Diffusion plays a critical role in determining the rate at which gases migrate through porous systems. Accurate estimates of diffusion coefficients are essential if gas transport is to be accurately modeled and better techniques are needed that can be used to measure these coefficients non-invasively. Here we present a novel method for using prompt gamma activation analysis to determine the binary diffusion coefficients of a gas in a porous system. Argon diffusion experiments were conducted in a 1 m long, 10 cm diameter, horizontal column packed with a SiO2 sand. The temporal variation of argon concentration within the system was measured using prompt gamma activation analysis. The binary diffusion coefficient was obtained by comparing the experimental data with the predictions from a numerical model in which the diffusion coefficient was varied until the sum of square errors between experiment and model data was minimized. Predictions of argon concentration using the optimal diffusivity fit experimental measurements with an R 2 of 0.983. © 2012 Elsevier B.V. All rights reserved.