A Memory and Logic Device
HP researchers demonstrate logic with memristors
Source: “ ’Memristive’ switches enable ‘stateful’ logic operations via material implication”
R. Stanley Williams et al.
Nature 464: 873-876
Results: Researchers at Hewlett-Packard have shown that nanoscale circuit elements called memristors, which have previously been made into memory devices, can perform full Boolean logic, the type used for computation in computer processors.
Why it matters: Memristor logic devices are about an order of magnitude smaller than devices made from transistors, so they could pack more computing power into a given space. Memristor arrays that perform both logic and memory functions could eliminate the need to transfer data between a processor and a hard drive in future computers, which would save energy.
Methods: HP researchers fabricated memristors of various sizes on a silicon substrate by growing metal nanowires, coating them with titanium dioxide, and topping them with another series of metal nanowires oriented perpendicular to the first group. A memristor is formed where the two layers of wires cross. Each wire was connected to some test circuitry that the researchers used to bring current into the system. Using this test system, the researchers showed that a pair of memristors could serve as a logic gate: it processed information by switching one of the memristors on or off to create a 1 or a 0, depending on the initial states of the two elements. The memristors could also serve as a “latches”–that is, they retained data, “remembering” whether they had recently switched to a 1 or a 0.
Next steps: The basic material properties of the metal oxides used to make memristors are still not well understood, so it’s not clear whether the devices will be as reliable as silicon transistors. HP Labs researchers are working toward introducing a memory product based on memristors in 2013, and researchers will learn more as they develop facilities to make them.
A simple, inexpensive treatment could reduce the cost of solar power
Source: “Efficient Black Silicon Solar Cell with a Density-Graded Nanoporous surface”
Howard Branz et al.
Applied Physics Letters 95:123501-123503
Results: A simple chemical technique can create a highly antireflective surface on silicon solar cells. The new method for making this so-called “black silicon” results in cells that convert 16.8 percent of the light that hits them into electricity, which is comparable to the efficiency of many commercial solar cells. It’s a significant improvement over the previous record for solar cells made of black silicon, which was 13.9 percent.
Why it matters: The technique could make crystalline solar cells, the most common type, cheaper to make, because it is less expensive than producing the antireflective coatings now used to keep photons from bouncing off the cells and going to waste. Previously developed methods for making black silicon may be impractical for large-scale manufacturing because they are more complex or involve slow and costly equipment. The new research demonstrates that the antireflective surface can be readily made using equipment already on hand at solar-cell factories.
Methods: The researchers submerged a silicon wafer in an acidic solution containing trace amounts of gold. Chemical reactions generated gold nanoparticles, which then catalyzed reactions that etched holes of varying depths into the wafer. This created a porous structure that blurs the boundary between the surrounding air and the bulk silicon, reducing reflection.
Next steps: The researchers are working to increase the cell efficiencies further and performing more detailed calculations to determine how the process will affect the cost of solar power.
Going bald? Lab-grown hair cells could be on the way
These biotech companies are reprogramming cells to treat baldness, but it’s still early days.
Meet Altos Labs, Silicon Valley’s latest wild bet on living forever
Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.
Meta’s new learning algorithm can teach AI to multi-task
The single technique for teaching neural networks multiple skills is a step towards general-purpose AI.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.