I read it as: for the same calculation, what would have previously taken 5,000 CPU cores * hours ran on a supercomputer can now be run in 2 milliseconds on a desktop. Or about roughly 10 orders of magnitude faster
This is an interesting result. It makes me wonder if you could grow a silicon ingot in this way. If so it would cut the cost of silicon solar cells significantly (a big chunk of their cost is the cost of the ingot, the ingot cost is a function of time to produce, producing them quickly would get more ingots per unit time from a given reactor).
Another crystal structure that would be useful would be sapphire for things like sapphire screens and other covers.
> This is an interesting result. It makes me wonder if you could grow a silicon ingot in this way.
No really. The way we grow silicon crystals is already very efficient and would not be improved by this sort of thing. We grow them by pulling a seed (in carefully controlled conditions, with a bit of twisting and other tweaks, but still), adding dissolutions steps would waste a lot of time. Besides, the defects they get rid of by dissolution (say, a magnesium atom on a calcium site, which we call antisite defects) do not exist in silicon, in which all atoms are identical.
> Another crystal structure that would be useful would be sapphire for things like sapphire screens and other covers.
Again, sapphire is something that is quite easy to grow. And like for silicon, it does not have any antisite defects. One of the limitations with growing sapphire is its purity, and it would not be improved by additional dissolution steps.
I am not saying that similar steps could not be used to help produce industrially relevant crystals, just that silicon and sapphire are very unlikely to be examples of that. This is a way to grow in a lab in weeks a crystal that would take millions of years to grow in nature. It would not really help with crystals that grow already very efficiently, and whose growth is controlled by other mechanisms.
i admit to being pretty ignorant about thermochemistry
is there a relevant difference between crystals approaching equilibrium with a melt and crystals approaching equilibrium with a solution? because i had the impression that the mechanisms were the same and the line between these processes was actually a little bit fuzzy, what with things like hydrate crystals and aquo complexes
(i think you can't melt dolomite tho, at least near atmospheric pressure)
> The way we grow silicon crystals is already very efficient and would not be improved by this sort of thing. We grow them by pulling a seed
One interesting remaining unsolved problem is highly homogenous doping. Such silicon is needed for very high-current and high-voltage switching (like invertors on large wind turbines or HVDC power lines).
Creating it is so hard that we have to use FREAKING NUKES.
Seriously, a chunk of pure silicon is irradiated by neutrons in a reactor, transmuting some silicon atoms into phosphorus. It's apparently the only way to do it, which has always seemed wild to me.
Saw some beautiful dolomite bluffs along the Katy Trail in Missouri earlier this year. (Until I saw a sign talking about them I had assumed they were limestone.)
These are along the Missouri River — so no doubt a historically flooding/drying spot that they mentioned in the article.
Really glad to see this work and result. My PhD dissertation 25 years ago (pub cited in group's Science paper) focused on laboratory synthesis of dolomite at lower temperatures (less than hydrothermal), but since then focused on virtual (kinetic Monte Carlo simulations) experiments as a means of understanding crystal growth. Really happy to see this fundamental progress.
This could potentially spell a long-term sustainability of most climbing crags in the world, which are on some form of limestone/dolimite or similar.
What happens is that thousands of climbing shoes with tiny bits of dirt eventually polish many holds, altering the routes and always for the worse since those really polished parts that everybody has to use become unreliable and everybody avoids them like plague.
I was thinking of spraying a thin film of some acid or similar to make surface rough again, but this eventually wears down the rock (although it could probably last few centuries till you run out of it).
Anybody knows about anything actually being done, instead of me just fantasizing?
TIL that the mineral dolomite is not named after the Dolomite mountains in Italy, but the other way round, the mineral being named after the geologist who discovered it (https://en.wikipedia.org/wiki/D%C3%A9odat_Gratet_de_Dolomieu). Which of course begs the question of how the mountains were called before dolomite was discovered, apparently "pale mountains" ("monti pallidi").
42 comments
[ 3.2 ms ] story [ 107 ms ] threadIs this phrase equivalent to "Each atomic step would take 5,000 hours on a desktop. Now, it takes 2 CPU milliseconds on a supercomputer."? ^^
That poor dog
Another crystal structure that would be useful would be sapphire for things like sapphire screens and other covers.
No really. The way we grow silicon crystals is already very efficient and would not be improved by this sort of thing. We grow them by pulling a seed (in carefully controlled conditions, with a bit of twisting and other tweaks, but still), adding dissolutions steps would waste a lot of time. Besides, the defects they get rid of by dissolution (say, a magnesium atom on a calcium site, which we call antisite defects) do not exist in silicon, in which all atoms are identical.
> Another crystal structure that would be useful would be sapphire for things like sapphire screens and other covers.
Again, sapphire is something that is quite easy to grow. And like for silicon, it does not have any antisite defects. One of the limitations with growing sapphire is its purity, and it would not be improved by additional dissolution steps.
I am not saying that similar steps could not be used to help produce industrially relevant crystals, just that silicon and sapphire are very unlikely to be examples of that. This is a way to grow in a lab in weeks a crystal that would take millions of years to grow in nature. It would not really help with crystals that grow already very efficiently, and whose growth is controlled by other mechanisms.
is there a relevant difference between crystals approaching equilibrium with a melt and crystals approaching equilibrium with a solution? because i had the impression that the mechanisms were the same and the line between these processes was actually a little bit fuzzy, what with things like hydrate crystals and aquo complexes
(i think you can't melt dolomite tho, at least near atmospheric pressure)
One interesting remaining unsolved problem is highly homogenous doping. Such silicon is needed for very high-current and high-voltage switching (like invertors on large wind turbines or HVDC power lines).
Creating it is so hard that we have to use FREAKING NUKES.
Seriously, a chunk of pure silicon is irradiated by neutrons in a reactor, transmuting some silicon atoms into phosphorus. It's apparently the only way to do it, which has always seemed wild to me.
https://en.m.wikipedia.org/wiki/Dolomite_(mineral)
These are along the Missouri River — so no doubt a historically flooding/drying spot that they mentioned in the article.
https://mse.engin.umich.edu/people/whsun
https://scholar.google.com/citations?user=PD6xwoEAAAAJ&hl=en
-Prof. H. Farnsworth
What happens is that thousands of climbing shoes with tiny bits of dirt eventually polish many holds, altering the routes and always for the worse since those really polished parts that everybody has to use become unreliable and everybody avoids them like plague.
I was thinking of spraying a thin film of some acid or similar to make surface rough again, but this eventually wears down the rock (although it could probably last few centuries till you run out of it).
Anybody knows about anything actually being done, instead of me just fantasizing?
Obligatory snoop dog video lol
...and here I stopped reading. Thw White Cliffs of Dover are famously chalk, and not dolomite.