Probing the Microscopic Origin of Toughness in Multiple Polymer Networks

TL;DR

This work interrogates the microscopic origins of toughness in multiple polymer networks by coupling high-resolution photon correlation imaging with 3D coarse-grained molecular dynamics. The authors demonstrate that double-network elastomers exhibit early, delocalized microscopic rearrangements and nonlocal stress redistribution well ahead of crack propagation, in contrast to single networks where damage and rearrangements localize near the crack tip and drive failure. Simulations reveal that damage in the matrix network redistributes load over a broad region, delaying localization and enabling larger extensions before macroscopic failure, thereby explaining the superior toughness. These findings highlight the importance of spatiotemporal stress redistribution and enhanced microscopic dynamics as the key mechanisms behind toughness in DN elastomers, with implications for designing damage-tolerant soft materials.

Abstract

Multiple polymer networks, such as double-network elastomers comprising a sacrificial and a matrix network, exhibit exceptional mechanical resilience, commonly attributed to the formation of an extended damage zone before a crack can grow. However, the microscopic mechanisms underlying their toughness remain poorly understood. Here, we combine advanced light scattering methods and molecular dynamics simulations to explore the microscopic relaxation dynamics and stress redistribution at the polymer strand scale of single-network and double-network elastomers under uni-axial loading. Dynamic light scattering experiments show that microscopic rearrangements and bond-breaking events are localized near the crack tip in single networks, readily causing the crack to advance. In contrast, double networks exhibit delocalized microscopic rearrangements well ahead of and not directly correlated with crack propagation, enabling the dissipation of energy over broader regions and timescales. Numerical simulations of the damage zone show that bond breaking in the matrix network of double networks leads to widespread stress redistribution, mitigating catastrophic damage localization. This enhanced ability to redistribute stress in a non-local manner allows a much larger extension before localized macroscopic failure occurs, explaining the superior toughness of double networks. Our findings identify early, delocalized bond-breaking events combined with more efficient dissipation pathways through enhanced microscopic rearrangements as the key microscopic mechanisms responsible for the outstanding toughness and extensibility of multiple elastomer networks.

Figures (22)

• Figure 1: Architecture and mechanical properties of single and double polymer networks. a) Architecture of the network model used in our simulations and obtained through a protocol analogous to the experimental one: a sacrificial network (blue, left snapshot) is first synthesized and then swollen by a linear factor $\lambda_0$ by adding monomers (red, middle snapshot), which are finally polymerized and cross-linked to form the matrix network (right snapshot). $\lambda_0 = 1.7, 2$ in our experiments and simulations, respectively. b) Main graph: experimental engineering stress $\sigma_\mathrm{E}$vs stretch $\lambda=L/L_0$ for single (SN) and double (DN) pre-notched networks, with $L$ and $L_0$ the sample length along the stretching direction during and at the beginning of the test, respectively. Shaded regions indicate where microscopic rearrangements increase significantly, as discussed in the text. Inset: image of a SN sample, the notch is circled. The sample size along the vertical direction is 10 mm.
• Figure 2: Notch propagation in single and double networks. Colorized speckle images at selected $\lambda$ values, for a SN (a,b) and a DN (c,d), showing the crack shape and trajectory in a traction test. In b) and d), the white solid line is the crack tip trajectory, while yellow crosses show its position at intervals of one minute, starting from the position of the crack tip at the $\lambda$ value of the corresponding left panels. The dotted line in b) is a parabolic fit to the red pluses marking the crack shape. While the cracks are similar in single and double networks at the onset of propagation (a,c), they differ in shape and trajectory as they propagate through the sample.
• Figure 3: Crack speed and microscopic dynamics prior to rupture. Left axis: time dependence of the microscopic rearrangements as quantified by the activity $A$ (see text for definition) averaged over the half-plane ahead of the crack tip, for scattering in the forward (FS$_x$, FSy) and backward (BS) directions, see Sec. Materials and Methods and Supplementary Information for details. $\Delta t=0$ corresponds to macroscopic rupture and the top and bottom panels refer to representative single and double networks, respectively. The right axis and red line show $v_\mathrm{c}$, the propagation speed of the crack tip, with zooms on the behavior right before rupture in the insets. The dotted line in the insets is the threshold below which $v_\mathrm{c}$ is considered to be negligible for the analysis of Fig. \ref{['fig:exp-dyn-act-maps']}.
• Figure 4: Correlation between crack propagation and microscopic dynamics depends on network architecture. Color-coded number of occurrences of ($v_\mathrm{c}$, $\left<A\right>$) pairs obtained by binning measurements as those shown in Fig. \ref{['fig:exp-dynact_vs_time']}, for SNs (top) and DNs (bottom), in the FS$_x$ geometry. Data obtained from tests on three SN and four DN samples in the 3000 s preceding macroscopic failure. Note the excess of large values of activity at vanishing $v_\mathrm{c}$ in DNs as compared to SNs.
• Figure 5: Spatial dependence of the activity in the half plane ahead the crack tip. Spatial maps of $A$, for SNs (a) and DNs (b), in a moving reference frame where the crack tip corresponds to $(x,y) = (0,0)$ (red cross). Data are obtained by averaging FS$_x$ measurements for all samples, in time intervals with negligible notch propagation speed, $v_c \leq \qty{2.5}{\micro\meter\second^{-1}}$, and within 3000 s before macroscopic rupture. (c) Activity as a function of distance $r$ to the crack tip, obtained by azimuthally averaging the data of a) and b). The shaded areas show the standard deviation resulting from sample to sample variations.