The Effects of Processing on the Microstructure of Copper Thin Films on Tantalum Barrier Layers
- PDF / 1,316,757 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 70 Downloads / 209 Views
Mat. Res. Soc. Symp. Proc. Vol. 391 01995 Materials Research Society
sputtering. The base pressure for the chamber was less than 3 x 10-7 Torr and the Ar pressure was 2.3 mTorr. For all the samples, the Cu and Ta deposition rates were held constant at about 80 and 23 nm/min., respectively. Cu deposition temperature was varied from ambient temperature, roughly 30 'C, to 250 'C. Ta was deposited either at 30 'C or 100 'C. None of the samples were annealed. Grain size was characterized in each of the films by grain area measurements from plan-view transmission electron micrographs (TEM). A minimum of 1000 grains were traced for each sample, the traces were digitized, and grain area was measured by image analysis software. Grain areas were converted to equivalent circular diameters before statistical analysis. Samples were not heated in TEM sample preparation, which consisted of dimpling and ion-milling at 77 K. Microscopy was done on a Philips 430ST TEM, operating at 300 kV. Preferred orientation was measured by pole figure analysis. In this technique, the diffracted intensity of a particular set of planes is measured as a function of the angle of inclination from the film normal and the in-plane rotation. Since grains with a particular texture must have the measured planes at a specific orientation, the degree and type of texturing can be inferred from the intensity measurement. Thin films deposited on amorphous substrates typically exhibit textures with axial symmetry, thus measuring intensity as a function of tilt angle (the fiber plot) is generally sufficient to characterize the pole figure. In this paper, fiber plots from {111 } Cu planes are presented, however, {200 }diffracted intensity was also measured from most samples to confirm the indexing of specific peaks. Following the work of Knorr [13], data reduction included subtraction of non-Bragg intensity, normalization to Bragg intensity measured from a randomly oriented powder sample in order to correct for geometric defocusing, an absorption correction and normalization by equal area integration. This results in normalized intensity that is expressed as a fraction or multiple of random intensity, where intensity from a completely randomly oriented sample would have a value of one. Intensity was measured on a four circle diffractometer using either a conventional CuK, source or synchrotron radiation with a wavelength of 1.180 A at the Stanford Synchrotron Radiation Laboratory (SSRL). For the absorption correction applied to the scans in this study, .tt, the product of the linear absorption constant and the film thickness, was measured experimentally from a peeled film for CuK, radiation. That value was converted to pt, the product of the density and film thickness, using the mass absorption coefficient for CuKa radiation. Then .tt for a wavelength of 1.180 A was calculated by multiplying pt by an appropriate mass absorption coefficient [14]. RESULTS AND DISCUSSION Typical fiber texture components in Cu films have been well characterized by Tracy and Knorr [1
Data Loading...
