First, the rotational barriers of diol 1 and diether 2 were computationally modelled to verify the presence and influence of an intramolecular hydrogen bond in the planar transition state. The DFT method M06-2X/6-31G* was selected as it provides a good balance of accurate steric and hydrogen bonding energies. The symmetry of the atropisomeric N,Ndiimide platforms enabled analysis of a substructure containing a single N-arylimide rotor (Fig. 2), which greatly shortened calculation times. A dihedral angle driver was used to systematically calculate the energy of structures along the rotational barrier energy surface (Fig. 3) and the GS and TS structures were identified. The calculated rotational barriers of the phenol and benzyl ether (18.6 and 26.4 kcal/mol) rotors were very similar to the experimental values (21.3 and 27.5 kcal/mol). More importantly, the computational studies were able to replicate the significantly lower rotational barrier of the phenol rotor.

Fig. 2 Calculated GS and TS structures (M06-2X/6-31G*) and barrier heights for the phenol N-arylimide rotors allowed to form (blue lines) or constrained to prevent (black lines) the formation of intramolecular hydrogen bonds.

The computational studies also enabled examination of the role of intramolecular hydrogen bonding on the rotational barrier. In support of our hypothesis, a strong well-defined hydrogen bond was formed in the transition state (Fig. 2) as evidenced by an extremely short =O-to-H distance (1.557 Å). For comparison, the barrier was also calculated for a geometry in which the OH group was constrained (Fig. 2, black lines) to point away from the imide carbonyl and was unable to form an intramolecular hydrogen bond. The barrier for the nonhydrogen bonding phenol rotor was much higher (26.4 kcal/mol) in comparison to the unconstrained hydrogen bonding phenol rotor (18.6 kcal/mol). The higher barrier of the non-hydrogen bonding geometry was very similar to the benzyl ether rotor (26.4 kcal/mol) confirming that the phenol and ether substituents have similar steric effects on the TS in the absence of a hydrogen bond. This confirmed that the differences in rotational barriers between diol 1 and diether 2 were primarily due to the ability of diol 1 to form an intramolecular hydrogen bond.

Interestingly, the phenol rotor could also form an intramolecular hydrogen bond that stabilized the ground state (Fig. 2). However, this GS hydrogen bond was much weaker than the TS hydrogen bond as was evident from its greater length (1.856 Å) and non-linear geometry (144°). Corroboration was provided by a comparison of the ground states and transition states for the constrained and unconstrained systems (Fig. 2, grey arrows). The GS hydrogen bond was 3.0 kcal/mol; whereas the TS hydrogen bond was 10.8 kcal/mol.

The addition of various hydrogen bonding guests, provided a unique means to control the rotational barrier of diol 1 and also verified the importance of the TS hydrogen bond. Small molecule guests were used as opposed to changing the bulk solvent to minimize variations in solvent polarity or viscosity. The guests compete for and disrupt the intramolecular hydrogen bond raising the rotational barrier, as shown in Fig. 1b. Small molecule hydrogen bonding guests were selected with varying hydrogen bond accepting abilities: acetic acid, acetone, THF, pyridine, triethylamine (TEA), and DMSO.

The rates of rotation were measured by monitoring the syn/anti ratio of anti-enriched samples of diol 1 and diether 2 as they approached equilibrium in solution at 18.2 °C and 70.1 °C, respectively. Anti-enriched samples of diol 1 were prepared by crystallization. 32 Anti-enriched samples of diether 2 were prepared by treating the anti-enriched diol 1 with benzyl chloride and potassium carbonate in dry DMF at 0 °C. 32 All of the measurements were carried out in a solvent (72:28 (v/v), benzene-d6:acetonitrile-d3) that would not disrupt the hydrogen bonding interactions, while yielding distinct peaks in the 1 H NMR spectra for the syn-and anti-conformers of diol 1 and diether 2.

In each case, the addition of the hydrogen bonding guests (200 equivalents) increased the rotational barrier of diol 1 (Fig. Please do not adjust margins

Please do not adjust margins

3). By comparison, the rotational barrier for diether 2 did not change significantly on addition of 200 equivalents of the same hydrogen bonding guests. The sharp contrast in the sensitivities of diol 1 and diether 2 to the addition of hydrogen bonding guests were consistent with the proposed mechanisms. For diol 1, the hydrogen bonding guests disrupted the TS stabilising intramolecular hydrogen bond raising the rotational barrier. For diether 2, the hydrogen bonding guests have little or no effect on the rotational barrier as diether 2 cannot form an intramolecular hydrogen bond. Quantitative analysis of the effects of each guest provided additional support for the role of the intramolecular hydrogen bond in diol 1. The measured rotational barriers in the presence of each guest were plotted against the hydrogen bond accepting ability of the guest, as measured by the Kamlet-Taft β parameter (Fig. 3). An excellent linear correlation was observed between the rotational barriers of diol 1 and the β parameters of the guests. DMSO, which was the strongest hydrogen bond acceptor in the series, showed the most dramatic effect raising the barrier of diol 1 by 2.3 kcal/mol. Weaker hydrogen bond acceptors such as acetic acid, THF, and acetone much smaller effects raising barrier by only 0.6-0.7 kcal/mol. Fig. 3 Correlation of the measured rotational barriers of diol 1 (diamonds) and diether 2 (squares) in the presence of 200 eq of guests against the Kamlet-Taft hydrogen bond accepting parameter (β) of the guests. The rotational barriers were measured in 72:28 (v/v) benzene-d6:acetonitrile-d3 at 18.2 and 70.1 °C, respectively. The error in the measurements were within the data markers. Acetone was not tested against diether 2 because its boiling point is lower than 70.1 °C. The points labelled no guest were assigned the β of acetonitrile.

To test the possibility that the guests were raising the rotational barriers of 1 by hydrogen bonding to and stabilising the ground state, 1 H NMR titration studies were conducted. The measured association constants of DMSO and acetone for diol 1 in benzene-acetonitrile (72:28, v/v) were 8.9 x 10 -1 and 1.0 x 10 -2 M -1 , respectively. These association constants are very low (Ka < 1). Thus, the stabilization of the ground states by the guests is small or negligible and could not explain the dramatic changes in the rotational barriers in the presence of these guests.

The dynamic hydrogen bonding interactions of the guests enabled the reversible turning ON and OFF of the braking effect. For this demonstration, CD2Cl2 was selected as the solvent system so that the polar hydrogen bonding guest (DMSO) could be removed by aqueous extraction. The rate of rotation of a solution of anti-enriched 1 in CD2Cl2 was measured for two complete ON-OFF cycles by adding and removing DMSO (Fig. 4). Due to the rapid isomerization of diol 1 in the absence of guest, this study was started with the brake engaged (with DMSO) and the temperature was kept at -10 °C for the entire experiment. The rate of rotation sped up and slowed down by two orders of magnitude with the removal and then addition of DMSO. After two cycles, the diol reached the syn/anti equilibrium ratio, and the rate of interconversion could no longer be observed. In conclusion, the rotational barrier of a N,N'diarylnaphthalenediimide molecular rotor can be tuned using guests with different hydrogen bonding abilities. The key to this system is the presence of an intramolecular hydrogen bond in the transition state that catalyses bond rotation. The guests disrupt the intramolecular hydrogen bond, raising the rotational barrier. The magnitude of the rate acceleration could be modulated by the hydrogen bond accepting ability and number of equivalents of the hydrogen bonding guest. In future work, we aim to use this rotor as an enzyme model system to study the role of transition state stabilization by hydrogen bonding.