Heteroatom Effects on Quantum Interference in Molecular Junctions: Modulating Antiresonances by Molecular Design

Controlling charge transport through molecular wires by utilizing quantum interference (QI) is a growing topic in single-molecular electronics. In this article, scanning tunneling microscopy-break junction techniques and density functional theory calculations are employed to investigate the single-molecule conductance properties of four molecules that have been specifically designed to test extended curly arrow rules (ECARs) for predicting QI in molecular junctions. Specifically, for two new isomeric 1-phenylpyrrole derivatives, the conductance pathway between the gold electrodes must pass through a nitrogen atom: this novel feature is designed to maximize the influence of the heteroatom on conductance properties and has not been the subject of prior investigations of QI. It is shown, experimentally and computationally, that the presence of a nitrogen atom in the conductance pathway increases the effect of changing the position of the anchoring group on the phenyl ring from para to meta, in comparison with biphenyl analogues. This effect is explained in terms of destructive QI (DQI) for the meta-connected pyrrole and shifted DQI for the para-connected isomer. These results demonstrate modulation of antiresonances by molecular design and verify the validity of ECARs as a simple “pen-and-paper” method for predicting QI behavior. The principles offer new fundamental insights into structure–property relationships in molecular junctions and can now be exploited in a range of different heterocycles for molecular electronic applications, such as switches based on external gating, or in thermoelectric devices.


S3
3-bromo-1-(triisopropylsilyl)pyrrole (3.30 mL, 12.7 mmol, 1 eq.) was dissolved in anhydrous THF (100 mL) in a dry flask under Ar. The mixture was degassed (Ar, 10 min) and cooled to −78 °C before n BuLi (2.0 M in hexanes, 14 mL, 28 mmol, 2.2 eq.) was added carefully over 10 min. The mixture was stirred at −78 °C for a further 15 min then dimethyl disulfide (2.5 mL, 28 mmol, 2.2 eq.) was added and the reaction was allowed to warm slowly to RT. After 1 h TLC indicated the reaction was complete, and it was quenched by careful addition of deionized H2O (30 mL). The quenched mixture was degassed (Ar, 3 h), venting the exhaust gases through an oxidizing trap, to minimize the presence of foul-smelling volatile by-products prior to work-up. The mixture was diluted with additional deionized H2O (70 mL) then extracted with CH2Cl2 (3 × 75 mL). The combined organic layers were washed with brine (2 × 75 mL) then dried (MgSO4) before the solvent was removed in vacuo affording a yellow oil (3.96 g). The crude product was purified by column chromatography (5 cm Ø, 200 mL SiO2, eluting with hexane until a higher-Rf impurity eluted * , then gradient elution from hexane to 94:6 hexane/EtOAc) which afforded 8 as a pale yellow oil, which darkened upon standing under ambient conditions (2.54 g, 74%

3-methylthiopyrrole (9)
This reaction was not conducted under dry conditions or inert atmosphere. 8 (2.53 g, 9.39 mmol, 1 eq.) was dissolved in THF (50 mL) and stirred at RT. TBAF (1.0 M in THF, 10 mL, 10 mmol, 1.06 eq.) was added and the reaction was stirred at RT for 45 min, then quenched with saturated NH4Cl(aq) (100 mL) and extracted with toluene (3 × 50 mL). The combined organic layers were washed with deionized water (3 × 50 mL) and brine (50 mL) then dried (MgSO4) before removing the solvent in vacuo to afford a brown oil (2.46 g). NMR spectroscopy and mass spectrometry indicated that this was a mixture of 9 and TIPS-OH † . Using a kugelrohr, the TIPS-OH was removed by distillation (ca. 60 °C, ca. 0.3 mbar), leaving 9 of sufficient purity for the following reactions (< 5 mol % TIPS-OH remained based on 1 H NMR integrals) as a brown oil (690 mg, 65% ‡ ). This material was used without further purification owing to concerns over the longterm stability of the N-unsubstituted pyrrole.
This reaction was not conducted under dry conditions, but care was taken to exclude oxygen. 3-(Methylthio)phenylboronic acid (375 mg, 2.23 mmol, 1.5 eq.), 4-bromothioanisole (302 mg, 1.49 mmol, 1 eq.), EtOH (4 mL), toluene (12 mL) and an aqueous solution of Na2CO3 (2 M, 3.35 mL, 6.7 mmol, 4.5 eq.) were added to a flask, stirred at RT and placed under Ar. The mixture was degassed (Ar, 45 min), Pd(PPh3)4 (86 mg, 0.075 mmol, 0.05 eq.) was added and the mixture was degassed for 5 more min, then heated to 90 °C. All solids suspended in the mixture had dissolved upon reaching this temperature. The reaction was stirred at 90 °C for 4 h, after which time TLC indicated the reaction was complete. The mixture was cooled to RT then diluted with 10 mL sat. NaHCO3. The organic layer was separated, and the aqueous layer extracted with EtOAc (3 × 20 mL). The combined organic layers were dried (MgSO4) before the solvent was removed in vacuo to afford a brown oil (576 mg). The crude product suspended on celite (10 mL, from CH2Cl2) then purified by column chromatography (3 cm Ø, 200 mL SiO2, gradient elution from hexane to 19:1 hexane/EtOAc) which afforded a white solid. Remaining impurities were removed by recrystallization; the impure material was dissolved in a minimum of CH2Cl2 upon which hexane was layered. The solvent was allowed to slowly mix and evaporate, and after several days small white crystals formed in the concentrated solution. After filtration and drying in vacuo, 3 was obtained as a white solid (154 mg, 42%). A second crop of crystals was grown from the dried filtrate in the same manner, yielding further 3 as a white solid (72 mg).

3,3'-bis(methylthio)biphenyl (4)
Based on a published procedure, 6 although a higher temperature than the reported 80 °C was required for the reaction to proceed.

This reaction was not conducted under dry conditions, but care was taken to exclude oxygen.
3-(Methylthio)phenylboronic acid (375 mg, 2.23 mmol, 1.5 eq.), 3-bromothioanisole (0.2 mL, 1.49 mmol, 1 eq.), EtOH (4 mL), toluene (12 mL) and an aqueous solution of Na2CO3 (2 M, 3.35 mL, 6.7 mmol, 4.5 eq.) were added to a flask, stirred at RT and placed under Ar. The mixture was degassed (Ar, 45 min), Pd(PPh3)4 (17 mg, 0.015 mmol, 0.01 eq.) was added and the mixture was degassed for 5 more min, then heated to 80 °C. All solids suspended in the mixture had dissolved upon reaching this temperature. The reaction was stirred at 80 °C for 18 h, after which time TLC indicated no reaction had occurred. Additional Pd(PPh3)4 (69 mg, 0.060 mmol, 0.04 eq.) was added and the temperature increased to 90 °C. After 5 h, TLC indicated the reaction was complete. The mixture was cooled to RT then diluted with sat. NaHCO3 (10 mL). The organic layer was separated, and the aqueous layer extracted with EtOAc (3 × 10 mL). The combined organic layers were dried (MgSO4) before the solvent was removed in vacuo to afford a brown solid (570 mg). The crude product was purified by column chromatography (3 cm Ø, 200 mL SiO2, gradient elution from hexane to 9:1 hexane/EtOAc) which afforded an off-white solid. Remaining impurities were removed by recrystallization; the impure material was dissolved in a minimum of CH2Cl2 upon which hexane was layered. The solvent was allowed to slowly mix and evaporate, and after several days white needles formed in the concentrated solution. After filtration and drying in vacuo, 4 was obtained as white needles (248 mg). A second crop of crystals was grown from the dried filtrate in the same manner, yielding further 4 as a white solid (72 mg). Total yield: 320 mg (87%).

2.1: Experimental methods
The calculation of the junction length was based on a previously reported method. 8 In brief, the junction displacements were calculated by correcting the piezo stretching rate and tip snapback distance. Firstly, we use the pure mesitylene solvent as a blank experiment and determine the piezo stretching rate by calibrating the direct tunneling distance distribution in mesitylene to the reported value (range from 10 −3.5 G0 to 10 −5.5 G0, 0.36 nm). We then use the determined piezo stretching rate to analyze the statistical junction displacements. The final junction displacements were obtained by adding the snapback distance (0.5 ± 0.1 nm). 8 Figure S10. Comparison of logarithmically binned conductance histograms of (a) molecules 1 and 5, 1 and (b) molecules 3 and 5, with molecular structures (above).

3.1: Computational methods
Geometry optimization: The geometry of each molecule studied was relaxed to a force tolerance of 10 meV/Å using the SIESTA implementation 9 of density functional theory (DFT), with a double-ζ polarized basis set (DZP) and the Generalized Gradient Approximation (GGA) functional with Perdew-Burke-Ernzerhof (PBE) parameterization. A real-space grid was defined with an equivalent energy cut-off of 250 Ry. To calculate molecular orbitals of gas phase molecules (table 3), we employed experimentally parameterized B3LYP functional using Gaussian g16 10 with 6-311++g basis set and tight convergence criteria.
Electron transport: To calculate electronic properties of the device, from the converged DFT calculation, the underlying mean-field Hamiltonian H was combined with the quantum transport code, Gollum. [11][12] Table S1. Torsion angle (θ) between aromatic rings of molecular wires 1-4 in the DFT-optimized junction geometry between two gold electrodes and in gas phase.

Molecule
Aryl-aryl torsion angle (θ) in gas phase molecules    Figure 3a in the main text and correspond to the ground state geometry of the junctions. The orange curve corresponds to an alternative conformation of 2 in which the dihedral angle Au-S-C-C is set to be the same as that of ground state junction geometry of 1.
The differing junction geometries of 1 and 2 ( Figure S13) mean that the dihedral angle Au-S-C-C at the anchoring point is different in each case. It is known that this angle can influence the electrical conductance of molecular wires. 13 To rule out the possibility that the relative conductance of the two isomers was significantly affected by this dihedral angle, an alternative conformation of 2 in a junction ("2p") was investigated, in which the dihedral angle Au-S-C-C was constrained to be the same as that of 1 in its ground state junction conformation. These calculations show that this new junction conformation is energetically ca. 1.25 eV less favorable than the ground state geometry. The resulting DFT-based transmission function ( Figure S15) for this new 2p configuration shows no significant difference in the HOMO-LUMO gap compared to the ground state