Practical, Scalable Quantum Logic Gate Demonstrated
Alexander E. Braun, Senior Editor -- Semiconductor International, 5/20/2008 7:57:00 AM
 |
The research team, led by Stanford Assistant Professor Jelena Vuckovic, produced the gate by introducing an InAs quantum dot within an optical cavity in a photonic crystal and a GaAs chip precisely drilled with holes to give it the capability to trap photons and have them interact with the quantum dot.
According to Vuckovic, there has been similar work in atomic physics for the past two decades. “However, in those experiments, an atom must be trapped inside a large resonator,” she said. “This requires a complicated apparatus, and trapping times for atoms are on the order of a couple of seconds, so this isn’t really a very practical system that one can use afterwards for building something on a larger scale.” The group began work on a semiconductor equivalent of a quantum logic gate five years ago. Instead of an atom, the researchers used a quantum dot, a ball of a small bandgap semiconductor embedded in a large gap semiconductor. Vuckovic describes it as an artificial atom because the electron trapped inside of the dot can only occupy discrete energy levels, much like the discrete orbits of an electron in an atom. Because a quantum dot is based on semiconductor technology, it made possible the implementation of an optical resonator around it using standard semiconductor microfabrication techniques.
 |
| When two photon streams are incident on a cavity, one affects the other, resulting in communication of data. (Source: Stanford University) |
The experiment showed that two photons at the same frequency and amplitude (identical bits) result in a 12.6° phase shift. Because a 180° phase shift is necessary to fully flip the signal polarization given the control bit, the interaction must be repeated multiple times. If photons can be used at different frequencies and powers, larger phase shifts are possible.
While the research group’s target is to produce an array version of the system, research still remains to be done. Current coupling efficiency into the cavity is ~2-5%, which is far from ideal. While coupling two cavities via a waveguide would provide considerably higher efficiencies, there are problems associated with the quantum dots themselves, one being that they are not identical. As Vuckovic explained, “If we have multiple nodes with quantum dots, we must develop techniques that will enable us to tune them into resonance with each other to allow the production of scalable logic. We’ve developed a number of methods to tune quantum dots, as well as cavities, in resonance with each other on the chip.” Although results so far achieved are satisfactory, to actually do logic several of these elements must be scaled, making coupling efficiency crucial.
Were a quantum computer to be produced, one should not expect to run Microsoft Office on it; paradoxically, it would offer no speed or processing benefit for such tasks. There are very specific applications for a quantum computer, such as unsorted database searches. For instance, if you want to find a drug that will bind to receptor A but not to B and C, this presents a numerically tedious problem that could be attacked on such a computer with two key algorithms — the search algorithm and the factoring algorithm; other problems could also be mapped to it.
How such a computer would look depends on the implementation platform. Some are based on ion traps in resonators, or spintronic implementations. The Stanford approach uses a GaAs chip with different nanostructures, such as the optical resonator, with dimensions on the order of 5.0 µm that are interconnected via other nanostructures that allow the resonators to communicate. While it would resemble a typical semiconductor circuit, it would be a GaAs-based optical chip; the computer itself consisting of optical cavities on a grid connected by waveguides.
According to Vuckovic, this is the first time that such strong interactions between photons on a semiconductor chip have been observed, demonstrating a quantum equivalent of a classical computer’s logic and gate. Additional elements are still needed to build a functional quantum computer. “We need a quantum form of memory to store these quantum bits,” Vuckovic said. “We’re working to develop that because classic memory elements won’t work.” Vuckovic added that considerable R&D will be needed because, as yet, nobody has successfully produced a semiconductor-based quantum memory.
There is no consensus as to what the most promising platform for quantum memory would be. The need is to somehow store the state of the photon in some sort of quantum dot or some other quantum device inside a semiconductor, where all of its properties (polarization, etc.) are preserved, and then retrieve it successfully. Anything that can do this for a few microseconds would be considered a breakthrough in quantum memory today.
