Valleytronics Offers Laser-Based Computing at Femtosecond Speeds

​Someday in the not too distant future our current dependency on electric charge (pun fully intended) for computing tasks will seem like cave painting. Here we are in 2015 encoding just the smallest bits of information using enormous buckets full of positive and negative paint, when all it really takes to represent information is the smallest quantum drip.

Physicists at Berkeley Labs (and elsewhere) are honing on a new information technology that's poised to become one of those desired quantum drips—a superfast, low-power method of information encoding dubbed "valleytronics." In a way it's analogous to the quantum spin number of spintronics, except valleytronics stashes its information in what's known as the quantum valley number, or wave quantum number.

Also like spintronics, storing information in a valleytronics scheme really only requires a single particle. Contrast this with the electric charge we usually depend on now. It's volatile, fragile. Information in conventional electronics is sorted using states of high and low electric current (the paint buckets), and keeping that information intact and then actually using it requires loads of power. All of the added juice needed to do actual computing using electric charge means heat.

Heat adds up, eventually reaching a tipping point where the heat needed to maintain higher and higher computing speeds becomes a barrier to actual computing. Electronics are then self-limiting.

This is why engineers and physicists are currently engaged in a broad quest to identify and refine alternate information encoding schemes. At stake is nothing less than Moore's Law, that famous (or infamous) promise of ever-increasing/ever-doubling computing speeds.

So: valleytronics. Understand that a given electron is described by four different numbers. The first one, n, corresponds to the energy state an electron can be found in—in other words, which of the ascending orbitals or electron shells it currently calls home. Closest in to the atom's nucleus is 1, and as we move outward from there, the value of n increases as positive whole numbers. (The other three quantum numbers correspond to angular momentum and spin.)

Electrons, as they wobble around the atomic nuclei in their respective orbitals, have a state of lowest energy, which is called a ground state. This is the innermost shell and as we jump up from shell to increasing shell, the electrons will have progressively higher energies.

The ground state isn't some fixed value, however; it's not an absolute "0." (The exception would be an electron at 0 degrees Kelvin, if that were a real thing.) Valleytronics depends on tweaking this minimum energy state, creating a "valley" for the particle to rest in. This can be done in crystals, where there's no sitting still for an electron (no state of 0 momentum). So it's always jittering around in an energetic, overcaffeinated valley.

​

Some materials can have multiple different valleys and this is where we can encode information. A diamond, for example, has six possible valleys of minimum energy. The Berkeley team is beyond diamonds though, instead using a two-dimensional semiconductor material that can be controlled using laser pulses to simulate tiny magnetic fields.

One way to manipulate the ground state energy of a particle is by putting it into smaller and smaller boxes. Imagine a prisoner pacing back and forth in a jail cell with walls that are slowly closing in, with every bit of reduced space (increased confinement, technically) translating to more energetic pacing. It's like that only for a subatomic particle.

The key to getting the electrons to jump back and forth between energy states (two distinct valleys in the Berkeley Lab's 2D material) is known as the optical Stark effect. The general Stark effect, which the team approximated using laser pulses, boosts and limits electron energy states by the application of an electric field pointing in some direction, which you could just imagine as a giant arrow. By the tip of the arrow, electrons will fall into a higher valley, and by the tail, the electrons will fall into a lower valley.

"This is the first demonstration of the important role the optical Stark effect can play in valleytronics," Feng Wang, a Berkley condensed matter physicist and the coauthor of a new paper describing the work, offered in a statement. "Our technique, which is based on the use of circularly polarized femtosecond light pulses to selectively control the valley degree of freedom, opens up the possibility of ultrafast manipulation of valley excitons for quantum information applications."

How fast? 50 femtoseconds. To give some idea, switching speeds with conventional electronics setups are usually in the nanosecond range, so it's a difference of a 1 with nine zeroes in front of it (nano) and a 1 with 15 zeros in front of it (femto). That's absurdly fast. Quantum fast.

Indeed, where it might count the most is in quantum computing, where information is stored in single-particle (or arrays of entangled particles, for parallel computation) superpositions of different states. So: we might a lone particle cruising around a 2D semiconductor within two valleys at the same time. Try that with buckets of paint.