In 1998, Bruce Kane proposed a solid state quantum computer architecture that has the potential to be scalable to kilo-qubits or even mega-qubits.[1] It utilizes an array of single 31P dopant atoms embedded in a 28Si lattice. The challenge for this concept is the placement of one and only one 31P atom at each site, ~10 nm below the surface and separated ~20 nm, with about 1 nm precision in the Si substrate. To avoid ion straggling, the P ions should be deposited at low energy, ~100 eV. Even the best commercial ion guns do not have sufficiently small phase space to accomplish this task. Deposition of one and only one ion in a site is also problematic.
A laser cooled and trapped single atom in a magneto optic trap, followed by resonant photo-ionization, provides a deterministic ion source with a small enough phase space to achieve the required precision in deposition [2]. Current laser technology does not allow for the direct cooling and trapping of P atoms, as the transition at 177 nm is in the vacuum ultraviolet region. However, the 3s23p2 3P2® 3s3p3 3Do3 transition in silicon at 221.7 nm is a cycling transition that may be used for the laser cooling and trapping of silicon atoms. In particular, the radioactive isotope 31Si would beta decay after deposition into 31P. Such a deterministic single ion source of31Si could provide the desired 31P qubits for a Kane quantum computer.
We will report on the status of our experiment. We have verified spectroscopically favorable conditions for trapping of all four isotopes of Si and for efficient resonance ionization.[3] The magneto-optic trapping of Si atoms is underway. The ion optics for extraction of single Si+ ions and low energy implantation into a Si substrate are being designed.
