Characterizing the Metal–SAM Interface in Tunneling Junctions

. This paper investigates the influence of the interface between a gold or silver metal electrode and an n -alkyl SAM (supported on that electrode) on the rate of charge transport across junctions with structure Met(Au or Ag) TS by comparing measurements of current density, J (V), for Met/AR = Au/thiolate (Au/SR), Ag/thiolate (Ag/SR), Ag/carboxylate (Ag/O 2 C), and Au/acetylene (Au/C ≡ CR), where R is an n -alkyl group. Values of J 0 and β (from the Simmons equation) were indistinguishable for these four interfaces. Since the anchoring groups, A, have large differences in their physical and electronic properties, the observation that they are indistinguishable in their influence on the injection current, J 0 (V = 0.5) indicates that these four Met/A interfaces do not contribute to the shape of the tunneling barrier in a way that influences J (V). (Met TS -AR) based on 0 (V) and β for three groups, A, chosen to be very different in their electronic and geometrical structure. We measured J 0 (V) and β for five SAMs formed by allowing terminal alkynes (1-hexyne, 1-octyne, 1-decyne, 1-dodecyne, 1-tetradecyne) to react with gold, and compared the rates of charge tunneling through these alkyne-based junctions to those through junctions composed of n -alkanethiolates 26 and n -alkanecarboxylates 27 of comparable lengths on gold and silver. SAMs with composition AuSR and AuC ≡ CR (R = n -alkyl) have very similar geometrical structure. 28 we assume transport by hole tunneling). The results indicate that the HOMO for these two types of junctions is localized on the anchoring atoms (here, -S- and -C ≡ C-); this

Understanding the relationship between the structure of the insulating organic molecules in junctions of the form Met TS /A(CH 2 ) n T//Ga 2 O 3 /EGaIn, and rates of charge transport across these junctions by tunneling, requires understanding the influences of the interfaces between the electrodes and the self-assembled monolayer (SAM). [1][2][3][4] (Here, Met is the "metal" electrode and A and T are "anchoring" and "terminal groups.") The supporting information summarizes previous studies of this and other relevant systems of SAM-bound tunneling junctions, generally organized in terms of the injection current, J 0 (V), and the attenuation parameter, β, of the simplified Simmons equation, 5 Eq. 1.
J(V) = J 0 (V)e -βd = J 0 (V)10 -βd /2.303 (1) Determining the influence of the interface between the SAM and electrode on the shape (for a simple rectangular barrier, the height and width) of the tunneling barrier, and of the current density across that barrier, has motivated a number of investigations. [6][7][8][9][10][11][12][13][14][15][16][17][18] Recent studies of singlemolecule break junctions have been interpreted to indicate that the presence of covalent Au−C σ-bonds-formed using trimethyltin (-SnMe 3 )-terminated n-alkyl groups, 19,20 and SnMe 3terminated aromatics [19][20][21] or trimethylsilyl (TMS)-terminated conjugated systems 22 -increases rates of charge transport across these junctions by approximately a factor of 10-100, relative to amine or thiolate anchoring groups. One possible inference from the increase is that the Au−C σ-bond, and the absence of resistive anchoring heteroatoms, increases "conductivity" (although the meaning of this word is not entirely clear for tunneling junctions). Other reports have suggested that the strength of the interaction between the anchoring atom and the metal electrodes influences rates of charge transport, with stronger binding interactions (i.e. Au/SR and Au/NH 2 R) leading to higher measurements of conductance than weaker interactions (i.e. Au/O 2 CR and Au/NCR). 23,24 In contrast, using a large-area junction, Cahen and coworkers established that a substantial difference between two types of bonds between the electrode and the SAM did not influence rates of tunneling. 25 Using n-alkyl-SAMs on silicon, and Hg top electrodes, they demonstrated that a change in the interaction at the SAM-Hg interface-from a van der Waals interaction (-CH 3 //Hg) to a covalent bond (S/Hg)-did not change rates of charge transport. 25 This paper summarizes a study of the so-called "bottom" (Met TS -AR) interface, based on characterizing J 0 (V) and β for three groups, A, chosen to be very different in their electronic and geometrical structure. We measured J 0 (V) and β for five SAMs formed by allowing terminal alkynes (1-hexyne, 1-octyne, 1-decyne, 1-dodecyne, 1-tetradecyne) to react with gold, and compared the rates of charge tunneling through these alkyne-based junctions to those through junctions composed of n-alkanethiolates 26 and n-alkanecarboxylates 27 of comparable lengths on gold and silver. SAMs with composition AuSR and AuC≡CR (R = n-alkyl) have very similar geometrical structure. 28 We conclude that the rate of tunneling transport through Ag/(AgO x /)O 2 CR, Ag/SR, Au/SR, and Au/C≡CR interfaces are-using an "EGaIn" top electrode (that is, Ga 2 O 3 /EGaIn)indistinguishable. This work indicates that in these large-area junctions, the details of the chemical binding at the Met/A interface do not significantly influence the injection current or current density: remarkably, the total variation in J 0 across the four systems examined ( Figure 1) is no more than a factor of 2. This result does not require that there be no differences in the electronic structure of the interface, or that differences in the Met/A interfaces do not influence the shape of the tunneling barrier associated with those interfaces; rather, they demonstrate that these differences-whatever their nature-do not influence tunneling currents.

RESULTS AND DISCUSSION
In contrast to the ionic interaction of Ag/O 2 CR 27, 29-31 and the covalent, but weak (~30 kcal/mol, 32  This study also demonstrated that these SAMs are susceptible to oxidation during (and possibly after) formation of the Au-C≡CR bond, and require careful handling (see the supporting information for more details). In this current work, SAMs were prepared by immersion of template-stripped gold substrates in an anhydrous solution of n-alkyne (~6 mM in hexadecane) for 48 hours at room temperature under an atmosphere of nitrogen. We monitored the SAM for oxygen contamination using both X-ray photoelectron spectroscopy (XPS) and contact angles with water ( Figure S1 and Table S1). Electrical measurements were performed using EGaIn (eutectic Ga-In; 74.5% Ga, 25.5% In) top electrodes over a potential window of ±0.5 V ( Figure   2). (We and others have described the EGaIn electrode extensively. 26,[38][39][40][41][42] )

Comparing the Electrical Properties of n-Alkyl SAMs having A = -C≡C-and A = -S-on Gold
Electrodes. Figure

Comparing the Electrical Properties of n-Alkyl SAMs having A = -C≡C-, A = -S-, and A = -O 2 C-on Gold and Silver Electrodes.
Comparisons of the results summarized in Figure 2 with those published previously, 3 Table I) report a factor of 10 difference in the conductance between Au-C and Au-S contacts using STM, and Venkataraman and coworkers (ref 21 in Table   I) report a factor of 100 difference in conductance between Au-C and Au-NH 2 contacts using STM. It is not clear that one can extract a consensus from these data concerning the influence of the group A on tunneling across junctions having a Met/A interface.
The conclusion from the work presented here-that a change in the Met/A interface does not influence rates of charge tunneling-is, however, in agreement with one other large-area junction measurement. Cahen and coworkers observed-using a Hg large-area junction (ref 14 in Table   I Table I); they changed the bottom electrode/SAM covalent contact (from Au-CNR to Au-SR, where R = oligoacenes), while keeping the van der Waals top contact unchanged, and reported just a factor of ~3 difference in conductance between the two interfaces using CP-AFM. iii) The nature of the contacts between the molecules and the metal in break junctions is not defined, and the structure of the metal tip is also unclear. These uncertainties limit the ability to compare large-area and break-junction measurements.

CONCLUSIONS
The

Supporting Information
Detailed experimental procedure, histograms of current densities, and summary of junction measurements. This material is available free of charge via the Internet at http://pubs.acs.org.