Cu diffusion in single-crystal and polycrystalline TiN barrier layers: A high-resolution experimental study supported by first-principles calculations

Dense single-crystal and polycrystalline TiN/Cu stacks were prepared by unbalanced DC magnetron sputter deposition at a substrate temperature of 700 degrees C and a pulsed bias potential of -100 V. The microstructural variation was achieved by using two different substrate materials, MgO(001) and thermally oxidized Si(001), respectively. Subsequently, the stacks were subjected to isothermal annealing treatments at 900 degrees C for 1 h in high vacuum to induce the diffusion of Cu into the TiN. The performance of the TiN diffusion barrier layers was evaluated by cross-sectional transmission electron microscopy in combination with energy-dispersive X-ray spectrometry mapping and atom probe tomography. No Cu penetration was evident in the single-crystal stack up to annealing temperatures of 900 degrees C, due to the low density of line and planar defects in single- crystal TiN. However, at higher annealing temperatures when diffusion becomes more prominent, density-functional theory calculations predict a stoichiometry-dependent atomic diffusion mechanism of Cu in bulk TiN, with Cu diffusing on the N sublattice for the experimental N/Ti ratio. In comparison, localized diffusion of Cu along grain boundaries in the columnar polycrystalline TiN barriers was detected after the annealing treatment. The maximum observed diffusion length was approximately 30 nm, yielding a grain boundary diffusion coefficient of the order of 10(-16) cm(2) s(-1) at 900 degrees C. This is 10 to 100 times less than for comparable underdense polycrystalline TiN coatings deposited without external substrate heating or bias potential. The combined numerical and experimental approach presented in this paper enables the contrasting juxtaposition of diffusion phenomena and mechanisms in two TiN coatings, which differ from each other only in the presence of grain boundaries. (C) 2015 AIP Publishing LLC.