Molecular Processing of Polymers with Cyclodextrins

Abstract

We summarize our recent studies employing the cyclic starch derivatives called cyclodextrins (CDs) to both nanostructure and functionalize polymers. Two important structural characteristics of CDs are taken advantage of to achieve these goals. First the ability of CDs to form noncovalent inclusion complexes (ICs) with a variety of guest molecules, including many polymers, by threading and inclusion into their relatively hydrophobic interior cavities, which are roughly cylindrical with diameters of ∼ 0.5 − 1.0 nm. α-, β-, and γ-CD contain six, seven, and eight α-1,4-linked glucose units, respectively. Warm water washing of polymer-CD–ICs containing polymer guests insoluble in water or treatment with amylase enzymes serves to remove the host CDs and results in the coalescence of the guest polymers into solid samples. When guest polymers are coalesced from the CD–ICs by removing their host CDs, they are observed to solidify with structures, morphologies, and even conformations that are distinct from bulk samples made from their solutions and melts. Molecularly mixed, intimate blends of two or more polymers that are normally immiscible can be obtained from their common CD–ICs, and the phase segregation of incompatible blocks can be controlled (suppressed or increased) in CD–IC coalesced block copolymers. In addition, additives may be more effectively delivered to polymers in the form of their crystalline CD–ICs or soluble CD–rotaxanes. Secondly, the many hydroxyl groups attached to the exterior rims of CDs, in addition to conferring water solubility, provide an opportunity to covalently bond them to polymers either during their syntheses or via postpolymerization reactions. Polymers containing CDs in their backbones or attached to their side chains are observed to more readily accept and retain additives, such as dyes and fragrances. Processing with CDs can serve to both nanostructure and functionalize polymers, leading to greater understanding of their behaviors and to new properties and applications.

References

As seen in Fig. 10 in both non-aromatic carbon spectral regions the least shielded, downfield resonance belongs to as-received PET, the most shielded, upfield resonance comes from coalesced or included PETs, and the precipitated PET [ 59 ] seems to contain ∼30 and ∼70% material resonating at the frequencies of these downfield and upfield, as-received, and coalesced or included PET peaks, respectively. We expect [ 122 ] carbonyl carbons terminating ethylene glycol fragments whose -$CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds have g± conformations to resonate upfield from those with t conformations. This is consistent with conclusions drawn previously from modeling the conformations of included PET chains [ 78, 84 ] and the FTIR analysis [ 21 ] of as-received and coalesced PET conformations.

We would normally expect the methylene carbon resonances of these PET samples to exhibit the same order of resonance frequencies, because they are γ to the carbonyl carbons and are either t or g± to them conformationally across the same -$CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds. This is, in fact, what we observe in Fig. 10. However, a recent solid-state ¹³ C-NMR study of PETs by Kaji and Schmidt-Rohr [ 123 ] has convincingly established that the resonance frequencies of methylene carbons in PETs are insensitive to the conformations of the -$CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds, and instead seem to depend only upon the conformations of the -$CH_{2}\hbox{--}CH_{2}$- bond connecting them. On highly crystalline and predominantly amorphous PET samples with ¹³ C-enriched methylene carbons they were able to separately observe methylene carbon resonances belonging to t and to $g\pm - CH_{2}\hbox{--}CH_{2}$- bonds. They found that in both PET samples the methylene carbons belonging to $t-CH_{2}\hbox{--}CH_{2}$- bonds resonated ∼ 2 ppm upfield from those methylene carbons with $g\pm\hbox{-}CH_{2}\hbox{--}CH_{2}$- bonds, even though the -$CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds are predominantly t in the highly crystalline PET and significantly g± in the nearly completely amorphous PET sample. As a consequence, we can conclude that our coalesced and included PETs have predominantly t -$CH_{2}\hbox{--}CH_{2}$- bonds, as-received PET predominantly g±, and our precipitated PET [ 59 ] sample seems to have about 30% g± and 70% -$t \hbox{-} CH_{2}\hbox{--}CH_{2}$ bonds. Again, this is consistent with our molecular modeling [ 78, 84 ] and FTIR [ 21 ] results, so as-received PET has predominantly $g\pm\hbox{-}CH_{2}\hbox{--}CH_{2}$- bonds and substantial amounts of t -$CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds, while coalesced, included, and precipitated PETs have preponderantly t-$CH_{2}\hbox{--}CH_{2}$- and $g\pm\hbox{-}CH_{2}\hbox{--}O$- and -$O\hbox{--}CH_{2}$- bonds.

The ¹³ C-observed ¹ H spin-lattice relaxation times observed in the rotating frame $[T_{1\rho}^{1}(H)]$, which reflect motions in the kHz frequency regime, are presented as a function of temperature for our PET samples in Fig. 11 . Generally the coalesced sample has the longest and the as-received sample the shortest $T_{1\rho}^{1}(H)$, indicating an increasing kHz mobility for PET chains in the coalesced, precipitated, and as-received samples, respectively. Also note that the $T_{1\rho}^{1}(H)s$ of as-received PET show a marked sensitivity to temperature at T ∼ T _g , which is largely unobserved in the coalesced and precipitated samples. Initially, the molecular motion increases with temperature resulting in shorter $T_{1\rho}^{1}(H)s$ for all the PET samples. However, the $T_{1\rho}^{1}(H)$ of the as-received PET reaches a minimum near T _g , and further heating results in molecular motions that are too rapid for efficient nuclear spin energy transfer, so T _{1 ρ}(¹ H) increases for T > T _g . This is consistent with the presence and absence of a glass transition observed in the DSC scans of as-received and coalesced (see Fig. 7 ) or precipitated [ 59 ] PETs, respectively. Thus, both macroscopic (DSC) and microscopic (NMR) observations point to the absence of a glass transition in the noncrystalline regions of coalesced or precipitated PETs.

At room temperature, the ¹³ C spin-lattice relaxation times, $T_{1}^{13}(C)$ in seconds, for the as-received (asr), precipitated (ppt), and coalesced (coa) PETS are C = O → 31.8 s (asr) and 36.2 s (coa, ppt); nonprotonated aromatic → 28.6 s (asr) and 36.2 (coa, ppt); protonated aromatic → 14.4 s (asr) and 21.0 s (coa, ppt); and CH ₂ → 7.1 s (asr) and 9.6 s (coa, ppt).

Coalesced and precipitated PETs have longer $T_1^{13}(C)s$ than as received PET. Thus, motions in the MHz frequency regime are also more restricted in the coalesced and precipitated PETs, compared with as-received PET, possibly because of both their higher crystallinities and the tighter packing of kink conformers in their noncrystalline regions. Temperature dependencies similar to those observed in the rotating frame for ¹ Hs, $T_{1\rho}(^{1}H)\ s$, are also observed for the $T_{1}^{13}(C)s$. This behavior implies that the noncrystalline regions of coalesced and precipitated PETs are distinct from the amorphous regions in as-received PET, because only in as-received PET are the kHz, MHz motions important to $T_{1\rho}(^{1}H), T_{1}(^{13}(C))$ relaxations sensitive to whether or not the sample is below or above its T _g . This is again consistent with the failure to observe a glass transition by DSC for coalesced [ 21 ] and precipitated [ 33, 59 ] PETs.