Single-molecule conductance of a chemically modified, π-extended tetrathiafulvalene and its charge-transfer complex with F4TCNQ

We describe the synthesis and single-molecule electrical transport properties of a molecular wire containing a π-extended tetrathiafulvalene (exTTF) group and its charge-transfer complex with F4TCNQ. We form single-molecule junctions using the in situ break junction technique using a homebuilt scanning tunneling microscope with a range of conductance between 10 G0 down to 10−7 G0. Within this range we do not observe a clear conductance signature of the neutral parent molecule, suggesting either that its conductance is too low or that it does not form a stable junction. Conversely, we do find a clear conductance signature in the experiments carried out on the charge-transfer complex. Due to the fact we expected this species to have a higher conductance than the neutral molecule, we believe this supports the idea that the conductance of the neutral molecule is very low, below our measurement sensitivity. This idea is further supported by theoretical calculations. To the best of our knowledge, these are the first reported single-molecule conductance measurements on a molecular charge-transfer species.


General methods
Reagents for synthesis were used as purchased. Argon was used as inert gas when necessary. Column chromatography was carried out using silica gel (40-63 μm) from Fluka. Analytical thin layer chromatography (TLC) was done using aluminium sheets (20 x 20 cm) precoated with silica gel RP-18W 60 F254 from Merck, or aluminium foils (20 x 20 cm) covered with nano-silica gel from Fluka.
UV active compounds were detected with a UV lamp from CAMAG at wavelengths λ = 254 or 366 nm. NMR spectra were recorded on a Bruker Avance 300 spectrometer at 298 K using partially deuterated solvents as internal standards. Coupling constants (J) are denoted in Hz and chemical shifts (δ) in ppm. Multiplicities are denoted as follows: s = singlet, d = doublet, t = triplet, m = multiplet, br = broad. Matrixassisted laser desorption ionization (coupled to a time-of-flight analyzer) experiments (MALDI-TOF) were recorded on a Bruker REFLEX III spectrometer. Infrared spectra were recorded in a Bruker Tensor 27 spectrophotometer with an ATR device. Absorption spectra were recorded on a Shimadzu UV-vis-NIR 3200 spectrophotometer. Electrochemical measurements were carried out on an Autolab PGSTAT30 potentiostatgalvanostat equipped with an electrochemical analysis software for windows version 4.8, in a 3-electrode single compartment cell. A glassy carbon electrode (GCE) was used as working electrode; as the counter electrode a platinum wire was used; the reference electrode was a solution of Ag/Ag + in acetonitrile.
Compounds 1-3 [1], and 4 [2] were prepared following the reported procedures and showed the same characterization properties as described therein.

Synthesis and characterization of compounds 5
To

Break junction experiments
A home-built scanning tunneling microscope (STM) as described previously [3] was used. As substrates commercially available gold samples on quartz (Arrandee), and as tip a freshly cut gold wire was used. For all measurements, a bias voltage (V), of 200 mV was applied between the tip and the substrate. A linear current-to-voltage converter with two amplification stages was used to measure the current (I) in the circuit. For the measurements presented in Figure   3 and Figure   S8 Figure S5: The 2D histograms of the same data as in Figure S3. We notice in experimental run (a) a less prominent feature close to 10 −3 G 0 than in (b). The low conductance group, however, appears very similar in the two runs. This strengthens the idea that the low group corresponds to transport across the CT complex from thiol to thiol. In both runs the percentage of plateaus in the low conductance group was 10% (the total number of plateaus are: run a = 637/6443 (10%); run b = 485/4868 (10%).

Theoretical methods
We employed the DFT-based transport method described in detail in reference [4], which is built on the quantum chemistry code TURBOMOLE 6.1 [5]. In all our calculations we used the BP86 exchange-correlation functional [6] and a split-valence basis set with polarization for all non-hydrogen atoms def-SVP [7].
In order to construct the junction geometries, we first relaxed the molecules in the gas phase. Then, we placed the relaxed molecules between two finite clusters of 20 (or 19) gold atoms and performed a new geometry optimization in which the molecule and the four (or three) outermost gold atoms on each side were relaxed, while the other gold atoms were kept frozen. The total energies were converged to a precision of better than 10 −6 atomic units, and structure optimizations were carried out until the maximum norm of the Cartesian gradient fell below 10 −4 atomic units. Subsequently, the size of the gold clusters was extended to about 120 atoms on each side in order to describe the metalmolecule charge transfer and the energy level alignment correctly. The information obtained on the electronic structure of the junctions within DFT was then transformed into the different transport properties by using non equilibrium Green's function techniques, as described in detail in reference [4]. To overcome the well-known problems related with the HOMO-LUMO gap underestimation in DFT, we applied the DFT + Sigma method [8] as described in detail in reference [9].