Self-Assembly of a Functional Oligo(Aniline)-Based Amphiphile into Helical Conductive Nanowires

A tetra(aniline)-based cationic amphiphile, TANI-NHC(O)C5H10N(CH3)3+Br– (TANI-PTAB) was synthesized, and its emeraldine base (EB) state was found to self-assemble into nanowires in aqueous solution. The observed self-assembly is described by an isodesmic model, as shown by temperature-dependent UV–vis investigations. Linear dichroism (LD) studies, combined with computational modeling using time-dependent density functional theory (TD-DFT), suggests that TANI-PTAB molecules are ordered in an antiparallel arrangement within nanowires, with the long axis of TANI-PTAB arranged perpendicular to the nanowire long axis. Addition of either S- or R- camphorsulfonic acid (CSA) to TANI-PTAB converted TANI to the emeraldine salt (ES), which retained the ability to form nanowires. Acid doping of TANI-PTAB had a profound effect on the nanowire morphology, as the CSA counterions’ chirality translated into helical twisting of the nanowires, as observed by circular dichroism (CD). Finally, the electrical conductivity of CSA-doped helical nanowire thin films processed from aqueous solution was 2.7 mS cm–1. The conductivity, control over self-assembled 1D structure and water-solubility demonstrate these materials’ promise as processable and addressable functional materials for molecular electronics, redox-controlled materials and sensing.


Synthesis of EB TANI-PTAB
Scheme S1 -Synthesis pathway for TANI precursor

4-((diphenylmethylene)amino)-N-phenylaniline (1)
Activated molecular sieves (100 g, 3Å) were added to a solution of N-phenyl-p-phenylene diamine (1.05 eq, 27.65 g, 150 mmol) and benzophenone (1 eq, 25g, 135 mmol) in toluene (150 mL) under nitrogen atmosphere. The mixture was refluxed for 24 hours and a strong yellow colour developed. The sieves were removed, washed with THF and combined with the toluene mixture. The solvent was removed to give a brown oil that solidified into a bright yellow solid when methanol (100 mL) was added. The suspension was filtered and the filtrate concentrated, and filtered again. The combined solids were recrystallized from methanol to yield yellow crystals; with concentration of the mother liquor yielding a second crop (70 % yield). 1 6.60 (d, J = 8.6 Hz, 2H). 13

tert-butyl (4-((diphenylmethylene)amino)phenyl)(phenyl)carbamate (2)
To a solution of 1 (1 eq.) in THF under a nitrogen atmosphere was added di-tert-butyl dicarbonate (1.2 eq.) and dimethylaminopyridine (0.1 eq.). The solution was refluxed for 24hrs, after which time the mixture was allowed to cool to room temperature and ethanol (2x volume of THF) was added. The mixture was cooled in the fridge overnight and filtered, yielding fine pale-yellow needles. Concentration of the filtrate yielded a second crop of crystals (78% yield). 1

tert-butyl (4-bromophenyl)(4-((diphenylmethylene)amino)phenyl)carbamate (3)
To a solution of 2 (1 eq, 500 mg, 1.11 mmol) in dichloromethane (10 mL) was added tetrabutylammonium tribromide (1.1 eq.) in one shot. The mixture was stirred at room temperature for 1 hour before sodium sulfite (aq., 22%, 5 mL) was added. The mixture was stirred a further 30 minutes at room temperature before sodium hydroxide (2 M, 5 mL) was added with stirring. The mixture was then separated and the organic phase washed with deionised water (3 x 20 mL) followed by drying over anhydrous MgSO4. The solvent was evaporated and ditert-butyl dicarbonate (1.1 eq) and dimethylaminopyridine (0.1 eq) were added to the residue. The mixture was taken up in THF (100 mL) and refluxed overnight. After cooling to room temperature, ethanol (200 mL) was added and the mixture was cooled in the fridge. Filtration, followed by concentration of the mother liquor to recover a second crop yielded the product as fine pale-yellow needles (69 % yield

tert-butyl (4-aminophenyl)(phenyl)carbamate (4)
To a solution of 3 (1 g, 2.23 mmol, 1 eq) in a THF-methanol mixture (1:2.5, respectively) was added ammonium formate (12 eq., 1.68 g, 26.8 mmol) and palladium on carbon (10 % Pd by weight, 2.5 mol% Pd vs. 3, 60 mg) under a nitrogen atmosphere. The mixture was refluxed for 24 hours, after which time completion was confirmed by TLC (1:1 ethyl acetate:n-hexane). The reaction was cooled to room temperature and the solvent removed. The residue was taken up in dichloromethane (50 mL) and filtered through a plug of celite. The celite was washed with DCM and the filtrate evaporated. The residue was stirred in n-hexane (150 mL) for 1 hour followed by filtration and washing with hexane to yield the product as an off-white powder (96 % yield

6-bromo-N-(4-((4-((4-(phenylamino)phenyl)amino)phenyl)amino)phenyl)hexanamide (8a)
To a solution of 7 (400 mg, 0.47 mmol, 1 eq.) in anhydrous dichloromethane (40 mL) under a nitrogen atmosphere was added trimethylsilyliodide (0.34 g, 1.7 mmol, 3.6 eq.) dropwise. The mixture was stirred for 1 hour before anhydrous methanol (approx. 0.5 mL) was added dropwise causing precipitation of a pale solid. Methanol addition was stopped when no further precipitate was produced, and the mixture was stirred for a further 30 minutes. Triethylamine (1 mL) was added causing a pale purple colour to develop. This mixture was stirred for 15 minutes before centrifugation and washing with diethyl ether. After 3 washing and centrifugation steps, the supernatant was discarded and the off-white LEB-state Boc-deprotected solid product 8a was dried in vacuum overnight, and stored under an inert atmosphere (74% yield). 1

Supplementary analysis
For comparison to computationally modelled UV-Vis transitions, the UV/Vis spectrum of EB TANI-PTAB was recorded in acetonitrile, to avoid the affects of self-assembly observed in aqueous solution.  Figure S7. CD spectrum of EB TANI-PTAB (4 mM, aqueous). Observed signal closely resembles the UV-Vis spectrum of EB TANI-PTAB, and is suggested to arise from macroscopic sample alignment 1 that was inadvertently induced during filling of the narrow pathlength cuvette required to study the very highly absorbing solution and obtain reliable data. No bisignate features occur for the EB state material.

Computational chemistry: DFT modelling of EB TANI-PTAB Computational details
Calculations which accurately capture the structural and electronic properties of aniline oligomers remain a significant challenge, 2 and were undertaken here to aid the interpretation of experimentally observed properties, focusing on only the simpler emeraldine base (EB) oxidation state of TANI, which has been treated by similar methods previously. 3 The doped species, TANI(CSA)2-PTAB, was not considered, as the modelling of charged radical species with associated anions poses additional computational challenges, outside the scope of the present work. Calculations were performed to support the assignments of experimentally observed UV-Vis transitions, and to predict the orientation of the associated transition dipole moments, for EB TANI-PTAB. The geometry of EB TANI-PTAB was optimised using the B3LYP 4-7 functional with a standard 6-31G* basis set but with only the five spherical harmonic components of the polarization functions and a polarizable continuum model (PCM) solvation field 8,9 with water as solvent, as implemented in Gaussian 09, revision D.01. 10 Timedependent DFT (TD-DFT) calculations with the CAM-B3LYP 11 functional (as suggested by Tozer et al. [12][13][14] for complexes with increased charge transfer character) were then used to calculate the excitation energies for the first 20 singlet transitions, using the B3LYP-optimised geometry as input. From these calculations the transition dipole moments were also extracted. Visualisation of the transition dipole moment in 3D space was accomplished using GabEdit 15 to overlay the (x, y, z) coordinates of the transitions as produced by TD-DFT calculations on the optimised Cartesian coordinates of TANI-PTAB.
The output of TD-DFT calculations for the first 10 excited states is detailed below, with images of the contributing molecular orbitals (MOs) ( Figure S7). While one would expect the TD-DFT to capture well what is happening experimentally, i.e. several MOs of appropriate symmetry contributing to transitions to an excited state, it is worth noting that the standard (non-TD) B3LYP calculation gives a single point HOMO-LUMO gap in good agreement with experimentally observed transitions. This is no longer the case when performing a CAM-B3LYP single point calculation on the same geometry (Table 1). Additionally, CAM-B3LYP optimised geometries are generally in poorer agreement with available structural data (unpublished results). In contrast, TD-DFT calculations with CAM-B3LYP on the B3LYP optimised geometry give a closer match to the experimentally observed transition energies than do TD-DFT calculations with B3LYP, and this agrees with in-house benchmarking for different functionals (unpublished). We have therefore used this mixed method approach.