Thesis presented January 17, 2023

Abstract:
This PhD thesis deals with the experimental investigation of charge and spin dynamics in silicon-based quantum dot arrays, confining either electrons or holes. The work presented was carried out in collaboration with the CEA-LETI, where the samples were fabricated on 300-mm SOI (Silicon-On-Insulator) substrates using an industrial-level CMOS platform. With this technology, quantum dots are confined inside silicon nanowires etched in the SOI. The compatibility of these quantum devices with microelectronics production lines can eventually play a key role in the development of a large-scale quantum computing platform based on semiconductor quantum bits (qubits). In this prospect, the development of efficient and scalable qubit readout and manipulation schemes is a crucial step. Ideally, one would like to manipulate the spin of an electron, or of a hole, via a simple modulation of the gate voltage (in the range of tens of GHz), and to read the spin state via a radiofrequency reflectometry technique (typically in the range of several hundred MHz), which can be implemented by connecting a gate, or an ohmic contact, to an LC resonator. Such an idea has motivated several experiments carried out within the framework of this thesis. A first experiment was carried out on an n-type array with 2×3 quantum dots. It compares two readout schemes based on gate reflectometry. The first one, based on a dispersive readout mechanism, requires no additional control gates, facilitating the scale-up to large qubit arrays. The second one, based on charge-sensing readout, requires additional readout components, and hence additional control gates. On the other hand, this second scheme is less sensitive to the tunnel coupling between neighbouring quantum dots. As shown in this thesis, it also allows for fast charge detection, a necessary condition for single-shot qubit readout. Regarding spin manipulation, in this thesis I was able to measure signatures of electron spin resonance induced by an electric-field modulation. This observation confirmed the existence of a spin-orbit coupling mechanism for electrons. However, the spin-orbit interaction turned out to be too weak to enable the observation of Rabi oscillations.
Holes in silicon have an intrinsically stronger spin-orbit coupling than electrons. Therefore, holes are better suited for electrically-driven spin manipulation. In this thesis, I present an experimental study on a p-type device with six gates, demonstrating independent and simultaneous single-shot readout of the charge states of two quantum dots defined by the two central gates. The readout is carried out by means of rf reflectometry through two large hole quantum dots positioned at the ends of the silicon channel and acting as charge sensors. In a following experiment, an extension of the same readout technique was applied to a four-gate p-type device in which we have been able to demonstrate the coherent electrical control of a qubit based on a single hole and to achieve a coherence time close to 100 microseconds, well beyond the state-of-the-art.
In order to minimize the number of control and readout gates, we studied and demonstrated the functionality of an elementary building block consisting of a double quantum dot defined in a p-type device with two gates. The first dot hosted a hole spin qubit and the second one was used for the readout of that qubit via dispersive reflectometry. The readout scheme used did not require any coupling to a Fermi reservoir, thereby offering an extremely compact and potentially scalable solution. In conclusion, this thesis work has largely focused on the exploration of different possible solutions for the readout of spin qubits in silicon nanowire SOI devices containing linear or bilinear arrays of quantum dots. In particular, my interest has focused on the development of solutions compatible with future large-scale integration.

Keywords:
qubit, silicon, quantum, array

On-line thesis. ​