Skip to content

Korteweg Stress

A gradient-dependent capillary contribution to fluid stress that distributes interfacial force across a finite density or composition transition under a specified constitutive model.

Version
v1 · 2026-10-03 · History
Domain-specific #
13365
Domain group
Natural Sciences
Origin domain
Physics
Subdomains
Continuum Fluid Mechanics, Capillarity → Physics

Core Idea

Korteweg stress is the gradient-dependent capillary part of a fluid's continuum stress in a diffuse-interface model. Instead of concentrating an interfacial effect solely in a jump condition on an infinitely thin surface, the model represents a transition in a field such as density or composition over finite thickness and couples that field's spatial variation to momentum balance. The capillary contribution acts through its stress divergence, or an equivalent force representation. Its detailed tensor depends on the chosen free energy, order parameter, constitutive assumptions and sign convention; no single constant-coefficient formula is the whole identity.[1][2]

In a particular one-component isothermal liquid–vapor model, a square-gradient free-energy term \(\frac{1}{2}\kappa|\nabla\rho|^2\) leads to density-gradient and density-Laplacian terms in the stress. A binary or miscible-liquid model instead often uses composition variation, with different field evolution and constitutive justification. The common abstraction is gradient-derived or gradient-posited capillary momentum stress, not proof that every mixture has a permanent equilibrium surface tension.[2][1][3]

Structural Signature

Sig role-phrases: smoothly varying fluid field — constitutive gradient relation — capillary stress contribution — momentum-balance effect — model-bound transition region.

  • Fluid field. A density \(\rho\), composition \(c\), or other justified fluid order parameter varies through a transition zone. For the single-component capillary case, density distinguishes the bulk phases; in a binary-fluid treatment, composition can play that role. The two fields do not obey identical evolution equations.[1]
  • Gradient constitutive relation. The chosen constitutive law makes the fluid stress depend on spatial derivatives of that field. An illustrative constant-\(\kappa\) square-gradient energy gives terms involving \(\nabla\rho\otimes\nabla\rho\) and \(\nabla^2\rho\) in one convention, but coefficients and signs must be read from the actual model. A raw gradient alone is not yet a stress law.[2][1]
  • Momentum coupling. The stress appears within the total continuum stress, or equivalently as a capillary force in the momentum equation. It is not a substitute for the Cauchy-stress concept: pressure, capillary/reversible and viscous/dissipative parts can coexist in the same total stress balance.[1]
  • Diffuse transition. The contribution is distributed over a nonuniform region rather than applied only as a boundary jump. Where the modeled field is spatially uniform, its gradient-dependent interfacial contribution disappears or can be absorbed into a bulk pressure convention; total pressure or viscous stress need not vanish.[1][2]
  • Model qualification. The fluid, field, free-energy or phenomenological coefficient, temperature assumptions, and treatment of energy balance determine what the term means. In non-equilibrium models, thermodynamic consistency can require accompanying energy-flux terms rather than adding a capillary tensor to momentum in isolation.[1][3]

What It Is Not

It is not ordinary viscous stress. Viscosity models dissipative response to velocity gradients; Korteweg stress represents capillary response tied to density or composition gradients. Both can occur in a viscous interfacial fluid, but setting a viscosity coefficient does not define its Korteweg contribution.[1]

It is not the entire Cauchy stress or an extra force outside continuum mechanics. Authors may separate pressure, viscosity and capillarity algebraically; Korteweg names the capillary/gradient-dependent contribution. Nor is it identical to a sharp-interface surface-tension jump. The latter applies a traction condition at a geometric surface, while the diffuse model resolves an interfacial region. The two can connect in a suitable limit but are not interchangeable without scale and constitutive assumptions.[1][2]

It is not a universal constant-\(\kappa\) tensor, a guarantee of interface stability, or a claim that miscible liquids possess a fixed immiscible-style surface tension. Chen and colleagues explicitly said the relevant miscible-flow coefficient values, signs and physical derivation were uncertain in their model; their simulated interface-like effects are conditional and transient.[3]

Scope of Application

The literal scope is fluid continuum modeling with nonuniform density or composition and a specified gradient-dependent capillary constitutive law. One well-developed case is a single-component liquid–vapor transition modeled through a finite density profile. Hosseini and colleagues derive a constant-coefficient isothermal stress in their chosen free-energy model and show why recovering the full Navier–Stokes–Korteweg continuum target matters for their numerical representation.[2]

Another case is a mixing region between miscible liquids. Kostin and colleagues' original abstract describes slowly diffusing glycerol and water with a steep concentration gradient and possible transient effective interfacial effects. Chen and colleagues' full original study uses a proposed Korteweg stress in a miscible-droplet displacement model; its front and rear respond differently as concentration gradients evolve. This is not the assertion that the miscible and liquid–vapor tensors have one identical molecular origin or calibrated coefficient.[4][3]

The stress label is justified only if the specified model couples the gradient term to momentum. A density or concentration gradient seen in data, by itself, is evidence for an interface-like transition, not evidence that a particular Korteweg coefficient or stress law has been established.[1][3]

Clarity

The abstraction distinguishes where capillary effects are represented. A sharp-interface model treats a surface as a boundary and expresses a stress jump there. A Korteweg model treats the transition as a finite zone and places a capillary stress in the continuum balance throughout it. This comparison clarifies what a claimed “surface tension” in a miscible liquid actually means: perhaps a model-dependent transient stress effect, not an equilibrium material constant of a permanent interface.[1][3]

It also separates field variation from stress. Knowing \(\nabla\rho\) or \(\nabla c\) is insufficient without the constitutive relation, and knowing a tensor formula is insufficient without its pressure convention and energy assumptions. In a uniform bulk region, the interfacial gradient term may vanish while pressure remains. This resolves the seed's misleading inference that the whole stress disappears wherever density is uniform.[1][2]

Manages Complexity

Fluid interfaces can bend, merge or break, forcing a sharp-boundary formulation to manage moving surfaces and jump conditions. A diffuse-interface model tracks a smooth order-parameter field and its gradient stress in the bulk equations, reducing one kind of geometrical bookkeeping. The NIST authors identify near-critical flows, moving contact lines and topology changes as settings where a finite interfacial zone matters.[1]

The substitution is not free. A resolved transition requires a constitutive free energy or phenomenological law, parameter calibration, sufficiently appropriate spatial resolution and compatible mass/energy balances. The Korteweg term compresses capillary action into a continuum field relation, but does not erase model uncertainty—especially for miscible-liquid coefficients or the question whether an effective tension analogy is valid.[1][3]

Abstract Reasoning

Start with a proposed fluid model and ask whether its stress depends on the spatial structure of a density or composition transition. If the nonuniform field contributes a capillary term to momentum, then it has the Korteweg signature in the model's stated convention. In a one-component constant-\(\kappa\) free-energy example, varying the square-gradient energy produces a density-gradient stress; a calculation that omits that term cannot be claimed to recover the same capillary continuum target.[2]

Then bound the inference. If a miscible droplet's simulated shape changes when the postulated concentration-gradient term is varied, that supports sensitivity within that model. It does not identify the correct physical coefficient, prove a first-principles derivation for the liquid pair, or show that every observed shape is caused by Korteweg stress. Chen and colleagues explicitly call for comparison with experiments and flag the surface-tension analogy's limitations.[3]

Knowledge Transfer

Within fluid mechanics, the diagnostic roles transfer from liquid–vapor density transitions to some binary or miscible composition transitions: a nonuniform field, a gradient-dependent constitutive term, momentum coupling and a finite transition region. What does not transfer automatically is the specific tensor, coefficient or thermodynamic derivation. The NIST analysis treats density and composition as different order parameters with different evolution laws, while Chen's miscible model retains phenomenological uncertainty.[1][3]

Outside continuum fluids, gradients can produce stresses or costs in other theories, but using the name Korteweg Stress literally requires the fluid capillarity constitution described here. A general relationship between gradient fields and distributed response might be a future-prime question; no currently inspected live prime has been asserted as a strict parent of this named entry.[1]

Examples

Liquid–vapor diffuse interface in a specified one-component model. Hosseini, Dorschner and Karlin begin with a non-ideal fluid whose isothermal free energy contains \(\frac{1}{2}\kappa|\nabla\rho|^2\) for constant capillarity coefficient \(\kappa\). Their §2.1 derives a Korteweg tensor with isotropic gradient/Laplacian and anisotropic gradient-product contributions and sets a Navier–Stokes–Korteweg target for two-phase dynamics. A droplet's liquid–vapor transition is a finite density profile, not merely a zero-thickness jump. This is one specified model, not the tensor of every fluid.[2]

Mapped back: The field is density \(\rho\); gradient dependence follows from the stated square-gradient free energy; its capillary tensor contributes to momentum balance; the transition is the finite liquid–vapor zone; and isothermal/constant-\(\kappa\) and equation-of-state choices bound the example.

Miscible droplet displacement. Chen, Wang and Meiburg model a drop displaced through another miscible liquid in a Hele–Shaw/porous-medium setting. They add a postulated concentration-gradient stress to their flow equations and compare simulated fronts with and without it. Stronger assumed gradient effects change fingering and shape in the modeled cases, but the original paper notes that coefficient magnitude/sign and the analogy with a permanent immiscible interface remain uncertain.[3]

Mapped back: The field is composition across a transient mixing layer; gradient dependence is posited rather than copied from the liquid–vapor derivation; its stress divergence affects modeled flow; the layer is finite and diffusing; and phenomenological coefficient uncertainty is a condition of the result, not an omitted detail.

Structural Tensions

Sharp-interface economy versus diffuse-zone fidelity. A zero-thickness boundary with surface-tension jump is compact when transition thickness is negligible. It cannot at the same time resolve internal interfacial gradients or topology changes without further treatment. Distributing stress through a finite zone can represent those features, but it demands constitutive inputs and resolved gradients. Diagnostic: Is the physical question controlled by a zone whose thickness/gradient structure matters, or is a sharp traction jump sufficient at the modeled scale?[1]

Constitutive fidelity versus portable phenomenology. A stress derived from a specified single-component free energy has clearer thermodynamic constraints, but those assumptions do not automatically apply to a transient miscible mixing layer. A flexible composition-gradient term can explore miscible-flow behavior, but its uncertain coefficients weaken prediction and causal identification. One cannot claim both unrestricted cross-fluid portability and model-specific derivational warrant from the same formula. Diagnostic: For this fluid and field, is the term derived or empirically constrained, or is it being borrowed by analogy with unverified coefficients?[1][3]

Structural–Framed Character

Korteweg Stress sits toward the structural side within its fluid-mechanical domain, but the named identity remains domain-specific. Its gradient-to-stress mapping is mathematical; nevertheless its variables are fluid density or composition, capillary constitutive behavior and momentum balance. Evaluative weight is low: the term does not itself label a flow good or bad; usefulness depends on the modeling question. Human-practice dependence is moderate: the physical relation is not created by a convention, but its decomposition, order parameter, coefficient choice and validation are modeling practices. Institutional origin is a scientific research lineage, not a regulatory definition. Vocabulary travel of “stress” and “gradient” is broad, whereas the exact Korteweg name does not retain its meaning outside capillary continuum fluids. Import versus recognition therefore matters: seeing a gradient response in another field is an analogy; recognizing this stress requires its fluid constitutive and momentum roles.[1][2]

Its character: a domain-specific constitutive-stress abstraction with a potentially portable gradient-response skeleton. The skeleton is not promoted to a prime here because the existing catalog has no inspected necessary genus matching it, and mathematical resemblance alone would erase the fluid-specific physics.[1]

Structural Core vs. Domain Accent

At the most portable level, a nonuniform field contributes a structured response to a local balance law. That skeleton is an unadmitted future-prime question, not an already verified live parent edge. The complete Korteweg identity needs a fluid order parameter, capillary/gradient constitutive stress and mechanical coupling through momentum balance. In the liquid–vapor instance, a free-energy density gradient supplies one derivation; in the miscible instance, a composition-gradient term may be phenomenological and transient.[1][3]

Removing the fluid and interfacial physics leaves only a generic gradient-to-response analogy. Conversely, calling a classical velocity-gradient viscosity tensor Korteweg because both are stresses drops the capillary mechanism. Those boundaries keep the named node domain-specific while allowing later graph curation to consider a suitable higher-order stress or gradient-response parent if one is genuinely established.[1]

No strict typed parent relation is asserted in the current DAG.

Neighborhood in Abstraction Space

Korteweg Stress sits in a sparse region of the domain-specific corpus (64th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Domain-Specific Measurement Parameters (36 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • Newtonian viscous stress: depends on deformation-rate/velocity gradients and models dissipation; Korteweg contribution depends on density/composition structure and capillarity.[1]
  • Sharp-interface surface tension: a boundary traction jump at a geometric interface; related in appropriate limits, but not identical to distributed gradient stress.[1]
  • A concentration gradient alone: it supplies an input, not a stress without a constitutive relation and momentum coupling.[3]
  • A universal constant-\(\kappa\) formula: single-component, isothermal assumptions and tensor conventions cannot silently be exported to mixtures.[2][1]
  • Permanent equilibrium tension between fully miscible liquids: the original miscible studies discuss transient or effective interfacial effects with uncertain parameters and imperfect analogy.[4][3]

References

[1] D. M. Anderson, G. B. McFadden and A. A. Wheeler, Diffuse-Interface Methods in Fluid Mechanics, NISTIR 6018 (1997), original-author institutional report, Introduction; §§2–4 on stress, density/composition fields, sharp-interface limit and miscible flows. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u ↩v ↩w ↩x ↩y

[2] S. A. Hosseini, B. Dorschner and I. V. Karlin, “Towards a Consistent Lattice Boltzmann Model for Two-Phase Fluids”, Journal of Fluid Mechanics 953:A4 (2022), original article §§1–2.1, especially equations (2.1)–(2.7). registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[3] Ching-Yao Chen, Lilin Wang and Eckart Meiburg, “Miscible Droplets in a Porous Medium and the Effects of Korteweg Stresses”, Physics of Fluids 13:2447–2456 (2001), original author-hosted article, Introduction, §§II–IV and Conclusions. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n

[4] Ilya Kostin, Martine Marion, Rozenn Texier-Picard and Vitaly A. Volpert, “Modelling of Miscible Liquids with the Korteweg Stress”, ESAIM: Mathematical Modelling and Numerical Analysis 37:741–753 (2003), original publisher abstract; full article not fetched here. registry ↩a ↩b