Slashdot Mirror
Storing Qubits In Nuclei
bednarz writes "Scientists have demonstrated what is being called the 'ultimate miniaturization of computer memory,' storing data for nearly two seconds in the nucleus of an atom of phosphorus. The hybrid quantum memory technique is a key step in the development of quantum computers, according to the National Science Foundation. An international team of scientists demonstrated that quantum information stored in a nucleus has a lifetime of about 1¾ seconds. 'This is significant because before this technique was developed, the longest researchers could preserve quantum information in silicon was a few tens of milliseconds. Other researchers studying quantum computing recently calculated that if a quantum system could store information for at least one second, error correction techniques could then protect that data for an indefinite period of time.'" Here's the NSF press release with pictures of the apparatus. They claim that this technique is promising because it "uses silicon technology" seems a bit of a stretch — the silicon the researchers employed was a painstakingly grown crystal of extremely high purity.
I heard BGC3 has already patented this idea.
Everything old is new again.
In Liberty, Rene
An international team of scientists demonstrated that quantum information stored in a nucleus has a lifetime of about 1¾ seconds
Just as long as it takes me to c..
... compute 2 + 2
The plural of nucleus is nuclei, please!
Unselfish actions pay back better
The first thing I thought was all I need now is a miniturized keyboard and mouse and the worlds smallest lcd screen.
"The stupid neither forgive nor forget; the naive forgive and forget; the wise forgive but do not forget." -Thomas Szasz
It wasn't so much that we thought "souvlaki" was a latin plural when the dish was clearly of Greek origin that bothered the restaurant owner so much as our constant bickering whether the singlar was "souvlakum" or "souvlakus".
(with apologies to Wayne and Shuster)
In Liberty, Rene
1¾ seconds should be enough for anyone!
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
Here's the NSF press release
Anyone else read that as the "Not Safe For [the] Press" release?
"I'm not sure I like the fugnutish tone you used in your post!" -RogL (608926)-
Not only that, but it's not like the silicon used in today's chips is low grade crap. The purity standards for electronic grade silicon are pretty insane considered to the standards of most things we think of as "pure", including pharmaceuticals. (Seven to eight 9's purity is not uncommon). And yet its produced in great volumes relatively cheaply.
nucleus = singular
nuclei = plural
nucleii = ???
However, this isn't the first time short term memory has been used in computing. I can remember (pardon the pun) memory which had to be refreshed, so I'd imagine using that concept would fix the "short" timespan.
However, that's not the only important number. What about latency?
Insert
It's a front. The researchers all ganged up and wrote a bunch of nonsensical papers, then they used the grant money for blackjack and hookers.
I haven't had time to read the nature article quite yet, but it would appear that magnetic moment coherence information is transfered from electrons, which decohere quickly, to nuclei, which decohere much more slowly. Magnetic moments on nuclei in the solid-state and in the absence of local motions can maintain coherences for minutes to hours -- this is not surprising. However, I can't tell from this summary how this is different from DNP, a well established method. Maybe because it was done in silicon?
To run Vista?
Last time I looked, single-crystal silicon technology (what's used in chips except for things like amorphous-silicon memory) consists EXPLICITLY of "painstakingly grown crystal of extremely high purity".
- A defect in the crystal structure results in the failure of every component that the defect is present in.
- Carefully-controlled Minuscule fractions of impurity atoms selectively substituted for silicon atoms define the active regions. Unplanned impurities change the characteristics, resulting in components that don't behave according to design.
So existing silicon technology is exactly what is required. Bednarz's concerns are off the mark. The purity and crystalline nature of the component won't impose any extra costs, because it's what is already done.
Some OTHER requirement MIGHT make it costly. But that's a separate issue.)
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
I bet that refreshing a qbit will face all kinds of problems... More specificaly, the uncertainty principle forbids refreshing qbits, unless you want them to behave like clasical bits.
A computer based on this would have to make the hole quantum calculation on 1 3/4 seconds, all the way into a classical result. That would be enough to break RSA if there was a big enough computer, but I guess that isn't enough for everybody. Also, I don't know if it is possible to couple several nuclei and create a computer out of those qbits.
Anyway, that is an awesome result. Maybe people can create some error corretion and increase that time, and maybe somebody already knows how to couple nuclei. Since I am not a physicist, I don't know about the state of the art.
Rethinking email
'in' takes the ablative, so let's go all the way and use the ablative plural, nucleis.
occultae nullus est respectus musicae - originally a Greek proverb
Silicon for electronics has additional requirements. It isn't simply that it has to be ultra-pure for the element, but it also needs to be ultra-pure for the specific isotope. Further, there have to be minimal flaws in the crystalline structure across the entire wafer for any reason whatsoever. That gets complicated when you consider that modern chip making uses all kinds of techniques for doping, stressing and god-knows-what-elsing to improve performance, though there are other factors. If pharmaceuticals can be improved in microgravity, where the smallest unit you care about is an entire complex molecule, a process that is sensitive to the displacement of single atoms is necessarily going to be much more sensitive to any issues full Earth gravity is going to throw into the mix. (I'm not saying that that's a real problem, only that if the oft-quoted case for pharmaceuticals is true, it must be hundreds if not thousands of times more so for wafer production.)
The way silicon wafer production tends to work is to assume a moderate rejection rate. Chip makers test the chips and if they fail QA may simply be re-stamped at a lower grade. The best-known example of this was the early production of the 486SX and the 487, which were just 486DXes in which either the main CPU or the coprocessor had failed in testing. Those from Britain may also be familiar with Sir Clive Sinclair buying rejected chips on the grounds that most rejects for industrial use were perfectly acceptable for home use, and the cost of replacing was still cheaper than buying the better quality chips.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Somehow you store a qbit which is both 0 and 1. Then you try to retrieve it. Problem is, as soon as you do so, it collapses to either 0 or 1. So how do you know that what you stored is what you got back?
You don't retrieve it in a way that causes the entanglement to collapse. You instead transfer the enganglement to another particle which then participates in the next step of the computation (or perform that computational step on the nucleus that has been acting as a storage medium).
The first one corresponds to a memory (with a destructive read - because you can't COPY entanglement, so the qbit itself DOES collapse when the information is transferred out).
The second one corresponds to a bit in a datapath register where the computation takes place in the register logic rather than in a nearby hunk of logic. (I.e. the old "accumulator" style of processor typical through the 1960s.)
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
A quantum computer is cleverly built to operate in both the "crashed" and "running" states simultaneously. Since the "running" state is the only one that responds to the user's actions, users never have to interact with the "crashed" state. This makes Windows run much better.
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
A defect in the crystal structure results in the failure of every component that the defect is present in.
Every component, or one out of one. That's something like 100.00001863%
Wow, you're still using a Pentium? I feel sorry for you.
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
Throw some DRAM-style refreshing in, and it could be viable at even that lifespan.
> A defect in the crystal structure results in the failure of every component that the defect is present in.
That's not entirely true. The last time I checked, which has admittedly been some time, the defect rate was usually on the order of a part per billion or so. Excellent, certainly, but far from perfect considering that still means billions of billions of ... defects in each wafer. The key is that small, isolated defects are tolerable, so you only need to junk parts with high concentrations or an unlucky distribution. This is why the exact same processor die can have different speed limits.
Another thing worth pointing out is that having a controlled rate of SiO2 defects in the silicon crystal was actually found to be beneficial. Again, though, I don't know if that still hold for, say, =65nm processes.
The point being that without knowing the how "extremely high" the purity needs to be it is impossible to say definitively whether existing tech is good enough. However, if I were designing this I would certainly try to make it work!
I barely understand it myself. A decent explanation by someone who gets it and can communicate in human language would be good :D
But from what I can gather... they do everything that standard binary computers do, but other things as well. Instead of just working on binary bits (which can be 1 or 0), they operate on qubits (which can be 1 or 0, in one dimension, but also have a quantum spin, adding another dimension). So essentially, they work with complex (2-dimensional) numbers as their basic level of computation, rather than just working on a single (onoff) dimension.
If you think of the few manipulations we can do on a bit: AND it with another bit, OR it, XOR it, negate it... and then think of everything that's come from those simple principles: adding numbers, doing logical ANDs and ORs (as opposed to bitwise ANDs and ORs), making decisions and carrying out actions based on that logic, doing simple multiplications, divisions, sorting, etc... all the way up to complex layers above this, like operating systems and mp3 codecs. Now imagine where we'd be if the most fundamental building block, the bit, was 2-dimensional instead of 1-dimensional. It's a whole new ballgame.
One of the few currently known benefits of quantum computers is this: the advanced cryptography we use to protect secrets is based on how long it would take a computer to find the cryptographic key by doing lots of brute-force calculations. Usually, security types are happy with something that would take a lot of expensive computers ten years to crack. Quantum computers, though, by using their extra dimensions to calculate, can do these so much faster, that they're no longer safe. The way it seems to work is that, with a standard computer, you have to do lots of seperate calculations, which, all together, take an exponential amount of time. Whereas, with a quantum computer, it can be reduced to a simpler function that can all be done at once, or at least, in a way that doesn't require every part of the sum to be done seperately.
At a guess, I'd say they'll probably be good for anything involving complex numbers, too. In particular, the Dirac codec, Realvideo9 (or is it 10?) and other wavelet codecs, and probably a lot of things that we use graphics cards' processors and vector processors like MMX/3dNow/Altivec for now.
Anyway. This is just what I've been able to pick up, and might well be wrong on a little/most of it. Someone with more maths training can probably get a better idea from wikipedia:
http://en.wikipedia.org/wiki/Quantum_computer
You might also want to read up on complex numbers.
If noah was already building stuff with qubits, he was getting a lot more help from god than we all thought.
I think that remark about high purity silicon is by the editor kdawson, not the submitter bednarz. I don't know where he came up with "painstakingly grown crystal of extremely high purity" - it's not in the NSF press release. But searching in Nature reveals the phrase "P-31 donors in isotopically pure Si-28 crystal" in the abstract. So maybe the isotopic (number of neutrons) purity of their material is above and beyond the chemical (number of protons) purity of standard microelectronic silicon.
Every component, or one out of one.
In this context a component is a circuit element on the chip (i.e. a transistor or the like) while the chip is an "assembly", not a "component".
Of course most chips don't have redundancy and fail if any of the (millions of) components on them are defective.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
Information? What information? There is no fucking information. It's both 1 and 0 from beginning to end, for crying out loud.
The information is not stored as the state of the particle but as the entanglement of the states of the set of particles. The state of a set of N entangled qbits encodes 2^N separate possible sets of states and thus 2^N bits of information, while one operation on the set can perform 2^N operations in parallel, one on each of the possible combinations of states.
Now there are a limited number of things you can do. And to get to a usable output you have to perform computations that reduce the allowed states to the set that contains the answers you want, of which you only get to observe one when you finally collapse the wave functions. But that's enough to do some very useful computations.
Like perhaps using a chip containing N entangled qbits and suitable supporting structures to find a factor of a positive composite number M = ((2^N)-1) in O(N) time. Goodbye RSA encryption, hello Big Brother.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
This is what I think happened.
A lab assistant pinned a periodic table chart at the end of the hallway.
Then one of the main researchers got to throw a dart at it blindfolded and it came up as Phosphorus.
"Looks like this time around we'll be using Phosphorus."
"Ok your turn...Ahhh Silica!
Let's get down to the beach, get some sand and shove some K into it and zap it with high voltage and see what it does!"
Seriously! Phosphorus is not the easiest thing to work with. Why that?
Don't be apathetic. Procrastinate!
This could be the "ultimate miniaturization of computer memory", if not for the fact that each nucleus is wrapped in 15 electrons and about a trillion times its own volume of empty space. Unless, of course, they've found a way to contain degenerate matter and selectively polarize individual nuclei therein -- and I'm thinking compressing matter to degeneracy would tend to shorten those T1 times pretty substantially.
While this is certainly a neat result, calling it the "ultimate miniaturization" is silly press-release-talk. For sure, I know of quantum dots in GaAs approaches to quantum computing that store qubits in few or even single electrons. Though the current approach for GaAs can only store qubits for a few hundred microseconds at best, the storage time for Si/SiGe heterostructures could be as long as a few seconds (that hasn't yet been measured, as far as I know, so a few seconds is just a prediction). Beyond that I think that there are efforts to use photons for qubits, but I don't know much about them. An electron is certainly smaller than a nucleus, as much as it can be said to have size, and photons don't have mass, so it's hard to imagine what size they'd be.
Given that nuclear power comes from splitting atoms, I think you'd have to be combining atoms to "make" a nuclear battery. You'd be putting in as much energy as the later nuclear fission would give out (including what's lost as heat or light or whatever you don't use), plus some in manufacturing. We're probably not close to exhausting our resources of uranium etc., and fusion might eventually arrive before that happens, so it'd all be a bit pointless at this stage I guess. Definitely worth thinking about though, yeah :)
Fission and fusion can both be either endothermic or exothermic, depending on what you're working with. If you're working with atoms lighter than Iron, you get energy from fusion, and fission requires providing energy. If you're working with atoms heavier than Iron, you get energy from fission, and fusion requires providing energy. If you're working with Iron, you have to provide energy for both.
There's no failure quite as dissatisfying as a complete and total solution to the wrong problem.