LAUSR

Abstract Previous research simulating gold nanowires predicts a wide range of elastic moduli and does not agree on the effect of nanowire cross-sectional size on elastic modulus. However, these studies do not use the same potentials, possibly explaining the inconsistencies and relative errors. Many of the previous studies used potentials that were not parameterized for elastic properties or the unique geometry found in nanowires. However, no study has attempted to explain the source of these simulation discrepancies. Therefore, the purpose of this study was to use molecular dynamics to determine the effect of potential type on elastic modulus and to determine whether this may explain the discrepancies in previous studies. Findings indicated that potential type had a significant effect on the predicted elastic modulus during uniaxial tension. The purpose of development and potential specific parameterization significantly affected the accuracy relative to experimental data. Further, those potentials most accurate in predicting elastic properties from simulations of bulk periodic samples were not necessarily the most accurate for nanowires. Therefore, the sources of error in previously reported studies appear to be caused by using an unsuitable potential type that was not verified for the unique geometry of nanowires. The most accurate potentials were those that considered bulk elastic properties, surface energies, and thermal properties. In addition, the study tested the effect of temperature and demonstrated significant softening of the modulus with increased temperature, as observed in previous studies. Finally, the study also considered two temperature equilibration methods (the NPT barostat and the NVT thermostat) and demonstrated that for small cross-section nanowires (less than 4 nm) the NVT thermostat resulted in large initial stresses that could affect predicted values.