Toward Reproducible Domain-Wall Conductance in Lithium Niobate Single Crystals

Conductive domain walls (DWs) in lithium niobate (LiNbO3, LNO) are promising constituents for potential applications in nanoelectronics, due to their high conductance, as compared to the surrounding bulk material, their high local confinement at the nanometer scale, and the ability to be created quasi-on-will through dedicated high-voltage poling. However, electrically contacting the DWs unavoidably leads to the formation of a potential barrier between the DW itself and the electrode material. Thus, the focus of this work is the investigation of the various factors influencing the electronic transport across that barrier, namely, the type of electrode material, the quality of the LNO surface (atomically-smooth versus mirror polished), the quality of the crystal lattice (i.e., the presence of higher concentrations of lithium and oxygen vacancies VLi and VO), and the magnitude of the applied voltages during the domain-wall conductivity (DWC) enhancement procedure. It is found that all the above-mentioned factors have a significant impact on the current-voltage characteristics of the DW-electrode system. For example, the metal electrodes deposited onto the surface of the LNO crystal, once, impede the DW motion, while, secondly, stabilizing the DWs inclination across the LNO crystal. Another important finding is the major role played by large negative voltages in the DWC-enhancement procedure that strongly influences the near-surface structure of the DW, and hence the qualitative characteristics of the formed potential barrier, such as characteristic voltage and saturation current. The application of moderate voltages from –50 V to –100 V is also found to influence the structure of the near-surface DW. The creation of a variety of vacancy defects inside the bulk LNO that accompanies the formation of an atomically-smooth surface, is found to have far more influence on the DW charge transport than the quality of the surface, due to the formation and repulsive interaction of a multitude of spike domains stemming from these defects. In summary, the results demonstrate the importance of providing known and reproducible sample surface conditions and identifying promising directions for implementing reproducible domain wall conductivity