Silicon spin qubits: Exploring an epitaxial materials platform for controllable valley splitting and charge noise mitigation

点击翻译按钮获取中文摘要

The spin of single electrons is one of the most promising candidates to define qubits in a highly integrated and scalable semiconductor environment. Spins localized in silicon-germanium by means of gate-defined quantum dots have recently seen major advances in manipulation fidelities and ideas for scalable quantum computing architectures. In our previous project, we, ourselves, have demonstrated a state-of-the-art qubit in highly enriched 28Si/SiGe, longest echo coherence time and high valley splitting. However, there is a consensus that charge noise, potential disorder and valley splitting of these Si/SiGe-based qubits need to be improved in terms of reproducibility to implement further scalability. The material physics of the qubit devices has been scarcely addressed so far, although it is foreseeable that it contributes significantly to the currently observed limits.With respect to charge noise and potential fluctuations, our previous experiments point to the special role of the interface between the Si/SiGe heterostructure and the oxide dielectric. We propose here to replace the amorphous aluminum oxide - with all its charged defects - currently used worldwide by alternative oxides. Our focus will lie on an oxide that is grown by molecular beam epitaxy, in-situ, epitaxially and monocrystalline on the Si/SiGe heterostructure. This should minimize impurities and defects at the interface and in the dielectric and be significantly reflected in our noise analyses of single qubits. The valley splitting for spin qubits in Si/SiGe is currently dominated by local short-range atomic impurities such as steps. Here, we want to exploit that an electric field perpendicular to the Si/SiGe quantum well can, additionally, significantly contribute to the valley splitting. Therefore, we will extend our surface gate controlled qubit devices by a global backgate. This will enable us to apply considerably higher electric fields to the heterostructure than is possible in the current state of research. Experimentally, a significant increase of the valley splitting is expected, which on the one hand will considerably facilitate the manipulation of spin qubits and at the same time - due to the global backgate - promises lower fluctuations of the splitting in the local space. Within the project we will also introduce locally strongly localized annealing for the activation of implanted donors in the Ohmic contacts of the qubit components by means of an in house-designed laser scanner to replace the currently worldwide used high temperature annealing step of the entire component. The resulting significant reduction of volume diffusion in the Si/SiGe/dielectric qubit structure will result in reduced charge noise and potential fluctuations as well as in reproducibly high valley splitting.