Molecular diffusion plays a very important role in various reservoir processes, especially in the oil-recovery processes where convective forces are not dominant or when direct frontal contact and mixing are not possible. For example, in heavy-oil and bitumen recovery, injected light hydrocarbons can diffuse into the oil beyond the potential fronts and/or convective zones and promote the effectiveness of the displacement process, reducing in-situ viscosities and in turn enhancing the oil recovery.Similarly, diffusive mixing can also be a dominant mechanism in the gas-redissolution process, even in lighter-hydrocarbon systems. For example, it controls how much gas will be dissolved in oil and how long it will take to dissolve, in the absence of mechanical/convective mixing, as in the case of reservoir repressurization. The extent of dissolution of a gas into oil is governed by its solubility, but the rate is controlled by both molecular diffusivity and solubility. Thus, accurate determination of these parameters is essential to design and understand displacement processes.Despite the significance of diffusion in various aspects of oil recovery, there are very few experimental studies available in the literature addressing the diffusion of gas in heavy oils. Experimental work is most commonly based on the pressure-decay concept. However, the parameter inversion in these tests relies on an error-function solution that neglects the transient processes at the gas/oil interface and assumes constant-saturation concentration. This assumption is not appropriate when decay in pressure is large because pressure in the gas cap changes continuously as gas is dissolved in the oil, and hence the gas solubility varies with time. One of the major issues related to this experimental process is that it takes a long time (order of several days to several months) to achieve steady-state (converged) solution to determine diffusivity.In this work, we have • Experimentally investigated the diffusion of methane in heavy oils as well as light oils by use of a pressure-decay test • Captured properly the variation in gas concentration in oil at the gas/oil interface with time by expressing gas solubility in terms of Henry's constant in the mathematical model • Developed the exact solution of the 1D pressure-decay (transient-diffusion) model with pressure-dependent gas/oilinterface concentration and shown that after a long time, pressure decays exponentially in time with an exponent that depends on diffusivity as well as solubility • Presented the inversion technique to determine the diffusivity and other parameters from late-transient-pressure data, and shown the convergence in their estimates • (Most importantly) developed a cutoff criterion permitting us to stop the experiments while still being able to extract the converged diffusivity values (this is important in situations when the experiment is stopped prematurely for technical or other reasons)