We use the hyperbolic subdiffusion equation with fractional time derivatives (the generalized Cattaneo equation) to study the transport process of electrolytes in media where subdiffusion occurs. In this model the flux is delayed in a non-zero time with respect to the concentration gradient. In particular, we obtain the formula of electrochemical subdiffusive impedance of a spatially limited sample in the limit of large and of small pulsation of the electric field. The boundary condition at the external wall of the sample are taken in the general form as a linear combination of subdiffusive flux and concentration of the transported particles. We also discuss the influence of the equation parameters (the subdiffusion parameter and the delay time) on the Nyquist impedance plots.
Using the quasistatic approximation, we show that in a subdiffusion-reaction system with arbitrary nonzero values of subdiffusion coefficients, the reaction front evolves in time as , with being the subdiffusion parameter and being controlled by the subdiffusion coefficients. To check the correctness of our analysis, we compare approximate analytical solutions of the subdiffusion-reaction equations with the numerical ones.
We consider the subdiffusion-reaction process with reactions of a type (in which particles are assumed to be mobile, whereas are assumed to be static) in comparison to the subdiffusion-reaction process with reactions which was studied by Sokolov, Schmidt, and Sagués [ Phys. Rev. E 73, 031102 (2006)]. In both processes a rule that reactions can only occur between particles which continue to exist is taken into account. Although in both processes a probability of the vanishing of particle due to a reaction is independent of both time and space variables (assuming that in the system with the reactions, particles are distributed homogeneously), we show that subdiffusion-reaction equations describing these processes as well as their Green's functions are qualitatively different. The reason for this difference is as follows. In the case of the former reaction, particles and have to meet with some probability before the reaction occurs in contradiction with the case of the latter reaction. For the subdiffusion process with the reactions we consider three models which differ in some details concerning a description of the reactions. We base the method considered in this paper on a random walk model in a system with both discrete time and discrete space variables. Then the system with discrete variables is transformed into a system with both continuous time and continuous space variables. Such a method seems to be convenient in analyzing subdiffusion-reaction processes with partially absorbing or partially reflecting walls. The reason is that within this method we can determine Green's functions without a necessity of solving a fractional differential subdiffusion-reaction equation with boundary conditions at the walls. As an example, we use the model to find the Green's functions for a subdiffusive reaction system (with the reactions mentioned above), which is bounded by a partially absorbing wall. This example shows how the model can be used to analyze the subdiffusion-reaction process in a system with partially absorbing or reflecting thin membranes. Employing a simple phenomenological model, we also derive equations related to the reaction parameters used in the considered models.
Anomalous diffusion of an antibiotic (colistin) in a system consisting of packed gel (alginate) beads immersed in water is studied experimentally and theoretically. The experimental studies are performed using the interferometric method of measuring concentration profiles of a diffusing substance. We use the -subdiffusion equation with the fractional Caputo time derivative with respect to another function to describe the process. The function and relevant parameters define anomalous diffusion. We show that experimentally measured time evolution of the amount of antibiotic released from the system determines the function . The process can be interpreted as subdiffusion in which the subdiffusion parameter (exponent) decreases over time. The -subdiffusion equation, which is more general than the “ordinary” fractional subdiffusion equation, can be widely used in various fields of science to model diffusion in a system in which parameters, and even a type of diffusion, evolve over time. We postulate that diffusion in a system composed of channels and a matrix can be described by the -subdiffusion equation, just like diffusion in a system of packed gel beads placed in water.
We present a method of deriving a boundary condition for diffusion at a thin membrane from experimental data. Based on experimental results obtained for normal diffusion of ethanol in water, we show that the derived boundary condition at a membrane contains a term with a Riemann-Liouville fractional time derivative of order . Such a form of the boundary condition shows that a transfer of particles through a thin membrane is a “long-memory process.” The presented method is an example that an important part of the mathematical model of physical processes may be derived directly from experimental data.