Shrinking the truly random

Computer chip makers continuously strive to pack more transistors in less space, yet as the size of those transistors approaches the atomic scale, there are physical limits on how small they are able to make the patterns for the circuitry.

Now, taking advantage of a germanium wafer coated with a layer of virtually pristine graphene -- a sheet of carbon arranged just one atom thick -- a team of engineers from the University of Wisconsin-Madison and the University of Chicago has devised a simpler, reproducible and less expensive manufacturing approach using directed self-assembly.

Directed self-assembly is a large-scale, nano-patterning technique that can increase the density of circuit patterns and circumvent some limitations of conventional lithographic processes for printing circuits on wafers of semiconductors such as silicon.

Electrical engineer Zhenqiang "Jack" Ma and materials engineer Michael Arnold of UW-Madison, chemical engineer Paul Nealey of the University of Chicago, and their students published details of the advance in the Aug. 16 edition of the journal Scientific Reports.

Their work could mean a boost in functionality for semiconductor electronics and in capacity for data storage.

To achieve the incredibly tiny size required for the circuitry in future semiconductor electronics, manufacturers are developing directed self-assembly, which enables the fabrication of intricate, perfectly ordered polymer patterns for circuitry.

For directed self-assembly, the researchers use conventional chemical techniques to define a pre-pattern. When chains of molecules known as block copolymers self-assemble on the pre-pattern, they follow the pattern to form well-ordered features.

The researchers' new method is much faster, and reduces the number of steps in the process to just two: lithography and plasma etching.

In the first demonstration of their technique, the researchers used electron beam lithography and a mild plasma etching technique to pattern one-atom-thick graphene stripes on a germanium wafer. Then they spin-coated the wafer with a common block copolymer called polystyrene-block-poly(methyl methacrylate).

When heated, the block copolymer self-assembled completely in just 10 minutes -- compared to 30 minutes using conventional chemical patterns -- and with fewer defects. The researchers attribute this rapid assembly to the smooth, rigid, crystalline surfaces of germanium and graphene.

Their new method takes advantage of a phenomenon called density multiplication. The researchers used electron beam lithography to first create a larger master template with sparse patterns that guide the orientation of their block copolymers.

When they directed the block copolymer to self-assemble, it did so in a way that enhanced the resolution of the original template -- in this case, by a factor of 10. The best previous enhancement by density multiplication was a factor of four.

While the stripe pattern was a simple demonstration of their technique, the researchers also showed it works with more architecturally complex or irregular patterns, including those with abrupt 90-degree bends.

"These templates offer an exciting alternative to traditional chemical patterns composed of polymer mats and brushes, as they provide faster assembly kinetics and broaden the processing window, while also offering an inert, mechanically and chemically robust, and uniform template with well defined and sharp material interfaces," says Nealey.

The technique enables them to combine the uniformity and simpler processing of traditional "top-down" lithographic methods with the advantages of "bottom-up" assembly and greater density multiplication, and offers a promising route for large-scale production at significantly reduced cost.

"Using this one-atom-thick graphene template has never been done before. It's a new template to guide the self-assembly of the polymers," says Ma. "This is mass-production-compatible. We opened the door to even smaller features."

Random number generators are crucial to the encryption that protects our privacy and security when engaging in digital transactions such as buying products online or withdrawing cash from an ATM. For the first time, engineers have developed a fast random number generator based on a quantum mechanical process that could deliver the world's most secure encryption keys in a package tiny enough to use in a mobile device.

In The Optical Society's journal, Optica, the researchers report on their fully integrated device for random number generation. The new work represents a key advancement on the path to incorporating quantum-based random number generators -- delivering the highest quality numbers and thus the highest level of security -- into computers, tablets and mobile phones.

"We've managed to put quantum-based technology that has been used in high profile science experiments into a package that might allow it to be used commercially," said the paper's first author, Carlos Abellan, a doctoral student at ICFO-The Institute of Photonic Sciences, a member of the Barcelona Institute of Science and Technology, Spain. "This is likely just one example of quantum technologies that will soon be available for use in real commercial products. It is a big step forward as far as integration is concerned."

The new device operates at speeds in the range of gigabits per second, fast enough for real-time encryption of communication data, such as a phone or video calls, or for encrypting large amounts of data traveling to and from a server like that used by a social media platform. It could also find use in stock market predictions and complex scientific simulations of random processes, such as biological interactions or nuclear reactions.

Shrinking the truly random

The random number generators used today are based on computer algorithms or the randomness of physical processes -- essentially complex versions of rolling dice over and over again to get random numbers. Although the numbers generated appear to be random, knowing certain information, such as how many "dice" are being used, can allow hackers to sometimes figure out the numbers, leaving secured data vulnerable to hacking.

The new device, however, generates random numbers based on the quantum properties of light, a process that is inherently random and thus impossible to predict no matter how much information is known. Although other researchers have developed quantum random number generators, they have all been either larger or slower than the device reported in the Optica paper.

"We have previously shown that the quantum processes taking place exhibit true randomness," said Valerio Pruneri, who led the collaborative research effort. "In this new paper, we made a huge technological advance by using a new design that includes two lasers that interfere with each other in a confined space. This makes the device smaller while keeping the same properties that were used in the past experiments."