Search

Mechanical failure such as fatigue is a process with complexities arising from the microstructures of materials and the underlying physics. Single crystals without surfaces and imperfections thus offer an ideal platform to probe the process of mechanical degradation by excluding the microstructural complexity. We explore the failure mechanisms of single-wall carbon nanotubes under constant, monotonic and cyclic tensile loads. Simulations with the full-atom details show that, under high strain, failure is dominated by bond breakage, while bond recovery and Stone–Wales transitions are promoted as strain decreases or temperature increases. By assuming the process as a single-step or multiple-step reaction, kinetic models based on the Arrhenius law are developed for the process of mechanical degradation. The models well predict the life expectation and distribution, which conform to the exponential law for single-step processes and are complicated by the memory effect resulted from the multi-step mechanisms. We find that the nanotube could conform to the damage-free, memory-less limit dominated by bond breakage at room temperature and high strain, featuring an exponential distribution of life to failure, while at elevated temperatures where the Stone–Wales transitions are activated, the memory effect becomes significant and the life distribution can be better fitted by the log-normal or Weibull model. The work offers an atomistic view of the failure process and clarifies the underlying physics of the probabilistic models used for life prediction. • Unveil failure of carbon nanotubes by atomistic simulations. • Establish multi-step kinetic models for prediction of life and mechanistic shift. • Elucidate the physics behind the probabilistic life models and the memory effects. [ABSTRACT FROM AUTHOR]