Suppressing Si Valley Excitation and Valley-Induced Spin Dephasing for Long-Distance Shuttling

We present a scalable protocol for suppressing errors during electron spin shuttling in silicon quantum dots. The approach maps the valley Hamiltonian to a Landau-Zener problem to model the nonadiabatic dynamics in regions of small valley splitting. An optimization refines the shuttling velocity profile over a single small segment of the shuttling path, allowing for fast shuttles and ensuring the spin and valley degrees of freedom are unentangled. The protocol reliably returns the valley state to the ground state at the end of the shuttle, after which the spin phase accumulation becomes predictable and is compensated with a single virtual z rotation on the spin. The time cost and complexity of the error suppression is minimal and independent of the distance over which the spin is shuttled, and the maximum velocities imposed by valley physics are found to be orders of magnitude larger than shuttling speeds achieved experimentally so far. This protocol offers a chip-scale solution for high-fidelity quantum transport in silicon spin-based quantum computing devices.