On the Multi-Dimensional Divergence-Curl Problem and Its Connection with Pseudo-Harmonic Fields

Technical Summary: On the Multi-Dimensional Divergence-Curl Problem and Its Connection with Pseudo-Harmonic Fields

Problem Statement
The paper addresses the solvability of the multi-dimensional divergence-curl problem for a vector field $v$ in a domain $\Omega$. The objective is to find a solenoidal vector field satisfying a prescribed curl function $f$ under homogeneous Dirichlet boundary conditions (the no-slip condition):
$$\begin{cases} \text{curl } v = f, \\ \text{div } v = 0, \\ v|_{\partial\Omega} = 0. \end{cases}$$
While the two-dimensional case is well-understood—requiring the orthogonality of $f$ to harmonic functions—the extension to higher dimensions ($n \geq 3$) with the no-slip condition presents significant challenges. Previous studies have addressed the problem with weaker boundary conditions, without boundary conditions, or with prescribed tangential/normal components, but the specific case of the homogeneous no-slip condition in higher dimensions requires a refined solvability criterion.

Methodology
The author employs the language of differential forms on Riemannian manifolds to generalize the problem. The system is reformulated as finding a $k$-form $\omega$ such that $d\omega = f$ and $d^*\omega = 0$, subject to $\omega|_{\partial M} = 0$. Here, $d$ is the exterior derivative and $d^*$ is the codifferential.

The core theoretical framework relies on the vector Laplacian representation $\Delta = d^*d + dd^*$. The paper distinguishes between:

1. Harmonic fields: Forms in the kernel of both $d$ and $d^*$ (where $\Delta \omega = 0$).
2. Pseudo-harmonic fields: Forms defined by the condition $dd^*\omega = 0$.

The author argues that for boundary value problems with the no-slip condition, the necessary and sufficient solvability condition is not orthogonality to harmonic fields, but rather orthogonality to the broader class of pseudo-harmonic fields (the kernel of $dd^*$). This distinction is crucial because, unlike on manifolds without boundary where $dd^*\omega = 0$ is equivalent to $d^*\omega = 0$ (due to adjointness), the presence of a boundary alters the relationship, making the orthogonality to $\text{Ker}(dd^*)$ the correct constraint.

Key Contributions and Results

1. General Solvability Criterion:
   The paper establishes a theorem for bounded Lipschitz domains stating that the problem admits a unique solution if and only if:

   • $\text{div } f = 0$ (a necessary condition from the properties of the exterior derivative).
   • $f$ is orthogonal in $L^2$ to all pseudo-harmonic fields $h$ (i.e., $\int_\Omega (f, h) dx = 0$ for all $h$ such that $dd^*h = 0$).

   The proof utilizes a decomposition of the space and the properties of the adjoint operator. Uniqueness is demonstrated via Stokes' formula, showing that if $f=0$, the solution must vanish.

2. Explicit Construction in Spherical Domains:
   To make the abstract criterion concrete, the author applies the theory to the exterior of a sphere in $\mathbb{R}^3$. By expanding the fields in vector spherical harmonics, the author constructs a countable family of pseudo-harmonic fields $\{a_n\}$ that satisfy $dd^*a_n = \text{curl}^2 a_n = 0$.

   • These fields are explicitly identified as $\Phi_{\ell m} / r^{\ell+1}$, where $\Phi_{\ell m}$ are toroidal vector spherical harmonics.
   • The paper proves these fields are divergence-free and harmonic ($\Delta a_n = 0$) but possess non-zero curl, distinguishing them from standard harmonic fields in the context of the boundary conditions.

3. Integral Solvability Conditions:
   For the exterior of a sphere, the abstract orthogonality condition translates into a specific countable set of integral constraints on the vorticity function $f$:
   $$\int_{B_{r_0}} f(x) \cdot \frac{\Phi_{\ell m}(x/|x|)}{|x|^{\ell+1}} \, dx = 0, \quad \ell = 1, \dots, \infty, \, m \in [-\ell, \ell].$$
   The paper demonstrates that if these conditions are met, a unique solution exists and is given explicitly by the Biot–Savart–Laplace formula:
   $$v(x) = -\frac{1}{4\pi} \int_{B_{r_0}} \frac{(x-y) \times f(y)}{|x-y|^3} \, dy.$$
   Furthermore, the author provides $L^p$-$L^q$ estimates for the solution and its gradient, relying on the Hardy–Littlewood–Sobolev inequality and Calderón–Zygmund singular integral theory.

4. Handling Unbounded Domains and Asymptotics:
   The paper clarifies that in unbounded domains, the orthogonality condition does not require $f$ to be orthogonal to constants (unlike the 2D case). It also analyzes the modification of solvability conditions if the boundary condition at infinity is changed from vanishing velocity to a prescribed uniform flow, showing that the orthogonality conditions remain largely intact, with only specific modes affected.

Significance and Claims
The paper claims to extend the known solvability criteria for the divergence-curl problem from two dimensions to higher dimensions by replacing the orthogonality to harmonic functions with orthogonality to pseudo-harmonic fields. The primary significance lies in providing a rigorous, necessary, and sufficient condition for the existence of solutions under the physically relevant no-slip boundary condition.

The author highlights the practical relevance of these findings for numerical solutions. In computational hydrodynamics, enforcing orthogonality to an infinite countable family of fields is often infeasible. The paper suggests that imposing orthogonality to a finite subset of these pseudo-harmonic fields allows for an approximate solution where the boundary value is "approximately zero," corresponding to the "partial-slip" condition. This offers a theoretical basis for handling the constraints on the right-hand side of the divergence-curl system in numerical schemes.

The work does not propose new experimental applications or future research directions beyond the mathematical framework established, focusing strictly on the analytical solvability and the explicit construction of the required orthogonality conditions.