Self-diffusion in gamma-TiAl: an experimental study and atomistic calculations

This paper presents the results of combined experimental and theoretical st udies of Ti and Al self-diffusion in the intermetallic compound gamma-TiAl. Ti self-diffusion coefficients were measured over a temperature range of 1 184-1691 K using the radiotracer Ti-44 and the serial sectioning method. Th e diffusion coefficients show a non-Arrhenius behaviour with significant up ward deviations at high temperatures (above similar to 1470 K). The pre-exp onential factor and the activation energy of Ti self-diffusion determined i n the low temperature range (<1470 K) are (D-Ti*)(0) approximate to 1.5 x 1 0(-6) m(2)/s and Q(Ti)* approximate to 250 kJ/mol, respectively. By combini ng the obtained Ti self-diffusion coefficients with interdiffusion coeffici ents measured previously by Sprengel et al. (W. Sprengel, N. Opkawa, H. Nak ajima, Intermetallics 4 (1996), 185) and using the Darken-Manning equation, Al self-diffusion coefficients in TiAl were evaluated. Within the scatter of the data points, Al self-diffusion follows the Arrhenius law with (D-Al* ) approximate to 2.1 x 10(-2) m(2)/s and Q(Al)* approximate to 360 kJ/mol. The energies of point defect formation and migration in TiAl were calculate d by molecular statics with embedded-atom potentials. Several diffusion mec hanisms in TiAl were considered, including migration of Ti and Al along the ir own sublattices by single vacancy jumps, 3-jump vacancy cycles and the a nti-structural bridge mechanism. The activation energies of Ti and Al self- diffusion by different mechanisms were evaluated as functions of the alloy composition. The calculations predict that, in the compositions studied exp erimentally, Ti diffusion at low temperatures is dominated by the vacancy m echanism, while at higher temperatures the anti-structural bridge mechanism can essentially contribute to the overall diffusivity. Al diffusion occurs by the vacancy mechanism with some contribution of 3-jump cycles and anti- structural bridges. These predictions are in good agreement with the experi mental results of this work. (C) 1999 Elsevier Science Ltd. All rights rese rved.