Linear Viscoelastic Properties of Putative Cyclic Polymers Synthesized by Reversible Radical Recombination Polymerization (R3P)

Abstract

Linear viscoelastic properties in both melt and solution states are reported for a series of poly(3,6-dioxa-1,8-octanedithiol) (polyDODT) made by reversible radical recombination polymerization (R3P) under conditions designed to produce linear (LDODT), cyclic (RDODT), and linear–cyclic mixtures (LRDODT). PolyDODT is amorphous (T_g < −50 °C) and highly flexible (entanglement molecular weight M_(e,lin) ≈ 1850 g/mol for LDODT). PolyDODT's low T_g and low M_(e,lin) enable characterization over a wide dynamic range and a wide range of dimensionless weight-average molecular weight Z_w = M_w/M_(e,lin). Measurements at temperatures from −57 to 100 °C provide up to 18 decades of reduced frequency, which is necessary to characterize RDODT melts with Z_w from 23 to 300. The two highest-molecular-weight polymers in the present RDODT series have such high M_w (406k and 556k g/mol) that mass spectrometry, NMR spectroscopy, and even chemical assays for chain ends are unable to rule out up to 2 mol % of linear contaminant. By studying the samples in solution (using dilution to reduce Z_w), we could compare their dynamics with those of previously established high-purity polystyrene (PS) rings (limited to Z_w ≤ 13.6). RDODT solutions with Z_w < 15 (concentrations <5 wt % for RDODT-406k and 556k) have dynamic moduli G* that accord with LCCC-purified PS rings in terms of the frequency dependence (including the absence of a plateau), the progression of shapes of G* as a function of Z_w, and the linear scaling of their zero-shear viscosity η₀ with M_w. The shape of G* as a function of Z_w for solutions of RDODT-406k and -556k also accords with lower Mw RDODT melts (which have ≤1.3 mol % of linear contaminant). Thus, the measurement of the linear viscoelastic properties of appropriate concentrations of high M_w (>200k g/mol) putative cyclic polymers, in which linear chains evade spectroscopic detection, may provide an alternative means (though not fully proven) of validation of sample purity. When Z_w > 15 (including all seven RDODT melts and eight of their solutions), G* has a rubbery plateau. This suggests that the onset of entanglement-like behavior in rings requires 4–5-fold greater Z_w than is required for linear chains. Further, the plateau moduli of RDODT samples are indistinguishable from G⁰_N of the corresponding LDODT (melt or matched-concentration solutions). In entangled linear polymers, the observation that G⁰_N is independent of Z_w follows from limitations on lateral fluctuations due to neighboring chains becoming independent of position along a given chain. The present results for RDODT suggest that this holds for sufficiently long endless chains, too. While the RDODTs have the same G⁰_N as entangled LDODTs, when Z_w > 60, the terminal relaxation, if reached at all, of RDODT extends to orders of magnitude lower frequency than an entangled linear polymer of the same Z_w. Consequently, the viscosity of RDODT with Z_w > 60 increases with Z_w much more strongly than the 3.4 power observed for entangled linear polymers. Finally, these novel polymers, with a disulfide-linked backbone and broad relaxation time distribution, may prove important in relation to biodegradable elastomers and materials with exceptional low-frequency dissipation, extending at least 12 decades below the onset of the rubbery plateau.

Additional details