An important goal in supramolecular chemistry is the development of stimuli-responsive systems in which the strength of the intermolecular interactions can be altered by external signals such as changes in light, temperature, pH or voltage of an electrode. Our group is exploring the latter possibility, primarily by trying to develop systems in which electron transfer perturbs the strength of H-bonding interactions between molecules. The underlying principle is straightforward: an oxidation that decreases the negative charge on a H-acceptor (A) or a reduction that decreases the positive charge on a H-donor (D) will weaken a H-bond. Alternatively, reduction that increases the negative charge on a H-acceptor or oxidation that increases the positive charge on a H-donor will increase the strength of a H-bond. However, in the latter case, it is possible that oxidation or reduction could also lead to full proton transfer. If this occurs across the H-bond, the primary H-bonds will remain, but the secondary H-bonds will change. This can lead to an increase in unfavorable secondary interactions, which would counteract the effect of the initial proton transfer. However, with proper design, proton transfer could lead to an increase in favorable secondary interactions, which would enhance the effect of initial transfer. The goal of this project is to do the latter.

For this work, the 3 H-bond DAD array, 1, that contains a N-methyl-4,4'-bipyridinium or “monoquat” redox couple (see figure) has been synthesized. Compound 1 forms a three H-bond dimer with the non-electroactive ADA array, 2, in CH₂Cl₂. Typically, DAD-ADA arrays such as this have relatively weak association constants of ~10² M⁻¹ in non-competitive solvents such as CH₂Cl₂ due to the three, favorable primary H-bonds (shown as solid, double-headed, green arrows in the figure) being counterbalanced by four unfavorable secondary interactions (shown as dashed, double-headed, red arrows). Preliminary NMR studies suggest stronger binding in this case (>10³ M⁻¹), probably due to the greater acidity of the amide NH’s in 1 because of the strongly electron-withdrawing nature of the pyridinium. In the absence of 2, initial cyclic voltammetry studies of 1 display the expected two, sequential, 1 e⁻ reduction waves of the monoquat redox couple in CH₂Cl₂, corresponding to reduction of 1 to the radical and then the quinoidal anion. Addition of 2 results in no change in the E_1/2 of the first reduction, but the second reduction shifts ~0.24 V positive. Further addition of 2 causes no additional change in the E_1/2, consistent with a 1:1 complex. The 0.24 V positive shifts indicate a ~10⁴ increase in binding strength upon overall 2 e⁻ reduction of 1. Combined with the initial association constant of >10³ M⁻¹, this indicates an association constant of >10⁷ M⁻¹ in the fully reduced state. We believe the most likely explanation of such strong H-bonding in a 3 H-bond array is that the second reduction induces proton transfer across the central H-bond in the complex, thus converting the DAD-ADA array to a DDD-AAA array. The latter is expected to have significantly stronger H-bonding because all of the secondary interactions, in addition to the primary interactions, are favorable. If confirmed by further studies, the association constant in the reduced form would represent, to the best of our knowledge, the largest achieved to date in a redox-dependent H-bonding system, thereby demonstrating the utility of incorporating proton transfer to amplify the effect of electron transfer in these systems.