Entangled Polymer Dynamics in Attractive Nanocomposite Melts

Abstract

We investigate single chain dynamics of an entangled linear poly­(ethylene oxide) melt in the presence of well-dispersed attractive nanoparticles using high-resolution neutron spectroscopy at particle volume fractions as high as 0.53. The short-time dynamics shows a decrease of the Rouse rates with particle loading, yet the change remains within a factor of 2, with no evidence of segment immobilization as often hypothesized. The apparent reptation tube diameter shrinks by ≈10% from the bulk at a 0.28 particle volume fraction when the face-to-face interparticle distance approaches the single chain size. The tube diameter is remarkably concentration-independent at higher loadings where all chains are essentially bound to particle surfaces. These direct experimental observations on the microscopic chain dynamics in attractive nanocomposites are distinct from their nonattractive counterparts and account for some of the unusual dynamic behaviors of the nanoparticles as well as rheology in the composites.

Insights from modeling into structure, entanglements, and dynamics in attractive polymer nanocomposites

“…The simulation predictions are in agreement with different experimental efforts in attractive nanocomposites, such as for PMMA/nanosilica mixtures by Jouault et al 17 up to f E 30% loading. The same unperturbed behavior is observed not only in attractive PEO/ nanosilica mixtures, 72 up to f E 53%, but also in an athermal PS/nanosilica mixtures, up to f E 32% loading. 60 We use the single chain static structure factor within the qR g { 1 regime as an alternative way to calculate the R g , as it behaves as follows, S sc (x) = N[1 À x 2 /3 + O(x 4 )] with the dimensionless x = qR g .…”

“…Such behavior was also observed in experimental measurements in PEO/nanosilica composites. 72 However, this behavior was different to that observed in nanocomposites containing nonattractive interaction between polymers and NPs, 75 where d app remained constant up to f = 20%. Beyond that volume fraction, d geo coincided with d app (red and green lines coincide for f 4 30%) as depicted in Fig.…”

“…Moreover, the tube diameter, characterizing dynamical aspects of entangled polymer matrices, was found to increase with NP loading from neutron spin echo measurements on nanosilica/ PEP composites (which can be considered as nanocomposites with nonattractive interactions). 75 In attractive PEO-nanosilica nanocomposites, however, a different behavior on the tube diameter was observed recently by Senses et al 72 Furthermore, for a strongly attractive nanocomposite material such as poly(2vinylpyridine) (P2VP) polymers and nanosilicas, weakly adsorbed chains eventually desorbed from the NP surface, while strongly adsorbed chains remained bound for experimental time scales available to elastic recoil recovery (ERD) and Rutherford Backscattering spectrometry (RBS). 76 The boundlayer thickness was found up to R g /2 distance from the NP surface and was affected by the strength of attraction.…”

“…59 Unperturbed dimensions were also found for an attractive poly(ethylene oxide) (PEO)-nanosilica composite with an even smaller ratio (R g /R NP = 0.28), 71 at large NP volume fractions of up to f = 53%. 72 In contrast, PS chains were found to expand by up to 20% in the presence of PS-crosslinked NPs (for ratios R g /R NP = 1.6-5.7) 55 or carbon nanotubes. 58 A dramatic expansion of polymers by almost 60% at high loading was observed for poly(dimethyl siloxane) (PDMS) in the presence of soft polysilicate R NP = 1 nm NPs (ratios R g /R NP = 6-8).…”

“…Secondly, we calculate and evaluate polymer and NP dynamics for different NP loadings, while focusing on the bound layer which affects dynamic properties and reinforcement [105][106][107] (Section 3.3). Thirdly, we investigate the entanglement network for different NP loadings in an attempt to address open questions formulated earlier by Senses et al 72 for attractive nanocomposites (Section 3.4). While Senses claimed that the bound layer is responsible for a constant d app at high f, we find a behavior that originated, based on a mean field assumption, wholly from the geometrical confinement.…”