Accelerated Quantum Simulations of Polymer Aging and Degradation
We have conducted extensive quantum simulations to predict the susceptibility of polydimethylsiloxane (PDMS) under tensile loads (i.e., stretching) to hydrolytic reactions. Initial results under uniform strain revealed that the PDMS electronic state is highly sensitive to deformations of the Si-O backbone (Figure 1). Predictions for the energy gap between the highest-occupied and lowestunoccupied molecular orbitals (i.e., the HOMO-LUMO gap, E Gap ) and visualization of these orbitals reveals the distinct possibility that strained PDMS chains have an electronically driven anisotropy in their susceptibility to hydrolysis. First, we observe that EGap begins to drop precipitously as the PDMS chain is stretched beyond 10% (labeled 1.1 in the figure). The increased accessibility of these unoccupied states generally will promote chemical degradation, as excited electronic states due to bond breaking will form more easily. Second, we observe that the geometric character of the LUMO state changes under stretching of 20% (strains of 1.20) and larger, whereby the orbital density localizes to the interior of the O-Si-O bond angles along the backbone. This LUMO localization coupled with a substantial reduction in EGap indicates that the interior O-Si-O bond angles may be potential sites for chemistry due to environmental water or by roaming acids or bases. This is in contrast to a chemical attack on the backside of the O-Si-O angle, where there is practically no density of low-energy unoccupied states. The enhanced electronic susceptibility likely couples with steric shielding by the two methyl groups that also protect the backside of the Si-O chain. These detailed results can be used as inputs for coarse-grained simulation approaches, by helping us to define the susceptibility of specific polymer chain location to scission and/or backbiting reactions, based on the local geometry of the Si-O backbone. Definition of this type of order parameter with application to larger scale models is the subject of our future work.