Formula and Method for Stokes' Law
Stokes' law, published by George Gabriel Stokes in 1851, describes the viscous drag force on a small sphere moving slowly through a fluid: Fd = 6πμrv, where μ is the fluid's dynamic viscosity, r is the sphere's radius, and v is its velocity relative to the fluid. For a sphere falling (or rising) freely under gravity, that drag force grows with speed until it exactly balances the sphere's net weight — the point where it stops accelerating and falls at a constant terminal velocity. This calculator solves for that terminal velocity, the drag force at that speed, and the Reynolds number that tells you whether Stokes' law is actually valid for your inputs.
Deriving the terminal velocity from a force balance
At terminal velocity the sphere has zero acceleration, so the three forces acting on it — weight, buoyancy, and drag — sum to zero: (4/3)πr³ρpg = (4/3)πr³ρfg + 6πμrv. Solving for v cancels a factor of πr and rearranges to v = (2/9) r² (ρp − ρf) g / μ, where ρp is the sphere's density and ρf is the fluid's density. If the sphere is denser than the fluid (ρp > ρf), gravity wins and it settles downward; if it is less dense — an air bubble or an oil droplet in water, for example — buoyancy wins and it rises instead. Once v is known, plugging it back into Fd = 6πμrv gives the drag force at that terminal speed.
Checking validity with the Reynolds number
Stokes' law assumes "creeping flow," where viscous forces dominate over inertial ones and no turbulent wake forms behind the sphere. That assumption is only accurate when the particle Reynolds number, Re = ρfvd/μ (d is the diameter), stays below about 1 — ideally under 0.1 for high precision. This matters for unit choices too: keep diameter in meters, densities in kg/m³, and viscosity in Pa·s (or convert consistently) so every term cancels correctly; water at 20°C has a viscosity of about 1.002 mPa·s (0.001002 Pa·s), which is why that value is the default here. If your Reynolds number comes out above 1, Stokes' law underestimates the true drag and you should switch to an empirical drag-coefficient correlation (such as the intermediate or Newton's-law drag regimes) instead of trusting this formula.