impossible with analog images. We can adjust the digital signal to enhance the contrast or to reduce the noise. Alternatively, we can store the digital information and process it m
25、athematically. The availability of cheap memory and fast computers permits on-line image processing and the rapid extraction of quantitative data from the scanning image. Because of developments in computer technology, there is great interest in recording analog TEM images via a TV camera in order t
26、o digitize them and charge-coupled device (CCD) cameras are already available for on-line viewing and processing, particularly of HRTEM images. In attempting to compare the properties of detection and recording devices we often use the concept of the “detection quantum efficiency“ or DQE. If the det
27、ector is linear in its response, then the DQE is defined simply as 2 2 = in in out out N S N S DQE 4.1.2.1 where S/N is the signal-to-noise ratio of the output or input signal. So a perfect detector has a DQE of 1 and all practical detectors have a DQE 0.5) even at low input signal levels. ? The dyn
28、amic range of a CCD is high, making it ideal for recording diffraction patterns which can span an enormous intensity range. The major disadvantages are the speed and the expense. These devices can also be used as two-dimensional arrays for parallel-collection electron energy-loss spectrometers rathe
29、r than the more conventional one-dimensional silicon diode arrays. 4.1.2.3.D. Faraday Cup In conventional TEM there isnt much need to know the beam current, but in the AEM it is essential since there is often a need to compare analytical results obtained under identical beam current conditions. A Fa
30、raday cup is a detector that simply measures the total electron current in the beam. We dont use it for any imaging process, but rather as a way of characterizing the performance of the electron source, as we saw in Chapter 5. Once the electrons enter the Faraday cup, they cannot leave except by flo
31、wing to ground through an attached picoammeter that measures the electron current. A Faraday cup is a black hole for electrons. You can easily construct a Faraday cup to go in an SEM, but it is more difficult to design one that fits in the stage of a TEM. Fortunately, some manufacturers now incorpor
32、ate a Faraday cup in the specimen holder. You can measure the current by deflecting the beam into the cup or partially extracting the holder (Figure 4.1.2.4B). These cups are not ideal because they dont trap all the electrons. A dedicated Faraday-cup holder is shown in Figure 4.1.2.4A. The entrance
33、aperture is small and the chamber is relatively deep and lined with a low-Z material to minimize backscatter. If you tilt it slightly, the electrons have little chance of being scattered directly back. With such a holder you can only find the hole if you can image the upper surface with SE or BSE de
34、tectors, and if these are not available then you must have a cup with a hole in the lower surface too. When the cup is not tilted, the electrons go straight through; if you tilt the cup, then all the electrons are trapped as shown in Figure 4.1.2.4A. The way to ensure that you are measuring the maxi
35、mum current is to look at the picoammeter reading as you tilt the cup. Figure 4.1.2.3. (A) A single cell in a CCD array showing the storage of charge in the potential well under one pixel. If we vary the applied potential to rows of pixels in sequence, as in (B), one pixel row is shifted to the para
36、llel register, and is read out pixel by pixel, after which the next row is moved to the parallel register, and so on. The stored charge in each pixel is thus fed into an amplifier and digitized. If you dont have a Faraday cup, it is possible to get an approximate reading of the current by just measu
37、ring the current through an insulated line from a bulk region of the specimen and correcting for electron backscatter. Backscattering is independent of the accelerating voltage and approximately linear with atomic number up to about Z = 30. For example, the backscatter coefficient for Cu is about 0.
38、3 and for A1 it is about 0.15. It is also possible to deflect the beam onto the last beam-defining diaphragm and measure the current via an insulated feed-through (also correcting for backscatter). A B Figure 4.1.2.4. (A) Schematic diagram of a Faraday cup in the end of a side-entry specimen holder.
39、 The entrance aperture has to be found using SEs or BSEs. In (B) the holder is retracted slightly so the electrons fall into a cup on the tip of the holder. The electron current is measured as it goes to ground through a picoammeter attached to the outside of the holder. 4.1.2.4. WHICH DETECTOR DO W
40、E USE FOR WHICH SIGNAL? As we mentioned at the start of the section, the principal electron signals that we can detect are the forward-scattered electrons, which as well see in Section 9 are the most common TEM images, and the BSE and SE signals from the beam-entry surface of the specimen. Figure 7.
41、5. The various electron detectors in a STEM. Scintillator-PM detectors are invariably used for SE detection and semiconductor detectors for the BSE. The on-axis and annular forward-scattered detectors may be either type, depending on the microscope. Semiconductor detectors are only sensitive to elec
42、trons with sufficient energy (5 keV) to penetrate the metal contact layer. So we use these detectors mainly for high-energy forward-scattered imaging and high-energy BSE imaging. Because of the surface contact layer we dont use semiconductor detectors for low-energy SEs and a scintillator-PM system
43、is required. Remember that the scintillator may also be coated with Al to prevent visible light from generating noise. This coating would also prevent low-energy SEs from being detected. So for SE detection, either there must be no coating, or the electrons must be accelerated to an energy high enou
44、gh to penetrate the coating; we achieve the latter by applying a high kV (10 kV) positive bias to the scintillator. The capacitance is relatively high for semiconductor detectors, so they are not the detector of choice in dedicated STEMs where high scan-rate TV images are the normal viewing mode, i.
45、e., where you need a quick response. The scintillator-PM system is again preferred under these circumstances. As most microscopes move toward TV-rate display of scanning images it is likely that the scintillator-PM will be used increasingly for forward-scattered TEM imaging. Semiconductor detectors
46、may only be used for BSEs, which is not a major imaging mode in TEMs. A summary of all the various electron detectors in a TEM/STEM is given in Figure 7.5. Sometimes we examine specimens which themselves exhibit cathodo- luminescence under electron bombardment. A mirror is used to focus the light in
47、to a scintillator-PM system, and one such design is shown in Figure 4.1.2.6. This setup effectively prevents detection of all other signals, including X-rays, because the mirror occupies all the free space in the TEM stage. So you have to dedicate the TEM to CL detection alone and ignore other signa
48、ls. Figure 7.6. Cross section of a mirror detector below a thin cathodoluminescent specimen that collects light and focuses it into a spectrometer-PM system. The CL signal is usually very weak and so the detector has to be as large as possible, and it takes up much of the free volume in the TEM stag
49、e. 4.1.2.5. IMAGE RECORDING 4.1.2.5.A. Photographic Emulsions Although film is the oldest recording medium, it still retains sufficient advantages that we continue to use it in virtually all TEMs. Photographic emulsions are suspensions of silver halide grains in a gel. Electrons strike the halide, ionize it, and transform it to silver. The emulsions are usually supported on a polymer film or (very rarely) glass plates. Unlike polymer film, glass plates do not outgas and do not shrink during prepumping or processing. Ho