Rectifiability of entropy productions for weak solutions of the 2D eikonal equation with supercritical regularity

Xavier Lamy, Elio Marconi · 2025

Read Rectifiability of entropy productions for weak solutions of the 2D eikonal equation with supercritical regularity by Xavier Lamy, Elio Marconi (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-2509-16692.txt.

Introduction

Xavier Lamy and Elio Marconi study weak solutions of a key equation in a 2025 math paper. The equation forces a field to have unit length and no net outflow. Such fields arise as sharp limits in models of layered materials. A long-standing guess says that a by-product of these fields must sit on the lines where the field jumps.

The pair prove that guess for a wide class of cases. Their proof allows rougher fields than past work could handle. Only one borderline case stays open. The result firms up the math behind models like the Aviles-Giga energy. It also shows how rough a field can be while its by-products stay tame.

Catalog record

CREATORXavier Lamy, Elio Marconi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-16692

Browse the full library · About the Tower of Babel

Full text

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

Rectifiability of entropy productions for weak solutions of the 2D eikonal equation with supercritical regularity Xavier Lamy, Elio Marconi Abstract Weak solutions m : Ω ⊂ ℝ 2 → ℝ 2 m\colon\Omega\subset{\mathbb{R}}^{2}\to{\mathbb{R}}^{2} of the eikonal equation | m | = 1 ​ a.e. and ​ div ​ m = 0 , \displaystyle|m|=1\text{ a.e. and }\mathrm{div}\>m=0\,, arise naturally as sharp interface limits of bounded energy configurations in various physically motivated models, including the Aviles-Giga energy. The distributions μ Φ = div ​ Φ ​ ( m ) \mu_{\Phi}=\mathrm{div}\,\Phi(m) , defined for a class of smooth vector fields Φ \Phi called entropies, carry information about singularities and energy cost. If these entropy productions are Radon measures, a long-standing conjecture predicts that they must be concentrated on the 1-rectifiable jump set of m m – as they do if m m has bounded variation (BV) thanks to the chain rule. We establish this concentration property, for a large class of entropies, under the Besov regularity assumption m ∈ B 1 / p p , ∞ ⇔ sup h ∈ ℝ 2 ∖ { 0 } ∥ m ( ⋅ + h ) − m ∥ L p | h | 1 / p < ∞ , \displaystyle m\in B^{1/p}_{p,\infty}\hskip 10.00002pt\Leftrightarrow\hskip 10.00002pt\sup_{h\in\mathbb{R}^{2}\setminus\{0\}}\frac{\|m(\cdot+h)-m\|_{L^{p}}}{|h|^{1/p}}<\infty\,, for any 1 ≤ p < 3 1\leq p<3 , thus going well beyond the BV setting ( p = 1 p=1 ) and leaving only the borderline case p = 3 p=3 open. 1 Introduction For an open set \Omega\subset{\mathbb{R}}^{2} , we consider weak solutions m\colon\Omega\to{\mathbb{R}}^{2} of the eikonal equation \displaystyle|m|=1\text{ a.e. in }\Omega,\hskip 10.00002pt\dv m=0\text{ in }\mathcal{D}^{\prime}(\Omega). (1.1) If \Omega is simply connected, this is equivalent to the existence of a Lipschitz function u\colon\Omega\to{\mathbb{R}} satisfying m=i\nabla u and \displaystyle|\nabla u|=1\text{ a.e. in }\Omega\,, which is classically referred to as the eikonal equation. If m\colon\Omega\to{\mathbb{R}}^{2} is a C^{1} solution of the eikonal equation ( 1.1 ) then the chain rule provides a whole family of conservation laws: \dv\Phi(m)=0 for all \Phi\in C^{1}(\mathbb{S}^{1};{\mathbb{R}}^{2}) such that \partial_{\theta}\Phi(e^{i\theta})\cdot e^{i\theta}=0 for all \theta\in{\mathbb{R}} . For a general weak solution m\colon\Omega\to{\mathbb{R}}^{2} of the eikonal equation ( 1.1 ), the distributions \dv\Phi(m) may not be zero, and they carry information on how singular m is. They were first introduced in the context of the eikonal equation in [ 7 ] , and called entropy productions by analogy with the theory of scalar conservation laws. We denote by \displaystyle\mathrm{ENT}=\big\{\Phi\in C^{1,1}(\mathbb{S}^{1};{\mathbb{R}}^{2})\colon\frac{d}{d\theta}\Phi(e^{i\theta})\cdot e^{i\theta}=0\;\forall\theta\in{\mathbb{R}}\big\}\,, (1.2) the set of all C^{1,1} entropies. Weak solutions m of the eikonal equation ( 1.1 ) whose entropy productions \dv\Phi(m) are finite Radon measures play an important role in the theory of the Aviles-Giga energy [ 3 ] . The structure of these finite-entropy solutions is not fully understood, but it is known that they share with functions of bounded variation (BV) several fine properties. Note that, if m\in BV(\Omega;{\mathbb{R}}^{2}) is a weak solution of ( 1.1 ), then by the BV chain rule its entropy productions are measures concentrated on the \mathcal{H}^{1} -rectifiable jump set J_{m} . For a general finite-entropy solution, denote by \nu the supremum measure \displaystyle\nu=\bigvee_{\Phi\in\mathrm{ENT},\,\|\Phi\|_{C^{1,1}}\leq 1}|\dv\Phi(m)|\,. In [ 6 ] the authors prove that the jump set J_{m}\doteq\left\{x\in\Omega:\limsup_{r\to 0}\frac{\nu(B_{r}(x))}{r}>0\right\}\,, (1.3) is \mathcal{H}^{1} -rectifiable and m admits left and right L^{1} traces \mathcal{H}^{1} -a.e. along J_{m} . According to a long-standing conjecture on the Aviles-Giga energy [ 3 ] , entropy productions should be concentrated on that jump set. Among all weak solutions of ( 1.1 ), the finite-entropy solutions can be characterized, at least locally, in terms of Besov B^{1/3}_{3,\infty} regularity [ 8 ] . For s\in(0,1) and p\geq 1 , a map m\in L^{p}(\Omega) has the Besov regularity B^{s}_{p,\infty} if and only if the seminorm \displaystyle|m|_{B^{s}_{p,\infty}}=\sup_{h\in{\mathbb{R}}^{2}\setminus\{0\}}\frac{1}{|h|^{s}}\|m(\cdot+h)-m\|_{L^{p}(\Omega\cap(\Omega-h))}\,, is finite [ 14 , §2.5.12] . Between the spaces BV(\Omega;\mathbb{S}^{1}) and B^{1/3}_{3,\infty}(\Omega;\mathbb{S}^{1}) lies the intermediate scale of spaces B^{1/p}_{p,\infty}(\Omega;\mathbb{S}^{1}) , for 1<p<3 . We prove that the concentration conjecture is true for solutions of ( 1.1 ) with that intermediate regularity, and for entropies in the class \displaystyle\widetilde{\mathrm{ENT}}\doteq\left\{\Phi\in\mathrm{ENT}:\frac{d}{d\theta}\Phi(-e^{i\theta})=-\frac{d}{d\theta}\Phi(e^{i\theta})\;\forall\theta\in{\mathbb{R}}\right\}, (1.4) which corresponds to odd entropies plus constants. This restriction is due to the same technical reasons as in [ 12 ] (where this class of entropies is denoted by \mathcal{E}_{\pi} ). The fundamental entropies introduced in [ 10 ] to establish a sharp lower bound for the Aviles-Giga energy (see also [ 1 ] ) are odd, and therefore covered by our result. Theorem 1.1 . Let \Omega\subset{\mathbb{R}}^{2} a bounded open set, m\colon\Omega\to\mathbb{R}^{2} a weak solution of the eikonal equation ( 1.1 ), and assume that m\in B^{1/p}_{p,\infty}(\Omega) for some p\in[1,3) . Then the entropy productions of m are 1-rectifiable, \displaystyle\dv\Phi(m)=\mathbf{n}\cdot\big(\Phi(m^{+})-\Phi(m^{-})\big)\,\mathcal{H}^{1}\llcorner J_{m}\,, for all \Phi\in\widetilde{\mathrm{ENT}} . To describe the ideas behind Theorem 1.1 , let us consider first a solution m\in{\mathrm{BV}}(\Omega;\mathbb{S}^{1}) of ( 1.1 ). Then, by the BV chain rule we have |\dv\Phi(m)|\leq C_{\Phi}|m^{+}-m^{-}|^{3}d\mathcal{H}^{1}\llcorner J_{m}. for every entropy \Phi\in\mathrm{ENT} . Moreover, for any p\in[1,3) we have \int_{J_{m}}|m^{+}-m^{-}|^{p}d\mathcal{H}^{1}\leq\|m\|^{p}_{B^{1/p}_{p,\infty}}\,. As a consequence, the contribution of jumps smaller than a threshold \delta>0 is controlled by \displaystyle|\dv\Phi(m)|(\Omega\setminus(J_{m}\cap\{|m^{+}-m^{-}|\geq\delta\}) \displaystyle\leq C_{\Phi}\int_{J_{m}\cap\{|m^{+}-m^{-}|\leq\delta\}}|m^{+}-m^{-}|^{3}d\mathcal{H}^{1}\leq C_{\Phi}\delta^{3-p}\|m\|^{p}_{B^{1/p}_{p,\infty}}. (1.5) Here we assumed that m\in BV(\Omega;\mathbb{S}^{1}) to ensure that \dv\Phi(m) is concentrated on J_{m} , but the estimate does not depend on the total variation of m . Moreover, a structure result proved in [ 12 ] about the continuous part of the entropy production allows to interpret it as being generated by infinitesimally small jumps. It is therefore natural to conjecture that the estimate ( 1 ) should be true for solutions m not necessarily of bounded variation. We prove indeed a similar estimate in Proposition 4.1 , and the main result then follows by letting \delta\to 0 . To technically implement these ideas, we actually have to argue along trajectories of a Lagrangian representation of m , also introduced in [ 12 ] , and which can only provide information on entropies in ( 1.4 ). The control on the continuous part of entropy production at the level of these Lagrangian trajectories is obtained by using a singular family of entropies, whose entropy productions are uniformly bounded thanks to the supercritical Besov regularity assumption. 2 Entropy productions and Besov regularity In this section we prove that the supercritical Besov regularity assumption provides uniform control on entropy productions over families of entropies which are unbounded in C^{1,1} . Proposition 2.1 . Let \Omega\subset{\mathbb{R}}^{2} a bounded open set, m\colon\Omega\to\mathbb{S}^{1} with \dv m=0 and assume that m\in B^{1/(2+\alpha)}_{2+\alpha,\infty}(\Omega) for some \alpha\in(0,1) . Then we have \displaystyle\bigg(\bigvee_{\|\Phi\|_{C^{1,\alpha}}\leq 1}|\dv\Phi(m)|\bigg)(\Omega)\leq C\|m\|^{2+\alpha}_{B^{1/(2+\alpha)}_{2+\alpha,\infty}(\Omega)}\,, for some C=C(\alpha)>0 , where the supremum of measure is taken over all \Phi\in\mathrm{ENT} such that \|\Phi\|_{C^{1,\alpha}}\leq 1 . Proposition 2.1 can be interpreted as an interpolation between the estimates \displaystyle\bigg(\bigvee_{\|\Phi\|_{C^{1,1}}\leq 1}|\dv\Phi(m)|\bigg)(\Omega) \displaystyle\lesssim\|m\|^{3}_{B^{1/3}_{3,\infty}(\Omega)}\,, \displaystyle\bigg(\bigvee_{\|\Phi\|_{C^{0,1}}\leq 1}|\dv\Phi(m)|\bigg)(\Omega) \displaystyle\lesssim\|m\|^{2}_{B^{1/2}_{2,\infty}(\Omega)}\,. The first of these estimates is proved in [ 8 , Proposition 3.10] , and the second can be established using similar calculations which rely on commutator estimates for the function \displaystyle w_{\epsilon}=1-|m_{\epsilon}|^{2}=(|m|^{2})_{\epsilon}-|m_{\epsilon}|^{2}\,, where the subscript \epsilon denotes convolution at scale \epsilon . (In the context of the eikonal equation, arguments based on commutator estimates were introduced in [ 5 ] .) The interpolation argument is however a bit involved. In particular, the constant C=C(\alpha) we are able to obtain in Proposition 2.1 blows up as \alpha\to 0 or 1 , even though

Read the complete text