Researchers Create Silicon-Based Quantum Bit 46
angry tapir writes "Researchers at the University of New South Wales in Australia have created the world's first working quantum bit based on a single atom in silicon. The research team was able to both read and write information using the spin, or magnetic orientation, of an electron bound to a single phosphorous atom embedded in a silicon chip. In February, UNSW researchers revealed they had successfully created a single-atom transistor using a single phosphorous atom in a silicon crystal."
ONLY A BIT ?? (Score:1, Funny)
Seems like a bit of waste !!
Re: (Score:3, Insightful)
so did a single vacuum tube when that was first invented...
Re: (Score:2)
The difference being that you can't do anything with a single bit, although this will of course lead to better things, but a single vaccuum tube, a diode, wire, and a speaker will make a radio (without the tube you need headphones). A single vaccuum tube is an amplifier..
Re: (Score:3)
A single vaccuum tube is an amplifier.
So's a single transistor.
Both become "bits" only when you choose to interpret their outputs as binary.
Re: (Score:2)
I'm not talking about bits, I'm talking about sound amplification. You can hook an input to a tube's cathode, a speaker to its anode, a power source to its heater, and it will amplify the input. You need at least two transistors to do that.
Re: (Score:2)
Ah, so saying "bit" was a typo. Okay. Anyway, you only need one transistor to make a linear amplifier.
Re: (Score:2)
Re: (Score:1)
Quantum computers aren't intended for general purpose computing or math operations
Not entirely true. There is a lot of work being done on general purpose quantum computing architectures at the moment. In fact there are already some quantum compilers out there such as http://tph.tuwien.ac.at/~oemer/qcl.html [tuwien.ac.at]. Also, a number of algorithms do actually require math operations to be performed by the quantum processor. Remember that measuring a qubit collapses its wavefunction so it is often important to do math operations on qubits before they are measured.
Useful links (Score:5, Informative)
Actual research article [nature.com]
Wikipedia article on the architecture proposal [wikipedia.org]
Bit not a Qubit (Score:5, Informative)
Re:Bit not a Qubit (Score:5, Informative)
They show relatively clear Rabi oscillations [wikipedia.org], which are a definite proof of the quantumness of the evolution of their system (which has nothing to do with entanglement). So, yes, this is a genuine qubit, albeit not a perfect one.
Re:Bit not a Qubit (Score:5, Funny)
No, they've demonstrated a single Qubit.
TFS makes things mucky by mentioning single electron transistors too, which are a completely different beast.
The problem with quantum computing isn't demonstrating single qubits though. The problem is in getting a reasonable number in a superposition. Most I've ever seen in a QC that actually does computations is 7 qubits.
Just to get an idea of the scale we need, Shor's algorithm, the one which we could use to crack RSA encryption in polynomial time, needs 2*N qubits minimum. So to crack RSA1024 we'd need 2048 qubits all in a state of superposition.
I'm of the opinion adding more qubits to a superposition is going to be an exponentially hard problem.
Re: (Score:2, Interesting)
I'm of the opinion adding more qubits to a superposition is going to be an exponentially hard problem.
History tells otherwise. [quantenblog.net]
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
So at some level we will be able to store more than off/on information? How many states can we expect? 3? 10?
Re: (Score:1)
A qbit is a quantum binary digit, so there will only two states by definition. Otherwise, it's not a binary digit. Eventually, we might find a way to pack more than 1 qbit per physical device, but we're having enough trouble just getting a usable number of qbits.
On the other hand, regular transistors are a technology that is quite mature, and most people don't realize they're actually analog devices. We just pick a threshold current, where anything above is on and anything below is off, in order to make
Re: (Score:2)
So at some level we will be able to store more than off/on information?
You could do that now, but you would need a completely new archetecture. You could have a three state machine, positive, negative, and off. It would work in trinary rather than binary (000 001 002 010 011 012 020 021...)
back in the fifties through seventies they had "infinite state" analog computers in many universities. I built a primitive one when I was 12 (more of an electric slide rule than a computer). But the big fancy ones did rea
Re: (Score:2)
I have often thought of the ternary/trinary architecture with the positive/negative/off or a no charge, half charge, full charge type of breakdown.
An infinite state medium of storage increases chances for corruption. I was wondering if there are multi-state quantum particles at the atomic or subatomic level. That would make for REALLY small mediums of storage. A 100-state subatomic particle would result in TBs of storage in the smallest imaginable locations.
Re: (Score:2)
I have often thought of the ternary/trinary architecture with the positive/negative/off or a no charge, half charge, full charge type of breakdown.
That makes the circuits much harder to design.
For traces that are short enough that you can just treat them as capacitive/resistive loads rather than transmission lines, the nice thing about a binary circuit is that you just drive the output towards one of the two voltage rails. Driving it harder just gives you the result you want faster, or drive it slower if that's okay.
With a 3-rail circuit there would be plenty of situations where driving too fast towards the middle rail would result in reading '0' or
Re: (Score:2)
Re:payback (Score:5, Funny)
It's called PHOSPHORUS (Score:1)
...not phosphorous. That's something different.
Geez. For a geek site, I'd expect better. Just look up in Wikipedia if unsure.
Dr. Morello comes off as a great marketer but (Score:2)
provides little details on how it was done. I suspect an incremental but important improvement being hyped as revolutionary.
I have a question: (Score:2)
Re:I have a question: (Score:5, Funny)
can you run linux on it?
Nope. Linux requires at least a two bit computer to run.
Re: (Score:1)
|linux> = kernel|1> + distro|2> + windowmanager|3> + shell|4>
So it seems to be in a superposition of 4 qubits (and I'm sure I'm forgetting a few), so: Yes it can run linux, or better: linux IS quantum computing.
Re: (Score:1)
*note to myself: use preview funtion*
Re: (Score:2)
Yeah, that just ruined the joke. ;)
Time for us all to learn QCL, Q-Lisp... (Score:2, Funny)
...or some other quantum computer programming language.
(Of course, I suppose that to really do quantum software development correctly, you have to learn to be great at it and totally suck at the same time. Well, at least while no one's watching...)
Impressive when you consider... (Score:2)
Re: (Score:2)
... that the Core i7 processor containing 731,000,000 transistors weighs approximately 386 g. 731,000,000 Silicon atoms weighs roughly 34 fg (femtograms) - using Avogadro's Constant and assuming I did my calculation correctly. That's a pretty huge space savings, even if you were only using binary computation.
...and then it got hot and separated the atoms too far from each other to interact. D'oh!