SiGe Spin-Qubits

 

Principal Investigator: Lars Schreiber Background Research Focus References

  

Background

Copyright: © Lars Schreiber

Quantum computation poses an important challenge: In contrast to classical computing, every physical degree of freedom involved in the process of quantum computing must be under precise control. Its base unit, the qubit, has to be coherently manipulated (for 1-qubit gates) and coupling of these qubits has to be feasible (for 2-qubit gates) while the disturbance of the environment is kept to a minimum.

The spin of an electron in a solid-state fulfills all these conditions and is thus ideal to form a qubit [1]. Individual electrons can be electrostatically trapped in a silicon-germanium (SiGe) heterostructure (see figure). This all-electrical, scalable approach is not only compatible with standard silicon computer technology, but more significantly the silicon-based qubit can be efficiently decoupled from its environment. Being the most significant environmental disturbance, a million nuclear spins typically interact with the electron spin. This uncontrolled nuclear spin bath limits the usefulness of spin qubits by destroying the quantum effects. In SiGe, however, the nuclear bath can be removed by isotopical purification. Thus, the SiGe spin qubit stays preserved enabling quantum computation. Recently spin-to-charge read-out by Pauli blockade and very long spin relaxation times were demonstrated [2]. The quality of the SiGe heterostructures has been raised clearing the path to coherent singlet-triplet oscillations in in SiGe [3]. Our samples are fabricated in the local clean room facility and at the Helmholz Nanoelectronic Facility at Forschungszentrum Jülich. Measurements are performed in ³He/⁴He dilution refrigerators at a temperature of about 10 mK.

 

Research Focus

We test different encodings to represent a qubit by electron spins for scalable quantum computation in SiGe. The simplest encoding is to use the spin-up and spin-down state of a single electron spin in a magnetic field as the qubit 0 and 1 states. This approach is taken in the famous proposal by Loss and DiVincenzo [4] (see figures) and also pursued in GaAs [5]. Single qubit operations are performed by electron spin resonance. Since local AC magnetic fields are hard to achieve locally [6], we invest spin manipulation by AC electric fields in an inhomogeneous magnetic field generated by a micro-magnet (see figures).

We improve the quality of the electrostatically defined quantum dots by replacing the dopant layer by electron inducing top gates and aiming at smaller electrostatic gate patterns (in cooperation with FZ Jülich and IHT RWTH Aachen).

We investigate the electrical and magnetic sources of spin decoherence in silicon. Therefore we analyze the role of charge-noise on qubit gate fidelities. Furthermore we compare the qubit decoherence in samples with a natural abundance of silicon isotopes (²⁹Si exhibits a nuclear spin) and isotopically purified samples.