Direct measurements of sputtering have been made using alkali metal targets bombarded by inert gas ions in high vacuum, the sputtered atoms being counted by a surface ionization detector. Measurements were made in the range 0-1800 ev ion energy for the ions ${\mathrm{He}}^{+}$, ${\mathrm{Ne}}^{+}$, ${\mathrm{A}}^{+}$, and ${\mathrm{Xe}}^{+}$ incident upon Na and K target surfaces. Secondary electron coefficients $\gamma$ were measured simultaneously. It was found that over most of the energy range covered, the measured sputtering coefficient ${\theta }^{\prime }$ (number of sputtered atoms reaching the detector per positive ion incident on the target) was linear in $\ln E$ for ${\mathrm{He}}^{+}$ and ${\mathrm{Ne}}^{+}$ and in $\surd E$ for ${\mathrm{A}}^{+}$ and ${\mathrm{Xe}}^{+}$, where $E$ is the ion energy. At energies less than about 150 volts the sputtering rate appeared to be controlled by the amount of kinetic energy which could, on the average, be transferred to a single surface atom in a direct two-body collision. Thus at low energies ${\mathrm{Ne}}^{+}$ produced the most sputtering from a Na surface, whereas ${\mathrm{A}}^{+}$ produced the most from a K surface. At high energies, the sputtering rate was in the same order as (but not proportional to) the mass of the ion. The effect of surface contamination was also studied. An old surface showed a lower sputtering coefficient than a fresh surface and a higher and more erratic secondary electron coefficient, the effect being most pronounced for the light ions. The absolute sputtering coefficient $\theta$ (total number of sputtered atoms per incident ion) could not be measured directly, but it was estimated from the geometry that for 1000-volt ions the value of the coefficient ranged from about 1.0 for ${\mathrm{Xe}}^{+}$ on K to 0.05 for ${\mathrm{He}}^{+}$ on Na.

• Received 28 September 1953

DOI:https://doi.org/10.1103/PhysRev.93.719