The work in this thesis is basically divided into two related but separate investigations. The first part treats simple chemical reactions of adsorbate molecules on metallic surfaces, induced by means of a scanning tunneling probe (STM). The investigation serves as a parameter free extension to existing theories. The theoretical framework is based on a combination of density functional theory (DFT) and non-equilibrium Green's functions (NEGF). Tunneling electrons that pass the adsorbate molecule are assumed to heat up the molecule, and excite vibrations that directly correspond to the reaction coordinate. The theory is demonstrated for an OD molecule adsorbed on a bridge site on a Cu(110) surface, and critically compared to the corresponding experimental results. Both reaction rates and pathways are deduced, opening up the understanding of energy transfer between different configurational geometries, and suggests a deeper insight, and ultimately a higher control of the behaviour of adsorbate molecules on surfaces. The second part describes a method to calculate STM images in the low bias regime in order to overcome the limitations of localized orbital DFT in the weak coupling limit, i.e., for large vacuum gaps between a tip and the adsorbate molecule. The theory is based on Bardeen's approach to tunneling, where the orbitals computed by DFT are used together with the single-particle Green's function formalism, to accurately describe the orbitals far away from the surface/tip. In particular, the theory successfully reproduces the experimentally well-observed characteristic dip in the tunneling current for a carbon monoxide (CO) molecule adsorbed on a Cu(111) surface. Constant height/current STM images provide direct comparisons to experiments, and from the developed method further insights into elastic tunneling are gained.

No description available.

No description available.

No description available.

No description available.

No description available.

The scanning tunneling microscope (STM) is one of the most advanced surface science tools capable of atomic resolution imaging and atomic manipulation. Unfortunately, STM has many time-consuming bottlenecks, like probe conditioning, tip instability, and noise artificing, which causes the technique to have low experimental throughput. This dissertation describes my efforts to address these challenges through automation and machine learning. It consists of two main sections each describing four projects for a total of eight studies. The first section details two studies on nanoscale sample fabrication and two studies on STM tip preparation. The first two studies describe the fabrication of graphene-based Josephson Junction devices and the factorial optimization of patterned carbon nanotube forest synthesis. The second two studies focus on the factorial optimization of electrochemical STM tip etching and automated STM tip functionalization via in-situ silicon nanocolumn growth. The second section details four studies on the use of neural networks for STM image and spectroscopy analysis. The third two studies are on the effectiveness of convolutional neural networks for identifying images of conditioned STM tips on the Au(111) surface and on the detection and metrology of atomic scale defects in single crystal tungsten diselenide, a transition metal dichalcogenide. The fourth two studies are on the use of variational autoencoders to autonomously classify scanning tunneling spectra of various materials, molecules, and surface structures and to identify bismuth and nickel atoms from cross sectional STM images of doped gallium arsenide.

No description available.

No description available.

The electronic and geometrical structure of single difluoro-bora-1,3,5,7-tetraphenyl-aza-dipyrromethene (aza-BODIPY) molecules adsorbed on the Au(111) surface is investigated by low temperature scanning tunneling microscopy and spectroscopy in conjunction with ab initio density functional theory simulations of the density of states and of the interaction with the substrate. Our DFT calculations indicate that the aza-BODIPY molecule forms a chemical bond with the Au(111) substrate, with distortion of the molecular geometry and significant charge transfer between the molecule and the substrate. Nevertheless, most likely due to the low corrugation of the Au(111) surface, diffusion of the molecule is observed for applied bias in excess of 1 V. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.