Accurate determination of multiple sets of single molecular conductance of Au/1,6-hexanedithiol/Au break junctions by ultra-high vacuum-scanning tunneling microscope and analyses of individual current-separation curves.

The effect of the binding sites of the terminal groups -S on gold on currents through a single molecular junction (MJ) of Au/1,6-hexanedithiol/Au was studied by measuring current-separation (i-s) curves during repeated formation of a break junction in UHV-STM. Three different single molecular conductance (SMC) values (i.e. G(m)(HC), G(m)(MC) and G(m)(LC)) were found by a careful analysis of corrected current histograms for background tunneling currents using a previously developed robust statistical analysis. Here, HC, MC and LC represent a single MJ with high, medium and low conductance, respectively. These three SMC values are attributed to three different contact modes (i.e. strong-strong, strong-weak (or weak-strong) and weak-weak bindings at the two ends). In addition to these three SMC values due to the different contacts, another lower SMC value was newly observed in the corrected histogram. The presence of the fourth SMC is specific to MJs of alkanedithiols and is attributable to LC of a single alkylene chain with gauche rich conformation, which has a lower SMC value than that of LC with all-trans conformation as proposed previously (Fujihira M et al 2006 Phys. Chem. Chem. Phys. 8 3876). Due to the effects of the contact and the conformational change, it was difficult to determine six different SMC values corresponding to two different conformations (i.e. gauche-rich versus all-trans) with three different contacts (i.e. HC, MC and LC). In addition to this complexity, the current steps corresponding to HC, MC and LC almost always appeared in this order in measured i-s curves during separation. The current step observed here could not only be a contribution from a single molecule, but also contributions from a few groups of molecules that happen to link gold atoms of the substrate with those of the tip apex. Therefore, the SMC value for HC obtained as a peak or a set of peaks in the current histogram could be based upon the sum of the current of HC and those of MCs and LCs coexisting in parallel, unless every MJ would change successively from HC to MC and MC to LC. Namely, the currents through coexisting MCs and LCs would raise the intrinsic current observed for HC itself, while those through coexisting LCs would raise the intrinsic current for MC. To avoid such errors in determining the true SMC, we demonstrate here a new method based upon analyses of individual i-s curves referred to as jump height analyses of individual i-s curves. By this method, the true SMC of LC(all-trans) was determined to be 1.6 nS (i.e. G(m) (LC, all-trans) of 2.1 × 10(-5)G(o)) without ambiguity in spite of the possible presence of LCs(gauche rich) in parallel.