Fick S Law Of Diffusion

Working With Diffusion Flux in Real Systems

Fick's first law is J = -D(dC/dx). You measure the concentration gradient across a known distance, multiply by the diffusion coefficient, and you get the flux. The negative sign means the substance moves from high to low concentration. That's it. People make it sound deeper than it is. The diffusion coefficient D is the thing that kills you. It changes with temperature, solvent, molecular size, and everything else. If you're using tabulated values from a paper for a different solvent or a slightly different temperature, your numbers will be off by 20 to 40 percent without you noticing. I've seen people spend three weeks debugging a microfluidic design only to realize the D value they plugged in was for water at 25°C while their experiment ran at 37°C in a buffer with 150 millimolar salt. The viscosity change alone accounts for most of the discrepancy.

Fick S Law Of Diffusion in practice

The second law, dC/dt = D(d²C/dx²), tells you how concentration changes over time at a given point. This is the one you actually need for transient problems. Steady-state is straightforward. Transient is where things get messy. Here's a specific problem I ran into last year. We were modeling gas permeation through a polymer membrane in a lab-scale separator. The membrane was 50 micrometers thick. I calculated the steady-state flux using Fick's first law with D = 1.2 × 10^-7 cm²/s and got a reasonable number. Then I measured the actual flux and it was half what the calculation predicted. I spent two days checking my math before I realized the membrane had a skin layer—a dense, nearly impermeable surface layer about 5 micrometers thick that formed during manufacturing. The bulk of the membrane had the D value I was using. The skin layer did not. The effective diffusion path wasn't 50 micrometers of uniform material. It was a 5-micrometer barrier on top of 45 micrometers of more permeable polymer. The workaround was simple once I knew what to look for. I treated the membrane as two resistances in series. The overall permeability became the sum of the individual layer resistances rather than a single D divided by a single thickness. That cut my error from 50 percent down to under 5 percent.

There's a common misconception that Fick's law applies broadly to any diffusion-like process. It doesn't. It assumes an ideal dilute solution, a homogeneous medium, and no convective flow. When you're dealing with concentrated solutions, the chemical potential gradient is the real driving force, not the concentration gradient. The Maxwell-Stefan formulation handles that, but nobody uses it because it requires mutual diffusion coefficients that are nearly impossible to measure directly. Another thing people miss: the diffusion coefficient isn't always constant. In porous media, in crowded biological systems, in polymer blends, D can depend on concentration itself. If D varies with C, the simple analytical solutions to Fick's second law fall apart and you're solving a nonlinear partial differential equation. Finite difference methods work, but you need a small enough spatial step that you're not introducing numerical diffusion that swamps the physical process you're trying to model. A rule of thumb is keeping dx²/dt below 2D to maintain stability in an explicit scheme, though implicit methods relax that constraint significantly. If you need a quick reference for D values, the CRC Handbook of Chemistry and Physics has tables for common solvent-solute pairs at 25°C. For temperature dependence, the Stokes-Einstein relation gives you a starting point: D = kT/(6r). It's approximate. It assumes spherical particles in a continuum fluid. It breaks down at the nanometer scale and in viscous media. Still, it's useful for rough estimates when experimental data is unavailable.

The biggest practical limitation of Fick's framework is that it ignores coupling effects. In multiphase systems, temperature gradients drive mass flow (the Soret effect), and concentration gradients can drive heat flow (the Dufour effect). These are normally small, but in systems with steep thermal gradients—like during rapid cooling of a thin film or in geothermal brine transport—they become significant. If your calculated flux consistently runs 10 to 15 percent high compared to measurements and you're working in a system with temperature variation, check whether thermal diffusion is contributing. For most engineering applications, Fick's law is sufficient if you respect its assumptions and verify your D values against your actual conditions. It's not elegant. It's not universal. But it's the baseline you start from before layering on corrections for non-ideality, convection, or coupled transport.