Introduction
Electron Backscatter Diffraction (EBSD) is an established microanalysis tool for characterizing a material’s crystallographic microstructure. In practice, EBSD typically requires a significant sample tilt value (≈ 70°) to improve the yield of diffracted backscattered electrons towards the EBSD detector. The sample geometry is often not ideal for other characterization techniques within the Scanning Electron Microscope (SEM). The notable exception to this is the Energy Dispersive Spectroscopy (EDS) detector, which can be positioned to efficiently collect X-ray data from both a flat and a highly tilted sample. However, the techniques are more limited by a tilted sample geometry. Wavelength Dispersive Spectrometry (WDS) and Cathodoluminescence (CL) are two examples of this. These techniques can provide information to complement the EBSD (and EDS) characterization but require a non-tilted sample surface for data acquisition. Correlative microscopy is an excellent approach for combining these different techniques for more comprehensive characterization. In correlative microscopy, datasets are collected from the same area of interest on a sample using each characterization modality. The data is then spatially correlated so that each representative pixel on the sampling grid from the area of interest has data values from each analytical technique of interest. To facilitate this correlative microscopy, tools for alignment and analysis are available in the OIM Analysis™ Software.
Results and Discussion
EBSD and CL data were collected from a cadmium telluride (CdTe) solar cell material to demonstrate this correlative analysis capability. The CdTe films were grown using radio frequency magnetron sputtering. After deposition, the film received a 30-minute CdCl2 treatment in air at 387 °C to passivate the grain boundary structure and grow the grains within the microstructure. Due to the film’s roughness, a focused ion beam was used at a 1° glancing angle to mill a flat region on the surface for analysis. The EBSD data was collected using a Velocity™ Super EBSD Detector, operating at 20 kV and 2,000 indexed points per second at 20 kV beam energy and 6.4 nA beam current at 70° sample tilt (note the required 68.5° stage tilt + 1.5° from the FIB glancing angle into the surface). The CL data was collected from the same area using a Gatan Monarc Pro CL System. A hyperspectral map was collected from 700 nm to 1100 nm with a 0.2 second dwell time. DigitalMicrograph® was used to visualize the hyperspectral data. A primary Gaussian peak was identified at 807 nm, and a greyscale image was generated using this wavelength along with the Secondary Electron (SE) image collected with the CL data.
A correlative microscopy tool within OIM Analysis associates the images from the complementary techniques with the EBSD data. In this example, the 807 nm CL and SE images were selected for correlation. The intensity range of this data can be associated with each image. Spatial correlation between these maps and the EBSD data is achieved using a Quadratic Bivariate correlation method. This method requires at least nine features to be identified in both the correlated and EBSD data. In this case, the SE maps collected during both acquisitions were used, as the structure contained voids that were easily identifiable in both images. This approach allows the correlated data to be mapped to the EBSD data and does not require the same sampling step size for both techniques. Figure 1 illustrates the correlated SE images from both the EBSD and CL acquisitions, showing the correlative alignment.
Figure 1. Correlated SE images from a) EBSD and b) CL acquisitions.
Figure 2 shows the EBSD grain map, where grains are determined from the measured orientations and then randomly colored to show grain morphology. A grayscale image of the correlated CL map of the
807 nm wavelength emission, was generated from the correlated values in OIM Analysis is shown in Figure 3. The CL detector detected this map of the intensity of the light at this wavelength. The light was generated by recombining charge carriers within the CdTe material and expected to be correlated to the primary band gap of the material.
Figure 2. EBSD grain map, where grains are determined from the measured orientations and then randomly colored to show grain morphology.
Figure 3. Grayscale image of the correlated CL map of the 807 nm wavelength emission, generated from the correlated values in OIM Analysis.
Figure 4. The EBSD and CL data are visualized together in this EBSD Image Quality map with grayscale contrast. The CL intensity data at 807 nm is colored using a white-to-red coloring scheme across the intensity distribution.
Figure 4 shows how the EBSD and CL data can be visualized together. This image illustrates the EBSD Image Quality map with grayscale contrast while the CL intensity data at 807 nm is colored using a white-to-red coloring scheme across the intensity distribution. The correlation allows for the analysis of relationships between the EBSD and CL data. For example, the highlighting tool within OIM Analysis can be used to measure the CL intensity across different grain boundary types in the microstructure. In this example, random high-angle grain boundaries had lower CL signal levels than twin boundaries within the CdTe. These results suggest that twin boundaries are beneficial to CdTe conversion efficiency by reducing charge recombination sites within the material. CL signal can also be correlated with crystallographic orientation. Figure 5 shows a scalar texture Inverse Pole Figure (IPF) plot of the CL intensity values as a function of orientation. This map shows some relationship between the crystal orientation and CL signal intensity, with the (001) orientation relative to the surface-normal direction having higher CL intensities.
Figure 5. A scalar texture IPF plot of the CL intensity values as a function of orientation.
Conclusion
This application shows how meaningful data can be extracted from the correlation of CL and EBSD data and, in broader terms, the usefulness of correlative microscopy in general. CL is particularly interesting, as solar cell materials’ properties depend on compositional uniformity and defect concentrations, which can be measured in detail with CL. EBSD provides crystallographic microstructure characterization to complement these measurements. The correlative features within OIM Analysis offer powerful tools to align, visualize, and measure the relationships between these analytical techniques, and provide new insight into material performance.