I am not as hesitating as saabluster.
This is one of the very very view areas that I will definitively say: It will NEVER happen.
Seriously, the idea of using a thing rod (<1cm) for making cicular leds is ridiculous on several levels.
1. Each led size would require a whole substrate handling toolchain. Something like that can cost dozends of millions. (Not to mention that handling tiny weeny wavers would suck in general)
2. Serialization of fabrication: Each step would need to be done for each led indiviudally, instead of 100s of leds in one go.
1+2 would mean that the fabrication would be extremely expensive (even in mass production environment, 100s of $ per device. The idea of it being cheaper than losing 20-30% of the substrate area due to cutting away is ludicrious.
But thats not really the KO-criterium. What is is:
3. THere is a reason you never use the outermost regions of substrates. Any bulbs being grown drop in quality to the edges. Making very thin rods would only increase the issues with the crystal quality. The whole idea of "not needing to mill it into a round shape" depends on using the whole substrate up to the outermost um, which will certainly end up in failure (as edge effects will also not help with any of the more esoteric steps of fabrication). To even in the best case, you would need to trip the outermost edges anyway - which totally kills the whole "no milling/cutting" concept.
Any led made that way would certainly have MUCH worse performance characteristics than a normally created one, so even in a university setting I could only imagine it as a proof of non-viability
But no pressure Micheal :naughty:
