Josh Kacher,  Post doc – UC Berkeley

One of the perks of having done my undergraduate degree at Brigham Young University was that I got to know the people at EDAX pretty well. Over the years since I’ve moved out of Utah, I’ve been able to collaborate with them on a number of different projects, which has been a great benefit to my work. Recently, I had the opportunity to spend a week with them working on a particularly challenging material, rhenium.

To give some background of the material since it’s one that the majority of people aren’t familiar with, rhenium is a high density refractory metal with a HCP crystal lattice. It maintains good ductility over a wide range of temperatures, making it an attractive material for applications in extreme environments. Unfortunately, it’s expensive, rare, and we’re running out of it. My work is focused on obtaining a fundamental understanding of the material in hopes of being able to dilute or replace it in some of its current applications. Unfortunately for me, it also tends to have a difficult to characterize microstructure due to difficulties in sample preparation and multiscale twinning mechanisms active during deformation.

The samples I looked at were slices from a compressed rhenium pillar (4 mm tall, 3 mm diameter). Since any mechanical grinding tends to induce twinning in the material, all sample preparation was done using a twin-jet electropolishing setup using a perchloric solution as the electrolyte. Differential etching rates dependent on crystal orientations resulted in quite a bit of topography on the polished sample. To get an idea of the feature sizes I was interested in, we first made use of the new PRIAS imaging feature, which made for some fun 3D-looking maps of the surface. Below are a few of the PRIAS images, along with the region from the EBSD pattern used to form the image:

The ridges in the images mark where the grain or twin boundaries are located. As can be seen, the microstructure consists of both larger twins, on the scale of hundreds of nanometers, as well as small crisscrossing twins, on the scale of tens of nanometers.

With this information, we were ready to get started doing full EBSD scans. We settled on a 100 nm step size over a 70×70 micron area. Although we knew some of the twins would be missed, a step size small enough to capture them all would have been prohibitively slow.

As can be seen, virtually every grain in the scan area had some level of twinning, some even developed multiple twinning systems. All the twins were identified as {111}<100> type and had an average twin width of ~700 nm. At this point, I already had most of the information I was looking for, but something interesting stood out to me in the orientation map. There seems to be many locations where the twins transmitted directly across the boundary. Although I had seen similar behavior in TEM characterization, it always involved nanotwins, not twins at the scale seen in this scan. To try to discover what was going on, six different areas were chosen for more in depth analysis.

Of these six interactions, four show twin transmission and two show the twins stopped at the boundary. The easiest factor to take into account was the misorientation across the boundary. It turns out that this was also the best as those boundaries with misorientation angles below 23° were found to allow twin transmission while those with higher misorientation angles acted as barriers. The next thing to check was the alignment of the twin planes and shear vectors associated with the twinning on either side of the boundary. This was done using what’s known as the m’ factor, which is calculated by:

m’=cos θ cos Ψ

where θ  is the angle between twin plane normals on either side of the boundary and ψ is the angle between shear vectors. Calculating m’ for all six twinning events gave:

Twin system	1	2*		3	4	5	6
Boundary misorientation	15°	23°		72°	13°	18°	36°
m’	0.94	0.91	0.91	0.19	0.95	0.95	0.97
Transmits?	Y	Y	Y	N	Y	Y	N

* Two twinning systems interacting at the boundary

The maximum m’ value possible is 1, so all in all of the cases, only twins with a high m’ value developed at the boundary. So, not only was it possible to predict whether a twin would transmit across a boundary, we could also predict which twinning system would nucleate if given the incoming twin and boundary angle. Twin 6 presents a particularly interesting case. Even though the interaction has a high m’ factor, or in other words, has a closely aligned available twin plane on the other side of the boundary, the twin still did not transmit. That is, the misorientation angle seems to be a better indication of the barrier strength of the boundary than the m’ factor.

To sum up, what began as a relatively routine effort to characterize the deformed microstructure of rhenium turned out to be an interesting study on twin/grain boundary interactions in HCP materials. There’s still significant work left to do to really understand what’s going on, but the EBSD data provides a great starting point. A big thanks to the people at EDAX for letting me come spend such an enjoyable and productive week with them!