Identifying\nQuantum Interference Effects from Joint\nConductance–Thermopower Statistics

Although quantum\neffects are thought to dominate the heat and charge\ntransport through molecular junctions, large uncertainties in chemical\nstructure, lead-molecule coupling strengths, and energy levels make\nit difficult to definitively identify these effects from the measured\nthermopower <i>S</i> and conductance <i>G</i> distributions\nalone. Here, we develop a simple statistical method to identify destructive\nquantum interference features (nodes) through the anticorrelation\nbetween simultaneously measured <i>G</i> and <i>S</i> values. We find these correlations can be used to unambiguously\nidentify far-detuned nodes, even when <i>G</i> and <i>S</i> distributions alone cannot. As an example, we consider\nseveral para- and meta-configured systems, including benzenediamine\nand diiodo-terphenyl-based junctions, finding that nodes can be identified\nin ensembles with broad level-alignment and lead-molecule coupling\ndistributions, and with significant anodal transport contributions,\nincluding from vacuum tunneling. The efficacy and limitations of this\nmethod are analyzed.