Navier–Stokes Equations in and on a Riemannian Manifold
You have watched cream swirl into coffee for longer than you expect. You have seen a river smooth on the surface but violent underneath at a bridge pier. You have felt an airplane shake in turbulence and wondered why it is so hard to predict. The same two equations run in each: one for mass and one for momentum. Together they are the Navier–Stokes equations.
Two facts from freshman physics (mass is conserved, force equals mass times acceleration) applied to a moving blob of fluid produce the equations in a few lines. The open question concerns the solutions. Give the fluid smooth initial conditions: does it stay smooth forever or can something catastrophic happen in finite time? In two dimensions smooth initial data produces smooth solutions for all time. In three dimensions nobody knows. The gap has been open since regularity theory was first studied rigorously in the 1930s.[1]
The Clay Mathematics Institute listed it as one of seven Millennium Prize Problems in 2000 with a million-dollar prize for a proof either way.[2] The difficulty has a specific name: vortex stretching. In three dimensions rotating structures in the flow can stretch and amplify themselves because the velocity field that drives the stretching is determined nonlocally by the vorticity throughout the domain via the Biot–Savart law. A small concentration of vorticity anywhere affects the whole flow. A regularity proof would have to control that feedback loop and no one has managed it.
This post builds the equations from first principles and then examines vortex stretching from three vantage points. Differential forms on a Riemannian manifold show what happens to viscous diffusion when space is curved. Functional analysis states the regularity gap in terms of Sobolev spaces and the Leray energy inequality. Geometric algebra encodes vortex stretching as a grade-2 source term in a compact multivector evolution equation for the combined velocity-vorticity state. Together they show where the 3D proof stalls and why the 2D proof goes through.
Interactive: drag the Reynolds number slider to watch the flow structure change.
1. Continuum Mechanics Axioms
The Navier–Stokes equations rest on two conservation laws. Before writing them as differential equations we fix the continuum setting and the notation for control volumes and material derivatives.
The fluid is modelled as a continuous medium: at every point in the domain, quantities such as density , velocity , and pressure are smooth functions. Molecular structure below the mean free path is ignored.
For a scalar field , the material derivative following a fluid parcel with velocity is
This measures the rate of change of as seen by a parcel moving with the flow.
A control volume is a fixed region with smooth boundary and outward unit normal . For a quantity with density and flux , the net flux out of is
For any smooth scalar field and material volume moving with a velocity field :
Proof. Let be the flow map with and . Pull back to the reference domain:
where is the Jacobian. Differentiating under the integral:
The Jacobi formula gives (Euler's expansion formula). Pulling back to :
where the last equality uses the product rule . Equivalently the boundary term arises directly from the Leibniz rule for moving domains. The divergence theorem then converts it to .
The stress tensor is a symmetric -tensor encoding internal forces: it maps a unit normal vector to the traction force per unit area . For a Newtonian fluid decomposes into an isotropic pressure part and a viscous part.
For an incompressible fluid ():
Proof. Mass conservation states that the total mass in any material volume is constant: . Applying [reynolds-transport] with :
Since this holds for every and the integrand is continuous, the localisation lemma gives the pointwise continuity equation:
Expanding using the product rule: . For an incompressible fluid is constant in space and time. Hence and and so . Since we conclude .
For an incompressible fluid with body force per unit volume:
Proof. Newton's second law for a material volume states that the rate of change of momentum equals the net force (surface tractions plus body forces). Apply [reynolds-transport] to the momentum density (component by component):
The surface traction force is , which by the divergence theorem equals . Newton's law therefore gives:
Localise: . For the left side, expand using constant and :
Hence .
The Cauchy momentum equation is the general form of Newton's second law for continua. The divergence remains unexpanded. Expanding it requires a constitutive law relating to the flow kinematics, which the next section supplies.
2. Constitutive Law and Navier–Stokes in
Closing the Cauchy momentum equation requires specifying how the stress tensor depends on the velocity field . For a Newtonian fluid this dependence is linear in the symmetric part of the velocity gradient. Substituting into [cauchy-momentum] then yields the Navier–Stokes equations.
The strain rate tensor measures the symmetric part of the velocity gradient:
Its eigenvalues give the principal rates of extension and compression of fluid parcels.
A fluid is Newtonian if the viscous stress is linear in the strain rate:
where is the dynamic viscosity and is the second viscosity coefficient.
The Stokes hypothesis sets . This gives zero bulk viscosity. Under this assumption and incompressibility ():
This is a widely used approximation and is exact for monatomic ideal gases at equilibrium.
Proof. Step 1 (expand div T). Substitute with :
Step 2 (vector identity). For any smooth vector field, the identity
Componentwise: . Under incompressibility , the second term vanishes:
Step 3 (substitution into Cauchy). Therefore . Inserting into [cauchy-momentum] [eq:cauchy-eq]:
Dividing both sides by , setting and writing again for the force per unit mass yields the Navier–Stokes momentum equation.[5]
For characteristic velocity , length scale , and kinematic viscosity :
measures the ratio of inertial forces to viscous forces. Flow transitions from laminar to turbulent as increases.
Non-dimensionalising with velocity scale and length scale gives:
As the viscous term vanishes and the equation approaches the Euler equations for inviscid flow.
3. Geometric Reformulation of the Flat Case
This section rewrites the R³ Navier–Stokes equations in the language of differential forms. The geometric structure lifts directly to an arbitrary Riemannian manifold. No new physics appears. Index notation enters for k-form components and boldface is retained for vector fields.
Let denote Euclidean space with the standard metric . The flat and sharp maps are the metric isomorphisms between vector fields and 1-forms:
Worked example. For in Cartesian coordinates:
The vorticity 2-form is then:
Also , where is the codifferential on k-forms in dimension n.
The exterior derivative satisfies . The codifferential is its formal adjoint with respect to the inner product on forms: , where is the Hodge star. On a Riemannian manifold is independent of the metric and depends on it through .
The vorticity 2-form associated to a velocity field is . Its components are the classical vorticity vector components: . The Hodge dual recovers the classical curl vector field. The map is nonlocal: Section 4 recovers from by convolution with the Biot–Savart kernel.
(i.e., is coclosed).
Proof. By the definition of the codifferential on 1-forms in , applying the Euclidean Hodge star:
In Cartesian coordinates, . Then . Applying again: . Hence .
The Hodge Laplacian (or Laplace–de Rham operator) on is . For a coclosed 1-form (i.e., ), this reduces to .
Sign convention. This post uses the physicists' Laplacian (the convention of Section 2). On flat with the Euclidean metric, for a coclosed 1-form , a direct computation in Cartesian coordinates gives applied component-wise, whence . Equivalently: .
The nonlinear advection term arises from the Lie derivative of along . By Cartan's magic formula:
Here . Hence the last term is . A direct computation in coordinates (the Lamb identity) gives the advection 1-form:
The term is a gradient and is absorbed into the modified pressure . All subsequent equations use and take when they write . See the manifolds post on vector fields and flows for background on the Lie derivative.
The incompressible Navier–Stokes equations in are equivalent to:
where is the modified pressure and (see [hodge-laplacian]).
Proof. Apply the flat isomorphism to the NS momentum equation from [navier-stokes-r3], term by term.
Time derivative: since acts fiberwise by the Euclidean metric (which is time-independent in ).
Pressure term: For any smooth function , the musical isomorphism satisfies , since in Cartesian coordinates .
Advection term: From the Lamb identity [eq:lamb-eq] ([lie-derivative-advection]):
Absorb the exact term into the modified pressure . The advection contribution is then .
Viscous term: By the Hodge Laplacian identity ([hodge-laplacian]), on flat for a coclosed 1-form: . Hence .
Combination: Applying to [eq:ns-momentum] ([navier-stokes-r3]) and substituting:
The incompressibility constraint is equivalent to by [incompressibility-coclosed].
Setting and in the forms-NS gives the incompressible Euler equations:
By Arnold (1966)[6] these are the geodesic equations on the Lie group of volume-preserving diffeomorphisms of the fluid domain, equipped with the right-invariant metric. Arnold related the negative sectional curvature of this group in many directions to the practical unpredictability of the flow.
4. Biot–Savart and the Nonlocal Structure of Vorticity
In Section 3 introduced the vorticity 2-form . The NS equations show that drives the nonlinear term. Knowing everywhere in the domain determines . The relationship is nonlocal: a vortex ring on the other side of the domain still moves the fluid near you. This is the Biot–Savart law, and it makes the Navier–Stokes equations much harder in three dimensions than in two. The same integral kernel gives the magnetic field of a current distribution in electromagnetism (see Section 1 of the Maxwell post).
The Biot–Savart kernel is the vector-valued function :
It decays like and is the negative gradient of the Newtonian potential : one has since . The operator associating a velocity field to a vorticity field via convolution with is the Biot–Savart operator .
Let be divergence-free with vorticity compactly supported. Then:
Proof. Step 1 (vector potential). Since , the Helmholtz decomposition on with rapid decay gives a vector field such that . Impose the Coulomb gauge (this fixes the gauge freedom in uniquely up to harmonic fields, which vanish for compactly supported data). Under this gauge, the vector identity gives:
so satisfies component-wise.
Step 2 (consistency of the Coulomb gauge). The ansatz solves by the fundamental solution of the Laplacian. Verify the gauge: since , differentiating under the integral and integrating by parts (boundary term vanishes by compact support of ):
using .
Step 3 (recovery of u). Compute by differentiating under the integral. For a constant vector , the identity gives:
Integrating gives the Biot–Savart formula.
In 2D incompressible flow with scalar vorticity , the velocity is:
where . The kernel now decays like , slower than in 3D.
The Biot–Savart formula makes the Navier–Stokes equations an integro-differential system: the velocity at depends on the vorticity at every point in the domain. Three consequences bear directly on the Millennium Prize:
(i) Nonlocal driving of vortex stretching. The stretching term involves , which by Biot–Savart depends on all of . A concentration of vorticity anywhere in the fluid can strain and amplify vortex tubes far away.
(ii) BKM criterion in this language. The Beale–Kato–Majda criterion (see [bkm-criterion]) requires for blowup. Biot–Savart and the Calderón–Zygmund inequality give for . The inequality fails at . Beale, Kato and Majda replace it with a logarithmic estimate that bounds by times the logarithm of a higher Sobolev norm.
(iii) Planar flow. In 2D is a scalar. The vorticity equation reduces to advection-diffusion with no stretching term. The maximum principle then gives a uniform bound on for all time and the 2D Biot–Savart kernel converts it into control of . In 3D no such bound is available.
5. Navier–Stokes on a Riemannian Manifold
Each flat object in Section 3 has a Riemannian counterpart. This section lifts the forms-NS to an arbitrary closed oriented Riemannian manifold by substituting the Levi-Civita connection for the flat derivative, the Riemannian Hodge star for the Euclidean one and the Bochner Laplacian for the Hodge Laplacian. The resulting equation has a Ricci curvature correction.
On , the metric defines the flat and sharp maps between the tangent and cotangent bundles. For a vector field , the associated 1-form is:
The sharp map is its inverse. In local coordinates: , . See the Riemannian metric definition for context.
The covariant material derivative of a vector field along itself, using the Levi-Civita connection , is:
In flat with the standard connection this reduces to from Section 3.
The divergence of a vector field on is defined by , where is the codifferential with respect to the Riemannian metric. Incompressibility becomes:
The Bochner Laplacian (or connection Laplacian) on 1-forms is the operator:
where the trace is over the two connection indices. On flat this reduces to minus the component-wise scalar Laplacian, . See the Bochner theorem in the manifolds post.
On a Riemannian manifold , the Hodge Laplacian and the Bochner Laplacian are related by:
where acts on 1-forms by . Equivalently, for the corresponding vector field:
On flat , and hence . With the sign of [hodge-laplacian] this is the standard identity of Bochner (1946)[10] and Taylor Vol. III.[8] See the Riemann curvature tensor for the definition of Ric.
On the pressure gradient is , where is the exterior derivative and uses the Riemannian metric. In local coordinates . For the incompressible equations, satisfies a Poisson equation determined by the constraint .
For an incompressible Newtonian fluid on a closed oriented Riemannian manifold , with kinematic viscosity , modified pressure , and body force :
Proof. [eq:ns-forms-eq] from [ns-forms] with Riemannian replacements: the exterior codifferential becomes , the Hodge Laplacian becomes , and the pressure gradient becomes with applied. Applying [eq:weitzenbock-eq] ([weitzenbock]): . Applying throughout gives the vector-field form above.
The term in the manifold NS equation is a curvature-induced body force. When (positive Ricci curvature, as on the round sphere) it adds to the viscous diffusion and increases damping. When (negative Ricci curvature, as on a compact hyperbolic manifold) it reduces the effective diffusion. This sign belongs to the Hodge lift. Lifting the stress of [stokes-hypothesis] directly gives the viscous term for divergence-free (Ebin and Marsden 1970; Taylor[8]), with the Ricci term of the opposite sign. Chan, Czubak and Disconzi (2017) argue that this deformation-tensor form is the physical one. On flat , , and the covariant NS reduces exactly to the equation in [navier-stokes-r3].
The two canonical settings for the incompressible Navier–Stokes equations are:
(a) Flat space . With , and :
(b) Riemannian manifold . The curvature of enters through the Weitzenböck identity as a Ricci correction to the viscous term:
Case (a) is the formal specialisation of (b) to on the non-compact space .
Interactive: toggle between Euclidean and Levi-Civita transport around a closed loop. Drag to rotate. The gold arrow (Levi-Civita) rotates by π on the latitude circle; the white initial vector is shown for reference.
6. Millennium Prize Problem
In 2000 the Clay Mathematics Institute listed the existence and smoothness of solutions to the Navier–Stokes equations in as one of the seven Millennium Prize Problems, each with a $1,000,000 prize.[2] The question has two parts: do global smooth solutions always exist, and if not, what does a finite-time singularity look like?
A Leray weak solution to the incompressible NS on with initial data is a vector field satisfying:
- The NS equations hold in the distributional sense against all divergence-free test functions
- The energy inequality:
Leray (1934)[1] proved that such solutions exist globally for any . The Millennium Prize problem concerns whether these solutions are smooth.
A classical solution in the sense of the Clay problem is a function satisfying the NS equations pointwise with bounded energy, for all . Rapid decay is imposed on the initial data: for all multi-indices and all .
Let be divergence-free and . Then the 2D incompressible NS has a unique global solution. It is smooth for .
Proof. Step 1 (local solution). For with the standard local existence theory gives a unique solution on , smooth for . In 2D the space is subcritical and the local existence time is bounded below by a positive function of alone.
Step 2 (vorticity equation). In 2D the vorticity is the scalar . Taking the curl of the 2D NS equations yields:
There is no vortex stretching term: in 2D since is independent of .
Step 3 ( bound on ). Equation [eq:vorticity-2d] with divergence-free advection field satisfies the maximum principle for parabolic equations. Hence if also then for all :
Step 4 ( bound on ). From the Leray energy inequality . The vorticity satisfies and hence . Multiply the vorticity equation by and integrate:
(the advection term vanishes by skew-symmetry of ). Hence for all and uniformly in time.
Step 5 (global continuation). Suppose . By Step 4 on . By Step 1 the solution started at any exists for a time that depends only on . Taking extends it past . This contradicts the maximality of . Hence . Smoothness for is the parabolic smoothing of the local theory.
For with and divergence-free, there exists and a unique smooth solution to the NS equations. The maximal existence time satisfies for a constant depending only on . The Millennium Prize question is whether .
(The threshold suffices by the Sobolev embedding when in . Then the velocity field is Lipschitz and classical solutions are well-defined. See Taylor Vol. III[8] for the precise statement.)
The Beale–Kato–Majda criterion (1984)[7]: the smooth solution from [local-3d] fails to extend beyond time (i.e., ) if and only if:
where is the vorticity. Finite-time blowup requires the vorticity to blow up in .
The Clay Mathematics Institute problem, as formulated by Fefferman (2006)[2] lists four statements:
(A) For any smooth divergence-free initial data with rapid decay and zero force, the NS equations in have a smooth global solution with bounded energy.
(B) The same on the periodic domain .
(C) There exist smooth rapidly decaying initial data and a smooth rapidly decaying force on for which no smooth global solution with bounded energy exists.
(D) The same on .
A proof of any one of the four qualifies for the prize. The difficulty of (A) is that the energy inequality gives regularity in , but closing to requires control of the nonlinear term that the energy alone cannot provide in three dimensions.
7. Functional Analysis
The precise formulation of the Millennium Prize problem requires function spaces that capture the energy structure of the equations. This section introduces the Sobolev space framework and the Galerkin method that underlies Leray's existence proof.
For a domain and integer , the Sobolev space is:
with inner product . The norm controls up to weak derivatives in .
The natural function space for incompressible NS on a bounded domain is:
with inner product . The Leray projection is the orthogonal projection onto the closure of .
Multiplying the NS equation by a test function and integrating by parts gives the weak form: find such that for all :
where the trilinear form is .
For divergence-free and : .
Proof. By integration by parts:
using div and the homogeneous Dirichlet boundary condition.
Let be an orthonormal basis of Stokes eigenfunctions for . The -th Galerkin approximation is the projection onto the -dimensional subspace, satisfying the projected NS system for each coefficient .
For any and , the Galerkin approximations satisfy the uniform energy bound:
By the Banach–Alaoglu theorem a subsequence converges weakly in and weak-* in . A bound on in a dual space and the Aubin–Lions lemma give strong convergence in , enough to pass to the limit in the nonlinear term. The limit is a Leray weak solution. Leray (1934)[1] proved existence by mollifying the advecting velocity. The Galerkin construction is due to Hopf (1951) and was developed by Ladyzhenskaya (1969)[3] and Temam.[4]
The Stokes operator is a positive self-adjoint operator on , where is the Leray projection. Its eigenfunctions are the Stokes eigenfunctions with eigenvalues . By the Weyl law in dimension 3:
where depends on . The fractional powers define the domains , with and .
The energy estimate [eq:energy-ineq] gives . To improve to requires estimating the trilinear form in terms of the energy. In 2D this is possible via the Ladyzhenskaya inequality . In 3D the analogous inequality is and the resulting estimate does not close. This leaves a gap between what the energy controls and what regularity needs. The nonlocal structure of Biot–Savart ([biot-savart-formula]) produces this gap: controlling via the Calderón–Zygmund theorem requires , but the energy gives only and BKM requires control of .
8. Geometric Algebra
Clifford algebra gives an algebraic language for the Navier–Stokes equations. This section derives a single equation for the fluid multivector that contains both the momentum and the vorticity equations. Vortex stretching appears as a grade-2 source term present in 3D and absent in 2D.
The Clifford algebra is the associative algebra generated by subject to the relations .[9] It has basis (grades 0 through 3). The pseudoscalar is , and . The grade-k projection is denoted .
The geometric derivative of a vector field (identified with a grade-1 element of ) is:
The scalar part is grade-0; the bivector part encodes the vorticity as a grade-2 element. Under incompressibility , the geometric derivative is a pure bivector: .
The fluid multivector combines velocity and vorticity into a single Clifford element:
where is the grade-1 velocity and is the vorticity encoded as a grade-2 bivector. The Navier–Stokes and vorticity equations can both be extracted from a single equation for .
The Lamb identity expresses the advection term via the Clifford product:
Since is central and , the grade-1 part of is . The gradient term is again absorbed into the modified pressure .
The incompressible NS equations in are equivalent to:
Proof. From [navier-stokes-r3], substitute the Lamb identity ([rotor-advection]) for the advection term: . Absorbing into gives the result. The geometric algebra form makes explicit that advection acts on the velocity through the vorticity cross product.
Applying to the NS-GA equation gives the vorticity transport equation in geometric algebra form:
The term is vortex stretching: it amplifies vorticity when the flow strains vortex tubes in the direction of the vorticity vector.
The fluid multivector satisfies the compact evolution equation:
The grade-1 part is the NS momentum equation. The grade-2 part is the vorticity transport equation:
The term is a grade-2 source: its contribution to the growth of enstrophy is . In 2D this source vanishes and the grade-2 part satisfies homogeneous advection–diffusion.
Proof. The grade-1 part is the NS-GA equation from [ns-geometric-algebra] rewritten as . For the grade-2 part, apply (a constant) to [eq:vorticity-3d] ([vorticity-equation-ga]): . Combining both grades gives the single -equation. In 2D and is independent of and therefore .
The -equation shows the structural difference between 2D and 3D at the level of algebra. In 2D the source and the grade-2 component satisfies : vorticity is advected and diffused with no amplification. The maximum principle then gives a uniform bound on for all time and closes the regularity argument. In 3D is a nonzero grade-2 input that can drive unbounded growth of . By [eq:biot-savart-eq] ([biot-savart-formula]) the velocity gradient in the source is determined nonlocally by all of . The grade-2 source is therefore self-amplifying: a vorticity concentration anywhere in the domain feeds back into its own growth.
The BKM criterion [eq:bkm-eq] ([bkm-criterion]) states that blowup occurs iff . In the -evolution equation from [ns-multivector], this corresponds to the grade-2 source driving the norm of the grade-2 component to blow up. In 2D the source vanishes, the grade-2 component is bounded for all time and the 2D analogue of the Millennium Prize question is settled. In 3D the source is generically nonzero and the Millennium Prize asks whether it can accumulate enough to blow up.
9. Navier–Stokes on a Deforming Surface
The Riemannian setting of Section 5 assumes the geometry is fixed. Many fluid systems violate this: in liquid–liquid phase separation (LLPS), a fluid interface separates two phases and deforms as the bulk fluids move. It has its own surface tension, surface viscosity and bending energy. Protein condensates (stress granules, P-bodies, nucleoli), lipid membrane domains (liquid-ordered versus liquid-disordered) and immiscible drops all have a moving interface of this kind.
The central new ingredient is that the domain changes with time, which Section 5 does not treat. The material derivative of a tangent vector on the surface acquires an extra term proportional to the normal velocity times the shape operator, because normal motion of a curved surface stretches the tangent basis. The metric, the curvature and all derived operators depend on . The incompressibility condition must be revised: a tangential flow on a curved deforming surface can have nonzero surface divergence provided the surface itself expands normally at the matching rate.
A smooth evolving hypersurface is a family where each is a smooth closed embedded -manifold in , described locally by a smooth embedding . The induced metric is , the unit outward normal is , and the second fundamental form is (Gauss formula: ). The mean curvature is . The Weingarten equation gives .
The velocity of a surface point decomposes as:
where is the tangential velocity and is the normal velocity. The shape of as a set in depends only on ; changing only reparametrises the surface without altering its geometry. Throughout Section 9 the superscript T denotes tangential and matrix transpose is written .
For a scalar field (with following the flow), the surface material derivative is . For a tangent covector field , the covariant surface material derivative is:
Here is the tangential projection of the time derivative of along a material point, and the formula follows from with the Weingarten equation. The correction is absent for static manifolds. It comes from : normal motion of a curved surface stretches the coordinate basis.
For a smooth evolving hypersurface with velocity , the induced metric evolves as:
where is the surface rate-of-strain tensor. The first term couples the metric to the normal velocity via the second fundamental form and the second couples it to tangential shearing.
Proof. Differentiate at fixed :
Substitute . Normal contribution: by the Weingarten equation and :
and symmetrically . Tangential contribution: expanding via the Gauss formula and collecting Christoffel terms gives . Summing gives the result.
The area element evolves as:
For a surface-incompressible fluid (local area preservation along the flow):
Where (a plane or a minimal surface) or on a stationary surface () this reduces to .
Proof. Take the metric trace: . Since and , the result follows.
For a Newtonian surface fluid with surface shear viscosity , surface dilatational viscosity , and surface tension , the Scriven–Boussinesq surface stress tensor is:[11]
where is the trace-free surface strain rate. The three terms are respectively isotropic surface tension, deviatoric surface viscous stress and resistance to surface dilation.
For a surface-incompressible fluid on with surface mass density , the tangential momentum equation is:
where is the Gaussian curvature of , is the tangential stress jump from the surrounding bulk fluids, and is a surface body force.
Proof. Apply the Reynolds Transport Theorem on the moving surface to the momentum density on a moving patch , equate to the surface stress divergence plus bulk tractions. Expanding using the surface Bochner formula (the 2D Weitzenböck identity): on a 2D surface, . By the Weitzenböck identity from Section 5:
The Gaussian curvature term is the 2D instance of the Ricci term . With it has the sign of the deformation-tensor lift in [ricci-correction], opposite to the Hodge lift of [ns-manifold]. For a static surface (, , uniform) the equation reduces to the deformation-tensor form of the manifold NS with .
The Helfrich Hamiltonian for a fluid membrane is:[12]
where is the bending rigidity, is the spontaneous curvature (nonzero for asymmetric bilayers), and is the Gaussian bending modulus. By the Gauss–Bonnet theorem, is a topological invariant under smooth deformations and therefore affects only topology-changing events such as membrane fusion or fission. With as in [evolving-hypersurface] a normal displacement gives and . The first variation of is therefore
and for the Willmore bending force density is . The factor comes from the averaged mean curvature. With the summed convention and energy density the same force reads . Check on a cylinder of radius and length : there and give . Integrating the variation with over the area gives .
At the interface between bulk domains , momentum balance across the zero-thickness interface gives two conditions. Normal balance (generalised Young–Laplace with bending):
Tangential balance (Marangoni and surface viscosity):
The term drives Marangoni flow whenever surface tension is nonuniform.
Proof.
Step 1 (Thin pillbox). Surround by a pillbox of cross-sectional area and height . Integrate the Cauchy momentum equation over and send . Since the interface is massless (surface mass density times vanishes), the inertial term drops out, and the force balance gives:
The surface stress term comes from the lateral faces of via the surface divergence theorem. The bending force is the first variation of the Helfrich energy ([helfrich-energy]) under a normal displacement field.
Step 2 (Surface divergence decomposition). For any tangential symmetric tensor , the Gauss–Weingarten equations give:
The normal term is normal leakage: surface stress transfers momentum to the bulk via curvature.
Step 3 (Normal balance). Taking the -component of the jump condition from Step 1 and using Step 2:
The leading term: gives the Young–Laplace pressure (for a sphere of radius with outward normal and ), and the Helfrich bending correction adds the Willmore force density.
Step 4 (Tangential balance). Taking the -component of the jump condition and using Step 2:
Expanding term by term: (i) ; (ii) for the viscous deviatoric term, the 2D surface Weitzenböck identity ( on any 2D surface; cf. Section 5) gives:
and (iii) the dilatational term gives . Summing all three:
Therefore .
Three LLPS contexts illustrate the equations above.
(i) Lipid bilayer membrane domains. The liquid-ordered (Lo, cholesterol-rich) and liquid-disordered (Ld) phases have different surface viscosities . At the Lo/Ld domain boundary, a line tension (energy per unit length of the Lo/Ld contact line, analogous to surface tension but for a 1D boundary on the 2D surface) enters the tangential balance as a 1D Laplace pressure term , where is the geodesic curvature of the domain boundary (the rate at which the contact line bends within the surface , measured by the Levi-Civita connection of ). The Saffman–Delbrück length sets the crossover between 2D and 3D hydrodynamic regimes: proteins smaller than diffuse as if in a 2D fluid; larger proteins or domains couple strongly to the bulk.
(ii) Biomolecular condensates. The observation that germline P granules in C. elegans behave as liquid droplets that dissolve and condense through controlled phase separation established the physical basis of membraneless organelle formation.[13] Subsequent work by Hyman, Weber and Jülicher (MPI-CBG and MPI-PKS, Dresden) placed LLPS of intrinsically disordered proteins within a coherent framework of biological phase transitions.[14] Sabari et al. in the Young lab (MIT/Whitehead) then showed that transcriptional coactivators form condensates at super-enhancers. This links LLPS directly to gene regulation.[15] Alberti, Gladfelter and Mittag collect broad reviews of the physical principles, challenges and biological roles of condensates.[16] A condensate interface is a liquid–liquid boundary with no bilayer, and it is modelled with surface tension alone (). The viscosity contrast is large. Droplets of the P-granule protein LAF-1 have Pa s at physiological salt, more than times the viscosity of water.[17] Spatially varying protein concentration along the interface induces surface tension gradients . These drive Marangoni flows (the term in the tangential balance) that advect condensate material along the surface and can accelerate or slow coalescence.
(iii) Classical immiscible drops. A viscous drop of internal viscosity falling through an ambient fluid of viscosity is governed exactly by the bulk–surface coupling of [bulk-surface-coupling] with . At low Reynolds number the system reduces to two Stokes problems coupled at . The classical Hadamard–Rybczyński solution for the terminal velocity is:
As (rigid sphere limit) , the Stokes drag value. As (inviscid bubble) . This is times the Stokes value because internal circulation within the drop reduces the effective drag below that of a rigid sphere.
10. Conclusion
Differential forms (Section 3) show that incompressibility is co-closedness of and that the Euler equations are geodesic equations on an infinite-dimensional Lie group. The Riemannian lift (Section 5) shows that curved space modifies viscous diffusion by Ricci curvature: in the Hodge lift the Weitzenböck identity replaces the flat Laplacian with . There positive Ricci curvature adds damping and negative Ricci curvature removes it. The deformation-tensor lift reverses the sign of the Ricci term. Geometric algebra (Section 8) contains the same information in a single multivector equation: momentum is the grade-1 part, vorticity the grade-2 part, and vortex stretching is the source that is identically zero in 2D and generically nonzero in 3D.
Section 9 applies the same structure to an evolving surface , a 2D Riemannian manifold with a time-dependent metric. Since on any 2D surface the Ricci correction from Section 5 reappears in the surface NS as the Gaussian curvature term . The revised incompressibility condition is a theorem: it follows directly from the metric evolution equation when area is preserved along the flow. The bulk-surface stress jump then feeds the surface dynamics back into the bulk NS equations through the same Weitzenböck structure that first appeared in Section 5.
The Millennium Prize remains open because no known technique controls the grade-2 source long enough to rule out finite-time blowup, and no known technique constructs initial data that provably forces it.
References
- Leray, J. (1934). Sur le mouvement d'un liquide visqueux emplissant l'espace. Acta Mathematica, 63, 193–248. DOI
- Fefferman, C. L. (2006). Existence and Smoothness of the Navier–Stokes Equation. Clay Mathematics Institute Millennium Problems. PDF
- Ladyzhenskaya, O. A. (1969). The Mathematical Theory of Viscous Incompressible Flow. Gordon and Breach.
- Temam, R. (1977). Navier–Stokes Equations: Theory and Numerical Analysis. North-Holland.
- Chorin, A. J. and Marsden, J. E. (1993). A Mathematical Introduction to Fluid Mechanics. Springer.
- Arnold, V. I. (1966). Sur la géométrie différentielle des groupes de Lie de dimension infinie et ses applications à l'hydrodynamique des fluides parfaits. Annales de l'Institut Fourier, 16(1), 319–361. DOI
- Beale, J. T., Kato, T., and Majda, A. (1984). Remarks on the breakdown of smooth solutions for the 3D Euler equations. Communications in Mathematical Physics, 94(1), 61–66. DOI
- Taylor, M. E. (1996). Partial Differential Equations III: Nonlinear Equations. Springer.
- Hestenes, D. and Sobczyk, G. (1984). Clifford Algebra to Geometric Calculus. Reidel.
- Bochner, S. (1946). Vector fields and Ricci curvature. Bulletin of the American Mathematical Society, 52(9), 776–797. DOI
- Scriven, L. E. (1960). Dynamics of a fluid interface: equation of motion for Newtonian surface fluids. Chemical Engineering Science, 12(2), 98–108. DOI
- Helfrich, W. (1973). Elastic properties of lipid bilayers: theory and possible experiments. Zeitschrift für Naturforschung C, 28(11–12), 693–703. DOI
- Brangwynne, C. P., Eckmann, C. R., Courson, D. S., Rybarska, A., Hoege, C., Gharakhani, J., Jülicher, F., and Hyman, A. A. (2009). Germline P granules are liquid droplets that localize by controlled dissolution/condensation. Science, 324(5935), 1729–1732. DOI
- Hyman, A. A., Weber, C. A., and Jülicher, F. (2014). Liquid-liquid phase separation in biology. Annual Review of Cell and Developmental Biology, 30, 39–58. DOI
- Sabari, B. R., Dall'Agnese, A., Boija, A., Klein, I. A., Coffey, E. L., Shrinivas, K., Abraham, B. J., Hannett, N. M., Zamudio, A. V., Manteiga, J. C., Li, C. H., Guo, Y. E., Day, D. S., Schuijers, J., Vasile, E., Malik, S., Hnisz, D., Lee, T. I., Cisse, I. I., Roeder, R. G., Sharp, P. A., Chakraborty, A. K., and Young, R. A. (2018). Coactivator condensation at super-enhancers links phase separation and gene control. Science, 361(6400), eaar3958. DOI
- Alberti, S., Gladfelter, A., and Mittag, T. (2019). Considerations and challenges in studying liquid-liquid phase separation and biomolecular condensates. Cell, 176(3), 419–434. DOI
- Elbaum-Garfinkle, S., Kim, Y., Szczepaniak, K., Chen, C. C.-H., Eckmann, C. R., Myong, S., and Brangwynne, C. P. (2015). The disordered P granule protein LAF-1 drives phase separation into droplets with tunable viscosity and dynamics. Proceedings of the National Academy of Sciences, 112(23), 7189–7194. DOI