Nonunique tangent maps at isolated singularities of minimizing p p -harmonic maps

Jonas Hirsch · 2025

Read Nonunique tangent maps at isolated singularities of minimizing p p -harmonic maps by Jonas Hirsch (2025) 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-2509-02740.txt.

Introduction

Jonas Hirsch wrote a 2025 solo analysis paper. It gives the first example of an energy-minimising p-harmonic map with nonunique tangent maps. The singularity is isolated. Tangent maps are blown-up limits. They reveal a map's fine structure near a singular point.

Until now, nonuniqueness was known only for stationary maps. That is a weaker class. Hirsch constructs a target manifold with a special trait. For each allowed dimension and exponent range, some map into it minimises the p-energy. That map has an isolated singularity at the origin. It admits a whole continuum of distinct tangent maps there. The construction extends B. White's earlier stationary example for the quadratic case. Minimizers can act worse near singularities than theory once hoped.

Catalog record

CREATORJonas Hirsch
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-02740

Browse the full library · About the Tower of Babel

Full text

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

Nonunique tangent maps at isolated singularities of minimizing p p -harmonic maps Jonas Hirsch Abstract . The analysis of “tangent maps” at singular points of energy minimizing maps plays an important role in our understanding of the fine structure of the singular set. This note presents the first example of a minimizing (not just stationary) p p -harmonic map with nonunique tangent maps at an isolated singularity. We construct a n n -dimensional manifold N N such that for every admissible tuple p < m ≤ n + 2 p<m\leq n+2 , there exists a map from B 1 m B_{1}^{m} into N N that minimizes the p p -energy, has an isolated singularity at the origin and admits a continuum of distinct tangent maps. The construction builds upon and extends B. White’s example for p = 2 p=2 in the stationary case [ 9 ] . A tribute to Brian White Introduction Let \Omega\subset\mathbb{R}^{m} be a bounded domain and N a compact smooth Riemannian manifold of dimension n\geq 2 isometrically embedded in some Euclidean space \mathbb{R}^{l} . The p -energy of a map u\in W^{1,p}(\Omega,N):=\{v\in W^{1,p}(\Omega,\mathbb{R}^{l})\colon v(x)\in N\text{ a.e. }\} is E_{p}(u):=\int_{\Omega}\lvert Du\rvert^{p}. The map u is called minimizing in \Omega iff for all v\in W^{1,p}(\Omega,N) with u-v\in W^{1,p}_{0}(\Omega,\mathbb{R}^{l}) one has E_{p}(u)\leq E_{p}(v) . The direct method of the calculus of variations provides, for any given v_{0}\in W^{1,p}(\Omega,\mathbb{R}^{l}) the existence of at least one energy minimizing map u with the same boundary data. A map u locally minimizing the p -energy is of class C^{1,\alpha} outside of a set of Haussdorff dimension at most m-[p]-1 . This was shown for p=2 by Schoen and Uhlenbeck, in [ 7 ] , it had been extended to general p by F. Fuchs [ 2 ] , R. Hardt, F. Lin [ 4 ] , and S. Luckhaus [ 6 ] extended their result to general p . Appealing to the monotonicity formula and the sub-sequential compactness of harmonic maps, they showed that for any y\in\Omega and every sequence r_{i}\searrow 0 , a subsequence of the maps x\mapsto u(y+r_{i}x) converges locally in energy to a map u_{\infty}\colon\mathbb{R}^{m}\to N . Such a map u_{\infty} is zero homogeneous, itself locally minimizing the p -energy and called a tangent map of u at y . Hence there is a special interest in understanding this particular class of energy minimizer, more precisely the set of possible tangent maps at a given point y . For instance one could hope that a tangent map u_{\infty} provides a good picture of u near y . This would not be true if there were more than one tangent map at y meaning there are two different subsequences that converge to different tangent maps. Whether such pathological behavior is possible remains unclear to the best of the author’s knowledge, even though Brian White announced the existence of such an example in [ 9 ] . The uniqueness of tangent map has been shown (1) if there is one constant tangent map at y because then y is actually a regular point (2) if the target manifold is analytic and singularity y is isolated, proven in the groundbreaking work of L. Simon, [ 8 ] (3) if m=3 and N is two-dimensional in which case the Gulliver and White could prove a rate of convergence, [ 3 ] (4) and more recently with the help an epiperimetric inequality once again for analytic targets, [ 1 ] . Results (2) through (4) have been proven in the context of classical 2 -harmonic maps. However, only (1) holds for general p . However, the author expects that L. Simon’s result probably extends to general p . The aim of this note is to give an example for the pathological behaviour: Theorem 0.1 . For every p>1 , there is a constant m(p) such that for any m>m(p) there exists a smooth manifold N of dimension n\geq m+2 and a p -harmonic map U\colon B^{m}_{1}\to N with an isolated singularity at 0 and a continuum of distinct tangent maps. Furthermore U is the unique minimising map with these boundary data. The argument is inspired by the example of B. White for a stationary but not necessarily minimizing harmonic map with a continuum of distinct tangent maps. It is worth mentioning that the manifold N is kind of universal that for every compact set \mathcal{P} of parameters (p,m) one can choose N independent of the individual elements in \mathcal{P} . 1. The “universal” manifold Following B. White we consider the “universal” n+2 dimensional manifold N being the product \left(T^{1}\times\mathbb{R}\right)\times S^{n} associated with the metric (1.1) g=d\theta^{2}+dy^{2}+(B_{0}-V(\theta,y))\,g_{0}\,; where (T^{1}=\mathbb{R}/\mathbb{Z},d\theta^{2}) is the flat torus, g_{0} is the round metric on S^{n} , the large constant B_{0} satisfies at least B_{0}>\sup(|V|+|\nabla V|)+1\,. Even though most of our analysis does not depend on the specific choice of the potential V , we take it to be the very same as B. White proposed i.e. (1.2) V(\theta,y)=-e^{-\nicefrac{{1}}{{y^{2}}}}\sin\left(\theta+\nicefrac{{1}}{{y}}\right) This defines a complete metric on N for any n\in\mathbb{N} . We will always assume that n\geq m-1 but otherwise the result does not depend on the specific choice of n . In that sense the manifold is “universal”. Each map U taking values in N can be divided into the part u taking values in T^{1}\times R and v taking values in S^{n} . The associated p -energy of a map U=(u,v)\colon B_{1}^{m}\to N is henceforth E(U)=\int_{B_{1}}\left(|Du|^{2}+B(u)\,|Dv|^{2}\right)^{\frac{p}{2}}\,. Independent of the dimension of the spherical part S^{n} in N one has that the orthogonal group O(m) acts on B_{1}\subset\mathbb{R}^{m} and N . In the latter case by g\cdot(\theta,y,z)=(\theta,y,g\,z) where we consider g\in O(m)\subset O(n+1) by identifying \mathbb{R}^{m} with \mathbb{R}^{m}\times\{0\}\subset\mathbb{R}^{n} . We will be particular interested in maps that are O(m) -equivariant i.e. U(g\,x)=g\cdot U(x) . One realises that every equivariant map is of the form (1.3) U(x)=\bigl(u(|x|),\pm\frac{x}{|x|}\bigr) where u\colon(0,1)\to T^{1}\times\mathbb{R} is radial symmetric. The theorem is a corollary of the following two propositions, that will be proven in the next two sections. Proposition 1.1 . There are constants \epsilon_{1}=\epsilon_{1}(m,p),\,\overline{B}_{0}=\overline{B}_{0}(m,p,\epsilon_{1},|V|,|\nabla V|) such that any energy minimizing map U=(u,v)\colon B_{1}^{m}\to N with (1) equivariant boundary data, i.e. U(x)=(u_{0},x) on \partial B_{1} , u_{0}\in T^{1}\times\mathbb{R} constant. (2) E(U)\leq\epsilon_{1}\,B_{0}^{\nicefrac{{(p+1)}}{{2}}} (3) 4(m-1)<(m-p)^{2} and B_{0}>\overline{B}_{0} is equivariant. A consequence of the second proposition will be that the equivariant harmonic maps we are interested in will satisfy the assumption ( 2 ) on the energy. Remark 1.1 . The condition 4(m-1)<(m-p)^{2} is necessary in the sense that if 4(m-1)>(m-p)^{2} the hedge hog v_{0}(x)=\frac{x}{|x|} is not even a stable harmonic map from B_{1} into S^{m} , compare [ 5 ] . Proposition 1.2 . There is an equivariant p -harmonic map U(x)=(u(|x|),\frac{x}{|x|}) that is the unique minimizer with respect to its own boundary data and additionally satisfies (1) E(U)\leq CB_{0}^{\nicefrac{{p}}{{2}}} , (2) 0 is an isolated singularity and its tangent maps there are the continuum \left\{(\theta,0,\frac{x}{|x|})\colon\theta\in T^{1}\right\}\,. 2. Proof of Proposition 1.1 We will divide the proof into the following four substeps: Step 1: Let \tilde{U}=(\tilde{u},v)\colon B_{1}\to N be given such that \tilde{u}=u_{0} constant on \partial B_{1} for some u_{0}\in T^{1}\times\mathbb{R} then there exists u(x)=u(|x|) radial symmetric with u(1)=u_{0} such that E((u,v))\leq E(\tilde{U})\,. Equality implies that \tilde{u} is radial. Step 2:

Read the complete text