Recent improved measurements of kinetic-energy distributions of electrons ejected from metals and semiconductors by slow ions make it possible to examine the energy broadening inherent in the Auger neutralization process. Broadening, defined as the full width at half maximum of a convoluting Lorentzian function, is shown in all cases to vary linearly with ion velocity for ion energies less than 100 eV. Data for ${\mathrm{He}}^{+}$, ${\mathrm{Ne}}^{+}$, and ${\mathrm{Ar}}^{+}$ ions and the crystal surfaces Ni(111), Ge(111), GaAs(111), GaAs($\overline{1}\overline{1}\overline{1}$), and GaAs(110) show that the magnitude of broadening depends both on ion and solid and varies by a factor of 5 among the ion-solid combinations studied. For 4-eV ${\mathrm{He}}^{+}$ ions the broadenings are 0.96, 0.54, 0.37, 0.17, 0.31 eV for the surfaces listed above, respectively. The velocity-dependent broadening is the sum of components due to initial-state lifetime and nonadiabatic excitation of electrons from the filled band into states above the Fermi level from which they can participate in the Auger process. Other possible broadening components—those due to final-state lifetime, energy-level shifts in the atom near the surface, and variation of surface impact parameter—are shown to be small in the low-energy range. An analysis of earlier data for ${\mathrm{He}}^{+}$ on Ge(111) at ion energies up to 1000 eV indicates that the total transition rate does not increase appreciably above 100 eV, most likely as a result of the collapse of the barrier between the ion and the solid at higher ion energies. This analysis also shows that nonadiabatic excitation accounts for at most $\frac{1}{3}$ and initial-state lifetime for at least $\frac{2}{3}$ of the broadening at energies below 100 eV. The initial-state-lifetime broadening thus deduced shows the Auger neutralization process to be very rapid, having a transition rate for 4-eV ${\mathrm{He}}^{+}$ ions on Ge(111) of about 5×${10}^{14}$ ${\sec }^{-1\mathrm{}}$. The magnitude deduced for broadening due to nonadiabatic excitation is shown to be in the range expected. Possible reasons for the variation with ion and solid are suggested also.

• Received 19 February 1965

DOI:https://doi.org/10.1103/PhysRev.139.A526