College of Science and Engineering

Case studies

Nuclear Graphite

The sample

Graphite is hugely important for the construction of both historical and modern nuclear reactors [1], acting as a neutron moderator. Due to the harsh environment of the nuclear reactor, the material properties of the graphite can change throughout the service life. Understanding the microstructure evolution (e.g., porosity), is instrumental for the performance assessment of the material, allowing us to design a safer and longer lasting nuclear reactor.

Characterising over multiple length scales

Nuclear graphite has a ‘hierarchical microstructure’, meaning that we have features ranging from hundreds of micrometres right down to tens of nanometres. This presents a challenge in terms of analysing a large volume (>10mm3) to understand the context and broader structure but also enabling high resolution (<10nm3) in key areas. The combination of MicroCT and LaserFIB (the Hercules solution) can circumvent this challenge.

MicroCT is used to determine the microstructure at the millimetre scale on a bulk sample while also identifying important regions of interest for further analysis. The structure of dense grains and intergranular pores can be easily measured however, the measured porous volume from MicroCT alone gives a value of 12.8%, which clearly doesn’t match the value of 22.7% as reported in [2,3].

This is because many nanometre scaled porosities in the graphite cannot be resolved by MicroCT. Therefore, a higher resolution analysis (i.e., X-ray Microscopy (XRM)) is required to characterise the sample fully. However, a much smaller sample volume must be prepared to allow this analysis, such as the free-standing pillar.

How Hercules solves the problem

Site specific FIB sample preparation would be your go-to technique to prepare this kind of structure, perhaps a Plasma FIB given its large size. Unfortunately, the hardness and anisotropy of the graphite make it very hard to mill either with plasma or Ga+ FIB at any practical rate. The problem is even more complex if the region of interest is deeply buried in the sample.

Femtosecond Laser, on the other hand, is independent of the material hardness and therefore can mill far more rapidly, preparing this pillar in only 20 minutes. What’s more, due to the short pulses, the laser milling does not suffer from a large heat affected zone (HAZ), meaning the material near the cut will not be affected by the milling.

Correlative workflow

The data obtained in the MicroCT can be easily passed to the LaserFIB, allowing the regions of interest to be rapidly identified and prepared with the femtosecond laser, providing quicker throughput and more confidence in your sample preparation. The XRM data reveals that on top of the intergranular porosity, there is a significant amount of intragranular porosity. Extrapolating this to the remainder of the intragranular region identified in MicroCT, the porosity value is calculated as 20.5%, far closer to the expected value [2,3]. Without this multimode approach, offered by the Hercules solution, it wouldn’t be possible to measure the porosity and also understand the microstructure.

  • [1] J. Kane et al. Journal of Nuclear Materials, vol. 415, 2, pp189-197, 2011
  • [2] T.Oku, M.Eto and S. Ishiyama, Journal of Nuclear Materials, vol. 172, no. 1, pp 77-84, 1990
  • [3] T. Burchell, Carbon, vol 34, no.3, pp 297-316, 1996


Limitation of 2D

The diagram below the cross section of several 3D-shapes. If you presented just the cross section and were asked to guess what shape you could guess any of the three and depending on where we sectioned the shapes, we could also easily be confused as to the scale of the 3D volume.

Without context, a 2D cross-section can be misleading. This is an issue with traditional techniques in metallography and geology as microstructures can look very different when viewed from different directions; an example is from super-duplex stainless steel is illustrated here. Typically, a material is cut, polished and then only a section of the material is analysed, showing only part of the entire picture. Although this might be done in several orthogonal directions (as below) to provide more information this is not only time intensive, it also requires the sample to be destroyed, meaning no further testing can occur. Hercules offers a different option though, through a technique called Diffraction Contrast Tomography (DCT).

Why is Diffraction Contrast Tomography different from MicroCT?

Like medical x-rays, MicroCT is based on contrast caused by different materials absorbing different percentages of incident x-rays (Absorption Contrast Tomography) reducing the transmitted intensity. By rotating the sample, we get multiple projections allowing for a 3D-reconstruction of the sample.

This method alone isn’t very applicable if we wish to determine the microstructure of a single composition material.

Diffraction contrast tomography (DCT) is a non-destructive technique for the 3D characterisation of polycrystals containing up to a few thousand grains. While it is an established technique at synchrotrons ,  it is still an emerging technique for lab based x-ray sources which only becomes available,  as part of the ZEISS Hercules solution. When the sample rotates, some crystalline phases within the sample will fulfil the Bragg condition, meaning that some x-rays will be diffracted as well as absorbed. This gives both a dark spot in our transmitted image, indicating the real space position of the diffracting crystal, but also a diffraction pattern. By measuring the diffraction patterns at many angles and combining this with knowledge of the material being analysed, information about the orientation of the crystal can be  calculated  as well as the grain shape and size.

Why use DCT rather than Electron Back Scatter Diffraction (EBSD)?

EBSD, while fantastic at giving us fine detail about orientation and engineering strain, relies on cross sections that are time consuming to prepare and only obtains part of the available information about the sample. While we could use 3D-EBSD to reconstruct a 3D-volume, the scale at which this is practical is much smaller than DCT and more importantly is destructive.

Moreover, as DCT is non-destructive, we can take more than just a single snapshot. A tensile (need to check with Zeiss, do they have a better example that shows sample at different stages?) specimen of Al-4wt%Cu is illustrated here with the measurement done with DCT. Not only can we see the grain orientation, shown in the Inverse Pole Figure colour (as in EBSD), but also we can take snapshots showing us the evolution of the grain structure. The X-ray setup at Hercules can apply loads up to 5kN as well as heating or cooling the sample while it is being scanned.

Correlating DCT and EBSD

Of course, if extra detail is required, Hercules also has further capability to take a region of interest identified in the DCT scan and prepare this for further EBSD studies, giving more localised information on orientation and strain within the sample. This multiscale approach provides both the wider context of the sample through its life time as well as the fine detail at the critical moments.

  • [1]Modified from TF. Santos et al. Materials Research. 2016; 19(1): 117-131

Back to top