Optimal Regularity for Hölder continuous Hamiltonian Stationary Lagrangian graphs
Arunima Bhattacharya and W. Jacob Ogden · 2025
Read Optimal Regularity for Hölder continuous Hamiltonian Stationary Lagrangian graphs by Arunima Bhattacharya and W. Jacob Ogden (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-2510-22756.txt.
Introduction
Smoothness of special surfaces can fail at a sharp threshold. Arunima Bhattacharya and W Jacob Ogden prove that point in 2025. Their objects are Hamiltonian stationary Lagrangian graphs. Such graphs must stay merely Holder continuous at first. Holder continuity is a weak form of smoothness.
If the Holder exponent beats one third, the graph is smooth. This needs a supercritical phase condition. That condition gives a helpful convexity in the potential. The bound is best possible. At exactly one third, they build explicit singular solutions. Those solutions stay singular even under the strongest phase condition. This departs sharply from special Lagrangian theory.
Catalog record
| CREATOR | Arunima Bhattacharya and W. Jacob Ogden |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-22756
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Optimal Regularity for Hölder continuous Hamiltonian Stationary Lagrangian graphs Arunima Bhattacharya and W. Jacob Ogden Abstract . In this paper, we establish optimal regularity for Hölder continuous Hamiltonian stationary Lagrangian graphs in ℂ n \mathbb{C}^{n} . We prove that such a graph is smooth whenever its Hölder exponent is strictly larger than 1 3 \frac{1}{3} and the Lagrangian phase is supercritical, which yields semi-convexity of the potential. We establish the optimality of our result by constructing explicit singular solutions to the fourth order Hamiltonian stationary equation when the Hölder exponent of the graph is 1 3 \frac{1}{3} . The singular solutions exist even under the strongest convexity assumption on the Lagrangian phase, namely the hypercritical phase, which enforces convexity of the potential. This presents a striking departure from the theory of special Lagrangian graphs. 1. Introduction The theory of special Lagrangian graphs in \mathbb{C}^{n} has been extensively studied with many deep and elegant results. In contrast, the corresponding fourth order generalization, namely Hamiltonian stationary Lagrangian graphs, remains largely unexplored. In this work, we show that fundamental regularity properties of solutions to the special Lagrangian equation break down entirely in the Hamiltonian stationary setting. At the same time, we identify sharp conditions under which comparable regularity results continue to hold. In \mathbb{C}^{n}=\mathbb{R}^{n}\times\mathbb{R}^{n} , with the standard Kähler structure, the Lagrangian phase or angle, \Theta , of the Lagrangian submanifold L_{u}=\{(x,Du)\>|\ x\in\Omega\subseteq\mathbb{R}^{n}\} , can be expressed as \Theta=\sum_{i=1}^{n}\arctan\lambda_{i} where \lambda_{i} ’s are the eigenvalues of the Hessian of u and \Omega is a fixed bounded domain in \mathbb{R}^{n} . The mean curvature vector along L_{u} can be written as \vec{H}=J\nabla_{g}\Theta where \nabla_{g} is the gradient operator of L_{u} (see the work of Harvey-Lawson [ HL82 , (2.19)] ) and g is the induced metric from the Euclidean metric on \mathbb{C}^{n} , which can be written as g=I_{n}+(D^{2}u)^{2}. We say that u solves the Hamiltonian stationary equation when the Lagrangian phase \Theta is g -harmonic (cf. Oh [ Oh93 ] , Schoen-Wolfson [ SW03 , Proposition 2.2] ): (1.1) \Delta_{g}\Theta=0. The gradient graph of u is a Hamiltonian stationary Lagrangian submanifold, which is a critical point of the volume functional under Hamiltonian deformations, that is, variations generated by J\nabla_{g}\eta for compactly supported smooth functions \eta . When \Theta is constant, u solves the special Lagrangian equation (1.2) \Theta=c. The gradient graph of u is a special Lagrangian submanifold, which is a critical point of the volume functional under all compactly supported variations of the submanifold L_{u} . In this paper, we resolve the question of regularity of Hölder continuous Hamiltonian stationary Lagrangian graphs. Before we state our main results, we first clarify the notion of a weak solution to the Hamiltonian stationary equation. Definition 1 . We say that u\in C^{1}(\Omega) is a weak solution of ( 1.1 ) if (1) u\in W^{2,1}(\Omega) , (2) \sqrt{g}\in L^{1}(\Omega) , (3) \Theta\in W^{1,2}(\Omega) , and (4) for all compactly supported smooth functions \eta on the cylinder \Omega\times\mathbb{R}^{n} , u satisfies \displaystyle\int\langle\nabla_{g}\Theta,\nabla_{g}\eta\rangle\>d(\mathcal{H}^{n}\llcorner L_{u})=0. Note that, under the above conditions, \nabla_{g}\eta is well defined a.e. by projecting the Euclidean gradient onto the tangent plane of L_{u} . Our first result is the following. Theorem 1.1 . Let u\in C^{1,\beta} be a weak solution of ( 1.1 ) on the unit ball B_{1}\subset\mathbb{R}^{n} with \beta\in(\frac{1}{3},1) and |\Theta|\geq(n-2)\frac{\pi}{2}+\delta . Then u is smooth in B_{1/2} and satisfies the estimate \|u\|_{C^{k,\alpha}(B_{1/2})}\leq C(\alpha,k,n,\delta,\|u\|_{C^{1,\beta}(B_{1})}) where k\geq 2 . We prove the sharpness of the above Hölder exponent \beta by constructing the following counterexamples, which hold even under the strongest convexity condition on the Lagrangian phase, namely, |\Theta|>(n-1)\frac{\pi}{2} . Theorem 1.2 . On B_{1}\subset\mathbb{R}^{n} (n\geq 3) , for each positive integer k there exists a weak solution u^{(n,k)} to ( 1.1 ) with |\Theta|>(n-1)\frac{\pi}{2} such that u^{(n,k)}\in C^{1,\frac{1}{2k+1}}\setminus C^{1,\frac{1}{2k+1}+\varepsilon} for all \varepsilon>0 . Remark 1.1 . The above result shows that a convex solution u\in C^{1,\frac{1}{3}} to the Hamiltonian stationary equation need not be any more regular. This is a notable departure from the theory of special Lagrangians, for which Chen-Shankar-Yuan [ CSY23 ] established that convex viscosity solutions of the special Lagrangian equation must always be smooth. Remark 1.2 . The singular solutions constructed in this work have gradient graphs that are smooth submanifolds of \mathbb{C}^{n} . It is an interesting question whether there exists a Hamiltonian stationary Lagrangian graph with a hypercritical phase or with a convex potential that has a genuine geometric singularity, i.e., a gradient graph with unbounded curvature. In the case when \beta of Theorem 1.1 is 1, i.e., u\in C^{1,1}(B_{1}) , we obtain a stronger regularity result. Theorem 1.3 . Let u\in C^{1,1} be a weak solution of ( 1.1 ) on the unit ball B_{1}\subset\mathbb{R}^{n} . Then u is smooth on some open set U_{0} containing \left\{\,x\in B_{\frac{1}{2}}\>|\>|\Theta(x)|\geq(n-2)\frac{\pi}{2}\,\right\}, and satisfies the estimate \|u\|_{C^{k,\alpha}(B_{r})}\;\leq\;C\big(\alpha,k,n,r,\|D^{2}u\|_{L^{\infty}(B_{2r})}\big), where k\geq 2 and B_{2r}\subset U_{0} . We emphasize that we do not assume that |\Theta|\geq(n-2)\frac{\pi}{2} holds everywhere in B_{1} . Instead, we show that our result holds on an open set U_{0} that contains \{\,x\in B_{\frac{1}{2}}\>|\>|\Theta(x)|\geq(n-2)\frac{\pi}{2}\,\} . We elaborate on this point further in the discussion in Section 1.3.1 . Remark 1.3 . When n\leq 4 , any C^{1,1} solution of ( 1.1 ) is smooth as shown by Bhattacharya in [ Bha25 ] . When n>4 , let \lambda_{\min} and \lambda_{\max} denote the smallest and largest eigenvalues of the Hessian. If these satisfy \arctan(\lambda_{\min})>(\Theta-\pi)/n or \arctan(\lambda_{\max})<(\Theta+\pi)/n , then the conclusion of Theorem 1.3 follows from using a rigidity result of Ogden-Yuan [ OY25 , Corollary 1.1] and [ Bha25 ] . It is worth noting that in higher codimensions Lawson-Osserman [ LO77 ] constructed examples of Lipschitz minimal submanifolds, including graphical ones, that fail to be C^{1} . Combined with the absence of a rigidity result for special Lagrangian cones in dimensions n>4 , this indicates that imposing restrictions on the phase, or on the minimal or maximal eigenvalues of the Hessian of u , may be essential in establishing higher regularity for Lipschitz continuous Hamiltonian stationary Lagrangian graphs. Remark 1.4 . We would also like to point out that without the condition |\Theta|\geq(n-2)\frac{\pi}{2}+\delta , the result of Theorem 1.1 remains valid provided that one assumes \arctan(\lambda_{\min})>(\Theta-\pi)/n+\delta . This follows from [ OY25 ] and [ Bha25 ] . To highlight the distinction between our results and those established in special Lagrangian geometry, we first briefly describe how the range of the Lagrangian phase, \Theta , affects the concavity properties of the arctangent operator, which plays a key role in regularity. The phase (n-2)\frac{\pi}{2} is referred to as critical: indeed, the level set \{\lambda=(\lambda_{1},..,\lambda_{n})\in\mathbb{R}^{n}\>|\>\Theta=\sum_{i=1}^{n}\arctan\lambda_{i}\} is concave precisely when |\Theta|\geq(n-2)\frac{\pi}{2} [ Yua06 , Lemma 2.2] . When |\Theta|\geq(n-1)\frac{\pi}{2} , the phase is termed hypercritical. In this range, the concavity of the level set is immediate, as the condition enforces \lambda>0 , making the arctangent operator concave. Finally, when |\Theta|\geq(n-2)\frac{\pi}{2}+\delta , the phase is called supercritical. In this range, one can show that the Hessian of the potential u admits a uniform negative lower bound dependent on \delta ; as a consequence, the potential is semi-convex. For the special Lagrangian equation ( 1.2 ) with phase |\Theta|\geq(n-2)\frac{\pi}{2} , Warren-Yuan [ WY09b , WY10 ] and Wang-Yuan [ WY14 ] established Hessian estimates showing