A remark on isolated removable singularity of harmonic maps in dimension two

Changyou Wang · 2025

Read A remark on isolated removable singularity of harmonic maps in dimension two by Changyou Wang (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-2503-12804.txt.

Introduction

Changyou Wang published this mathematics note in 2025. It concerns harmonic maps, smooth ways of stretching one shape onto another while minimizing a kind of energy. The question is whether such a map, defined on a disk with one point removed, can misbehave at the missing point. Wang gives conditions under which the point is no obstacle: if the map's steepness obeys a self-similar bound near the hole, the map extends smoothly across the whole disk. On the entire plane, the same conditions force the map to be constant.

Catalog record

CREATORChangyou Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE13 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-12804

Browse the full library · About the Tower of Babel

Full text

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

A remark on isolated removable singularity of harmonic maps in dimension two Changyou Wang Abstract . For a ball B R ​ ( 0 ) ⊂ ℝ 2 B_{R}(0)\subset\mathbb{R}^{2} , we provide sufficient conditions such that a harmonic map u ∈ C ∞ ​ ( B R ​ ( 0 ) ∖ { 0 } , N ) u\in C^{\infty}(B_{R}(0)\setminus\{0\},N) , with a self-similar bound on its gradient, belongs to C ∞ ​ ( B R ​ ( 0 ) ) C^{\infty}(B_{R}(0)) . Those conditions also guarantee the triviality of such harmonic maps when R = ∞ R=\infty . 1. Introduction This short note address a question arising from the author’s recent study with Bang [ 1 ] on the question of rigidity for the steady (simplified) Ericksen-Leslie system in \mathbb{R}^{n} , which seeks to answer that if (u,d)\in C^{\infty}(\mathbb{R}^{n}\setminus\{0\},\mathbb{R}^{n}\times\mathbb{S}^{n-1}) solves (1.1) \begin{cases}-\Delta u+u\cdot\nabla u+\nabla p=-\nabla\cdot(\nabla d\odot\nabla d),\\ \nabla\cdot u=0,\\ \Delta d+|\nabla d|^{2}d=u\cdot\nabla d,\end{cases}\ \ {\rm{in}}\ \ \mathbb{R}^{n}\setminus\{0\}, and satisfies a self-similar bound (1.2) |u(x)|\leq\frac{C_{1}(n)}{|x|},\ \ |\nabla d(x)|\leq\frac{C_{2}(n)}{|x|},\ \forall x\in\mathbb{R}^{n}\setminus\{0\}, for some constants C_{1}(n),C_{2}(n)>0 , then (u,\nabla d)\equiv(0,0)\ {\rm{in}}\ \mathbb{R}^{n} ? We obtained in [ 1 ] some partial results towards the question concerning ( 1.1 ) and ( 1.2 ) mentioned above. In particular, among other results we proved that when n\geq 3 , there exists \varepsilon_{n}>0 such that if C_{1}(n),C_{2}(n)\leq\varepsilon_{n} then \nabla d\equiv 0 ; while u\equiv 0 when n\geq 4 , and is a Landau solution of the steady Navier-Stokes equation when n=3 . When n=2 , we constructed infinitely many nontrivial solutions of ( 1.1 ) and ( 1.2 ), that resemble the so-called Hamel’s solutions of steady Navier-Stokes equation in \mathbb{R}^{2} . A Liouville theorem on harmonic maps plays an important role in [ 1 ] , that is, for n\geq 3 if d\in C^{\infty}(\mathbb{R}^{n}\setminus\{0\},N) solves the equation of harmonic maps: (1.3) \Delta d+A(d)(\nabla d,\nabla d)=0\ \ {\rm{in}}\ \ \mathbb{R}^{n}\setminus\{0\}, and there exists an \varepsilon_{0}(n)>0 such that (1.4) |\nabla d(x)|\leq\frac{\varepsilon_{0}(n)}{|x|},\ \forall x\in\mathbb{R}^{n}\setminus\{0\}, then d must be a constant map. Here N\subset\mathbb{R}^{L} is a compact smooth Riemann manifold without boundary, and A denotes the second fundamental form of N . A natural question to ask is that whether this Liouville property remains to be true when n=2 . More precisely, Question 1.1 . Suppose d\in C^{\infty}(\mathbb{R}^{2}\setminus\{0\},N) solves ( 1.3 ) and satisfies ( 1.4 ) for some small constant \varepsilon_{0}(2) . Then d must be constant. To the knowledge of the author, this question has not been addressed from the literature. In contrast with n\geq 3 , ( 1.4 ) alone does not guarantee d has locally finite Dirichlet energy in dimension two, or, equivalently, E(d,B_{1}(0))=\int_{B_{1}(0)}|\nabla d|^{2}<\infty for the unit ball B_{1}(0)\subset\mathbb{R}^{2} . Thus, one may apply neither the celebrated theorem by Sacks-Uhlenbeck [ 3 ] on the removability of isolated singularity of harmonic maps in dimension two, nor the regularity theorem by Hélein [ 2 ] on weakly harmonic maps in dimension two. Observe that d(x)=\frac{x}{|x|}:\mathbb{R}^{2}\setminus\{0\}\to\mathbb{S}^{1} is a harmonic map, satisfying |\nabla d(x)|=\frac{1}{|x|} for x\not=0 and E(d,B_{1}(0))=\infty , while x=0 is a non-removable singular point. This example indicates that \varepsilon_{0}(2) in Question 1.1 must be sufficiently small. In this note, we will give a partial answer to Question 1.1. More precisely, let B_{R}(0)\subset\mathbb{R}^{2} be the ball in \mathbb{R}^{2} with center 0 and radius R , we first prove Theorem 1.2 . There exists an \varepsilon_{0}>0 such that if u:B_{R}(0)\setminus\{0\}\to N is a smooth harmonic map, satisfying (1.5) |\nabla u(x)|\leq\frac{\varepsilon_{0}}{|x|},\ \ \forall x\in B_{R}(0)\setminus\{0\}, and if, in addition, there exists r_{i}\to 0 such that (1.6) \lim_{i\to\infty}r_{i}\int_{\partial B_{r_{i}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\big)\,d\sigma=0, then u\in C^{\infty}(B_{R}(0),N) . As a direct consequence of Theorem ( 1.2 ), we establish Corollary 1.3 . There exists an \varepsilon_{0}>0 such that if u\in C^{\infty}(\mathbb{R}^{2}\setminus\{0\},N) is a harmonic map, satisfying (1.7) |\nabla u(x)|\leq\frac{\varepsilon_{0}}{|x|},\ \ \forall x\in\mathbb{R}^{2}\setminus\{0\}, and if, in addition, there exists r_{i}\to 0 such that (1.8) \lim_{i\to\infty}r_{i}\int_{\partial B_{r_{i}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\big)\,d\sigma=0, then u must be a constant map. 2. Proofs This section is devoted to the proofs of Theorem 1.2 and Corollary 1.3 . First, we need Lemma 2.1 . If u\in C^{\infty}(B_{R}(0)\setminus\{0\},N) is a harmonic map, then (2.1) \phi(r):=r\int_{\partial B_{r}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\big)\,d\sigma is constant for r\in(0,R) . Proof. Since u\in C^{\infty}(B_{R}(0)\setminus\{0\},N) solves the harmonic map equation ( 1.3 ), for any 0<r_{1}<r_{2}<R , we can multiply ( 1.3 ) by x\cdot\nabla u and integrate the resulting equation over B_{r_{2}}(0)\setminus B_{r_{1}}(0) to obtain \displaystyle 0 \displaystyle=\int_{B_{r_{2}}(0)\setminus B_{r_{1}}(0)}\Delta u\cdot(x\cdot\nabla u)=\int_{B_{r_{2}}(0)\setminus B_{r_{1}}(0)}(u_{j}x_{i}u_{i})_{j}-|\nabla u|^{2}-\frac{1}{2}x_{j}(|\nabla u|^{2})_{j} \displaystyle=\int_{\partial(B_{r_{2}}(0)\setminus B_{r_{1}}(0))}(x\cdot\nabla u)\cdot(\nu\cdot\nabla u)-\frac{1}{2}\int_{\partial(B_{r_{2}}(0)\setminus B_{r_{1}}(0))}|\nabla u|^{2}x\cdot\nu, where \nu denotes the outward unit normal of \partial(B_{r_{2}}(0)\setminus B_{r_{1}}(0)) . This implies that \displaystyle r_{2}\int_{\partial B_{r_{2}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{2}|\nabla u|^{2}\big)\,d\sigma=r_{1}\int_{\partial B_{r_{1}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{2}|\nabla u|^{2}\big)\,d\sigma. Since \displaystyle|\nabla u|^{2}=|\frac{\partial u}{\partial r}|^{2}+\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2} , it follows that (2.2) \displaystyle r_{2}\int_{\partial B_{r_{2}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\big)\,d\sigma=r_{1}\int_{\partial B_{r_{1}}(0)}\big(|\frac{\partial u}{\partial r}|^{2}-\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\big)\,d\sigma. This implies ( 2.1 ). ∎ Remark 2.2 . It is easy to check that if \displaystyle d(x)=\frac{x}{|x|}:\mathbb{R}^{2}\setminus\{0\}\to\mathbb{S}^{1} , then \phi(r)=-2\pi for all r>0 . Proof of Theorem 1.2 : From the condition ( 1.6 ) and ( 2.1 ) of Lemma 2.1 , we have that (2.3) \int_{\partial B_{r}(0)}|\frac{\partial u}{\partial r}|^{2}\,d\sigma=\frac{1}{r^{2}}\int_{\partial B_{r}(0)}|\frac{\partial u}{\partial\theta}|^{2}\,d\sigma holds for all 0<r<R . We will modify the original argument by Sacks-Uhlenbeck [ 3 ] in showing that x=0 is a removable singularity for u . First, we will show that u has finite Dirichlet energy, i.e., u\in H^{1}(B_{R}(0)) . For this, let 0<r_{*}<R_{*}\leq R be two given radius. Set K=\big[\frac{\ln(\frac{R_{*}}{r_{*}})}{\ln 2}\big]\in\mathbb{N} and define the annuals \displaystyle A_{m}=B_{2^{m}r_{*}}(0)\setminus B_{2^{m-1}r_{*}}(0),\ \ 1\leq m\leq K. Denote the radial harmonic function h_{m}(r):=a_{m}+b_{m}\ln r:A_{m}\to\mathbb{R}^{L} , where a_{m} and b_{m}\in\mathbb{R}^{L} are chosen according to the condition h_{m}(2^{m}r_{*})=\fint_{\partial B_{2^{m}r_{*}}}u\,d\sigma,\ \ h_{m}(2^{m-1}r_{*})=\fint_{\partial B_{2^{m-1}r_{*}}}u\,d\sigma, where \displaystyle\fint_{\partial B_{r}(0)}f\,d\sigma=\frac{1}{2\pi r}\int_{\partial B_{r}(0)}f\,d\sigma denotes the average of f over \partial B_{r}(0) . Note that the condition ( 1.5 ) implies {\rm{osc}}_{A_{m}}u\leq C\varepsilon_{0},\ \forall 1\leq m\leq K. Now, by multiplying ( 1.3 ) by u-h_{m} and integrating the resulting equation over A_{m} , we obtain \displaystyle\int_{A_{m}}|\nabla(u-h_{m})|^{2}=\int_{\partial A_{m}}\big(\frac{\partial u}{\partial r}-h_{m}^{\prime}(r)\big)\cdot(u-h_{m})+\int_{A_{m}}A(u)(\nabla u,\nabla u)\cdot(u-h_{m}) \displaystyle=\int_{\partial B_{2^{m}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m})-\int_{\partial B_{2^{m-1}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m}) \displaystyle\quad+\int_{A_{m}}A(u)(\nabla u,\nabla u)\cdot(u-h_{m}) \displaystyle\leq\int_{\partial B_{2^{m}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m})-\int_{\partial B_{2^{m-1}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m})+C\varepsilon_{0}\int_{A_{m}}|\nabla u|^{2}. Since h_{m} depends only on r , we can apply ( 2.3 ) to obtain that \int_{A_{m}}|\nabla(u-h_{m})|^{2}\geq\int_{A_{m}}\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\,d\sigma=\frac{1}{2}\int_{A_{m}}|\nabla u|^{2}. Hence we have (2.4) (\frac{1}{2}-C\varepsilon_{0})\int_{A_{m}}|\nabla u|^{2}\leq\int_{\partial B_{2^{m}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m})-\int_{\partial B_{2^{m-1}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{m}) By summing up ( 2.4 ) over 1\leq m\leq K , we obtain that \displaystyle(\frac{1}{2}-C\varepsilon_{0})\int_{B_{2^{K}r_{*}}(0)\setminus B_{r_{*}}(0)}|\nabla u|^{2} \displaystyle\leq\int_{\partial B_{2^{K}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{K}) (2.5) \displaystyle\quad-\int_{\partial B_{r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{1}). By Poincarè inequality, ( 2.3 ) and ( 1.5 ), the terms in the right hand side of ( 2.5 ) can be estimated by \displaystyle\big|\int_{\partial B_{2^{K}r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{K})\big| \displaystyle\leq C\big(\int_{\partial B_{2^{K}r_{*}}(0)}|\frac{\partial u}{\partial r}|^{2}\,d\sigma\big)^{\frac{1}{2}}\big(\int_{\partial B_{2^{K}r_{*}}(0)}|u-h_{K}|^{2}\,d\sigma\big)^{\frac{1}{2}} \displaystyle\leq C2^{K}r_{*}\Big(\int_{\partial B_{2^{K}r_{*}}(0)}|\frac{\partial u}{\partial r}|^{2}\,d\sigma\Big)^{\frac{1}{2}}\Big(\int_{\partial B_{2^{K}r_{*}}(0)}\frac{1}{r^{2}}|\frac{\partial u}{\partial\theta}|^{2}\,d\sigma\Big)^{\frac{1}{2}} (2.6) \displaystyle\leq C2^{K}r_{*}\int_{\partial B_{2^{K}r_{*}}(0)}|\nabla u|^{2}\,d\sigma\leq C\varepsilon_{0}^{2}, and, similarly, (2.7) \displaystyle\big|\int_{\partial B_{r_{*}}(0)}\frac{\partial u}{\partial r}\cdot(u-h_{1})\big|\leq Cr_{*}\int_{\partial B_{r_{*}}(0)}|\nabla u|^{2}\,d\sigma\leq C\varepsilon_{0}^{2}. Substituting the inequalities ( 2.6 ) and ( 2.7 ) into ( 2.5 ) yields (2.8) (\frac{1}{2}-C\varepsilon_{0})\int_{B_{2^{K}r_{*}}(0)\setminus B_{r_{*}}(0)}|\nabla u|^{2}\leq C\varepsilon_{0}^{2}. Thus, by choosing \varepsilon_{0}<\frac{1}{4C} and observing \frac{R_{*}}{2}\leq 2^{K}r_{*}\leq R_{*} , we obtain that (2.9) \int_{B_{\frac{R_{*}}{2}}(0)\setminus B_{r_{*}}(0)}|\nabla u|^{2}\leq C\varepsilon_{0}^{2}. Since ( 2.9 ) holds for any two 0<r_{*}<R_{*}\leq R , we conclude that (2.10) \int_{B_{\frac{R}{2}}(0)}|\nabla u|^{2}\leq C\varepsilon_{0}^{2}<\infty. Next, with the help of ( 2.10 ), we can repeat the above arguments to obtain the Hölder continuity of u near x=0 . In fact,

Read the complete text