Ellipsometry and fluorescence are used via measurements of film thickness and pyrenyl dye emission spectral shape, respectively, to characterize residual stress relaxation in polystyrene (PS) films. In particular, fluorescence of pyrene-labeled PS (MPy-PS) films, with ~1 mol% pyrene label, provide sensitivity to film stress relaxation and stiffness by the dependence of the ratio of the first to third vibronic peak intensities (I 1 /I 3) of the pyrenyl dye to nanosecondtime-scale molecular caging. Both techniques show that residual stress relaxation occurs over a period of hours despite the PS films being 15-40 °C above the film glass transition temperature (T g). Both techniques also show that film T g is unaffected by stress relaxation, even when stress relaxation is accompanied by measurable changes in thickness. Fluorescence shows that stress relaxation time follows an Arrhenius temperature dependence with an activation energy of ~110 kJ/mol, which is consistent with stress relaxation occurring by β-relaxation. Using a bilayer/fluorescence technique with bulk bilayer films, it was observed by I 1 /I 3 measurements that a 30-nm-thick MPy-PS layer located at a glass substrate interface is stiffer than a 30-nmthick MPy-PS layer located at a free surface. Over a 20 to 400 nm thickness range, fluorescence of MPy-PS films show a significant effect of substrate on molecular caging and hence stiffness, with stiffness increasing in the following order: free-standing films (no substrate) < films supported on a PDMS (soft) substrate < films supported on a glass (hard) substrate.

Polymer-tethered nanoparticles provide a strategy to improve particle dispersion in polymer nanocomposites and as materials themselves can exhibit self-healing behavior and enhanced mechanical properties. The few studies that previously characterized the glass transition temperature (T g) behavior of neat polymer-grafted nanoparticles in the absence of a polymer matrix largely focused on average T g response. We synthesized polystyrene-grafted silica nanoparticles (Si-PS) via ARGET ATRP, achieving the densely grafted state. Using differential scanning calorimetry, we investigated the brush molecular weight (MW) dependence of T g, T g breadth, heat capacity jump (ΔC p ), and fragility from 12 to 98 kg/mol. Compared with free PS chains of the same MW, brush T g increases by 1–2 °C, brush T g breadth remains unchanged within error down to 36 kg/mol and increases by 3–4 °C at brush MWs of 12 and 13 kg/mol, and brush ΔC p and fragility remain unchanged within error down to 52 kg/mol and then decrease with decreasing MW. Evidence of a significant T g gradient from near the nanoparticle graft interface to near the free chain end was obtained for the first time via fluorescence of a pyrenyl dye labeled at specific regions along the brush chain length. In relatively high MW brushes, T g = ∼116 °C near the graft interface and T g = ∼102 °C near the chain end. Comparisons are made to results recently reported for similar PS brushes densely grafted to a flat substrate, which indicate that a larger T g gradient is evident in a grafting geometry involving a flat interface as compared with a spherical nanoparticle interface. Other comparisons are also made with glass transition and fragility behaviors reported in the flat substrate geometry. Results of this study and others will help to better understand nanocomposites and tailor them for optimal properties.

The issue of how significantly and over what length scales stiffness or modulus can be modified by the presence of a substrate or nanoparticle interface is important in the design and performance of polymer nanocomposites and nanostructured polymers. Here, we provide the first comparison of stiffness gradient length scales in polymeric materials characterized by two techniques: fluorescence spectroscopy (which is sensitive to molecular caging and hence to modulus) and AFM (which, coupled with finite element analysis, provides a direct determination of modulus). After cooling samples from 140 °C at 1 °C/min, characterization is done at room temperature on model nanocomposites in which a polystyrene (PS) film is supported on both sides by glass substrates. In confined model nanocomposites, the local stiffness enhancement relative to bulk is the result of perturbations from both substrate interfaces. In a 266 nm thick PS model nanocomposite, perturbations to modulus extend ∼200 nm from each interface in a nonlinear compound effect. Both methods reveal a small (5+% by AFM) stiffness enhancement at the midpoint of a 266 nm thick model nanocomposite; the midpoint modulus increases with confinement and is 50% higher than bulk in a 60 nm thick model nanocomposite. In bulk model nanocomposites, stiffness gradients result from perturbations propagating from a single substrate interface that are damped by the bulk PS layer, and both methods indicate that stiffness gradients extend ∼80 nm from an interface. The two techniques show qualitative and quantitative agreement regarding stiffness gradient length scales and are correlated via a simple monotonic relationship between the fluorescence measurable and normalized modulus.

Stiffness-confinement effects are characterized via a non-contact, self-referencing fluorescence approach in polystyrene (PS) films labeled with trace levels of 1-pyrenylmethyl methacrylate. The pyrene fluorescence measurable I 1 /I 3 is sensitive to molecular caging, which increases with stiffness. At 140 °C, molecular caging and hence stiffness in single-layer PS films supported on silica is independent of thickness down to 240 nm and increases with decreasing thickness at 165 nm and below. In contrast, near T g at 100 °C and in the glassy state at 60 °C, molecular caging and hence stiffness in single-layer films is independent of thickness down to 63 nm and increases with decreasing thickness at 36 nm and below. In bulk bilayer films, perturbations originating at the substrate interface (free-surface interface) cause major increases (decreases) in caging and hence stiffness in 20-nm-thick substrate-adjacent (free-surfaceadjacent) layers. In contrast, in 40-nm-thick bilayer films, the 20-nm-thick substrate-adjacent and free-surface-adjacent layers exhibit little difference in caging and stiffness. Thus, the gradient in stiffness from a film interface depends significantly on confinement, which we hypothesize begins to occur when thickness becomes comparable to the combined length scales over which free-surface and substrate perturbations propagate inside the film. Bulk bilayer films were used to investigate the length scales associated with interfacial perturbations. At 100 °C and 60 °C, stiffness-gradient length scales extend ~45-85 nm from the substrate and ~35-85 nm from the free surface. At 140 °C, the stiffness-gradient length scales extend ~85-200 nm from the substrate and ≲ 20 nm from the free surface.

Many studies have established a major effect of nanoscale confinement on the glass transition temperature (T) of polystyrene (PS), most commonly in thin films with one or two free surfaces. Here, we characterize smaller yet significant intrinsic size effects (in the absence of free surfaces or significant attractive polymer-substrate interactions) on the T and fragility of PS. Melt infiltration of various molecular weights (MWs) of PS into anodic aluminum oxide (AAO) templates is used to create nanorods supported on AAO with rod diameter (d) ranging from 24 to 210 nm. The T (both as T and fictive temperature) and fragility values are characterized by differential scanning calorimetry. No intrinsic size effect is observed for 30 kg/mol PS in template-supported nanorods with d = 24 nm. However, effects on T are present for PS nanorods with M and M ≥ ∼175 kg/mol, with effects increasing in magnitude with increasing MW. For example, in 24-nm-diameter template-supported nanorods, T - T = -2.0 to -2.5 °C for PS with M = 175 kg/mol and M = 182 kg/mol, and T - T = ∼-8 °C for PS with M = 929 kg/mol and M = 1420 kg/mol. In general, reductions in T occur when d ≤ ∼2R, where R is the bulk polymer radius of gyration. Thus, intrinsic size effects are significant when the rod diameter is smaller than the diameter (2R) associated with the spherical volume pervaded by coils in bulk. We hypothesize that the T reduction occurs when chain segment packing frustration is sufficiently perturbed by confinement in the nanorods. This explanation is supported by observed reductions in fragility with the increasing extent of confinement. We also explain why these small intrinsic size effects do not contradict reports that the T-confinement effect in supported PS films with one free surface exhibits little or no MW dependence.