San Jose (CA) – IBM researchers today said that they were able to probe magnetic anisotropy in individual atoms, potentially opening the development of a new kind of data storage devices. In another announcement, the company unveiled a single-molecule switch that is expected to lead the way toward smaller, faster and more energy-efficient semiconductors.

According to IBM, technologies described today (with further details being provided in two papers scheduled for publication in Science tomorrow) could be significant milestones in developing single-atom data storage devices and molecular computers.

In its first report, IBM scientists describe progress in probing a property called magnetic anisotropy in individual atoms. This fundamental measurement is believed to have a substantial impact on technology as it determines an atom’s ability to store information. Previously, nobody had been able to measure the magnetic anisotropy of a single atom. Down the road, the company believes that it may be possible to build structures consisting of small clusters of atoms (or individual atoms) that could store magnetic information.

And, according to IBM, this technology could provide lots of storage capacity: "Nearly" 30,000 feature length movies or the entire contents of YouTube could fit in a device the size of an iPod, the firm said.

The development could also lead to "new kinds of structures and devices that are so small they could be applied to entire new fields and disciplines beyond traditional computing."

In their second report, IBM researchers outline the first single-molecule switch that is claimed to be able to operate flawlessly without disrupting the molecule's outer frame, which is a step toward building molecule-sized computing elements that are expected to be substantially smaller, faster and use less energy than today's computer chips and memory devices – IBM is envisioning these devices to be as fast as today's supercomputer

Switches inside computer chips act like a light switch to turn the flow of electrons on and off and, when put together, make up the logic gates, which in turn make up electrical circuits. Having ever smaller switches allows the circuits to be shrunk to ever smaller sizes, making it possible to pack more circuits into a processor and boosting speed and performance, the company said.

IBM did not indicate when these technologies could be commercially available.

Good so?

Aug 30, 2007 16:58     

Make something simple that works. Any proof of concept. Give the single-molecule switch to AMD also, they need it.
Erric report abuse vote down vote up Votes: +0

thats cool news!

Aug 30, 2007 18:30     

I wish they would also indicate the other companies that participated in the research, that would be cool too... As far as when we'll see this? My guess is at least 5 years, there are very serious problems in (all) technologies that challenge the norm that must be overcome I'm sure here. As well as an entirely new manufacturing process here I'd imagine... all speculation of course. More info!!
Kyle Schochenmaier report abuse vote down vote up Votes: +0

RE: thats cool news!

Aug 30, 2007 22:44     

I'd say you're dreaming if you think 5 years; more like 10-20 years. This technology may be a long way off but it's always great to see a long term vision. Thanks for the read!
ChOOkiE report abuse vote down vote up Votes: +0

come on Johnny Mnemonic!...

Aug 31, 2007 01:14     

come on Johnny Mnemonic!
shawn report abuse vote down vote up Votes: +0

If we were stuck on large super computers, mainfram...

Aug 31, 2007 02:16     

If we were stuck on large super computers, mainframes, or even todays desktops and laptops, I'd say there would be no need to push it, and the 10 to 20 years would be expected, but the mobility people are searching for, and the power consumption of that mobility will push this technology advancement so fast that a 5 year prediction is no surprise for me at all.
Virtualban report abuse vote down vote up Votes: +0

I/O problem

Aug 31, 2007 02:38     

While there has been a lot of research and many theoretical investigations on molecular switches and the like, the big problem remains unsolved yet: how to transfer the information from one order of magnitude to the other (so from the macroscopic world to the nanoscropic world and vice versa). You can't just plug a cable into a molecular switch and as of now there's nothing you can do to actually store the information in molecule and read it without an electron microscope or atomic force microscope etc. I also wonder how they want to put an experiment to detect magnetic anisotropy in a disk drive... Once these problems are solved though, it'll be great. :)

Johannes report abuse vote down vote up Votes: +0

entire new fields

Aug 31, 2007 03:57     

maybe they can make a computer that could be installed in your body through an injection, and be powered by your body's heat. next, it will connect to your brain and control it. :D you'll never see it happen
ceteras report abuse vote down vote up Votes: +0

Programmer / Analyst

Aug 31, 2007 05:01     

Pretty amazing stuff! Once they can make the technology commercially available, it will be interesting to see what happens to the existing companies that make their living selling storage devices (all types). If you now have a medium that can store off "tremendous" amounts of data as indicated, with transfer rates making it also viable for computer memory replacement, and it’s reasonably priced for the consumer, why would technology manufactures use anything else in their products? Can't wait till it hits the commercial market.

Gerald Knauss report abuse vote down vote up Votes: +0

The tech to read the atom magnetic state takes up t...

Aug 31, 2007 08:50     

The tech to read the atom magnetic state takes up the space of a large filing cabinet if not larger. This won't be ready for something the size of an laptop or even a home system for 30 - 40 years. It is nice the possibilities though.
Frank report abuse vote down vote up Votes: +0

History-Present

Aug 31, 2007 09:09     

Wow this is truly awe inspiring. But to those that said 10-20 and 30-40 years try again. I would venture a guess at 5-10 and Im sure most would agree if they were to look to history and the trends of "doubling" our technology.
Justin report abuse vote down vote up Votes: +0

'The tech' size issues.

Aug 31, 2007 09:15     

It used to be that lasers took up that much space, too. Computers used to fill rooms. I think 30-40 years is overly pessimistic. 5 years? Maybe for supercomputer data storage. That seems too soon for regular consumer availability. Take a look at the slow adoption of SSDs. Many people aren't willing to pay the premium price to adopt better tech early. I give it 10 years.
kw report abuse vote down vote up Votes: +0

year predictions

Aug 31, 2007 09:22     

it's understandably cautious to say that this technology won't be around commercially for 30-40 years, but 30 or 40 years ago they said some of things we have now wouldn't be a reality for centuries (think mapping of the human genome for example) So while 5 years may be a premature, 30-40 is potentially as unrealistic.
wearyhobo report abuse vote down vote up Votes: +0

can it get any smaller

Aug 31, 2007 10:07     

this will be the end to shrinking the transistor to increase performance. at least when u get down to things this small u won't be able to get much smaller.
hoho report abuse vote down vote up Votes: +0

Year Predictions

Aug 31, 2007 10:45     

As was said inthe History-Present comment, looking at past trends of technology development the technology will easy be available in 5-10 years
Mike report abuse vote down vote up Votes: +0

Re:I/O problem

Aug 31, 2007 11:32     

I would say that's the limitation with our thinking. Why can we think of a cable only to transfer the information; some news ways can also emerge out which would prove more faster than the cable. May be like the way a electron microscope works.
Amit Pal report abuse vote down vote up Votes: +0

5-10 years....? Put down the pipes....

Aug 31, 2007 11:50     

Some of you need to do a little more research on the effort needed to bring this idea into a viable product. While most major companies are having trouble passing the 20-40nm hurdle within the next 5 years, you expect them to magically shrink an order of magnitude, over 1000 times smaller?. Technology does not advance that way, and I believe, that before this process is perfected, another aproach with better techniques will be introduced. Look at holographic memory and the advancements over the past 15 years. By most of the assumptions, we should already have petabyte sugar cube storage chips.

EU critic report abuse vote down vote up Votes: +0

Re: EU critic

Sep 01, 2007 02:38     

This is not an evolutionary step, so technology advances (revolutionary) this way. Yes, there might be problems of read-write, but I believe those are smaller compared to the actual data storage :) If there will be need (and already there is), research will be made and this revolution will Just Arrive.
Virtualban report abuse vote down vote up Votes: +0

Innovation

Sep 01, 2007 21:25     

Another step forward for innovation and technology.
Tech report abuse vote down vote up Votes: +0

Smaller and Smaller

Sep 01, 2007 22:24     

the transistors can get smaller and smaller. who said we can't go into the picometer femtormeter field? from what i have seen in the last decade or so, we have come so far so fast, and eventually we won't be able to double technology, but right now i don't see us slowing down. and i don't see this atomic data storage being around for another 15 years or so. there is a lot of crap you have to overcome to do that: trememdous amount of money to study it, creating some way of testing it, creating a reasonably priced technology that will allow us to read and write to single atoms.

RaptorXT report abuse vote down vote up Votes: +0

People that say 10-20- or 30-40 years Yeah Right

Sep 01, 2007 23:58     

I dont think it will take that long or there will be no reson to even spend the time on it because 1.5" hard drives will be the ones in our desk top along with mabe 30 TB on a single 1.5" hard disk and on a .75" hard disk there will will be like 15 or 20 Tb on that disk which will be in our laptops so i dont know iff it will take 40 years we will probably see it come out to major industries in about 5-7 years We have such powerful computers out know that they are building themselves and they will also put together that atoms chip

Jeff report abuse vote down vote up Votes: +0

I wonder.

Sep 01, 2007 23:59     

will this be the end of compacting magnetic information or will they figure another way?
Anonymous report abuse vote down vote up Votes: +0

Server Farms

Sep 02, 2007 07:21     

Would be interesting to see how Server Farms will look like in future, a briefcase size!!!
Prajith Nambiar report abuse vote down vote up Votes: +0

Moore's Law

Sep 03, 2007 06:11     

Alright, lets look at this whole issue logically. First, why has Moore's Law proven out for so long? 1. Through each improvement of tech comes the ability to use better systems to design the next generation of systems. 2. Every time you design the next generation of systems, it becomes more and more difficult to overcome the laws of physics. These two issues cancel each other out so that you don't get the exponential advance in the technology curve. So, lets do some math. Assume that when the author said 30,000 full length feature movies, he was talking high def. It's not a highly accurate gauge, but for ease, I'll use 10GB per movie. 10GB X 30,000 = 300TB. So, it could be anywhere up to a Petabyte or so if you had a little more high def to it. Back in 1997, the biggest drives I could get when I built my first PC were about 9GB. Today, we are just about to get the first 1TB single drives. So, that is roughly an increase of 100 times in a decade. And, we probably want to hit a target of 500 times that just to make sure we can fit all of our 30,000 movies (that we didn't download from bittorrent lol) onto our drive. So, according to the math, it is possible that such a tech could show up just around the start of the 2020's, if the current increase trend continues. Just remember, Intel is going to struggle just to get their 80 core chip or so out in 5 years, so having consumer molecular compuers at that time just doesn't fly. Maybe, I could see the military having secret ones in about 8 years, but they have their work cut out for them. Some of you are saying 5-10, some are saying 20 , I think a good compromise would be about 15 years. At any rate, it is certainly an interesting time to be alive. IBM invented the hard drive and many other things. I don't see them as great business people, but they are tremendous researchers. I think if any company can get this technology to work, it will be them.

Stevo report abuse vote down vote up Votes: +0

re:entire new fields

Sep 04, 2007 10:16     

Hmm... replacing outdated equipment gets REALLY interesting. Anyway, some of the folks sounding off on Moore's Law should reconsider what it really means. Moore's "Law" is just an observation of tendencies within a given technology (transitors to be specific, although it seems to hold for other types of tech such as HDDs), not a physical law. A paradigm shift to a new type of technology for logic gates or bit storage COULD (notice I don't say WOULD) side step the "law" just as handily as the creation of transistors side stepped the limitations of tubes.

Jim L report abuse vote down vote up Votes: +0

Re: Moore's Law

Nov 03, 2007 16:41     

Back in 1997, the biggest drives I could get when I built my first PC were about 9GB. Today, we are just about to get the first 1TB single drives. Ok you mean that your going to buy your first 1TB drive right, because 1TB drives have been out for awhile now, and soon to hit 2TB. The 2TB drives are expected Q1 of 2008 and might go as high as 5TB by the end of 2008. Moore’s Law refers to transistors. And basically states that the transistor counts will double every 18 months. But most insiders agree that Moore’s Law will brake down in the near future do to the fact that you can only shrink a transistor so much and have it still be useful. In a recent interview with the creator of Moore’s Law he even admitted that the end is in site.

thomfrost report abuse vote down vote up Votes: +0