Production of Isotopically Pure Si Layers for Quantum Computers using Implanted Layer Exchange 

Jonathan England, Ella Schneider 

Surrey Ion Beam Centre, Advanced Technology Institute, University of Surrey 

A major attraction of donor based qubits in Si is their potential compatibility with CMOS technology with an obvious path to future mass production.  We are developing an implanted layer exchange (ILE) process to produce isotopically pure Si layers for the manufacture of quantum computers.  ILE can use a standard production implanter to implant 28Si into the surface of an Al film deposited on a substrate wafer.  A post implant, layer exchange anneal causes the 28Si atoms to diffuse through the Al and grow on the substrate.  ILE allows implantation energies to be used where beam transmission is maximised, requires order of magnitude lower fluences than enrichment by direct implantation into Si and, as ILE self-getters isobaric impurities and overcomes surface oxidation, negates the need for a UHV endstation. 

BF STEM images of the 28Si ILE enrichment process: (a) 28Si/30keV/6.7×1016 cm-2 ion implantation into Al on Si; (b) after 500°C / 1 min layer exchange anneal. 

This talk will describe the imperfect state of the art of the layer exchange process.  The elephant in the room is that by involving a layer of Al, the enriched Si layer promises to contain Al atoms that are far worse in spoiling T2* lifetimes of the donor spins than those of the 29Si that have been eliminated.  The ultimate ILE process would produce non-defective, single crystal, isotopically pure, ²⁸Si layers of uniform thickness containing no Al contamination.  The problem statement has been changed from one of isotopic enrichment to that of chemical purification.  Future purification studies will involve the use of gettering techniques that have reduced levels of metal contamination in the solar and CMOS industries. 

If ultimately perfected, aluminium free ILE would compete with other enrichment techniques to supply the substrates into which single ion implants would produce donor qubits.  We would like to discuss a scheme whereby a low level residual Al atoms might provide isolated acceptor spins opening the possibility of producing acceptor qubits without the need for a single ion implant step.