3D Elemental Mapping in the ESEM - A Combination of Serial Block-face SEM and EDS
- Fig. 1: Backscattered electron image of the block-face of an aluminum-copper alloy. Compositional contrast (1) and channeling contrast (2).
- Fig. 2: Sketch of the SBFSEM system: an ultramicrotome is installed in the specimen chamber of the microscope. The Everhart-Thornley detector is complemented by an EDS system for mapping.
- Fig. 3: 3D elemental maps for aluminum (a), copper (b), magnesium (c) and manganese (d). The bottom color bar gives the atomic percentage of each element. Region 1 is related to Mg2Si, region 2 to Al2CuMg and region 3 represents inhomogeneous precipitates composed of Al2Cu and Al3Mn
- Fig. 4: Color-coded 3D elemental reconstruction of the cubic ROI marked in fig. 3. The region numbers of fig. 3 correspond to the colors shown in the legend (1: Mg2Si, 2: Al2CuMg). Phases of different elemental composition (Al2Cu, Al3Mn) in the inhomogeneous precipitates (3) can be distinguished by different colors.
The combination of serial block-face scanning electron microscopy and energy dispersive x-ray spectroscopy (EDS) is demonstrated for 3D elemental analysis of materials, enabling investigations of large sample volumes compared to other tomographic techniques. The approach involves cutting a stack of 200 slices from an aluminum-copper alloy specimen and generating EDS maps after each cut. The resulting 3D reconstruction reveals phases of different elemental composition on the sub-micrometer scale.
Serial Block-face Scanning Electron Microscopy
Serial block-face scanning electron microscopy (SBFSEM) was introduced in 2004 . This method is particularly suitable for the low vacuum mode [2, 3] of an environmental scanning electron microscope (ESEM) with built-in ultramicrotome. It is based on a serial slicing and imaging process (block-face area conventionally 500 µm x 500 µm) typically using a backscattered electron (BSE) detector for compositional contrast . The method was originally developed for biological and medical applications  and has recently found its way into materials sciences [5, 6].
The low vacuum mode of the ESEM is necessary for stable imaging of electrically non-conductive specimens, while electrically conductive materials can be investigated in the conventional high vacuum mode. Figure 1 shows a micrograph of the block-face of an aluminum-copper alloy specimen with typical compositional contrast (1) and even channeling contrast (2).
The Idea: 3D Elemental Mapping
Compositional contrast provides basic information about the phase distribution of a material, while EDS enables qualitative and quantitative elemental analysis. The combination of SBFSEM and EDS thus appears to be a very promising approach. Similar methods have already been successfully applied in focused ion beam (FIB) technology [7, 8] or transmission electron microscopy (TEM) .
Materials and Methods
Analyses were carried out using an ESEM Quanta 600 FEG (FEI, Eindhoven, the Netherlands) and a serial block-face sectioning tool, 3View of Gatan, (Pleasanton, CA, USA), which is controlled by Digital Micrograph software.
The EDS maps were recorded with an X-Max Silicon Drift Detector (SDD) from Oxford Instruments Analytical, UK, featuring an 80 mm² active area and equipped with INCA software. Automatic sectioning and mapping was accomplished using the program "AutoIT" (version 18.104.22.168; 1999-2010 Jonathan Bennett) . The investigated aluminum-copper alloy specimen of the type EN AW‑2024 T351 (by AMAG, Ranshofen, Austria) was supplied by Dr. Thomas Koch, Vienna University of Technology.
Development of a New 3D Analytical Technique
The new approach is sketched in figure 2. The specimen (monolithic aluminum rod, precut in a conventional ultramicrotome) is cut by a diamond knife and then its block-face is imaged by an Everhart-Thornley detector. Additionally EDS maps are recorded. When these two processes are completed the specimen is moved in the z direction to repeat the cutting and imaging process until the desired stack size is reached. The resolution of this method is mainly affected by the electron energy and the material to be investigated .
An electron energy of 5 keV was used in order to minimize the interaction volume of the electrons in the material. X-ray data acquisition was performed with a resolution of 512 x 416 pixels. The evaluated line energies were 1.487 keV (K line) for aluminum, 0.928 keV (L line) for copper, 0.636 keV (L) for manganese, 1.254 keV (K) for magnesium and 1.740 keV (K) for silicon. A total of 200 slices with a thickness of 100 nm were cut enabling a recorded volume of (42.7 x 34.5 x 20.0) µm³. After data acquisition the elemental maps were quantified with a reduced resolution of 256 x 208 pixels using INCA software.
Results and Discussion
Figure 3 shows the 3D elemental maps for the selected elements aluminum (a), copper (b), magnesium (c) and manganese (d) (the silicon map is not shown). The bottom color bar gives the atomic percentage of each element. While region 1 in figure 3 can be assigned to Mg2Si precipitates, the majority of the precipitates are composed of Al2CuMg (marked as region 2). Heterogeneous precipitates (region 3) can be related to Al2Cu and Al3Mn. These results are in agreement with the properties of the Al 2024 aluminum-copper alloy described in . A 3D reconstruction was performed for a cubic region of interest (ROI) in the center of the 3D elemental maps. The regions of different elemental composition are shown in color-coded form in figure 4.
Summary and Outlook
This method enables a time saving quantitative elemental reconstruction of large sample volumes compared to FIB or TEM and represents an enhancement of mere EDS mapping. Precise cutting is performed without any contamination, e.g. by Gallium ions. This new technique has the potential to provide new insights into sliceable materials with complex chemical structures in the sub-micrometer scale.
Up to now it was only possible to combine the Everhart-Thornley SE detector with the EDS mapping process. Although the detected secondary electron (SE) are partially so called SE2 electrons  providing compositional contrast, it would be optimal to use a backscattered electron (BSE) detector. A further improvement would be the integration of the BSE detector, microtome control and EDS mapping in one user environment. Additionally, the compositional contrast of a BSE detector could be helpful for further data correction routines as mentioned in .
This work was financially supported by Austrian Cooperative Research (Project No. BMWFJ-98.175/0043-C1/11/2009). It was enabled by the very generous support of Dr. Mario Strasser and Benjamin Gossler from Oxford Instruments GmbH, Wiesbaden, Germany.
 Denk W. and Horstmann H.: PLoS Biol 2(11): e329 (2004)
 Danilatos G.D.: Academic Press 71, 109-250 (1988)
 Stokes D.J.: John Wiley & Sons, Ltd (2008)
 Goldstein J. et al.: 3rd ed. Kluwer (2003)
 Zankel A. et al.: J. Microsc. 233, 140-148 (2009)
 Reingruber H. et al.: J. Membr. Sci. 372, 6674 (2011)
 Schaffer M. et al.: Ultramicroscopy 107, 587-97 (2007)
 Scott K.: J. Microsc. 242, 86-93 (2010)
 Möbus G. and Inkson B.J.: Materials Today 10, 18-25 (2007)
 Hatch J.E.: American Society for Metals (1984)
Dr. Armin Zankel
Dipl.-Ing. Herbert Reingruber
Ing. Hartmuth Schröttner
Institute for Electron Microscopy, Graz University of Technology,
and Center for Electron Microscopy, Graz, Austria
Tel.: +43 316 873 8832
Fax: +43 316 811 596