Spins of donor electrons and nuclei in silicon are promising quantum bit (qubit) candidates which combine long coherence times with the fabrication finesse of the silicon nanotechnology industry. We outline a potentially scalable spin qubit architecture where donor nuclear and electron spins are coupled to spins of electrons in quantum dots and discuss requirements for donor placement aligned to quantum dots by single ion implantation.
INTRODUCTIONElectron and nuclear spins of donors in silicon have long been recognized as promising qubit candidates [1]. In isotopically purified 28 Si they exhibit long coherence times [2, 3] and their integration can benefit from the great fabrication finesse of silicon nanotechnology. Several prominent proposals for scalable quantum computer architectures with donor spin qubits have emerged [1,[4][5][6][7]. In the original Kane proposal, quantum information is stored in the nuclear spin of phosphorus atoms. Electrostatic gates facilitate transfer of quantum information form nuclear to electron spins and between electron spins, by modulation of the contact hyperfine interaction (A-gates), and the exchange coupling (J-gates), respectively. Recently, reliable detection of single electron spins [8] and the control of single electron and nuclear spins [9] states was reported for donors in silicon. Similar advances have also been reported for NVcenters in diamond [10,11] where there are many materials and integration challenges complementary to those for donors in silicon [12,13].Following Di Vincenzo's widely used criteria for the development of a large scale quantum computer [14], elements of quantum memory, quantum logic and efficient quantum communication channels have to be integrated. While single donor electron and nuclear spin readout and control have been demonstrated, the next difficult challenges are to master spin qubit coupling so that two and multi-qubit logic operations can be implemented. In early donor qubit proposals, coupling was envisioned along 1D chains of nearest neighbor coupled qubits. This ought to suffice for quantum logic demonstrations with several qubits even with limited coupling control [15] but severe limitations of nearest neighbor coupling have been pointed out [4,16]. Coherent shuttling of electrons between donors has been proposed as a path to circumvent nearest neighbor coupling challenges or to supplement nearest neighbor coupling with a longer range coupling option [4,17]. For electron shuttling, critical questions regard spin coherence of donor electron and nuclear spins during cycles of ionization and recombination. Other potential paths for long range transport of quantum information from donor spins include concepts of a spin bus [18], virtual phonon mediated coupling [19], coupling via nano-mechanical resonators [20] and spin to photon coupling in optical cavities [21] or via high Q microwave resonators [10,22, 23].In parallel to single donor spin control, control of electron spins in silicon and SiGe based quantum dots has also matured rap...