Spintronics Used To Create 3D Microchip

← Back to Stories (view on slashdot.org)

Spintronics Used To Create 3D Microchip

Posted by Soulskill on Friday February 1, 2013 @04:33AM from the import-independence-from-flatland dept.

Zothecula writes "A major obstruction to the development of practical 3D microchips is moving data and logic signals from one layer of circuitry to another. This can be done with conventional circuitry, but is quite cumbersome and generates a good deal of heat inside the 3D circuit. Physicists at the University of Cambridge have now developed a spintronic shift register that allows information to be passed between different layers of a 3D microchip. 'To create the microchip, the researchers used an experimental technique called ‘sputtering’. They effectively made a club-sandwich on a silicon chip of cobalt, platinum and ruthenium atoms (abstract). The cobalt and platinum atoms store the digital information in a similar way to how a hard disk drive stores data. The ruthenium atoms act as messengers, communicating that information between neighbouring layers of cobalt and platinum. Each of the layers is only a few atoms thick. They then used a laser technique called MOKE to probe the data content of the different layers. As they switched a magnetic field on and off they saw in the MOKE signal the data climbing layer by layer from the bottom of the chip to the top.'"

28 comments

Min score:

Reason:

Sort:

Magic moke by Anonymous Coward · 2013-02-01 04:37 · Score: 0

But once you let the magic moke out of the chip it won't work any more!
Sputtering is experimental? by Anonymous Coward · 2013-02-01 04:41 · Score: 3, Informative

Sputtering is experimental? News to me.
1. Re:Sputtering is experimental? by bityz · 2013-02-01 04:57 · Score: 1
  
  perhaps making chips by sputtering is experimental?
2. Re:Sputtering is experimental? by cellocgw · 2013-02-01 05:03 · Score: 3, Insightful
  
  Well, certainly, making aggrieved comments by sputtering about the OP is past the experimental stage.
  (disclaimer: I ran sputtering machines in 1974 and they sure as heck weren't new then)
  
  --
  https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
3. Re:Sputtering is experimental? by CanHasDIY · 2013-02-01 05:17 · Score: 2
  
  Sputtering is experimental? News to me.
  A statement pretty much everyone who ever owned a Chevy Corvair would be hard pressed to disagree with.
  
  --
  An enigma, wrapped in a riddle, shrouded in bacon and cheese
4. Re:Sputtering is experimental? by plus_M · 2013-02-01 05:24 · Score: 3, Interesting
  
  Maybe by "experimental technique" they meant "a technique that is used in experiments", rather than "a technique that is still in experimental stages". That is how I read it. We often call things such as X-ray crystallography "experimental techniques".
5. Re:Sputtering is experimental? by Lithdren · 2013-02-01 05:38 · Score: 2
  
  I have no background on this kind of thing so i'm sure someone will come along here and happily correct me if i'm wrong, but the link you provide about Sputtering doesn't seem to have anything to do with what they're doing here.
  They're using Sputtering, but in a new way. I'd consider that experimental for most pratical purpuses. One can make a fire and demonstrate quite easily that boiling water produces steam. However when Heron played around with the idea and invented a (rather impratical but somewhat functional) steam engine, i'd say he was working with something experimental.
6. Re:Sputtering is experimental? by Anonymous Coward · 2013-02-01 06:10 · Score: 0
  
  Ironic that you mention this a couple of days after I passed a Corvair broken down on the side of the road.
7. Re:Sputtering is experimental? by Reverand+Dave · 2013-02-01 08:16 · Score: 2
  
  Nope, sputtering has been a pretty standard part of the semiconductor process for at least 2 decades now...
  
  --
  I got here through a series of tubes
8. Re:Sputtering is experimental? by henryteighth · 2013-02-01 08:19 · Score: 2
  
  Maybe by "experimental technique" they meant "a technique that is used in experiments",
  Indeed. It is an "experimental technique" rather than a "theoretical technique" or a "computational technique", say. It's frustrating to read an abstract of a physics paper which sounds like the authors have performed a nifty measurement, only to find that in fact they are proposing an idea, or have performed a simulation, or theoretically analysed the problem. (Don't get me wrong, they're all equally important things, but not the same as performing an experiment). Thus, it's nice to emphasise one's "experimental technique".
9. Re:Sputtering is experimental? by tragedy · 2013-02-01 10:48 · Score: 1
  
  As ironic as ra-a-a-iane on your wedding day.
10. Re:Sputtering is experimental? by OolimPhon · 2013-02-02 00:42 · Score: 2
  
  Get off my lawn! I was involved in a sputtering project, production not experimental, in 1968.
  It is a well-known industrial method for various kinds of non-obvious plating, such as aluminum on PVC. In the experience I referred to above, aluminum, platinum and/or gold were layered onto glass and silicon substrates.
  It is how the aluminum interconnect layer is deposited on silicon chips, after all.
11. Re:Sputtering is experimental? by Reverand+Dave · 2013-02-04 04:38 · Score: 1
  
  We used AlCu sputtering and AlTi for out interconnects. AlCu looks cool, but is a really dirty process.
  
  --
  I got here through a series of tubes
I thought they were already? by Anonymous Coward · 2013-02-01 05:06 · Score: 0

I thought they already were 3D? I mean, I thought processors included layers of circuits, stacked on top of each other?
1. Re:I thought they were already? by tragedy · 2013-02-01 10:56 · Score: 1
  
  There are relatively few links between the layers, though. The idea here would be to go from a multi-core processor that's, for example, basically just 8 flat processors sitting one on top of the other with some shared components here and there to a multi-core processor where the 8 individual processors are arranged more like eight spheres clumped together with shared components filling in the gaps. Moving from 2D to 3D could mean packing all the actual components a lot closer to one another, allowing signals to travel shorter distances and, therefore, faster. That's the hope, anyway.
tell that to Seymour Cray by swschrad · 2013-02-01 05:07 · Score: 4, Informative

the 1-X had several layers of chips stacked under the epoxy in the ALU section. had a guy in a class who worked in chippewa falls show me a naked chip, pretty cool.
the technique has been around a while, and chip on chip with one reaching over the divide to another stack has been around for quite a while, too. called "dead bug" assembly.
sputtering has been around since the planar transistor, and before that, in putting the active layer on vacuum tube cathodes.

--
if this is supposed to be a new economy, how come they still want my old fashioned money?
Possibly the least laymans summary yet by AwesomeMcgee · 2013-02-01 05:36 · Score: 1

I read a lot of summaries on slashdot on topics I know nothing about, but they usually give me some idea what's going on. This one is giving me no clues but sounds really interesting. I'm not stupid, I just know the software and not the physical medium for squat, please explain this down to relatively smart persons laymans terms someone?
1. Re:Possibly the least laymans summary yet by Anonymous Coward · 2013-02-01 08:24 · Score: 0
  
  The abstract called it a data rachet. It may be easier to think of it as a staircase taking the data from one layer to the next. Each step is made of a cobalt-platinum magnet and the ruthenium spaces out the steps. (The ruthenium forms a break in the ferromagnetic coupling.) When they provide energy to the system (likely in a current pulse), the data (magnetic state) is moved up by one step. So they can take the output from one layer and walk it up to the next layer with much lower current densities and power requirements.
Come On, Slashdot! by Jane+Q.+Public · 2013-02-01 05:55 · Score: 1

Am I seriously the first one here to see the obvious opportunity for jokes about MOKE signals?

You guys are slipping.
1. Re:Come On, Slashdot! by Anonymous Coward · 2013-02-01 08:20 · Score: 0
  
  Am I seriously the first one here to see the obvious opportunity for jokes about MOKE signals?
  You guys are slipping.
  Wake me when they have Secure MOKE signals working.
2. Re:Come On, Slashdot! by metaforest · 2013-02-03 09:51 · Score: 1
  
  If it is magic MOKE, you see, then that signals that you need a new 3D chip.
I wonder by overshoot · 2013-02-01 06:06 · Score: 1

... if it would work better with osmium/iridium instead of platinum? Of course, there's prior art.

--
Lacking <sarcasm> tags, /. substitutes moderation as "Troll."
Hmm by Intrepid+imaginaut · 2013-02-01 06:21 · Score: 1

Probably an obvious question but what advantages are we expecting out of 3D microchips that can't be gotten already from massively parallel systems, just a footprint saving or is there more?
1. Re:Hmm by Anonymous Coward · 2013-02-01 08:36 · Score: 0
  
  The IC itself isn't the "chip"; the package is much bigger. The chip itself is so thin it's almost 2D. If you can stack chips, you can put a whole lot more transistors in a single IC of the same size.
Spintronics, by Anonymous Coward · 2013-02-01 06:44 · Score: 0

Sputtering, transmogrifying, discombobulating... I hope future chip-making keeps digging up funny words like that.
Costs... by alva_edison · 2013-02-01 07:10 · Score: 1

They effectively made a club-sandwich on a silicon chip of cobalt, platinum and ruthenium atoms
So back of the envelope math based on current die sizes, quoted thickness from the article (several hundred nanoscale layers) and $1600/troy oz price for Pt yields an added material cost of roughly $0.12/chip. So, cheaper than I was expecting.

--
He effected a bored affect.
Re:Shitty idea by Anonymous Coward · 2013-02-01 07:27 · Score: 0

Heaven forbid your chips are a few microns thicker... would the packaging of the chip, let alone the device be any bigger as a result? I would worry more about fitting a larger heat source, and hence more cooling components, into a small space than just the physical size of the actual silicon.