dc.description.abstracteng | In this thesis, three different aspects of the Si(111)-2x1 surface are investigated
in more detail: positive and negative buckling of the pi-bonded chains, mobility
of domain boundaries and signatures of phosphorus (P) atoms, which are present due to the
n-type doping of the samples. In the first part, domain boundaries are utilised
in order to determine whether two Si(111)-2x1 domains are buckled differently
considering distances between the ‘up’-atoms of frontally meeting pi-bonded
chains of the two domains. dI/dV measurements at 6K of domains with long,
defect-free pi-bonded chains of both buckling types do not only show the band
gap of pi-bonded chains with positive and negative buckling, but permit also the determination of the relative energetic
positions of the surface bands of the two buckling types.
The tip induced movement of domain boundaries between positively and negatively
buckled domains is treated in the second part of this thesis. At low positive
and negative voltages, the position of a domain boundary may be reversibly
altered by up to 8nm in favour of the domain with negative buckling. The
movement is largest in voltage regions where surface states are available for pi-
bonded chains with negative buckling but not for positively buckled pi-bonded
chains. The extent of movement is not influenced by the amount of tunnelling
current or the tip sample distance. As a comparison of measurements with the
tip scanning parallel to the pi-bonded chains to measurements with a perpendicular
scan direction indicates, the trigger for the movement is transmitted along
the pi-bonded chains and suppressed perpendicular to them.
The last and major part of this thesis is concerned with the signatures of the
dopant phosphorus atoms. These atoms are located, statistically distributed, at substitutional sites in bulk and surface. The Si(111)-2x1 surface offers four different
sites, and all signatures of P atoms in the surface layer can be assigned to one of
these sites. P atoms induce the same contrast pattern for positive and negative
buckling, except for the voltage values at which the contrast patterns occur.
These depend on the surface states which are at different energetic positions for
the two buckling types. We investigated pi-bonded chains with a length from
10nm up to more than 1 mm. In these limits, the chain length does not affect
the topographic contrast patterns of the signatures induced by P atoms at the
different surface sites.
Signatures of P atoms beneath the Si(111)-2x1 surface are attributed to three
subsurface layers. This is supported by counts of signatures which show even
numbers for signatures of P atoms in the surface layer and in the each of the
three subsurface layers, as can be expected from a large number of statistically
distributed donor atoms.
The origin of the contrast patterns at low negative voltages, which extend over
8nm in pi-bonded chain direction, cannot be the defect state of the P atoms
alone, as this defect state is highly localised within less than one surface unit
cell. The comparison of dI/dV measurements of P induced signatures to
dI/dV spectra of single metal atoms on metal surfaces shows that the origin
of these signatures are bound states which split-off from the surface states. As
electrons at the band edges of the surface states must have a minimal wavelength
of l = 7.4nm in pi-bonded chain direction, the bound states also show this
spatial extension.
We can distinguish between three cases where a bound state (or resonance)
occurs: The potential of P atoms located in the surface layer of free pi-bonded
chains causes a bound state which splits-off from the SCB and is located directly
beneath the surface band minimum. The potential of subsurface P atoms is
weaker and results in a surface state resonance in the lower part of the SCB.
The third case are P atoms in the surface layer of short pi-bonded chains. Their
potential also induces a bound state, but this bound state is split-off from the
SVB instead of the SCB, due to the presence of a Coulomb gap at EF which
prevents the coupling between the potential of the P atom and the SCB. | de |