Program
Manager
Dr Chris Pakes - University of
Melbourne
Participating
University of Melbourne Researchers
Students
Matt Lay (PhD)
Mr Giuseppe Tettamanzi (PhD)
Collaborating
Centre Researchers
Prof
David Jamieson - University of Melbourne
Prof Steven Prawer - University of Melbourne
Dr Jeff McCallum - University of Melbourne
Dr
Andrew Ferguson - University of New South Wales
Prof
Andrew Dzurak - University of New South Wales
Prof Robert Clark - University
of New South Wales
OTHER COLLABORATORS
Dr Sven Rogge - Kavli Institute of Nanoscience, TU Delft,
Netherlands
Dr Paul Alkemade - Kavli Institute of Nanoscience, TU Delft,
Netherlands
Dr David Saxey - University of Sydney, Australia
Prof Simon Ringer - University of Sydney, Australia
Mr Alessandro Potenza - University of Leeds, United Kingdom
Dr Chris Marrows - University of Leeds, United Kingdom
Program
Description :
The goal of this program is the measurement of single-charge and single-spin, and the
study of charge and spin-related effects that may influence the development of siliconbased quantum computer architectures.
The key objectives of the program are:
- To make use of scanning probes to explore the electrostatic environment of ion-implanted impurities in silicon.
- To develop techniques for the fabrication of a nanoscale Superconducting
Quantum Interference Device (SQUID), sufficiently sensitive for
single-spin measurement.
- To study isolated nanoscale
magnetic particles, which may have application in developing spin-based
quantum computer technology.
1. Charge trap and dopant imaging in ion-implanted Si:P devices
The phosphorus ion implantation process employed in top-down quantum computer development will induce defect sites that must be removed by rapid thermal annealing (RTA). Making use of Scanning Kelvin Microscopy (SKM), we can monitor the potential of the ion-implanted surface in response to RTA processing schemes proposed within the centre. Buried impurity islands are formed by implantation of P, B and Si-ions through photoresist masks into an underlying Si substrate, which incorporates a high-quality 5 nm SiO2 surface layer. Implanted islands are seen as circular regions where the potential is modified with respect to the surrounding un-implanted background (Figure 1). Following implantation, a significant SKM signal is observed, arising presumably from charge trapped on defects introduced in the Si and SiO2 layer. The signal is found to be weakest for the B-implanted islands, which is consistent with boron having the lowest atomic mass amongst the species under study, giving rise to reduced nuclear stopping. Following a 1000°C anneal for 5 seconds in argon gas, the potential profiles are modified. The Si-implanted islands disappear, giving us confidence that the RTA is sufficient to remove ion-implantation damage in the Si matrix. The noise level achieved with this technique approximates to the signal that would arise from an array of charge traps spaced by approximately 70 nm, which exceeds the typical inter-qubit spacing utilised in qubit test devices. For the P-ion and B-ion implanted surfaces following RTA, the sub-surface islands remain in the surface potential image, indicating a Si.significant signal arising from the impurities once they become activated by the RTA process. The inclusion of B-ion implanted samples in 2005 has allowed us to examine the difference in surface modification due to donor and acceptor impurity implantation, a useful test of SKM analysis. To this aim we have also removed the surface SiO2 layer via a hydrofluoric acid (HF) etch to avoid complications arising at the Si-SiO2 interface. For the P-islands a reduction in surface potential with respect to the undoped surface is observed, due to rising of the Fermi level towards the conduction band, as expected for a donor-rich environment. In contrast, the surface potential is enhanced for the B-islands, consistent with lowering of the Fermi level towards the valence band due to the introduction of acceptors.
Our experiments indicate that SKM may be very sensitive to buried phosphorus donors in silicon, and we are interested in the possibility that SKM may be sensitive to just a few underlying dopants. During 2005 we have developed a Kelvin-probe capability within our ultra-high vacuum (UHV) scanning probe microscope, which offers several advantages over ambient SKM experiments, including imaging with atomic resolution. Non-contact force microscopy in vacuum is a non-trivial technique because the AFM cantilever has a very high Q-factor, and phase-locked loop circuitry must be therefore employed to monitor the shift in resonant frequency of the cantilever in response to a force. To investigate the sensitivity of SKM to a low number of buried dopants, we are studying CoSi2 epitaxial Schottky diodes, that take the form of selfassembled nanoscale islands on the Si(111) 7×7 surface.
 |
Figure 1: (a) SKM surface potential image, revealing an implanted sub-surface array of P-doped islands in Si. (b-d) Profiles in surface potential across single ion-implanted islands, illustrating the effects of RTA processing and removal of the surface SiO2 layer, for P, B and Si doped islands respectively. |
 |
2. Dopant imaging in Si:P Fin-FET devices
An ability to image the spatial distribution of dopants in a fully-fabricated Si:P device would be of fundamental and commercial importance. In nanoscale devices, incorporating just a few dopants, the donor locations have a significant influence on the transport characteristics, governed by tunnelling thorough impurity sites, and tunnelling spectroscopy is an important tool in the centre to elucidate the electronic structure of introduced impurities. During 2005 Chris Pakes spent five weeks visiting the Kavli Institute of Nanoscience at the Delft University of Technology, where Sven Rogge directs a team examining transport in silicon Fin-Field Effect Transistors, which are fabricated by Serge Bieseman’s group at IMEC, Belgium. Extending an existing collaboration between our centre and the Electron Microscope Unit (EMU), University of Sydney, to include the Kavli Institute, we have commenced a project to make use of the EMU Atom Probe to image dopants in Si:P devices, fabricated both in Europe and Australia. The difficulty in this technique rests in preparation of a specimen suitable for atom-probe analysis, which must take the form of a sharp needle which is isolated from its surroundings by several tens of microns, yet has a tip with a radius of curvature of order 50 nm and is registered to the active area of the device to within a few tens of nanometres. Focused ion-beam processing, performed by Paul Alkemade (Delft) has demonstrated the ability to ion-beam mill the needle with sufficient accuracy (Figure 2), and furthermore shown that the FIB can excavate the surrounding material to isolate the specimen from the surrounding layer. These specimens will be tested using the Atom Probe facility early in 2006.
Figure 2: SEM image illustrating the formation of a Si:P FinFET specimen for atom-probe analysis by FIB-milling.
3. Superconducting nanoelectronics
Single spin measurement is a crucial goal that must be achieved if readout of quantum computer architectures employing spin-based qubits is to be realised. Submicron DC Superconducting Quantum Interference Devices (SQUIDs) may be a potential candidate for direct, magnetic spin
measurement with single-spin sensitivity. In 2005 we successfully completed the development of a process flow utilising a crossed beam Scanning Electron Microscope (SEM)/Focused Ion Beam (FIB) system to pattern superconducting nanostructures in thin Nb films, which are deposited by DC sputtering in an Ar-gas environment at the University of Leeds.
Our fabrication process utilises a 30 keV Ga ion beam to pattern both the functional component and contact structures. Subsequent measurement of the quantum behaviour of fabricated devices is performed at the University of Melbourne using a 4K low-noise measurement system, capable of voltage measurement with nV/√Hz sensitivity.
We have demonstrated Josephson behaviour in microbridge junctions (Figure 3) and shown that the
junction critical current can be tuned by varying the physical width of the junction (Figure 4). With this capability established, we are now addressing the fabrication of a sub-micron SQUID (Figure 5). The transport characteristics are found to be very different to the single junctions, with the SQUIDs exhibiting hysteretic behaviour at low temperature. A study of the SQUID characteristics as a function of the loop dimension is in progress.
In the narrowest microbridges that we have fabricated, typically with width below 100 nm, voltage steps appear in the transport characteristics at low temperature (Figure 6). These steps indicate an increase in dissipation, which is typical of the formation of a current-driven phase slip centre.
Phase slip behaviour occurs as a result of the superconducting order parameter vanishing locally at some point along a thin wire, allowing the order parameter to slip, and governs the mechanism by which superconductivity is suppressed in the limit where the wire width is comparable to the coherence length. In addition to being dynamically or thermally driven, phase slip centres can be activated quantum mechanically, in a process that has been proposed as the basis of a new superconducting qubit. In 2006 we shallexplore this as a further additional to the centre’s tool-kit of superconducting electronic devices.
|
