(600b) End-Pairing of Soluble Capped (-CH2-CH2-O-)n Oligimers Provides a Experimentally Accessible Direct Observation of a Hydrophobic Bond

Recently, Hammouda et al. (2004) showed that the aqueous solution clustering of (-CH2-CH2-O-)n polymers is sensitive to the capping of the linear chains by either ?OH or ?OCH3 or both. This suggests to us hydrophobic interaction, or bonding, of the caps in the ?OCH3 case, and further that manipulation of the capping groups might help in understanding hydrophobic interactions on a molecular scale (Pratt-2002). This would be important because hydrophobic interactions are widely expected to be central to self-assembly in aqueous solutions, and have been the target of several decades of theorizing and modeling, but so far there has been no direct mechanical observation of a primitive hydrophobic bond. Even the thermodynamic observation of isolated hydrophobic interactions has been limited (Tucker-Christian, Ashbaugh, Asthagiri-2008). Here we present extensive molecular dynamics simulations on ?OCH2CH3 capped (-CH2-CH2-O-)n oligomers that demonstrates a natural hydrophobic bond between end C-atoms, which appears distinctly in the atom-pair correlations that might be interrogated by large-angle X-ray and neutron scattering experiments. FIG. 1: Parallel tempering results for T=297.1K distinguish low-, moderate-, and high-extension regimes. The moderate-extension regime is associate with predominantly helical conformations typical of the polymer without the water medium; that regime is identified on the middle panel by the P(0)(r) obtained in a separate calculation without the water. The low-extension regime includes the contact peak on the left of these graphs, and is associated with the analogous contact-minimum in the CH4?CH4 potential-mean-force (pmf). That pmf has been studied in numerous previous simulation calculations over several decades. The solid curve represents short-distance ?Hydrophobic bonds' with window sampling that is consistent with the direct simulation results. The cumulative probability (red, right axis) shows that the interesting low-extension feature depends on about 1% of these data. That fraction reflects the compromise choice of chain length, here a 21-mer. The fraction might be higher for shorter chains but the correspondence of the low-extension feature to the CH4?.CH4 pmf might be poorer because of the stiffness of the chain. We expect that fraction to be lower for longer chains, but a comparison of this low-extension feature to the CH4?CH4 pmf should be closer.