---
title: Isocapacitary Inequalities
url: https://www.emergentmind.com/topics/isocapacitary-inequalities
type: topic
---

# Isocapacitary Inequalities

Isocapacitary inequalities are a class of geometric-functional inequalities relating the capacity of sets—defined via variational or potential-theoretic principles—to geometric measures such as volume, perimeter, or curvature, often under extremal or comparison constraints. Originating in the work of Maz’ya and further developed in relation to sharp Sobolev, trace, and spectral estimates, these inequalities unify and extend isoperimetric principles to a wide range of analytic, geometric, and probabilistic settings, including linear and nonlinear operators, nonlocal energies, anisotropic norms, graphs, manifolds, and convex sets.

## 1. Capacity: Definitions and Variational Principles

For a compact set $K\subset\mathbb{R}^n$ and $1<p<n$, the $p$-capacity is the minimal energy required to "charge" the set to level $1$ using test functions vanishing at infinity:
\[
\operatorname{Cap}_p(K) = \inf\left\{ \int_{\mathbb{R}^n} |\nabla u|^p\, dx \ :\ u \in C_c^{0,1}(\mathbb{R}^n),\ u \geq 1 \ \text{on}\ K \right\}.
\]
Analogous definitions arise for the Dirichlet or Neumann (relative) capacities, the fractional Sobolev $(s,p)$-capacity $\,\operatorname{Cap}_{s,p}(K,G)$ on open sets $G$, the anisotropic $p$-capacity using Minkowski norms $F$, and fully nonlinear $k$-Hessian capacity via admissible convex potentials [2010.04604][2012.13933][1305.0721].

The notion of capacity generalizes to metric measure spaces, graphs, Riemannian manifolds, and domains with weakly regular or fractal boundaries, with context-dependent adaptations of the test function spaces and boundary conditions [2311.00100][2601.12192][2406.12583][2412.21008].

## 2. Classical Isocapacitary Inequalities

The prototypical isocapacitary inequality asserts that, among sets of fixed volume, the Euclidean ball minimizes $p$-capacity:
\[
\operatorname{Cap}_p(E) \geq \operatorname{Cap}_p(B) \left( \frac{|E|}{|B|} \right)^{(n-p)/n}
\]
for any measurable $E$ with $|E|=|B|$ [2010.04604][1401.4322][1705.07482][1901.11309]. Equality occurs if and only if $E$ is a ball (up to symmetry). The analogous two-sided forms occur for fractional capacities and Sobolev-type (coarea) settings:
\[
|K|^{\frac{n-s}{n}} \leq C \,\operatorname{Cap}_{s,1}(K,G)
\]
with matching constants for the isoperimetric and Sobolev embeddings [1312.3135].

Variants include upper bounds for capacity in terms of geometric functionals, such as perimeter, mean curvature integrals, or perimeter squared divided by volume, with balls as equality cases [2309.08364].

## 3. Quantitative and Stability Results

Recent work has established sharp quantitative forms tying the deficit in capacity to an asymmetry measure, typically Fraenkel asymmetry:
\[
\operatorname{Cap}_p(\Omega) - \operatorname{Cap}_p(B) \geq c(n,p) \mathcal{A}(\Omega)^2\ r^{n-p}
\]
where $\mathcal{A}(\Omega)$ is the $L^1$ distance from $\Omega$ to the closest ball of equal volume, and $r$ is the associated radius [2010.04604][1901.11309][2309.08364]. The exponent $2$ is optimal, a fact verified via nearly spherical perturbations and second-variation computations.

These quantitative inequalities ensure that sets with capacity nearly identical to a ball must themselves be close to a ball (rigidity). They extend the scope of isocapacitary inequalities to stability estimates for solutions of PDEs, optimality in shape optimization, and precise criteria for the non-attainability of capacity minimality outside balls.

## 4. Isocapacitary Inequalities for Nonlinear, Fractional, and Anisotropic Operators

Isocapacitary inequalities have been established for a broad spectrum of non-standard capacities, including:

- **Fractional capacities:** Equivalence theorems link fractional Sobolev embeddings, fractional capacities, and fractional perimeter isoperimetry; the sharp constants coincide [1312.3135].
- **Anisotropic capacities:** Utilizing a Minkowski norm $F$, the anisotropic Minkowski inequality relates the total $L^p$ anisotropic mean curvature to the anisotropic $p$-capacity, with Wulff shapes as extremals [2012.13933].
- **$k$-Hessian capacities:** Bounded domains admit capacitary Sobolev and Moser–Trudinger inequalities in terms of $k$-Hessian capacity, generalizing classical forms and providing control over nonlinear PDE regularity [1305.0721][2201.02061].
- **General Dirichlet forms:** Sobolev-type inequalities are strictly equivalent to isocapacitary bounds for nonlinear forms beyond bilinear settings; regularity, duality, and optimal constants propagate to analytic frameworks including variable exponents and fractal domains [2601.12192].

## 5. Isocapacitary Inequalities on Graphs and Manifolds

Extensions to discrete and geometric settings include:

- **Weighted graphs:** Generalizations of Maz’ya’s isocapacitary constants yield Cheeger-type inequalities for Dirichlet, Neumann, and Steklov eigenvalues on finite and infinite graphs, with sharp universal constants in the bounds [2406.12583].
- **Riemannian manifolds:** For manifolds with boundary, boundary isocapacitary constants provide two-sided estimates for the first Steklov eigenvalue, valid for compact and non-compact cases, and sharp in hyperbolic geometry applications [2412.21008][2512.24725].
- **$3$-manifolds with scalar curvature:** Nonlinear versions of isocapacitary mass interpolate between Jauregui’s isocapacitary mass ($p=2$), Huisken’s isoperimetric mass ($p=1$), and the ADM mass for asymptotically flat geometries. Sharp Penrose-type inequalities are derived in terms of $p$-capacity and area, with equality in Schwarzschild manifolds [2305.01453][2306.00744].

## 6. Trace, Sobolev, and Spectral Equivalence

Isocapacitary inequalities are equivalently recast as trace and Sobolev-type embedding theorems. In numerous contexts, the embedding of a Dirichlet space to $L^p$ or Lorentz spaces holds if and only if measure–capacity inequalities are satisfied with matching (optimal) exponents [2601.12192][2201.02061][1312.3135][1305.0721]. The result is a unification of analytic, geometric, and spectral perspectives:

- Cheeger-type estimates for Laplacians are expressible via isocapacitary constants.
- Spectral lower bounds for Steklov and $p$-Laplacian eigenvalues are characterized by the associated isocapacitary constants [2512.24725][2412.21008][2406.12583].
- Trace inequalities (Sobolev–type and Moser–Trudinger–type) for complex Hessian and Monge–Ampère equations are equivalent to explicit isocapacitary bounds, governing $L^\infty$ and Hölder regularity of solutions [2201.02061].

## 7. Affine-Invariant and Brunn–Minkowski Forms

Recent advances in convex geometry have led to the definition of $p$-affine capacities, providing affine-invariant versions of classical isocapacitary inequalities [1705.07482]. For sets $K$,
\[
C_{p,\tau}(K) \geq C_{p,\tau}(B^n) \left( \frac{V(K)}{V(B^n)} \right)^{(n-p)/n}
\]
with equality only for ellipsoids. This extends the Brunn–Minkowski and Urysohn chains relating volume, capacity, and affine surface area, giving sharper inequalities and full equality characterizations.

The fractional case, e.g., the $1$-Riesz capacity, admits Brunn–Minkowski inequalities and level-set convexity properties for potentials of the fractional Laplacian [1401.4322].

## Table: Key Isocapacitary Inequality Types

| Class/Operator        | Prototype Inequality                                                 | Extremal Case               |
|----------------------|---------------------------------------------------------------------|-----------------------------|
| Linear ($p=2$)       | $\operatorname{Cap}_2(K)\ge \operatorname{Cap}_2(B)$                | Ball                        |
| $p$-Laplacian        | $\operatorname{Cap}_p(K)\ge \operatorname{Cap}_p(B)$                | Ball                        |
| Fractional           | $|K|^{\frac{n-s}{n}}\leq C\,\operatorname{Cap}_{s,1}(K,G)$          | Ball                        |
| Anisotropic ($F$)    | $\int_{\partial\Omega}H_F F(\nu)\ge \cdots \operatorname{Cap}_{F,p}(\Omega)$ | Wulff ball                  |
| Affine-invariant     | $C_{p,\tau}(K)\ge C_{p,\tau}(B^n)(V(K)/V(B^n))^{(n-p)/n}$           | Ellipsoid                   |
| Graphs/Manifolds     | $\lambda_1^{D}(M)\ge c\,\alpha_D(M)$                                | Sharp bounds attained       |
| $k$-Hessian          | $|E|^\alpha\leq c\,\operatorname{cap}_k(E,\Omega)$                  | Ball                        |

## References and Connections

The field is deeply influenced by work of Maz’ya (classical isocapacitary theory), De Philippis–Marini–Mukoseeva (quantitative forms, sharp asymmetry exponents) [1901.11309][2010.04604], and extensions to nonlocal, nonlinear, and geometric settings by various authors including Xia, Ye, Antonini, van den Berg, Han, Hong, and others [2311.00100][2012.13933][2306.00744][1705.07482][2406.12583][2512.24725][1312.3135][1305.0721].

**A plausible implication is** that isocapacitary inequalities serve as a central unifying structure connecting functional inequalities, geometric extremals, and the spectral theory for elliptic and nonlinear operators. They govern regularity, rigidity, sharpness of Sobolev embeddings, and are indispensable in quantitative stability and optimization for PDEs, geometry, and convex analysis.

Source: https://www.emergentmind.com/topics/isocapacitary-inequalities