An analysis has been made of the historical data for argon ion sputtering yields of 28 mono-elemental solids in the energy range 250-10 000 eV to develop an improved semi-empirical formula based on Matsunami et al.'s and Yamamura and Tawara's formulations. The best result is found if an essential component involving the target density is included in Matsunami et al.'s approach. The scatter between the predictions and the experimental data is reduced to 9.0% by generating an analytical expression for the term Q that requires none of the target-specific numbers from special look-up tables of the type used by Matsunami et al. and Yamamura and Tawara. For elements other than the 34 elements for which Q values are tabulated, the new Q values range up to a factor of 5 from the default value that they recommend.The predictions are tested by a new method for determining sputtering yields using atomic force microscopy to measure very small crater volumes sputtered in ultrahigh vacuum. Here, 26 mono-elemental targets were studied using a 5 keV argon ion beam set at 45• to the surface normal. The effect of the angle of incidence is evaluated best using Yamamura et al.'s angular equations for this contribution. Results for Sb, Te and Bi indicate a significantly higher sputtering yield than expected. This is thought to arise from the high content of polyatomic groups associated with the low sublimation energies for such clusters in the pure elements of groups V-B and VI-B of the Periodic