“…In this letter, we present evidence for the growth of stoichiometric Ti 1−x Al x N barrier on ͑001͒ Si by incorporating a minor amount of Al with x as small as 0.09 ͑analyzed by Rutherford backscattering spectroscopy, RBS͒, to significantly improve its thermochemical stability and capability as barrier layers for Cu metallization. 1͑b͒, in conjunction with the findings from TEM and GIXRD, indicates that the intensities of the surface Cu have drastically reduced by half due to its penetration through the barrier to form Cu 3 Si precipitates within the volume of the Si wafer; the step in the Si signal similarly represents Si mainly in the Cu 3 TiN has been partially dissociated to cause a decay of the N backscattering signal and the formation of a silicide ͑Ti 5 Si 3 ͒, consistent with the findings made by previous authors. The Ti 1−x Al x N film was grown without intentional heating by ultrahigh-vacuum reactive magnetron cosputter deposition in Ar/ N 2 discharge, and the Cu film was subsequently deposited on the barrier layer without breaking the vacuum.…”