Three in situ analysis techniques consisting of X-ray diffraction, elastic light scattering and resistance analysis have been used to investigate the failure temperatures of Cu diffusion barriers. Bilayer films consisting of 200 tint Cu on a 20-24-nm barrier (Ta, Ti7W93, Ta36Si12N52, TiN) on single and polycrystalline Si were analyzed as blanket layers and in damascene structures. The damascene structures were trenches with SiO2/Si side walls and Si(100) bottoms 0.6 mum deep, 0.23-0.62 mum in width (aspect ratios 0.97-2.6) and 5.1-80 mum in length. Barrier failure was determined by using the three techniques to monitor Cu silicide formation, which occurs at low temperatures < 250 degreesC when Cu is in direct contact with an unlimited supply of Si. The barriers were annealed at a temperature ramp rate of 3 degreesC/s from 100 to 1000 degreesC in N-2 In the first analysis method, failure temperatures were determined by monitoring the disappearance of the Cu(111) X-ray diffraction peak and appearance of Cu silicide diffraction peaks. Simultaneously, elastic light scattering was used, where barrier failure is indicated by increases in surface roughness monitored at two different lateral length scales. Increases in sheet resistance of the blanket films on poly-Si were also monitored simultaneously using a four-point probe technique. It was determined, with the three techniques, that the order of increasing effectiveness for the barriers studied was Th < Ti7W93 < Ta36Si12N52 < TiN. It was also shown that all barriers, except TiN, show a decrease in failure temperature as trench width decreases (aspect ratio increases).