Introduction

Stereocomplexation, which involves physical blending of both the enantiopure forms of a polymer, is a powerful strategy for improving polymer properties1,2,3,4. Stereocomplexes possess higher melting points5,6, mechanical strengths7, storage moduli8, crystallinities9, solvent resistances4, thermal stabilities10,11 and hydrolytic stabilities12 than the parent enantiopure polymers, which make them attractive materials for advanced applications13,14,15,16,17,18,19. However, a major issue associated with stereocomplexation is inefficient complexation due to (i) mismatch of the lengths of the two polymer chains of opposite chirality; (ii) flexibility-driven looping, knotting, coiling of individual chains; and (iii) bundling of chains of the same chirality leading to phase-separated chiral domains (Fig. 1a)20,21,22,23. For effective stereocomplexation, proper pairing of chains of equal sizes but opposite chirality is essential, which is improbable when two enantiopure polydisperse polymers are mixed postsynthetically. Additionally, the attractive noncovalent interactions between these chains of opposite chirality should overpower the intrachain interactions leading to coiling, looping, knotting, etc., and the attractive interchain interactions between chains of similar chirality leading to bundling to maintain their hybridization throughout their length. Therefore, in reality, by mixing two enantiomeric polymers, achieving a perfect and uniform stereocomplex is almost impossible, and this often results in imperfect and locally anisotropic stereocomplexes containing domains of pure chiral bundles (homocrystallites) and complexes of mismatched chain lengths. A significant milestone in this field would be to devise methods that avoid mixing preformed polymers but yield polymer stereocomplexes directly from monomers24,25,26,27,28,29. Polymerization of a racemic monomer with a chiral center in solution would yield 2n different stereoisomeric polymers for each n-mer. Although there are reports on the solution synthesis of stereocomplexes from racemic monomers using special catalysts/conditions, they are not generally applicable and pose difficulty in the unambiguous characterization of the polymer30,31,32,33. An ideal case would be the topochemical polymerization of an achiral monomer, which generates new chiral centers and yields equal amounts of two enantiopure polymers of opposite chirality within a crystal (Fig. 1b). Here, we report the topochemical synthesis of a perfect 1:1 stereocomplex polymer from an achiral monomer. Remarkably, the polymerization occurred in a single-crystal-to-single-crystal (SCSC) manner, which allowed us to unequivocally characterize the resultant stereocomplex at atomic resolution via single-crystal X-ray diffraction analysis.

Fig. 1: Schematics of conventional method and hypothesis of the work.
figure 1

a Conventional method of polymer-stereocomplexation showing the imperfect blending. b Hypothesis of the current work.

Full size image

Results

Topochemical polymerization is a powerful method for producing crystalline polymers in a green, solvent-free, and catalyst-free process34,35,36,37. The recently introduced topochemical ene-azide cycloaddition (TEAC) polymerization38,39,40 is more attractive because it creates a new chiral center in the triazoline linkage formed in the reaction. So far, TEAC polymerization has been employed only for the polymerization of a few peptides, and in all these cases, the cycloaddition reaction proceeded stereospecifically, as dictated by the molecular packing, which in turn was dictated by the chirality of the peptide monomer. We were curious to know the stereochemical outcome of TEAC polymerization of a meso-monomer: it can either yield two stereoregular polymers of opposite chirality (stereocomplexed polymers) or produce a polymer with alternately placed R and S repeat units to maintain the overall symmetry of the system. Myo-inositol functionalized with complimentary reactive groups (CRGs) at carbon atoms positioned on its plane of symmetry are known to arrange as stacks of 2D hydrogen-bonded sheets in crystals, placing the CRGs in close proximity in a head-to-tail fashion along the stacking axis (Fig. 2a)41,42. This fact prompted us to use this scaffold for designing meso-monomer M decorated with ene and azide for TEAC polymerization (Fig. 2b). Apart from our interest in the stereochemical outcome in the reaction of a meso-monomer, the use of skeletons other than peptides would establish the generality of the TEAC reaction.

Fig. 2: Packing arrangement of monomer M in its crystal.
figure 2

a Schematic representation of a meso myo-inositol derivative, its 2D H-bonded sheet and stacking of sheets leading to head-to-tail alignment. b Chemical structure of monomer M. c Conformational enantiomers of the monomer in the crystal. d H-bonded sheet in the ‘ac’ plane. Hydrogen atoms, other than those involved in H-bonding, are hidden for clarity.

Full size image

We synthesized meso-monomer M from the commercially available myo-inositol (Fig. 2b, Supplementary Figs. 1, 7–18). We obtained block-like single crystals of monomer M by slow evaporation of its saturated methanol solution over a period of 1–2 days (Supplementary Fig. 2). Single-crystal X-ray diffraction (SCXRD) analysis revealed that monomer M crystallized in the triclinic P−1 space group (Supplementary Fig. 3, Supplementary Tables 1 and 2) with one molecule in the asymmetric unit. The cyclohexyl ring adopts a chair conformation in the crystal. The β-N of the azide is involved in the intramolecular azide…oxygen interaction with the neighboring hydroxyl group43. Similarly, the alkenyl hydrogen is involved in the intramolecular C–H…O hydrogen bond with a nearby hydroxyl group. Although monomer M is a meso compound, in its crystal, it adopts two nonsymmetric conformations (conformational enantiomers), which are mirror images of each other (Fig. 2c). As anticipated, the monomer molecules arranged as H-bonded 2D sheets by utilizing its vicinal diols for zig-zag patterned H-bonding (Fig. 2d). Conformational enantiomers are self-sorted collinearly along the ‘a’ axis via weak interactions, forming molecular arrays, each of which containing only a particular conformational enantiomer. Along the ‘c’ axis, the chains of opposite conformational enantiomers are aligned alternately (Fig. 3). The zig-zag-patterned hydrogen bonding between the diol units of adjacent chains positions them in a slipped arrangement. The hydrogen-bonded sheets are perfectly stacked along the ‘b’ axis with the help of weak C–H…N, C–H…O, and van der Waals interactions (Fig. 3, Supplementary Table 2), such that each molecular column along the ‘b’ axis contains only one of the two conformational enantiomers. This stacking arranges the monomer molecules in a head-to-tail fashion along the ‘b’ axis (Fig. 3), reorganizing the alkene and azide groups of adjacent molecules in proximity (3.6 Å) and in a ready-to-react antiparallel arrangement.

Fig. 3: Analysis of stereoisomers formation.
figure 3

Head-to-tail arranged monomers arranged along the ‘b’ direction. The black dotted lines indicate the O–H…O hydrogen bonding, and the maroon lines indicate the proximity of the CRGs. The possibility of the formation of two stereoisomers is shown. Hydrogen atoms are hidden for clarity.

Full size image

The antiparallel arrangement suggests the regiospecific formation of 1, 4-triazolinyl linkages in the polymer (Fig. 3). As monomer molecules aligned for polymerization to form a particular chain are of similar conformational enantiomers (self-sorted arrangement), it is expected that all the repeating units of a particular chain will have the same absolute configuration (Fig. 3). Since the monomer crystal has an equal number of both conformational enantiomers, TEAC polymerization is expected to yield a perfect crystalline stereocomplex containing equal amounts of two enantiomeric chains of opposite chirality.

Crystals of monomer M remain stable at room temperature, exhibiting no reaction or decomposition for at least one month. The monomer crystals did not melt but were charred above 210 °C. We heated the crystals of monomer M to assess their reactivity at high temperatures (Supplementary section 2). The crystals kept at 80 °C for 4 days exhibited limited solubility in common solvents such as methanol, ethanol, DMF, isopropanol, acetone, DMSO, water, chloroform, DCM, toluene, and benzene. 1H NMR analysis of this partially soluble (heated) sample revealed only signals due to monomer M, indicating the insolubility of the product formed during heating. After 6 days of heating at 80 °C, the single crystals became completely insoluble in common solvents, which prevented characterization by solution-state NMR. However, a comparison of the solid-state 13C NMR spectra of the pristine crystals and the heated crystals confirmed the consumption of the alkene units in the latter. The signals at 134 and 118 ppm corresponding to the alkene group in the case of monomer M were diminished (disappeared) in the heated sample (Fig. 4a), indicating the consumption of alkene units in the TEAC reaction. Similarly, a comparison of the IR spectra of pristine crystals and heated crystals revealed a significant reduction in the signal representing the asymmetric stretching of azide at 2112 cm−1 in the latter (Fig. 4b), confirming the consumption of the azide group, presumably due to the TEAC reaction. Notably, at 100 °C, the reaction was complete within 36 h. However, heating the monomer crystals at 80 °C for 6 days preserved the high quality of single crystals.

Fig. 4: Comparison of pristine and heated samples.
figure 4

a Solid-state 13C NMR spectra of pristine and heated samples. b IR spectra of pristine and heated samples. c DSC analysis of pristine and heated samples. d PXRD analysis of pristine and heated samples.

Full size image

DSC analysis of pristine and heated single crystals elucidated the differences in their thermal properties. The pristine crystals revealed two exothermic peaks in the temperature ranges of 165–186 °C and 186-244 °C (Fig. 4c). While the former peak is attributable to the heat released during the exothermic ene-azide cycloaddition reaction, the latter corresponds to the decomposition of the product. Conversely, DSC analysis of heated single crystals revealed only one exothermic peak in the range of 186–244 °C, which is ascribed to decomposition (Fig. 4c). The absence of the peak in the 165–186 °C range suggested that the ene-azide cycloaddition reaction has already completed. The PXRD profile of the heated crystals was distinctly different from that of the pristine crystal, as anticipated in view of the topochemical reaction (Fig. 4d). The sharp peaks even in the heated crystals are suggestive of a crystal-to-crystal TEAC reaction. Examination of the heated crystal under a polarized microscope revealed a birefringence pattern, supporting the crystalline nature of the product (Supplementary Fig. 4). We checked the unit cell parameters of the heated crystal. We observed significant changes in the unit cell parameters of a heated crystal compared to those of a monomer M single crystal. While the unit cell parameter ‘a’ increased by 4.3%, ‘b’ and ‘c’ decreased by 13.3% and 1.8%, respectively (Supplementary Tables 1 and 3), resulting in a 4.3% reduction in the cell volume (Supplementary Tables 1 and 3). This large reduction in ‘b’ aligns with the prediction that polymerization occurs along the crystallographic ‘b’ direction.

We determined the crystal structure of the heated crystal using SCXRD analysis. The heated crystal retained the P−1 space group. As expected, the monomer underwent single-crystal-to-single-crystal (SCSC) TEAC polymerization along the ‘b’ axis, yielding a perfect 1:1 stereocomplex of two triazoline-linked enantiomeric polymers (Fig. 5a, Supplementary Fig. 5, Supplementary Tables 2 and 3). As expected, the reaction occurred in a regiospecific fashion, yielding 1, 4-triazolinyl-linked polymer products, PS and PR. For a particular polymer chain, the reaction is also stereospecific (Fig. 5a) such that all its triazoline linkages have the same absolute stereochemistry. Generally, because of their large molecular weight, the polymers formed via SCSC topochemical polymerization are insoluble in common solvents. In such cases, it is an accepted method to report the molecular weight per unit length of the polymer crystal by calculating it from the unit cell44,45. We adopted this procedure to calculate the molecular weight (MW), given the insolubility of the stereocomplex polymer crystals reported here. The MW per micrometer was estimated to be 278 kDa (Supplementary section 10).

Fig. 5: Crystal structure of stereocomplex.
figure 5

a Chemical structures and crystal structures of the polymer (PS and PR) are shown. b 3D view along ‘a’ axis showing stacks of layers laminated alternately along ‘c’ axis. c 3D view along ‘b’ axis. d O–H…O hydrogen bonds between polymer chains PS and PR.

Full size image

The polymer crystal also retains the hydrogen bonds present in the monomer crystal, and these zig-zag propagating hydrogen bonds connect the polymer chains along the ‘a’ and ‘c’ axes. Enantiopure polymer chains of one kind are parallelly arranged along the ‘a’ axis, forming a layer of enantiopure chains in the ‘ab’ plane (Fig. 5b). These enantiopure layers of opposite stereochemistry alternately stack along the ‘c’ axis (Figs. 5b, c). This alternately laminated arrangement makes a perfect 1:1 blend of two stereoregular polymers of opposite chirality, resulting in crystalline stereocomplexed polymers (Figs. 5b,c, Supplementary Fig. 6). It should be noted that conventional blending of two enantiopure polymers cannot produce such perfect stereocomplexes. In view of the peculiar hydrogen bonding pattern, each enantiopure chain connects with four enantiopure chains of opposite chirality via O–H…O hydrogen bonds (Fig. 5d).

In conclusion, by circumventing the inherent challenges of traditional methods, we demonstrated a simple protocol to generate a perfect stereocomplex of two polymers of opposite chirality. Our hypothesis that a perfect stereocomplex can be made via topochemical polymerization relies on the self-sorted assembly of monomers in the crystal that translates via a topochemical polymerization reaction into two enantiomeric polymers of opposite chirality. We used a meso-monomer and a reaction that generates a new chiral center in each linkage. Conformational enantiomers of this meso-monomer in its crystal showed self-sorted head-to-tail assembly with the reacting motifs in a ready-to-react arrangement. Topochemical polymerization of this meso-monomer yielded a perfect 1:1 blend of two enantiopure stereoregular polymers of opposite chirality in an SCSC fashion. The single-crystalline nature of the polymer provided atomic-level information about the stereocomplex. We believe that this proof-of-concept would encourage researchers to exploit different topochemical reactions for the synthesis of polymer stereocomplexes from a wide class of monomers and such research attempts would establish the generality of this methodology.

Methods

Details of chemicals, solvents, TLC, chromatography

All chemicals were purchased from commercial suppliers and were used without further purification. All solvents were used after distillation. Reactions were monitored by thin-layer chromatography using pre-coated silica gel plates (60 F254). Thin-layer chromatograms were visualized by heating the plates dipped in the ceric ammonium molybdate staining solution. Column chromatography was carried out using silica gel (230–400 mesh) as the stationary phase.

Details of NMR analysis

1H NMR, 13C NMR spectra were recorded on a 500 MHz and 125 MHz NMR spectrometer, respectively. Proton chemical shifts (δ) are relative to tetramethylsilane (TMS, δ = 0.0) as an internal standard and expressed in parts per million. Solid-state Cross Polarization Magic Angle Spinning Carbon-13 (CP-MAS 13 C) NMR spectra were recorded on Avance-III HD-700 MHz (Bruker) Two-Bay NMR spectrometer (16.4 T magnetic field) using 3.2 mm MAS probe at sample temperature approximately 10 °C. Cross-polarization (CP) was done to enhance the sensitivity of 1H decoupling. 1H pulse width of 3 μs was used (contact time is 2 ms). The MAS frequency was set to 10 kHz. 13C chemical shifts were indirectly calibrated to the methylene signal of adamantane (41.1 ppm on the TMS scale). Coupling constants (J) are given in Hertz. Protons and carbons were assigned using 2D NMR spectra such as COSY, HMQC, HMBC, and NOESY.

Details of IR, SCXRD, PXRD, and DSC analysis

IR spectra were recorded using IR Prestige-21 (Shimadzu) spectrometer. Melting points were determined by a Stanford Research Systems (EZ-Melt) melting point apparatus. Single Crystal X-Ray Diffraction (SCXRD) data measurements were carried on a Bruker-KAPPA APEX II CCD diffractometer with graphite-monochromatized (Mo αK = 0.71073 Å) radiation. The X-ray generator was operated at 50 kV and 30 mA. The X-ray data collection was analyzed by the SMART program (Bruker, version 5.631, 2004). All the data were corrected for Lorentzian polarization and absorption effects using SAINTPLUS and SADABS programs (Bruker, 2004). SHELXT and SHELXL-2014 were used for structure solution and full-matrix least-squares refinement on F2. Analysis of the crystal structure was done using the Mercury 3.9 software. All the hydrogen atoms were placed in geometrically idealized positions and refined in the riding model approximation with C-H = 0.95 Å, and with Uiso (H) set to 1.2Ueq (C). Powder X-ray diffraction (PXRD) analysis was done using an X’pert PRO (PANalytics) X-ray diffractometer using Cu as the anode material (Kα1 = 1.540598 Å). DSC analyzes were carried out using a DSC Q20 differential scanning calorimeter at a heating rate of 5 °C/min.