Abstract
In this chapter, we study diffusional transport in disordered media. Diffusional transport can be modeled either by solving the diffusion equation or by studying the time developments of random walks—both approaches produce the same results. We utilize the statistical approach to examine how random walkers spread over time in both free space and on percolation clusters. We introduce a scaling theory for the behavior of this process in space and time, extending previous scaling approaches and providing new tools and insights into anomalous transport phenomena.
Introduction
Standard diffusion, such as a drop of ink spreading in water, is well-described by Brownian motion where the mean squared displacement ($\langle r^2 \rangle$) grows linearly with time ($t$). However, many real-world systems—such as oil flowing through porous rock or proteins moving across a crowded cell membrane—are "disordered." In these environments, the path is obstructed by a complex, often fractal-like geometry. The core challenge is that the standard diffusion equation fails in these contexts. This study aims to bridge this gap by determining how the geometry of a disordered medium, modeled as a percolation cluster, fundamentally changes the relationship between time and distance traveled.
Methods and Simulation
The research employs a combination of statistical physics and computational simulation. Random walk simulations are conducted where "walkers" move randomly between connected sites on a lattice. Disordered media are generated through percolation modeling, specifically focusing on the "spanning cluster"—the giant connected component that appears at the critical threshold $p_c$.
To handle the limitations of finite simulation sizes, periodic boundary conditions are implemented, allowing walkers to "wrap around" the edges of the grid to collect better long-term statistics. The study relies on synthetic datasets generated through Monte Carlo simulations of 2D lattices (typically $50 \times 50$ or $100 \times 100$).
Key Insights: Anomalous Diffusion and Scaling
A central finding is the phenomenon of anomalous diffusion. At the critical point ($p_c$), diffusion is no longer linear. Instead, $\langle r^2 \rangle \propto t^{2k}$, where $2k \approx 0.66$. This indicates that the walker spreads significantly slower than a standard random walker (where $2k = 1$). This "compact" walk occurs because the fractal dimension of the random walk ($D_w$) on a percolation cluster is greater than 2; the walker effectively gets "stuck" in the intricate geometry of the cluster.
The behavior is scale-dependent. At short times, the walker "sees" the fractal disorder and moves anomalously. At very long times (for $p > p_c$), it eventually averages out the disorder and reverts to standard linear diffusion, albeit with a reduced diffusion constant. The scaling exponents derived, such as $k$, are universal, depending on the dimensionality of the space rather than the specific lattice geometry.
Results and Analytical Framework
The study successfully derived the walk dimension $D_w$ in terms of standard percolation exponents ($\nu, \mu, \beta$):
$$d_w = 2 + \frac{\mu - \beta}{\nu}$$
By rescaling simulation data according to the proposed scaling theory, the authors achieved "data collapse," where results from different probabilities and time scales fall onto a single curve. This confirms the validity of the universal scaling model. Furthermore, it was shown that in the anomalous regime, the effective diffusion constant $D$ is not constant but decays over time as the walker encounters increasingly large-scale obstacles.
Conclusion
While simulations are limited by finite lattice sizes, the use of periodic boundary conditions and scaling theory provides robust insights into diffusion in disordered media. Future work may explore self-avoiding walks, which add further complexity to the scaling behavior. This research provides a fundamental framework for understanding transport in complex, real-world fractal environments.
Summary updated