Conformational bifurcation drives dual transport regimes in molecular junctions: Unsupervised machine learning...

Control over charge transport in molecular-scale devices requires a deep understanding of how minute structural changes influence electronic properties. Here, we demonstrate dual transport regimes in tunnel junctions of n-alk-1-yne (CnA) molecules with gold electrodes driven by conformational bifurcation-the emergence of two nearly isoenergetic (planar and skewed) molecular conformers (dihedral angles $\alpha =180°\text{and}\alpha \approx 65°$ at the alkyne terminus in the gas phase). Although the energy differences are small, these subtle conformational differences manifest as distinct transport behaviors, uncovered through unsupervised machine learning, which identified two junction groups: "short" and "long" chains, with distinct attenuation factors ( ${\beta }_{\text{short}}\approx 1.0$ vs. ${\beta }_{\text{long}}\approx 0.74$ ) and contact conductances ( $G_{c,\text{short}}\approx 200\mu S$ vs. $G_{c,\text{long}}\approx 8\mu S$ ). This dramatic impact of the dihedral angle exceeds the impact of the inter-ring twist angle in biphenyl-based junctions and rivals changes induced by switching from gold to platinum electrodes or from monothiol to dithiol anchors in oligoacene and oligophenylene junctions. X-ray photoelectron spectroscopy (XPS) confirmed this bifurcation, linking the "short" and "long" groups to planar and skewed conformers, with dihedrals remarkably agreeing with the gas-phase values. This work establishes conformational bifurcation as a promising route for designing programmable nanotransport properties through anchor-group control.