The diffusion of Hg into monocrystalline slices was measured by using a radiotracer sectioning technique. Diffusion annealing was carried out, at temperatures ranging from 23 to 460C, in evacuated silica capsules within which a saturated vapor pressure of the metal was maintained during diffusion. A combination of anodic oxidation and dry mechanical lapping was used to determine the diffusion profiles. All of the profiles revealed 2 diffusion components which were both Fickian, and permitted a double complementary error function to be fitted to the data. When plotted on an Arrhenius graph, the results for fast diffusion could be described by E = 0.789eV and D = 2.65 x 10-6cm2/s, and were suggested to be due to either an interstitial or a defect mechanism which was associated with a residual impurity in the CdTe slice. The plot for slow diffusion comprised 2 straight-line components, with a change in slope at 300C. When a mathematical expression which consisted of the sum of 2 Arrhenius functions was fitted to the data, the results could be described by E1 = 1.572eV, D1 = 0.017cm2/s, E2 = 0.265eV and D2 = 2.84 x 10-13cm2/s. This was explained in terms of the operation of 2 diffusion mechanisms. The first was a defect mechanism which was associated with a residual impurity or interstitial that was similar to the one which had been proposed for fast diffusion, and which operated over the entire temperature range where measurements had been made. The second was attributed to a vacancy defect which became significant at temperatures above 300C. Isoconcentration measurements did not reveal a similar behavior. Arrhenius graphs of the surface concentration due to Hg were obtained for both diffusing components, and each curve exhibited a sharp kink; with a change of slope occurring at 300C. This again gave minimum values of C (slow) = 2.4 x 1018/cm3 and C (fast) = 1.1 x 1017/cm3 at that temperature. This behavior was suggested to be associated with the peritectic in the Cd/Hg and Cd/CdTe phase diagrams. The addition of Cd to the diffusion source reduced both the diffusivity and the Hg concentration at the surface of the slice. The diffusion at 460C appeared to be pressure-dependent.
E.D.Jones, V.Thambipillai, J.B.Mullin: Journal of Crystal Growth, 1992, 118[1-2], 1-13