Attributed to its strain-induced crystallization (SIC), natural rubber (NR) exhibits more excellent mechanical properties compared to other elastomeric materials and has been attracting numerous scientific and technological attention. However, a systematical understanding of the structure–mechanics relation of NR is still lacking. Herein, for the first time, we employ molecular dynamics simulation to examine the effects of the key structural factors on the SIC and mechanical properties at the molecular level. We examine the effects of phospholipid and protein mass fraction (ω), the strength of hydrogen-bond interaction (εH), and the strength of non-hydrogen-bond interaction (εNH) on structural morphology, dynamic behavior, and mechanical properties. NR tends to form local clusters due to the hydrogen-bond interaction formed between phospholipids or proteins and chain ends, which is absent in the case of cis-1,4-polyisoprene (PIP). The polymer chain mobility of NR is retarded due to the formed clusters or even physical network at great εH and high ω. Interestingly, we find that the stress–strain behavior of NR is greatly manipulated by εH and ω, as evidenced by the increase of the chain orientation and the SIC, compared with the cases of PIP. This underlying mechanism results from the alignment of the molecular chains induced by the formed clusters along the deformed direction, and the clusters during the deformation become more stable, particularly at great εH. Lastly, we adopt a machine learning algorithm named extreme gradient boosting via data augmentation, finding that εH has the most significant influencing weight factor on the stress–strain behavior of NR. In general, this work demonstrates a detailed molecular-level structure–mechanics relation of NR and provides some rational guidelines for experimentally designing and synthesizing biomimetic NR.

Through molecular dynamics (MD) simulation, the structure and mechanical properties of matrix-free polymer nanocomposites (PNCs) constructed via polymer-grafted graphene nanosheets are studied. The dispersion of graphene sheets is characterized by the radial distribution function (RDF) between graphene sheets. We observe that a longer polymer chain length L g leads to a relatively better dispersion state attributed to the formation of a better brickmud structure, effectively screening the van der Waals interactions between sheets. By tuning the interaction strength ε end−end between end functional groups of grafted chains, we construct physical networks with various robustness characterized by the formation of the fractal clusters at high ε end−end values. The effects of ε end−end and L g on the mechanical properties are examined, and the enhancement of the stress−strain behavior is observed with the increase of ε end−end and L g . Structural evolution during deformation is quantified by calculating the orientation of the graphene sheets and their distribution, the stress decomposition, and the size of the clusters formed between end groups and their distribution. Then, we briefly study the effects of time and temperature on the self-healing behavior of these unique PNCs in the rubbery state. Lastly, the self-healing kinetics is quantitatively analyzed. In general, this work can provide some rational guidelines to design and fabricate matrix-free PNCs with both excellent mechanical and self-healing properties.

No abstract