Powerful Nanoscale Computer Created

Very, very interesting:

The device is made of a compound known as duroquinone. This molecule resembles a hexagonal plate with four cones linked to it, "like a small car," explained researcher Anirban Bandyopadhyay, an artificial intelligence and molecular electronics scientist at the National Institute for Materials Science at Tsukuba in Japan. The machine is made of 17 duroquinone molecules. One molecule sits at the center of a ring formed by the remaining 16. The entire invention sits on a surface of gold.

Scientists operate the device by tweaking the center duroquinone with electrical pulses from an extremely sharp electrically conductive needle. The molecule and its four cones can shift around in a variety of ways depending on different properties of the pulse — say, the pulse's strength.

Since weak chemical bonds link the center duroquinone with the surrounding 16 duroquinones, each of those shifts too. Imagine, for instance, a spider in the middle of a web made of 16 strands. If the spider moves in one direction, each thread linked to it experiences a slightly different tug from all the others. Since duroquinone possesses four cones, each molecule essentially has four different settings. Since the central molecule can simultaneously control 16 other duroquinones, mathematically this means a single pulse at the machine can have 4¹⁶ — or nearly 4.3 billion — different outcomes.

The idea is to hook this new gadget up with other molecules — either copies of itself or different compounds other scientists have invented. For instance, researchers have created a host of machines just a molecule or so large over the last decade or two — motors, propellers, switches, elevators, sensors and so on. The new invention might offer a way to control all those other compounds to work as a whole. Indeed, Bandyopadhyay and his colleagues revealed they could hook up eight other such "molecular machines" to their invention, working together as if they were part of a miniature factory.

Bandyopadhyay added they could expand their device from a two-dimensional ring of 16 duroquinones around the center to a three-dimensional sphere of 1,024 duroquinones. This means it could perform 1,024 instructions at once, for 4¹⁰²⁴ different outcomes — a number larger than a 1 with 1,000 zeroes after it. They would control the molecule at the center of the sphere by manipulating "handles" sticking out from the core.

"We are definitely going to 3-D from 2-D immediately," Bandyopadhyay said.

That report is from Charles Choi at LiveScience.

Duroquinone is composed of carbon, hydrogen, and oxygen — specifically, C₁₀H₁₂O₂ — which suggests that it might become a key component of an early-generation nanofactory.

This tiny "brain-like" device could prove to be very big indeed.

UPDATE: BBC News has an article up that describes the development as a "tiny chemical 'brain' which could one day act as a remote control for swarms of nano-machines."

UPDATE 2: Commenting on the story for MSNBC, Alan Boyle writes, "If scientists can create assembles that can pass along instructions from one molecule to 16, then to 256, then to 4,096, and so on — pretty soon you could have nanofactories capable of churning out mega amounts of custom-designed molecules."

UPDATE 3: MSNBC also has a video posted that illustrates the nanoscale brain-like computer and even shows STM images of the device.

Mike Treder

Tags: nanotechnology nanotech nano science technology ethics blog

Comments

[kurt9]

Guys,

Hold your horses.

This device can only operate at temperatures near absolute zero. Like all of the other nano-mechanical stuff that has been done, none of it can work anywhere near room temperature.

I will be impressed (and convinced that it is real) once they can make such a system to work at room temperature. Until then, I maintain that the future is wet and squishy rather than dry and hard (think Blade Runner instead of Terminator).

[Mike Treder, CRN]

Kurt, the article at LiveScience states:

The device needs to be made in vacuum conditions at extremely cold temperatures — about -321 degrees F (-196 degrees C). Bandyopadhyay said it could be operated at room temperature, however.

So, yes, very low temps are required to build the device, but it can, according to the lead scientist, "be operated at room temperature."

[Brian]

The MSNBC article also mentions that the researchers believe within 18 months they expect to have 1024 of the devices working together. Mark Ratner indicated that Nanoink (maker of massively parallel dip pen lithography arrays) has one of the options for input and output [from macroscale to the devices and back]. A promising architecture and approach for more powerful computing and devices. they talk not just of computing but other kinds of devices.

[kurt9]

Mike,

I would like to know more about its operation. I would have to purchase the article for that. More likely, I can go to the university research library and read the article. If it works as advertised, it does represent a partial step to validating mechano-nanotechnology. However, full validation of this concept does require both fabrication as well as operation at room temperature.

[kurt9]

I do not want to be critical of the applications of this device or this technology. This could very well prove to be the next step in electronics technology. Certainly the information storage and manipulation capabilities of this goes way beyond anything envisioned by Drexler and company back in the day.

I noticed that this device was assembled at liquid Nitrogen temperature, not near absolute zero. Liquid nitrogen cryogenics is cheap, unlike the absolute zero regime. I can see a fab line using cryogenic vacuum environment to make these things commercially. In this sense, the manufacturing economies would be no different than conventional semiconductor or MEMS process technology.

[the oakster1]

I don't see your worry about cryogenic temperatures. Like materials acting differently on nanoscales, cooler temperatures at the nanoscales is not so hard or expensive. Of course, this device is still connected to the macroscale by an stm, but that is a problem that will surelly be soon corrected.

Drexler's daimondoid machiens work in vacuum - the coldest temperatures!

[Dan S]

I can't read original article (no subscription to PNAS), but it seems to be this device does not actually *computed* anything. It's not immediately apparent how to make switches or logic gates from it (is that architecture really Turing complete?).

[Dan S]

to oakster1: vacuum does not mean "coldest temperatures" - molecular machines in vacuum could still be heated by IR radiation and power provided to them also dissipates as heat. In fact devices from Drexler's Nanosystems are designed to operate at room temperature.

[Brian Wang]

I believe they are not yet computing. But there is a lot of related molecular electronics work where I believe suitable modifications can be made to the chemical strucure to enable computation and memory and other eight+ nanodevice types that they have identified to operate. It is a promising architecture which we can see how well it progresses over the next 18 months with the researchers goal of 1024 interconnected devices and collaboration on input and output methods.

[John Akers]

How does this stack up to quantum computers in terms of speed?

John Akers

[Brian]

Quantum computers can handle different kinds of algorithms. so for certain kinds of problems quantum computers would be superior to a molecular classical computer. A molecular classical computer like this would be could be a million times faster than the best petaflop machines that we have now. Estimates for the human brain are 100 teraflop to 20 petaflops.

http://vadim.oversigma.com/MAS862/Project.html

A large quantum computer with millions of qubits or more would be able to process certain problems faster.
If the qubits could allow say a problem to be solved as the square root of n * n where is the qubits versus the best classical algorithm which might be some exponential function or X**3 or something. then you could see when X**0.333 (where X is the flops of the molecular computer, change the functions based on the kind of problem) is less than n qubits.

[LarryLarryLarry]

This headline is disgraceful. You people are liars, plain and simple. There is no computer at all, much less a powerful one. You're all sitting around making up technologies which don't even exist yet, to be paired with this molecular arrangement, in order to make a "powerful computer".

Nothing good can come from wildly exaggerating and lying about technology when the facts are in plain view for all to see.

Pure hype and not even remotely convincing.