Slashdot Mirror
IBM Demos Single-Atom DRAM
An anonymous reader writes "A single-atom DRAM was demonstrated by IBM recently with a slow-mo movie of the atomic process of setting and erasing a bit on a single atom. Videos of atomic processes inside chips were not possible until now, leading to IBM's claim that its pulsed-STM (used to make the movie) will lead to a new atomic-scale semiconductor industry, and not just for memory chips, according to this EETimes story: 'The ultimate memory chips of the future will encode bits on individual atoms, a capability recently demonstrated for iron atoms by IBM's Almaden Research Center in San Jose, Calif., which unveiled a new pulsed technique for scanning tunneling microscopes (STMs). Pulsed-STMs yield nanosecond time-resolution, a requirement for designing the atomic-scale memory chips, solar panels and quantum computers of the future, but also for making super efficient organic solar cells by controlling photovoltaic reactions on the atomic level.'"
Does microwaving it make it go faster?
are we talking H or Uuq sized DRAM? because I don't want to be obsolete within a year.
Video explaining the process.
One atom ought to be enough for anybody.
(Sorry)
WALSTIB!
So this has not already happened (as the article implies) but is an idea for future development.
"Maybe this world is another planet's hell"
Aldous Huxley
A movie that you view? A movie that "moves you"?
I eat only the real part of complex carbohydrates.
Don't electronics become more susceptible as they become smaller? How much redundancy would be needed now that you only have a single atom to hold a bit of memory?
the processor on my computer runs on a single Atom already. I'm not impressed.
This gives new meaning to atomic writes.
Moore's Law.
It just means we'll start looking at sub-atomic particles as new storage methods...
This is a fantastic technical achievement. However, it has no meaningful direct link to ANY deployable technology. It is a measurement technique, and although the article does not say so, I'm sure it requires a temperature of somewhere below 1K, maybe below .001K. That is the only way they could be getting signals of these phenomena without getting swamped by thermal noise. All the stuff about single atom storage is boilerplate marketing hype. I assumes that they have a hot key to paste in how a new technology can be used for memory storage, or solar cells, or green technology or ...
Why is Snark Required?
I have enough trouble reading the print on MicroSD cards.
It doesn't apply, Moore's law only applies to transistor count on processors.
yes and no, just don't ever try to access the data...
You have to start worrying at quantum effects WAY before the scale which is currently in most modern computers. Interesting question though.
And while single-atom memory is an interesting feat, memory density isn't everything. It lets you get more capacity into less space, which can be nice. But if size was everything, I'd use my hard drive instead of my system memory and CPU cache. After all, it's easy to get a hard drive on the order of a couple terabytes while system memory is still typically on the order of a few gigabytes, and CPU cache is on the order of a few megabytes.
We Haskellers already use STM since a long time.
Call me when the sense hardware is only an atom per bit.
“Common sense is not so common.” — Voltaire
"... on the next Science"? Hmm... Maybe that means on the next cover of Science? Maybe a little editing could fix that? And maybe typographical errors like "moview" could be fixed? Perhaps by actually reading the summary, Timothy?
Put my fist through my alarm clock with its ding-dong death inside my ear. - The Blackjacks.
Slackers! Most atoms have way more electrons than that!
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
but also for making super efficient organic solar cells by controlling photovoltaic reactions on the atomic level
Where did that quote come from? All I saw was a vague mention of measuring the efficiency of solar cells. Not sure why they can't measure the efficiency of the ones they have already.
And while single-atom memory is an interesting feat, memory density isn't everything. It lets you get more capacity into less space, which can be nice.
yes, indeed, will let you get more capacity only when you fit the probe in the same space. For the time being, an STM is about this big.
As a research technique, is amazing. As an applicative discovery... a long way yet until the real-life consumer grade direct application will emerge (if ever)
But if size was everything
Hit the nail in the head here: latency and power consumption spring into my mind as well.
Questions raise, answers kill. Raise questions to stay alive.
For which we'll need new physics rather than new technology, unfortunately.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
yes, indeed, will let you get more capacity only when you fit the probe in the same space. For the time being, an STM is about this big.
I'd dearly love to know how they plan on locating any particular atom, let alone redirect the read/write head to it and only it.
Even if the atoms are arranged in an array, flat, how does an atom-scale read head know where it is pointing with sufficiently minuscule granularity? Do they intend to put markers on the surface nearby--oh no wait atoms. Well, they can probably have wires leading--oh no wait atoms. Well, maybe if they color--oh.
Well I guess they'll just have to have one atom surrounded by its own read-write logic, flash-style, and completely negate the whole point of having the actual storage on the atomic size. Oh no wait, that's not even what this research is about.
Seriously, I don't think this has much potential for engineering, as much as it may be clever science.
And by Moore's Law in 2 years it will be stored in a single proton.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
One baby step closer to Skynet. Long live our new immortal overlords.
Says we should skip atoms and go directly to cospatial nudged quanta.
You apply a voltage gradient. By some clever field manipulation, sensor placement, and computational wizardry you can address a spot in a three-dimensional lattice. It would probably work something like an MRI, if it had the bastard child of an STM.
Dynamic? The atom needs a refresh cycle?
Random Access? It is addressed by a row/column or similar structure?
Memory... ok, one bit, yes?
Doesn't "DRAM" strike anyone else as almost entirely not applicable?
I've fallen off your lawn, and I can't get up.
...which would quash Moore's Law. Hint: moving from hitting an engineering target of improving a well-understood process to inventing an entirely new one will change the rate of progress.
No-one has invented a buzzword to cover this yet, alongside all of the curve-jumping bullshit. I nominate progress-refraction as a suitable misunderstanding of a physical process for MBAs
Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
Next up on Slashdot, the RIAA is now demoing a single-atom DRM. Will the minuscule security it provides make a difference? Find out three redirects from the blog post we link!
I don't know if I should be up or down, but I definitely feel strange about your ideas.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
It actually might not. The thing about Moore's law is it's something of a self fulfilling prophesy. Semiconductor companies start designing CPUs before the processes that they're going to be using are available, with a reliable expectation that the process will be available by that time. The fabs aren't going to offer a process that nobody yet wants. So both of them assume a doubling in transistor count every 18 months. If this remains possible with some sort of subatomic process it will very likely continue at that rate.
And that single atom in a RAM cell doesn't count as a transistor?
If it performs the same functions, there is no reason to not apply the same law.
Oh, hey, look, we've got Transistors with THREE atoms.
ONE atom shouldn't be a problem. If it acts like a transistor, it's a transistor.
A transistor either acts as an amplifier or an electrical signal switch.
Therefore, a single-atom DRAM cell would be using single-atom transistors.
Want to try making this argument against someone *NOT* deeply involved in this industry?
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
I would bet more on tunnel-scanning versus anything else for the inner workings. It would be far easier to measure the blank spaces between atoms instead of the atom itself.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
TFA referes to "capability recently demonstrated for iron atoms by IBM's Almaden Research Center in San Jose, Calif"
I remember that place. It used to be one of the biggest research parks in the are. Then an few years ago it became Hitachi, say "Inspire the Next", research after Hitachi bought that division of IBM many years ago. I think they shut it down a few years ago, because it all became tall weeds, and now a brand new Lowe's store emerged in its place.
BTW, someone should collect slogans of Japanese companies: "Inspire the Next", WTF does that mean?
don't cut it off www.mgmbill.org
The article is about IBM's new pulsed STM tech, and notes that "it may enable atom-scale memory in the future". They did NOT demonstrate single-atom DRAM.
What a depressingly stupid machine.
And that single atom in a RAM cell doesn't count as a transistor?
No, it counts as a capacitor.
Want to try making this argument against someone *NOT* deeply involved in this industry?
Aren't we already doing that?
This is so horribly old hat. I mean, we know atoms for ages now so IBM needn't be smug about them. IBM, stop wasting our time and give the world a call as soon as single Higgs boson DRAM is available to retailers!
Now, what was I doing again? Yes, studying Xiph' Digital Media Primer For Geeks and appreciating sample videos with scarcely clothed women.
I hadn't the slightest objection to his spending his time planning massacres for the bourgeoisie... (P.G. Wodehouse)
Actually, Moore's law applies to the number of components on an integrated circuit (for a fixed cost). The original paper makes no mention of processors, and only talks about transistors as an example of the components you put on an IC. It directly applies to RAM, and any other kind of IC, because it's talking about process technology not about what you do with the ICs.
I am TheRaven on Soylent News
Nope, there are lots of subatomic places where we could store information. The spin on electrons, for example. Another simple alternative would be to use the photovoltaic effect to move electrons up and down energy levels. Fire a photon at the atom to move the electron up one energy level, measure its charge to find the current one.
Of course, when I say simple, I mean in terms of theoretical physics. In terms of engineering, it's quite the opposite.
I am TheRaven on Soylent News
"Seriously, I don't think this has much potential for engineering, as much as it may be clever science.'
Too bad your high UID pretty much guarantees your ignorance in this field, I've got several workable applications already in mind for this. Hello, 300+ irradiant lumens per watt output capability, good-bye HID lamps.
I think you need to study more quantum sciences before talking any further.
~Director of Research, EcogroLED USA.
Grow mushrooms instead, no need to study quantum sciences, all you need is a damp closet and a pile of horse-shit.
Single-atom DRM.
I wouldn't have considered either spin or energy level of the electron subatomic.
I was wondering where you were going to store information on a quark, for example.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
It's going to take a lot more then one atom to connect the memory and make it do something useful.