7 Conclusions Conclusions and proposal for future work are presented here. Chapter 7.1 Conclusions Conclusions specific to each result chapter have been given at the end of these chapters. Hence only summary of the key findings from result chapters in this thesis will be given here. Four key findings are observed from this thesis and they are as follows: (i) the structure of (63x63)R30o/6H-SiC(0001) The presence of Si on this carbon-rich 63 surface has been in debate by current contradicting results reported by various groups. Our STM observations show that the 63 surface consists of triangular bright protrusion with pseudo-6x6 periodicity (6x6 maxima) and atomic clusters with 3 periodicity (3 cluster). Both STM and photoemission study shows that there are significant amount of elemental Si still present on this 63 surface. STM observations show that these elemental Si atoms can be found in the 6x6 maxima (Section 4.2.3). The 3 clusters, which are topographically lower than 6x6 maxima, are found to be C-rich. Based on dimension measurements of 6x6 maxima and 3 cluster, bias-dependant STM imaging and photoemission study of this 63 surface, a structural model for 6x6 maxima is proposed. This proposed structure correlates well with four Si 2p components observed by photoemission study of this surface. The scaffold for 3 clusters is proposed based on dimension measurement of this cluster, vertical displacement between 6x6 and 3 that observed from STM imaging and also based on photoemission studies where significant amount of sp3-related C-C bond are 248 Conclusions detected. The final structure of 63 consists of Si-rich 6x6 maxima and carbon-rich 3 clusters. Since the 3 clusters are in general denser than the atomic clusters associated to 6x6 maxima, the 63 surface is relatively richer in carbon than its precursor phase i.e. the Si-rich (3x3)R30o phase. (ii) mechanism for graphene formation from 63 surface The 63 structure proposed in Chapter has direct implications on the epitaxial growth of graphene on 6H-SiC(0001). The 3 clusters which are C rich are the precursor for graphene formation since they are the main supply for carbon source. Both of unit cells of 3 clusters and graphene are rotated 30o with respect to SiC-1x1. STM imaging coupled with Co decoration and step height measurements at various stage of graphitisation reveal that while the 63 decompose to form graphene, an additional three SiC bilayers underneath the initial 63 surface also collapse. The decomposition of these three SiC bilayers generates another 63-like layer at the interface. This observation suggests that the 63 surface would have sufficient carbon to form a monolayer of graphene (carbon from BL : ML graphene). However the presence of Si-rich 6x6 maxima would mean that these features are needed to be eliminated at higher temperature in order for 63 surface transforms into graphene. The collapse of three BL is observed as the primary graphitisation mechanism adopted by the 6H-SiC(0001) surface. An effective kinetic barrier of 3.0 0.4 eV is deduced for the transformation to take place. A 63-like interface is preferred over Si rich (3x3)R30o since this interface is also carbon rich due to desorption of Si during the graphene formation. 249 Chapter The graphene monolayer is found resting 2.3 0.2 Å above the maxima of 63like interface. This distance is 20% larger than the covalent bond length of Si-C (1.89 Å). Because of this large separation between graphene and 6x6 maxima, we believe the graphene monolayer interact weakly with the underlying interface. (iii) nucleation dynamics of Co on graphite and graphene Co forms three-dimensional dome-shaped nanoclusters on all three surfaces i.e. the C-rich 63, epitaxial graphene and graphite. Photoemission studies show that they are metallic clusters and physisorbed on these three surfaces with no shift in binding energy for both adsorbate and hosts. The existence of both Si and C on the 63 surface created a number of non-equivalent adsorption sites for Co adsorption. High-resolution STM imaging shows that the Co clusters preferentially nucleate on the Si-rich 6x6 maxima. Photoemission seems to suggest that partial charge transfer from Co to Si-rich 6x6 maxima may occur at this adsorption site where increase of elemental Si upon adsorption of Co is observed. On the other hand, it is interesting to note that despite weak interaction between graphene and 63 interface, the 63 interface is found to influence the nucleation of Co on graphene surface. Due to partial charge transfer from the 63 interface to the graphene surface [1], the critical nucleus size, i* of Co clusters increases from i* equals to zero (which deduced based on Co/graphite study) to i* equals to 3. i* equals to was found for Co nucleation of 63 surface. 250 Conclusions (iv) surface intercalation of Co on graphene and 63 surface Co deposited at elevated temperature exhibits a completely different growth dynamics than the growth at room temperature. Instead of forming small compact nanoclusters, Co forms multilayer flat-top islands with steep edges on graphene surface. The average apparent height of the islands is about monolayers and the widths range between 20 nm and 80 nm. At elevated temperature, thermodynamics plays an important role where due to the desire of the surface to keep its surface energy minimal, formation of surface-supported Co islands is not observed. Instead Co islands buried underneath the graphene monolayer are observed. By tuning the growth conditions (temperature), this work shows that Co can have different functions on 63 surface. They can form compact 3D metallic clusters at room temperature or become a catalyst for graphitisation of 6H-SiC(0001) at temperature as low as 610oC as shown in Fig. 6.21. 251 Chapter 7.2 Future Work Future work arises from this thesis is proposed as follows: (i) resolving and calculation for 63 surface The structure for 63 has not fully understood yet, in particular the structural details for the carbon-rich 3 clusters. The carbon structure of 3 cluster should akin to those found on graphene since both of them have same orientation i.e. rotated 30oC from the bulk SiC-1x1. Another option to have a better understanding of this 63 surface is by monitoring the transition from Si-rich (3x3)R30o to this 63 surface using techniques such as STM. In particular, the collapse of step heights (i.e. number of SiC bilayer decompose as Si continually sublimated from the surface) as the surface converts from Si-rich phase to 63 phase will give an idea on the carbon concentration of the 63 surface. The finalised structure should be tested with calculation. The STM image and angle-resolved photoemission spectroscopy (ARPES) of this surface should also be simulated and compare with experimental results. Similar simulation should also be carried out by placing a graphene monolayer on top. The results should be compared against the ARPES results reported by others where band structures that obey the Dirac’s equation with a gap opening are observed [1]. 252 Conclusions (ii) i* for bright and dim cluster The i* for the dim clusters can be determined by carrying out a series of experiment where Co are deposited onto graphene for a fixed growth temperature and coverage but with varying arrival flux of Co, F. Using the relation between dim cluster F density, N with F i.e. N , i* can be determined from the slope of log-log plot of N D vs F. This relation has been previously described in Section 2.2, Chapter 2. Probing the i* will give deeper insight to the nucleation dynamics of Co on graphene at elevated temperature. The information on i* is also required to deduce barriers associated with nucleation i.e. diffusion, desorption and nucleation barrier as shown in Table 2.1 (Chapter 2). (iii) comparison of Co on 63 and graphite at elevated temperature growth The growth of Co on 63 and graphite at elevated temperature should also be carried out. Preliminary studies have shown that Co can catalyse the graphene formation of 6H-Si(0001) at much lower temperature (500oC versus 1200oC). This may provide a better route to prepare better epitaxial quality of graphene on 6H-SiC(0001) since surface defects associated with deep pits formation during graphitisation at 1200oC can be avoided. As graphite and graphene has similar surface energy and structure, growth of Co on graphite may also produce surface intercalation of Co. If this is indeed observed, better 253 Chapter quality of Co-GIC may be produced since we not have interference from 63 and the open space between graphite sheets is structurally more ordered than the open space between graphene and 63 interface. (iv) magnetic measurement for GIC-like transport. All three surfaces have narrow size distributions of metallic Co clusters. In particular, 63 surface shows stronger template effect on the nucleation of these metallic Co clusters where growth at room temperature produces smaller and higher density of Co clusters. Magnetic measurements such as X-ray magnetic circular dichroism (XMCD) and magneto-optic Kerr effect (MOKE) should be carried out to investigate the sizedependant magnetism of these clusters. For growth at elevated temperature, the bright clusters and dim clusters were found to be metallic and not oxidised despite exposed to air. Since the energy difference between them is huge i.e. 0.6 eV, the growth conditions can be tuned to preferentially grow one of them on the surface. Magnetic measurements should also be carried out to determine the magnetic properties of these bright and dim clusters. Based on the size measurements from the bias-dependant studies in Fig. 6.8, the average size for bright and dim clusters are 45 and 87 atoms respectively*. Since they are just a few hundreds atoms, they should display superior properties than the bulk. Moreover, their high density on the surface (more than 1012 clusters per cm2) is also suitable for high-density data storage. * Density of Co atoms based on face-centered cubic unit cell. Calculation of cluster volume were carried out based on width and height measurements and by assuming disc- and dome-shaped geometry for dim and bright cluster respectively. 254 . graphite and graphene has similar surface energy and structure, growth of Co on graphite may also produce surface intercalation of Co. If this is indeed observed, better Chapter 7 254 quality of. diffusion, desorption and nucleation barrier as shown in Table 2.1 (Chapter 2). (iii) comparison of Co on 63 and graphite at elevated temperature growth The growth of Co on 63 and graphite. nucleation of 63 surface. Conclusions 251 (iv) surface intercalation of Co on graphene and 63 surface Co deposited at elevated temperature exhibits a completely different growth dynamics