Taylor-Couette Flow: Unmixing Colors with Reversible Laminar Flow

Chaotic advection, a mixing mechanism arising from deterministic dynamical systems with high sensitivity to initial conditions, enables efficient mixing in low-Reynolds-number flows by creating exponential particle separation through stretching and folding; this principle was applied to improve DNA chip hybridization efficiency by 30-100 times compared to static diffusion alone, demonstrating that three-dimensional time-periodic flows can achieve rapid molecular mixing without requiring turbulence.

In laminar flow regimes where viscosity dominates over inertia, mixing can occur efficiently not through turbulence but through chaotic advection—where stretching and folding of fluid elements, combined with interfacial effects like the Marangoni effect and evaporation-driven convection, creates complex internal flows that dramatically enhance mixing rates beyond what molecular diffusion alone could achieve.

Chaotic advection in porous media occurs through exponential divergence of nearby trajectories, enabling efficient mixing via stretching and folding mechanisms analogous to baker's map dynamics used in bread-making. While steady low-Reynolds-number flows in two dimensions cannot exhibit periodic chaos, three-dimensional porous media can support chaotic advection, potentially representing the norm rather than exception. In body-centered cubic bead packings, varying mean flow orientation relative to packing symmetries controls chaos intensity, measured by Lyapunov exponents. Non-chaotic flows produce regular, periodic trajectories with minimal mixing, while chaotic orientations enable trajectories to braid and bend, getting transported to different spheres and creating net exponential stretching that leads to efficient chaotic mixing.

At low Reynolds numbers, fluid motion is time-reversible: particles follow reversed streamlines when movies are played backward. Two particles return to starting positions, but three or more particles exhibit chaotic behavior where small initial position differences lead to exponentially divergent trajectories. Dave Pine and Jerry Golub's experiments revealed a sharp strain amplitude threshold: below it, diffusion constants are essentially zero; above it, diffusion becomes anisotropic and increases linearly. The explanation involves three-body collisions causing chaos. However, this was debated as a crossover rather than a true phase transition. The random organization model simulation confirmed this is a dynamical phase transition with power-law critical slowing down.

At very low Reynolds numbers (creeping flow regime), fluid motion exhibits kinematic reversibility: reversing the driving forces causes the fluid to retrace its exact path backward. This was demonstrated experimentally using concentric cylinders where colored fluid precisely retraced its path upon reversing rotation. This property has profound implications for mixing: without molecular diffusion, mixing could theoretically be undone by reversing the flow. However, diffusion always causes irreversible mixing at the molecular scale. The distinction between simple shear (linear thinning) and pure shear (exponential thinning) determines mixing efficiency. Most mantle mixing occurs at plate boundaries where pure shear dominates, explaining why heterogeneity is most rapidly destroyed at mid-ocean ridges and subduction zones.

The Taylor-Couette flow is a configuration where fluid exists in the gap between two concentric cylinders. When the inner cylinder rotates while the outer cylinder remains stationary, one of the simplest laminar solutions derived analytically from Navier-Stokes equations shows non-zero azimuthal velocity (vθ) with zero radial velocity (vr) and zero axial velocity, representing a steady, unidirectional flow independent of time.

The Taylor-Couette instability occurs in fluids confined between concentric rotating cylinders with different angular velocities. As the angular velocity difference increases, the initially featureless flow becomes unstable and develops toroidal convective vortices between the cylinders. This phenomenon was studied extensively in laboratory settings and has practical applications, such as in traditional South Indian dough-making for dosa preparation, where the inner grinder rotates within a stationary outer cylinder.

Taylor flow research began nearly 80 years ago with Fairbrother and Stubbs studying bubble velocity in capillaries. In 1961, Taylor conducted experiments across capillary numbers (Ca = μU/σ), measuring film thickness and theoretically analyzing bubble shape, velocity, and pressure drop. This work was motivated by understanding blood flow in capillaries. Applications include: blood flow in capillaries, lung airway mucus plugs, membrane fouling removal, electronics cooling, fuel cells, oil/gas displacement, and chemical micro-reactors. About 80-90% of microfluidic work involves Taylor flow regime.

The lecture introduces Taylor-Couette flow, which is the rotational motion between two circular cylinders of infinite length. The inner cylinder rotates with angular velocity Omega 1, while the outer cylinder may rotate with angular velocity Omega 2. The cylinders have radii R1 and R2, with the space between them defined as Delta. The fluid between the cylinders has known parameters: density and viscosity. This is a classic system that can be observed in a laboratory setting.

Taylor-Couette flow is a fundamental experimental setup in fluid mechanics consisting of two concentric cylinders with an inner cylinder that can be rotated and an outer transparent cylinder. The fluid contained between these cylinders allows measurement of viscosity from the torque applied to the inner cylinder. This device is the most commonly used viscometric device for measuring fluid viscosity.

This lesson demonstrates how to solve the Navier-Stokes equations in cylindrical coordinates for fully developed laminar pipe flow, showing that the velocity profile is parabolic (u = -(-ΔP/ΔX)(R²/4μ)(1 - r²/R²)) and deriving key properties including maximum velocity, volume flow rate, average velocity, and Darcy friction factor (f = 64/Re).

This comprehensive section covers the complete analysis of Couette flow between two concentric cylinders. The problem involves fluid with viscosity μ between an inner cylinder rotating at constant angular velocity Ω and a stationary outer cylinder. The instructor begins with course logistics and introduces Navier-Stokes equations in cylindrical coordinates. Key assumptions are made: purely tangential flow (only u_θ component), two-dimensional (no z-variation), and axisymmetric (no θ-variation). The instructor first presents a simplified linear velocity profile solution, then transitions to the full Navier-Stokes solution. Mass conservation confirms u_θ varies only with r. The θ-component of Navier-Stokes reduces to μ × [∂/∂r(r × ∂u_θ/∂r)] = 0, integrating to u_θ = C₁r + C₂/r. Applying boundary conditions u_θ(r_i) = Ωr_i and u_θ(r_o) = 0 yields the complete velocity profile u_θ = Ωr_i × [r_o² - r²] / [r(r_o² - r_i²)]. The shear stress is calculated as τ_rθ = -2μΩr_i × r_o² / [r²(r_o² - r_i²)], and torque on both cylinders is T = 4πμΩr_i²r_o²L / (r_o² - r_i²), confirming moment balance. The radial pressure distribution is found from dp/dr = ρu_θ²/r, showing pressure increases radially outward to provide centripetal force.

The Navier-Stokes equations are fundamental to fluid dynamics and essential for solving exam problems. The course covers two main problem types: Navier-Stokes problems and problems from the second unit, with theoretical questions always included. The Couette flow problem involves two parallel plates separated by distance H, with a fluid of viscosity μ and density ρ between them. One plate is stationary while the other moves at velocity U₀. The fundamental equations are the continuity equation (conservation of mass) and the momentum equation (conservation of linear momentum), which can be separated into x, y, and z components. Stating assumptions is crucial, as failure to do so may result in penalized solutions. The four main assumptions are: steady flow (all time derivatives are zero), incompressible flow (density remains constant), Newtonian fluid (viscosity remains constant), and bidimensional flow (no velocity or variations in z-direction). The continuity equation for incompressible flow simplifies to the divergence of the velocity field equals zero. The velocity field is a vector with components u (x-direction), v (y-direction), and w (z-direction). For Couette flow, since fluid only moves in the x-direction, v = 0 and w = 0, implying ∂u/∂x = 0, meaning u is a function of y only. The momentum equation in the x-direction is used when fluid moves only in the x-direction. For steady flow, time derivatives are zero. From continuity, ∂u/∂x = 0, so its second derivative is also zero. Since v = 0 and w = 0, all terms involving these velocities are eliminated. The pressure gradient term can be eliminated in Couette flow because motion is caused by plate displacement, not pressure differences. The simplified equation is μ times the second derivative of u with respect to y equals zero. This is integrated twice with respect to y to find the velocity profile. The boundary conditions are: at y = 0 (stationary plate), u = 0 (no-slip condition), and at y = H (moving plate), u = U₀. The first condition gives C₂ = 0, and the second gives C₁ = U₀/H. The final velocity profile is u = (U₀/H)y, showing linear velocity increase from zero at the stationary plate to U₀ at the moving plate.

This comprehensive lesson covers the derivation of fundamental fluid mechanics equations in cylindrical coordinates. The key insight is that unit vectors e_r and e_θ change direction with θ (∂e_r/∂θ = e_θ, ∂e_θ/∂θ = -e_r), unlike rectangular coordinates where unit vectors are constant. For the continuity equation ∇·V = 0, expanding in cylindrical coordinates yields ∂v_r/∂r + v_r/r + (1/r)∂v_θ/∂θ + ∂v_z/∂z = 0, or equivalently (1/r)∂(r v_r)/∂r + (1/r)∂v_θ/∂θ + ∂v_z/∂z = 0. The extra term v_r/r arises from unit vector derivatives and would be missed by blindly extending rectangular equations. For Navier-Stokes, the convective term (V·∇)V has additional terms: -v_θ²/r in the r-component and (v_r²)/r in the θ-component, representing centrifugal effects. The velocity gradient tensor ∇V contains off-diagonal terms with extra contributions from unit vector derivatives. The viscous terms μ(∇²V + ∇(∇·V)) can be computed using the same methodology.

This video demonstrates the mathematical solution of Navier-Stokes equations for fluid flow in cylindrical coordinates, covering key simplifying assumptions including constant density, constant velocity, and symmetry conditions, with practical applications in calculating flow rates and pressure gradients using proper integration techniques and SI unit systems.