a compact globular geometry (Fig. 12b) is preferred both in vacuum and in solution.
Unfortunately, most of these calculations were performed with Cl
À counterions in
H 2 O, whereas our experimental investigations and Goddard’s computations were
carried out using PF
À
6 counterions in MeCN. Thus, a perfect comparison between
the computational methods cannot be made. However, one set of SA MD simulations with explicit MeCN solvent and PF
À
6 counterions still led to a collapse of the
π-stacked structure after ~500 ps, although it was partially recovered (up to 75%)
again after ~2,500 ps when using an OPLS-AA force field that applied more diffuse
charges to the pyridinium rings of CBPQT
4+ . Franco et al. concluded that the folded
secondary structure with extended D–A stacking is stabilized by crystal packing
effects rather than the weak forces such as charge transfer, π–π stacking, and the
C–H Á Á Á O hydrogen bonding interactions that we have claimed [114] persist in
solution. We side with the secondary structure (Fig. 12a) proposed by the Goddard
group because it agrees with the chemical shift data we observe over and over again
by
1 H NMR spectroscopy. We acknowledge that the information provided by
1 H
NMR spectra represents an average of all the dynamic structures that are sampled in
solution; indeed, the globular co-conformations (Fig. 12b) described by Franco
et al. most likely make fleeting contributions to the dynamic solution-state superstructures. We note, however, that the protons in the unencircled DNP unit of the
structure in Fig. 12b should resonate at higher frequencies than an isolated DNP unit,
on account of its location orthogonal to the shielding cone of the nearby BIPY
2+ unit.
Since the average chemical shifts of all alongside DNP are observed to resonate
uniformly and consistently at lower frequencies than their isolated counterparts
(Sect. 3.2), we conclude that the π-stacked superstructure (Fig. 12a) outcompetes
the globular superstructure (Fig. 12b), at least at 233 K in MeCN.
The (contradictory) results of modeling these charged D–A mechanically interlaced foldamers presents an enticing challenge to the chemical theory community to
build a robust methodology for predicting accurately the behavior of these secondary structures mediated by relatively strong intramolecular noncovalent bonding
interactions in polar solvents in the presence of soft counterions.
5 Conclusions and Outlook
We have described a class of oligo- and polyrotaxanes and pseudorotaxanes that
adopt well-defined folded secondary structures in the solid state and in solution as a
result of stabilizing donor–acceptor charge transfer interactions between the aromatic recognition units, aided and abetted by multiple C–H Á Á Á O interactions
between polyether chains and bipyridinium protons. In the solid state, oligomers
above a certain critical chain length (between three and five repeating units)
crystallize into a lattice that is essentially indistinguishable from an infinite polymer, offering a predictive glimpse at hypothetical polymers that are much more
difficult to obtain as single crystals. When these types of pseudorotaxanes are
290
C.J. Bruns and J.F. Stoddart
Unfortunately, most of these calculations were performed with Cl
À counterions in
H 2 O, whereas our experimental investigations and Goddard’s computations were
carried out using PF
À
6 counterions in MeCN. Thus, a perfect comparison between
the computational methods cannot be made. However, one set of SA MD simulations with explicit MeCN solvent and PF
À
6 counterions still led to a collapse of the
π-stacked structure after ~500 ps, although it was partially recovered (up to 75%)
again after ~2,500 ps when using an OPLS-AA force field that applied more diffuse
charges to the pyridinium rings of CBPQT
4+ . Franco et al. concluded that the folded
secondary structure with extended D–A stacking is stabilized by crystal packing
effects rather than the weak forces such as charge transfer, π–π stacking, and the
C–H Á Á Á O hydrogen bonding interactions that we have claimed [114] persist in
solution. We side with the secondary structure (Fig. 12a) proposed by the Goddard
group because it agrees with the chemical shift data we observe over and over again
by
1 H NMR spectroscopy. We acknowledge that the information provided by
1 H
NMR spectra represents an average of all the dynamic structures that are sampled in
solution; indeed, the globular co-conformations (Fig. 12b) described by Franco
et al. most likely make fleeting contributions to the dynamic solution-state superstructures. We note, however, that the protons in the unencircled DNP unit of the
structure in Fig. 12b should resonate at higher frequencies than an isolated DNP unit,
on account of its location orthogonal to the shielding cone of the nearby BIPY
2+ unit.
Since the average chemical shifts of all alongside DNP are observed to resonate
uniformly and consistently at lower frequencies than their isolated counterparts
(Sect. 3.2), we conclude that the π-stacked superstructure (Fig. 12a) outcompetes
the globular superstructure (Fig. 12b), at least at 233 K in MeCN.
The (contradictory) results of modeling these charged D–A mechanically interlaced foldamers presents an enticing challenge to the chemical theory community to
build a robust methodology for predicting accurately the behavior of these secondary structures mediated by relatively strong intramolecular noncovalent bonding
interactions in polar solvents in the presence of soft counterions.
5 Conclusions and Outlook
We have described a class of oligo- and polyrotaxanes and pseudorotaxanes that
adopt well-defined folded secondary structures in the solid state and in solution as a
result of stabilizing donor–acceptor charge transfer interactions between the aromatic recognition units, aided and abetted by multiple C–H Á Á Á O interactions
between polyether chains and bipyridinium protons. In the solid state, oligomers
above a certain critical chain length (between three and five repeating units)
crystallize into a lattice that is essentially indistinguishable from an infinite polymer, offering a predictive glimpse at hypothetical polymers that are much more
difficult to obtain as single crystals. When these types of pseudorotaxanes are
290
C.J. Bruns and J.F. Stoddart