Simulation Study on the Mechanical Properties of Lamella-Forming ABABA Linear Multiblock Copolymers

Xing Chu,Zhanwen Xu,Yu-cai Cao,Weihua Li
DOI: https://doi.org/10.1021/acs.macromol.4c01204
IF: 5.5
2024-01-01
Macromolecules
Abstract:The A ' BA '' BA ' multiblock copolymers usually exhibit superior mechanical performance due to the increased bridging conformations that could bridge multiple domains compared to their ABA triblock counterparts. However, the underlying molecular mechanisms are not completely understood due to the difficulty in experimental characterization. Moreover, for the A ' BA '' BA ' pentablock copolymers, the molecular conformations and the segregation between the blocks are significantly influenced by the length ratio tau of A ''-block relative to total A-blocks, and therefore there should be an optimal value of tau for the best mechanical performance. To elucidate the effect of the middle A ''-blocks on the mechanical properties and optimize the molecular architecture for A ' BA '' BA ' pentablock copolymers, we investigated the mechanical properties of lamella-forming A ' BA '' BA ' multiblock samples using self-consistent field theory (SCFT) coupled with coarse-grained molecular dynamics (CGMD) simulations. It is revealed that the bridging A ''-blocks significantly enhance the mechanical performance of A ' BA '' BA ' pentablock samples by bearing and transmitting stress during elongation. Furthermore, the mechanical properties of A ' BA '' BA ' samples are greatly affected by the length ratio tau of the A ''-block to the total A-blocks. Specifically, the maximal stress sigma(m) is mainly influenced by the bridging ratio together with the contact degree of A-blocks extending from adjacent B-domains, becoming the largest at tau approximate to 0.4 before the bridging ratio reaches the highest value. The elongation ratio epsilon(m) reaches its optimum value at a larger tau around 0.7. Combining the maximal stress and elongation ratio, the pentablock samples exhibit optimal toughness when tau is close to 0.6, which is estimated to be 50% larger than that of the A ' BA '' BA ' sample with simply equal A-blocks.
What problem does this paper attempt to address?