Slow-proton reemission from noble-gas solids.

Abstract

A 1-\ensuremath{\mu}sec pulsed proton beam is being used to study ${\mathrm{H}}^{+}$ thermalization and reemission from solid target surfaces in ultrahigh vacuum in order to help clarify analogous experiments using muon beams. Using a solid Ar target, vapor deposited on an \ensuremath{\approxeq}6-K Cu substrate, the reemission probability Y is 6\ifmmode\times\else\texttimes\fi{}${10}^{\mathrm{\ensuremath{-}}4}$ at a proton implantation energy ${\mathit{E}}_{\mathrm{H}}^{+}$=1.4 keV and falls with increasing energy to 3\ifmmode\times\else\texttimes\fi{}${10}^{\mathrm{\ensuremath{-}}4}$ at ${\mathit{E}}_{\mathrm{H}}^{+}$=5 keV and 2\ifmmode\times\else\texttimes\fi{}${10}^{\mathrm{\ensuremath{-}}4}$ at ${\mathit{E}}_{\mathrm{H}}^{+}$=15 keV. Ne exhibits a 25% larger yield, while the yield for Kr is a factor of 4 lower. The reemitted protons are slow, with kinetic energies of order 1 eV. The reemitted proton yield Y decreases with an \ensuremath{\approxeq}100-m time constant, presumably due to deposition of neutral contaminants associated with the incoming beam, and thus ruling out the possibility that the slow protons originate from surface contaminants. For Ar, the observed variation of Y with ${\mathit{E}}_{\mathrm{H}}^{+}$ is interpreted with the help of a Monte Carlo calculation of the stopping and backscattering of the incident protons.The observed magnitude of Y is significantly greater than the calculated backscattering yield at the higher values of ${\mathit{E}}_{\mathrm{H}}^{+}$. We therefore hypothesize that few-eV protons in the solid, which are considered ``stopped'' by the simulation, can diffuse a significant distance and escape into the vacuum. In our model, the diffusion length for few-eV protons in pristine solid Ar, ${\ensuremath{\lambda}}_{0}$, is found to be ${\ensuremath{\lambda}}_{0}$=(${50}_{\mathrm{\ensuremath{-}}20}^{+70}$) \AA{}. However, the diffusion length we deduce from our measurements and simulations varies with ${\mathit{E}}_{\mathrm{H}}^{+}$, possibly because of the interaction of the slow proton with its implantation trail. The vacancy density is computed to be too low for few-eV protons near the surface to be trapped at defects created by the energetic incoming particle. On the other hand, the proton neutralization probability could be dependent on the availability of free electrons in the ion trail of the implanted particle [O. E. Mogensen, J. Chem. Phys. 60, 998 (1974)]. Extension of our model to the case of positive muons suggests that an experiment to moderate 4-MeV ${\mathrm{\ensuremath{\mu}}}^{+}$ with a solid Ar target [Harshman et al., Phys. Rev. B 36, 8850 (1987)] may have underestimated ${\ensuremath{\lambda}}_{0}$ for ${\mathrm{\ensuremath{\mu}}}^{+}$ due to sample impurities. It appears that the prospects for making a slow ${\mathrm{\ensuremath{\mu}}}^{+}$ beam are better than we thought, but that remoderation of a few-keV ${\mathrm{\ensuremath{\mu}}}^{+}$ beam using an Ar surface might have an efficiency less than 1% due to the high muonium-formation probability.