Molecular dynamics study on melting point of tungsten nanostructures

We used molecular dynamics to study the melting point of tungsten nano-fibers. It is found that the nanoscale effect and helium impurity can significantly reduce the melting point. We use linear model and neural network model to fit the effect of two compound effects on melting point. For fuzz with 1%-4% helium and a diameter of 10-200 nm, its melting point can be reduced by 7.7%-16.4% with linear model and 6.1%-17.0% with neural network model. The actually reduction of melting point may be more than 20% due to the existence of a large number of holes in fuzz and the rough surface of fuzz. The reduction of melting point of tungsten fuzz can accelerate the dynamic phenomenon of fuzz annealing at low temperature, leading to the disappearance of fuzz. Our research is helpful to understand that the temperature limit of fuzz formation is lower than 2000 K to some extent.