Extension theorems for logarithmic Schrödinger and discrete Laplacian operators
J. J. Betancor, M. de León-Contreras, L. Rodríguez-Mesa · 2026
Read Extension theorems for logarithmic Schrödinger and discrete Laplacian operators by J. J. Betancor, M. de León-Contreras, L. Rodríguez-Mesa (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-2604-03638.txt.
Introduction
Jorge Betancor, Marta de León-Contreras, and Lourdes Rodríguez-Mesa defined logarithmic versions of two operators. One belongs to Schrödinger operators with a suitable potential; the other is the discrete Laplacian on integers. The work covers both the smooth setting and the discrete setting in one account. Both versions act nonlocally, so the authors built them through extension problems like those used for fractional Laplacians. Here the logarithm shows up as a boundary value, reached by a more involved route.
Catalog record
| CREATOR | J. J. Betancor, M. de León-Contreras, L. Rodríguez-Mesa |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65278 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.03638
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Extension theorems for logarithmic Schrödinger and discrete Laplacian operators Thanks: The authors are partially supported by the Grant PID2023-148028NB-I00 funded by MICIU/AEI/10.13039/501100011033 and by ERDF/EU”. J. J. Betancor, M. de León-Contreras, L. Rodríguez-Mesa Abstract . In this paper we consider logarithmic operators in two different contexts: the adapted to (continuous) Schrödinger operators and the classical discrete setting. The Schrödinger operator ℒ V \mathcal{L}_{V} on R d \mathbb R^{d} is defined as ℒ V = − Δ + V \mathcal{L}_{V}=-\Delta+V , where the potential V V is nonnegative and satisfies a reverse Hölder inequality and, as usual, Δ \Delta denotes the Euclidean Laplacian, while the discrete Laplacian Δ d \Delta_{d} on Z \mathbb Z is given by ( Δ d f ) ( n ) = f ( n + 1 ) − 2 f ( n ) + f ( n − 1 ) (\Delta_{d}f)(n)=f(n+1)-2f(n)+f(n-1) , n ∈ Z n\in\mathbb Z . Both logarithmic operators log ℒ V \log\mathcal{L}_{V} and log ( − Δ d ) \log(-\Delta_{d}) are nonlocal operators and we will define them through suitable extension problems. The extension problems for logarithmic operators are inspired by the one introduced by Caffarelli and Silvestre for the fractional Laplacian but, in this case, the logarithmic operators are obtained as the boundary values of the extension in a more involved way. † † footnotetext: Last modification: August 10, 2026. 2010 Mathematics Subject Classification 35J10, 42B37, 47D06 1. Introduction In this paper we show that logarithmic operators associated with Schrödinger operators on \mathbb{R}^{d} and with the discrete Laplacian on \mathbb{Z} can be obtained through suitable extension problems. These logarithmic operators are of non-local nature. The extension problem associated with the logarithmic Laplacian on \mathbb R^{d} has been recently introduced in [ 13 ] . Although this extension problem was inspired by the one developed by Caffarelli and Silvestre (see [ 10 ] ) for the fractional Laplacian, the logarithmic operator appears as the boundary values of the solution to the extension problem in a more involved way. The celebrated Caffarelli-Silvestre extension theorem can be stated as follows. Theorem A . Let \sigma\in(0,1) . We consider the Dirichlet problem \left\{\begin{array}[]{cc}\nabla\cdot(t^{1-2\sigma}\nabla u(x,t))=0,&(x,t)\in\mathbb{R}^{d}\times(0,\infty)\\ u(0,x)=f(x),&x\in\mathbb{R}^{d},\end{array}\right. where f\in\mathcal{S}(\mathbb{R}^{d}) , the space of Schwartz functions, and \nabla=(\partial_{x_{1}},\dots,\partial_{x_{n}},\partial_{t}) , and the associated Dirichlet to Neumann map, given by f\in\mathcal{S}(\mathbb{R}^{d})\mapsto-\lim_{t\to 0^{+}}t^{1-2\sigma}\partial_{t}u(\cdot,t). Then, the last mapping coincides with \frac{\Gamma(1-\sigma)}{2^{2\sigma-1}}(-\Delta)^{\sigma} , where (-\Delta)^{\sigma} is the \sigma- fractional Laplacian. The Caffarelli-Silvestre extension has opened numerous and new lines of research, one of which focuses on identifying operators that admit a representation through an extension problem. In this direction, Kwaśnicki and Mucha (see [ 33 ] ) proved that if \phi is a complete Bernstein function, the operator \phi(-\Delta) can be obtained from an extension problem. The results in [ 33 ] were extended to other differential operators in [ 3 ] by using Îto calculus. Extension problems have been studied in more abstract settings by Stinga and Torrea (see [ 39 ] ), and Galé, Miana and Stinga (see [ 30 ] ). Moreover, the fractional Laplace Beltrami operator on some noncompact manifolds has been defined through an extension problem in [ 4 ] and [ 8 ] . Arendt, ter Elst and Warma ( [ 2 ] ) considered the problem for sectorial operators in Hilbert spaces. The logarithmic Laplacian operator, \log(-\Delta), was studied in [ 16 ] , where a pointwise representation was obtained (see [ 16 , Theorem 1.1] ). Chen and Véron ( [ 15 ] ) analyzed the Cauchy problem associated with \log(-\Delta), and proved the existence of the fundamental solution of the logarithmic Laplacian in dimension d\geq 3 . Recently, Lee ( [ 35 ] ) extended the result about the fundamental solution of \log(-\Delta) by using an approach via the division problem. On the other hand, spectral properties for the logarithmic Laplacian have been studied in [ 14 ] , [ 16 ] , and [ 32 ] while the m- order logarithmic Laplacian was studied in [ 12 ] . Logarithmic operators have also been defined in several other settings. Logarithmic Bessel operators were analyzed in [ 29 ] , while Fall and Felli (see [ 25 ] and [ 26 ] ) proved unique continuation properties and essential self-adjointness of the relativistic Schrödinger operator with a singular homogeneous potential defined by H=(-\Delta+m^{2})^{s}-\frac{a(x/|x|)}{|x|^{2s}}-h(x),\quad m\geq 0, where the pointwise representation for (-\Delta+m^{2})^{s} , 0<s<1 , can be found in [ 26 , (1.3)] . On the other hand, Feulefack (see [ 27 , 28 ] ) introduced and studied the logarithm for the operator I+(-\Delta)^{s} , s\in(0,1] . Until very recently there were no results of the logarithmic Laplacian on manifolds. In [ 37 ] , the logarithmic operator \log(-\Delta+mI) , for m>1 , was defined on closed manifolds, while Chen ( [ 17 ] ) defined \log(-\Delta) on general Riemannian manifolds, and Chen and Xu (see [ 18 ] ) on general graphs. The logarithmic Schrödinger operator, \log(-\Delta+V) , where V is a nonnegative potential satisfying a reverse Hölder inequality, RH_{q} , with q>d/2 , has been defined recently by Betancor, Dalmasso, Fariña and Quijano in [ 7 ] . Here we summarize the main points in order to state our results, see Section 2 for more details. The Schrödinger operator, \mathcal{L}_{V}=-\Delta+V , can be defined through the following sesquilinear form T(f,g)=\int_{\mathbb{R}^{d}}\nabla f\;\overline{\nabla g}\;dx+\int_{\mathbb{R}^{d}}V\;f\;\overline{g}\;dx, for f,g\in D_{T}:=\{h\in L^{2}(\mathbb{R}^{d}):\>\nabla h\in L^{2}(\mathbb{R}^{d})\text{ and }\sqrt{V}h\in L^{2}(\mathbb{R}^{d})\}. The sesquilinear form T is closed and nonnegative and the domain D_{T} is dense in L^{2}(\mathbb{R}^{d}) . Therefore, there exists a selfadjoint operator \mathcal{L}_{V}:D(\mathcal{L}_{V})=D_{T}\subset L^{2}(\mathbb{R}^{d})\to L^{2}(\mathbb{R}^{d}) such that \langle\mathcal{L}_{V}f,g\rangle=T(f,g) , f,g\in D(\mathcal{L}_{V}) (see [ 38 , Theorem VIII.15] ). The space C^{\infty}_{c}(\mathbb{R}^{n}) of smooth functions with compact support in \mathbb{R}^{d} is contained in D(\mathcal{L}_{V}) and \mathcal{L}_{V}f=-\Delta f+Vf , for every f\in C^{\infty}_{c}(\mathbb{R}^{n}) . Hence, there exists a spectral measure E_{V} supported in the spectrum \sigma(\mathcal{L}_{V}) of \mathcal{L}_{V} such that \mathcal{L}_{V}f=\int_{[0,\infty)}\lambda\;dE_{V}(\lambda)f,\quad f\in D(\mathcal{L}_{V}), and {D(\mathcal{L}_{V})}=\left\{f\in L^{2}(\mathbb{R}^{d}):\>\int_{0}^{\infty}\lambda^{2}\;d\mu_{f,f}^{V}(\lambda)<\infty\right\}. Here, for every f,g\in L^{2}(\mathbb{R}^{d}), \mu_{f,g}^{V}(U)=\langle E_{V}(U)f,g\rangle , for every Borel subset U of \mathbb{R}. Notice that E_{V}(\{0\})=0 , because 0 is not an eigenvalue of \mathcal{L}_{V} . Thus, the logarithmic \log(\mathcal{L}_{V}) of \mathcal{L}_{V} is defined by \log(\mathcal{L}_{V})f=\int_{0}^{\infty}\log\lambda\;dE_{V}(\lambda)f, for every f\in D(\log(\mathcal{L}_{V}))=\left\{f\in L^{2}(\mathbb{R}^{d}):\>\int_{0}^{\infty}{|\log\lambda|^{2}}\;d\mu_{f,f}^{V}(\lambda)<\infty\right\}. In [ 7 , Theorem 1.1 (a)] , it was established that if f\in D(\mathcal{L}_{V}^{s_{0}})\cap D(\log(\mathcal{L}_{V})) , for some s_{0}\in(0,1], then \log(\mathcal{L}_{V})f=\lim_{s\to 0^{+}}\frac{\mathcal{L}_{V}^{s}f-f}{s}, where \mathcal{L}_{V}^{s}f=\int_{0}^{\infty}\lambda^{s}\;dE_{V}(\lambda)f, f\in D(\mathcal{L}_{V}^{s}):=\left\{f\in L^{2}(\mathbb{R}^{d}):\>\int_{0}^{\infty}\lambda^{2s}\;d\mu_{f,f}^{V}(\lambda)<\infty\right\}, s\in(0,1]. Thus, for every t>0 we can define T_{t}^{V}f=\int_{0}^{\infty}e^{-\lambda t}dE_{V}(\lambda)f , f\in L^{2}(\mathbb{R}^{d}) , so that the family \{T_{t}^{V}\}_{t>0} is the C_{0} -semigroup in L^{2}(\mathbb{R}^{d}) generated by \mathcal{L}_{V}. Then, for every t>0 , there exists a measurable function T_{t}^{V}:\mathbb{R}^{d}\times\mathbb{R}^{d}\to\mathbb{R} such that (1.1) T_{t}^{V}(f)(x)=\int_{\mathbb{R}^{d}}T_{t}^{V}(x,y)f(y)dy,\quad f\in L^{2}(\mathbb{R}^{d}),\quad x\in\mathbb{R}^{d}. Furthermore, by using the integral representation ( 1.1 ), each T_{t}^{V} can be extended as a contraction in L^{p}(\mathbb{R}^{d}) , so that \{T_{t}^{V}\}_{t>0} is a semigroup of contractions in L^{p}(\mathbb{R}^{d}) , for 1\leq p\leq\infty . However, \{T_{t}^{V}\}_{t>0} is not Markovian, that is, T_{t}^{V}1\neq 1 , t>0 . This fact leads to essential differences between the proofs of the results concerning the Schrödinger setting and those for the Laplacian. The following pointwise representation of \log(\mathcal{L}_{V}) was established in [ 7 , Theorem 1.2] . Theorem B . Let d\geq 3 and q>d/2 .