The future of quantum technologies could be written in sand, according to researchers from the University of Bristol in the UK. A group at the institution demonstrated high-fidelity quantum photonic circuits using waveguides made of silicon dioxide, or silica — the main constituent of sand — on silicon. The investigators constructed a two-photon quantum interference and a controlled-NOT gate device, the latter achieving an average logical basis fidelity of greater than 94 percent.The design of an integrated quantum photonic circuit on a chip includes a direction coupler that replaces the beamsplitter used in versions created with bulk optics. If the two waveguides are close enough together, the evanescent fields overlap, and operation similar to that of a beamsplitter results (a). The modeled transverse intensity profile of the guided mode superimposed on the waveguide structure (b) and the design of an integrated two-photon controlled-NOT quantum logic gate, showing the various inputs and outputs (c), are shown. Reprinted with permission of Science.Research team leader Jeremy L. O’Brien noted that this is a better performance than that of the same device implemented in bulk optics. Moreover, the integrated device actually could be working at an even higher level than the measurements indicate.“We think the main source of error is the alignment of the polarization-maintaining fibers in the fiber arrays we use to test the devices. If true, this would mean that the device itself is working far better than this,” O’Brien said.Properly exploited, quantum mechanical effects could lead to communication, computation and measurement devices with improved performance over what is currently available. Single-photon quantum circuits have been built out of bulk optics; however, as with discrete electronics, such devices are bulky and cannot be scaled to large arrays.One solution is the photonic equivalent of an electronic integrated circuit, with waveguides routing photons as needed. Such waveguides must be constructed out of a material that is almost totally transparent at the right wavelengths while also allowing single-mode operation at the core size of single-mode optical fibers. The most promising choice, the researchers noted, is silica doped at a low level to control its refractive index and grown on a silicon substrate.They started by growing a 16-μm-thick buffer layer of undoped silica on a silicon wafer, then growing a 3.5-μm waveguide core of silica doped with germanium and boron oxides. They patterned this layer into waveguides 3.5 μm wide. The length and layout of the waveguides varied according to the type of circuit desired. The researchers topped the core with a cladding layer of phosphorus- and boron-doped silica.The devices thus constructed were not exotic, with a bulk optics version of the controlled-NOT gate demonstrated years before, and the fabrication techniques are standard in the semiconductor industry. However, achieving quantum interference is more demanding than its classical counterpart, O’Brien said, and constructing the integrated devices required bringing together the waveguide design and fabrication expertise with the knowledge of single-photon quantum optics.After fabricating the devices, the group tested them using custom-made measurement equipment built by graduate student Alberto Politi. For example, the scientists generated pairs of quantum-encoded photons, coupled these into the controlled-NOT device using optical fibers and measured the output. They found that the measured results closely tracked the expected ideal outcome, with much of the error suspected to be arising off-chip.The group now is working to determine and reduce error sources and to implement more sophisticated circuits using the same architecture. One possibility is the use of electro-optically active materials in addition to or instead of silica, along with the integration of single-photon sources and detectors into the chip.“The ultimate goal is to have a platform for photonic quantum technologies,” O’Brien said, also noting that the end result could be a quantum computer.Science, May 2, 2008, pp. 646-649.