Thiosquaramide-Based Supramolecular Polymers: Aromaticity Gain in a Switched Mode of Self-Assembly

Abstract

Despite a growing understanding of factors that drive monomer self-assembly to form supramolecular polymers, the effects of aromaticity gain have been largely ignored. Herein, we document the aromaticity gain in two different self-assembly modes of squaramide-based bolaamphiphiles. Importantly, O → S substitution in squaramide synthons resulted in supramolecular polymers with increased fiber flexibility and lower degrees of polymerization. Computations and spectroscopic experiments suggest that the oxo- and thiosquaramide bolaamphiphiles self-assemble into "head-to-tail" versus "stacked" arrangements, respectively. Computed energetic and magnetic criteria of aromaticity reveal that both modes of self-assembly increase the aromatic character of the squaramide synthons, giving rise to stronger intermolecular interactions in the resultant supramolecular polymer structures. These examples suggest that both hydrogen-bonding and stacking interactions can result in increased aromaticity upon self-assembly, highlighting its relevance in monomer design.

Figure 1

(a) Hydrogen bonding increases the aromatic character in oxosquaramide; the resonance form on the right shows increased cyclic 2π-electron delocalization in the four-membered ring. (b) Carbonyls (C=O) typically form hydrogen bonds with small deviations from the lone-pair (xy) plane, but thiocarbonyls (C=S) can engage in hydrogen bonds with C=S···H angles of close to 90°. As a result, C=O- and C=S-containing synthons are expected to promote drastically different self-assembly modes. (c) Structures of the oxosquaramide (1a and 1b) and thiosquaramide (2a and 2b) bolaamphiphiles under study. Compounds 1a and 1b self-assemble into rigid fibers (top), while 2a and 2b self-assemble into short flexible rodlike structures (bottom). This disparity is attributed to the “head-to-tail” self-assembly of the oxosquaramides 1′ versus the “stacked” self-assembly of the thiosquaramides 2′.

Figure 2

(a) Cryo-TEM images of 1a (left) and 2a (right) in aqueous solution (580 μM) after overnight equilibration. Scale bar: 100 nm. (b) Histograms of length distributions of 1a (left) and 2a (right) (N = 50, with average lengths of 235 ± 118 nm for 1a and 41 ± 18 nm for 2a). (c) Histograms of width distributions of 1a (left) and 2a (right) (N = 50, with average widths of 5.8 ± 1.2 nm for 1a and 4.8 ± 1.3 nm for 2a). (d) End-to-end distance plots (⟨R²⟩) as a function of contour length for 1a (left) and 2a (right), respectively, determined by cryo-TEM (blue open circles). Least-square fits are shown as red lines. (e) Fiber contours of 1a (left) and 2a (right) analyzed from cryo-TEM images, where initial tangents were aligned (contour lengths of 252 ± 116 nm for 1a and 77 ± 17 nm for 2a).

Figure 3

(a) Experimental SAXS profiles of 1a and 2a (5 mg mL^–1). The curves are modeled with a form factor for homogeneous and flexible homogeneous cylinders for 1a and 2a, respectively. The blue curve is shifted vertically by multiplying by a factor of 10 to enable comparison of the two profiles. (b) UV–vis spectra of 1a and 2a (c = 30 μM) in H₂O (solid line) and H₂O–CH₃CN (6:4) (dotted line).

Figure 4

Orbital interactions of neighboring monomers in HOMO–LUMO and HOMO–LUMO+1 transitions for (a) head-to-tail (for 1′) and (b) stacked (for 2′) arrangements. The direction and magnitude of the computed transition dipole moments for each transition are indicated.

Figure 5

The degree of aggregation (α_agg) plotted as a function of the volume fraction of CH₃CN as determined from UV–vis denaturation experiments for 1a (a) at λ = 330 nm and 2a (b) at λ = 384 nm. Data for the various monomer concentrations (c = 15–40 μM) were fit with the equilibrium model. Spectral data can be found in the Supporting Information.

Figure 6

IR spectrum recorded in the N–H region (inset) and amide I and amide II regions in D₂O for both 1a and 2a (5.8 mM).

Figure 7

¹H–¹³C HETCOR experiments performed at a contact time of 2048 μs on 1a (a) and 2a (b) in their polymerized (H₂O) form. Highlighted areas are described in the text.

Figure 8

(a) Computed electron density difference (EDD) maps for the head-to-tail 1′ and 2′ dimers (note the larger lobes on 1′), as well as stacked 1′ and 2′ dimers (note the larger lobes on 2′). (b) Electrostatic potential maps (MEP) of 1′ and 2′ monomers (blue indicates electron density loss and positively charged; red indicates electron density gain and negatively charged).

Figure 9

Computed geometries in implicit solvation for the isolated monomers of 1′ and 2′ (bond distances in angstroms, values in bold font), the head-to-tail hexamer of 1′ (left, averaged bond distances for each of the monomeric units, values in italic font), and the stacked hexamer of 2′ (right, averaged bond distances for each of the monomeric units, values in italic font).