Learning from Solvophobically Driven Assemblies - Intermolecular Solvophobic Interactions
One way of looking at foldamer chemistry with its efforts in employing non-covalent interactions to stabilize collapsed conformations is classical supramolec-ular chemistry carried out intramolecularly. If molecular recognition between different molecules is the goal for the latter, recognition between different parts of a flexible molecule is the goal for the former. For specific, directional noncovalent interactions such as hydrogen bonds, the information for molecular recognition is encoded in the donor-acceptor motif. Thus, a D-D-A hydrogen bonding motif is most suitable to interact with an A-A-D motif. For nonspecific interactions such as solvophobic and van der Waals interactions, complementarity is still needed for selectivity but is required for the entire binding surfaces instead of a localized pair of interacting functional groups. As both forces are proportional to the interacting surface area, geometries (including the size, shape, and distribution) of the binding surfaces logically become the most important parameters one can use to modulate both the stability and the selectivity of interactions. We call this strategy of creating complementary solvophobic surfaces for molecular recognition geometrical manipulation. In 1991, Whitesides  pointed out that ''van der Waals and hydrophobic interactions . . . are ubiquitous in biological systems, but have been difficult to use by design in synthetic systems.'' A long road still lies ahead of chemists in approaching nature's abilities to use these nondirectional interactions for structural control. We are, however, much closer than we were 15 years ag°.
Solvophobic effects are most simply demonstrated when an organic solvent such as hexane (''oil'') is poured over an aqueous solution. In such a system, any "oil" molecule can freely approach other "oil" molecules and no geometrical constraint is present. The requirement of maximum solvophobic contact (or minimal solvophobic/solvophilic contact) naturally calls for complete phase separation of the two components, with the ''structure'' of the "product" determined by relative volumes of the two components and shape of the container used.
This is not the case when a head/tail amphiphile is dissolved in water. Contact among the hydrophobic tails is now restricted by the hydrophilic headgroups. Indeed, both the stability and the types of surfactant aggregates are determined by the two distinctively different yet inseparable parts of the molecule. Even with a simple topology, head/tail amphiphiles can form a variety of aggregates ranging from relatively simple structures such as micelles, vesicles, and reversed micelles to lyotropic liquid crystals with complex nanometer-sized phase-separated domains. Despite the large number of possibilities, the preferred aggregates can be understood from simple geometrical considerations. This is the critical packing parameter (Q = v/a0lc) proposed by Isrealachvili in which v is the volume of the hydrophobic tails, a0 the area of the hydrophilic headgroup, and lc the average critical length of the amphiphile . Spherical micelles are preferred with Q < 1/3 and cylindrical micelles with 1/2 < Q < 1/3. With even higher Q, other structures such as hexagonal, lamellar, bicontinuous cubic, and inverted hexagonal phases are favored. Simple geometrical consideration is able to predict the structure of the aggregates because maximal solvophobic contact allowed by the head and the tail is the major driving force in the system. Similar phase structures are displayed by flexible diblock copolymers and can be reliably predicted by the relative volume fractions of the individual blocks [27-31]. Even though it is not a solvophobic effect in block copolymers (as no solvent is present), minimal contact between chemically different components is the same as in surfactant aggregates and is a manifestation of the simple ''like dissolves like'' principle.
Limited geometrical manipulation gives limited selectivity. Surfactant micelles can solubilize a wide range of hydrophobic molecules with no discrimination. In order to have a higher level of structural control, rigid hydrophobic groups must be employed in the assembling process. When the hydrophobic groups involved have distinctive shape, curvature, and dimensions instead of being a simple flexible chain, only limited packing motifs are allowed in order to maximize solvopho-bic contact and complex assemblies with highly unique structures may result. This approach has long been used in rod-coil block copolymers in which the ''rod'' is a stiff, often aromatic-containing block and the ''coil'' represents a conventional flexible polymer [32-36]. Depending on the exact structures and ratio of the two blocks, various nanoscale objects such as bundles [37, 38], ribbons , tubules , and vesicles  can be formed. One such example was reported by Stupp and co-workers , which nicely demonstrated the principle of geometrical manipulation. In triblock copolymer 1, the rigid rod of biphenyl units tends to self associate into 2D structures . The polystyrene block serves to prevent infinite growth of the 2D aggregates and isolate the crosslinkable middle block of polybutadiene so that crosslinking only happens within a single aggregate. The result is anisotropic (2 x 8 nm), mushroom-shaped nanoclusters that are completely soluble in organic solvents.
Another field that frequently relies on geometrical manipulation of individual components is liquid crystalline materials . Solvents are involved in lyotropic but not in thermotropic liquid crystals. Nevertheless, the same interplay between enthalpy and entropy expressed in minimal contact between dissimilar components exists in both systems. Therefore, it is not difficult to imagine rod-like molecules such a 2 form nematic (Greek for soap, layer-like structures) phases and flat aromatic molecules with alkyl side chains (e.g. 3) form discotic phases . Geometrical control is continued to be employed by chemists in search of novel liquid crystalline phases and is especially productive in recent years when unusually shaped molecules such as dendrimers are introduced .
Recently, a number of novel nonpolymeric assemblies appeared in the literature based on geometrical manipulation of amphiphilic structures. With two dendritic headgroups and four hydrophobic tails, cone-shaped amphiphile 4 forms a structurally persistent micelle consisting of seven molecules . Unlike conventional micelles that have a distribution of sizes and highly dynamic in nature, micelles from 4 have stable, well defined structures resembling viruses and protein
80 | 3 Foldamers Based on Solvophobic Effects Scheme 3.2
aggregates. Hexabenzocoronene derivative 5 aggregates into 14-nm-diameter tubular objects, with the solvophilic tri(ethylene glycol) chains located on the inner and outer surfaces of the tube and solvophobic contacts among the aromatic and aliphatic groups to stabilize the bilayer structure . Rigid-flexible macrocycle such as 6 is particularly interesting with the hydrophilic poly (ethylene glycol) or PEG chains at the two ends of the hexaphenylene joined together by covalent bonds. Solvophobic contact between the aromatic segments is thus confined by the looped hydrophilic side chains. Macrocycles with high proportions of PEO chains assemble into donut-like nanoclusters  and those with slightly lower PEO portions give extremely long tubes by coiling of ribbon-like structures .
Geometrical manipulation may appear straightforward, but direct prediction of the final assembled structures is often difficult, as the "shapes" of most molecules cannot be represented by simple geometrical objects. Moreover, van der Waals interactions (which often play important roles in the assembling process) in rigid molecules such as fused aromatic rings not only have preferred orientations but also need to be balanced and compromised with other noncovalent forces present in the entire system. Nevertheless, the more closely the solvopho-bic surfaces can be approximated by simple geometrical shapes, the more predictable the end products will be. This is the case for mesoscale or macroscale self-assembly with components ranging from microns to centimeters in size . At such dimensions, molecular details are no longer important and assembling is driven by capillary forces - the macroscopic manifestation of solvophobic effects in the current case. 2D and even 3D objects with complex structures can be prepared predictably from components with patterned solvophobic and solvophilic surfaces. The assembling process has all the characteristics of the small-molecule counterpart including complementarity in binding surfaces, reversibility, and environmental dependency. Quite interestingly, the concept can be expanded to create mesoscopic "foldamers" with folding motifs resembling those found in natural proteins .
Was this article helpful?