Helium Microbeam Mixing of Bilayers.

This study is an experimental and theoretical investigation of room-temperature mixing of bilayers by micron-width He^+ ion beams. Bilayer targets, including Cu/Al, Cu/Si and Sb/Si, were irradiated at room temperature in the University at Albany's Dynamitron particle accelerator with 2-MeV He^+ ion beams ranging from 2 to 6 μm in width. At doses on the order of 10^ {19}/cm^2, RBS spectra revealed evidence of interface mixing in all targets to depths of several thousand A within the cylinder irradiated by the beam. Both RBS spectra and isometric RBS contour maps of the target also showed that mixing of the interface extends laterally well beyond the irradiated area. The interface mixing reaches a maximum in an annular region several times larger in diameter than the ion-beam. Standard theories of primary-recoil, secondary -cascade and thermal-spike mixing predicted interface widths two orders of magnitude smaller than observed for Cu/Al bilayers. Furthermore, He^+ irradiation of Cu/Al targets at liquid-nitrogen temperature did not produce interface mixing, further indicating that ballistic interpretations of the mixing are inadequate. Defect concentrations as a function of position and time were calculated by numerical solution of coupled rate equations for vacancies and interstitials in aluminum. The results of these calculations show that room-temperature He^+ mixing of Cu/Al results almost exclusively from interstitial migration. The numerically calculated concentration of interstitials within the damage cylinder was used to derive an approximate expression for interface width as a function of dose. Comparisons of these predicted values with the experimentally determined interface width as a function of dose agree, within uncertainties. In addition, the annular region observed on RBS maps is explained by the continued presence of a non-equilibrium concentration of interstitials after the ion beam is shut off. Interface mixing in Cu/Si targets, although qualitatively similar to that in Cu/Al bilayers, was significantly greater for similar doses. The higher rate is consistent with the fact that defect trapping in single-crystal Si is expected to be smaller than in polycrystalline Al.