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 M
w 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 G
N
o of the corresponding
LDODT (melt or matched-concentration solutions). In entangled linear
polymers, the observation that G
N
o 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 G
N
o 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.
Finall...