Melting silicon 'in reverse' can help purify it, result in cheaper electronics
By Vlad Savov 
posted August 2nd 2010 10:22AM
posted August 2nd 2010 10:22AM
Just our favorite combination of news: a mind-bending innovation that can have a very practical impact on our daily tech consumption. MIT scientists have found that silicon -- when combined in the right dosage with other metals -- can actually be made to melt by reducing its temperature. Typically, you'd require 1,414 degrees of Celsius heat to liquidize solid silicon, but the intermixed variant discussed here need only reach 900 degrees before its slow cooling process starts turning it gooey. The great advantage to this discovery is that because the impurities tend to separate off into the liquid part, there's now a practicable way to filter them out, meaning that things like solar cells won't require the same high grade of silicon purity for their construction -- which in turn might lead to us being able to afford them one day. Of course, that's getting way too far ahead of ourselves, as the research is still ongoing, but good news is good news no matter the timescale.
I agree I love these stories :-P
Man MIT is amazing.
@paul34
I agree...much better than Woman MIT =P
@paul34
No manual entry for MIT
@wickywills You mean Oven MIT?
@benjgvps
They only get used in Women MIT =P
**prepares for slap in face**
it's relatively cooler. :)
@Darkroom
Yes it is!
@Darkroom
900 degree is still not that cool.
I'm getting flustered talking about this hot story!
But cheaper for the consumer of for the manufacturer?
@dudedud999 Manufacturer at first, eventually consumer. It'll take a good while to recoup the cost of all the R&D and new factories that will have to be built to take advantage of the technology, when and if it's ready for prime time.
Presumably this occurs due to the presence of the metals and they way they effect the molecular structure of the mixture as a whole.
Similar to how precise Ethanol/Water azeotropes and Benzene/Ethanol/Water azeotropes have a lower boiling point by reducing surface tension.
Great story, keep em coming! Thanks.
this was not what I expected when I read "in reverse" but this is almost as interesting.
@Darkroom
When is anything
Made at MIT going to be more than research
You hear about it but never see anything using the very process they discovered
@Darkroom
Damn you MIT, why won't you love me?!
When Si crystal PV's go do down to $1/Watt I'm going to take buy 10KW of the junk, covering every square inch of my house's southern-facing roof. I'm going to take my AC off the grid, buy a 500W unit for the shed, and buy 2 electric cars. When they go down to $0.10/Watt I am going to implant them in every inch of my exposed skin.
I'll put this in the cures-for-cancer part of my brain, meaning "ignore it until something gets a pricetag".
@Darkroom Yes it is. Eventually I will be able to buy KIRFs for pennies.
Course 3 FTW
@manofchao5
Are you trying to say that nothing worthwhile has ever come out of research done at MIT? Stop trolling. R&D takes time, which is why it is called R&D.
@manofchao5
Seriously. Why have MIT? It's not like people who graduate from there, who come up with those useless processes/devices/inventions ever go do anything productive.
/s
When these type of stories don't pan out engadget should post stories about how no one is actually using these discoveries.
@shaynes
how am i trolling
yes they have done useful research but i just dont think that their research gets used soon enough
like it would usually take a few years but i rather it take only a year
and im not a troll you ass, this is probably only the 1st or second time i got downranked bad
i just think it takes too long getting to production
you can people's opinion but dont hate the person
@manofchao5
"hate"
First of all, this is nothing new, metals and other non-Si elements have been used to getter recombination lifetime killing impurities in Si in the melt for years. The fact that it "un-solidifes" with further cooling was also demo'ed years ago with different materials.
The other funny thing is that this researcher has been working on "Dirty Silicon" for years and very pure Si has gotten so cheap (and continues to get cheaper) that the industry has already passed him by.
Finally, I am sick of journalists repeating the fallacy that solar is too expensive. Used to be true, but that has been changing for a while now. I give you an article Engadget JUST linked to as proof:
http://inhabitat.com/2010/07/29/solar-power-is-cheaper-than-nuclear-for-the-first-time/
This article does not make sense:
I think the conclusion is wrong to say, "solar cells won't require the same high grade of silicon purity for their construction." What should be said is that you can get the same high grade of silicon purity with a lower cost method.
Or am I just missing something here?
@brutek I think you're right.
@Eternal Density
That's what I was thinking... poorly written article. You may start with a less pure material, but you still have to process it to the same purity; the process is just (potentially) easier/cheaper by this method.
I don't understand why everyone says solar is expensive. Sure, it's a really big loan that lasts a lot of years, but, hey, you've got no more electric bill, forever!