Length distributions in metallic alloys

We use the embedded-atom potential of Johnson to compute the length-distribution functions for a large number of fcc binary metallic alloys. From these distributions, we extract the mean lengths of the nearest-neighbor bonds, which compare well with recent extended x-ray-absorption fine-structure (EXAFS) experiments in ${\mathrm{Ni}}_{\mathit{x}}$ ${\mathrm{Au}}_{1\mathrm{-}\mathit{x}}$. In other cases, where EXAFS results are not available, we compare our results with the mean lattice parameter as determined by diffraction experiments. While the embedded-atom potential is accurate for some alloys (e.g., Ni-Au), we show that for alloys containing Pt, a simple central-force model is superior. The embedded-atom potential of Johnson predicts an unexpected contraction of the Au-Au distance in Ag-rich Au-Ag alloys. We point out that an important characteristic of any alloy potential is its ability to get the single and double defects correct.