Life at Low Reynolds Number: Why a Micron-Scale Swimmer Cannot Coast
Drop a coin into a swimming pool and it keeps going after you let go. Coasting is so ordinary that it is hard to notice it is a physical assumption. Shrink the coin to the size of a bacterium and the assumption fails completely: the moment you stop pushing, the object stops. Not slows — stops, within a distance far smaller than its own body. Everything strange about how micron-scale organisms and colloidal particles move follows from that single fact.
The number that decides which world you live in is the Reynolds number:
Re = ρUL / μ
where ρ is the fluid density, μ its dynamic viscosity, U a characteristic speed, and L a characteristic size. Read it as a ratio of two ways a fluid can carry momentum. The numerator is inertia — momentum carried along by the bulk motion of the fluid. The denominator is viscosity — momentum diffusing sideways through friction between neighboring layers. A swimmer in a pool has Re near a million: inertia dominates, and the wake it leaves behind outlives the stroke that made it. A 2 µm bacterium swimming at 30 µm/s in water has Re of order 10−5. Viscosity wins by five orders of magnitude.
What vanishing inertia actually removes
When Re is small enough, the inertial terms in the Navier–Stokes equations can be dropped altogether and the flow obeys the Stokes equations:
μ∇2u = ∇p and ∇·u = 0
Two features of these equations do all the work. First, there is no time derivative. The flow field u at this instant is fixed entirely by the boundary motion at this instant — the fluid has no memory of what happened a moment ago. Second, the equations are linear in u and p. Double the forcing and you double the flow; reverse the forcing and you reverse the flow exactly.
The practical consequence of the missing time derivative is the coasting distance. Push a sphere of radius a to speed U and then release it; it travels roughly ρa2U/μ before viscous drag stops it. For a 1 µm particle moving at 10 µm/s in water, that is about 0.01 nm — a tenth of an atomic radius. A colloid does not glide. It is dragged, and it stops when the dragging stops.
The consequence of linearity and time-reversibility is sharper still, and it has a name: the scallop theorem, stated by Edward Purcell in 1977. A swimmer whose body executes a sequence of shapes that looks the same played forwards and backwards — a hinge that opens slowly and closes quickly, like a scallop — returns to exactly where it started, no matter how fast or slow it moves. Speed cannot rescue it, because the equations contain no time. Net displacement at low Reynolds number requires a stroke that is not its own mirror image in time. Real microswimmers obey this: E. coli rotates a helical flagellar bundle, a shape sequence with a handedness that reversal does not reproduce; sperm propagate a bending wave along a tail, which travels in a definite direction.
Where the same physics shows up in the lab
Microfluidics lives here by design. In a 50 µm channel running water at 1 mm/s, Re is about 0.05, so two streams brought together side by side do not mix turbulently — they flow in parallel and exchange material only by molecular diffusion across the interface. That is inconvenient if you want to stir, and extremely convenient if you want a precisely controlled concentration gradient, which is why so much colloidal and cell-biology measurement is done in exactly this regime.
It also sets the rules for synthetic active colloids. A Janus particle that generates its own chemical gradient still swims in Stokes flow, so its far-field disturbance decays as a power law in distance rather than trailing behind it as a wake, and its neighbors feel that disturbance essentially instantaneously. This is why hydrodynamic interactions between microswimmers are long-ranged, and why collective behavior emerges at concentrations that would be unremarkable for passive particles. The dispersion of active swimmers in confined flows is one measurable consequence.
Note that Re says nothing about how a particle competes with its own Brownian motion — that is a separate comparison, governed by the Péclet number. A colloid can sit at Re ∼ 10−6 and still have Pe anywhere from 0.01 to 1000 depending on how hard it is driven. The two numbers answer different questions: Re asks whether inertia matters at all, Pe asks whether driving beats diffusion.
Takeaway
At low Reynolds number the fluid forgets instantly and reverses perfectly. Strategies that work by storing momentum — coasting, flicking, wake-riding — return nothing. What remains available is geometry: strokes with a built-in direction in time, gradients that a particle can climb, and boundaries that break symmetry. Every design decision in microswimmer engineering, from flagellar mimics to catalytic Janus caps, is ultimately a way of manufacturing an asymmetry that the Stokes equations cannot undo.