Quantum Hall devices on epitaxial graphene: towards large-scale integration

Doktorsavhandling, 2015

Quantum Hall devices have been used as the primary standard of electrical resistance for over two decades, and they are unlikely to be replaced in this role any time soon. The work presented in this thesis was being done towards the goal of establishing epitaxial graphene on silicon carbide as a new material of choice for these devices. Experiments on individual devices have already demonstrated that due to unique electronic properties of graphene and peculiarities of its interaction with the SiC substrate, quantum resistance standards based on epitaxial graphene can operate at higher temperatures, lower magnetic fields, or higher current densities, as compared to their state-of-the-art gallium arsenide counterparts. Here, we were aiming at developing the technology for reliable mass-production of the devices. One of the issues that we address is the carrier density control. We have found that photochemical gating, a technique which has previously been used for this purpose, becomes unreliable when the electron density needs to be lowered by more than 1016 m2. Instead, corona discharge can be used for efficient electrostatic gating, enabling us to sweep the carrier density from 4*1016 electrons*m-2 to 5*1016 holes*m^-2 and to observe the quantum Hall effect at low doping. The presence of bilayer patches in majority-monolayer samples is another important problem. We have observed both metallic and insulating behaviour of these patches while driving the monolayer into the quantum Hall regime. When the bilayer is metallic, we show that a patch completely crossing the Hall bar will break down the quantum Hall effect in a way that agrees with theoretical expectations. Further, we propose imaging these patches by optical microscopy as a way of avoiding them, by selecting substrates where the patches are sufficiently small and sparse. We demonstrate that, despite the optical contrast being less than 2%, the bilayer areas can be imaged in real time using digital post-processing. Also, we show that optical microscopy can be used to detect the steps that form on the SiC surface during graphene growth, and even measure their height: steps as low as 1.5 nm could be clearly seen. Finally, we have fabricated arrays of 100 Hall bars connected in parallel, devices which provide a low-ohmic quantum standard if every single Hall bar works correctly. We have chosen a substrate with a sufficiently low bilayer content, and adapted the geometry of the Hall bar to the shape of the patches. One out of for devices has performed correctly within the relative measurement precision of 10-4 in magnetic fields above 7 tesla. We see this as a confirmation that the quality of graphene was sufficiently high to enable ≥99% yield of working Hall bars.