Grain contrast imaging in FIB and SEM - Chalmersfy.chalmers.se/~f10mh/Halvarsson/EM_intro_course_files/Grain cont… · 3.1. Channeling contrast The depth of penetration of the incoming

Grain contrast imaging in FIB and SEM

S Canovic1, T Jonsson and M Halvarsson

Microscopy and Microanalysis, Department of Applied Physics, Chalmers University of Technology, SE-412 96 Göteborg, Sweden

E-mail: [email protected]

Abstract. Grain contrast imaging can be performed with several techniques. In order to be able to choose the most suitable one it is important to know which techniques are available and also to be aware of the strengths and weaknesses of each technique. In this work, the grain contrast imaging is performed with secondary electrons, backscattered electrons, forward scattered electrons, transmitted electrons in the scanning electron microscope, and with secondary electrons in the focused ion beam instrument. The advantages and disadvantages of each method are discussed in order to make it easier to choose the most appropriate technique for grain contrast imaging.

1. Introduction Grain contrast imaging can be performed by using a wide range of techniques and choosing the most suitable technique is not a straightforward task. Historically, etching and light microscopy has been used for imaging grain contrast. Instead of photons, electrons and ions can be used for imaging with a better resolution. The most widely used imaging techniques where electrons and ions are used as imaging sources are scanning electron microscopy (SEM) and focused ion beam microscopy (FIB), respectively. Imaging with secondary electrons (SE), backscattered electrons (BSE) and forward scattered electrons (FSE) are the most commonly used modes for imaging grain contrast in the SEM. Another way to image grain contrast with high spatial resolution is by using thin foil specimens, instead of bulk specimens, i.e. scanning transmission electron microscopy (STEM) in the SEM.

The aim of this paper is to illustrate grain contrast imaging with the techniques mentioned above and also to discuss the strengths and weaknesses of each technique.

2. Experimental Three different samples have been used in this work. These are: (i) a κ-Al2O3/TiC multilayer coating deposited on cemented carbide by using chemical vapour deposition (CVD); (ii) an oxidized pure iron sample with a thick iron oxide scale formed after oxidation in oxygen at 600°C for 1 hour and; (iii) a Ti(C,N) coating deposited on cemented carbide by using moderate temperature (MT) CVD.

Two different instruments have been used: (i) a Leo Ultra 55 FEG SEM equipped with an Everhart-Thornley (E-T) SE detector, a solid state BSE detector, an in-lens SE detector, an FSE detector (mounted on a Nordlys electron backscattered detector (EBSD)), and a STEM detector and; (ii) an FEI 200 THP FIB with a secondary electron detector. 1 To whom any correspondence should be addressed.

Electron Microscopy and Analysis Group Conference 2007 (EMAG 2007) IOP PublishingJournal of Physics: Conference Series 126 (2008) 012054 doi:10.1088/1742-6596/126/1/012054

c© 2008 IOP Publishing Ltd 1

3. Results and Discussion

3.1. Channeling contrast The depth of penetration of the incoming electrons is affected by the regular arrangement of atoms in crystalline materials. If the incoming electrons encounter a low density of atoms, there is a lower probability that these will return to the surface as BSEs. The channeling contrast is carried by the high-energy fraction of the BSEs [1]. The channeling contrast effect is less pronounced for samples with no or low tilt due to the higher fraction of low-energy BSEs in this case.

Figure 1 shows three images of the fine-polished Ti(C,N) sample. The image in figure 1(a) is acquired with the E-T detector. The contrast present in this image is due to electron channeling. Even though the positively biased E-T detector primarily collects SEs and low-energy BSEs, also some high-energy BSEs are detected [1], which gives channeling contrast. In addition, the SE2s (SEs created by BSEs leaving the sample) also contribute to the channeling contrast.

Figure 1(b) shows an image that is acquired with the in-lens detector, which is placed in the electron column. The collection solid angle for high-energy BSEs is larger for the in-lens detector than for the E-T detector, which is placed on the side. This could be one reason for the better contrast in figure 1(b) than in figure 1(a). Another reason could be the effect of the in-lens detection system. If more SE2 electrons are collected by the in-lens detector than the E-T detector, this will contribute to the higher contrast.

Figure 1: SEM images of the fine-polished Ti(C,N) surface acquired with (a) E-T (b) in-lens and (c) BSED. The images are presented as acquired at the microscope. The brightness, in (a) and (b), is set to zero in order to achieve maximum contrast. The acceleration voltage is 10 kV and the working distance is 3 mm.

Figure 1(c) is obtained with the BSED. The high grain orientation contrast in this image is

explained by the high collection efficiency of the high-energy BSEs, which carry the channeling contrast information.

The channeling contrast illustrated above is even more apparent when the sample is tilted 70° [1]. This effect is used for imaging with FSEs. Figure 2(a) shows an SEM image illustrating the strong grain orientation contrast of the fine-polished Ti(C,N) sample by using FSEs. In addition, the primary beam interacts with the lattice, creating a pattern of Kikuchi lines that can be detected with an EBSD system. By recording these lines a band contrast map can be created. The map represents the contrast in the Kikuchi line patterns. Thus, the patterns do not need to be indexed; they can be constructed directly, without knowledge about the crystal structure of the material. As grain boundaries produce less band contrast they appear dark in these maps. This is illustrated in figure 2(b), which shows the band contrast for the fine-polished Ti(C,N) sample. The advantages with FSE imaging are the high grain orientation contrast and the possibility to get more useful information from the EBSD analysis. The most important drawback with this technique is that the sample surface has to be very fine-polished, which is a relatively time-consuming process.

a b c

400 nm


2

Figure 2: (a) an SEM image, of the fine-polished Ti(C,N) surface, acquired with FSEs. (b) Band contrast of an EBSD pattern.

In order to investigate the resolution of grain contrast images of samples that are not tilted a series

of images were acquired. Figure 3 shows the fine-polished Ti(C,N) sample viewed at 200 kX magnification. Figure 3(a) (in-lens, 10 kV) and 3(b) (BSED, 10 kV) both show relatively high resolution, in the order of 10 nm. As mentioned above, the channeling contrast is carried by the high-energy BSEs. This means that only BSEs from the upper part of the pear-shaped interaction volume are contributing to the channeling contrast. If the acceleration voltage is increased to 30 kV (from 10 kV) the resolution does not decrease significantly, as is seen in figure 3(c) (BSED, 30 kV). Although the total interaction volume increases, the upper part of the interaction volume does not increase considerably. This leads to only a slight loss in resolution in the grain contrast BSED image obtained at the higher voltage. However, the contrast is lower in figure 3(c), which could depend on a lower fraction of high-energy BSEs for higher acceleration voltage.

Figure 3: SEM images of the fine-polished Ti(C,N) surface acquired with (a) in-lens (10 kV) (b) BSED (10 kV) and (c) BSED (30 kV). The working distance is 3 mm.

The procedure for SE and BSE imaging is fast and straightforward. One disadvantage for the in-lens SE images is the problem with contamination during acquisition. As the in-lens SE detector is very surface sensitive, contamination will be very apparent in the images. This means that the samples have to be cleaned carefully (e.g. with a plasma cleaner) and also the microscope chamber has to be clean.

3.2. Other contrast mechanisms Another way to image grains with high spatial resolution in the SEM is by imaging thin foil specimens, instead of bulk specimens, with a STEM detector. This leads to a reduction of the interaction volume without reducing the acceleration voltage. The STEM unit used in this work consists of an electron detector with the ability of bright and dark field imaging. A STEM bright field image of the Ti(C,N) coating (figure 4) exhibits a high grain orientation contrast. The use of high primary beam energies causes low aberration, which results in high-resolution (< 1 nm) images [2]. A disadvantage is that the production of a thin foil specimen is relatively time-consuming.

a b c

a b

150 nm

5 µm 4 µm


3

Figure 4: A STEM image, acquired at 30 kV acceleration voltage, of the Ti(C,N) coating showing a high grain orientation contrast.

Due to its high spatial resolution, the FIB can be used as an imaging tool. Figure 5(a) shows a FIB

image of an ion-milled cross section of iron oxide. The grain orientation contrast in this image is caused by the channeling of the incident ions between lattice planes of the specimen [3]. This means that different crystal orientations affect the number of secondary electrons escaping from the specimen, which makes the different grains appear with different contrast. However, FIB is not always suitable for imaging grain orientation contrast. Figure 5(b) shows a FIB image of a κ-Al2O3/TiC multilayer coating. No grain orientation contrast is visible in the thicker, dark κ-Al2O3 layers, which are polycrystalline, as shown in the TEM micrograph in figure 5(c). Some advantages of using the FIB are the relatively good spatial resolution (~5 nm) and the high grain orientation contrast for many materials. Some disadvantages are the difficulties of imaging non-conducting samples and also the damage created by the ion beam impact on the samples.

Figure 5: (a) A FIB image of an ion-milled cross section of iron oxide, (b) a FIB image of a κ-Al2O3/TiC multilayer coating, (c) a TEM micrograph showing two polycrystalline κ-Al2O3 layers.

4. Concluding remarks There are several factors influencing the choice of method for grain contrast imaging. One important factor to be considered is the type of material to be imaged. For example, the FIB cannot be used for imaging too non-conducting samples. Another determining factor is the availability of instruments. Further, since some of the methods require more sample preparation than others the availability of the sample preparation equipment and time has also to be taken into account. The spatial resolution necessary is also important to consider before choosing method.

Acknowledgments This work was financed by the Swedish Foundation for Strategic Research program CROX, the National Graduate School in Materials Research and Knut and Alice Wallenberg foundation.

References [1] J. Goldstein et al., Plenum Press, New York, 2003. [2] J. P. Vermeulen, A Novel STEM Detector System, 2005 Imaging & Microscopy 1 22–23 [3] M. W. Phaneuf, Micron 30 (1999), 277-288.

a b c

1 µm

5 µm 4 µm 1 µm

Iron oxide

Iron

κ-Al2O3 TiC/κ-Al2O3

substrate

TiC

TiC κ-Al2O3


4

Documents

Grain contrast imaging in FIB and SEM - Chalmersfy.chalmers.se/~f10mh/Halvarsson/EM_intro_course_files/Grain cont… · 3.1. Channeling contrast The depth of penetration of the incoming