Viscosity-Stratified Lift-Up Mechanism

• Viscosity-stratified lift-up mechanism is a process where wall-normal velocity fluctuations interact with mean temperature gradients to generate viscosity and temperature streaks in wall-bounded flows.
• The mechanism operates via two pathways— a modified classical lift-up route and a new scalar-mediated channel—whose cooperation or competition depends on the underlying flow geometry.
• Numerical simulations and operator decomposition analyses show that stratification amplitude, Prandtl number, and flow configuration critically influence streak amplification and transition behavior.

Searching arXiv for the primary paper and closely related work on viscosity-stratified lift-up mechanisms. The viscosity-stratified lift-up mechanism is a streak-generation pathway in wall-bounded shear flow in which wall-normal velocity fluctuations interact with a mean temperature gradient to produce temperature, and hence viscosity, streaks that feed back into the streamwise momentum equation. In the formulation developed for viscosity-stratified Poiseuille and Couette flows, this pathway appears alongside a viscosity-modified version of the classical lift-up mechanism and changes both the efficiency and wall-normal localization of streamwise streak amplification (Madhusudanan et al., 28 Aug 2025).

1. Classical lift-up under viscosity stratification

In the classical lift-up mechanism, wall-normal and spanwise vortices extract energy from mean shear and amplify streamwise velocity fluctuations $u$ through the non-normal interaction between wall-normal velocity $w$ and the base shear $U'(z)$. In the viscosity-stratified setting considered for wall-bounded channels with temperature-dependent viscosity, this mechanism is not simply retained unchanged: viscosity variation alters the linear operators in the governing equations through terms involving viscosity and its gradients, and it also changes the mean shear profile itself, making $U'(z)$ asymmetric when a temperature difference is imposed across the walls (Madhusudanan et al., 28 Aug 2025).

The streamwise-constant dynamics are written in terms of coupled equations for $\widehat{w}$, $\widehat{u}$, and $\widehat{\theta}$: $\begin{split} \frac{1}{Re} \bm{L}_{OS} \widehat{w} &= -ik_y\frac{\partial }{\partial z}\widehat{f}_y - k^2\widehat{f}_z \ \frac{1}{Re} \bm{L}_{SQ} \widehat{u} &= - U' \widehat{w} +\frac{1}{Re} \bm{M}_{u \theta} \widehat{\theta} + \widehat{f}_x \ \frac{1}{RePr} \bm{L}_\Theta \widehat{\theta} &= - \Theta' \widehat{w} + \widehat{f}_\theta . \end{split}$

This formulation shows that viscosity stratification modifies streak formation at two levels. First, it changes the classical $w \mapsto u$ lift-up route through the altered Orr–Sommerfeld and Squire operators and the modified base shear. Second, it introduces an additional scalar-mediated coupling through $\widehat{\theta}$. A central consequence is that the most efficient streak amplification need not occur where viscosity is lower. In stratified Poiseuille flow, the most efficient amplification associated with the modified classical mechanism is found in the more-viscous half of the channel, a result attributed to asymmetry-induced interaction between otherwise non-contributing symmetric singular modes and anti-symmetric modes. In Couette flow, by contrast, this effect is absent and the amplification remains more efficient in the less-viscous half, roughly as in the unstratified case (Madhusudanan et al., 28 Aug 2025).

2. Operator decomposition and the new pathway

The paper isolates the streak response through the transfer-function decomposition

$w$0

Within this decomposition, $w$1 is the viscosity-modified classical lift-up route and $w$2 is the new viscosity-stratified lift-up route (Madhusudanan et al., 28 Aug 2025).

The modified classical mechanism is

$w$3

The new mechanism is

$w$4

Its physical sequence is explicit. Wall-normal velocity fluctuations $w$5 act on the mean temperature gradient $w$6 through the term $w$7 in the scalar equation, generating temperature streaks. Because viscosity is temperature-dependent, these are simultaneously viscosity streaks. The induced scalar field then enters the streamwise momentum equation through the coupling operator $w$8, producing additional streamwise streak amplification. This is the defining content of the viscosity-stratified lift-up mechanism (Madhusudanan et al., 28 Aug 2025).

The scaling of $w$9 distinguishes it from the purely shear-mediated route. Its importance increases with the Prandtl number $U'(z)$0 and with the temperature or viscosity contrast $U'(z)$1. At sufficiently high $U'(z)$2 or sufficiently large stratification, the viscosity-stratified mechanism can dominate streak generation over the classical mechanism. This makes the mechanism not merely a correction to classical lift-up, but a separate amplification channel whose relative weight depends on scalar transport and constitutive coupling (Madhusudanan et al., 28 Aug 2025).

3. Cooperation, competition, and geometry dependence

The observable streak field is determined by the combined response of $U'(z)$3 acting on optimal forcing. The key result is that the two mechanisms do not combine in a geometry-independent way: they may either cooperate or compete, and the sign of that interaction differs between Poiseuille and Couette flow (Madhusudanan et al., 28 Aug 2025).

In viscosity-stratified Poiseuille flow, the streaks generated by $U'(z)$4 and $U'(z)$5 cooperate in the more-viscous half and compete in the less-viscous half. The consequence is enhanced net amplification in the more-viscous region and a shift in the wall-normal location of streak-energy maxima. The resulting trend can be counterintuitive, because it implies that increased viscosity may enhance rather than suppress turbulence-related streak activity in part of the flow. The paper attributes this behavior to stratification-induced asymmetry and to mode interactions revealed by singular-value decompositions of the relevant operators (Madhusudanan et al., 28 Aug 2025).

In viscosity-stratified Couette flow, the overall trend is reversed. Cooperation between the two mechanisms occurs in the less-viscous half, and that is where larger streak amplification is found. The asymmetry that is decisive in Poiseuille flow is much less significant in Couette flow because the mean shear is symmetric. Accordingly, the classical lift-up pathway remains dominant in the less-viscous region, and the new scalar-mediated pathway reinforces it there rather than shifting activity toward the more-viscous side (Madhusudanan et al., 28 Aug 2025).

This Poiseuille–Couette contrast is a defining feature of the mechanism. A common simplification is to treat viscosity stratification as a monotone modifier of streak amplification, but the operator analysis shows that the sign and localization of the effect depend on base-flow geometry and on the relative phasing of the two streak-generating routes.

4. Numerical substantiation and robustness

The theoretical predictions were substantiated by numerical experiments and by direct numerical simulations in minimal channels. In the minimal-channel DNS, Poiseuille flow displayed higher energy in the more-viscous half, whereas Couette flow displayed higher energy in the less-viscous half, matching the theoretical trend based on the interaction of $U'(z)$6 and $U'(z)$7 (Madhusudanan et al., 28 Aug 2025).

The parametric dependence of streak amplification was examined as a function of spanwise wavenumber $U'(z)$8, Prandtl number, and viscosity profile. These calculations showed the isolated contribution of $U'(z)$9, the separate and net pathway amplification, and the conditions under which the viscosity-stratified route becomes dominant. The same framework explains why the wall-normal location of energetic streaks can move as stratification is varied, rather than simply intensify at a fixed position (Madhusudanan et al., 28 Aug 2025).

Robustness was also tested against different constitutive laws for viscosity variation. The reported trends hold for viscosity–temperature relations based on an exponential fit for liquids and on Sutherland’s law for gases. This matters because the mechanism depends on viscosity gradients and scalar coupling rather than on a single special constitutive choice. A plausible implication is that the mechanism is structurally tied to viscosity stratification itself, not to one narrowly tuned viscosity law, although the quantitative amplification remains model-dependent (Madhusudanan et al., 28 Aug 2025).

5. Relation to nonlinear optimal perturbations and transition scenarios

A related line of work on nonlinear optimal perturbations in a viscosity-stratified channel showed that the early evolution of perturbation growth differs qualitatively from the constant-viscosity case and that nonlinearity is essential for capturing the full dynamics. In that study, the Orr mechanism operated at low perturbation amplitudes and a modified lift-up mechanism at high amplitudes. The nonlinear optimal perturbation initially contained more energy on the hot, less-viscous side, with a stronger initial lift-up there, but as the flow evolved the dynamically important activity shifted to the cold, more-viscous side, where wide high-speed streaks of low viscosity grew and persisted and strengthened the inflectional quality of the velocity profile (Thakur et al., 2020).

That nonlinear-optimal picture is consistent with the later operator-level identification of geometry-dependent streak amplification in the more-viscous region of Poiseuille flow. The 2020 study also emphasized that linear optimal perturbations miss most of the relevant physics: at low initial energy the linear optimal remained localized near the hot wall and did not reproduce the later cold-wall dominance seen in the nonlinear dynamics. The Prandtl number did not qualitatively affect those findings over the time horizons studied, whereas in the later transfer-function analysis the importance of the explicitly scalar-mediated route $U'(z)$0 increases with $U'(z)$1 (Thakur et al., 2020).

The connection to transition is direct. Persistent wide streaks and intensified inflectional structure on the more-viscous side provide a route to secondary instability. The nonlinear-optimal study positioned these observations relative to earlier DNS by Zonta et al. (2012) and Lee et al. (2013), which had found that turbulence tends to be suppressed near hot, less-viscous walls and enhanced near cold, more-viscous walls; the viscosity-stratified lift-up framework provides a linear-operator explanation for part of that asymmetry (Thakur et al., 2020).

6. Broader context, analogies, and delimitations

The viscosity-stratified lift-up mechanism belongs to a broader class of scalar-mediated hydrodynamic amplification processes, but it should not be conflated with all instabilities induced by viscosity gradients. In an inertialess falling film with continuous viscosity stratification, a surface-mode instability exists even in the Stokes limit. There the feedback loop involves advection of the base viscosity gradient, production of a viscosity perturbation through an advection–diffusion equation, and a phase shift that causes perturbation vorticity to lag the interface displacement. The instability exists only within a finite Péclet-number window, and both the base-state viscosity gradient and the perturbation viscosity are necessary. That mechanism was described as structurally resembling the surfactant-driven Marangoni instability in creeping two-layer flows (Gundavarapu et al., 9 Apr 2026).

Miscible inclined-film studies identify another related but distinct pathway: overlap modes arise when the critical layer of the disturbance coincides with the mixed layer containing a viscosity gradient, enabling efficient energy extraction from the base flow. In that setting, wall slip can delay the dominant surface mode and may stabilize or destabilize overlap modes depending on the position of the mixed layer relative to the wall (Ghosh et al., 2016). These results reinforce the general point that viscosity gradients can enter stability problems through explicit perturbation coupling terms rather than through mean-profile modification alone.

A useful contrast is provided by stable density stratification in the three-dimensional Boussinesq equations about Couette flow. There, buoyancy coupling transforms the streak dynamics into an oscillatory system for the zero streamwise mode, eliminating the algebraic lift-up effect and yielding exponential decay under viscosity and thermal diffusion (Zelati et al., 2023). This juxtaposition delimits the specificity of the viscosity-stratified lift-up mechanism: stratification does not have a universal effect on lift-up. In some settings, scalar coupling creates a new streak-generation route; in others, it replaces algebraic growth by bounded oscillation.

Accordingly, the viscosity-stratified lift-up mechanism is best understood as a specific wall-bounded, streamwise-constant, scalar-coupled amplification pathway in which temperature-induced viscosity variation feeds back into the Squire dynamics. Its significance lies in showing that streak formation in variable-viscosity shear flows is governed not only by classical non-normal shear extraction, but also by viscosity-mediated coupling whose strength, sign, and localization depend on geometry, Prandtl number, and stratification amplitude (Madhusudanan et al., 28 Aug 2025).