Many attempts to obtain high current rectification ratios (RRs) in molecular electronics are triggered by a potentiometer rule argument, which predicts that a strongly asymmetric location of the dominant molecular orbital yields large RR-values. Invoking this argument, molecular junctions based on alkane monothiols (CnT) can be expected to exhibit high RRs; the HOMO of these molecules is localized on the thiol terminal group bonded to one electrode. The extensive current-voltage (I-V) results for CP-AFM (conducting probe atomic force microscope) CnT junctions of various molecular lengths (n = 7, 8, 9, 10, and 12) and different metallic contacts (Ag, Au, and Pt) are consistent with conduction dominated by the HOMO, but the measured RR ∼ 1.5 is much smaller than that predicted by the potentiometer rule framework. Further, the linear shift in the HOMO position with applied bias, γ, which gives rise to rectification, is also smaller than expected, and critically, γ has the opposite sign from potentiometer rule predictions. Companion ab initio OVGF (outer valence Green's function) quantum chemical calculations provide important insight. Namely, a linear Stark shift γm is calculated for the HOMO of CnT molecules for electric field strengths (106-107 V cm-1) typical of molecular junctions, and the sign of γm matches the sign of the experimental γ for junctions derived from transport measurements, suggesting that the Stark effect plays an important role. However, the magnitude of the measured γ is only 10-15% of the computed value γm. We propose that this implies that the contacts are far from optimal; they substantially screen the effect of the applied bias, possibly via molecule-electrode interface states. We predict that, with optimized contacts, the rectification ratios in CnT-based junctions can reach reasonably high values (RR ≈ 500). We believe that Stark shifts and limited current rectification due to non-ideal contacts discussed here for the specific case of alkane monothiol junctions are issues of general interest for molecular electronics that deserve further consideration.
Bibliographical noteFunding Information:
C. D. F. acknowledges nancial support from the U.S. National Science Foundation (CHE-1708173). I. B. acknowledges nan-cial support from the Deutsche Forschungsgemeinscha (DFG grant BA 1799/3-1) and computational support by the State of Baden-Württemberg through bwHPC and the Deutsche For-schungsgemeinscha (DFG) through grant no. INST 40/467-1 FUGG.76
© 2018 The Royal Society of Chemistry.