A helix functions as a grain selector. I think that only the crystal with axial alignment is fastest in general at propagating through a spiral.

Animation of the effect: https://www.youtube.com/watch?v=br9iaeYYxSM

Paper on the development of single crystal casting: http://www.tms.org/superalloys/10.7449/2008/Superalloys_2008...

By the principle of elimination by bottleneck. As the crystals grow into the pigtail, the lagging crystals hit the already-formed leading ones or the mold surface, leaving a primary, single crystal grain to exit the pigtail and grow across the blade proper. This would be almost impossible to happen without the pigtail because multiple crystal grains would grow in parallel across the blade.

Here's a technical explanation.[1] But here's an example of triggering crystal formation in a supersaturated solution.[2] Classic high school chemistry experiment. Dropping a solid crystal into the supersaturated solution causes the whole liquid to turn into a solid in seconds.

That's the basis of single-crystal casting. The solidification process relies on operating very close to the melting point of the material. The idea is to create an alloy mixture that's ready to crystallize, then starting the process by dropping in a crystal and letting the liquid crystallize in a very controlled way. The heated zone is moved through the liquid, as in zone refining (or the solid is pulled from the heated zone, or gas cooling is used) so that the crystal growth takes place in only a narrow layer.

This is decades old, and one of the reasons jet engines last a long time now.

[1] http://www.asminternational.org/documents/10192/6019788/amp1... [2]

Where's [2]? Thanks!

Oops!

https://www.youtube.com/watch?v=2Kud9eVbmRY

Merci!

How does this compare with pulling a silicon boule? Any chip people care to chime in?

I don't fully understand the process, but this paper has an explanation and diagrams: http://www.journalamme.org/papers_vol32_1/32110.pdf