Rapid serial sectioning for 3D characterization of materials

Sectioning a material sufficiently large should provide an accurate assessment of the extremes in the size of any constituent to determine how sensitive the material will be to fatigue-induced failure.

1310ilsrapidserialf1

By William Clark, Ph.D., Larry Walker, Ph.D., and Tissa Gunaratne, Ph.D., Clark-MXR Inc.

Destructive serial sectioning is a technique used to characterize the structure and composition of opaque materials in three dimensions (3D). Two-dimensional layers of a target material are successively exposed and imaged to provide information on the material’s structure and composition. The process is repeated hundreds of times to create the data set that is then assembled to provide a profile of the material in 3D. Sometimes referred to as 3D tomography, serial sectioning is about the only way to obtain information on the composition and structure of opaque materials.

There are several techniques used to characterize materials in 3D using serial sectioning:

a) Focused ion beam milling (FIB) followed by scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), and electron backscatter diffraction (EBSD) requiring two charged particle beams,

b) Mechanical systems that expose each layer by etching, grinding, polishing, or dicing, and

c) Atomic probe systems using atomic force microscopy.

In comparison, only one type of particle beam (photons) is needed to both section and characterize materials with ultrafast pulses of light. Other advantages are:

1. It can be done in air, that is, it does not require a vacuum.

2. It has the ability to section materials at a rate that is five orders of magnitude higher than other high-resolution ablative tools.

3. It has the ability to section any material, even materials as hard as diamond and as soft as polymer. Mechanical techniques are limited to materials that can be etched, ground, sliced, or polished.

4. It can expose successive layers while minimizing the physical and/or chemical properties of the native structure.

5. It has the ability to precisely control the amount of material ablated, thereby achieving feature size and resolution on a scale that can be as small as tens of nanometers (FIGURE 1).

1310ilsrapidserialf1
FIGURE 1. Resolution capabilities of several tomographic techniques for analyzing opaque materials [1].


6. It can perform layer-by-layer in situ analysis of exposed layers using linear and nonlinear spectroscopic techniques. For example, Rayleigh scattering and photoacoustic imaging for structural analysis, laser-induced breakdown spectroscopy (LIBS) [2] provide elemental analysis, coherent anti-Stokes Raman spectroscopy (CARS) and/or surface-specific vibrational sum frequency generation spectroscopy (Vib-SFG) for molecular species identification, or pump-probe spectroscopies such as transient absorption to study material dynamics. These are all within the capabilities of the Clark-MXR model ShapeShifter nonlinear spectrometer using the same pump laser [2].

7. Femtosecond serial sectioning can do all this without staining the sample (since photons do not embed in the material) without needing a vacuum (since these processes can be done in a normal atmospheric pressure) and with automatic image registration.

Why is 3D material characterization important? The following is a quote from Ref. [3]:
"For many materials, the rarely occurring features located at the tails of the size distribution (in inhomogeneous materials) govern properties such as fatigue life or shear strength, and it is therefore critical to section large volumes of material to gain access to the microstructural statistics.”

In other words, it is important to section a volume of a material sufficiently large to provide an accurate assessment of the extremes in the size of any constituent to determine how sensitive the material will be to fatigue-induced failure. It is also important to do this with sufficient resolution to obtain accurate information on the size of each feature. Thus, the thickness of each layer must be substantially smaller than the mean size of a constituent of interest. As an example, the authors characterized a Ni-based superalloy used in the manufacture of jet aircraft engines. Clearly a failure of an engine in flight due to a stress-induced crack would be catastrophic. This group used the same Clark-MXR model CPA-Series that is used in the model UMW-2110 ultrafast micromachining workstation.

FIGURE 2 illustrates the power of using a Clark-MXR model UMW-Series ultrafast micromachining workstation to create a 3D image of an EEPROM IC chip. Each 2D layer was exposed by machining a 0.2-micron-thick layer of material followed by photographic imaging, as illustrated in Fig. 1 of Ref. [3].

1310ilsrapidserialf2

FIGURE 2. 3D image of an EEPROM generated by micromachining with a Clark-MXR Model UMW-2110 ultrafast micromachining workstation.

The dimension of each 2D layer was 300 microns × 500 microns × 0.2 microns thick, and 200 layers were machined and imaged in slightly less than 200 minutes. Five of these 200 layers are shown in FIGURE 2 to illustrate changes in the structure of a 2D layer. It would have taken more than 40 hours to obtain a similar size data set using a charged particle beam technique that we obtained with the UMW-2110 in slightly over 3 hours!

In summary, the model UMW ultrafast micromachining workstation can not only micromachine materials. With only minor modification, it can also do 3D characterization of materials at rates that far exceed what can be achieved with other tools on the market today.

References
1. Illustration courtesy of Tresa M. Pollock and McLean P. Echlin.
2. www.cmxr/Products/LaserSystems/Spectrometers.html
3. McLean P. Echlin, Naji S. Husseini, John A. Nees, and Tresa M. Pollock, "A new femtosecond laser-based tomography technique for multiphase materials," Adv. Mater. 23, 2339-2342 (2011) DOI: 10.1002/adma.201003600.


More in Micromachining