Identifying Clean and Contaminated Atomic-Sized Gold Contacts Under Ambient Conditions Using a Clustering Algorithm

Molecular electronics studies have advanced from early, simple single-molecule experiments at cryogenic temperatures to complex and multifunctional molecules under ambient conditions. However, room-temperature environments increase the risk of contamination, making it essential to identify and quantify clean and contaminated rupture traces (i.e., conductance versus relative electrode displacement) within large datasets. Given the high throughput of measurements, manual analysis becomes unfeasible. Clustering algorithms offer an effective solution by enabling the automatic classification and quantification of contamination levels. Despite the rapid development of machine learning, its application in molecular electronics remains limited. In this work, we present a methodology based on the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm to extract representative traces from both clean and contaminated regimes, providing a scalable and objective tool to evaluate environmental contamination in molecular junction experiments.