Classification of the ruled surfaces that are critical points of the Dirichlet energy

Rafael López · 2026

Read Classification of the ruled surfaces that are critical points of the Dirichlet energy by Rafael López (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14384.txt.

Introduction

López classifies every ruled surface in Euclidean space that is a critical point of the Dirichlet energy. Such critical points mark a kind of optimal shape under a variational energy. Ruled surfaces are formed by sweeping a straight line through space, so the result pins down exactly which line-swept shapes meet the energy condition. The paper gives explicit parametrisations of all such surfaces, completing the classification in full.

Catalog record

CREATORRafael López
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE29249
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.14384

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Classification of the ruled surfaces that are critical points of the Dirichlet energy Rafael López Abstract . We classify all ruled surfaces in Euclidean space that are critical points of the Dirichlet energy, obtaining explicit parametrizations of these surfaces. 1991 Mathematics Subject Classification Primary 53A10; Secondary 53C42 1. Introduction and statement of the results Let \Sigma be a surface in Euclidean space \mathbb{R}^{3} given as the graph z=u(x,y) of a smooth function u defined in a bounded domain \Omega\subset\mathbb{R}^{2} . The Dirichlet energy of u is the integral \int_{\Omega}|Du|^{2} . The function u is a critical point of the Dirichlet energy for all volume-preserving variations of \Sigma if and only if u is a critical point of the functional E[u]=\int_{\Omega}|Du|^{2}+\Lambda\int_{\Omega}u, where \Lambda\in\mathbb{R} stands for a Lagrange multiplier. Notice that the integral \int_{\Omega}u is the volume enclosed by \Sigma with the solid cylinder \Omega\times\mathbb{R} . The Euler-Lagrange equation of E is u_{xx}+u_{yy}=\frac{\Lambda}{2}, (1) where u=u(x,y) . In the particular case \Lambda=0 , the function u is harmonic. A surface \Sigma satisfying locally ( 1 ) is called a stationary surface of the Dirichlet energy. In this paper we want to study stationary surfaces of the functional E under some geometric properties using techniques of differential geometry. Here we follow the seminal paper [ 10 ] . Instead to use the function u , it is more suitable to write E[u] in terms of the surface \Sigma . If \nu=(\nu_{1},\nu_{2},\nu_{3}) is the unit normal of \Sigma , then \nu=(-Du,1)/\sqrt{1+|Du|^{2}} . Thus the Dirichlet energy becomes \mathcal{F}(\Sigma)=\int_{\Sigma}\left(\dfrac{1}{\nu_{3}}-\nu_{3}\right)\,d\Sigma, (2) where d\Sigma is the area element of \Sigma . More generally, we can consider functionals of type \mathcal{F}(X)=\int_{\Sigma}F(\nu)\,d\Sigma , where X\colon\Sigma\to\mathbb{R}^{3} is an immersion of \Sigma into \mathbb{R}^{3} and F\colon U\subset\mathbb{S}^{2}\to\mathbb{R}^{+} is a positive smooth function on the unit sphere \mathbb{S}^{2} . These functionals are called anisotropic because the energy depends on the normal direction \nu of \Sigma . Anisotropic energies appear in fluid phenomenon when the surface tensions of interfaces depend on \nu [ 12 ] . Critical points of \mathcal{F} for compactly supported volume-preserving variations are characterized by the property that the function \Lambda given by \Lambda:=2HF-\mbox{div}_{\Sigma}DF is constant, where DF is the gradient of F in \mathbb{S}^{2} . The function \Lambda is called the anisotropic mean curvature of \Sigma . Thus, critical points of \mathcal{F} , or stationary surfaces, are surfaces with constant anisotropic mean curvature (CAMC). Usually, for general anisotropic energies, it is assumed that F is elliptic in the sense that the matrix D^{2}F+F\,\mbox{Id} is positive definite, where D^{2}F is the Hessian of F . This allows to define the Wulff shape associated to \mathcal{F} as the map \xi\colon\mathbb{S}^{2}\to\mathbb{R}^{3} given by \xi(\nu)=DF(\nu)+F(\nu)\nu . This map is the parametrization of a convex surface, which it is an ovaloid if U=\mathbb{S}^{2} . We point out that the Dirichlet energy is elliptic and its Wulff shape is the paraboloid z=x^{2}+y^{2} , after translations and rescalings in \mathbb{R}^{3} . Notice that the integrand in ( 2 ) is only defined on a hemisphere of \mathbb{S}^{2} . We ask for those CAMC surfaces of the Dirichlet energy ( 2 ) with some particular geometric property. A first example are the rotational surfaces, that is, surfaces that are axially symmetric about an axis of \mathbb{R}^{3} . This reduces the equation \Lambda=ct. in an ordinary differential equation. Exactly, CAMC surfaces of the Dirichlet energy that are axially symmetric about the z -axis are given by u(x,y)=c_{1}\log(\sqrt{x^{2}+y^{2}})+\frac{\Lambda}{8}(x^{2}+y^{2})+c_{2},\quad c_{1},c_{2}\in\mathbb{R}. When c_{1}=0 we obtain the Wulff shape and when c_{2}=0 we have the anisotropic catenoid of the Dirichlet energy. For the interesting reader in the differential-geometric viewpoint of critical points of anisotropic energies, we refer that works of Koiso and Palmer [ 6 , 7 , 9 , 8 ] and references therein. See also the recent papers [ 1 , 3 , 4 , 5 , 11 ] . Other interesting family of surfaces are the ruled surfaces. These surfaces are generated by moving a straight line in Euclidean space \mathbb{R}^{3} and they can be parametrized by X\colon I\times\mathbb{R}\to\mathbb{R}^{3},\quad X(s,t)=\alpha(s)+t\beta(s), (3) where \alpha\colon I\to\mathbb{R}^{3} is a regular curve, called the directrix which we can suppose to be parametrized by arc-length. The curve \beta is a curve on the unit 2 -sphere \beta\colon I\to\mathbb{S}^{2} and it indicate the direction of the rulings at each point \alpha(s) . Moreover, we can assume \langle\alpha(s),\beta(s)\rangle=0 for all s\in I . In case that \beta is a constant curve, \beta(s)=\vec{w} , the surface is called cylindrical . In this paper, we give a complete classification of the ruled CAMC surfaces of the Dirichlet energy ( 2 ). Ruled surfaces of cylindrical type are studied separately. We prove in Sect. 3 that any cylindrical CAMC surface is a plane ( \Lambda=0 ) or the directrix \alpha is a parabola, that is, the surface is a parabolic cylinder (Thm. 3.1 ). For non-cylindrical CAMC surfaces, we have that \{\beta,\beta^{\prime},\beta\times\beta^{\prime}\} is an orthonormal frame. Since \langle\alpha(s),\beta(s)\rangle=0 , there are smooth functions a,b\colon I\to\mathbb{R} such that \alpha(s)=a(s)\beta^{\prime}(s)+b(s)\beta(s)\times\beta^{\prime}(s). (4) The classification of the ruled non-cylindrical CAMC surfaces is the following. Theorem 1.1 . Let \Sigma be a ruled non-cylindrical surface in \mathbb{R}^{3} parametrized by ( 3 ) where \alpha is given by ( 4 ). If the anisotropic mean curvature \Lambda of the Dirichlet energy is constant then \beta is a great circle of \mathbb{S}^{2} . Moreover, up to a rotation about the z -axis and translations of \mathbb{R}^{3} , the surface \Sigma is one of the following surfaces: (1) Case \Lambda=0 . (a) A plane. (b) A helicoid. The curve \beta is the horizontal equator \beta(s)=(\cos s,\sin s,0) and \begin{split}a(s)&=c_{1}\cos s,\quad c_{1}\not=0,\\ b(s)&=b_{1}\tan(s),\quad b_{1}\not=0.\end{split} (5) (c) The curve \beta is the vertical equator \beta(s)=(\cos s,0,\sin s) and \begin{split}a(s)&=c_{1}\cos(s+c_{2}),\quad c_{1},c_{2}\in\mathbb{R},c_{1}\not=0,\\ b(s)&=b_{1}\tan(s),\quad b_{1}\not=0.\end{split} (6) (d) The curve \beta is a (non-horizontal, non-vertical) great circle of \mathbb{S}^{2} and \begin{split}a(s)&=c_{1}\cos s+c_{2}\sin s\\ &-\frac{b_{1}m}{1-m^{2}}\cos s\cot^{-1}(\sqrt{1-m^{2}}\cot s)+\frac{b_{1}}{m\sqrt{1-m^{2}}}\sin s,\\ b(s)&=\frac{b_{1}}{\sqrt{1-m^{2}}}\tan^{-1}(\sqrt{1-m^{2}}\tan s).\end{split} (7) Here c_{1},c_{2},b_{1},m\in\mathbb{R} with b_{1}\not=0 , m\in(0,1) . (2) Case \Lambda\not=0 . Then \beta is the vertical equator \beta(s)=(\cos s,0,\sin s) and \begin{split}a(s)&=c_{1}\cos s+c_{2}\sin s-\frac{\Lambda b_{1}^{2}\cos(2s)}{4\cos s},\quad c_{1},c_{2}\in\mathbb{R},\\ b(s)&=b_{1}\tan s,\quad b_{1}\not=0.\end{split} (8) The proof of Thm. 1.1 will be carried out in Sect. 4 separating the case \Lambda=0 (Subsect. 4.1 ) from \Delta\not=0 (Subsect. 4.2 ). 2. Preliminaries In this section, we recall the calculation of the anisotropic mean curvature of a surface \Sigma when the energy functional is of type \mathcal{F}(X)=\int_{\Sigma}F(\nu_{3})\,d\Sigma , and next, we particularize to the Dirichlet energy ( 2 ). A first observation is that the equation \Lambda=ct. is preserved by translations of \mathbb{R}^{3} and by dilations (with different constant \Lambda ), but not by rigid motions in general. This is because the functional \mathcal{F} depends on the unit normal \nu . However, the equation \Lambda=ct. is invariant by rotations about an axis parallel to the z -axis. Let h\colon T\Sigma\times T\Sigma\to\mathbb{R} be the second fundamental form of \Sigma , h(v,w)=-\langle d\nu(v),w\rangle , v,w\in T\Sigma . The anisotropic mean curvature \Lambda is given by \Lambda=\frac{h(v_{1},v_{1})}{\mu_{1}}+\frac{h(v_{2},v_{2})}{\mu_{2}}, (9) where \{v_{1},v_{2}\} is an orthonormal frame of T\Sigma and \mu_{1} and \mu_{2} are the principal curvatures of the Wulff shape given by \frac{1}{\mu_{1}}=(1-\nu_{3}^{2})F^{\prime\prime}(\nu_{3})+\frac{1}{\mu_{2}},\quad\frac{1}{\mu_{2}}=F-\nu_{3}F^{\prime}(\nu_{3}). The principal directions \{E_{1},E_{2}\} of \mu_{1} and \mu_{2} are E_{1}=e_{3}-\nu_{3}\nu,\quad E_{2}=\nu\times E_{1}, where e_{3}=(0,0,1) . Here it is understood that \nu_{3}^{2}\not=1

Read the complete text