Appendix A

Al, Ga, In and Sn/Si{111}

The discussion of the Si-rt3-B phase in Chapter 3 should be prefaced by information on the Si-rt3-Al, -Ga, -In, and -Sn phases, which is summarized here.

1. It has been known for some time that the adsorption and heat treatment of group-III and -IV elements on Si{111} produces rt3 LEED patterns [122,123], and it was thought that these adsorbates occupied the H_3 site.
2. Research by Northrup [124-127] renewed interest; his total-energy calculations showed that the T_4-site was more stable than the H_3 site, if the surface was allowed to relax, for Al, In, and Si adsorbates.
3. Photoemission [128,129] and x-ray photoelectron-diffraction [130,131] studies could not conclusively rule out either the H_3 or T_4 site.
4. Inverse photoemission analyses [131,132], showed, however, that the calculated unoccupied bands for the T_4-model agree well with the experimentally determined bands for the Si-rt3-Al, -Ga, and -In phases.
5. Nogami et.al.[133] then verified, using STM, the T_4 site for the Si-rt3-Ga phase by comparing the registry of the rt3 unit cell with the bulk and with that of the now known 7*7 unit-cell.
6. Zegenhagen et.al.[134] determined the T_4 structure for the Si-rt3-Ga phase using STM, TEC and XSWM, and found relaxations occurring down to the fourth layer.
7. SXD work [135] determined the T_4 structure for the Si-rt3-Sn phase, with Si-atom relaxations occurring down to the fourth layer.
8. The Si-rt3-In structure was confirmed by ion scattering [136,137] to be a T_4-phase, although all of the structural parameters were not determined.
9. Huang et.al.[32] and Kawazu et.al.[138,139] reported dynamical LEED analyses for Si{111}rt3*rt3-30-Al and -Ga showing similar conclusions.

Appendix B

Pb/Si{111}

The properties of Pb on Si{111} and on Ge{111} surfaces are similar, and they are interesting for several reasons [140]:

1. Pb is a large atom to study stress effects with [141].
2. Pb is both immiscible in and nonreactive with Si and Ge [142].
3. Pb has a low Debye temperature, which can be exploited in the study of temperature-related surface-effects [143],
4. Pb atoms occupy the T_4 site and the surface relaxes for both the 1/3 ml phases of Si-rt3-Pb and Ge-rt3-Pb [144]-[149].
5. A 4/3 ml Si-rt3-Pb phase exists which is an incommensurate close-packed Pb layer [123,150,151,152]. This incommensurate rt3 is different from the high-coverage commensurate Ge-rt3-Pb phase [144,145,146] see Figures 4.2 and 4.3. These structures form the basis for subsequent Pb{111} overlayers. .
6. The epitaxial growth of Pb on Si{111}7*7-Pb and Si-rt3-Pb surfaces yields Schottky barrier-heights which differ by 0.2 eV [153]. Hence, the interface structure of metal/semiconductor systems can be dominant over bulk-properties.

Appendix C

Sb and Bi/Si{111}

The group-V elements are of great interest as bulk dopants, delta-dopants, and surfactants [154] on semiconductor surfaces [155]. Additionally, it is interesting to compare the rt3 phases of group-V elements on Si{111} with those of group III, group IV, and Au, because similar models have been proposed for all of them.

1. While As forms a 1*1 structure on Si{111} [156], Bi and Sb form rt3 phases on Si and Ge{111} surfaces.
2. Initial research by Kawazu et.al.[157, 158] reported the formation of a 1/3-ml Si-rt3-Bi structure and the epitaxial growth of Bi(0001) at higher coverages.
3. Studies using X-ray diffraction [159,160], XPS [161], X-ray photoelectron diffraction [162,163], and ARUPS [164,165] concluded that a 1-ml rt3 phase exists for both Sb and Bi. These phases contained trimers of the adsorbent, but the trimer's registration to the Si lattice was not determined.
4. Martensson et.al.[166], using STM and total energy calculations, confirmed the Si-rt3-Sb structure to contain a trimer centered on T_4-sites of Sb atoms on off-T_1 sites (T_4-T_1 model), see Figure 4.6.
5. Elswijk et.al.[167], also using STM, found the Si-rt3-Sb trimer was registered to the Si bulk in the H_3-T_1 configuration.
6. X-ray diffraction and SEXAFS results confirmed the T_4-T_1 registration for the 1-ml Si-rt3-Sb phase [168,169].
7. Dynamical LEED studies determined the T_4-T_1 structure for 1-ml Si-rt3-Bi, plus additional substrate relaxations [75].
8. Total energy calculations [170] confirmed the stability of the 1-ml trimer structures.
9. Our own Keating-energy analysis confirms the elastic stability of the simple-trimer structures for Bi and Sb on Si and Ge{111}, see Tables C.1, C.2, C.3, and C.4. This stability arises from the ability of the adsorbate to saturate the Si dangling-bonds with only minor substrate bond-bending. It should be noted, however, that the elastic-energy analysis cannot distinguish between a H_3 and a T_4 trimer-center.
10. Apart from the determination of a unique structure for both Sb and Bi on Si and Ge{111}, great significance lies in the confirmation of a trimer-based structure. This is because similar models have been proposed for the Si{111}rt3-Ag and -Au surfaces, but they have not been widely accepted.
11. Finally, two additional simple-adatom 1/3 ml T_4-structures were also determined for Bi on Si{111} and Ge{111} by LEED [74,171].

Appendix D

Ag/Si{111}

The structure of Si{111}rt3-rt3-30-Ag has been controversial for several years. One of the uncontested facts is that when Ag is deposited on a Si{111}7*7 surface which is heated during or after deposition to several hundred degrees, a Si-rt3-Ag phase is formed. Models have been proposed which are similar to those proposed for the Si{111}rt3*rt3-Au surface, see Chapter 5, but it appears now that Au and Ag have different effects on Si{111}.

Various models have been proposed for the Si-rt3-Ag phase. Most contain Ag but no Ag-Ag bonds, and are consistent with the seemingly honeycomb structure seen in STM images [172]. The models have either a Si-honeycomb layer above a Ag-trimer layer above the Si double-layer [173,174,175]; or a Si trimer-layer above a Ag-honeycomb layer above the Si double-layer [176,177,178]; or a simple Ag-honeycomb layer (H_3 sites) [179,180,181]; or a Si-honeycomb layer above a honeycomb chained-trimer (HCT) Ag-layer above a Si-trimer layer above the Si double-layer [110,182,-192]. The last model (HCT) has been sported with and without the overlying Si-honeycomb layer and is most widely accepted. However, LEED I(V) calculations of the HCT model do not agree with published experimental-spectra [193].

LEED dynamical analysis has proposed a different model and demonstrated its validity. Fan et.al.[31] noted the creation of an ``all Si-rt3''structure by rapid anneal and quench of an ion-bombarded Si{111} sample (with no Ag) before a 7*7 structure developed. LEED spectra of this ``all Si-rt3'' and of the Si-rt3-Ag are identical. Previously, Yang et.al.[70] had noted the striking similarity between LEED spectra of the Si-rt3-Ta and of the Si-rt3-Ag surfaces. If the Si-rt3-Si, -Ag, -Ta have identical spectra, then neither Ag nor Ta can be part of the long-range order of the system. Similar results have been reported for the ``all Ge-rt3'', Ge-rt3-Ag, and Ge-rt3-Li surfaces. Therefore, these metals must produce impurity-induced rt3 surfaces [194]. Fan et.al.[31] have determined the ``all Si-rt3'' surface to have one top-layer Si atom missing, to be relaxed, and to feature no Ag atoms in the unit-cell.

The present state of affairs is obviously confusing, but the five major-models proposed contain similarities. The necessity for Si vacancies in several of the proposed models, the lack of agreement on the Ag coverage, the lack of agreement on Si coverage [195,196], and conflicting information about where Ag resides leads one to believe that any one of the proposed models maybe correct. They also may all be incorrect.