Understanding creep suppression mechanisms in polymer nanocomposites through machine learning

Area of Science:

• Materials Science
• Polymer Physics
• Computational Materials Science

Background:

• Understanding interfacial dynamics in complex materials like polymer nanocomposites (PNCs) is crucial but challenging.
• While local structure influences dynamics in homogeneous systems, interfacial effects in heterogeneous systems are less understood.
• Interfaces are known to alter molecular packing, but their impact on dynamics extends beyond immediate proximity.

Purpose of the Study:

• To quantitatively distinguish the roles of polymer packing and other factors in modifying dynamics near nanoparticle surfaces in PNCs.
• To investigate the relationship between structure and dynamics at interfaces using advanced computational and experimental methods.
• To develop a predictive model for creep behavior in PNCs based on their static structure.

Main Methods:

• Combined molecular dynamics simulations and experimental techniques (e.g., creep compliance measurements).
• Utilized a machine-learned structure indicator ('softness') to decompose polymer dynamics.
• Analyzed dynamics as a function of distance from nanoparticle surfaces.

Main Results:

• Achieved good qualitative agreement between simulations and experiments for glassy structure and creep compliance.
• Decomposed polymer dynamics into packing-dependent and packing-independent contributions to the free energy barrier.
• Found that both packing-dependent and independent barriers are elevated near nanoparticles and decrease with applied stress.
• Demonstrated that slow interfacial dynamics result from both altered packing and modified structure-dynamics relationships.

Conclusions:

• The slow dynamics observed at polymer-nanoparticle interfaces are not solely attributable to differences in polymer packing.
• Changes in the fundamental relationships between material structure and dynamics play a significant role.
• The developed decomposition method accurately predicts strain-time creep curves for PNCs from static configurations.
• Provides new insights into creep suppression mechanisms at polymer-nanoparticle interfaces.