Researchers Put 'Spin' in Silicon

ccellist writes "Physorg.com is reporting on the University of Delaware and Cambridge NanoTech's experiments regarding 'spintronics,' or the ability to use information about electron spin in atoms of silicon to encode information, much like we use information about an electron's charge state in computers today. 'Spintronics' research hopes to usher in a new age of computer speed and performance by measuring and even controlling the angular momentum displayed by all electrons, and using this information to encode data. Researchers for the first time have successfully conducted the spin of electrons in a custom-made silicon chip, a process known as 'spin transport.'"

Comments

Putting a spin on it

[aussie_a]

I'd like to see the Luddites try putting a spin on this one.

Re:Putting a spin on it

[Old Benjamin]

bad... or maybe good.

And here I thought that

[zappepcs]

spin control was fully in the political domain? Does this mean that politicians will soon begin to understand the Intarweb tubes?

Re:And here I thought that

[athakur999]

Will these computers work in the Fox News offices? As we all know, that is a "No Spin" zone.

Angular momentum

[suv4x4]

Quoting from a movie I saw once. The editor tells the article author how he edited his work:

Editor: "I replaced atom with molecule here and there, atom repeats too much"
Author: "But... it's not the same thing at all!"
Editor: "Oh come on! Who'll know the difference. Molecule, atom.. same thing to me."

So, in the light of this, particle "spin" isn't about an electron actually spinning, and thus "angular momentum" as seen in the article text, so that's pretty hilarious replacement.

Another thing you may want to know for future articles: quark colors also aren't actual colors.

Re:Angular momentum

[Hemogoblin]

It was here that the thaum, hitherto believed to be the smallest possible particle of magic, was succesfully demonstrated to be made up of /resons/ (Lit.: 'Thing-ies') or reality fragments. Currently research indicates that each reson is itself made up of a combination of at least five 'flavours', known as 'up', 'down', 'sideways', 'sex appeal' and 'peppermint'.

- Terry Pratchett, Lords and Ladies

Re:Angular momentum

[Nymz]

In this case I'm not sure if either term "spin" or "angular momentum" is accurate. Perhaps they should create a unique name to describe the currently unknown action. I know they like to name these things after the first person to imagine, define, discover, or prove it, but that hasn't stop naming based on unknowns before. Some examples...

UFO - unidentified flying object, but now used in the positive sense. Think about someone asking "do you believe in UFOs", could you possibly say "no, I don't believe in flying objects that cannot be identified"?

Dark Matter - originally to explain why equations don't calculate out equally, and suggested some unaccounted principle or force that we don't know about or understand. Now, it's actually matter.

SIDS - suden infant death syndrome, once a catch-all phrase put on death certificates that meant unknown cause. Now studies blame various aliments like second hand smoke for Sids. Technically, if second hand smoke was the cause, then it would be the cause itself, and not Sids, but trying to explain that is more difficult than explaining the end of the millenium is on December 31st 2000, not 1999.

Re:Angular momentum

[digitalderbs]

nah, it's all appropriate. The angular momentum is on the magnetic moment produced at the electron from an applied magnetic field, due mainly to the Zeeman interaction. The spin itself is the magnetic moment of the electron. Changes in the magnetic field (non parallel) apply a torque on the moment and change its orientation. The motions are described by the quantum angular momentum equations .. i.e. commutation relations. These commutation relations are the quantum analog to the classical angular momentum equations.

Re:Angular momentum

[Anonymous Coward]

Also see the "not the comp.text.sgml FAQ":

Part 5. Terminology.

Q. I've tried reading the (XML | SGML | XSL | XPATH | DSSSL | ...) specification, but it doesn't make any sense! There's too much jargon!

A. Specification authors deliberately obfuscate the text of ISO and W3C standards to ensure that normal people (e.g., Perl programmers) can't use the technology without assistance from the so-called "experts" who designed the specs.

Fortunately, there is a handy translation table you can use:

ISO/W3C terminology : Common name

attribute : tag
attribute value : tag
attribute value literal : tag
attribute value specification : tag
character reference : tag
comment : tag
comment declaration : tag
declaration : tag
document type declaration : tag
document type definition : tag
element : tag
element type : tag
element type name : tag
entity : tag
entity reference : tag
general entity : tag
generic identifier : tag
literal : tag
numeric character reference : tag
parameter entity : tag
parameter literal : tag
processing instruction : tag
tag : command

With the help of this table, even Visual Basic programmers should have no trouble deciphering ISO prose.

Sorry for the formatting, Slash keeps complaining about "junk characters" (spaces are not junk!)

Re:Angular momentum

[chrisb33]

Unfortunately, the "hilarious" joke is on you, since spin actually does refer to angular momentum. It can even be interchanged with "orbital" (extrinsic) angular momentum while obeying momentum conservation, and (as the article mentions) can be measured using a magnetic field.

A quick look at wikipedia before posting is usually helpful.

Re:Angular momentum

[suv4x4]

Unfortunately, the "hilarious" joke is on you, since spin actually does refer to angular momentum.

Well, I admit defeat, but at least now we all collectively have something to laugh about.

Re:Angular momentum

[noidentity]

'So, in the light of this, particle "spin" isn't about an electron actually spinning, and thus "angular momentum" as seen in the article text, so that's pretty hilarious replacement.'

I offer "angular momentum enhancement" as a politically-correct alternative to "spin control".

It's actually quite straightforward.

[Anonymous Coward]

This sort of physics is actually pretty easy to comprehend, so I'm not sure how it could be twisted.

Basically, we can think of an atom as a sea urchin. So around this atom, we have a number of spikes. These can be considered the electrons of the atom. Now, this is a major simplification of Schroedinger's equation, but essentially each spike represents the probabilty of locating an electron within a volume of space.

Now, these spikes come in pairs, in order to balance each other. They're on opposite sides of the sea urchin atom. They slide around the atom, but they're always opposite to one another. This is typically called the Pauli exclusion principle by physicists.

So to an external viewer looking at the atom, it appears as though the spike pairs are spinning around the circumference of the atom. Relative to the viewer's position, these spike pairs are moving either clockwise or counterclockwise around the atom. This is how we get the two different spins, which thus can be used to repesent binary information.

Now, when you're dealing with larger atoms, with many electron pairs, the interaction between the electrons leads to a greater degree of electron stability and predictability. Thus electron pairs will still spin around the atom, but they'll travel in a path that's actually quite consistent, relative to the other electron pairs. By focusing only on certain electron pair paths, and the direction that they spin around the atom relative to the viewer's position, we can store large amounts of binary data per atom.

Re:It's actually quite straightforward.

[Nullav]

I'm a bit confused by your analogy. I was convinced that electrons didn't orbit in any sort of path, and that they just constantly moved to random positions in the shell.

Re:It's actually quite straightforward.

[Anonymous Coward]

You're actually thinking of the outdated Rutherford-Moore model of the atom. While it was practical back in its day, it doesn't adequately describe much of the observed phenomena since that time. That's why we need to look towards models like those of quantum mechanics, where we're not dealing with randomness (because we're not in actuality), and instead we're dealing with probability distributions.

Think of the sea urchin spikes as the probability density function of a three-dimensional Bell curve. We have a greater probability of finding an electron closer to the atom thus resulting in the greater spread at the bottom of the spike. On the other hand, we have a small probability of finding an electron the further away we go from the atom, represented by the smaller spread (ie. the point) of the spike.

Re:It's actually quite straightforward.

[beyondkaoru]

quantum mechanics is very zen. as described in the model, you can't really think of electrons (or particles in general) as having position -- they don't. they have probabilities of being somewhere. ...and then you take the fourier transform of it and demonstrate the uncertainty principle, it's really cool. kind of trippy actually.

i'm a college student who took an introductory class on quantum computing with a really cool professor.

Re:It's actually quite straightforward.

[grammar fascist]

Thanks for that - very helpful.

Things like this are why I still come to Slashdot.

Re:It's actually quite straightforward.

[maxume]

The Academics Anonymous group meeting is next door.

Re:It's actually quite straightforward.

[Ceriel Nosforit]

Posting just to agree with the sibling. Thanks, both of you. Visual descriptions like that are a fantastic aid to get a starting grasp of what in the world the theory is talking about.

I've run across mentions of the Pauli exclusion principle before, but with this 'model' in hand I'll take another look at it.

Expanding markets

[HotBBQ]

I have a far more entertaining idea for spinning silicone. It mostly involves cheesy stage names and tassels.

Re:Expanding markets

[Anonymous Coward]

+1 excited state achieved

Sounds like MRI

[necro81]

After spin injection, electrons in the silicon were then subjected to a magnetic field, which caused their spin direction to "precess" or gyrate (much like gravity's effect on a rotating gyroscope), producing tell-tale oscillations in their measurement.

This sounds like the process used in Magnetic Resonance Imaging. In MRI, they use a BIG magnet to create a very strong magnetic field in a person's body. The main field is usually 1-3 Tesla, depending on the scanner (for reference, Earth's magnetic field is 30-50 microtesla). Then they use smaller magnets to establish a gradient in that main field, and RF pulses to query the spin precession of atoms in the body. In the case of human imaging, I think they focus on the spin precession of a hydrogen nucleus (a proton) in water. In function MRI, they focus on hemoglobin (which contains a little ferromagnetic iron, ya see), to determine where blood is most present. See this for an exhaustive overview of how it works.

Their spintronics methods sound similar, except it's focused on a much smaller volume (a chip instead of a human body), and are tuned to the electrons in doped silicon. Very cool.

Re:Sounds like MRI

[Eternauta3k]

which contains a little ferromagnetic iron, ya see

AFAIK, iron in hemoglobin isn't ferromagnetic

Re:Sounds like MRI

[belg4mit]

MRI has nothing to do with blood per se (PET is what you use to image activity).
It's just the medical/consumer friendly term for NMR (nuclear magnetic resonance).
It also has nothing to do with electromagnetism (e.g; ferromentism), but instead on
the magnet moment resulting from unpaired nucleons that occur in some isotopes of
various elements e.g; 1H and 13C.

Re:Sounds like MRI

[Artifakt]

The hemoglobin molecule in humans and related life forms, uses one individual iron atom. Four of these proteins are used in each red blood cell. The iron atoms are held by the carbon chains in the proteins at distances that put them much farther apart than they would be in an iron crystal, and not directly bonded to each other, so there are no crystal domains that can be lined up by a magnetic field.
    To put it another way, normally, what scientists mean by 'ferromagnetic' is both that the domains will line up inside a crystal under sufficient magnetic field from an external source, and that they will stay that way (at least partially), so the object becomes permanently magnetized. Iron atoms in hemoglobin will respond in groups to a sufficiently strong external field, but normally, this grouping doesn't count as a domain, as it's not in a crystal at all. Using the word domain this way is kind of like calling a bunch of seagulls a flock when they are spread out all over the coastline at multi-mile intervals. It's not totally prohibited to use the word domain this way, but it's far from standard. More definitively, the iron atoms won't 'remember' the field and stay magnetized, so the second test for ferromagnetism is a definite and unequivocal no, and they are normally best called paramagnetic.
        I suppose a field strong enough that would pull all the iron in a body into one concentrated spot and force it into such proximity that it crystalized there is possible. This would be a the 1.0 Ian McKellan unitary field, of course.

Big mistake in the summary

[OeLeWaPpErKe]

"much like we use information about an electron's charge state in computers today"

Obviously the charge of an electron is constant (in case you get confused eV is a measurement of velocity, not charge). What we use in computers today is the QUANTITY of electrons "flowing" (these days tunnelling may be a better term) through non-conductive layers.

And no non-conductive is exactly where it needs to be. The edge of today's transistor is a non-conductive, but very small, silicon edge.

Re:Big mistake in the summary

[Anonymous Coward]

The spin of the electron is constant as well.

eV is a measure of energy, not velocity.

The benefit of spintronics is you can transmit information without having current flow, generating heat. Of course you can do spintronics with current flow as well, but this doesn't really gain you anything.

Re:Big mistake in the summary

[l2718]

Obviously the charge of an electron is constant (in case you get confused eV is a measurement of velocity, not charge). What we use in computers today is the QUANTITY of electrons "flowing" (these days tunnelling may be a better term) through non-conductive layers.

Man, there are so many errors here I don't know where to begin. The Electron-Volt (eV) is a unit of energy (the work required to move an electron across a potential different of one Volt). Digital computers do not depend on the magnitude of the current, but on its abssence or presence. In fact, the goal is to have as little current as possible (less losses due to heat and radiation) -- we are nearing single-electron transistors. "Spintronics" would instead carry the information in the spin state (up or down) of an electron. The reference to "charge" probably stems from memory, where information is stored in the magnetization state of a small amount of matter.

Re:Big mistake in the summary

[glwtta]

Regardless, this: "we use information about an electron's charge state in computers today" is still a ridiculous statement.

Re:Big mistake in the summary

[MattskEE]

Digital computers do not depend on the magnitude of the current, but on its abssence or presence.

What you have just described is in fact a measure of magnitude of current, where a current above a certain threshold produces an output voltage high enough to be considered a 1, and a current below another threshold produces an output voltage corresponding to a 0. This is how early digital electronics worked.

But in modern CMOS digital circuits, 1's and 0's are represented by voltage levels, not current magnitudes, nor current's absence or presence. When the output of a digital circuit is changing, some current will flow out of or into it until the voltage level reaches whatever thresholds have been set for 1's and 0's. The current is actually charging or discharging the intrinsic capacitances of whatever circuits are connected to it. The more current each digital circuit is capable of sourcing or sinking, the faster the output voltage will be able to change between a 1 and 0, or vice versa, and the faster the total clock frequency can be.

Re:Big mistake in the summary

[OeLeWaPpErKe]

Man, there are so many errors here I don't know where to begin. The Electron-Volt (eV) is a unit of energy (the work required to move an electron across a potential different of one Volt). Digital computers do not depend on the magnitude of the current, but on its abssence or presence. In fact, the goal is to have as little current as possible (less losses due to heat and radiation) -- we are nearing single-electron transistors. "Spintronics" would instead carry the information in the spin state (up or down) of an electron. The reference to "charge" probably stems from memory, where information is stored in the magnetization state of a small amount of matter. You, my man, are an idiot : 1) what is the difference between an electron at 1 keV and 100 keV ? The speed. 2) absence or presence ? You haven't touched a circuit in 20 years, have you ? Maybe you have a point in a "my first circuit" setup, but in the real world 0's also have electrons flowing 3) "nearing single electron transistors" ? You wouldn't be making this statement if you had EVER used a transistor. It wouldn't have any use for a single electron transistor. It's an absurd concept. A transistor is an amplifier. 4) charge and "magnetization state" are completely, completely different things, oh and btw, I've yet to see the first magnetic memory in operation. hard disks are generally referred to as external storage, not memory I guess slashdot will be slashdot.

Re:Big mistake in the summary

[lachlan76]

As the other poster said, eV is a unit of energy, not velocity. Velocity is in ms^-1. 1ev ~ 1.6e-19 joules.

Obviously the charge of an electron is constant (in case you get confused eV is a measurement of velocity, not charge). What we use in computers today is the QUANTITY of electrons "flowing" (these days tunnelling may be a better term) through non-conductive layers.

What the article means is that we use the charge of an electron to transmit information at this point, rather than the spin or other properties. While we are not able to change the charge of an electron, we can change the charge of a flow of electrons by modifying the amount of current flowing.

Re:Big mistake in the summary

[ccellist]

Thanks for clearing that up for me. I guess I misunderstand the way the 1's and 0's look like at the atomic level. So you're saying that below a certain threshold of electrons flowing the IC yields a zero ("false"), and above that a 1 ("true")?

Clever

[heinousjay]

That's the first Fox News joke that actually made me laugh.

Spintronics

[AmishElvis]

is a stupid name. It sounds like a fad exercise routine.

Re:Spintronics

[Anonymous Coward]

Or an awkward system for modifying the transmission of a car.

FUD

[MillionthMonkey]

If my laptop has a spin-based CPU and I flip it upside down do all the white pixels turn black and vice versa?

This is completely not an issue with a charge-based CPU.

Where is...

[drgruney]

John C. Dvorak when you need him?

http://en.wikipedia.org/wiki/Silicon_Spin

Re:Where is...

[Bones3D_mac]

Gah, someone actually beat me to it!

"Silicon Spin" was probably the last good thing Dvorak did before he went all psycho on the tech industry.

(Of course, considering the TV abomination that is now "G4", I'd probably go a little psycho too...)

Terribly confusing summary. Is that Chinese?

[zymano]

Here is a better explanation. This topic has been covered before many times.

http://scientificamerican.com/article.cfm?articleI D=959FBD96-E7F2-99DF-341F959A7DA2A292&chanID=sa003

http://www.sciam.com/article.cfm?articleID=0007A73 5-759A-1CDD-B4A8809EC588EEDF

No Spin Zone

[kybred]

Keep Bill O'Reilly away from these guys!

Where's the good stuff?

[slackarse]

FTFA:

"We hope we're with spintronics where Bell Labs was with semiconductor electronics in 1948," Appelbaum said. Does that mean we're about 30 years away from being able to buy something "cool" based on this?

Spin State Energies

[Doc Ruby]

Is there an energy difference between the two spin states, in joules (or electron-volts, whatever the unit), the way there's an energy difference between an atom with an electron in two different orbit shell states?

If so, how much is the difference? Is there a way to move an electron from the higher energy spin state to the lower one that consumes less energy than the state difference? A way to move the electron from lower energy to higher energy spin state that is less than the difference? Can those moves be done on lots of electrons? How about lots of them per second, but serially, selecting which electrons in the stream are affected, not all of them?

Re:Spin State Energies

[orclevegam]

This is way outside my field, but I can give it a shot. From what I understand spin isn't a question of energy state, rather it's a property of electrons, like magnetism is for some metals (this is probably a terribly innacurate comparison, so any physicists out there cut me some slack). An electron can either have up spin or down spin, so you have the capability to represent a bit with the spin state of a electron. Now, what makes spintronics looks so good is two properties. One, because you aren't shoving electrons around inside conductors (or through non-conductors) you don't have to worry about all that nasty waste heat, and second spin state propagates through materials at high speeds (I don't remember quite what, and it's been a while since I've seen the article on it, but I seem to remember it being better than normal electron propagation speeds, although that might only have been in superconductors).

Re:Spin State Energies

[Doc Ruby]

Yes to all that, but I believe the two different spin states have different energy levels - even if the subatomic mechanics are currently unknown. And that it takes energy, however small, to change the states - but that one "direction" of change requires more, and that perhaps the other direction returns at least some of that energy.

There is a mechanical change, some "stuff" is in a different physical location, when changing electron spin. The two states might have the same energy level, changing between them might be "frictionless" and thereby consume the same amount of energy for a single cycle as for a trillion. But this is matter/energy in space, like anything else.

I'll keep looking for the precise answer to my half-informed question.

we're too young for magnetic memory...

[sash]

oh and btw, I've yet to see the first magnetic memory in operation. Yeah, me too - the only one I've ever seen was gathering dust on a shelf of the Physics datacentre, 15 years ago. And I haven't seen puch cards in operation either!

It's so nice those few times when technology makes me feel young :-P

Cowboy Fucking Neil HOORAH!!!!!

[therufus]

Okay, call it offtopic, whatever. But I reckon Teh Cowboy has been waiting for a decade to post an article with ANY Vanilla Ice reference possible.

ALL HAIL TEH NEILZORZ!!!!!

Okay, now I'm going to be in negative karma for this, but it's worth it.