TEM Imaging and TKD Mapping
Interaction of Nanoparticles Incorporated in a Nickel Matrix
- Fig. 1: Effect of particle type on Ni matrix. TEM BF (top, particles in bright contrast), TKD IQ maps (middle, HAGBs, Sigma3 and Sigma9 twin boundaries are white, red and mint, respectively), TKD IPF maps (bottom, grain boundaries are black, particles are black).
- Fig. 2: Particle incorporation in detail. EDS maps of Al, Ti and Si distribution, respectively (top and bottom, particles are shown as element intensities in grey scale, particle size is shown schematically by white circles) with Ni grain boundary overlay (HAGBs, grain and twin boundaries are black, mint and red, respectively), TEM BF (middle, particles in bright contrast).
- Table 1: Influence of particle type on nickel matrix grain size.
The potential of SEM-based transmission Kikuchi diffraction is shown by the example of particle-reinforced nickel matrix composite layers. The effect of different nanoparticle types on the matrix microstructure is evidenced by grain size distribution and grain boundary character analysis. Advantages and limitations are discussed. Automated orientation mapping on electron transparent samples can be carried out in the same way as for conventional EBSD, but with spatial resolution improvement.
Particle-reinforced metal matrix composites (MMCs) are widely applied in engineering. Electro-codeposition is a typical technique for producing MMC coatings. Improvement of material properties depends on type, size and content of the reinforcement. Reducing size down to nano-scale is particularly promising because matrix properties are highly susceptible in the vicinity of well-dispersed nanoparticles. Key issues for developing tailored materials are grain size distribution and boundary characters. A common approach to evaluate MMC microstructures is scanning electron microscopy (SEM), combined with energy-dispersive X-ray spectroscopy (EDS) and electron backscatter diffraction (EBSD) . EBSD provides the crystal orientation located within the interaction volume of the incident electron beam. Maps are generated by scanning the beam for a spatial description of grains, boundary characters and microtexture. The variety of statistical tools suitable for quantification of features is advantageous over transmission electron microscopy (TEM); disadvantageous is the limited resolution. The interaction volume is about 100x30x5 nm3 with the anisotropy caused by sample tilt. SEM-based diffraction on electron transparent samples referred to as transmission Kikuchi diffraction (TKD) , also known as t-EBSD  allows improved resolution. The specimens are mounted in the SEM chamber similarly as bulk samples but horizontally aligned at the level of the EBSD detector’s top. No tilt correction and dynamic focus is required for orientation mapping. A tailored sample holder prevents transmitted electrons to strike the sample stage and generate detrimental background. Patterns are projected from the bottom side.
Thus, spatial resolution is primarily limited by beam expansion in the TEM lamella (~5 nm in 150 nm thick nickel). TKD results on nanoparticle reinforced nickel coatings illustrate the prospects of the method.
Materials and Methods
MMC films were electro-codeposited under typical conditions in a Watts Ni plating bath  with similar particle load (20 g/l) and particle size (~50 nm): Al2O3 (Micropolish B, Buehler), TiO2 and b-SiC (both Nanostructured & Amorphous Materials Inc), respectively. TEM specimens were exfoliated from polished inox substrates, grinded and Ar ion polished (Gatan PECS). The TEM (Hitachi H8100) was operated at 200 kV for bright-field (BF) imaging. Pattern quality (IQ) and grain orientation (IPF) maps were obtained in a SEM (Zeiss NEON 40EsB) equipped with an EBSD detector (EDAX DigiView III) using TKD of 25 kV electrons (beam current 20 nA). Data acquisition and analysis (EDAX TSL OIM 5.31) was done using a sampling step size of 20 nm. The coherent domain size in the matrix was verified by XRD (Bruker AXS D5000) with Rietveld refinement.
Effect of Nanoparticles on the Microstructure of Nickel Composite Films
The influence of particle incorporation reveals TEM BF imaging (fig. 1, 2). The matrix grain size depends on the electrical behavior and dispersibility of the particles in the electrolyte. Due to best dispersion, Al2O3 particles provide the highest number of nucleation sites triggering the smallest nickel grain size. Agglomerated TiO2 particles generate a lower number of nucleation sites. Consequently, a certain number of matrix grains grew larger which is even more pronounced in the case of SiC particles with semi-conducting behavior. The grain boundary character according to the coincidence site lattice (CSL) model was recorded by TKD to quantify the size distribution. Analyzed TKD data sets comprise larger sample areas than shown in figure 1, where details of the composed maps are presented. Table 1 summarizes statistically verified mean grain sizes. A rather low spread in the Ni- Al2O3 composite contrasts the bimodal distribution caused by the two other particle types. Mainly the Ni-TiO2 composite evidences two distribution maxima at 130 nm and 500 nm, i.e. a population of small grains around particle agglomerates and large grains in between. Grain growth in the vicinity of particle agglomerates is detailed in figure 2. The particles within the boundary network are traced by EDS distribution maps of Al, Ti and Si, respectively. It should be emphasized that the EDS signal source comprises the entire sample thickness while the TKD signal arises only from the exit plane on sample bottom.
Nickel is highly prone to twin formation due to its low stacking fault energy. Twins are associated with favorable properties since energies of twin boundaries (fcc S3 and S9 according to the CSL concept) are lower than that of random high angle grain boundaries (HAGB). Hence, developing a stable energy configuration is associated with grain growth, rotation or twinning . TKD data allow the separation between HAGBs and twin boundaries in the composite matrix (fig. 1, middle row). Grain sizes were determined basing on 15° tolerance angle HAGBs, and subsequently with exclusion of recrystallization twins. Indicated as boundary length fractions, twin proportions evidence a correlation with the type of incorporated particles (table 1). The values are larger than ~0.3 as commonly expected for nickel. For comparison, the coherent domain size of the nickel matrix was measured by XRD. Results imply that the TKD analysis did not reveal all twins.
Prospects and Limitations of TKD
Further improvement of the lateral resolution is feasible by lowering the sampling steps down to some nanometers, however, with drawbacks of sample drift and contamination. Additionally, adjusting accelerating voltage, working distance and sample tilt [5-7] lowers the resolution limit. Tilting (~20°) in opposite direction to the usual EBSD array can be beneficial at a low working distance . However, an adequate sample thickness proved to be the most crucial factor. Patterns become either noisy or weak with reverse contrast if the specimen is either too thin or too thick [5, 6]. Consequently, resolution also depends on specimen density and atomic number as clearly evidenced by Monte Carlo simulation . Another aspect is the pattern quality at grain boundaries. Contributions from neighboring lattices and a local higher dislocation density result in weaker patterns. HAGBs appear as delineated by darker pixels in pattern quality maps of EBSD and TKD data, respectively; and complicate band detection and indexing . A typical example is the TKD analysis of the fine-grained Ni-Al2O3 composite in this report. Likewise, a TKD analysis of a Ni/P-diamond composite with matrix grains below 50 nm evidenced only several individual larger grains with appropriate confidence . In conclusion, the main issue of successful TKD analysis is the preparation of TEM lamellae of suitable homogeneous thickness without bending. On that base, automated orientation mapping with a spatial resolution improvement of up to one order of magnitude is feasible. TKD uses the existing EBSD hardware and the post-processing tools and allows simultaneous EDS collection. The results are comparable to those achieved using automated diffraction techniques in the TEM, but with the advantages and flexibility accustomed from a SEM.
The authors thank Anne Schulze for the careful preparation of TEM specimens and Thomas Mehner for doing the XRD analysis.
 Thomas Lampke, Bernhard Wielage, Dagmar Dietrich, and Anette Leopold: Details of crystalline growth in co-deposited electroplated nickel films with hard (nano) particles, Appl. Surf. Sci. 253, 2399-2408 (2006)
 Patrick W. Trimby: Orientation mapping of nanostructured materials using transmission Kikuchi diffraction in the scanning electron microscope, Ultramicroscopy 120, 16-24 (2012)
 Robert R. Keller, and Roy H. Geiss: Transmission EBSD from 10 nm domains in a scanning electron microscope, J. Microsc. 245 (3), 245–251 (2012)
Valerie Randle, Paulo R. Rios, and Yan Hu: Grain growth and twinning in nickel, Scripta Mater. 58, 130-133 (2008)
 Seiichi Suzuki: Features of transmission EBSD and its application, JOM 65, 1254–1263 (2013)
 Seiichi Suzuki: Evaluation of transmission EBSD method and its application to observation of microstructures of metals, J. Japan. Inst. Met. Mater. 77 (7), 268-275 (2013)
 Rik van Bremen, Diego Ribas Gomes, Leo T.H. deJeer, Václav Ocelík, and Jeff Th.M. deHosson: On the optimum resolution of transmission-electron backscattered diffraction (t-EBSD), Ultramicroscopy 160, 256–264 (2016)
 Stuart I. Wright, Matthew M. Nowell, René de Kloe, and Lisa Chan: Orientation precision of electron backscatter diffraction measurements near grain boundaries, Microsc. Microanal. 20(3), 852-63 (2014)
 Dagmar Dietrich, Amir Sadeghi, Andrea Sendzik, Anne Schulze, Thomas Mehner, Harry Podlesak, Daniela Nickel, Ingolf Scharf, and Thomas Lampke: A new insight into the phosphorus distribution of nanocrystalline Ni-Ni3P-diamond composites, J. Electrochem. Plating Techn.: DOI: 10.12850/ISSN2196-0267.JEPT1645 (2013)