“…This assumption is correct for various systems including, for example, Cu exposed to air 18,63 or coated with Ta 64 or Ti, 34 and also provides a method to determine the lower bound for k, since less diffuse scattering would lead to a larger value for p which, in turn, results in a larger value for k. 60 Correspondingly, we first fit the measured room temperature resistivity data with the FS model using a fixed p ¼ 0, but allow different effective mean free paths for the two sets of samples. The fitting provides values for the mean free path of 33.0 6 0.4 and 37.6 6 0.5 nm for annealed and asdeposited samples, respectively, with a tungsten bulk resistivity of q o ¼ 5.33 lX-cm at 295 K. These k values are close to the previously reported k ¼ 39.6 nm for W(001) with partial (p ¼ 0.3) specular scattering, 65 but are considerably larger than k ¼ 19.1 nm with p ¼ 0.11 for W(011) layers, 60 and k ¼ 15.5 or 19.1 nm from bulk density functional calculations. 44,60 The physical reasons for these differences are not completely clear, but may be attributed to anisotropy effects.…”