Self-diffusion of alpha-iron in a large temperature gradient

Abstract

Experimental results of an attempt to measure the thermal gradient self-diffusion in iron are given. A temperature gradient of 2500°C/cm was placed on each of three iron samples for 550 hr. Platinum wire markers were placed 0.01 in. apart, normal to the sample length; the marker spacings were measured before and after the diffusion anneal. No marker shift greater than the 0.0001-in. probable error was observed; whereas marker shifts of the order of 0.001 in. were anticipated from theoretical calculations. Each of several alternative possible conclusions which may be drawn from these results are discussed: 1. (1) self-diffusion in α-iron is not via a vacancy mechanism, or 2. (2) if the vacancy mechanism predominates in α-iron, then either 1. (a) the activation energies for formation, E f , and migration of vacancies, E m , are nearly equal, or 2. (b) the role of the barrier atoms must be explicitly considered in the process of thermal activation for an atom jump in a temperature gradient, in which case the mass flow is not proportional to the difference in E f and E m . The model of thermal gradient self-diffusion is discussed and several experiments are suggested. In addition, a calculation is given for the dependence of thermal gradient self-diffusion on the separation of vacancy sources and sinks, using parallel grain boundaries normal to the temperature gradient as the only sources and sinks which are important.