Focal Sets in Hypersurface Geometry

• Focal sets of hypersurfaces are geometric loci where normal vectors develop singularities, signaling curvature degeneracies and caustic formations.
• They arise in various geometries—including Riemannian, affine, and Finsler—providing a framework to study wavefronts, integrable systems, and submanifold theory.
• Explicit curvature formulas and rigidity conditions in isoparametric cases offer practical insights into their structure and applications in geometric analysis.

A focal set of a hypersurface is the geometric locus in the ambient manifold where normals to the hypersurface develop singularities, typically corresponding to degeneracy of the exponential map along the normal bundle. Focal sets arise in Riemannian, Finsler, affine, and even algebraic-geometric settings, and play a foundational role in the study of wavefronts, caustics, integrable systems, and the global theory of submanifolds.

1. General Definition and Construction

Let $\Sigma^n \subset \mathbb{R}^{n+1}$ (or more generally, a manifold $M^{n+1}$) be a smooth, oriented hypersurface with unit normal field $\nu$. The classical focal set is constructed as the set of points

$q - r_i(q)\,\nu(q), \qquad i = 1,\ldots,n,$

where $q \in \Sigma$ and $r_i(q)$ are the principal radii of curvature, reciprocals of the principal curvatures $\kappa_i$ at $q$. For each $i$, the image $S_i$ of the map $q \mapsto q - r_i(q)\nu(q)$ is called a focal sheet, and the union over $i$ gives the focal set of $\Sigma$ (Georgiou et al., 24 Jan 2026).

In full generality, let $\Gamma$ be a submanifold of the space of oriented lines—identified with the tangent bundle $T\mathbb{S}^n$ via $(\xi, \eta)$ with $|\xi|=1$, $\eta\cdot\xi=0$—and define the evaluation map

$$\Psi: T\mathbb{S}^n \times \mathbb{R} \to \mathbb{R}^{n+1}, \quad \Psi((\xi, \eta), r) = \eta + r\xi.$$

A focal point of $\Gamma$ is a value $p$ at which $d\Psi$ fails to be surjective, and the focal set $F(\Gamma)$ is the union of all such $p$ (Georgiou et al., 24 Jan 2026).

2. Focal Sets in Isoparametric and Space-Form Geometry

Isoparametric hypersurfaces—those with constant principal curvatures—produce particularly tractable focal sets, often called focal submanifolds. For $M^n \subset N^{n+1}$, the focal set is the image under the normal exponential map of those vectors at which the differential degenerates. For each principal curvature $\kappa_i$, the focal distance along the normal is the first zero of the equation

$$y''(t) + K_i(t) \, y(t) = 0, \quad y(0)=1,\, y'(0)=-\kappa_i,$$

where $K_i(t)$ is the ambient curvature in the normal plane. In space forms, these are given by explicit trigonometric or hyperbolic expressions: $\begin{aligned} &K=-1: & y(t) &= \cosh t - \kappa_i \sinh t \implies t_i = \frac{1}{\kappa_i}\,\artanh(\kappa_i),\ &K=+1: & y(t) &= \cos t - \kappa_i \sin t \implies t_i = \frac{1}{\kappa_i}\arctan(\kappa_i). \end{aligned}$ Each $t_i$ determines a focal sheet $F_i$ of codimension $\geq 2$, and the entire focal set is the union over $i$ (Dominguez-Vazquez et al., 2021, Reis et al., 2017, He et al., 2017).

3. Curvature Formulas and Rigidity

The geometry of focal sets is controlled by explicit curvature formulas. In the Euclidean case, for a hypersurface $\Sigma$ with distinct, simple radii $r_1,\ldots,r_n$, the sectional curvature of the focal sheet $S_1$ in the 2-plane spanned by the principal direction fields $X_1,X_2$ is

$$K_{S_1}(\Pi) = \frac{X_1(r_1 - r_2) - 1}{(r_1 - r_2)^2},$$

where $s_{12}=r_1 - r_2$ is the astigmatism function, interpreted as the signed normal separation between the two sheets, and $X_1(r_1 - r_2)$ is the directional derivative (Georgiou et al., 24 Jan 2026).

For higher symmetry, the classical Bianchi theorem states that when $s_{ij} = r_i - r_j$ is constant, all focal sheets $S_i$ have constant negative sectional curvature $K = -1/(s_{ij})^2$ in the corresponding 2-planes, generalizing the pseudospherical character from the $n=2$ case to arbitrary dimension (Georgiou et al., 24 Jan 2026).

Rigidity occurs under lower curvature bounds: if the ambient manifold has $\operatorname{sec} \geq 1$ or $\operatorname{Ric} \geq n-1$, the focal radius of any hypersurface is at most $\pi/2$ and equality forces total geodesy, with ambient isometry to sphere or projective space (Guijarro et al., 2016).

4. Focal Sets in Isoparametric Functions and General Riemannian Manifolds

Given an isoparametric function $f: N \to \mathbb{R}$, its regular level sets are tubes of constant mean curvature, and its singular level sets $M_-, M_+$ are the focal submanifolds, each of codimension at least $2$. These focal submanifolds are always minimal in $N$, and, in the case where every level has constant principal curvatures, have constant eigenvalues in all normal directions—the “austere” property (Ge et al., 2010, Li et al., 2015, Ge, 2011).

There is a filtration of possible curvature identities for focal varieties, depending on the number of constant-order mean curvatures $Q_k$ satisfied by tubes around the focal submanifold. As $k$ increases, stronger curvature equalities (e.g., on Ricci or scalar curvature) are forced (Ge, 2011).

In Finsler geometry, focal submanifolds of isoparametric hypersurfaces are anisotropic-minimal and satisfy a Finsler–Cartan formula for the eigenvalues of their shape operators (He et al., 2017).

5. Singularities, Caustics, and Symplectic Characterizations

Focal sets (caustics) generally contain singularities: these arise where principal radii coincide, leading to ridges, umbilic points, or higher degeneracy loci. The curvature formulas above apply away from such singular sets but can be used to analyze and predict their local behavior using catastrophe theory and the geometry of Jacobi fields (Georgiou et al., 24 Jan 2026). In the symplectic model, isotropic submanifolds of $T\mathbb{S}^n$ correspond to normal congruences of hypersurfaces, and the focal set construction is recast in terms of the degeneracy of the evaluation map (Georgiou et al., 24 Jan 2026).

Affine analogs appear as bifurcation sets of the affine distance function for codimension-2 submanifolds in hypersurfaces. The regularity and singularity types are classified for generic cases, and the locus can collapse to a line when the submanifold is umbilic and normally flat (Craizer et al., 2016).

6. Focal Sets in Broader Geometric and Algebraic Settings

In symmetric and Damek–Ricci spaces, isoparametric hypersurfaces admit focal sets generalizing those in space forms. Notably, in non-compact symmetric spaces of rank at least $3$, inhomogeneous isoparametric hypersurfaces with non-austere focal sets exist, highlighting that minimiality and austerity can decouple (Dominguez-Vazquez et al., 2021).

For hypersurfaces in projective algebraic geometry covered by linear spaces (with tangent constant along these subspaces), the focal locus is the branch locus of the natural congruence, and is of codimension at least $2$ in the resolved space. When this locus vanishes, $X$ is a projective bundle (e.g., cones or developables), and this is reflected in the vanishing of the Hessian determinant (Poi et al., 2019).

7. Applications and Further Directions

Focal sets underpin the study of wavefronts and caustics in geometric optics, drive algorithms in computer–aided geometric design (offsets and toolpath generation), control the evolution of singularities in mean curvature flow, and structure integrable systems (e.g., via Bäcklund transformations between focal sheets or via the geometric theory of solitons) (Georgiou et al., 24 Jan 2026, Reis et al., 2017). There is a vast interplay between minimality, convexity, eigenvalue constraints, and the global topological structure of focal sets, with ongoing research exploring generalizations in Finsler, affine, and projective settings, and classification in non-homogeneous or exotic ambient geometries.