The diffusion coefficient ($D$) quantifies how rapidly particles, molecules, or atoms spread through a medium due to random thermal motion. It is the central transport parameter governing everything from drug delivery in aqueous solutions to dopant migration in silicon wafers and carburization in steel.
This calculator resolves $D$ across two fundamentally different physical regimes: hydrodynamic diffusion in liquids via the Stokes-Einstein relation, and thermally activated jumping in crystalline solids via the Arrhenius law. It eliminates the manual handling of Boltzmann's constant, unit conversions (mPa·s → Pa·s, nm → m, kJ/mol → J/mol), and exponential arithmetic that commonly introduces order-of-magnitude errors.
Required Parameters
For the Stokes-Einstein regime (liquids):
- Absolute temperature ($T$) in Kelvin.
- Dynamic viscosity ($\eta$) of the solvent, in mPa·s (centipoise-equivalent).
- Hydrodynamic radius ($r$) of the diffusing particle, in nanometres.
For the Arrhenius regime (solids):
- Pre-exponential factor ($D_0$), entered as a mantissa and a base-10 exponent in m²/s.
- Activation energy ($E_a$) of the diffusion process, in kJ/mol.
- Absolute temperature ($T$) in Kelvin.
Theoretical Foundation & Formulas
The Stokes-Einstein Equation
Derived by Albert Einstein in his 1905 paper on Brownian motion, this relation couples the microscopic thermal energy $k_B T$ with the macroscopic Stokes drag on a spherical particle moving through a viscous continuum:
$$D = \frac{k_B T}{6 \pi \eta r}$$
Here $k_B = 1.380649 \times 10^{-23}$ J/K is the Boltzmann constant. The model assumes a rigid spherical particle much larger than the solvent molecules, no-slip boundary conditions, and a Newtonian fluid. Deviations appear for non-spherical macromolecules or when $r$ approaches solvent molecular scale.
The Arrhenius Equation for Solid-State Diffusion
Atomic diffusion through a crystal lattice proceeds by discrete jumps over potential-energy barriers, giving a strongly temperature-dependent rate:
$$D = D_0 \exp\left(-\frac{E_a}{R T}\right)$$
where $R = 8.314$ J/(mol·K) is the universal gas constant. $D_0$ captures the attempt frequency and geometric jump factors, while $E_a$ represents the energy barrier for a single atomic hop (vacancy or interstitial mechanism).
Derived Transport Metrics
From $D$, the calculator computes several practical quantities. The three-dimensional mean squared displacement over time $t$:
$$\langle r^2 \rangle = 6 D t$$
The characteristic diffusion length is:
$$L = \sqrt{2 D t}$$
And the particle mobility (Stokes-Einstein regime), linking diffusion to drift under force:
$$\mu = \frac{D}{k_B T}$$
Reference Data: Typical Diffusion Coefficients
| System | Medium | Approximate $D$ (m²/s) | Regime |
|---|---|---|---|
| O₂ in air (25°C) | Gas | $2.0 \times 10^{-5}$ | Gaseous |
| Glucose in water (25°C) | Liquid | $6.7 \times 10^{-10}$ | Stokes-Einstein |
| Hemoglobin in water (20°C) | Liquid | $6.9 \times 10^{-11}$ | Stokes-Einstein |
| Na⁺ ions in water (25°C) | Liquid | $1.3 \times 10^{-9}$ | Stokes-Einstein |
| Carbon in γ-iron (1000°C) | Solid | $3 \times 10^{-11}$ | Arrhenius |
| Copper self-diffusion (1000°C) | Solid | $1 \times 10^{-16}$ | Arrhenius |
| Silicon self-diffusion (1200°C) | Solid | $1 \times 10^{-18}$ | Arrhenius |
Typical viscosities at 25°C useful for Stokes-Einstein input: water 0.89 mPa·s, ethanol 1.07 mPa·s, glycerol ~940 mPa·s, blood plasma ~1.2 mPa·s.
Engineering Analysis & Real-World Application
Interpreting the Magnitude of D
A useful heuristic: gases cluster near $10^{-5}$ m²/s, dilute aqueous solutes near $10^{-9}$ m²/s, and solid-state systems span $10^{-12}$ down to $10^{-20}$ m²/s. The calculator highlights which phase regime your result falls into. If a Stokes-Einstein calculation returns $D \sim 10^{-11}$ m²/s, that correctly flags a large macromolecule such as a protein or nanoparticle.
Temperature Sensitivity
In the Stokes-Einstein regime, $D$ scales roughly linearly with $T$, but the viscosity $\eta$ itself drops exponentially with temperature — so in practice, warming water from 20°C to 40°C nearly doubles solute diffusivity. In the Arrhenius regime, the sensitivity is far more dramatic: raising $T$ from 800 K to 1000 K can increase $D$ by three to five orders of magnitude, which is why thermal processing windows are so narrow in semiconductor and metallurgical work.
Size and Barrier Effects
Halving the hydrodynamic radius doubles the liquid-phase diffusivity — critical for drug formulation and colloid engineering. In solids, lowering $E_a$ by just 20 kJ/mol at 1000 K multiplies $D$ by roughly an order of magnitude, explaining why interstitial atoms like carbon diffuse vastly faster than substitutional atoms like nickel in iron.
Timescale Estimation
The time to diffuse 1 μm metric is particularly valuable. For a protein in water ($D \sim 10^{-11}$ m²/s), this is ~50 ms — compatible with cellular signalling. For carbon in austenite at 1000°C, diffusing 1 μm takes ~17 seconds, setting realistic case-hardening schedules.
Frequently Asked Questions
The equation assumes a continuum solvent, but it fails when the diffusing species approaches the size of the solvent molecules themselves. For small ions or gas molecules in water, measured $D$ values can be 2–3× higher than Stokes-Einstein predicts because the "no-slip" boundary condition becomes invalid at molecular scales.
It also breaks down in highly non-Newtonian fluids, glassy or supercooled liquids near $T_g$, and for elongated or flexible macromolecules where a single hydrodynamic radius is a poor descriptor. For rod-like DNA or unfolded proteins, use Kirkwood-Riseman or bead-model corrections instead.
$E_a$ depends on both the host lattice and the diffusion mechanism. Interstitial diffusion (C, H, N in metals) typically shows $E_a$ between 80–160 kJ/mol. Vacancy-mediated substitutional diffusion is higher, usually 200–350 kJ/mol, because it requires both vacancy formation and migration energies.
Consult a tabulated source such as Smithells' Metals Reference Book or ASM Handbook Vol. 22 for verified $D_0$ and $E_a$ pairs. Never mix values from different references, since $D_0$ and $E_a$ are fit jointly to experimental Arrhenius plots and are statistically correlated.
The MSD is the ensemble-averaged squared distance a particle travels from its origin in time $t$. The $\sqrt{\text{MSD}}$ gives the typical — not maximum — displacement. Crucially, diffusion is sub-linear in distance: doubling your observation time only increases the typical distance by $\sqrt{2}$, not 2.
This is why diffusion is an inefficient transport mechanism over macroscopic distances, and why biological systems rely on active transport or convective flow beyond the ~100 μm scale. The 3D factor of 6 in $\langle r^2 \rangle = 6Dt$ reflects contributions from all three spatial axes ($2Dt$ each).
Professional Conclusion
Manual computation of diffusion coefficients is error-prone: unit conversions between mPa·s and Pa·s, and between kJ/mol and J/mol, routinely introduce three-order-of-magnitude mistakes. The exponential nature of the Arrhenius law amplifies any slip in $E_a$ or $T$ into dramatic errors in predicted $D$.
This calculator enforces SI-consistent arithmetic, exposes the derived quantities (MSD, diffusion length, mobility, characteristic timescales) that practitioners actually need, and classifies the result against known phase benchmarks. Whether you are designing a drug release profile, a thin-film deposition cycle, or a metallurgical heat treatment, it replaces an error-prone hand calculation with a physically rigorous, immediately interpretable estimate.