Massachusetts Institute of Technology
Department of Materials Science and Engineering
                         77 Massachusetts Avenue, Cambridge MA 02139-4307
                              3.21 Kinetics of Materials—Spring 2006
                                          February 27, 2006
                          Lecture 7: Solutions to the Diffusion Equation—I.
                                             References
1.	 Balluffi, Allen, and Carter, Kinetics of Materials, Sections 4.3–4.5.
                                            Key Concepts
•	 When the diffusivity D is concentration-dependent, the diffusion equation is nonlinear and closed-
   form solutions to practical problems don’t exist. The “Boltzmann–Matano” method is a graphical one
   for using a measured c(x)
                          � �profile     from a diffusion-couple experiment to determine D(c), using the
                                  � c1	                                                 � cL
            �           1
   relation D(c1 ) = − 2τ dcdx
                                c1 cR x(c) dc after setting the position x = 0 such that cR x dc = 0.
•	 Examination of asymmetry in an interdiffusion profile c(x) gives useful information about trends in
   the concentration dependence of D(c): D will be larger on the side with the shallower c(x) profile,
   and D will be smaller on the side with the steeper c(x) profile (see KoM Exercise 4.2).
•	 When D is time dependent (e.g., when temperature changes occur� during a diffusional process), a
   simple approach using a time-weighted diffusivity defined by τD = 0t D(t ) dt allows Fick’s second
                                                                                ′    ′
   law to be transformed into the alternate linear form ∂τ∂cD = ∇2 c. Familiar solution methods to solving
   the diffusion equation such as error functions and point sources can be readily adapted to cases where
   D is time dependent.
•	 In crystals and other anisotropic materials, D is generally anisotropic. Because D relates two vectors,
   D is a second-rank tensor quantity. Note however that symmetry considerations dictate that for cubic
   crystals, D is isotropic.
•	 The mathematical description of anisotropic diffusion depends on the choice of coordinate axes. Fre
   quently, the most convenient choice is parallel to high-symmetry crystal axes.
•	 When anisotropic diffusion is described in special coordinate axes termed principal axes, the diffu
   sivity tensor is diagonal, and diffusive fluxes along each principal axes are effectively uncoupled.
•	 Given a diffusivity tensor, finding its eigensystem (eigenvalues and eigenvectors) determines its prin
   cipal axes and the prinicipal values of the diffusivity tensor along the diagonal in the principal axis
   coordinates.
•	 Crystal symmetry dictates the form of the diffusivity tensor in the crystal axis system, i.e., where the
   non-zero terms will be, and which non-zero terms must be equal.
•	 A scaling transformation, KoM Eq. 4.64, permits solutions for isotropic D to be readily adapted to
   cases in which D is anisotropic (see KoM Exercise 5.9).
                            Related Exercises in Kinetics of Materials
Review Exercises 4.1–4.8, pp. 91–97.