Materials in the form of particulate or engineered objects are relatively easy to characterize in terms of their composition, shape, size, and size distribution, provided they are larger than 1 μm in average dimension. It is often assumed that nanometer-scale materials are just smaller versions of larger-sized materials made with the same precursors and procedures, but the chemical and physical interactions of submicrometer- or nanometer-scale materials are in fact different—offering new materials with novel properties. The very small scale of these new materials makes them very difficult to accurately characterize, and because of their unique properties, it is critical for materials scientists and engineers to conduct confirmatory analysis to verify the properties of their materials.
Fortunately, advances in analytical instrumentation, especially field emission scanning electron microscopy (FESEM), permit characterization of submicrometer-sized materials. Several challenges need to be addressed in order to successfully accomplish a comprehensive, high-quality material characterization. These include: calibrating the scalar measurement tools of the FESEM, understanding the limitations of sample handling and preparation relative to other methods, the inability to use automated methods either in terms of image analysis software or other optical size distribution methods, and using crystal structural analysis to strengthen elemental analysis when small analysis volumes preclude standards for fully quantitative analysis.
Size calibration considerations
The first step toward obtaining high-quality size measurements of nanometer-scale objects is to calibrate the measurement instrument using a traceable measurement standard. Historically, scanning electron microscopes (SEMs) were typically calibrated using items such as transmission electron microscope (TEM) sample grids. As most measurements conducted on SEMs were of objects tens or hundreds of micrometers in dimension, the small variation of production and handling size of TEM sample grids was considered to be inconsequential. With the current generation of FESEM instrumentation and the ability to measure objects that may only be single nanometers in dimension, the historic level of calibration is not sufficient.
Proper quality assurance methods require calibration of instrumentation using a standard reference material from a recognized measurement laboratory, such as the National Institute of Standards and Technology (NIST) in the United States. NIST has, in the past, produced a scanning electron microscopy standard reference material (SRM) identified as SRM 484G. Unfortunately, SRM 484G is out of production, and although the NIST website indicates a new alternative is being worked on, it is not currently available. Additionally, if one has access to a SRM 484G, the rulings with the smallest separation on the SRM are 500 nm. Therefore, in the new age of nanometer-scale materials, even the SRM 484G is only marginally sufficient.
Figure 1 – Lines on the 250-nm pitch ruler of the Metrochip. (Photos by Craig S. Schwandt, Ph.D.)
An alternative option is a microscopy measurement standard that is fabricated by another laboratory that utilizes a key component certified by a measurement laboratory such as NIST. A good example is the Metrochip® Microscope Calibration Target fabricated by Metroboost (Santa Clara, CA). The Metrochip is an etched polysilicon on silicon substrate SRM, where each target is not individually certified; however, the electron beam lithography mask used to create the SRM targets has been certified by NIST. This particular SRM has rulers with 250-nm pitch, which are guaranteed to be less than 5 ppm over the length of the 4.44-mm ruler, i.e., 22 nm (Figure 1). Similar products are available from Electron Microscopy Sciences (Hatfield, PA), which can be purchased with individual certification from Physikalisch-Technische Bundesanstalt (PTB), the German equivalent of NIST. The model 150-2DUTC provides certified pitch uncertainties of ±1.4 nm at the 95% confidence level.
FESEMs and SEMs that are used to provide measurement information should have their X and Y scanning coil rasters calibrated using one of these SRMs. Ideally, one would also like to characterize the size of new sample materials along with a previously well-characterized sample using the same instrument and analysis conditions. An added level of confidence can be achieved if the previously well-characterized sample was characterized using an alternate measurement method.
Figure 2 - Secondary electron images acquired at the same beam conditions. a) A 0.2-μm alumina filter was captured without use of a conductive coating. The image exhibits signs of charging and was acquired with multiple frames to reduce the effects of charging. b) The same filter after it was coated with AuPd for 30 sec. Charging is reduced, although there is some morphology change. c) The filter was coated for an additional 30 sec. Although charging is no longer an issue, clearly the plasma-coating process has altered the finest details of the sample morphology. Scalar tools of image analysis software can easily be utilized with these images to measure the features within them.
A more important concern is changes in sample dimensions produced by sample preparation and selection of electron beam conditions. Nonconductive samples may need to be coated with a conductive film in order to capture photomicrographs suitable for obtaining measurements. It is often best to characterize samples without a conductive coating, if possible. However, the small-diameter high-electron flux beams of FESEM instruments easily produce charging issues on nonconductive samples. In addition, the electron beam also cracks residual hydrocarbons in the sample chamber, causing significant carbon contamination to the sample. Therefore, electron beam conditions must be balanced to minimize these effects, while still providing photomicrographs of sufficient quality with which to make meaningful size measurements.
If alternative beam conditions cannot be found which reduce sample charging, the alternative is to coat them; however, coatings can easily be applied too thickly (e.g., metal coatings), obscuring the true dimensions of the objects or altering the morphology. Experimentation may be required in order to balance coating thickness with electron beam conditions (Figure 2).
Numerous image analysis software packages conduct size and shape distribution analyses in semiautomated, if not fully automated, fashion, provided the objects to be measured are isolated from each other, and there is sufficient gray-scale contrast between the objects and the substrate. Larger particles are usually relatively easy to disperse for analysis, so that the majority of particles do not contact one another and have sufficient visual contrast, and can easily be characterized for size and shape. However, nanometer-scale materials typically cluster into aggregates that are difficult to separate.
Size and shape algorithms do not successfully distinguish the boundaries of individual objects that are clustered, and therefore cannot be used in fully or semiautomatic mode. Although manual measurement of objects may be relatively slow, the measuring process provides the analyst the ability to see shapes and feature information that might otherwise be lost with an automated process (Figure 3).
FESEM analysis provides a complete image of size, shape, and presence of coatings or finer-size occlusions, which may not be seen with light microscopy or liquid-based particle analysis of submicrometer-sized particles. Additionally, the use of backscattered electron imaging in the FESEM provides qualitative compositional information. The number of beam electrons backscattered from the sample is proportional to the average atomic number of the sample; therefore, materials with a higher average atomic number appear brighter than those with a lower average atomic number. Thus, backscattered electron imaging provides information about sample composition and heterogeneity.
Figure 3 – This secondary electron image reveals a range of smaller-sized particles that would not be accurately recognized by methods that do not visualize the sample.
FESEMs commonly have energy or wavelength dispersive X-ray spectrometers attached (EDS or WDS), which can be used to confirm the elemental composition of nanometer-scale objects. The small-diameter beam of high-resolution FESEMs permits placement of the beam in single-point mode over objects as small as 5 nm. The information gathered is semiquantitative primarily because standard reference materials appropriate to the nanotechnology sample are rarely available. However, in most cases, the question is more about whether or not certain elements are constituents of the sample. Additionally, when all major constituents are analyzed, the X-ray analyses can be compared against ideal stoichiometric expectations.
Some FESEMs have electron backscatter diffraction (EBSD) systems attached. These systems determine crystal structure, grain size, and orientation. When combined with elemental data from the same grains, the overall degree of quantification improves the results because the crystal structure constrains the stoichiometry. This is analogous to EDS analysis in the TEM, except that FESEM samples do not require the same extent of sample preparation as TEM samples.
As is the case for imaging nanometer-scale objects, the method of mounting for compositional analysis is important. However, if objects are dispersed on polished, low-atomic-number substrates, the X-ray background is negligible.
Characterization of submicrometer- to nanometer-sized objects using FESEM methods can provide more comprehensive results than those obtained through the use of other object-sizing methods. Results include detailed images of particles, dimensions of irregularly shaped particles, information about occlusions or multicomponent objects, and compositional information, as well as crystal structural information from an FESEM equipped with an EBSD system. Reliable results are produced by systems that have been calibrated with measurement SRMs traceable to a recognized measurement laboratory.
Craig S. Schwandt, Ph.D., is Director of Industrial Services and Senior Research Scientist at McCrone Associates, Inc., and co-teaches several courses at Hooke College of Applied Sciences, LLC, 850 Pasquinelli Dr., Westmont, IL 60559, U.S.A.; tel.: 630-887-7100; e-mail: firstname.lastname@example.org ; www.mccrone.com