Angular distributions of gold sputtered from a (111) crystal: Dependence of spot shapes and of spot and background yields on the primary ion mass and energy and on the target temperature

Abstract

Angular distributions of gold sputtered from a single crystal with a (111) surface have been investigated as a function of the target temperature (15–550 K), the primary ion energy (0.1–270 keV) and the ion mass (He, Ne and Xe). Using the collector technique in combination with backscattering analysis of the deposits, absolute angular-differential sputtering yields were measured in a {11̄0} plane as well as at various azimuthal angles off this plane. The angular distributions showed the characteristic so-called 〈110〉 and 〈001〉 spots superimposed on an apparently random background. The essential features of the emission pattern, i.e. the shape of the spots and their relative contribution to the total sputtering yield, were found to depend on all experimental parameters. For the first time we have observed, at low primary ion energies, characteristic spots at glancing angles of emission (presumably emission due to correlated sequences in 〈101〉 directions parallel to the (111) surface, terminated by a defocusing collision). The angular dependence of spot and background emission can be described by cosine power distributions, from which the partial sputtering yields due to the spots, ΔY 110 and ΔY 001, and the background, ΔY bg, are derived by integration. The (total) sputtering yields Y thus obtained by summation over all partial yields agree well with available literature data. At low primary ion energies the relative contribution of spot emission to the total sputtering yield amounts to as much as 50%, but decreases with increasing energy to about 25% (15%) for He and Ne (Xe) impact. This finding implies that the angular distributions of particles sputtered from single crystals are sensitive to the energy and momentum distributions in the collision cascade. The ratio of the partial sputtering yields, ΔY 110 ΔY 001 , increases with increasing mean range〈 x 〉 of the primary ions, from about unity for 〈 x〉 = 1 nm to a maximum of 4–5 between 10 and 20 nm and then decreases slowly for larger 〈 x 〉. This result suggests that the range of 〈110〉 focusons may amount to a few nm. Whereas the background distribution remains unaffected by variations of the target temperature, the 〈110〉 spot and the 〈001〉) spots become broader as the temperature increases. At 15 K the half-width of the 〈110〉 spot is only 5°, but twice as large at 550 K. Spot shape analysis suggests that the 〈110〉 spot is composed of two contributions, one being strongly dependent on temperature (“focuson”) and the other one independent thereof. At ion energies between a few keV and about 50 keV the yield ratio R 110 = ΔY 110 Y shows a clearly detectable temperature effect with d R 110/d T = −(2 ± 0.5) × 10 −4/K, which is attributed to a decreasing focuson contribution to the total sputtering yield. High excitation densities in the collision cascade apparently also reduce the relative 〈110〉 spot yields. This effect of “dynamic randomization” of the crystalline structure was observed under Xe bombardment at energies above about 3–10 keV.