Active Matter: When Turbulence Emerges Without Inertia

A film of filaments can bend, break into domains and generate vortical flow even when inertia is negligible. Calling it turbulence does not mean every atmospheric or high-speed pipe mechanism reappears unchanged under a microscope. The interesting question is how matter consuming energy locally builds collective motion at larger scales and then loses regularity. The driving force is no longer supplied only through boundaries; it is distributed among the constituents.

Sanchez and colleagues built active matter in 2012 from microtubules and molecular motors powered by ATP. Their assembly produces collective motion and connects chemical supply to flow. This does not imply that every filament suspension shares one regime: assembly, interfaces and fuel availability matter. ^[1]

Reynolds number compares inertia, approximately ρU²/L, with viscosity, approximately ηU/L². A small value does not prohibit instability; it says inertial acceleration is not the dominant force-balance term. In an explicitly educational example, ρ = 1000 kg/m³, U = 10 µm/s, L = 100 µm and η = 10⁻³ Pa·s yield Re = 10⁻³. These numbers do not describe a particular preparation. They show how lively, disordered motion can remain viscosity dominated.

Stokes balance then determines the velocity associated with present stresses almost instantaneously. Those stresses still evolve because orientations move, bend and respond to flow. A dynamic loop survives removal of inertia: orientation → stress → velocity → changed orientation. Mistaking negligible inertia for absent collective dynamics misses the mechanism. Time dependence can reside in structural relaxation rather than acceleration of mass.

A population of rods aligned without a preferred end is described by a director n identified with −n. An ordinary arrow would incorrectly count two states where the material has one. In two dimensions, Q = S(nn − I/2) preserves this symmetry. S describes alignment strength and n its direction. When S is small, a precise direction becomes unreliable, especially at defect cores. This prevents low-confidence orientation estimates from becoming attractive but artificial vortex maps.

The active-nematic framework connects orientation, flow, defects and confinement. The review by Doostmohammadi and colleagues helps compare microscopically different systems. Coefficients and symmetries must still be identified before transferring conclusions from motor–filament mixtures to living tissue. ^[2]

Walk around a defect while following rod orientation. If θ increases by π, the director returns to a physically equivalent orientation and q = +1/2. If it decreases by π, q = −1/2. The familiar comet and trefoil shapes follow from this geometry. Our diagram constructs θ = q arg(x + iy) and draws unoriented segments, leaving the center blank because the orientation-only model does not specify S there. Fluid rotation and director charge are different measurements.

A +1/2 and −1/2 pair has zero total charge, so pair creation need not violate the global constraint imposed by domain and boundary conditions. Defects may separate, reorganize orientation and annihilate. Their number alone cannot determine transport intensity: films with similar defect density may differ in temporal correlations and wall interactions. Structure and dynamics should be measured together rather than replacing the whole description with a singularity count.

Under this convention, ζ sets active-stress amplitude, Γ represents substrate friction and fᵉˡ elastic forces. Extensile or contractile signs must be tied to the definitions of Q and σᵃ. Constant activity in a perfectly uniform medium need not create a bulk force: force involves stress divergence and therefore orientation gradients. A small bend can generate a flow that further amplifies the bend. The instability is collective even though individual motors obey local rules.

Elasticity resists orientation gradients over a length ℓ while activity can amplify them. Dimensional comparison yields ℓₐ. This scaling fixes neither a universal coefficient nor a complete spectrum. In our reduced calculation, quadrupling |ζ| halves ℓₐ; no measured physical diameter is plotted. If K or concentration changes along with ATP, attributing every size change to ζ alone is an identification error. An attractive data collapse does not remove that problem.

Friction introduces another length. Comparing η∇²u with Γu gives ℓₛ = √(η/Γ). For variations much larger than ℓₛ, friction can limit long-range hydrodynamic coupling. These lengths are not interchangeable: ℓₐ compares activity with elasticity; ℓₛ compares viscosity with friction. Changing the substrate can therefore alter patterns without directly changing motor chemistry. Domain size and boundary conditions must also enter the comparison.

The comparison field comes from ψ = sin x sin y and u = (∂ᵧψ, −∂ₓψ). Its divergence vanishes because mixed derivatives cancel, while vorticity is 2 sin x sin y under this convention. We have constructed vortical cells without solving an active-matter equation. That counterexample matters: visual resemblance to an experimental movie does not establish mechanism. Time-dependent predictions, correlations and perturbation responses must also be tested.

A velocity field estimated from tracking or image correlation has finite resolution and noise. Differentiating it to obtain vorticity amplifies small-scale errors. A reproducible protocol should specify spatial spacing, time interval, filtering and boundary treatment. A spectral maximum that shifts strongly when smoothing changes may be an artifact. A length robust across methods and controlled perturbations is stronger evidence than one especially striking image.

The 2017 study by Doostmohammadi and colleagues examines the onset of mesoscale turbulence in an active model. It provides a starting point for distinguishing regular patterns, intermittency and developed disorder without assuming a direct analogue of a pipe’s critical Reynolds number. ^[3]

Hillebrand and Alert reported a discontinuous transition in their studied model in 2025, including coexistence and hysteresis. Their results highlight hydrodynamic coupling and confinement. They do not establish that every active material has an identical threshold or transition order. ^[4]

This is our illustrative hysteresis normal form, not a reproduction of the cited model’s parameters. A = 0 is stable for μ < 0, while a nonzero stable branch can exist for −1/4 < μ < 0. In that interval two stable states coexist, separated by an unstable branch. Increasing μ from rest and decreasing it from motion can follow different paths. History here means attraction basins and perturbations, not a mysterious memory.

Saw and colleagues connected nematic defects with stress and cell extrusion in studied epithelia. This mechanobiological connection does not automatically transfer to another tissue, where adhesion, division, death and biochemical regulation must be considered. A defect’s shape alone cannot determine an individual cell’s fate. ^[5]

A physical explanation is useful when its level of applicability is clear. An orientation tensor can summarize many details and predict selected stresses without replacing signaling pathways. Conversely, identifying a protein does not automatically determine tissue dynamics. The scientific task is to build measurable links between levels: perturb activity, observe stress and test cellular response with controls that distinguish spatial correlation from causation.

A convincing protocol would vary fuel, substrate friction and domain size separately. It would measure orientation, velocity, defect density and correlations under consistent analysis rules. It would repeat upward and downward ramps with declared waiting times, compare initial states and examine different domain sizes. Hysteresis disappearing after longer waits may indicate a slow transient rather than stationary coexistence. A phase diagram therefore requires temporal discipline as well as parameter control.

The conceptual gain is substantial: a fluid can become disordered without following the usual inertial route. Visible vortices arise through coupling among energy consumption, orientation and force. A spiral is neither a universal motor nor an explanation by itself. Its curvature, singularities and dynamics become useful when they test that coupling. Only then does an attractive pattern become a tool for understanding transport, self-organization and selected aspects of living mechanics.

Multiply Stokes balance by u and integrate over a periodic incompressible domain. Pressure supplies no net power; integration by parts turns viscosity into positive dissipation on the opposite side of the balance. In a reduced model neglecting elastic work, active input balances viscous and frictional losses. A complete model with evolving orientation additionally requires elastic work and relaxation; omitting these can confuse changing stored energy with active power.

For a mode varying over length ℓ, viscous-to-frictional dissipation scales as η/(Γℓ²) = (ℓₛ/ℓ)². At ℓ = 10ℓₛ this estimate is about one hundredth; at ℓ = ℓₛ/10 it is about one hundred. Changing dominance helps explain different correlations on different supports. These are scaling estimates, not measurements, useful for choosing model terms before fitting a movie.

For f(A) = μA + A³ − A⁵, a small perturbation around A* obeys dδA/dt = f′(A*)δA. At A* = 0, f′ = μ. For nonzero states, setting y = A*² and using μ = y² − y gives f′ = 2y(1 − 2y). Thus y > 1/2 is stable and 0 < y < 1/2 unstable. This independent calculation justifies solid and dashed branches rather than treating the picture as proof.