Atomic Scale Memory

maddugan writes "Technology Research News is reporting that researchers from the University of Wisconsin at Madison have put the theoretical to the test by using single silicon atoms to represent the 1s and 0s of computing. This is equivalent to storing the contents of 7,800 DVDs in one square inch of material."

Comments

Feh. Only ONE bit per atom...

[Mad Bad Rabbit]

Only half joking: Researchers at U.Michigan hope to
store up to 10 bits per atom, by using Rydberg states.

http://www.aip.org/enews/physnews/1999/split/pnu 42 9-2.htm

>:K

Eh?

[ShooterNeo]

This isn't actually very useful : what we want is atomic scale logic gates, not data storage. In fact I'd venture to say that this technology is NOT what we will be using in the future for extremely dense memory. Why? Because its 2 dimensional and requires an independent readout head (that is MECHANICAL). Making it work anywhere but a vacuum may be impossible. (though that is not a real problem : making a disk drive that has an internal vacuum is quite feasible) A solution that is thousands, even millions of times faster would be a system that reads itself : i.e. a 3 dimensional array of logic gates to form a molecular version of ram. In addition, you could cram far more bits per gram of material used for the media. (I can't say per square inch because that would be misleading) In addition, storage capacity is not what our computers need more of : its performance (especially in accessing all those gigs of storage).

Re:Eh?

[athlon02]

You forget, we don't just need more performance, we need smaller memory... If these people could make a type of RAM or non-volatile memory from this stuff then you could store your entire DVD collection, your entire CD collection, 1000's of photos from a 2 or 3Mp digital camera, tons of setup files for commonly used apps you have, etc, etc. all on a postage stamp sized media and still have plenty of room left over. Personally, I've been waiting for something like that for a LONG time. I mean, combine something like that with a PDA, cell phone, and firewire camera all at once, and that'd be quite an interesting device, that would be as small as the Sony Clie's I've seen, but rival the space of my desktop machine with an 80GB Seagate Barracuda IV, possibly with better transfer speeds too!!

Yes, it's a lot of forward thinking and so forth, but I await the day when such things are common place and reasonably priced.

And as for these guys working on this project, more power to them, if they can do it, albeit, I'm not holding my breath for it to happen any time soon.

Reminds me of...

[teetam]

a juvenile idea/dream I had when I was a small kid to use the electron's spin (+half or -half) to store the same binary information. A single atom could store a lot more bits this way.

Now, if I could only do it!!!

Atomic scale computers?

[too_bad]

This article [http://www.eetimes.com/at/news/OEG20020319S0029]
talks about using a cell matrix which configure their neighbours at run-time,
something like the game-of-life or a more generic turing machine sort of thing.
This has lot of applications, including a highly programmable FPGA which
is very simple to fabricate or even complicate circuiry.

What really attracted my attention was the passage at the end:
>Cell Matrix has been working with nanotechnology groups, hoping to forge a
>new computing substrate from some type of atomic-level fabrication technique.
>Macias was impressed with work at Hewlett-Packard Laboratories in which a
>matrix defined by erbium disilicide wires that address rotaxane molecules
>has been proposed as an atomic-level route to massively dense FPGAs.

Could this new research be an answer to these people ? Probably combining
the two technologies, not only do we have a massive memory-device, but
a massive computing device : Imagine an FPGA (or an ASIC) with a million
times more density!

karma whoring

[flollywebfrog]

In 1959 Richard Feynman said that all the information accumulated in all the books in the world could theoretically fit in a cube 1/200th of an inch on a side.

You can read the transcipt of the speech from when he made that prediction.

Feynman worked on developing the atomic bomb, he won a nobel in physics and is known as much for his scientific research as for his story telling.

Theoretical density issues

[ssyladin]

The article states that the storage capacity of this new material/system is about equal to 7800 DVDs. Just to get nit-picky and technical, and to educate people some, this number will probably be lower.

When DVDs are burned and read, you don't simply read raw data off. The information is, of course, encoded. The DVD (and CD for that matter) specification says to use Reed-Solmon encoding. Saving the long math, RS encoding is about the most advanced error-correcting scheme that can be implemented in low-cost hardware today. By encoding data this way, your DVD (or CD) can become fairly scratched, but still play. RS protects against multiple-point errors. However, there is a price to pay - for every ~33k byte block on a DVD, almost 5K bytes are used in the parity checks for the DVD. See this file for more gritty details about DVDs. This means your 4.7GB DVD really holds about 5.48 GB of raw data.

Now, why is this relevant? Harddrives use their own error correcting schemes too. Manufacturers have the luxury of creating their own encoding systems since they're the ones that provide the read/write mechanisms. You can't pull the platter out of one harddrive and stick it in another. Hard drives typically use CRC (cyclic redundancy check) encoding schemes. I know you have all gotten CRC errors on a floppy way back when - that's what it stands for. Anyway, CRC is much less efficent when you compare the protected data to parity information ratios. While I wasn't able to pull the actual numbers from the Internet or my old math books, you can find a discussion and sample math here.

When you boil it down and relate all this information to our magical harddrive, the maximum usable density of the data would hover between 85%, or 6630 DVDs/in^2, to 60%, a measly 4680 DVDs/in^2, of the listed capacity. This is all assuming that the ideal lab conditions are maintained for a consumer level product.

As always, beware what the numbers tell you. However, if this can fly, then it would be an awesome step forward. Once you get Windows 2010 installed, you might even have a few Gig to play around with!

seen this before

[aaw]

Reading through the article, I noticed the following tidbit:

Eventually, "instead of moving them, we [picked] up the atoms," using a scanning tunneling microscope, he said.

This reminded me of an invention touted by one of my old professors at MIT. Low and behold, a search at the US patent office turned up this patent filed in 1994 for a high density dimer memory device which utilizes a scanning tunneling microscope

. . .which in response to being placed in intimate contact with the lower atom of a selected dimer results in an interatomic bond which accommodates pulling the lower atom upward and thus pulling the upper atom downward so as to effect a change in the dimer angle.

Looks like this isn't so novel after all.

Re:that's it?

[glwtta]

From what I understand, the Human Genome, though it represents a massive amount of data, is also highly redundant with huge sectons of 'legacy code' that doesn't seem to have any function or purpose.

The actual genetic code is not much at all - roughly 3 billion base-pairs, considering you only need two bits for a basepair, you can fit the whole thing on a CD. When they say that things are "redundant" or "legacy" what they really mean is that they have no idea what is going on - contrary to what some people will have you believe, very little (comparitevly) is known about the genome.

I'd be interested in knowing approximately how much actual information is functional in numerical terms. If one knew that, one could say exactly how much data it takes to create a human being.

Like I said, the actual genetic code is very small (~750 MB), the next level of complexity is annotation on some of its function and variation (essential to any sort of understanding of what it does) and this amount of data isn't a set size, we (a small biotech) have roundabout a terabyte of it, Celera has 100TB

The real complexity with humans starts around the protein stage, and proteomics is far younger than genomics. Once we start studying/simulating biological processes on the cell-wide scale, then we'll get into the astronomical numbers for storage and computing power, which will see use for "atomic scale" technologies.

Anyway, with humans it's not really the number of bits packed into small spaces that's impressive, it's the amount of information packed into those bits.