Comparison of inter-diffusion coefficients for Ni/Cu thin films determined from classical heating analysis and linear temperature ramping analysis by means of profile reconstruction and a numerical solution of Fick's law

Classical inter-diffusion studies assume a constant time of annealing when samples are annealed in a furnace. It is assumed that the sample temperature reaches the annealing temperature immediately after insertion, while the sample temperature immediately drops to room temperature after removal, the annealing time being taken as the time between insertion and removal. Using the above assumption, the diffusion coefficient can be calculated in a number of ways. In reality, the sample temperature does not immediately reach the annealing temperature; instead it rises at a rate governed by several heat transfer mechanisms, depending on the annealing procedure. For short annealing times, the sample temperature may not attain the annealing temperature, while for extended annealing times the sample temperature may reach the annealing temperature only for a fraction of the annealing time. To eliminate the effect of heat transfer mechanisms, a linear temperature ramping regime is proposed. Used in conjunction with a suitable profile reconstructing technique and a numerical solution of Fick's second law, the inter-diffusion parameters obtained from a linear ramping of Ni/Cu thin film samples can be compared to those obtained from calculations performed with the so-called Mixing-Roughness-Information model or any other suitable method used to determine classical diffusion coefficients. (C) 2009 Elsevier B. V. All rights reserved.