Curing-induced filler aggregation in epoxy-amine systems

Abstract

Hypothesis: The macroscopic properties of polymer composites are governed by the dispersion and aggregation states of filler particles within a crosslinking matrix. Although curing transforms a liquid precursor into a solid network, its influence on filler aggregation remains insufficiently understood. We hypothesize that curing induces effective attractive interactions between filler particles, leading to aggregation even in non-Brownian systems. Experiment: Fluorescent polystyrene beads were homogeneously dispersed in a bisphenol F epoxy resin. Curing was initiated by adding trimethylhexamethylenediamine, and the evolution of the three-dimensional particle configurations was quantitatively examined using confocal laser fluorescence microscopy before and after completion of curing. Findings: Aggregation was enhanced during curing despite the absence of conventional attractive forces. The aggregation increment cannot be described solely by filler volume fraction but is governed by the mean interparticle gap $H$. Data collapse onto a linear scaling with the reduced gap parameter, identifying a geometric control parameter for curing-induced aggregation. This scaling demonstrates that curing dynamically generates an effective interaction whose spatial range scales with particle size, consistent with previously predicted rigidity-percolation-induced attractions. These findings establish a geometric criterion for predicting final dispersion states in curing polymer composites.

Figures (6)

• Figure 1: (a) Temporal evolution of the curing ratio, $\alpha$, determined from the peak intensity ratios in the FT-IR spectra. Measurements were conducted at $t = 0, 1, 2, 3, 6, 24, 96,$ and $144$ h. (b) Temporal evolution of viscosity $\eta$. The viscosity increases to approximately 40 Pa$\cdot$s (nearly two orders of magnitude higher than the initial value) within about 2 h, after which it rises sharply, rendering further measurements difficult.
• Figure 2: Three-dimensional reconstructions from confocal fluorescence microscopy images at the initial curing stage (a) and after curing (b) for $a = 8\,\rm{\mu m}$ and $\varphi = 0.16$. Three-dimensional filler configurations at the initial curing stage (c) and after curing (d) for $a = 5\,\rm{\mu m}$ and $\varphi = 0.065$. The particle color represents the cluster size. ParaView was employed for three-dimensional reconstruction.
• Figure 3: Cluster size distribution $F(s)$ at the initial curing stage and after curing under each experimental condition: (a) $a = 8\,\rm{\mu m}$, $\varphi = 0.12$; (b) $a = 8\,\rm{\mu m}$, $\varphi = 0.16$; (c) $a = 5\,\rm{\mu m}$, $\varphi = 0.065$; (d) $a = 5\,\rm{\mu m}$, $\varphi = 0.13$. The black line with filled circles represents the distribution at the onset of curing, whereas the red line with filled triangles denotes the distribution after curing. Inset: Enlarged views of the large $s$ region in the $F(s)$ distributions for each experimental condition.
• Figure 4: Aggregation fraction, $\Psi_{\rm agg}$, at the early and final stages of curing. Panels (a) and (b) correspond to $\varphi = 0.065$ with particle diameters $a = 5\,\rm{\mu m}$ and $8\,\rm{\mu m}$, respectively. Panels (c) and (d) correspond to $a = 5\,\rm{\mu m}$ at $\varphi = 0.12$ and $a = 8\,\rm{\mu m}$ at $\varphi = 0.13$, respectively. Error bars represent standard deviations obtained from independent measurements.
• Figure 5: (a) Dependence of $\Delta \Psi_{\rm agg}$ on the mean interparticle gap $H$. $H_c$ denotes the critical gap at which $\Delta \Psi_{\rm agg}$ becomes approximately zero. (b) $\Delta \Psi_{\rm agg}$ plotted as a function of the reduced gap $\delta H/a$. Data obtained for different particle sizes collapse onto a single linear relationship. This scaling behavior is consistent with a simple geometric rescaling argument, in which the probability distribution of interparticle gaps follows a common form when normalized by particle size.