Gregory's ServerStupid question related to the whole LK-99 business: why is it so hard to predict properties of materials where you know the internal structure? Is it one of those quantum computational explosion things? Like, if we had large quantum computers could you just quickly say "nope, won't superconduct"?
|
9 comments
@ZachWeinersmith tbh I’ve been wondering how the first question wasn’t “could this be ferromagnetic,” I assume there’s a good reason it wasn’t….  @priyachandwriter @ZachWeinersmith According to the first verifiable attempts to at replication; LK-99 appears to actually contain ferromagnetic components and to _ not _ be superconducting. Guo et al. 2023, "Ferromagnetic half levitation of LK-99-like synthetic samples" - https://arxiv.org/abs/2308.03110 @ZachWeinersmith because scale and boundaries  @ZachWeinersmith In a solid, even if you assume nuclei are fixed, you have a lot of electrons all interacting with each other and that's difficult to compute. @ZachWeinersmith Well, part of it is that you never do know the internal structure, or at least it's harder than it seems. Knowing the chemical composition isn't the same as knowing the internal structure, and even if you know the crystal structure you don't really know how big the crystals are, etc. And structure isn't static - temperature makes things move and vibrate, and how they move and vibrate matters!  @ZachWeinersmith A tool typically applied in solid state physics is Density Functional Theory, which allows to reduce the complexity explosion of many-particle problems to basically a 1 particle simulation. It is relatively easy to show that the Density Functional exists, and that it is equivalent to a many-particle N-dimensional simulation, but nobody has found the exact formula for that Functional yet - so all DFT simulations are estimates... @ZachWeinersmith the question is : how can a composite material produce cooper pairs? Easily asked but difficult to answer also because temperature plays an important role too. A big quantum computer may still not be enough.  @ZachWeinersmith I think it's (at least) two different issues, one of which is 'we don't really know how properties scale down at this level, only observational data and theories that seem to work pretty well' and the other is 'know the internal structure? I want to live in your reality'. This is mostly just me observing what people who know a lot more than I about this stuff seem to be saying, though, my materials science knowledge is...minimal. |
@ZachWeinersmith My husband does quantum mechanical computations of materials, in close collaboration with experimental physicists and chemists. They can tell him things like: there is about X% of element Y in there and we baked it/annealed it/... for n hours. That's a long way from a crystal structure, and yet a good guess for that is needed as a starting point, otherwise the computations get stuck in a local energy minimum.
As far as I can tell, quantum computers wouldn't help much with that.