Using Deep Learning to Identify Molecular Junction Characteristics

• DOI: 10.1021/acs.nanolett.0c00198
• Publication Date: 2020-04-03T15:58:17Z
• Authors: Tianren Fu; Yaping Zang; Qi Zou; Colin Nuckolls; Latha Venkataraman
• Keywords: Break junction

The scanning tunneling microscope-based break junction (STM-BJ) is used widely to create and characterize single metal-molecule-metal junctions. In this technique, conductance continuously recorded as a metal point contact broken in solution of molecules. Conductance plateaus are seen when stable molecular junctions formed. Typically, thousands created measured, yielding distinct versus extension traces. However, such traces rarely analyzed individually recognize the types Here, we present deep learning-based method identify show that it performs better than several commonly recently reported techniques. We demonstrate identification from mixed measurements with accuracies high 97%. also apply model an situ electric field-driven isomerization reaction [3]cumulene follow over time. Furthermore, our can remain accurate even key parameter, average conductance, eliminated analysis, showing goes beyond conventional analysis existing methods.