Program Manager
Dr Jeff McCallum - University of Melbourne

MATERIALS CHARACTERISATION PROGRAM RESEARCHERS
Staff
Prof David Jamieson
Prof Steven Prawer
Dr Chris Pakes
Mr Paul Spizzirri
Dr Sergei Rubanov

Students
Mr Matthew Lay (PhD)
Mr Daniel Pyke (MSc)
Mr Byron Villis (PhD)

Collaborating Centre Researcher:
Prof Andrew Dzurak - University of New South Wales
Prof Michelle Simmons - University of New South Wales
A/Prof Alex Hamilton - University of New South Wales
Dr Eric Gauja - University of New South Wales
Mr Dane McCamey - University of New South Wales
Dr Wayne Hutchison - UNSW@ADFA
Prof Chennupati Jagadish - Australian National University

Other Collaborators:
Prof Chennupati Jagadish - Australian National University

PROGRAM DESCRIPTION
The aim of this program is to investigate the types and quantities of defects introduced
into silicon under the processing conditions used to fabricate the Kane quantum
computer by the top-down approach and to advise the Centre on methods of defect
minimisation.

1. 2005 Overview
A major highlight of the Materials Program for 2004 has been the contribution made to
the successful measurement of electrically detected magnetic resonance in ion implanted Si:P nanostructures. Full details of this achievement appear elsewhere in this annual report but it is a testament to the efforts of everyone involved that the desired result was achieved in the first set of fabricated samples and that our best guesses as to what would be required turned out to be largely correct. In the area of deep level transient spectroscopy analysis of interface and bulk traps in the silicon dioxide on silicon structures required for the quantum computer, it has been a year of trials and tribulations which culminated with our first successful measurement of the interface trap density for the 5 nm oxides utilised in the current generation of quantum devices. Earlier measurements were plagued by variability in the breakdown potential of the oxides when biased in accumulation together with an eventual realisation that the exposed 5nm oxides were exhibiting a time-dependent degradation that caused the breakdown voltage to reduce with time. Changing to a more sophisticated layer structure allowed the successful measurement of the interface state density to be achieved but we will need to continue monitoring the stability of the new samples into 2006 before we can be sure that a long term solution has been found. In 2005, the electron microscopy work of the group has largely concentrated on analysis of the ion beam damage that occurs during P implantation of the nanoscale structures needed for quantum computer fabrication. The project has identified the amorphisation threshold for P implantation in the low energy regime and has provided an insight into the relationship between the lateral distribution of damage and the edges of the implantation mask. This information is particular important for devices involving ion implanted source/ drain leads and dopant clusters.

2. Deep Level Transient Spectroscopy
Deep level transient spectroscopy (DLTS) in its various forms offers a sensitive means of identifying and quantifying the bulk charge traps and interface states introduced by
the phosphorus implantation process and monitoring their removal through subsequent
annealing steps. The technique has more than adequate sensitivity to detect defects
in the ion fluence regime ~1011 cm-2 of interest and it is readily able to probe the near-oxide interface region of metal-oxide semiconductor (MOS) devices formed using
processing steps and layered structures similar to those required in the quantum
computer project. In earlier reports we discussed implementation of DLTS on the
SULA system at the University of Melbourne and presented analyses of bulk charge
traps and interface states in relatively thick, 50 nm or greater, thermally-grown oxide
layers. During 2005, we have concentrated on DLTS measurements on the 5 nm oxides
required for the quantum computer. The 5 nm oxides turned out to be more difficult
than their thicker counterparts and MOS capacitors formed using the thin oxides exhibited a higher variability from device to device coupled with a long term degradation
which led after several months to consistent breakdown of the capacitors when placed
under a modest positive bias. Unlike the 50 nm oxides which are stable over several
years, exposed 5 nm oxides evidently degrade over a period of months and the
rate of degradation is largely independent of the storage conditions used and whether
or not metal gates have been deposited on the oxide. To circumvent this problem

Figure 1 CV profiles for MOS capacitors formed using 5 nm oxides. The exposed 5 nm oxide (red curve) exhibits premature breakdown and current flow through the oxide under a modest positive bias while the encapsulated oxide (blue curve) is stable.

Figure 2 (a) Schematic representation of the MOS capacitor structures with an encapsulated 5nm field oxide. (b) DLTS measurements of interface trap densities for Hpassivated thermally grown oxides of thickness 50 nm (blue curve) and 5nm (red curve).

Figure 3 Active fraction of implanted phosphorus versus implant dose, as obtained from various measurement techniques, for 15 keV implants and high-quality 5nm thick thermal oxides: 4pt probe measurements on commercial system at ANU (red datum), MOSFET threshold voltage measurements performed at UNSW (green datum), spreading resistance measurements performed by Schenkel’s group at Lawrence Berkeley National Laboratory (black data).

3. Phosphorus Activation
The state of play in our endeavours to investigate phosphorus activation in the lowconcentration near-surface regime relevant to the quantum computer are summarised in Figure 3. Four point probe measurements on the commercial Jandel system at ANU for a phosphorus dose of 1 × 1013 cm-2 at 15 keV (red data point) indicate an active fraction of greater than 80% under the standard P-activation anneal used in the Centre. Our findings are in agreement with the spreading resistance data of Schenkel’s group at Lawrence Berkeley National Laboratory for this dose. For lower implant doses, there is insufficient control available on the ANU system to obtain reliable measurements but previous threshold voltage measurements on ion implanted MOSFETs performed at UNSW indicated a high active fraction for doses as low as 2 × 1012 cm-2 (green data point). In contrast, the spreading resistance measurements of Schenkel’s group show a steadily decreasing active fraction with decreasing dose. To sort out the discrepancies, in 2006 we are planning to undertake a series of measurements on Hall bar structures at UNSW and to implement optical detection of dopants on the new optical spectrometer system set up in 2005 by Paul Spizzirri at Melbourne University. This latter technique, in particular, exhibits extraordinary sensitivity.

4. Electron Microscopy
A significant transmission electron microscopy (TEM) study in 2005 examined the lateral straggle of implanted phosphorus. The statistical variation in the location of the phosphorus qubits due to lateral straggle of the implanted ions is an important consideration in quantum computer fabrication. To gauge the lateral implant profile a special TiAu implantation mask with 49 nm apertures was fabricated at UNSW. A scanning electron microscopy image of the mask is shown in Figure 4(a). 14 keV P ions were implanted into our standard 5nm SiO2 on Si substrates through the mask to a fluence of 1 × 1015 cm-2 to create locally amorphised regions which have sufficient contrast to be readily observed via TEM. Figure 4(b) shows a low magnification image of the cross-section TEM sample prepared using the focussed ion beam lift-out technique. A magnified cross-sectional TEM image of the implanted sample is shown in Figure 5. Clearly visible in the image are the thick Au protective layer that was added prior to sectioning, the TiAu mask, the 5nm SiO2 layer, the crystalline Si substrate and the locally amorphised regions under the mask openings. Figure 6 shows details of one of the amorphous regions and its relationship to the overlying TiAu mask. The amorphous region was found to extend ~ 6 – 8 nm laterally beyond the edge of the implantation mask which can be compared with a lateral straggle for the phosphorus ion of ~11 nm predicted by the Monte Carlo code SRIM.

Figure 4 (a) A scanning electron microscopy image of the TiAu implantation mask. (b) A low magnification image of the cross-section TEM sample prepared using the focussed ion beam lift-out technique.
	Figure 5 A magnified cross-sectional TEM image of the implanted sample showing the thick Au protective layer that was added prior to sectioning, the TiAu mask, the 5nm SiO2 layer, the crystalline Si substrate and the locally amorphised regions under the mask openings.
	Figure 6 A detail of one of the amorphous regions showing its relationship to the overlying TiAu mask.